(Topics in Geobiology 47) Carrie L. Tyler, Chris L. Schneider - Marine Conservation Paleobiology-Springer International Publishing (2018)

Topics in Geobiology 47
Carrie L. Tyler
Chris L. Schneider
Editors
Marine
Conservation
Paleobiology
Topics in Geobiology
Volume 47
Series Editors
Neil H. Landman
American Museum of Natural History, New York, NY, USA
Peter J. Harries
North Carolina State University, Raleigh, NC, USA
The Topics in Geobiology series covers the broad discipline of geobiology that
is devoted to documenting life history of the Earth. A critical theme inherent in
addressing this issue and one that is at the heart of the series is the interplay between
the history of life and the changing environment. The series aims for high quality,
scholarly volumes of original research as well as broad reviews.
Geobiology remains a vibrant as well as a rapidly advancing and dynamic field.
Given this field’s multidiscipline nature, it treats a broad spectrum of geologic,
biologic, and geochemical themes all focused on documenting and understanding
the fossil record and what it reveals about the evolutionary history of life. The Topics
in Geobiology series was initiated to delve into how these numerous facets have
influenced and controlled life on Earth.
Recent volumes have showcased specific taxonomic groups, major themes in the
discipline, as well as approaches to improving our understanding of how life has
evolved.
Taxonomic volumes focus on the biology and paleobiology of organisms – their
ecology and mode of life – and, in addition, the fossil record – their phylogeny and
evolutionary patterns – as well as their distribution in time and space.
Theme-based volumes, such as predator-prey relationships, biomineralization,
paleobiogeography, and approaches to high-resolution stratigraphy, cover specific
topics and how important elements are manifested in a wide range of organisms and
how those dynamics have changed through the evolutionary history of life.
Comments or suggestions for future volumes are welcomed.
More information about this series at http://www.springer.com/series/6623

Carrie L. Tyler • Chris L. Schneider
Editors
Marine Conservation
Paleobiology
123
Editors
Carrie L. Tyler Chris L. Schneider
Department of Geology and Environmental Department of Earth
Earth Science and Atmospheric Sciences
Miami University University of Alberta, Edmonton
Oxford, OH, USA Alberta, Canada
ISSN 0275-0120
Topics in Geobiology
ISBN 978-3-319-73793-5 ISBN 978-3-319-73795-9 (eBook)
https://doi.org/10.1007/978-3-319-73795-9
Library of Congress Control Number: 2018935181
© Springer International Publishing AG, part of Springer Nature 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, express or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by the registered company Springer International Publishing AG part
of Springer Nature.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
We believe that a book of this nature is timely and important in highlighting the
contributions of conservation paleobiology in solving modern ecological crises. The
consequences of millennia of human impacts on ecosystems including extinction,
biodiversity loss, and changes in community structure and species composition
remain largely unknown. These trends may yet be reversible, but developing
effective strategies for conservation, remediation, restoration, and policy calls for a
thorough understanding of long-term ecological processes and the ability to examine
ecosystems before, during, and after perturbation. There is thus an urgent need
for a cross-disciplinary synthesis of modern and historical perspectives. Although
paleontological and modern data have traditionally been viewed as disparate data
types, fossils can yield high resolution data suitable for community analysis,
comparable to modern assemblage data.
We are pleased to present research from some of the foremost and upcoming
conservation paleobiologists. Contributions range from case studies with direct
application to reviews and meta-analyses, providing broader implications to con-
servation efforts, and to training of future generations. In this volume, we aim to:
1. Define the goals of conservation paleobiology and our role in conservation
science.
2. Highlight how conservation paleoecology can be used to identify and understand
ecosystem crises.
3. Provide case studies demonstrating applications to modern communities.
4. Stimulate novel applications of paleontological approach to neontological data.
5. Encourage cross-disciplinary dialogue and application of research to manage-
ment and conservation.
6. Explore the future of conservation paleoecology.
We have the capability, technology, and innovation to enact positive change
in the biosphere. There are a multitude of efforts undertaken by societies and
governments to mitigate the effects of global climate change. Marine protected areas
and reserves, for example, provide safe havens for biodiversity and contribute to
v
vi Preface
sustainable populations for harvested resources. However, conservation biology is

a crisis-driven discipline, with little time and often insufficient financial resources
to respond. Furthermore, conservation science studies systems already undergoing
change. In almost all cases, the baseline condition of the ideal ecosystem is
gone, and in many of those cases, given the trajectory of climate change and
related ocean deterioration, the baseline may no longer be achievable. Conservation
paleobiology can provide timely and critical insights to conservation, management,
and restoration efforts, providing data on timescales otherwise unavailable and of
true pre-impact conditions.
We must collaborate if we are going to maintain a productive marine biosphere
for future humanity, and we challenge conservation paleobiology as a field to
truly be interdisciplinary not just on paper, but in action. We continue to work
towards a better future, to which conservation paleobiology can make significant
contributions.
Oxford, OH, USA Carrie L. Tyler

Edmonton, AB, Canada Chris L. Schneider
Acknowledgements
We are deeply grateful to all of the reviewers who devoted their time, expertise,
and experience to improving each chapter, despite their busy schedules. We wish to
thank a few of the advocates of this new frontier of research, particularly Gregory
Dietl, Karl Flessa, Jeremy Jackson, Patricia Kelley, and Sue Kidwell. To all of the
authors, thank you for your contributions; we are thrilled to share such high-quality
work with the scientific community. We are particularly grateful to Greg Dietl for
thoughtful discussions, which influenced the shaping and content of this volume.
We also thank Lindsey Leighton for his advice on editing a special volume such as
this one. Thank you also to the editorial staff at Springer, without whom this volume
could not have been published.
vii
Contents
An Overview of Conservation Paleobiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Carrie L. Tyler and Chris L. Schneider
Should Conservation Paleobiologists Save the World on Their Own
Time? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Gregory P. Dietl and Karl W. Flessa
Conceptions of Long-Term Data Among Marine Conservation
Biologists and What Conservation Paleobiologists Need to Know. . . . . . . . . . 23
Jansen A. Smith, Stephen R. Durham, and Gregory P. Dietl
Effectively Connecting Conservation Paleobiological Research
to Environmental Management: Examples from Greater
Everglades’ Restoration of Southwest Florida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Michael Savarese
Using the Fossil Record to Establish a Baseline
and Recommendations for Oyster Mitigation in the Mid-Atlantic U.S.. . . . 75
Kristopher M. Kusnerik, Rowan Lockwood, and Amanda N. Grant
Coral Reefs in Crisis: The Reliability of Deep-Time Food Web
Reconstructions as Analogs for the Present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Peter D. Roopnarine and Ashley A. Dineen
Exploring the Species –Area Relationship Within a Paleontological
Context, and the Implications for Modern Conservation Biology . . . . . . . . . . 143
Matthew J. Pruden and Lindsey R. Leighton
Marine Refugia Past, Present, and Future: Lessons from Ancient
Geologic Crises for Modern Marine Ecosystem Conservation . . . . . . . . . . . . . . 163
Chris L. Schneider
ix
x Contents
Training Tomorrow’s Conservation Paleobiologists . . . . . . . . . . . . . . . . . . . . . . . . . 209

Patricia H. Kelley, Gregory P. Dietl, and Christy C. Visaggi
A Conceptual Map of Conservation Paleobiology: Visualizing
a Discipline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Carrie L. Tyler
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Contributors
Gregory P. Dietl Paleontological Research Institution, Ithaca, NY, USA

Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY,
USA
Ashley A. Dineen Invertebrate Zoology and Geology, Institute for Biodiversity
Science and Sustainability, California Academy of Sciences, San Francisco, CA,
USA
Stephen R. Durham Department of Earth and Atmospheric Sciences, Cornell
University, Ithaca, NY, USA
Paleontological Research Institution, Ithaca, NY, USA
Karl W. Flessa Department of Geosciences, University of Arizona, Tucson, AZ,
USA
Amanda N. Grant School of Earth Sciences and Environmental Sustainability,
Northern Arizona University, Flagstaff, AZ, USA
Patricia H. Kelley Department of Earth and Ocean Sciences, University of North
Carolina Wilmington, Wilmington, NC, USA
Kristopher M. Kusnerik Division of Invertebrate Paleontology, Florida Museum
of Natural History, Gainesville, FL, USA
Lindsey R. Leighton Department of Earth and Atmospheric Sciences, University
of Alberta, Edmonton, Alberta, Canada
Rowan Lockwood Department of Geology, The College of William and Mary,
Williamsburg, VA, USA
Matthew J. Pruden Department of Earth and Atmospheric Sciences, University
of Alberta, Edmonton, Alberta, Canada
xi
xii Contributors
Peter D. Roopnarine Invertebrate Zoology and Geology, Institute for Biodiversity

Science and Sustainability, California Academy of Sciences, San Francisco, CA,
USA
Michael Savarese Marine and Ecological Sciences, Florida Gulf Coast University,
Ft. Myers, FL, USA
Chris L. Schneider Department of Earth and Atmospheric Sciences, University of
Alberta, Edmonton, Alberta, Canada
Jansen A. Smith Department of Earth and Atmospheric Sciences, Cornell Univer-
sity, Ithaca, NY, USA
Carrie L. Tyler Department of Geology and Environmental Earth Science, Miami
University, Oxford, OH, USA
Christy C. Visaggi Department of Geosciences, Georgia State University, Atlanta,
GA, USA
An Overview of Conservation
Paleobiology
Carrie L. Tyler and Chris L. Schneider
Abstract The field of conservation paleobiology was formally established in

the early 2000s, as a growing body of literature substantiating the fidelity of
paleontological data on a variety of spatial and temporal scales emerged, and
paleontologists became increasingly aware of the potential insights that the fossil
record could provide into the current biodiversity crisis. Conservation paleobiology
contributes a temporal scope and historical perspective lacking from the relatively
short time spans covered by modern ecological studies, progressively in demand in
the face of changing climate and environmental degradation. The increasing number
of conservation paleobiology studies in the past decade validates the potential
contributions of this field to conservation efforts, and fall within a range of temporal
categories (e.g., “near-time” and “deep-time”). Data are not restricted to fossils,
but can also include historical reports and archeological evidence (conservation
archeobiology). Although temporal resolution often declines with increased assem-
blage age, ancient ecosystems document responses to, and recoveries from, global
crises. Thus, the field of conservation paleobiology, when considered in concert with
historical ecology and conservation biology, has the potential to positively affect
future ecosystems and biodiversity.
Keywords Conservation science · Conservation biology · Geohistorical data ·

Environmental history · Conservation archeobiology
C. L. Tyler ()
Department of Geology and Environmental Earth Science, Miami University, Oxford, OH, USA
e-mail: [email protected]
C. L. Schneider
Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta,
Canada
© Springer International Publishing AG, part of Springer Nature 2018 1

C. L. Tyler, C. L. Schneider (eds.), Marine Conservation Paleobiology,
Topics in Geobiology 47, https://doi.org/10.1007/978-3-319-73795-9_1
2 C. L. Tyler and C. L. Schneider
1 Defining and Establishing Conservation Paleobiology

as a Discipline
The importance of records documenting previous ecosystem states is increasingly

recognized in conservation and ecosystem management (e.g., Willis and Birks
2006; Knowlton and Jackson 2008; Louys et al. 2012), particularly in identifying
restoration targets for degraded or ailing ecosystems (Jackson and Hobbs 2009).
Ideally, baseline conditions should be established prior to human disturbance;
however, human activities have been altering the Earth since the Late Pleistocene
(Jackson et al. 2001; Boivin et al. 2016), increasing in scope and intensity with
the advent of agriculture 10,000 years ago, and later by the rise of large-scale
industrialization. Marine ecosystems have arguably been altered far more than those
on land (Jackson et al. 2001; Knowlton and Jackson 2008; Ban and Alder 2008),
and as a result, “pristine” ocean habitats have effectively vanished and may not
be recoverable. Furthermore, global and local environmental changes continue to
occur, with shifting baselines creating a moving target for conservation scientists
and resource management (Pauly 1995; Jackson et al. 2001). Yet few modern
ecological studies consider the historically unaltered state of ecosystems (Jackson
et al. 2001; Willis and Birks 2006; Froyd and Willis 2008), thereby limiting our
understanding of marine ecosystem functioning predating anthropogenic effects.
In modern ecosystems, long-term ecological studies are sparse and predomi-
nantly focused on terrestrial botanical systems (although this is improving with the
rise of programs such as the Long-Term Ecological Research Network). Generally,
ecosystem monitoring rarely extends beyond the past few decades (Jackson and
Hobbs 2009), and possibly the longest-running example of ecological monitoring,
the Park Grass experiment in Rothermstead, southern England, has been running
since 1856 (Silvertown et al. 2006). Designed to address agricultural questions, this
environmental monitoring experiment has continued for over 160 years. However,
many ecological theories require data on far longer timescales to test (Louys 2012),
e.g., neutral theory (Hubbell 2001). While broadly applicable and valuable insights
into ecological processes can be gained through modern observations, long-term
data from other types of habitats and ecosystems are necessary to determine the
generality of these findings. Geohistorical and paleontological data can, therefore,
provide information essential to conservation science, particularly given the benefits
of understanding long-term ecological processes, and that true “pristine” habitats
likely no longer exist.
As a nascent discipline, conservation paleobiology is currently ideally poised for
deliberate and thoughtful definition, and determination of agreed upon goals through
rigorous academic debate (Tyler 2018). Conservation paleobiology provides a deep-
time perspective for a biosphere in flux and has been defined as the “application of
the methods and theories of paleontology to the conservation and restoration of
biodiversity and ecosystem services” (Dietl and Flessa 2011; Dietl et al. 2015). As
this definition includes some of the goals of conservation science—the conservation
and restoration of biodiversity and maintenance of ecosystem services (Soule 1985;
An Overview of Conservation Paleobiology 3
Meine et al. 2006)—Dietl (2016) suggested that it may be beneficial to approach

conservation paleobiology as a sub-discipline within conservation science. This
approach would likely promote the dissemination of paleontological data, ideas,
and information in conservation science (Dietl 2016). As conservation scientists
and resource managers remain relatively unaware of potential geohistorical contri-
butions (Smith et al. 2018; Savarese 2018), this paradigm shift would be potentially
transformative to the field. It should be noted that not all perceive conservation
paleobiology as a value-driven science (see Dietl and Flessa 2018).
The rapid rise of conservation paleobiology is due in part to crucial work
validating the ability of paleontology to contribute to ecology and conservation
science, and the growing realization that skeletal remains can be used to examine
recent changes in ecosystems over years to decades (Dietl et al. 2011; Behrensmeyer
and Miller 2012). Taphonomic investigations assessing the quality of the recent
fossil record and identifying the limits of its applicability to neontological studies
demonstrate high fidelity between live communities and fossil assemblages (see
Kidwell and Tomašových 2013, and Kidwell 2015 for review). Likewise, critical
autecological and synecological studies, such as functional morphology, community
ecology, and predator–prey interactions, are necessary to understand past systems.
Discussions and colloquia have produced a plethora of ideas on the relevance of
fossil data to crises in the modern biosphere (e.g., Willis and Birks 2006; Dietl
and Flessa 2009; Dietl et al. 2011; Dietl and Flessa 2011; Louys 2012; Rick
and Lockwood 2013; Gatti et al. 2015; Dietl et al. 2015; Dietl 2016; Fordham
et al. 2016). A brief reflection on the growth and development of the discipline
illustrates that an awareness of paleontological applications to conservation began
just prior to the turn of the century, followed by rapid growth over the last
decade, particularly in recent years (Table 1). Research applicable to conservation
biology obviously began much earlier, chiefly in the last 20 years (Louys 2012).
However, immediately prior to 2001, significant momentum was achieved in the
paleontological community introducing the term “conservation paleobiology” (pre-
sumably prompting the organization of the 2001 North American Paleontological
Convention topical session “New Uses for the Dead: Paleobiological Contributions
to Conservation Paleobiology”), as the pursuit of research relevant to biodiversity
and conservation became widespread. A search of titles, abstracts, and keywords
in the Web of Science citation database (30th October 2016, WoS Core Collections,
ts D “conservation” and ts D “paleobiology”) suggests that the phrase “conservation
paleobiology” was first used in the title of a peer-reviewed journal article in only
2009 (Simões et al. 2009), reflecting its growing popularity in reference to research
applicable to historical ecology and conservation science at research symposia
and in discussions among paleoecologists. In 2011, a conservation paleobiology
workshop funded by the National Science Foundation appears to have spurred
a subsequent burst of research activity as evidenced by the numerous books,
symposia, and topical sessions in subsequent years (Table 1).
Table 1 Timeline for the growth and development of conservation paleobiology

Year Event
2001 North American Paleontological Society Symposium: “New Uses for the Dead:
Paleobiological Contributions to Conservation Paleobiology” (Organizers M
Droser and KW Flessa)
2005 Committee on the Geological Record of Biosphere Dynamics; National Research
Council; “The Geological Record of Ecological Dynamics: Understanding the
Effects of Future Environmental Change” (Chair KW Flessa)
2009 Paleontological Society Short Course at the Geological Society of America Annual
Conference: “Conservation Paleobiology: Using the Past to Manage for the Future”
(Organizers GP Dietl and KW Flessa); First published use of “Conservation
Paleobiology” in the title of a peer-reviewed journal article (Simões et al. 2009)
published in the journal Historical Biology
2011 NSF-Funded Workshop (Paleontological Research Institution, Ithica, New York):
“Conservation Paleobiology: Opportunities for the Earth Sciences”
2012 Publication of the book “Paleontology in Ecology and Conservation” (Author and
Editor J Louys)
2013 Geological Society of America Symposium: “Conservation Paleobiology—The
Microfossil Record” (Organizers S Goldstein and JH Lipps)
2013 Ecological Society of America Symposium: “Resilience, Disturbance, and
Long-Term Environmental Change-Integrating Paleoecology into Conservation
Management in the Anthropocene” (Organizer AW Ireland)
2014 North American Paleontological Society Symposia: “The Microfossil Record: The
Past is the Key to the Future (or Present) in Conservation Paleobiology”
(Organizers P Hallock and LS Collins); “Conservation Paleobiology: Ecosystem,
Community, and Species Response to Environmental Change” (Organizers CL
Tyler, SN Casebolt, R Terry)
2015 Global Change Biology Annual Symposium (UC Berkeley): “Integrating biology
and Paleobiology to Enhance Conservation of Terrestrial Ecosystems on a Rapidly
Changing Planet”; Society for Vertebrate Paleontology Conservation Paleobiology
Symposium – Annual Meeting
2016 4th International Sclerochronology Conference: “Environmental Monitoring and
Conservation Paleobiology” (Organizer GP Dietl)
2017 AAAS Annual Meeting Symposium “Conservation Paleobiology: Finding
Solutions in the Fossil Record” (Organizer S Kidwell)
Significant professional and academic activities expressly dedicated to the topic of conservation
paleobiology
2 Data in Conservation Paleobiology
Given the very recent establishment of the discipline, the accomplishments of the
last decade are particularly impressive. A large body of literature already exists,
and research with applications to conservation continues to grow. Geohistorical data
applied to conservation science include a wide range of data types and continuity
with sub-fossil and modern assemblages, employed to understand ancient and
modern ecosystems (Rick and Lockwood 2013). Particular emphasis has been
History of Life
Conservaon Paleobiology
Animal and plant macrofossils, shells, animal bones,
teeth, pollen, diatoms, trace fossils, microfossils, etc.
Past 2 million yrs
Past 10,000 yrs

Conservaon Archeobiology
Shell middens, tools, bone, teeth,
pollen, and plant remains
Past 500 yrs
Environmental History
Historical documents
and records, personal
journals, log books, and
photographs
Fig. 1 Range of data types within historical ecology that are commonly applied to conserva-
tion science. If conservation paleobiology data are considered a sub-discipline within conservation
science, historical ecology must be extended beyond human origins. Conservation paleobiology
can then further be divided into “near-time” considering the past 2 million years (within the dashed
line), and “deep-time,” potentially including all of the history of life (Dietl and Flessa 2011; Rick
and Lockwood 2013). Overlapping spheres illustrate overlapping temporal scales, for example, a
conservation paleobiology study may compare fossil species abundance and biodiversity from the
early Pleistocene to the modern
placed on human–environment interactions (Balée 2006; Szabó and Hédl 2011)

and natural variation pre- and post-human activities (Dietl and Flessa 2011). Data
can generally be grouped into three categories (Rick and Lockwood 2013): (1)
environmental history covering the past 500 years, (2) conservation archeobiology,
encompassing the past 10,000 years, and (3) conservation paleobiology, generally
from the Pleistocene and Holocene, but potentially including the entire history of
life (Fig. 1).
Historical records can provide valuable insights into ecosystem conditions prior
to European colonization, and baseline data for restoration targets. Written historical
records include a wide range of species and ecosystem data preserved mainly
in personal journals, log books, and photographs (Szabó and Hédl 2011; Rick
and Lockwood 2013). These data sources can provide important perspectives,
particularly considering historical shifts in ecological and biodiversity baselines of
large marine vertebrates. For example, written accounts from the 1700s have been
used to estimate population size and distribution of sea otters and the now extinct
Stellar’s sea cow (Hydrodamalis gigas) in kelp forest ecosystems, to reconstruct
long-term human impacts and baseline conditions (Jackson et al. 2001).
Similarly, archeological data can also be a valuable source for baseline con-
ditions. Conservation archeobiology includes data derived directly from human
activities, for example, shell middens, tools, bones, teeth, pollen, and plant remains
(Rick and Lockwood 2013). Such data can be used to understand demographic and
distribution shifts, translocations (movement of organisms by humans), extinctions,
and environmental reconstruction (Hofman et al. 2015). For example, Consuegra et
al. (2002) compared modern and sub-fossil Paleolithic genomes of Atlantic salmon
and found that the specimens from Iberian middens contained a previously common
mitochondrial genome that is now extinct. They cautioned that the use of modern
population data alone may not be representative of demographics during the last
glacial maximum. In this example, the addition of archeological data provided
a more complete understanding of salmon history, including the exposure of a
previously unrealized extinction and warn against a potential genetic bottleneck in
the declining Iberian salmon population.
Dietl and Flessa (2009, 2011) distinguished “near-time” and “deep-time” studies
(Dietl and Flessa 2009; Dietl and Flessa 2011), which roughly correspond to
intervals pre- and post-2 million years ago. Therefore, deep-time data may be
considered a fourth category in addition to those of Rick and Lockwood (2013),
as these fossil remains are likely to represent less similar communities, and to
have undergone more extensive alteration than their younger counterparts (Fig. 1).
Paleontological information from the past 2 million years includes similar species
and ecosystems to that of today’s oceans and may, therefore, be directly relevant and
applicable to current conservation (Smith et al. 2018; Dietl and Flessa 2011; Dietl
et al. 2015). These studies include establishing baseline conditions and capture the
response to human arrival (e.g., decline in oyster sizes between Late Pleistocene,
colonial, and modern times along the US Atlantic coast; Kusnerik et al. 2018),
and impacts of human-induced stressors (the history of overfishing and habitat loss
since pre-Roman times in the Mediterranean; Sala et al. 2012; Boada et al. 2015).
Studies of deep-time, prior to 2 million years ago, are differentiated by the greater
dissimilarity between ancient and modern ecosystems and species (for a thorough
review, see Dietl and Flessa 2009; Dietl and Flessa 2011; Dietl et al. 2015). A few
examples of deep-time studies include: (1) The varied responses of ecosystems
and species to perturbations, especially in terms of extinction, survivorship, and
recovery; (2) The range of conditions leading to stable ecosystem states; (3)
The conditions under which stable states persist; (4) Evolutionary and ecological
processes in the absence of human interference; and (5) The evolutionary and
ecological consequences of biotic invasions.
Although deep-time data come with some caveats that are comparatively negligi-
ble in near-time studies (for instance, the temporal resolution of paleobiological data
generally declines with age before present), the resulting “big-picture” perspective
can provide valuable information on how the marine biosphere has responded to,
and recovered from, global crises in the past, such as major climate shifts and mass
extinction events. In this regard, deep-time studies alone can answer questions about
the marine biosphere before, during, and after a global catastrophe (e.g., applying
the characteristics of successful refugia in past mass extinctions to marine reserves;
Schneider 2018). Ancient ecosystems and species may also differ significantly from
those of the modern, and an exact correspondence with present taxa is often not
possible. This is perhaps best emphasized through the recognition of the Cambrian,
Paleozoic, and Modern evolutionary faunas put forth by Sepkoski (1984), in which
the more ancient the assemblage, the generally more dissimilar from modern
systems in taxonomic dominants. However, assuming uniformitarianism, ecological
processes and evolutionary forces acting within the biosphere generally followed the
same principles. Therefore, even in ancient ecosystems that differ substantially in
composition from those of the present, valuable insights can still be gained. For
example, there have been two prior periods with major barrier reef systems: the past
few million years, and the Devonian Period. Conservation efforts in modern reef
systems may, therefore, benefit not only from the research of Pleistocene interglacial
reefs (e.g., Greenstein and Pandolfi 2008), but also the future investigation of the
causes and process of global reef demise in the Late Devonian with a focus on
applicability to modern crises in tropical oceans.
3 Looking Forward
Conservation is defined in the Merriam-Webster’s Dictionary as “a careful preser-

vation and protection of something; especially: planned management of a natural
resource to prevent exploitation, destruction or neglect.” If we wish to continue
to declare ourselves “conservation paleobiologists,” our research goals must con-
tribute to preservation and planned management. Paleobiological studies frequently
include only a brief phrase or statement about the potential relevance of their
findings to modern systems and conservation science (Burnham 2001; Kelley et al.
2018; Savarese 2018). Conservation paleobiology must move beyond potential and
towards applications to narrow the implementation gap. Relevant paleobiological
information needs to be presented, and made available, to the appropriate audiences,
i.e., resource managers, government agencies, and stakeholders (Smith et al. 2018;
Kelley et al. 2018; Savarese 2018). Furthermore, as conservation paleobiology
and conservation science have developed in parallel as two disparate disciplines,
managers are typically unaware of what conservation paleobiology has to offer.
Savarese (2018) provides a thorough discussion on establishing connections and
communication with decision-making groups. Some of these issues could also
be mitigated by enacting conservation paleobiology as a sub-discipline within
conservation science, as suggested by Dietl (2016).
Numerous principal objectives of conservation paleobiology have been proposed

by Dietl and others (Dietl and Flessa 2009; Dietl and Flessa 2011; Dietl et al. 2015)
and can be summarized as follows:
1. Establish baselines to compare pre- and post-anthropogenic disturbance condi-
tions
2. Examine response to perturbations, both natural and anthropogenic in origin
3. Document historical range of variability within ecosystems, and predict poten-
tial geographic range shifts in response to climate change
4. Provide information for the determination of realistic restoration targets
5. Distinguish between anthropogenic and non-anthropogenic change
6. Recognition of ecological legacies only explainable by prior events and condi-
tions
7. Identify vulnerable species in urgent need of protection and evaluating their
extinction risk
8. Identify invasive species
9. Detect recent shifts in species abundance
10. Assess changes in genetic diversity
11. Identify future refugia in anticipation of climate change
Ultimately, we must strive to contribute to conservation in a changing world.
Global and local environmental change will continue to occur, and conservation
scientists and ecologists are therefore trying to quantify and predict biotic changes
using systems either already altered or currently in flux. As paleontologists, we have
the unique and critical perspective of time and change that is lacking in the study of
modern systems. As conservation paleobiologists, we can provide crucial baselines,
insights into pre-impact conditions and ecosystems states, and perspectives on the
“hows” and “whys” of biosphere resilience through global catastrophes. Change is
inevitable. Whether or not the oceans will be able to provide services in the future is
uncertain. Thus, conservation paleobiology may prove essential for maintaining the
marine biosphere along with its resources and ecosystem services, and ultimately,
the survivorship of our own species.
References
Balée W (2006) The research program of historical ecology. Annu Rev Anthropol 35:75–98
Ban N, Alder J (2008) How wild is the ocean? Assessing the intensity of anthropogenic marine
activities in British Columbia, Canada. Aquat Conserv Mar Freshwat Ecosyst 18:55–85
Behrensmeyer AK, Miller JH (2012) Building links between ecology and paleontology using
taphonomic studies of recent vertebrate communities. In: Louys J (ed) Paleontology in ecology
and conservation. Springer, New York, pp 69–91
Boada J, Arthur R, Farina S et al (2015) Hotspots of predation persist outside marine reserves in
the historically fished Mediterranean Sea. Biol Conserv 191:67–74
Boivin NL, Zeder MA, Fuller DQ et al (2016) Ecological consequences of human niche
construction: examining long-term anthropogenic shaping of global species distributions. Proc
Natl Acad Sci 113:6388–6396
Burnham RJ (2001) Is conservation biology a paleontological pursuit? PALAIOS 16:423
Consuegra S, García de Leániz C, Serdio A et al (2002) Mitochondrial DNA variation in
Pleistocene and modern Atlantic salmon from the Iberian glacial refugium. Mol Ecol 11:2037–
2048
Dietl GP (2016) Brave new world of conservation paleobiology. Front Ecol Evol 4:21
Dietl GP, Flessa KW (eds) (2009) Conservation paleobiology: using the past to manage for the
future. The Paleontological Society Papers vol. 15, Lubbock, TX
Dietl GP, Flessa KW (2011) Conservation paleobiology: putting the dead to work. Trends Ecol
Evol 26:30–37
Dietl GP, Flessa KW (2018) Should conservation paleobiologists save the world on their own time?
In: Tyler CL, Schneider CL (eds) Marine conservation paleobiology. Springer, Cham, pp 11–22
Dietl GP, Kidwell SM, Brenner M et al (2011) Convservation paleobiology: opportunities for Earth
science. Ithica, New York
Dietl GP, Kidwell SM, Brenner M et al (2015) Conservation paleobiology: leveraging knowledge
of the past to inform conservation and restoration. Annu Rev Earth Planet Sci 43:79–103
Fordham DA, Akçakaya HR, Alroy J et al (2016) Predicting and mitigating future biodiversity loss
using long-term ecological proxies. Nat Clim Chang 6:909–916
Froyd CA, Willis KJ (2008) Emerging issues in biodiversity and conservation management: the
need for a palaeoecological perspective. Quat Sci Rev 27:1723–1732
Gatti G, Bianchi CN, Parravicini V et al (2015) Ecological change, sliding baselines and the
importance of historical data: lessons from combing observational and quantitative data on
a temperate reef over 70 years. PLoS One 10:e0118581
Greenstein BJ, Pandolfi JM (2008) Escaping the heat: range shifts of reef coral taxa in coastal
Western Australia. Glob Chang Biol 14:513–528
Hofman CA, Rick TC, Fleischer RC, Maldonado JE (2015) Conservation archaeogenomics:
ancient DNA and biodiversity in the Anthropocene. Trends Ecol Evol 30:540–549
Hubbell SP (2001) The unified neutral theory of biodiversity and biogeography. Princeton
University Press, Princeton
Jackson ST, Hobbs RJ (2009) Ecological restoration in the light of ecological history. Science
325:567–569
Jackson JBC, Kirby MX, Berger WH et al (2001) Historical overfishing and the recent collapse of
coastal ecosystems. Science 293:629–638
Kelley PH, Dietl GP, Visaggi CC (2018) Training tomorrow’s conservation paleobiologist. In: Tyler
CL, Schneider CL (eds) Marine conservation paleobiology. Springer, Cham, pp 207–223
Kidwell SM (2015) Biology in the Anthropocene: challenges and insights from young fossil
records. Proc Natl Acad Sci U S A 112:4922–4929
Kidwell SM, Tomašových A (2013) Implications of time-averaged death assemblages for ecology
and conservation biology. Annu Rev Ecol Evol Syst 44:22.1–22.25
Knowlton N, Jackson JBC (2008) Shifting baselines, local impacts, and global change on coral
reefs. PLoS Biol 6:1–6
Kusnerik KM, Lockwood R, Grant AN (2018) Using the fossil record to establish a baseline and
recommendations for oyster mitigation in the mid-Atlantic U.S. In: Tyler CL, Schneider CL
(eds) Marine conservation paleobiology. Springer, Cham, pp 75–103
Louys J (2012) Paleontology in ecology and conservation: an introduction. In: Louys J (ed)
Paleontology in ecology and conservation. Springer, Berlin Heidelberg, pp 1–8
Louys J, Wilkinson DM, Bishop LC (2012) Ecology needs a paleontological perspective. In: Louys
J (ed) Paleontology in ecology and conservation. Springer, New York, pp 23–38
Meine C, Soule M, Noss RF (2006) “A mission-driven discipline”: the growth of conservation
biology. Conserv Biol 20:631–651
Pauly D (1995) Anecdotes and the shifting baseline syndrome of fisheries. Trends Ecol Evol 10:430
Rick TC, Lockwood R (2013) Integrating paleobiology, archeology, and history to inform
biological conservation. Conserv Biol 27:45–54
Sala E, Ballesteros E, Dendrinos P et al (2012) The structure of Mediterranean rocky reef
ecosystems across environmental and human gradients, and conservation implications. PLoS
One 7:e32742
Savarese M (2018) Effectively connecting conservation paleobiological research to environmental
management: examples from Greater Everglades’ restoration of southwest Florida. In: Tyler
Schneider CL (2018) Marine refugia, past, present, and future: lessons from ancient geologic
crises for modern marine ecosystem conservation. In: Tyler CL, Schneider CL (eds) Marine
conservation paleobiology. Springer, Cham, pp 161–206
Sepkoski JJ (1984) A kinetic model of Phanerozoic taxonomic diversity. III. Post-Paleozoic
families and mass extinctions. Paleobiology 10:246–267
Silvertown J, Poulton P, Johnston E et al (2006) The park grass experiment 1856-2006: its
contribution to ecology. J Ecol 94:801–814
Simões MG, Rodrigues SC, Kowalewski M (2009) Bouchardia rosea, a vanishing brachiopod
species of the Brazilian platform: taphonomy, historical ecology and conservation paleobiology.
Hist Biol 21:123–137
Smith JA, Durham SR, Dietl GP (2018) Conceptions of long-term data among marine conservation
biologists and what conservation paleobiologists need to know. In: Tyler CL, Schneider CL
Soule ME (1985) What is conservation biology? Bioscience 35:727–734
Szabó P, Hédl R (2011) Advancing the integration of history and ecology for conservation. Conserv
Biol 25:680–687
Tyler CL (2018) A conceptual map of conservation paleobiology: visualizing a discipline. In: Tyler
Willis KJ, Birks HJB (2006) What is natural? The need for a long-term perspective in biodiversity
conservation. Science 314:1261–1265
Should Conservation Paleobiologists Save
the World on Their Own Time?
Gregory P. Dietl and Karl W. Flessa
Abstract Conservation paleobiology is an applied science. Academically based

conservation paleobiologists can put their science to work in field applications,
informing environmental policy and management, or both. This essay provides
some thoughts on how conservation paleobiologists can engage with stakeholders
and policymakers. Advocacy is one option, but it is not required. The decision to
advocate for a particular conservation policy or goal is a personal one and drawing
on one’s scientific expertise and experience is part of such advocacy. Marshalling
facts in support of one’s values can strengthen advocacy if one is explicit about
one’s values and committed to the ethical norms of science. Styles of engagement
with the policy world will vary: from arbiters of scientific information to advocates
for particular solutions to conservation problems. Engaging with policymakers and
implementing policy also means accepting that not all facts are equally important
and that compromise is often necessary. There are ample resources available to those
who decide that they want to get involved: take advantage of workshops, formal
programs, and publications about science communication and the science–policy
interface. Making decisions based on evidence is not a partisan issue—don’t be
afraid to engage.
Keywords Conservation biology · Conservation science · Objectivity · Science

advocacy · Science communication · Values
G. P. Dietl ()
Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY, USA
K. W. Flessa
Department of Geosciences, University of Arizona, Tucson, AZ, USA

12 G. P. Dietl and K. W. Flessa
1 Always Academicize?
Stanley Fish’s 2008 book, Save the World on Your Own Time, made an unabashed
public defense of what he thought academics could “legitimately (as opposed to
presumptuously)” claim to be able to do as part of their job. Fish argued that it
was imperative always to “academicize” issues. By this he meant that we should
always separate an issue from contexts that pose a choice of what to do—describe
a specific policy option, not advocate for or against it. According to Fish (p. 169),
once you start engaging in “discussions designed to produce action in the world, you
are surely doing something, but it is not academic, even if you give it that name.”
Fish made a valid point, but one that seems at odds with the job description of
conservation scientists (including conservation paleobiologists) whose research sits
at the nexus of science and politics. Today it is increasingly accepted by public
and policymakers alike that scientists should become more closely involved in
the search for solutions to major environmental problems (Baron 2016). These
requests for increased participation, however, raise important questions about how
a conservation paleobiologist could engage with the public in the process of solving
conservation problems.
Here, our goal is to raise awareness in the conservation paleobiology community
about the options available in making the leap from quietly describing the facts
(Fish’s academicizing) to entering the policy world to inform decision-makers and
help implement policy. We also offer practical advice for advocating responsibly
and how to begin. We felt compelled to share our thoughts on the obviously deeply
personal decision to venture into the policy world because we believe fear of
“crossing an illusory line into advocacy” (Garrard et al. 2016) might stop many
conservation paleobiologists from participating in the process.
2 To Advocate, or Not to Advocate
Are scientists uniquely qualified to offer trustworthy advice about which policies
might best help conserve biodiversity or ecosystem services? Only a decade ago,
many scientists would have answered that such active advocacy should be avoided
(Lach et al. 2003), fearing that doing so might damage one’s scientific credibility,
objectivity, and impartiality, while at the same time eroding the public’s confidence
in their work (e.g., Lackey 2007; Nelson and Vucetich 2009). But, the times are
changing. The last decade has witnessed a radical shift in the mindset of many
scientists (Baron 2016). Discussing the value of outreach by scientists to the public
and policymakers, Nancy Baron quipped that the question is no longer “Should I
do this?, but rather, How do I do this?” (Baron 2016, p. 518). For instance, Singh
et al.’s (2014) survey of more than 500 scientists from more than 30 countries
worldwide who attended ecological conferences, such as the annual Ecological
Society of America meeting, showed that many of them wanted to engage more
Should Conservation Paleobiologists Save the World on Their Own Time? 13
with policymakers as advocates for political action. Nelson and Vucetich (2009)
have also shown that many traditional arguments against science advocacy (e.g., it
conflicts with science because the purpose of science is to remain neutral and value-
free; Nielsen 2001), really say more about how you should advocate responsibly
than whether you should advocate in the first place.
Our own view is that there is nothing inappropriate about wanting to speak out
about the very real conservation implications of your research. Because conserva-
tion paleobiologists (and all scientists) are citizens first and scientists second, they
should have the right to advocate1 (or not at all) as long as it is done responsibly.
3 Speaking Honestly to Power
What does it mean to be a responsible advocate? For a conservation paleobiologist—

or any scientist—to sustain a credible and effective presence in the public sphere,
we believe that they must, above all else, be honest—to their science and to their
values. To paraphrase the late climate scientist Stephen Schneider, first you need to
make your own values and biases “conscious,” then, you have to make them explicit,
and not assume that others hold them. Some introspection will help. For starters, try
asking yourself: What are my personal values and prior beliefs and how do they
bias my choice of methods and interpretations (Garrard et al. 2016)? Let’s say that
you value biodiversity for its own sake and believe that it is currently threatened.
Given these personal positions you should be mindful that: (1) your methodological
decisions (e.g., sampling design, analytical approach) may enhance your ability to
interpret your results as “positive”—that is, to find what you expected (or wanted)
to find; and, (2) because biodiversity issues are often emotionally charged, your
feelings may come into play when interpreting your data (e.g., when choosing to
focus on (or ignore) data that are consistent/inconsistent with a preferred policy
option) (Wiens 1997). “No one is exempt from prejudices and values, but the people
who know when they are bringing in values and make their biases explicit are more
likely to provide balanced assessments—and to be able to single out those who do
not” (Schneider n.d.).
When speaking honestly to power, it’s also good to be conscious of who you
are speaking for. A lot of environmental decision-making involves “stakeholders”—
those individuals, groups, agencies, landowners, industries, etc., whose interests will
be affected by the decisions that must be made. Academic scientists have a long
tradition of advising stakeholders, but are they stakeholders themselves?—not in the
traditional sense. Our livelihoods are rarely affected by the pending decisions over
1
We also prefer to use the term “advocate” and not Sara Goldrick-Rab’s (2014) related
term “scholar activist.” In our opinion, the distinction matters because being a conservation
paleobiologist—unlike a scholar activist—means that we start with a goal in mind (e.g., the
conservation of biodiversity and ecosystem services), and work for change to achieve this goal
(Dietl 2016).
land use, emissions, recovery plans, protected areas, etc. We are usually careful to
avoid speaking for our employers, funding sources, or professional societies. We
speak for ourselves (after all each of us is a citizen entitled to preferences), but must
protect the credibility of scientists everywhere by advocating for a policy option
only when there is a rigorous and objective scientific basis for taking that position.
4 From Pure Scientist to Honest Broker
The way that you advocate is your choice. Suppose a politician asks you whether she
should support a proposed regulation to stop the spread of an invasive species. What
should you do?—try to convince her to support the proposal; only tell her about the
current scientific findings on the issue; answer her questions about the science of the
issue; or, explain the merits of the proposed regulation as well as several alternatives
to it? This is not an easy choice to make, made even more difficult because most
scientists are politically naïve and not aware that they even have a choice.
In his 2007 book, The Honest Broker: Making Sense of Science in Policy and
Politics, Roger Pielke stressed that scientists have choices in what role they play in
the policy process. His taxonomy is not prescriptive; its utility lies in illustrating
many, though likely not all, of the roles available to scientists in the policy realm.
Pielke argued that when there is widespread public agreement and low scientific
uncertainty concerning a potential decision the “linear model” of science and poli-
tics, in which science advice comes before and compels political action (Jasanoff
2008), applies and scientists can provide input as “Pure Scientists” who simply
provide knowledge or as “Science Arbiters” who serve as advisors in developing
policy. A pure scientist makes no effort to engage in advocacy or to promote certain
policies outside of how they might communicate by way of peer-reviewed scientific
papers. As Pielke explained, “ : : : research results in findings that are placed into
a reservoir of knowledge where they will be available to all decision-makers” (p.
15). Like the pure scientist, the science arbiter expresses no opinion on policy
preferences, nor does the science arbiter proactively discuss policy at all. Objectivity
is the scientist’s primary goal. Yet the science arbiter departs from the pure scientist
in that they make themselves available to policymakers by providing answers to
questions that can be addressed using the tools of science. We suspect that many
conservation paleobiologists would feel most comfortable as a pure scientist or
science arbiter, as these roles reflect entrenched dichotomies between science and
society. Pielke, however, argues that far from keeping politics out of science, acting
in accord with the linear model encourages the mapping of political interests onto
science, i.e., it leads to a politicization of science—whether the issue is climate
change or biodiversity loss—when partisan lobbyists “cherry-pick” scientific results
as a means to negotiate for a desired political outcome (see Pielke 2006).
When conflicts in values and a high degree of scientific uncertainty are con-
tributing to disagreements, scientific findings alone cannot lead to a decision or
action. For example, with many conservation issues, high scientific uncertainty
makes it difficult to exclude a wide range of future outcomes. In such contexts,

the expectations that science will resolve political conflict almost always fall short
because science provides an “excess of objectivity” that can be used to support
a range of “subjective, political positions on complex issues such as climate
change : : : or endangered species” (Sarewitz 2000, p. 90). According to Pielke,
scientists thus face the question of whether to try to broaden or reduce the range
of policy options on the table. If scientists choose to act as “Issue Advocates,” they
aim to reduce the scope of choices toward one preference or a small set of related
preferences. Issue advocates thus attempt to steer policymakers and public opinion
to promote a favored issue position (e.g., support for the Aichi biodiversity targets2 ),
one they believe is best according to their objective scientific judgment. On the
other hand, if scientists seek to broaden the range of options, they become “Honest
Brokers of Policy Alternatives” that clarify existing options and identify new
options. By expanding thinking, the honest broker encourages decision-makers to
think beyond personal values and ideologically driven biases to consider alternatives
that may make sense from diverse viewpoints.
Inevitably, to have any real-world impact, most conservation paleobiologists will
find themselves filling one (or a combination) of Pielke’s advocacy roles. And you
may fit one category on one issue (or day) and another category on another issue (or
day). We highlight two case studies from among our own ranks. Both are lessons on
how to embrace serendipity.
In the media blitz that followed the publication of a Nature paper (Barnosky et al.
2012) reporting that our planet faced a “planetary-scale tipping point,” conservation
paleobiologists—Tony Barnosky and Liz Hadly—found themselves unexpectedly
being catapulted into the political fray when the governor of the state of California—
Jerry Brown—called. He wanted to ask them questions about the science in the
paper, but, importantly, also to translate the science in the paper into something
that policymakers could use (Gewin 2014; Barnosky et al. 2014). The resulting
consensus document (Barnosky et al. 2014) informed climate policy discussions
(Fig. 1) and helped facilitate two international agreements to reduce greenhouse-
gas emissions and develop green technology—a memorandum of understanding
between two of the world’s largest economies, China and California,3 and the
Pacific Climate Pact between the states of California, Oregon, Washington, and
British Columbia4 (see Barnosky et al. 2014 for further details). Throughout the
process, Barnosky and Hadly had to make personal decisions on how they would
get involved. They have been open about their preferred form of advocacy, what
they call “informative advocacy” (Hadly and Barnosky 2014), which is similar in
spirit to Peilke’s honest broker role. According to Hadly and Barnosky, informative
advocacy:
2
https://www.cbd.int/sp/targets/.
3
http://gov.ca.gov/news.php?id=18205.
4
http://gov.ca.gov/news.php?id=18284.
Fig. 1 California’s governor, Jerry Brown, presenting consensus statement at NASA’s Ames
Research Center in 2013. Reprinted by permission from Macmillan Publishers Ltd.: Nature, vol.
511: 402–404, copyright 2014
: : : inject[s] : : : scientific realities into the many different categories of information that
decision makers must take into account when formulating policy. Informative advocacy
also has a second goal that is critical: learning from decision makers about the kind
of information they need. This back-and-forth dialog ultimately opens new doors for
decision makers to formulate solutions to complex problems, and new doors for scientists
to understand how their science is socially relevant. (Hadly and Barnosky 2014)
Informative advocacy attempts to expand the range of options, recognizing that

there are multiple “scientifically sound paths to desired destinations” (Hadly and
Barnosky 2014).
Note that Barnosky and Hadly depended on the media to make the connection
between the science and politics of the issue and that initial connection was
made at a high level. Their experience contrasts with our second example of how
conservation paleobiologists are more likely to become engaged.
Karl Flessa’s story starts at a particular place: the mudflats of the Colorado
River delta. His lab’s work documented, in a scientifically rigorous fashion, what
everyone should have expected: that turning off the supply of Colorado River water
had a profoundly deleterious effect on its estuary in the upper Gulf of California
(Calderon-Aguilara and Flessa 2009). Others had already documented the damage
to the river’s riparian zone. Along the way, Flessa facilitated collaborations among
Mexican and US scientists from universities, conservation groups, and agencies. He
would have likely been happy to continue in his roles as conservation paleobiologist
and science-facilitator.
But drought happened. Southwestern water agencies wanted to see if a dormant

desalination plant at the US-Mexico border could help supplement dwindling
supplies. But even a test run of the plant would increase the salinity and decrease
the amount of water that supplied an ecologically important Mexican wetland.
Conservation groups and agencies from the USA and Mexico came up with
a possible solution and asked Flessa to organize a binational science team to
monitor the wetland during the desalination plant’s trial run. No harm resulted. The
conservation groups were pleased, the scientists got to publish papers (Glenn 2013),
and the water agencies now include some of the costs of environmental protection
in their estimates of the economic feasibility of this desalination plant.
But then an earthquake happened in 2010. It damaged Mexico’s system of
irrigation canals and the USA agreed to store Mexico’s annual allocation of water
in a US reservoir, Lake Mead. This was a good deal for the USA as well. Keeping
the water level high in Lake Mead provided some insurance against a shortage that
would decrease the delivery of water to Arizona, Nevada, and eventually, California.
A five-year agreement, called Minute 319 (IBWC 2012) (a “Minute” is an explicit
agreement within the confines of the 1944 water treaty between the USA and
Mexico) allowed Mexico to continue to store part of its allocation in Lake Mead,
stipulated the degree to which Mexico and the USA would share in any future water
shortages or surpluses and, importantly for this story, arranged for a transboundary
environmental flow (Fig. 2) to see what its effects might be on the hydrology and
environment of the Colorado River delta’s riparian zone (Flessa et al. 2013). Flessa
and others organized another binational science team to monitor the effects of this
flow. The flow benefited the trees, birds, groundwater, and people of the riparian
zone (Flessa et al. 2016; Bark et al. 2016). Scientists deployed to respond to the
policymakers’ needs to account for the effects of the environmental flows. And the
delta science team is now playing a major role in designing scenarios for future
flows. Policy is shaping the science that gets done; science is shaping the policy that
develops.
Did analyses of dead clams (i.e., conservation paleobiology) result in the first
transboundary environmental flow to the Colorado River delta? No. Did it help
scientists take a more active role in decision-making over transboundary water? Yes.
Did it move Flessa closer to the policy sphere? Yes, but that was his choice. In the
context of Pielke’s taxonomy, did Flessa act as a science arbiter, an issue advocate
or an honest broker?—a little of each, depending on the circumstances.
These two examples illustrate three essential points. First, exactly which type of
advocacy you feel comfortable with, and competent at, will vary among individuals.
That is fine—it’s a personal decision. You need to find your comfort zone (Flessa
2017). Pielke’s taxonomic scheme makes it clear that you have a range of options to
consider—all of which have their roles in informing policy decisions—about what
best suits your personal situation. The important point to remember is that your
choice matters. Second, pathways to engagement with policymakers vary. A few
will be high-profile, media-intense events. More are likely to involve on-the-ground
collaborations with agencies or conservation groups. And, third, because natural
resource issues are often politically charged, all productive engagements will require
reciprocity: that scientists understand the needs and limitations of the policymakers,
and that policymakers understand the capabilities and limits of the scientists.
Fig. 2 Gates of Morelos Dam opening on March 23, 2014, sending the first transboundary
environmental flow into the Colorado River delta. Photo Credit: Karl Flessa
5 Keeping It Real
Decisions about environmental policies and their implementation are rarely—if

ever—made on the basis of scientific facts alone. This reality is not a matter of
“facts” vs. “feelings.” Facts occur in a social context and they never speak for
themselves. “If you come from an angle of ‘science is the truth you just don’t
understand the facts’, people will not be willing to listen to you” (Xavier Basurto, in
Gewin 2017, p. 426). Accept that policy decision-making is not always rational and
some facts are more important than others. The facts of climate change will not win
over the denialists. The facts of evolution will not win over the creationists. And, in
some cases, it’s not even a deficit of scientific facts that stands in the way of action.
Ideology and competing stakeholder interests can stand in the way. But, this does
not mean that scientists should not engage with the public or with policymakers. It
simply means that you can’t just assume that facts will always win the day.
In addition, there may be many legitimate “facts” at play. For example, the
Colorado River’s simulated spring flood of 130 million cubic meters was much less
than the 320 million cubic meters called for by the conservation-group scientists.
Constraints on delivery mechanisms were also facts: the amount delivered was
the maximum that the channel and canals could provide without flooding fields
and urban areas. The farmer’s irrigation contracts were also facts. The water was
delivered in late March/early April—months before the spring floods of the pre-
dam past. Delivering environmental water later was not possible because channel
and canal capacity was already dedicated toward meeting irrigation needs at that
time. As a “new” user of the river’s water, the environmental groups needed to
accommodate the facts on the ground. They decided to avoid letting the perfect be
the enemy of the good. They compromised.
6 Overcoming the Fear Factor
By this point, many of you who want to bridge the gap between your conservation
paleobiology research and the “real world” may be thinking to yourself that you
don’t know where to begin. Baron (2016, p. 518) evocatively called this gap—
between desire and action—a “valley of death” that can stop scientists from
engaging in policy issues. As academic scientists, you probably have little formal
training in public engagement (Kelley et al. 2018) and may find the prospect of
trying to inject your scientific findings into a public debate a little terrifying—or
even a lot. Fortunately, there are resources to turn to for help to get you started.
Nancy Baron’s 2010 book, Escape from the Ivory Tower: A Guide to Making Your
Science Matter, and Lee Badgett’s 2016 book, The Public Professor: How to Use
Your Research to Change the World, both offer accessible and practical advice to
scientists looking to engage with the policy world outside of academia. For instance,
Badgett boils down the things that effective “public professors” do well, and that
everyone else—from senior scientists who have reached a point in their careers
when they decide to focus on channeling their scientific expertise toward the public
good to newly minted PhDs who share their generation’s commitment to community
service—should learn how to do better: (1) seeing the bigger picture of a policy
debate to identify a role for your research; (2) developing personal relationships
with users of scientific information—such as journalists, policymakers, and business
leaders—who will be more willing to trust what you have to say if they already
know you; and (3) communicating5 your research outside academic circles (e.g.,
volunteering your expertise to an advocacy group, writing an op-ed piece, getting
your institution to write a media release about your work, reaching out to the agency
personnel responsible for implementing environmental policies, participating in
legislative visit days, testifying at legislative hearings). Honing and enhancing
these public engagement skills should be a high priority for all conservation
paleobiologists.
Two widely respected organizations that provide training opportunities to bridge
the gap are COMPASS and the Leopold Leadership Program. The mission of
COMPASS is “to help scientists effectively share their knowledge in the public
discourse and decision-making.”6 According to co-founder Jane Lubchenco the
Leopold Leadership program7 was envisioned as “a leadership and communication
training program to help environmental scientists become more effective communi-
5
A review of the extensive literature on the effective communication of science is beyond the scope
of this essay. Nevertheless, how you deliver the facts can be as important as the facts themselves.
Randy Olson’s 2009 book, Don’t Be Such a Scientist: Talking Substance in an Age of Style, and
Cornelia Dean’s 2012 book, Am I Making Myself Clear?: A Scientist’s Guide to Talking to the
Public, provide readable introductions.
6
http://www.compassonline.org/.
7
http://leopoldleadership.stanford.edu/.
cators of science to the public and policy makers.”8 Since their founding back in the
late 1990s, these programs have helped train thousands of scientists, improving the
flow of accurate, credible scientific information to policy makers, and the general
public alike on critical conservation issues. Conservation paleobiologists should
follow their lead and take advantage of the training opportunities that these programs
offer.
Engaging in the policy arena is not easy. If you decide that using your science to
shape policy is what you want to do, the opportunities are many and resources are
there to help. Don’t let fear paralyze your desire to get involved.
7 Later Is Too Late
In the coming years, conservation paleobiologists will (and should) continue to

academicize conservation issues, big and small. But, if conservation paleobiologists
ever had the luxury of remaining above the political fray, they no longer have it.
We know, as scientists, that the status quo is producing the sixth mass extinction
event. We believe that conservation paleobiologists who desire to inform public
policy to help solve this problem must do more than academicize the issues involved
and increasingly engage in the process. This need is greater now than ever before.
The examples we have highlighted illustrate that conservation paleobiologists can
emerge as powerful agents of change. More of us, however, still need to learn how
to bridge the gap between our scholarly work and the public that so urgently needs
it. Never underestimate the worth of your contribution.
Acknowledgments We thank Carrie Tyler and Chris Schneider for their invitation to contribute
to this volume and their thoughtful suggestions to improve the paper.
References
Badgett MVL (2016) The public professor: how to use your research to change the world. New
York University Press, New York, p 256
Bark RH, Robinson J, Flessa KW (2016) Tracking cultural ecosystem services: water chasing the
Colorado River restoration pulse flow. Ecol Econ 127:165–172
Barnosky AD, Hadly EA, Bascompte J et al (2012) Approaching a state-shift in Earth’s biosphere.
Nature 486:52–56
Barnosky AD, Hadly EA, Dirzo R, Fortelius M, Stenseth NC (2014) Translating science for
decision makers to help navigate the Anthropocene. Anthr Rev 1:160–170
Baron N (2010) Escape from the ivory tower: a guide to making your science matter. Island Press,
Washington, p 272
8
http://oregonstate.edu/ua/ncs/archives/1998/aug/new-aldo-leopold-program-help-scientists-
reach-out.
Baron N (2016) So you want to change the world? Nature 540:517–519

Calderon-Aguilera LE, Flessa KW (2009) Just add water? Transboundary Colorado River flow
and ecosystem services in the upper Gulf of California. In: López-Hoffman L, McGovern ED,
Varady RG, Flessa KW (eds) Conservation of shared environments: learning from the United
States and Mexico. University of Arizona Press, Tucson, pp 154–169
Dean C (2012) Am I making myself clear?: A scientist’s guide to talking to the public. Harvard
University Press, Cambridge, p 288
Fish S (2008) Save the world on your own time. Oxford University Press, New York, p 189
Flessa KW (2017) Putting the dead to work: translational paleoecology. In: Dietl GP, Flessa KW
(eds) Conservation paleobiology: science and practice. University of Chicago Press, Chicago,
pp 283–289
Flessa KW, Glenn EP, Hinojosa-Huerta O, de la Parra-Renteria CA, Ramirez-Hernandez J, Schmidt
JC, Zamora-Arroyo F (2013) Flooding the Colorado River Delta: a landscape-scale experiment.
Eos 94:485–486
Flessa KW, Kendy E, Schlatter K (2016) Minute 319 Colorado River delta environmental
flows monitoring interim report, International Boundary and Water Commission. http://
www.ibwc.state.gov/Files/Minutes%20319/2016_EFM_InterimReport_Min319.pdf.
Garrard GE, Fidler F, Wintle BC, Chee YE, Bekessy SA (2016) Beyond advocacy: making space
for conservation scientists in public debate. Conserv Lett 9:209–212
Gewin V (2014) Hello, Governor. Nature 511:402–404
Gewin V (2017) Communication: post-truth predicaments. Nature 541:425–427
Glenn EP (ed) (2013) Special issue: Colorado delta wetlands. Ecol Eng 59:1–184
Goldrick-Rab S (2014) On scholarly activism. Contexts (blog). https://contexts.org/blog/on-
scholarly-activism/. Accessed 4 Dec 2016
Hadly EA, Barnosky AD (2014) Beyond science communication: informative versus prescrip-
tive advocacy. ConsenusForAction Blog. http://consensusforaction.stanford.edu/blog/beyond-
science-communicatio.html. Accessed 5 Jan 2016
International Boundary and Water Commission [IBWC] (2012) Interim international cooperative
measures in the Colorado River Basin through 2017 and extension of Minute 318 cooperative
measures to address the continued effects of the April 2010 earthquake in the Mexicali Valley,
Baja California. http://www.ibwc.state.gov/Files/Minutes/Minute_319.pdf.
Jasanoff S (2008) Speaking honestly to power. Sci Am 96:240
Kelley PH, Dietl GP, Visaggi CC (2018) Training tomorrow’s conservation paleobiologists. In:
Tyler CL, Schneider CL (eds) Marine conservation paleobiology. Springer, Cham, pp 207–223
Lach D, List P, Steel B, Shindler B (2003) Advocacy and credibility of ecological scientists in
resource decisionmaking: a regional study. Bioscience 53:170–178
Lackey RT (2007) Science, scientists, and policy advocacy. Conserv Biol 21:12–17
Nelson MP, Vucetich JA (2009) On advocacy by environmental scientists: what, whether, why, and
how. Conserv Biol 23:1090–1101
Nielsen LA (2001) Science and advocacy are different—and we need to keep them that way. Hum
Dimens Wildl 6:39–47
Olson R (2009) Don’t be such a scientist. Talking substance in an age of style. Island Press,
Washington, p 206
Pielke RA (2006) When scientists politicize science. Regulation 29:28–34
Pielke RA (2007) The honest broker: making sense of science in policy and politics. Cambridge
University Press, Cambridge, p 188
Sarewitz D (2000) Science and environmental policy: an excess of objectivity. In: Frodeman R
(ed) Earth matters: the earth sciences, philosophy, and the claims of community. Prentice Hall,
Upper Saddle River, pp 79–98
Schneider, SH (n.d.) The Scientist-Advocate. In: “Mediarology”. https://
stephenschneider.stanford.edu/Mediarology/mediarology.html. Accessed 31 Jan 2017
Singh GC, Tam J, Sisk TD, Klain SC, Mach ME, Martone RG, Chan KMA (2014) A more
social science: barriers and incentives for scientists engaging in policy. Front Ecol Environ 12:
161–166
Wiens JA (1997) Scientific responsibility and responsible ecology. Conserv Ecol 1:art16
Conceptions of Long-Term Data Among
Marine Conservation Biologists
and What Conservation Paleobiologists
Need to Know
Jansen A. Smith, Stephen R. Durham, and Gregory P. Dietl
Abstract Marine conservation biologists increasingly recognize the value of long-

term data and the temporal context they can provide for modern ecosystems. Such
data are also available from conservation paleobiology, but the enormous potential
for integration of geohistorical data in marine conservation biology remains unre-
alized. The lack of a common language for data integration and a tendency in each
field to measure different variables, at scales that may differ by orders of magnitude,
make integration difficult. To better understand how conservation paleobiology can
maximize its potential, we conducted a survey of marine conservation biologists
working in the United States.
The respondent population included 90 marine conservation biologists from
a variety of workplaces (e.g., governmental, academic) and experience levels
(<5 years to >25 years). Survey responses indicated that our fields share common
conservation goals (e.g., conservation of biodiversity and ecosystem services) and
use long-term data in similar ways (e.g., to establish baselines and elucidate trends
and patterns). Respondents, however, mostly considered “long term” to refer to
decadal timescales and rarely mentioned geohistorical data.
Overall, the survey results suggest conservation paleobiologists have much
work to do before geohistorical data are regularly accepted and applied in
marine conservation biology. We highlight four takeaways from the results of
J. A. Smith ()
S. R. Durham
Present address: Paleontological Research Institution, Ithaca, NY, USA
G. P. Dietl

24 J. A. Smith et al.
our survey that can help conservation paleobiologists integrate their data into
marine conservation practice. (1) Conservation paleobiologists must improve
their communication with marine conservation biologists inside and outside of
academia. (2) One of the most promising areas for integration is investigating
climate change and its ecological implications. (3) The types of long-term data
that marine conservation biologists want and need are deliverables conservation
paleobiologists can provide. (4) Conservation paleobiologists must be proactive in
addressing the barriers that hinder the application of long-term data in conservation
practice.
Keywords Conservation barriers · Conservation paleobiology ·

Cross-disciplinary communication · Environmental stressors · Geohistorical
data · Survey
1 What is “Long Term”?
In recent years, long-term data (LTD) have been increasingly recognized for their
valuable contributions to improving the understanding of ecology, environmental
change, ecosystem service management, and conservation of biodiversity (Strayer
et al. 1986; Elliott 1990; Starzomski et al. 2004; Lindenmayer et al. 2012; Krebs
2015; Reed et al. 2016). In particular, LTD can help to elucidate the many important
ecological processes that occur slowly or infrequently, are temporally variable,
are dependent on historical events, and exhibit time lags (Strayer et al. 1986;
Davis 1989; Likens 1989; Carpenter and Turner 2001; Starzomski et al. 2004;
Lindenmayer et al. 2012). Even so, there is no consensus on how to define “long
term.” No fewer than seven definitions exist (Starzomski et al. 2004), with the most
simplistic among them defining the term as a distinct period of time (e.g., >10 years;
Lindenmayer et al. 2012). Others, however, have applied more restrictive, context-
specific definitions (Strayer et al. 1986; Starzomski et al. 2004; Lindenmayer et al.
2012). For instance, Franklin et al. (1990) considered long-term to be the recurrence
interval of a disturbance (e.g., floods, wildfires, volcanic eruptions) whereas others
have used the organism-specific scale of generation time, which can vary from days
for a bacterial community to centuries for trees in a forest (Strayer et al. 1986;
Carpenter and Turner 2001; Lindenmayer et al. 2012). Although a single definition
is likely overly restrictive, it is critical to communicate clearly what is meant by
“long term” between studies and, more importantly, between disciplines, as the
future success of restoration and conservation efforts likely lies in our ability to
assess and integrate information across multiple timescales (Starzomski et al. 2004).
No larger discrepancy exists in the definition of “long term” than that between
neontological and geohistorical studies. In a geohistorical sense, “long term” may
mean millions of years (Dietl et al. 2015) whereas neontologically the term is rarely
Conceptions of Long-Term Data Among Marine Conservation Biologists. . . 25
used to imply temporal scales beyond years or decades (Lindenmayer et al. 2012).
Although rarely given equal consideration to neontological data, when geohistorical
data are used they can greatly improve our understanding of long-acting processes
and dynamics (Dietl et al. 2015). For example, in understanding the effects of
climate change, geohistorical data have played a central role in understanding the
changes that have occurred during the last century (Dietl et al. 2015), including
biotic responses to climate change (e.g., changes to population size, de Bruyn et al.
2009; range shifts, Dawson et al. 2011).
Despite this potential, the integration of geohistorical data into conservation
practice has been slow (Fritz et al. 2013; Wolkovich et al. 2014; Holmes
2015), even with calls for integration from paleontologists and neontologists
alike (Dietl et al. 2015; Price and Schmitz 2016; and references therein).
For instance, even as geohistorical case studies are becoming more common,
Reed et al. (2016) reported a general paucity of ecological data on the
implications of climate change. That is, ecological data predating recent climate
change are generally unavailable from traditional sources (e.g., ecosystem
monitoring, experimental studies). This disjunction is driven by various
factors including a proliferation of discipline-specific jargon, a general lack
of interaction between researchers in different disciplines, and, traditionally,
a tendency in each discipline to measure different variables at timescales
that may differ by orders of magnitude (Polly et al. 2011; Fritz et al. 2013;
Dietl et al. 2015; Price and Schmitz 2016). These barriers are beginning
to break down, but much work remains to achieve full cross-disciplinary
integration.
Here we report the results of a survey of marine conservation biologists that
we conducted in an attempt to quantify and qualify their conception of “long
term” and to identify opportunities for collaboration between marine conservation
biologists and conservation paleobiologists. We wanted to know: How is “long
term” defined within the marine conservation community? What types of long-
term records do marine conservation biologists use? And, what barriers limit the
application of LTD in conservation practice? Addressing questions such as these is
important for enabling conservation paleobiologists to identify types of long-term
data that marine conservation biologists need and to develop corresponding outreach
and research directions in conservation paleobiology that are likely to improve the
integration of geohistorical and long-term ecological data into ongoing conservation
practices.
2 Survey Implementation
The web survey (Appendix 1) was open for responses during three periods.
The initial solicitation (September–November 2015) targeted marine laboratories
and researchers in the United States with academic, governmental, and non-
governmental affiliations. In response to the distribution of respondents from the

first solicitation, which included few academics, two additional solicitations were
sent out during January 2016 and April 2016, to the Ecological Society of America
listserv (ECOLOG-L) and to members of the National Association of Marine
Laboratories, respectively, to reach more academics (see Appendix 2 for a detailed
description of survey population selection and distribution). We report the combined
results of all three solicitations.
The survey consisted of 23 questions on six broad topics: demographics, conser-
vation goals, long-term data, environmental stressors, baselines, and the challenges
of applying LTD in marine conservation. Several question types were used, ranging
from multiple-choice to short written responses. Survey responses were categorized
for each question and all three authors reviewed each of the categorizations (see
Appendix 3 for details on response categorization). Analyses using subsets of the
data (e.g., gender, workplace) were not conducted due to small sample sizes. A table
of the survey responses can be downloaded at http://doi.org/10.7298/X4VM4965.
3 Survey Responses and What They Mean for Conservation

Paleobiologists
A total of 90 marine conservation researchers and professionals completed the

survey and they represent a broad cross-section of the marine conservation biology
community in the United States.1 The respondent population consisted of a roughly
equal number of men (n D 46, 51%) and women (n D 39, 43%; Fig. 1a) and
the majority of respondents identified as white (n D 78, 87%; Fig. 1b). Most
respondents worked in Governmental (n D 39, 43%) or Academic (n D 31;
34%; Fig. 1c) positions and have graduate degrees (n D 64; 71%; Fig. 1d). All
respondents who indicated their level of education completed at least 1 year of
college (n D 85, 94%; Fig. 1d). Respondents identified with fields such as Fisheries
1
The results presented here are focused on the entire population of survey respondents and do not
consider subsets that may lead to subtle differences in conservation ideologies and predispositions,
such as workplace (Braunisch et al. 2012; Laurance et al. 2012; Cook et al. 2013; Pietri et al.
2013) and gender (Kellert and Berry 1987; Czech et al. 2001; Dougherty et al. 2003; Bremner and
Park 2007; Mobley and Kilbourne 2013). Preliminary analyses show minor differences between
subsets, but these differences were not statistically comparable due to small sample sizes. When
differences did occur, they were relatively minor and did not change the interpretation of the survey
results as a whole. For instance, respondents identifying their workplace as Governmental tended
to select shorter timescales (e.g., months) for LTD compared to those who selected Academic (e.g.,
millennia), but both groups chose the decadal scale most often. Similarly, when asked to rank the
importance of environmental stressors, women were more likely to give individual stressors higher
importance ranks than men, but both genders agreed on the overall order of importance.
Fig. 1 (a) Respondent gender. (b) Respondent ethnicity. (c) Respondent workplace. (d) Highest
level of education completed by respondents. (e) Respondents’ fields of study; FS Fisheries
science, MB Marine biology, CES Conservation and environmental sciences, EAS Earth and
atmospheric sciences. Values in parentheses in (a) and (b) equal number of responses
science (n D 51; 22%) and Marine biology (n D 66; 29%; Fig. 1e) and filled a variety
of conservation-related roles, ranging from Research (n D 49; 46%) to Resource
management (n D 21, 20%) and project Administration (n D 8, 7%; Fig. 2a). The
survey population also included professionals, across all levels of experience (e.g.,
<5 years, n D 20, 22%; >25 years, n D 22, 24%; Fig. 2b), whose work focused on
a range of levels of biological organization (e.g., ecosystem, n D 36, 40%; Fig. 2c).
Specific responses were variable but, in general, respondents agreed on what
overarching goals drive marine conservation biology, what constitutes “long term”,
the importance of LTD, how to use LTD, and the challenges faced when applying
LTD. Perhaps unsurprisingly, conservation paleobiology parallels this consensus in
many ways, and the survey results highlight several opportunities for conservation
paleobiologists to enhance and expand on the application of LTD in marine
conservation.
Fig. 2 (a) Respondents’ conservation related roles within their employing organization; MC
Management (conservation), MR Management (resources). Percentages do not sum to 100% due to
rounding. (b) Years of experience working in marine conservation biology. (c) Level of biological
organization at which respondents work or study
Fig. 3 (a) Types of marine conservation biology goals identified by respondents; MBD Mainte-
nance of biodiversity, MES Maintenance of ecosystem services, MESF Maintenance of ecosystem
structure and function, HP Habitat protection, SUS Sustainability. (b) Categories of approaches
described by respondents to achieve their conservation goals. Percentages do not sum to 100% due
to rounding
Conservation Goals
The two conservation goals most commonly mentioned by survey respondents were
the Maintenance of biodiversity (n D 62, 26%) and Maintenance of ecosystem
services (n D 51, 21%; Fig. 3a). These goals are well matched by the overarching
goal of conservation paleobiology, which is to conserve and restore biodiversity and
ecosystem services (Dietl et al. 2015; Dietl 2016). Additionally, geohistorical data
have been applied in attempts to achieve the remaining three types of conservation
goals mentioned by survey respondents: Maintenance of ecosystem structure and
function (n D 34, 14%; e.g., Smith et al. 2016), Protection or restoration of
habitats (n D 18, 7%; e.g., Volety et al. 2009), and Sustainability (n D 31, 13%;
e.g., Pandolfi et al. 2003). The prevalence of these types of conservation goals in
responses highlights the shared principles and philosophical underpinnings of the
two disciplines (Dietl 2016) and should form the basis for future collaboration.
To achieve their conservation goals, respondents indicated several types of
approaches that they apply, including Management (n D 67, 39%; e.g., marine
protected areas), Technology (n D 46, 27%; e.g., satellite tagging), Mathematics

(n D 20, 12%; e.g., statistical and modeling approaches), Research (n D 23,
13%; e.g., understanding ecosystem processes), and Other (n D 17, 10%; e.g.,
education; Fig. 3b). In several instances and for different types of approaches (e.g.,
Management, Mathematics, Research), respondents discussed cross-disciplinary
collaboration as an integral component in achieving their conservation goals
(n D 12, 7%). Similarly, regardless of the type of approach, geohistorical data were
rarely mentioned (n D 5, 3%). Surprisingly, some novel conservation practices for
which geohistorical data have been useful were not mentioned by respondents (e.g.,
assisted migration, Marris 2011; rewilding, Burney et al. 2012), perhaps indicating
a lack of general awareness, acceptance, or implementation of these ideas in the
marine conservation community.
Long-Term Data
There is no doubt that the marine conservation biologists we surveyed consider LTD
to be important, particularly as the majority of respondents (n D 60, 67%) indicated
that they use LTD compared to only nine (10%) who do not (21 respondents,
23%, did not indicate whether they use LTD). In addition to agreement that LTD
are “essential,” “vital,” and “incredibly important,” respondents most commonly
indicated that “long term” refers to the decadal scale (n D 69, 49%; Fig. 4a).
Relatively few respondents selected timescales of centuries (n D 21, 15%) or
longer (n D 18, 13%) when asked to define “long term” (Fig. 4a), confirming
the documented difference in conceptions of the term between neontological and
geohistorical disciplines (e.g., Lindenmayer et al. 2012; Dietl et al. 2015). Some
respondents did indicate that longer-term temporal data (i.e., beyond decadal
timescales) could be useful, particularly when considering humans and their effects
on ecosystems. For example, one respondent stated, “ : : : to set goals about the future
state of the ecosystem, we need to be aware of conditions of the past (i.e., before
industrialization, overfishing activities, etc.).” Another respondent added: “Marine
ecosystems work on different time scales than human systems : : : ” Given these
responses, and others like them, marine conservation professionals may be receptive
to the integration of geohistorical data (Durham and Dietl 2015), particularly as they
are often the only source of perspectives beyond the limited time frame of direct
human observation (NRC 2005; Dietl et al. 2015).
The distribution of timescales on which respondents considered their sources
of LTD to be useful in conservation practice—primarily decades (n D 204; 34%),
years (n D 157, 26%), and months (n D 68, 11%; Fig. 4b)—suggests an opening
for the integration of geohistorical data. Sources of data on the scale of centuries
(n D 63, 10%) and longer (n D 47, 8%; Fig. 4b) were rare. Furthermore, the
Fig. 4 (a) Histogram of respondents’ selections of timescales to which the term “long term”
applies. (b) Histogram of respondents’ selections of timescales on which their sources of long-
term data can provide information. Percentages do not sum to 100% due to rounding. (c) Types of
long-term data sources mentioned by respondents
distribution of timescales on which LTD sources were considered by respondents

to be informative (Fig. 4b) was shifted towards shorter timescales relative to how
they defined “long term” (Fig. 4a). This discrepancy is likely due to the types of
data being used. Most LTD sources identified by respondents were categorized
as Modern observational (n D 142; 49%; e.g., monitoring data) as opposed to
Historical (n D 23; 8%; e.g., photographs) or Geohistorical (n D 20; 7%; Fig. 4c;
e.g., sediment cores).2 It is not surprising that the sources of LTD familiar to marine
conservation biologists are limited to timescales of decades or less, particularly
given the challenges associated with maintaining real-time observational studies
(e.g., funding, data continuity; Lindenmayer et al. 2012) and the general paucity
of long-term studies in ecology prior to the 1980s (Strayer et al. 1986; Likens 1989;
Franklin et al. 1990). Thus, “questions about longer time scales must be approached
by means other than direct long-term studies” (Strayer et al. 1986, p. 4). What
does not seem to be generally realized in the marine conservation community is
that the application of geohistorical data is one such approach (NRC 2005; Dietl et
al. 2015; Durham and Dietl 2015). Indeed, the applications of LTD (regardless of
definition) identified by respondents—setting Baselines (n D 50, 52%), establishing
the Natural range of variability (n D 19, 20%), and detecting Trends and patterns
(n D 25, 26%; Fig. 5a)—are well-aligned with those identified by conservation
paleobiologists (Dietl et al. 2015). The complementary nature of our methods for
obtaining and using LTD should provide fertile grounds for collaboration.
Conservation paleobiologists should be further encouraged because the types
of LTD that marine conservationists want, but do not currently have access to,
including Abiotic (n D 15, 21%) and Biotic (n D 38, 54; Fig. 5b) variables, map
neatly onto the types of geohistorical data that already exist and continue to be
collected. Considering mollusks, for example, sclerochronological techniques can
generate data on Abiotic variables such as Temperature (n D 3, 20%; e.g., Goodwin
et al. 2001) and Water chemistry (n D 10, 67%; e.g., Cintra-Buenrostro et al. 2012;
Fig. 5c), as well as on Biotic variables including species’ Distributions (n D 11,
29%; e.g., Smith and Dietl 2016) and Interactions (n D 4, 11%; e.g., Cintra-
Buenrostro et al. 2005; Fig. 5d). Separately, or in combination, these data can be
applied to assess human impact on ecosystems. For instance, geohistorical evidence
of millennial-scale variability in predator-prey interactions suggested that fishing
pressure pre-dated pollution stress in benthic ecosystems of Long Island Sound,
USA, indicating that ongoing efforts to restore these ecosystems by addressing
nitrogen pollution alone may not succeed if harvest pressure is not also further
controlled (Casey et al. 2014).
2
Three of the four major types of data sources identified in marine historical ecology (sensu
Lotze and McClenachan 2014; Jackson and McClenachan 2017), a sister field of conservation
paleobiology, were also conspicuously absent. Data types largely absent were geological (e.g.,
sediment cores), archaeological (e.g., middens), and historical narrative (e.g., accounts of explor-
ers), whereas the fourth, modern scientific and fisheries data (i.e., Modern observational data of
this study), was mentioned commonly.
b c
Fig. 5 (a) Types of reasons given by respondents in describing how long-term data are used. (b)
Types of data that respondents would like more of or that are not currently available. (c) Breakdown
of Abiotic data types from 5b. (d) Breakdown of Biotic data types from 5b. Values in parentheses
in (b), (c), and (d) equal number of responses
Environmental Stressors
When respondents were asked to rank the five major stressors identified in the
2005 Millennium Ecosystem Assessment report (pollution, habitat change, climate
change, overexploitation, and invasive species; MEA 2005) in order of importance,
they most often gave climate change a rank of one (highest importance; n D 36 #1
ranks, mean D 2.18) and also ranked habitat change as highly important (n D 32
#1 ranks, mean D 2.27; Table 1). The reported distribution of stressor importance
matches well with the distribution of respondents’ inclinations for how the LTD
sources they provided can be used to address the stressors (Fig. 6a). Whereas
Fig. 6 (a) Number of long-term data sources mentioned by respondents that they believe can
be applied to each of the environmental stressors identified in the 2005 Millennium Ecosystem
Assessment Report (MEA 2005). (b) Distribution of timescales identified by respondents as
necessary to address each environmental stressor
respondents indicated that data on timescales of decades or shorter would suffice

to understand the effects of pollution, habitat change, overexploitation, and invasive
species, they recognized that data on timescales of centuries, millennia, or longer
are needed to address climate change (Fig. 6b). For pollution, habitat change,
overexploitation, and invasive species, the responses approximated a bell curve with
a median between years and decades. With climate change responses, however, the
median was between decades and centuries and the distribution was skewed towards
longer timescales (e.g., thousands to millions of years; Fig. 6b).
The shape of the climate change distribution suggests that marine conservation
biologists may welcome data on longer timescales (i.e., centuries, millennia) to
address climate change, particularly because the majority of LTD sources provided
by respondents were considered to be informative on shorter timescales (i.e.,
decades or less; Fig. 4b). In light of the importance placed on climate change by
respondents (Table 1), the perceived need for data on timescales longer than decades
to address biotic responses to climate change (Fig. 6b), and the limited number of
LTD sources provided by respondents for timescales greater than decades (Fig. 4b),
Table 1 Respondent ranks 1 2 3 4 5 Unsure Mean

of environmental stressor
importance (1 D most Pollution 19 25 19 11 8 1 2.60
important, 5 D least Habitat change 32 20 14 8 8 0 2.27
important) Climate change 36 21 9 9 8 0 2.18
Overexploitation 27 19 13 10 12 2 2.60
Invasive species 6 28 23 16 9 1 2.96
conservation paleobiologists have an immediate and unmistakable opportunity to

provide highly relevant geohistorical data (MacDonald et al. 2008; Willis et al.
2010). Unfortunately, based on their descriptions of the timescales of data needed
to address pollution, habitat change, overexploitation, and invasive species (i.e.,
predominantly decades or shorter; Fig. 6b), respondents were seemingly unaware of
the valuable contributions that geohistorical data can make to understanding their
effects (Dietl and Flessa 2011; Dietl et al. 2015; and references therein). Thus,
the utility of geohistorical data for helping to address these stressors may be more
difficult for marine conservation biologists to accept.
Emphasizing the interactions between climate change and the other four stressors
may be a viable path to integrate geohistorical data on all of the stressors
because respondents did highlight the importance of interactions between each
of these stressors and climate change (Table 2). For example, one respondent
stated: “All of these problems are difficult to repair because we do not understand
the interactions of multiple drivers and stressors at the ecosystem level.” Hence,
because climate change is a long-acting process, data must be from equally long
timescales to understand these complex interactions (Dietl et al. 2015). In particular,
differentiating between causal factors (e.g., human versus natural) is imperative
for reaching conservation solutions, especially given the high incidence of non-
additive relationships (e.g., synergistic, antagonistic) between stressors (Crain et
al. 2008; Darling and Côté 2008). Each interacting stressor can complicate matters
further, as a meta-analysis by Crain et al. (2008) revealed that the addition of
a third stressor can double the number of synergistic interactions. As such, the
rapidly changing climate we now face can amplify the effects of the other stressors,
particularly as we move towards an uncertain future with novel ecosystems (Willis
et al. 2010). For example, the poleward spread of nonnative pests and pathogens
from the tropics (Bebber et al. 2013) may be exacerbated by a warming climate,
just as has been documented during climate shifts in the deep past (Labandeira
and Currano 2013). Thus, when interacting with marine conservation biologists,
conservation paleobiologists should emphasize that it is not only data on climate
change and its biotic effects that are required but also data on the remaining four
stressors, especially on the effects of those stressors over timescales (i.e., centuries
or longer) that are comparable to the temporal footprint of climate change.
Table 2 Total number of mentions for each environmental stressor in interactions and number of
mentions with other stressorsa
Habitat Climate Invasive
Total mentions (%) change change Overexploitation species
Pollution 29 (16) 25 25 21 17
Habitat change 47 (26) – 39 21 20
Climate change 51 (28) – – 30 18
Overexploitation 33 (18) – – – 17
Invasive species 19 (11) – – – –
a
In 16 instances, respondents selected all five environmental stressors
Baselines
The majority of survey respondents gave equal weight to spatial and temporal
data for use in setting baselines (n D 43, 66%; Fig. 7a), which is encouraging
for conservation paleobiology, particularly as respondents who favored one type
of data over the other placed higher value on temporal (n D 9, 75%) rather than
spatial data (n D 3, 25%; Fig. 7b). Even so, many of the baselines they described
were from Modern observational data (n D 65, 42%; e.g., monitoring data) and
Reference sites (n D 32, 21%; e.g., “pristine” reference sites) rather than Historical
(n D 25, 16%; e.g., written records) or Geohistorical (n D 7, 5%; e.g., death
assemblages) sources (Fig. 7c). Given the general consensus that pristine habitats
no longer exist (Jackson et al. 2012; Wiens and Hobbs 2015), this practice may be
problematic as the comparisons being made do not reflect “healthy” systems and are
subject to short-term variation (Kopf et al. 2015). Thus, conservation paleobiologists
have an opportunity to provide temporal baselines from geohistorical data (Dietl et
al. 2015; Dietl et al. 2016), which can be valuable for smoothing out the short-
term variation that is intrinsic to “snapshot” modern surveys (Kowalewski et al.
1998). For instance, the time-averaging and taphonomic inertia inherent to death
assemblages in live-dead studies (sensu Kidwell 2007) provide a means to collect
data on average conditions over decades or centuries, rather than days or weeks
(Kidwell and Tomašových 2013). Geohistorical records can undoubtedly provide
baseline data for marine conservation (MacDonald et al. 2008; Willis et al. 2010;
Dietl et al. 2015), but conservation paleobiologists should not limit themselves to
static baselines concerned with past species diversity and abundance.
In application, baselines are used to define targets for restoration actions;
they are the “natural” set of conditions on which goals are established. Many
baselines tend to be static and habitually fail to consider humans as a part
of the ecosystem (Kopf et al. 2015). As aptly described by one respondent:
“Conservation often aims at retaining the status quo or restoring some historic
state : : : this mindset is problematic : : : reaching an idealized historic state may be
impractical or impossible” (italics added). Kopf et al. (2015, p. 803) suggested
Anthropocene baselines—“a concept to represent an ecological and theoretical shift
from a fixed preindustrial-era reference condition to a dynamic point of reference
a b
Fig. 7 (a) Number of respondents disagreeing (No), agreeing (Yes), or conditionally agreeing
(Depends) on the statement that spatial and temporal data have equal value in establishing baseline
data or reference conditions. (b) Proportion of data types that respondents who disagreed in 7a
believe are more valuable for establishing baseline data or reference conditions. (c) Types of data
used by respondents to establish baselines. Values in parentheses in (a) and (b) equal number of
responses
for human-dominated ecosystems”—as a solution to this problematic mindset.

Although this concept may prove to be useful in some management situations,
it is reliant on subjective judgment and requires practitioners to recognize the
threshold at which ecosystems shift irreversibly to new conditions (i.e., regime
shifts). The Anthropocene baselines concept does incorporate human influences
and emphasizes a triage approach to conserving biodiversity but the new baselines
remain equally static to those from the pre-industrial era. And, for better or worse,
many managers and researchers already rely on modern baselines that are human-
influenced (Fig. 7c).
Instead, the geohistorical record can provide data on ecosystem function and
structure (Eronen et al. 2010; Polly et al. 2011; Fritz et al. 2013; Dietl and Flessa
2017), the maintenance of which is an increasingly important goal for restoration
(see Price and Schmitz 2016) and one that was recognized by survey respondents
(Fig. 3a). A function-based restoration approach does not discount the importance
of biodiversity but instead shifts the focus to traits with direct links to interactions
between the abiotic and biotic components of the ecosystem—that is, using an
ecometric approach (Polly et al. 2011)—with less emphasis on the presence or
absence of specific species (e.g., Dietl et al. 2016). This shift is achieved by
studying the form and function of organisms, particularly as they combine to
establish functional diversity, which ultimately controls ecosystem processes and
services (Price and Schmitz 2016; Dietl and Flessa 2017). For instance, as climate
change induces the formation of novel communities and ecosystems that humans
have not experienced (MacDonald et al. 2008; Willis et al. 2010), a trait-based,
integrative approach will be needed in order to understand ongoing changes in
ecosystem structure, health, and processes and to anticipate future changes during
the Anthropocene (Starzomski et al. 2004; Eronen et al. 2010; Polly et al. 2011;
Fritz et al. 2013; Price and Schmitz 2016; Dietl and Flessa 2017). Conservation
paleobiologists can help establish the basis for these evaluations.
Challenges
The application of LTD to conservation practice is challenging and the predominant

barriers to the utilization of these data that were identified by respondents (Funding,
n D 48, 36%; Data availability, n D 38, 28%; Communication, n D 14, 10%; and
Institutional, n D 13, 10%; Fig. 8) are largely unchanged from those acknowledged
decades ago (see Strayer et al. 1986). Even with the establishment of the Long
Term Ecological Research program by the National Science Foundation in 1980,
these challenges persist. As one respondent explained, “In the project that I work on
where we have high quality data sets going back nearly 40 years, we are regularly
having to partially refund [sic] the project from new funding sources. When you
have a recognized long term data collection program that has these struggles, it
is no wonder that there is a paucity of good long term databases directly linked to
marine conservation.” The utilization of geohistorical data may help overcome these
barriers to applying LTD in marine conservation practice.
One of the prominent institutional barriers for marine conservation biologists
is the necessity to maintain studies for years or decades to obtain LTD because,
as pragmatically described by one respondent, “ : : : it often takes a long time to
yield meaningful results.” Such studies produce LTD that are valuable, but their
longevity requires dedicated leadership, consistent funding, and, not least of all,
some good luck (Strayer et al. 1986; Likens 1989; Lindenmayer et al. 2012). The
substantial time and effort it takes to collect LTD in real time is a serious obstacle
for marine conservation biologists. In particular, several respondents believed
that LTD collection is hindered by the perception that “[long-term monitoring]
isn’t as sexy as some cutting edge work or new venture.” Consequently, the
relatively rapid and inexpensive assessments (e.g., retrospective studies, Davis 1989;
Fig. 8 Types of challenges described by respondents for applying long-term data in conservation
practice
live-dead studies, Kidwell 2007) that are possible using geohistorical records may
be a useful and resonant selling point for conservation paleobiology. In the right
settings (e.g., coastal marine environments), geohistorical records are essentially
available on demand. Whereas it is not feasible to monitor an entire coastline, a
retrospective geohistorical approach can often provide valuable, location-specific
data at relatively low costs (e.g., Dietl and Smith 2016).
Large potential and low costs of geohistorical data notwithstanding, it remains a
considerable challenge to secure funding for conservation (Katsanevakis et al. 2015;
Fig. 8). Indeed, one respondent noted, “there is a general lack of socio-political
will to invest in long-term monitoring, conservation, and management : : : budgets
associated with science and research, particularly climate research, continue to
be cut : : : ” Given this troubling trend in funding (see also Kearney et al. 2014;
Gauchat 2015; Nadelson and Hardy 2015; Besley 2016; Farrell 2016), conservation
science needs innovative solutions, which may be found in cross-disciplinary
collaborations (Boulton et al. 2005; Price and Schmitz 2016; Kelley et al. 2018).
Marine conservation biologists may not, however, immediately recognize the value
of collaborations with conservation paleobiologists, particularly given their lack of
awareness of the contributions geohistorical data can make to conservation practice
(e.g., on-demand data for all environmental stressors, reduced funding burden).
Thus, it will likely be up to conservation paleobiologists to initiate these cross-
disciplinary collaborations. In doing so, we should emphasize that, by working
together with LTD on complementary timescales, marine conservation biologists
may improve their chances of obtaining funding (particularly if they can achieve
their conservation goals at lower cost) and, in doing so, simultaneously reduce the
previously mentioned institutional barriers (CPW 2012; Boyer et al. 2017; Flessa
2017).
Another barrier to the application of LTD in marine conservation that was

discussed by respondents, and applies equally to conservation paleobiology, is the
“research-implementation gap” (sensu Arlettaz et al. 2010)—the fact that ideas,
data, and techniques generated in the research community are often not applied
in conservation practice without explicit outreach efforts from researchers. That is,
publishing data in peer-reviewed articles and theorizing within academia are not
sufficient to impact conservation practice (Arlettaz et al. 2010; Braunisch et al.
2012; Cook et al. 2013; Pietri et al. 2013; Roux et al. 2015; Flessa 2017; Kelley
et al. 2018), or, as described by one respondent, there is a need to be better about
“ : : : spreading word of the data being collected and where it can be obtained.” This
communication barrier is symptomatic of conservation biology as a whole and must
be acknowledged by conservation paleobiologists as they enter the applied realm
of conservation practice. Rather than wait and hope for practitioners to discover
our data, which is uncommon (Braunisch et al. 2012; Griffiths and Dos Santos
2012; Laurance et al. 2012; Boyer et al. 2017), conservation paleobiologists should
proactively seek out practitioners as they build cross-disciplinary collaborations
and ask what kinds of data are needed (Flessa 2017). Opening this dialogue will
allow for geohistorical studies guided by conservation-specific research questions
that generate data practitioners will apply.
4 Takeaways for Conservation Paleobiologists
The responses to our survey confirm that “long term” has a different meaning for
marine conservation biologists and conservation paleobiologists. Respondents most
commonly chose the decadal scale to define “long term” whereas many conservation
paleobiologists apply the term to much longer timescales (e.g., centuries and longer;
Dietl et al. 2015). Despite this difference, our survey results confirm that our fields
are united by the goals that drive them (Dietl 2016) and nearly identical conceptions
of the utility of LTD, if not its definition. Adopting a cross-disciplinary approach
by integrating data from our complementary sources and timescales is feasible and
stands to help us achieve our common goals. We highlight four takeaways from the
results of our survey that can help conservation paleobiologists work towards this
integration.
1. Conservation paleobiologists must improve their communication with marine
conservation biologists inside and outside of academia. Respondents rarely
mentioned geohistorical data, and were seemingly unaware of their potential to
contribute to conservation practice despite the growing number of case studies
in conservation paleobiology. Furthermore, respondents indicated that data on
decadal timescales are needed to address most ecosystem stressors and the
most frequently mentioned barrier to the application of LTD was a lack of
funding (often discussed in reference to maintaining costly long-term monitoring
studies). This pattern of responses indicates that either marine conservation
biologists are not aware of geohistorical data, or that they do not consider them
important for addressing conservation problems.
2. One of the most promising areas for integration is investigating climate change
and its ecological implications. Although they did not mention geohistorical
data, many respondents did recognize that in order to understand climate
change, and its effects on ecosystems worldwide, they need data on centennial
to millennial scales. Given that their responses also suggested they do not
currently have data on these longer timescales, this may be one focus of
marine conservation biology where the geohistorical perspectives offered by
conservation paleobiology would be immediately welcome.
3. The types of LTD that marine conservation biologists want and need are deliv-
erables conservation paleobiologists can provide. Respondents wanted more
LTD on temperature, water chemistry, and species abundances, distributions,
and interactions. These data are available from marine geohistorical records,
so broadcasting our ability to gather them may be beneficial for establishing
collaborations with marine conservation biologists.
4. Conservation paleobiologists can help address the barriers that hinder the appli-
cation of LTD in conservation practice. Retrospectively providing LTD on biotic
and abiotic variables can help alleviate the challenges with data availability,
funding, and institutional barriers that were identified by respondents. Whereas
marine conservation biologists often must maintain studies for years or decades
to build a long-term dataset, conservation paleobiologists may be able to provide
suitable data at considerably lower costs and without the need to gather the data
in real time.
5 Moving Forward
Our survey results suggest conservation paleobiologists have much work to do

before geohistorical data are regularly accepted and applied in marine conservation
biology. We were startled by the lack of awareness of geohistorical data, and their
applications in conservation practice, demonstrated by survey respondents. Despite
the continued growth of conservation paleobiology as a field over the last decade,
our outreach to the marine conservation biology community so far has not been
enough. We must redouble our efforts to integrate our data into conservation practice
to achieve the full potential of our field. The onus is on conservation paleobiologists
to demonstrate the utility of geohistorical data.
Acknowledgments We would like to thank the editors, Carrie Tyler and Chris Schneider, for their
invitation to participate in this volume and two reviewers, Michelle Casey and Michael Savarese,
whose comments improved the manuscript. We also thank all of those who assisted with survey
distribution and those who completed the survey.
Appendix 1: Survey Questions
Demographics and Professional Information

1. With which race/ethnicity do you identify?
– White
– Black
– Asian
– Hispanic
– Other: _____________
2. With which gender do you identify?
– Male
– Female
– Other: _____________
3. Please list up to three fields (e.g., fisheries biology, historical ecology, etc.) with
which you identify.
4. Which of the following best describes your workplace?
– Government
– Nongovernmental organization
– Academia
– Other:___________
5. Which of the following best describes your highest completed level of education?
a. Doctorate
b. Master’s
c. Bachelor’s
d. Other: _____________
6. How many years of experience do you have in marine conservation?
– <5
– 5–10
– 10–15
– 15–20
– 20–25
– >25
7. Please describe your work as it relates to marine conservation in one sentence or
less.
8. At what level of biological organization does your work primarily focus?

– Population
– Species
– Community
– Ecosystem
– Biome
– Other: _____________
Goals and Approaches
9. Please list up to three primary goals of the field of marine conservation (e.g.,
preservation of biodiversity) in your opinion.
10. What are the cutting-edge approaches currently being practiced in marine
conservation to achieve the goals you mentioned in the previous question?
Long-term Data: Definitions and Sources
11. In your opinion, to what timescales does the phrase “long-term data” typically
refer in the conservation community?
– Days
– Weeks
– Months
– Years
– Decades
– Centuries
– Thousands of years
– Tens of thousands of years
– Hundreds of thousands of years
– Millions of years
– Unsure
12. In your opinion, what is the importance of long-term data for achieving the
goals of marine conservation?
13. If you use long-term data, how do you use it and why do you use it? Or, if
long-term data are not considered in your work, why not?
14. Please list five sources of long-term data and indicate whether you have used
each one in your own research.
15. Considering the sources of long-term data you listed previously, at what time
scale(s) are data from these sources most useful?
Days Weeks Months Years Decades Centuries Millennia 104 years 105 years 106 C years Unsure
Source A [ ] [] [] [] [] [] [] [] [] [] []
Source B [ ] [] [] [] [] [] [] [] [] [] []
Source C [ ] [] [] [] [] [] [] [] [] [] []
Source D [ ] [] [] [] [] [] [] [] [] [] []
Source E [ ] [] [] [] [] [] [] [] [] [] []
Long-term Data and Ecological Stressors

16. The Millennium Ecosystem Assessment (2005) identified the following five
most-important stressors in ecosystems. Please rate each stressor’s importance
in marine conservation biology (one being highest importance and five being
lowest importance).
Importance
1 2 3 4 5 Unsure
Pollution [] [] [] [] [] []
Habitat change [] [] [] [] [] []
Climate change [] [] [] [] [] []
Overexploitation [] [] [] [] [] []
Invasive species [] [] [] [] [] []
17. Given that these stressors interact in complex ways, please identify and
briefly describe the interaction that is most pressing to understand in marine
conservation, in your opinion (e.g., the additive interaction between invasive
species and climate change).
18. Which of the long-term data sources that you identified previously do you
believe can be used to address the five stressors or their interactions?
A B C D E
Pollution [] [] [] [] []
Habitat change [] [] [] [] []
Climate change [] [] [] [] []
Overexploitation [] [] [] [] []
Invasive species [] [] [] [] []
Unsure [] [] [] [] []
19. Please select one or more timescales of data that would be needed to best
address each stressor, in your opinion.
Days Weeks Months Years Decades Centuries Millennia 104 years 105 years 106 Cyears Unsure
Pollution [] [] [] [] [] [] [] [] [] [] []
Habitat
change [] [] [] [] [] [] [] [] [] [] []
Climate
change [] [] [] [] [] [] [] [] [] [] []
Over-
exploi-
tation [] [] [] [] [] [] [] [] [] [] []
Invasive [] [] [] [] [] [] [] [] [] [] []
species
Temporal vs. Spatial Data

One important reason for using long-term temporal data is to produce baselines
against which current conditions in ecosystems can be compared, but spatial data
also are frequently used as references against which to judge current conditions at a
specific location. The following two questions are intended to help us understand the
balance between use of temporal and spatial data to produce baselines and reference
conditions in marine conservation.
20. If you use reference conditions or baselines in your research/conservation work,
please list three types of data sources that you use to produce them (e.g.,
reference sites, monitoring records, etc.).
21. In your opinion, are spatial and temporal data of equal value in establishing
reference conditions and baselines? Please explain briefly.
Problems and Challenges with Applying Long-term Data
22. Are there types of long-term data that would be useful, but that aren’t currently
available or of which you would want more? If so, please give an example.
23. What barriers (e.g., communication, funding, data availability, etc.) have you
experienced (or do you perceive to exist) in applying long-term data to marine
conservation?
Appendix 2: Survey Population Selection
First Solicitation
In order to establish our survey population, we searched the internet for organi-
zations conducting research or management in marine systems. All institutions,
agencies, laboratories, etc. were based in the United States and included National
Estuarine Research Reserves (e.g., Chesapeake Bay NERR), Sea Grant programs
(e.g., Alaska Sea Grant), governmental departments (e.g., Alaska Department of
Fish and Game) and their divisions (Division of Habitat), and academic marine
laboratories (e.g., Darling Marine Center, University of Maine). A full list of all
organizations contacted for the survey can be downloaded at http://doi.org/10.7298/
X4VM4965.
For each organization, we contacted the director, president, or positional equiva-
lent via email prior to the activation of the survey (n D 202). If we received a positive
response (agreement to distribute the survey) from an organization (n D 54), we sent
a solicitation to the contact upon activation of the survey. If we did not receive an
initial response (n D 136), we sent a second email to the contact with a solicitation
at the time of survey activation to encourage participation. We did not contact those
who responded negatively (n D 12) to the initial solicitation. The survey was open
September–November 2015.
Second Solicitation
We opened the survey a second time during January 2016. In this period, we sent a
solicitation to the Ecological Society of America listserv, ECOLOG-L, in an attempt
to reach marine conservation biologists who may not have been reached by our
first solicitation. ECOLOG-L is distributed internationally, however the solicitation
explicitly requested participation from researchers and managers working in the
United States of America.
Third Solicitation
During April 2016, we opened the survey for a third time, with a goal of increasing
participation from the academic demographic. We sent a solicitation to the President
of the National Association of Marine Laboratories (NAML) who kindly agreed to
distribute the survey to the directors of the member laboratories. NAML includes
governmental laboratories but its more than 50 members are primarily associated
with academia. A full list of all organizations contacted for the survey can
be downloaded at http://doi.org/10.7298/X4VM4965. Visit http://www.naml.org/
index.php for more information on NAML.
Appendix 3: Categorization of Responses
We categorized the survey responses for all free response and short answer questions
prior to analysis. For each question, all three authors reviewed each of the categories
to which responses were assigned and when disagreements occurred the categories
were discussed until a consensus was reached. Similarly, responses to free response
and short answer questions were reviewed collectively and placed within categories
after the authors reached agreement. Responses were categorized for the following
11 questions: 3, 7, 9, 10, 12, 14, 17, and 20–23.
Question 3: Please list up to three fields/scientific disciplines (e.g., fisheries
biology, historical ecology) with which you identify. Fields and disciplines listed
by respondents were grouped into five categories: Fisheries science, Marine biology,
Conservation and environmental sciences, Earth and atmospheric sciences, and
Other. The use of keywords by respondents facilitated this categorization. For

instance, any response including “fishery” (e.g., fisheries biology, fishery manage-
ment) was categorized as Fisheries science and any inclusion of “conservation”
or “restoration” (e.g., conservation biology, ecological restoration) was considered
Conservation and environmental sciences. Marine biology was applied generally
and was inclusive of responses such as “marine ecology” and “estuarine ecology”.
The Earth and atmospheric sciences category included responses such as “Geog-
raphy” and “Geology”. Responses grouped as Other included “biogeochemistry,”
“genetics,” and “molecular biology.”
Question 7: Please describe your work as it relates to marine conser-
vation in one sentence or less. Responses to this question were grouped into
six categories—Research, Education, Management (conservation), Management
(resources), Administration, and Other—and were not mutually exclusive. Many
respondents indicated that they conducted Research (e.g., “I am a marine ecolo-
gist studying : : : ”) and also filled Education (e.g., “Educating bay stewards”) or
Management roles. Keywords were particularly useful when distinguishing between
Management (conservation) and Management (resources). Responses were grouped
under Management (conservation) when the emphasis was on the preservation or
restoration of biodiversity or ecosystems (e.g., “Assess status and trends of ecosys-
tem health in our local estuaries”) whereas responses in the Management (resources)
category focused on ecosystem services and fisheries activities (e.g., “...implement
resource management actions : : : ”). Administration was differentiated from these
categories based on the level at which the respondent was working. For example,
“Chair of several science or technical advisory committees to coastal policy groups”
was considered Administration and “ : : : developing strategies to protect and restore
salmon habitat” was considered Management (resources). Responses categorized as
Other were generally too vague to fit any of the aforementioned categories (e.g.,
“Working hard today to ensure a better future tomorrow”).
Question 9: Please list up to three primary goals in the field of marine
conservation (e.g., preservation of biodiversity). Responses were grouped into
six categories: Maintenance of biodiversity, Maintenance of ecosystem services,
Maintenance of ecosystem structure and function, Habitat protection, Sustain-
ability, and Other. These groupings were not mutually exclusive, as responses
such as “Conservation of ecosystem function and services” were considered both
Maintenance of ecosystem services and Maintenance of ecosystem structure and
function. Maintenance of biodiversity, which was taken to be inclusive of all types
of diversity (e.g., genetic, species, ecosystem), ecosystem services, and structure and
function were distinguished based on phrasing and keywords used by respondents—
most prominently the category names themselves. Habitat protection was applied
in a general sense (i.e., not necessarily implying human exclusion from nature).
The Sustainability category included responses mentioning management practices
in a general sense (e.g., “smart management”) as well as responses that explicitly
mentioned sustainability (e.g., “long-term sustainability”). We acknowledge that the
concept of sustainability can be complicated (Callicott and Mumford 1997), but use
it here in a broad sense to mean the current and continued coexistence of humans
and the ecosystems in which they are embedded. Several respondents also included
educational goals (e.g., “education”), focused on research (e.g., “To study the impact
we have had : : : ”), or gave vague responses (e.g., “Understand marine ecosystems”);
these were considered Other.
Question 10: What are the cutting-edge approaches currently being prac-
ticed in marine conservation to achieve the goals you mentioned in question 9?
List no more than three. Responses to this question were variable and ranged from
data collection tools (e.g., “drones”) to practices (e.g., “adaptive management”) and
management actions (e.g., “marine protected areas”). Thus, responses were grouped
into the broad categories of Management, Technology, Mathematics, Research,
and Other. Many of the responses included multiple types of approaches and
others described approaches that spanned more than one of the approach categories
(e.g., the response, “ : : : statistical and modeling approaches combined with field
data from long term studies : : : ”, was categorized as Mathematics and Research).
Responses in the Management category included decision making (e.g., “utilization
of diverse data sets to make management decisions”) and management actions (e.g.,
“Marine Protected Areas”) as well as policy changes (e.g., “use laws and politics
to control the human activities”). Technology approaches referred to improving
(e.g., “greater computing power”) as well as adapting existing technology to
conservation practice (e.g., “use of drones”). Mathematics approaches were most
commonly related to improved modeling (e.g., “Modeling approaches combined
with community-based monitoring : : : ”) and analysis (e.g., “spatial analysis”). The
Research category primarily included descriptions of applying data to conservation
practice (e.g., “interdisciplinary collaborative research : : : studying how major river
freshwater plumes effect [sic] early life stage survival in marine environments”),
some more theoretical considerations (e.g., “ecosystem processes understanding”),
and citizen science (e.g., “Developing crowd-sources data [sic] and information
products”). The Other approaches included responses that were too broad to fit other
categories (e.g., “genetics”) or did not fit the previous categories (e.g., “education”).
Question 12: In your opinion, what is the importance of long-term temporal
data for achieving the goals of marine conservation? Responses to this question
were classified into one of three commonly described categories (Strayer et al. 1986;
Lindenmayer et al. 2012; Dietl et al. 2015)—Baselines, Trends and patterns, Range
of variability—and a fourth category, Other, for miscellaneous responses. In many
cases responses included components of multiple categories and were tallied in
each of those categories. Responses in the Baseline category typically referred to
using LTD to inform decision making in the future (e.g., “To combine with known
conditions to be able to model and predict future outcomes”). Responses classified
as Trends and patterns implied that LTD are important for determining trajectories
and removing short-term variation (e.g., “identifies long-term trends in populations
or water quality. Eliminates the noise of year-to-year variation : : : ”). Range of
variability most commonly included responses that highlighted the dynamic nature
of populations and ecosystems (e.g., “critical for detecting natural dynamics of
ecosystems : : : ”). The vast majority of responses fell in one of these three categories
and two remaining responses were grouped as Other (e.g., “Convincing policy
makers : : : ”).
Question 14: Please list five sources of long-term data and indicate whether
you have used each one in your own research. The respondent-provided sources of
long-term data were grouped into four categories, Modern observational, Historical,
Geohistorical, and Other, related to those described for sources of data in marine
historical ecology (Lotze and McClenachan 2014; Jackson and McClenachan 2017).
In marine historical ecology, “archaeological” is given equivalent status as a data
source, however, here it was subsumed under Geohistorical due to similarities
in timescales and the small number of responses including these data. Modern
observational included monitoring data and any contemporaneously collected data
such as “seabird productivity data,” “Weather station data,” and “fishery catch data.”
Historical (e.g., “historical documents”) was distinguished from Geohistorical (e.g.,
“Paleontological”) by its association with records kept by people (e.g., “historical
documents”), as opposed to records in nature (e.g., “sediment cores”). The Other
category included various responses including organizations (e.g., “NOAA”) and
variables (e.g., “pH”) that were too broad to categorize otherwise.
Question 17: Given that these stressors interact in complex ways, please
identify and describe the interaction that is most pressing to understand in
marine conservation, in your opinion (e.g., the additive interaction between
invasive species and climate change)? In 2005, the Millennium Ecosystem
Assessment identified five stressors—pollution, habitat change, climate change,
overexploitation, and invasive species—as the most important threats to ecosystems
and it has subsequently been noted that these stressors often interact in complex
ways (Crain et al. 2008; Darling and Côté 2008). Many respondents identified
multiple interactions or interactions between three or more stressors they found to
be important. Consequently, responses to this question were assessed in two ways.
First, the total number of mentions for each stressor was tallied. Second, interactions
between stressors were tallied. When three or more stressors were mentioned, each
unique pairing was tallied (e.g., a respondent mentioning climate change, habitat
change, and pollution resulted in tallies for climate change-habitat change, climate
change-pollution, and habitat change-pollution).
Question 20: If you use reference conditions or baselines in your
research/conservation work, please list three types of data sources that
you use to produce them (e.g., references sites, monitoring records, etc.)?
Responses were categorized into five groups: Modern observational, Reference
sites, Historical, Geohistorical, and Other. These categories were chosen to reflect
those used in Question 14. Responses classified as Modern observational commonly
included mentions of monitoring (e.g., “monitoring records”). “Reference sites”
was also a frequently given response and formed the basis for the Reference
sites category; such responses were not considered Modern observational because
they implied a spatial component rather than continued observation at one or a
few sites. Similarly, responses in the Reference sites category were distinguished
from responses in the Historical and Geohistorical categories by the mention or

implication of spatial rather than temporal data. Historical included baselines from
human-produced sources including “Literature,” “historical data,” and “historical
accounts.” Responses in the Geohistorical category often included mentions of
paleontological data (e.g., “paleobiology”) and geological sources (e.g., “sediment
cores”). The Other category included responses giving methods (e.g., “Hindcast
Circulation and Climate Models”) or variables (e.g., “ocean conditions”) that could
not be linked unequivocally to one of the aforementioned categories.
Question 21: In your opinion, are spatial and temporal data of equal
value in establishing reference conditions and baselines? Please explain briefly.
Responses to this question were categorized at two levels. First, responses were split
into three groups—Yes, No, and It depends—with respect to whether respondents
found spatial and temporal data to be of equal value. Second, the No category was
also subdivided into two groups based on whether respondents found Temporal or
Spatial data to be of greater value for establishing baselines.
Question 22: Are there types of long-term data that would be useful, but
that aren’t currently available or you would want more of? If so, please give an
example. Responses to this question were assessed at three levels. First, responses
were divided into those saying Yes, No, or Unsure to the initial question. Second,
Yes responses were categorized into Abiotic, Biotic, or Other (e.g., “rate or process
data”) groups. Third, the Abiotic and Biotic groups were further subdivided into
the specific types of data identified by respondents. For the Abiotic subgroup, data
types included Temperature (e.g., “Deep-ocean temperatures.”), Water chemistry
(e.g., “Nutrient concentration of seawater.”), and Other (e.g., “seismic”). For the
Biotic subgroup, data types included Species abundance (e.g., “abundance of key
species”), Species distribution (e.g., “species distribution data : : : ”), Interactions
(e.g., “predator-prey relationships”), and Other (e.g., “species extinction rates”).
Question 23: What barriers (e.g., communication, funding, data availability,
etc.) have you experienced (or do you perceive to exist) in applying long-term
data to marine conservation? Responses to this question were grouped into four
categories—Funding, Data availability, Communication, and Institutional—similar
to those identified by conservation biologists (e.g., Strayer et al. 1986; Lindenmayer
et al. 2012) and a fifth category, Other, for miscellaneous responses. Many
respondents identified multiple barriers and each was tallied under the appropriate
category (e.g., “Funding and agency interest” was categorized as Funding and
Institutional). Responses categorized as Data availability discussed barriers related
to data accessibility or lack of data (e.g., “lack of data availability”, “True long-
term data is often not available”). The Communication category encompassed
responses at the level of disciplines (e.g., “ : : : communication may be one barrier,
with researchers not recognizing how certain other disciplines might value their
contributions”) as well as general challenges such as “Communicating long term
data can also be difficult if the data is collected on timescales not easily processed
by human minds.” Responses classified as Other included such impediments as
education (e.g., “educational barriers”) and politics (e.g., “Playing political small
ball : : : ”).
References
Arlettaz R, Schaub M, Fournier J et al (2010) From publications to public actions: when

conservation biologists bridge the gap between research and implementation. Bioscience
60:835–842
Bebber DP, Ramotowski MA, Gurr SJ (2013) Crop pests and pathogens move polewards in a
warming world. Nat Clim Chang 3:985–988
Besley JC (2016) The National Science Foundation’s science and technology survey
and support for science funding, 2006–2014. Public Underst Sci:0963662516649803.
https://doi.org/10.1177/0963662516649803
Boulton AJ, Panizzon D, Prior J (2005) Explicit knowledge structures as a tool for overcoming
obstacles to interdisciplinary research. Conserv Biol 19:2026–2029
Boyer AG, Brenner M, Burney D et al (2017) Conservation paleobiology roundtable: from promise
to application. In: Dietl GP, Flessa KW (eds) Conservation paleobiology: science and practice.
University of Chicago Press, Chicago, pp 291–302
Braunisch V, Home R, Pellet J et al (2012) Conservation science relevant to action: a research
agenda identified and prioritized by practitioners. Biol Conserv 153:201–210
Bremner A, Park K (2007) Public attitudes to the management of invasive non-native species in
Scotland. Biol Conserv 139:306–314
Burney DA, Juvik JO, Burney LP et al (2012) Can unwanted suburban tortoises rescue native
Hawaiian plants? The Tortoise 1:104–115
Callicott JB, Mumford K (1997) Ecological sustainability as a conservation concept. Conserv Biol
11:32–40
Carpenter SR, Turner MG (2001) Hares and tortoises: interactions of fast and slow variables in
ecosystems. Ecosystems 3:495–497
Casey MM, Dietl GP, Post DM et al (2014) The impact of eutrophication and commercial fishing
on molluscan communities in Long Island Sound, USA. Biol Conserv 170:137–144
Cintra-Buenrostro CE, Flessa KW, Dettman DL (2012) Restoration flows for the Colorado River
estuary, México: estimates from oxygen isotopes in the bivalve mollusk Mulinia coloradoensis
(Mactridae: Bivalvia). Wetl Ecol Manag 20:313–327
Cintra-Buenrostro CE, Flessa KW, Guillermo A-S (2005) Who cares about a vanishing clam?
Trophic importance of Mulinia coloradoensis inferred from predatory damage. PALAIOS
20:296–302
Conservation Paleobiology Workshop [CPW] (2012) Conservation paleobiology: opportunities for
the earth sciences. Report to the Division of Earth Sciences, National Science Foundation.
Paleontological Research Institution, Ithaca
Cook CN, Mascia MB, Schwartz MW et al (2013) Achieving conservation science that bridges the
knowledge-action boundary: achieving effective conservation science. Conserv Biol 27:669–
678
Crain CM, Kroeker K, Halpern BS (2008) Interactive and cumulative effects of multiple human
stressors in marine systems. Ecol Lett 11:1304–1315
Czech B, Devers PK, Krausman PR (2001) The relationship of gender to species conservation
attitudes. Wildl Soc Bull 29:187–194
Darling ES, Côté IM (2008) Quantifying the evidence for ecological synergies. Ecol Lett 11:1278–
1286
Davis MB (1989) Retrospective studies. In: Likens GE (ed) Long-term studies in ecology. Springer,
New York, pp 71–89
Dawson TP, Jackson ST, House JI et al (2011) Beyond predictions: biodiversity conservation in a
changing climate. Science 332:53–58
de Bruyn M, Hall BL, Chauke LF et al (2009) Rapid response of a marine mam-
mal species to Holocene climate and habitat change. PLoS Genet 5:e1000554.
https://doi.org/10.1371/journal.pgen.1000554
Dietl GP (2016) Brave new world of conservation paleobiology. Front Ecol Evol 4:21.
https://doi.org/10.3389/fevo.2016.00021
Dietl GP, Durham SR, Smith JA et al (2016) Mollusk assemblages as records of past and present
ecological status. Front Mar Sci 3:169. https://doi.org/10.3389/fmars.2016.00169
Evol 26:30–37
Dietl GP, Flessa KW (2017) Conservation paleobiology in the Anthropocene In: Dietl GP, Flessa
KW (eds) Conservation paleobiology: science and practice. University of Chicago Press,
Chicago, pp 303–305
Dietl GP, Smith JA (2016) Live-dead analysis reveals long-term response of estuarine bivalve
community to water diversions along the Colorado River. Ecol Eng 106:749–756
Dougherty EM, Fulton DC, Anderson DH (2003) The influence of gender on the relationship
between wildlife value orientations, beliefs, and the acceptability of lethal deer control in
Cuyahoga Valley National Park. Soc Nat Resour 16:603–623
Durham SR, Dietl GP (2015) Perspectives on geohistorical data among oyster restoration
professionals in the United States. J Shellfish Res 34:227–239
Elliott JM (1990) The need for long-term investigations in ecology and the contribution of the
Freshwater Biological Association. Freshw Biol 23:1–5
Eronen JT, Polly PD, Fred M et al (2010) Ecometrics: the traits that bind the past and present
together. Integr Zool 5:88–101
Farrell J (2016) Corporate funding and ideological polarization about climate change. Proc Natl
Acad Sci 113:92–97
pp 283–289
Franklin JF, Bledsoe CS, Callahan JT (1990) Contributions of the long-term ecological research
program. Bioscience 40:509–523
Fritz SA, Schnitzler J, Eronen JT et al (2013) Diversity in time and space: wanted dead and alive.
Trends Ecol Evol 28:509–516
Gauchat G (2015) The political context of science in the United States: public acceptance of
evidence-based policy and science funding. Soc Forces 94:723–746
Goodwin DH, Flessa KW, Schone BR et al (2001) Cross-calibration of daily growth increments,
stable isotope variation, and temperature in the Gulf of California bivalve mollusk Chione
cortezi: implications for paleoenvironmental analysis. PALAIOS 16:387–398
Griffiths RA, Dos Santos M (2012) Trends in conservation biology: progress or procrastination in
a new millennium? Biol Conserv 153:153–158
Holmes G (2015) What do we talk about when we talk about biodiversity conservation in the
Anthropocene? Environ Soc Adv Res 6:87–108
Jackson JBC, Alexander KE, Sala E (2012) Shifting baselines: the past and the future of ocean
fisheries. Island Press, Washington
Jackson JBC, McClenachan L (2017) Historical ecology for the paleontologist. In: Dietl GP,
Flessa KW (eds) Conservation paleobiology: science and practice. University of Chicago Press,
Chicago, pp 87–100
Katsanevakis S, Levin N, Coll M et al (2015) Marine conservation challenges in an era of economic
crisis and geopolitical instability: the case of the Mediterranean Sea. Mar Policy 51:31–39
Kearney S, Murray F, Nordan M (2014) A new vision for funding science. Stanford Soc Innov Rev
Fall:50–55
Kellert SR, Berry JK (1987) Attitudes, knowledge, and behaviors toward wildlife as affected by
gender. Wildl Soc Bull 15:363–371
Kelley PH, Dietl GP, Visaggi CC (2018) Training tomorrow’s conservation paleobiologists. In:
Tyler CL, Schneider, CL (eds) Marine conservation paleobiology. Springer, Cham, pp 207–223
Kidwell SM (2007) Discordance between living and death assemblages as evidence for anthro-
pogenic ecological change. Proc Natl Acad Sci 104:17701–17706
and conservation biology. Annu Rev Ecol Evol Syst 44:539–563
Kopf RK, Finlayson CM, Humphries P et al (2015) Anthropocene baselines: assessing change and
managing biodiversity in human-dominated aquatic ecosystems. Bioscience 65:798–811
Kowalewski M, Goodfriend GA, Flessa KW (1998) High-resolution estimates of temporal mixing
within shell beds: the evils and virtues of time-averaging. Paleobiology 24:287–304
Krebs CJ (2015) One hundred years of population ecology: successes, failures and the road ahead.
Integr Zool 10:233–240
Labandeira CC, Currano ED (2013) The fossil record of plant-insect dynamics. Annu Rev Earth
Planet Sci 41:287–311
Laurance WF, Koster H, Grooten M et al (2012) Making conservation research more relevant for
conservation practitioners. Biol Conserv 153:164–168
Likens GE (ed) (1989) Long-term studies in ecology: approaches and alternatives. Springer, New
York
Lindenmayer DB, Likens GE, Andersen A et al (2012) Value of long-term ecological studies.
Austral Ecol 37:745–757
Lotze HK, McClenachan L (2014) Marine historical ecology: informing the future by learning
from the past. In: Bertness MD, Bruno JF, Silliman BR et al (eds) Marine community ecology
and conservation. Palgrave Macmillan Sinauer Associates, Sunderland, pp 165–200
MacDonald GM, Bennett KD, Jackson ST et al (2008) Impacts of climate change on species,
populations and communities: palaeobiogeographical insights and frontiers. Prog Phys Geogr
32:139–172
Marris E (2011) Rambunctious garden: saving nature in a post-wild world. Bloomsbury USA, New
York
Millennium Ecosystem Assessment [MEA] (2005) Ecosystems and human well-being: synthesis.
Island Press, Washington
Mobley C, Kilbourne W (2013) Gender differences in pro-environmental intentions: a cross-
national perspective on the influence of self-enhancement values and views on technology.
Sociol Inq 83:310–332
Nadelson LS, Hardy KK (2015) Trust in science and scientists and the acceptance of evolution.
Evol Educ Outreach 8:9
National Research Council [NRC] (2005) The geological record of ecological dynamics:
understanding the biotic effects of future environmental change. National Academy Press,
Washington
Pandolfi JM, Bradbury RH, Sala E et al (2003) Global trajectories of the long-term decline of coral
reef ecosystems. Science 301:955–958
Pietri DM, Gurney GG, Benitez-Vina N et al (2013) Practical recommendations to help students
bridge the research-implementation gap and promote conservation: graduate students and the
research-implementation gap. Conserv Biol 27:958–967
Polly PD, Eronen JT, Fred M et al (2011) History matters: ecometrics and integrative climate
change biology. Proc R Soc B Biol Sci 278:1131–1140
Price SA, Schmitz L (2016) A promising future for integrative biodiversity research: an increased
role of scale-dependency and functional biology. Philos Trans R Soc B 371:20150228
Reed DC, Rassweiler AR, Miller RJ et al (2016) The value of a broad temporal and spatial
perspective in understanding dynamics of kelp forest ecosystems. Mar Freshw Res 67:14–24
Roux DJ, Kingsford RT, McCool SF et al (2015) The role and value of conservation agency
research. Environ Manag 55:1232–1245
Smith JA, Auerbach DA, Flessa KW et al (2016) Fossil clams reveal unintended carbon cycling
consequences of Colorado River management. R Soc Open Sci 3:160170
Smith JA, Dietl GP (2016) The value of geohistorical data in identifying a recent human-induced
range expansion of a predatory gastropod in the Colorado River delta, Mexico. J Biogeogr
43:791–800
Starzomski BM, Cardinale BJ, Dunne JA et al (2004) Contemporary visions of progress in ecology
and thoughts for the future. Ecol Soc 9:14
Strayer D, Glitzenstein JS, Jones CG et al (1986) Long-term ecological studies: an illustrated
account of their design, operation, and importance to ecology. The New York Botanical Garden,
Institute of Ecosystem Studies, Millbrook
Volety A, Savarese M, Hoye B et al (2009) Landscape pattern: present and past distribution
of oysters in South Florida coastal complex (Whitewater Bay/Oyster Bay/Shark to Robert’s
Rivers). Report to the South Florida Water Management District, Fort Myers, FL, Florida Gulf
Coast University
Wiens JA, Hobbs RJ (2015) Integrating conservation and restoration in a changing world.
Bioscience 65:302–312
Willis KJ, Bailey RM, Bhagwat SA et al (2010) Biodiversity baselines, thresholds and resilience:
testing predictions and assumptions using palaeoecological data. Trends Ecol Evol 25:583–591
Wolkovich EM, Cook BI, McLauchlan KK et al (2014) Temporal ecology in the Anthropocene.
Ecol Lett 17:1365–1379
Effectively Connecting Conservation
Paleobiological Research
to Environmental Management:
Examples from Greater Everglades’
Restoration of Southwest Florida
Michael Savarese
Abstract Much of the research accomplished by paleontologists can be categorized

as conservation paleobiology. Unfortunately, these works often go unrealized or
under-appreciated because the environmental professionals best positioned to use
those results are detached from the science, as managers or decision-makers,
or are completely unaware of the scientific discipline. Commonly, a research
program with management implications is undertaken without briefing the relevant
agencies in advance and without their input with respect to research objectives
and design. Academics typically work within or among their academic insti-
tutions with a predestined pathway toward peer-review publication in scholarly
journals. Those papers often go unread by agency professionals, and universities
at times of promotion rarely value technical reports. Partnership and collaboration
with agencies is essential. Agency professionals have very specific environmental
management objectives and priorities, often influenced by resource limitations.
Academic scientists should adapt their research programs to accommodate agency
research needs and priorities, and be willing to adopt research designs that best
achieve agency objectives, even if those agencies are unable to financially support
the effort. Often agencies can support research through in-kind match for field and
laboratory work, or with research staff assistance. When partnerships are strong,
“request for proposals” (RFPs) from state and local governmental agencies can be
customized for specific talents found among academics.
University scientists and agency professionals in Southwest Florida have devel-
oped an appreciation of conservation paleobiology and a culture of cooperation.
Effective steps for generating such a productive relationship include: (1) devel-
oping and actively participating in “management collaboratives,” working groups
composed of agency professionals, university scientists, and non-governmental
organization (NGO) professionals with stewardship commitments to their region’s
management and restoration needs; (2) attending and presenting conservation
M. Savarese ()
Marine and Ecological Sciences, Florida Gulf Coast University, Ft. Myers, FL, USA

56 M. Savarese
paleobiological work at environmental science and restoration conferences; (3)

inviting agency professionals to market and solicit their environmental science
priorities and then engaging agency representatives as co-investigators in those
studies; and (4) reserving time for professional service for those same agencies
and NGOs. Universities can incentivize conservation paleobiological research by
valuing technical reports as community-engaged scholarship, particularly if the
work results in a management decision or practice that helps the environment.
Greater Everglades’ restoration efforts have benefitted greatly from conservation
paleobiological approaches because of these team-building efforts. A couple of
examples are presented.
Keywords Conservation paleobiology · Environmental management ·

Environmental restoration · Greater Everglades restoration
1 Introduction
Conservation paleobiology is a relatively new discipline that is still inventing

itself. Paleontology’s value in testing or refining concepts in ecology and evolution
is well established, and these applications have obvious relevance to matters in
conservation biology. In this mode, paleontological data can provide a deep time,
extensive record of life’s responses to environmental drivers (Dietl and Flessa 2011;
Conservation Paleobiology Workshop 2012). The geohistorical record provides a
richer database with which to test ecological processes, particularly when neonto-
logical records are not robust. Those same data can be used to validate ecological
systems’ models; if models withstand testing in post-diction mode (i.e., models
that successfully predict situations in the past), their application for prediction is
more secure. Finally, that deeper record serves to establish historical precedents: for
example, mass extinctions have occurred in the past because of similar drivers; they
can happen again.
More importantly, and less appreciated, is paleontology’s potential to provide
a near-time record of life’s response to recent environmental or anthropogenic
change (Dietl and Flessa 2011). Society did not have the foresight to characterize
natural conditions prior to human development and industrialization. Consequently,
restoration managers do not have access to information about those pristine, pre-
anthropogenic conditions. Methods and concepts from paleontology, and from the
allied disciplines of sedimentology, stratigraphy, and geochemistry, provide a mech-
anism to infer those pre-anthropogenic conditions. This gives paleontology a direct
role in the development of environmental management and restoration strategies.
It’s important to recognize that the pre-industrialized world was not necessarily
pristine. Significant anthropogenic influences can date back well into aboriginal
times (Denevan 1992; Cronon 1996; Knowlton and Jackson 2008). But here
too paleontology can directly assist environmental management; the geohistorical
record is often of high enough quality to include paleoecological and archaeological
data, illuminating shifting baselines throughout human history (e.g., Jackson et al.
Effectively Connecting Conservation Paleobiological Research. . . 57
2001). With this perspective, environmental managers can better define their
restoration target.
As a result, classically trained paleontologists and environmental geoscientists
are predisposed to assist in environmental management and restoration. Our pro-
fessional training and culture, however, inhibit our effectiveness. Our work is often
unrecognized, not applied, or not appreciated simply because we are trained in a
different academic setting and are employed in different professional networks.
I have been fortunate to be employed at an institution of higher education, Florida
Gulf Coast University (FGCU), whose science is genuinely interdisciplinary—I
am one of a few geoscientists who do not reside in a Department of Geological
Sciences; my colleagues and students come from numerous scientific disciplines;
and we are all training students for careers and higher education in environmental
science and management (FGCU offers a BA and an MS in Environmental Studies,
a BS in Marine Science, an MS in Environmental Science). Additionally, our
university was founded with environmental stewardship and literacy as components
of its mission, and from its very inception in 1997, the university has been
closely partnered with environmental agencies and non-governmental organizations
(NGOs) from federal down to local geographic levels. The agencies and organiza-
tions with whom we work closely are numerous (see Table 1 for a comprehensive
Table 1 Organizations engaged in Greater Everglades’ management and restoration benefitting

by conservation paleobiologic research
Agency or Organization Affiliation
U.S. Army Corps of Engineers (ACOE) Federal
U.S. Fish and Wildlife Service Federal
Everglades National Park Federal
Ten Thousand Islands and Florida Panther National Wildlife Refuges Federal
Big Cypress National Preserve Federal
Rookery Bay National Estuarine Research Reserve Federal and State
Charlotte Harbor National Estuary Program Federal
U.S. Geological Survey Federal
National Resources Conservation Services Federal
Florida Fish & Wildlife Conservation Commission State
Florida Department of Environmental Protection State
South Florida Water Management District State
Collier-Seminole State Park State
Fakahatchee Strand State Park State
Lee County County
Collier County County
Collier County Coastal Zone Management County
City of Naples Municipal
City of Fort Myers Beach Municipal
Conservancy of Southwest Florida Not-for-profit
Audubon Florida Not-for-profit
Florida Wildlife Federation Not-for-profit
58 M. Savarese
list). Finally, the university serves as the only institution of higher education
in Southwest Florida, making us the clearinghouse for the western Everglades.
Consequently, we are intimately involved in many aspects of Greater Everglades’
restoration.
These, by happenstance, are ideal circumstances under which to be a conser-
vation paleobiologist, but they are rarely united at one university. Consequently,
my experiences at FGCU have provided me with insights that are shared in this
manuscript and that can serve as a model for other institutions.
2 Defining the Problem
The process of environmental management typically proceeds in this way: an

environmental problem requiring resolution is identified; the already existing,
relevant science is brought forward; additional science needs are identified; those
science efforts are implemented and their results integrated with the previous
science; and this collectively informs best management practices. Aspects of this
process are captured in crafted and often-revised management plans. Those plans
then drive resource (i.e., money, staff time, facilities) allocation and prioritization.
Environmental scientists engaged in this process produce the technical reports that
fulfill the science needs. They are also sitting at the table with the managers when
management plans are developed and when priorities are defined.
As university academics and as scientists outside the environmental profession,
we generate peer-reviewed journal articles; those published studies might consider
societal and environmental implications, but rarely are they focused on those
issues, and the environmental profession is typically not consulted or aware of our
activities. As tenure track professors, the currency of scholarly success is the number
of peer-reviewed papers and the impact factor held by the journals. Technical reports
are not valued similarly, and the societal impact of the environmentally relevant
work, assuming the science did affect environmental management and restoration, is
not credited. Some universities do define scholarly productivity broadly, sometimes
to include “community-engaged scholarship” (Saltmarsh et al. 2009; Whitmer
et al. 2010). I’m not aware, however, if promotion and tenure review processes
are evaluating this type of scholarship fairly and equitably. Finally, the university
scientist is not sitting at the table when science needs are identified and management
plans are developed.
Ultimately, we as scientists have to place anticipated and well-received deliver-
ables in the right hands, and our academic institutions must value and reward this
type of scholarship.
3 Ensuring Success as a Conservation Paleobiologist
As individual scientists or as working groups within departments, there are a number

of practices that can enhance our role in environmental management.
Developing Partnerships and Collaborative Teams
We are accustomed to reaching out to other academics within our discipline for
collaboration—teaming up with people that bring needed skills to the effort. We
are unaccustomed to partnering with people that have an environmental problem
that requires management or correction. As paleontologists, we have unique and
helpful skills, but we are not always aware of how our talents can be best applied to
environmental science and management. Conservation paleobiologists must engage
environmental agencies and the scientists and managers they employ. Regardless
of where a faculty member’s home institution is located, there are state and local
governmental agencies and NGOs that are charged with environmental management
and stewardship (for Southwest Florida, see Table 1). Federal environmental
agencies may be located at greater distances, but probably still close enough for
direct interaction. These agencies should be engaged. Meeting agency professionals
at their places of employment, arranging visits to the university for briefings, or
presenting a sample of work to those agency professionals are all valuable means to
developing collaborative relationships. Most agencies and NGOs employ someone
charged with overseeing scientific research (with the term “research” appearing in a
position’s title). This is the most logical point of first contact.
Assuming you, as a conservation paleobiologist, are flexible in the kinds of
scholarship you undertake, soliciting research questions from the agencies is a way
to ensure the agency is fully vested in the work you’ll pursue. If they have a need
that you can fill, then your results will undoubtedly be appreciated and applied in
societally useful ways. Often a research question of interest to you is closely related
to one of interest to the agency. Combining dual interests may be more cost-effective
and allow both to be pursued in concert.
This approach can also result in research support. Agencies often have dis-
cretionary funding they can provide outside of competitive bidding and formal
competition. These monies could be directed toward projects. Requests for propos-
als (RFPs) are often crafted by agencies; collaboration ensures you’re a recognizable
face if and when a call for proposals goes out; and, even without genuine dollars,
agencies do have in-kind resources that can facilitate your work (e.g., staff time,
field access, supplies).
When applying for funds extramurally and when the work is relevant to an
agency’s management mission, agency professionals should be invited to become
a member of the proposal writing team. Their agency may not permit this; however,
even something as simple as a letter of endorsement to append to the proposal can
improve the proposal’s competitiveness.
60 M. Savarese
Becoming or Engaging a Liaison
At FGCU, many, if not most, of the environmental science faculty members serve
as “liaisons” to specific agencies and NGOs. Faculty members serve as points
of contact for an agency. I, for example, serve as the liaison to Rookery Bay
National Estuarine Research Reserve for FGCU; I sit on Reserve committees,
represent the university’s interests and capabilities, and pass research needs and
opportunities on to our students and faculty. This ensures communication between
the agency and university is strong, that opportunities are broadcasted, and that
personal and professional relationships develop. Not every university scientist
has the right personality for serving as a liaison; some are more predisposed to
the role. Establishing a formal plan, one that defines service responsibilities for
liaisons within your department or work group, distributes the service workload
and guarantees many organizations are engaged. The university and our department
value these liaison functions as service, and they are considered at times of review
and promotion.
Participate in “Management Collaboratives”
Because environmental problems affect multiple stakeholders, management of those

problems requires involvement from multiple agencies and organizations. For
this reason, environmental agencies often meet as “management collaboratives,”
representatives from each of the stakeholder groups participate in regular meetings
to coordinate efforts and interpret results. (The U.S. Army Corps calls these
“planning charrettes.”) This is a mainstay of environmental management: by law,
all stakeholder groups must be considered, and, because of resource limitations,
agencies must pool their efforts and dollars to remain effective and efficient.
For example, both ACOE and the SFWMD, a federal and state partnership,
facilitate Everglades’ restoration projects. Their oversight process requires that
each restoration project be developed by a Project Delivery Team (PDT), which
ultimately produces a Project Implementation Report (PIR). Teams are composed of
members from all stakeholder groups. Lesser restoration and management efforts,
ones not garnering federal and state monies, often employ the same management
collaborative model. Serving as a conservation paleobiologist on such collaboratives
positions our science well and broadly informs the environmental community of our
value and capability.
Compose Technical Reports in Addition to Peer-Reviewed

Journal Articles
Perhaps it is unreasonable to assume the university tenure and promotion process

will reward technical reports or less-traditional products generated by community-
engaged scholarship. That said, it makes good sense to prepare two kinds of products
from a conservation paleobiological investigation: the technical report that will
be read and utilized by the environmental management community, and a peer-
reviewed journal paper that will not be read by the environmental community, but
has the respectability of the university community. The two are rarely identical in
structure and purpose, but one is readily transformed to the other. Interestingly,
some agencies require peer-review validation before scientific results are employed.
ACOE, for example, requires its contributing scientists to publish in peer-reviewed
journals before results are employed in management or restoration.
Present Your Findings to Stake Holder Groups
As with academic audiences, oral presentation of information to the management

community is often more influential than when presented in written form. For this
reason, volunteering to give a talk at the agency or to the management collaborative
is important. Management decisions are more readily influenced through oral
communication than in writing. Oral presentations also generate conversation and
personal connection, which increases the likelihood of appreciation and adoption.
Attend and Present at Environmental Science and Restoration

Conferences
Conservation paleobiological studies are most commonly given to the wrong

professional societies. I too am guilty of this, most often presenting my research
at professional meetings catering to geoscientists exclusively. The more appropriate
societies (e.g., Society for Ecological Restoration [SEC], Association of Environ-
mental Studies and Sciences [AESS], Coastal and Estuarine Research Federation
[CERF]) take me out of my comfort zone and introduce me to new people. Most
environmental scientists and managers have their academic training in the biological
or ecological sciences. Most are unfamiliar with the geological sciences, or assume
geology reveals deep time and therefore is irrelevant to modern historical problems.
Presenting our work to these audiences informs these disciplines of our science’s
capabilities.
62 M. Savarese
Train our Students
Most of us teach paleontology in traditional ways, de-emphasizing or excluding

entirely the value of conservation paleobiology. Our courses must be restructured
to accommodate conservation paleobiology for training of our geoscience students.
Though I have no statistics to emphatically make this statement, I suspect there are
more employment opportunities, both in the scientific workforce and in academia,
for paleontologists doing conservation paleobiology. A geology student with strong
training in sedimentology, stratigraphy, geochemistry, and conservation paleobi-
ology makes a strong candidate for the environmental science and management
profession.
Additionally, topical units covering conservation paleobiology should appear in
courses provided to environmental science and biology students. I would argue
that this is the most important audience to affect. A student with a degree in
environmental science that enters the workforce and already recognizes the value of
conservation paleobiology promotes the incorporation of the discipline. At FGCU, I
teach a graduate level course to our MS Environmental Science students entitled
“Conservation Paleobiology.” At the undergraduate level, taught principally to
BS Marine Science students, I offer a paleontology course (titled “Geobiology”)
that includes conservation paleobiologic examples throughout the semester-long
curriculum.
Finally, conservation paleobiological research projects, particularly ones advo-
cated by an agency, make ideal thesis topics for undergraduate and graduate
students. Such a project is in essence a scholarly experience and a professional
internship packaged together. The student works closely with both a faculty member
and one or more environmental professionals—perhaps one of those professionals
serves on the student’s thesis committee—and the student improves her or his
employment opportunities with that agency once the degree is complete.
Reward Faculty for Conducting Community-Engaged

Scholarship
In an ideal world, the tenure and promotion process should value community-
engaged scholarship, where the research does not necessarily result in a peer-
reviewed publication, but where the work’s societal effect and usefulness can be
documented in other ways. This may be a difficult change to advance in light
of institutional tradition, but it is worth exploring. One way to market these
accomplishments, even if they cannot be relied upon when being reviewed for
promotion, is to develop a well-documented case study within your portfolio that
outlines the project’s value and how it influenced some environmental management
practice. A letter of affidavit should accompany the case study coming from
someone within the agency that can speak directly about the work’s influence.
Promote and Reward Community Service for Work

with Environmental Agencies and NGOs
Developing relationships with environmental agencies is time consuming. Those

investments, however, typically result in other, non-scholarly products that serve the
greater community. My experience is that these collaborations are valuable service
commitments; I’m providing helpful services to the agency by participating in meet-
ings and management collaboratives, and through conversation and presentation.
These are deserving of recognition at times of promotion and annual review.
4 Case Studies from Greater Everglades’ Restoration
Admittedly, I have benefitted greatly from geographic placement: FGCU sits within
the western region of the Greater Everglades, and the counties of Southwest Florida
are developing at an alarming rate while available space is limited. This means
environmental management must effectively compromise societal expansion with
the conservation of natural resources and ecosystem services. Despite the growth
and the never-ending desire to live in “paradise,” at least during the northern
hemisphere winter, the environmental conservation ethic here is strong: citizens
recognize the value of our natural systems, if for nothing else than the region’s
economic prosperity. This means the public is generally agreeable and appreciative
though it does not necessarily mean there is adequate financial investment in
management and restoration.
Greater Everglades’ restoration is principally about “getting the water right”
(Sklar et al. 2005; U.S. Department of the Interior Office of Everglades Restora-
tion Initiatives 2016). Our low-lying topography and seasonally intense rainfall
translates into an historic proliferation of fresh and brackish water wetlands. As
uplands become increasingly limited for development, wetlands are encroached
or consumed. Surface hydrology is dominated by sheet flow, meaning that subtle
barriers created through development (e.g., roads) can grossly influence water
delivery and flow through wetlands. These wetlands historically supported vast
and diverse populations of wildlife, with many species currently threatened or
endangered. Complicating all of this are the products of water management from
the previous century: a vast network of canals, weirs, and impoundments that
have effectively drained wetlands, delivering unnatural volumes of freshwater (i.e.,
freshets) to sensitive estuaries. Greater Everglades’ restoration is essentially about
replumbing the system, returning at least some portion of the sheet flow back onto
the landscape.
Figure 1 shows the problem graphically. The pre-anthropogenic situation allowed
Lake Okeechobee to overflow its natural levees, delivering sheet flow to the
downstream wetlands and eventually to coastal estuaries. Shark Valley’s River
of Grass is the most notable of the affected waterways. (For a more thorough
64 M. Savarese
Fig. 1 The two maps show the historic flow patterns through the Greater Everglades prior
to human engineering and the current situation, whereby flows entering Lake Okeechobee are
diverted. The geographic locations for the two case studies detailed herein (Caloosahatchee River
and Picayune Strand) are shown here. Graphics provided by ACOE, Jacksonville District
review of the historic hydrology and the problems and successes of Everglades’
restoration, see Grunwald 2006.) In order to protect property and to better serve
agriculture south of Lake Okeechobee: (1) canals were constructed to reduce risk
to the population center in southeast FL; and (2) the Caloosahatchee River, in the
west, and St. Lucie River, in the east, were artificially connected to the lake to serve
as shunts for quick removal of freshwater to the estuaries. As a result, sheet flow
has become much reduced throughout most of the region, and some estuaries are
bombarded with massive freshwater flows when water levels threaten human life
and property. Greater Everglades’ restoration is tasked with restoring as much of
the natural hydrology as possible, while still protecting society from flooding.
The two most critical problems for Southwest Florida are managing flow through
the Caloosahatchee River (Case Study 1) and restoring sheet flow through the
Picayune Strand (Case Study 2). Both have benefitted, in part, by insights provided
by conservation paleobiology.
Case Study 1: Water Management of the Caloosahatchee River
The Caloosahatchee River estuary has suffered from the voluminous releases of
freshwater from Lake Okeechobee. Two severe impacts have been documented.
One, the freshets stress benthic ecosystems, particularly oyster reefs and seagrass
beds, that historically flourished in the estuary. Two, the flushing of freshwater,
accompanied by enriched nutrients, causes harmful algal blooms which trigger
hypoxia events and fin- and shellfish kills throughout the estuary and well out
into the Gulf of Mexico (Anderson et al. 2008; Vargo et al. 2008). The South
Florida Water Management District (SFWMD) wanted to document the effects on
the benthos and then use this information to better regulate the freshwater releases
from the Lake.
An FGCU graduate student, Jorge Agobian, and I proposed a comparative
taphonomic study of mollusk life and death assemblages to document the effects
of seasonal salinity changes, influenced most dramatically by freshwater releases on
the infaunal and epifaunal mollusks residing in the estuary. The SFWMD was aware
of the suspected influence on oyster productivity and upon the distribution of the
seagrass Vallisneria americana (tape grass), but did not have a clearly documented
cause and effect relationship between patterns of salinity change and estuarine
health (Doering et al. 2002; Barnes et al. 2007). Following a methodological
and conceptual approach developed by Susan Kidwell (2007, 2009), mollusk
assemblages were collected and analyzed quarterly from six locations along the
estuary’s salinity gradient (Fig. 2). Estimates of the time averaging of mollusk death
assemblages from the taphonomically active zone, acquired through radiocarbon
dating, demonstrated that death assemblages are taphonomically inert and represent
600 years of accumulation with all samples dating from prior to the middle
1900s. They, therefore, clearly reflect the recent pre-anthropogenic environmental
Fig. 2 (a) Map of the Caloosahatchee River estuary in Southwest Florida. The location of the area
is outlined in Fig. 1. Seven water quality and benthic invertebrate monitoring sites are shown along
the river’s salinity gradient. (b) Aerial photograph of the Franklin Locks, the downstream-most
water control structure on the river that regulates freshwater releases from Lake Okeechobee into
the estuary
66 M. Savarese
condition and the salinity pattern prior to the engineered connection of the river to
Lake Okeechobee. When the compositions of the life and death assemblages were
compared using Pearson product-moment and Spearman rank-order correlations,
and a detrended correspondence analysis, the influence of lowered salinities was
evident within the life assemblages, characterized by faunas tolerant of significantly
lower salinity.
The study was made possible because of an established relationship we had
with the District. One of the District’s lead scientists at the time was an estuarine
ecologist specializing in malacology and someone with whom we had worked
previously. Though the District did not support the research with funding, it did
make critical water quality and benthic monitoring data available, and assisted
in the study’s design and implementation. The study, though never published in
a peer-reviewed journal (see thesis: Agobian 2010), was submitted in technical
report form to the District, a final presentation was made to their staff, and the
District scientist served as an ex-officio member of Agobian’s thesis committee. The
ultimate measure of success though is how the results were employed. The District
used these results, coupled with the effects of freshets on oysters and seagrasses,
to alter their water release practices through the locks on the Caloosahatchee River.
Water is now released more modestly with prescribed flows to generate tolerable
salinities, and excessive releases, if necessary, are reserved for non-reproductive
months for benthic mollusks.
Case Study 2: Picayune Strand Restoration Project
Picayune Strand is the most ambitious restoration project funded through the
Comprehensive Everglades Restoration Plan (CERP), estimated, upon completion,
to cost $350 M (at time of publication of the Project Implementation Report; U.S.
Army Corps of Engineers 2004). The restoration plan that was adopted is currently
being implemented. That plan’s design benefited immensely from the inclusion
of a conservation paleobiological approach, and it was active engagement in a
management collaborative that resulted in its incorporation.
The environmental problem requiring repair was caused in the late 1960s by
a housing development project financed by Gulf American Land Corporation,
South Golden Gate Estates, in former freshwater wetlands located in eastern
Collier County north of the Ten Thousand Islands. The project was halted but not
until after the infrastructure to support the housing community was emplaced: a
network of 279 miles of elevated roads and the construction of 48 miles of canals
affecting 70,000 hectares of wetland habitat (Fig. 3). The canal system abruptly
and significantly lowered the water table to transform wetlands to uplands, and
now delivers the bulk of the freshwater to one downstream estuary, Faka Union
Bay. Those same canals rob estuaries to the west of freshwater sheet flow. Western
estuaries are essentially receiving too little freshwater, and experience significantly
elevated salinities, while Faka Union Bay experiences freshwater conditions for
much of the summer and fall rainy season. The elevated roads create a hindrance
Fig. 3 Images of Picayune Strand. (a) The satellite image shows: the system of roads and canals
that effectively drained the wetlands, and the trunk canal that delivers the freshwater to Faka Union
Bay in the Ten Thousand Islands. (b) Aerial photo of region marked by yellow box in A showing
the mouth of the main canal as it enters Faka Union Bay
to the restoration of sheet flow because of their damming effect. Any restoration
design, and the one ultimately chosen, necessitates the lowering of road grade and an
interruption of canal flow. Because this restoration project involved both federal and
state monies, the lead agencies overseeing the project’s design and implementation
are ACOE and the SFMWD. Their process requires all stakeholders to be involved,
and, for that reason, a management collaborative was employed; the SFWMD and
ACOE nomenclature calls these “project development teams” (PDT; U.S. Army
Corps of Engineers 2012).
When the environmental problem was first identified and prioritized, it was
conceived as a freshwater wetland project. At the time, because of my former
involvement with the SFWMD, I was invited as an FGCU representative for the
PDT. The group’s early conversations, in part because of my participation, resulted
in the extension of the project’s scope to the downstream estuaries. It was for this
aspect of the project that conservation paleobiology made a significant contribution.
Without this participation in the management collaborative, conservation paleobi-
ology would have been a completely unknown discipline among the environmental
scientists and managers. The tenor of the group and the project’s ultimate design
was transformed.
68 M. Savarese
Fig. 4 Photograph of an oyster reef exposed at low tide. When oyster reefs become emergent, they
trap red mangrove propagules, which may transform the reef to a mangrove-forested island
In order to restore the hydrology, salinity targets were needed for the affected
downstream estuaries. Unfortunately, prior to the emplacement of the canals and
roadways in the 1960s, no one had the foresight to monitor the estuarine water
quality; no record of pre-anthropogenic salinity variability existed. The estuaries
of the Ten Thousand Islands are ideal habitat for oyster productivity; oyster reefs
are prolific (Savarese et al. 2003; Savarese and Volety 2008). In fact, oyster reef
development throughout the last 3200 years of the late Holocene created the Ten
Thousand Island geomorphology; because oyster reef accretion exceeded the rate of
sea-level rise through this interval of time, the coast prograded and the oyster reefs
succeeded into mangrove-forested islands (Parkinson 1989), thereby generating the
Ten Thousand Island estuarine-scape (Figs. 4 and 5). The eastern oyster, Crassostrea
virginica, tolerates a wide range of salinities from near fresh to fully marine, but its
maximum productivity sits within the brackish range, from 15 to 25 ppt (Tolley et al.
2003; Volety et al. 2009). The conservation paleobiological approach was predicated
on the assumption that oyster reef development would have been most prolific
within this salinity range. The loci of reef development during pre-anthropogenic
history, when water management practices had not yet influenced salinity, would be
the geographic targets for those brackish water salinities. Mapping of oyster reefs
today, integrated with an understanding of reef history, through stratigraphy, was the
methodology proposed. Hydrologic modelers could then model various freshwater
flow scenarios through the wetlands to deliver the appropriate salinities throughout
the estuarine-scape.
Fig. 5 Google Earth image of the Ten Thousand Islands. Faka Union Bay, the recipient of the
canal system’s freshwater outfall, has become a freshwater point source during the rainy season;
the estuaries to the west have had their freshwater sheet flow beheaded, creating unnatural high
salinities; further west, at Blackwater Bay, sheet flow has been relatively undisturbed
A first step, involving both conservation paleobiology and estuarine ecology, con-
cerned the mapping of existing oyster reefs and an assessment of oyster health and
productivity on those oyster reefs. The reef-distribution maps comparing an estuary
far enough west to have avoided effects of canalization, Blackwater Bay, with Faka
Union Bay, the estuary receiving excessive freshwater flow, show that modern reefs
with the greatest productivity are spatially skewed. In Blackwater Bay, greatest reef
number and greatest oyster productivity resides in the central region of the estuarine
track, in Homologue 3 (Fig. 6). (Regions of the same homologue in neighboring
estuaries exhibit similar geomorphology and experienced similar hydrology prior to
anthropogenic influence.) This is where the estuarine geomorphology is richest in
oyster-reef produced landforms—the now mangrove-forested islands. Faka Union
Bay, on the other hand, shows greatest oyster productivity in Homologue 5, while
relict reefs and reef-produced landforms are most extensive in Homologue 3 (Fig. 7).
Conditions for best growth and reproduction are displaced downstream in Faka
Union relative to Blackwater Bay, a condition one would predict because of the
volume of freshwater output into this estuary.
70 M. Savarese
Fig. 6 Satellite image showing the distribution of oyster reefs within Blackwater Bay estuary
(in yellow), located far enough west in the Ten Thousand Islands to be unaffected by sheet flow
beheading. Homologues 1–5 (H1, H2, etc.) are regions in the estuary that have specific and com-
parable geomorphology and pre-anthropogenic salinities to neighboring estuaries. Consequently,
similar homologues should have comparable states of reef development. In Blackwater Bay, oyster
reefs are most numerous and most productive in H3
The second conservation paleobiologic step was to interpret the history of reef
development in both estuaries in Homologue 3. Both Blackwater and Faka Union
Bay oyster reefs in Homologue 3 have an extensive history, dating back at least
2000 ybp, with high oyster productivity. This indicates Homologue 3 throughout
the Ten Thousand Islands maintained the ideal salinity for maximum oyster health in
the pre-anthropogenic condition. Delivering brackish water salinities to this medial
region through sheet flow restoration was justified, and the hydrologic modelers
targeted the needed salinities on this region of the estuarine-scape.
The ultimate restoration design chosen (Fig. 8) involves a number of engineered
features, included the lowering of road grade, the filling or plugging of canals, and
the emplacement of pumps and spreader canals. The pumps and spreader canals
permit quick transferal of freshwater and sheet flow initiation from the region
to the north, the existing housing development of North Golden Gate Estates, to
the Picayune Strand during times of excessive rainfall. The restoration design is
adaptive in that freshwater flow through the pump stations can be regulated to ensure
the appropriate salinities are delivered to the estuaries.
Fig. 7 Similar satellite image as in Fig. 6 but for Faka Union Bay, the estuary receiving excessive
freshwater. The greatest oyster reef productivity is shifted downstream and resides in H5
Alternative 3D
FAKA UNION
CANAL
MERRITT
PRAIRIE
CANAL
CANAL
I-75
800 cfs Elements
1200 cfs
N 2600 cfs -Three Spreaders
MILES -Three Pump
0 2 Stations
-225 Miles of
Road Removed
-83 Canal Plugs
U.S 41
Fig. 8 Restoration design “Alternative 3D,” the engineering plan that was ultimately chosen for
Picayune Strand restoration. The solution involves the removal of roads, the filling or plugging of
canals, and the construction of three pump stations and associated spreader canals
72 M. Savarese
5 Conclusions
Paleontologists have a wealth of unrealized experience and knowledge to share

with the environmental management and restoration world. Those skills are often
unappreciated by the environmental management community for two fundamental
reasons: the environmental profession is filled predominately by professionals that
lack training in geoscience, and therefore lacks an understanding of our abilities
to interpret environmental history; and the paleobiologic community does not
effectively communicate or collaborate with the environmental field. The keys
to our productive involvement in environmental science and management include
self-promotion, marketing, and relationship building, over the short term, and a
transformation in how we teach and to whom we teach paleontology, over the long
term. Conservation paleobiology is societally relevant and necessary. It provides a
new purpose for our science and, ultimately, greater opportunities for employment
of our students.
Acknowledgments Many agencies over the years have generously supported conservation pale-
obiologic research for Greater Everglades’ management and restoration. The granting agencies
include: South Florida Water Management District, Florida Department of Environmental Protec-
tion, City of Naples, Charlotte Harbor National Estuary Program, National Science Foundation,
National Oceanic and Atmospheric Administration, Environmental Protection Agency, and the
U.S. Department of Education (for support of student involvement). A number of individuals from
the environmental management world are owed thanks for their appreciation and advocacy for
conservation paleobiology: Michael Bauer, Peter Doering, Kim Dryden, Michael Duever, Kevin
Godsea, Patty Goodman, Gary Lytton, Ananta Nath, Janet Starnes, and Clarence Tears. Thanks to
my graduate students who over the years have advanced conservation paleobiologic applications.
They include Jorge Agobian, Kim Andres, Amanda Booth, Nicole Fronczkowski, Brian Hoye,
and Sasha Wohlpart. Special thanks to the Fall, 2013 Conservation Paleobiology graduate course
at FGCU. The conceptual framework of this paper developed through the course’s curriculum
and conversations. This manuscript benefitted greatly from the thoughtful reviews by Rowan
Lockwood and Lynn Wingard; thanks to both of them.
References
Agobian JN (2010) The impact of water management practices in the Caloosahatchee River:
mollusk assemblages as indicators of environmental change. M.S. Thesis, Florida Gulf Coast
University
Anderson DM, Burkholder JM, Cochlan WP et al (2008) Harmful algal blooms and eutrophication:
examining linkages from selected coastal regions of the United States. Harmful Algae 8:39–53
Barnes TK, Volety AK, Chartier K et al (2007) A habitat sustainability index for the eastern oyster
(Crassostrea virginica), a tool for the restoration of the Caloosahatchee estuary, Florida. J
Shellfish Res 26:949–959
Conservation Paleobiology Workshop (2012) Conservation paleobiology: opportunities for the
Earth sciences. 2011. Report to the Division of Earth Sciences, National Science Foundation.
Cronon W (1996) The trouble with wilderness: or, getting back to the wrong nature. Environ Hist
1:7–28
Denevan WM (1992) The pristine myth: the landscape of the Americas in 1492. Ann Assoc Am
Geogr 82:369–385
Evol 26:30–37
Doering P, Chamberlain RH, Haunert DE (2002) Using submerged aquatic vegetation to establish
minimum and maximum freshwater inflows to the Caloosahatchee estuary, Florida. Estuaries
25:1343–1354
Grunwald M (2006) The swamp: the Everglades, Florida, and the politics of paradise. Simon and
Schuster, New York
Jackson JBC, Kirby MX, Berger WH et al (2001) Historical overfishing and the recent collapse of
coastal ecosystems. Science 293:629–638
Kidwell SM (2007) Discordance between living and death assemblages as evidence for anthro-
pogenic ecological change. Proc Natl Acad Sci 104:17701–17706
Kidwell SM (2009) Evaluating human modification of shallow marine ecosystems: mismatch in
composition of molluscan living and time-averaged death assemblages. In: Dietl GP, Flessa KW
(eds) Conservation paleobiology: using the past to manage for the future. The Paleontological
Society papers, vol 15. pp 113–139
Knowlton N, Jackson JB (2008) Shifting baselines, local impacts, and global change on coral reefs.
PLoS Biol 6(2):e54. https://doi.org/10.1371/journal.pbio.0060054
Parkinson RW (1989) Decelerating Holocene sea-level rise and its influence on Southwest Florida
coastal evolution: a transgressive/regressive stratigraphy. J Sediment Petrol 59:960–972
Saltmarsh J, Giles DE Jr, Ward E et al (2009) Rewarding community-engaged scholarship. New
Dir High Educ (147):25–35
Savarese M, Volety A (2008) Oyster reef health in Pumpkin and Fakahatchee Estuaries: baseline
monitoring for Ten Thousand Islands Restoration. South Florida Water Management District,
Technical Report
Savarese M, Volety A, Tolley G (2003) Influence of watershed alteration on oyster health and
oyster-reef habitat: management implications for the Faka-Union and Estero Bays. South
Florida Water Management District, Technical Report
Sklar FH, Chimney MJ, Newman S et al (2005) The ecological-societal underpinnings of
Everglades restoration. Front Ecol Environ 3:161–169
Tolley G, Volety A, Savarese M (2003) Shellfish research and adaptive resource management in
Southwest Florida: oysters as sentinels of ecosystem health. World Aquaculture 34:64–66
U.S. Army Corps of Engineers (2004) Project documents: Picayune Strand Restoration Inte-
grated Project Implementation Report (PIR)/Environmental Impact Statement (EIS). Ever-
gladesRestoration.gov. URL: http://141.232.10.32/pm/projects/docs_30_sgge_pir_final.aspx.
U.S. Army Corps of Engineers (2012) SMART restoration planning guide: the project
delivery team. EvergladesRestoration.gov. http://planning.usace.army.mil/toolbox/
smart.cfm?Section=2&Part=1.
U.S. Department of the Interior Office of Everglades Restoration Initiatives
(2016) EvergladesRestoration.gov: Restoring America’s Everglades. http://
www.evergladesrestoration.gov/index.htm.
Vargo GA, Heil CA, Fanning KA et al (2008) Nutrient availability in support of Karenia brevis
blooms on the central West Florida Shelf: what keeps Karenia blooming? Cont Shelf Res
28:73–98
Volety AK, Savarese M, Tolley SG et al (2009) Eastern oysters (Crassostrea virginica) as an
indicator for restoration of Everglades’ ecosystems. Ecol Indic 9(Suppl 6):S120–S136
Whitmer A, Ogden L, Lawton J et al (2010) The engaged university: providing a platform for
research that transforms society. Front Ecol Environ 8:314–321
Using the Fossil Record to Establish
a Baseline and Recommendations
for Oyster Mitigation in the
Mid-Atlantic U.S.
Kristopher M. Kusnerik, Rowan Lockwood, and Amanda N. Grant
Abstract Eastern oyster populations throughout the Mid-Atlantic region of the

USA have been in decline for centuries due to overharvesting, disease, increased
sediment pollution, and habitat destruction. By studying Pleistocene fossil oyster
assemblages, it is possible to reconstruct baseline conditions and develop rec-
ommendations for oyster mitigation. Fossil assemblages were studied from five
Pleistocene sites located in Maryland, Virginia, and North Carolina. Reconstructions
of paleosalinity and temperature were used to identify modern and colonial sites
with similar environmental parameters for comparison. Shell height and life span
in Chesapeake Bay oysters declined significantly from the Pleistocene to today, at
the same time that ontogenetic growth rates have increased. This pattern is driven
by age truncation, in which both harvesting and disease preferentially remove the
larger, reproductively more active and primarily female members of the population.
By contrast, Pleistocene oysters from North Carolina did not differ significantly, in
shell height, life span, or growth rates, from modern oysters.
Although oyster management in the Mid-Atlantic States has focused historically
on protecting and supplementing early life stages, this study recommends three
potential management solutions to the age truncation revealed by comparison with
Pleistocene oysters. Possible solutions include (1) implementation of a maximum
size or slot limit on the fishery, (2) establishment of marine protected areas (MPA),
or (3) significant lowering of exploitation rates.
K. M. Kusnerik ()
Division of Invertebrate Paleontology, Florida Museum of Natural History, Gainesville, FL, USA
R. Lockwood
Department of Geology, The College of William and Mary, Williamsburg, VA, USA
A. N. Grant
School of Earth Sciences and Environmental Sustainability, Northern Arizona University,
Flagstaff, AZ, USA

76 K. M. Kusnerik et al.
Keywords Crassostrea virginica · Pleistocene · Growth rates · Chesapeake

Bay · Fishery · Aquaculture · Aquatic resource management · Oyster reef ·
Restoration · Virginia · Maryland · North Carolina
1 Introduction
The Eastern Oyster (Crassostrea virginica) plays a vital role in the ecosystem of
the Chesapeake Bay and Mid-Atlantic regions (Mann et al. 2009a). Oysters are
ecosystem engineers that build habitat for fish and other invertebrate species, boost
water quality by filtering bacteria and contaminants, and represent an important
component of the food web (Haven and Morales-Alamo 1970; Meyer and Townsend
2000; Cressmann et al. 2003; Hoellein et al. 2015). Historically, oyster harvests have
provided a key economic resource for the region (Paolisso and Dery 2010; Rick and
Lockwood 2013).
The Pleistocene record provides evidence of widespread, thriving oyster reefs
predating human settlement in the region, during the formation of the proto-
Chesapeake Bay (Hargis and Haven 1995; USGS 1998; Reshetiloff 2004; Rick
and Lockwood 2013). The earliest record of human harvest of Mid-Atlantic oysters
dates back to the Late Archaic (ca 2500–2000 cal yr. BC) in the form of oyster
middens, or archaeological deposits of kitchen waste material (Waselkov 1982;
Custer 1989; Thompson and Worth 2011). Native American harvesting occurred
for thousands of years, across the bay region, and is assumed to have involved
harvesting of small clusters of shells from easily accessible reefs (Rick et al. 2014,
2016).
European settlement of the region began with the Jamestown Colony, which was
established along the James River (Virginia) in 1607 (Rountree et al. 2007; Horn
2008). English settlers relied heavily on oysters as a food source, leaving a record of
oyster harvesting in the form of dozens of shells recovered from an abandoned well
within the settlement (Kelso 2004; Harding et al. 2008, 2010a). Although oyster
harvesting has been a key component of the regional economy for thousands of
years, these harvests have declined massively over the last 150–200 years and now
represent a mere 1% of peak productivity (Rothschild et al. 1994; Harding et al.
2008; Beck et al. 2011; Wilberg et al. 2011).
By the late 1800s, harvest by industrial dredging caused a massive reduction
in reef height through the removal of shell material faster than living oysters
could replenish it (Hargis and Haven 1999). Natural oyster reefs have effectively
disappeared in the modern bay, due to overfishing, disease, increased sediment
input, and habitat destruction, leaving oyster populations depleted and in need of
serious mitigation efforts (Rothschild et al. 1994; Mann and Powell 2007).
Oyster populations in the region have also been impacted by two prevalent
parasitic diseases that increase oyster mortality rates (Carnegie and Burreson
2009). Dermo disease is caused by the parasite Perkinsus marinus, is prevalent
in intermediate salinity (12–15 ppt) warmer waters, and was first documented
Using the Fossil Record to Establish a Baseline and Recommendations. . . 77
in the Chesapeake Bay in 1949 (Andrews 1996; Burreson and Ragone Calvo
1996; CTDOAG 2016a). MSX (Multinucleated Sphere Unknown) is caused by the
spore-forming protozoan Haplosporidium nelsoni, prefers more saline conditions
(>15 ppt) and was first documented in the Mid-Atlantic region in 1957 (CTDOAG
2016b) as a result of the intentional introduction of the Japanese oyster (Crassostrea
gigas) to Delaware Bay (Andrews and Wood 1967). Increased sediment influx from
land clearance has further complicated the situation by hastening habitat destruction
in areas of optimal oyster growth (Hargis and Haven 1999).
Early management approaches to Mid-Atlantic oysters focused on helping the
fishery recover from natural and anthropogenic problems primarily for the benefit
of the local economy. These approaches included the genetic enhancement of
broodstock, the release of spat raised through aquaculture, and planting of dead
shell (cultch) to act as substrate for settlement (Bartol and Mann 1997, 1999a,
1999b; Mann and Evans 1998; Southworth and Mann 1998; Wesson et al. 1999;
Luckenbach et al. 1999; Mann 2000; Southworth et al. 2000). More recent efforts
have embraced a broader strategy, prioritizing the mitigation of ecosystem services
via protected areas, reduction of harvesting, and large-scale three-dimensional reef
restoration (Luckenbach et al. 2005, Coen et al. 2006, 2007; North et al. 2010; Beck
et al. 2011; Grabowski et al. 2012).
Pleistocene fossil reefs may provide insight into how oyster ecosystems func-
tioned before human intervention in the Mid-Atlantic U.S. Although natural oyster
reefs have dwindled to extinction in the modern Chesapeake Bay, Pleistocene
oysters can yield information on shell size and growth rates that may prove vital
for mitigation efforts. The goal of this study is to quantify oyster size and growth
rates in the Mid-Atlantic region, across three timescales—Pleistocene, colonial, and
modern—to assess the implications of this conservation paleobiological approach
for oyster mitigation.
2 Methods
Pleistocene Localities
Samples of Pleistocene oysters were examined from the Virginia Museum of

Natural History (VMNH) and field collections across five localities, distributed from
southern Maryland to North Carolina (Fig. 1; Table 1). The northernmost of the
Chesapeake Bay sites, Wailes Bluff (WB), is located at the mouth of the Potomac
River in St. Mary’s County, Maryland. This site, which has since been covered by
a seawall, yielded molluscan material collected by L. W. Ward in 1971 (VMNH
71LW93). Bulk samples were collected from C. virginica shell layers distributed in
a sandy silt matrix, thought to represent the late Pleistocene Tabb Formation (Fig. 2;
Thompson 1972; Belknap 1979; Cronin 1979; Wehmiller and Belknap 1982; Rader
and Evans 1993). Many of the molluscan species identified at Wailes Bluff are now
Fig. 1 Map of five localities

in Maryland, Virginia, and
North Carolina, from which
Pleistocene fossil oysters
were sampled (WB Wailes
Bluff, CP Cherry Point, HP
Holland Point, SP Stetson Pit,
LC Lee Creek)
restricted to more southerly latitudes, suggesting a paleotemperature warmer than

today (Blake 1953). Several brackish water taxa have been documented at this site,
supporting a paleosalinity of approximately 15–30 ppt (Blake 1953; Cronin 1979).
Cherry Point (CP, also known as Norris Bridge) is a middle Pleistocene site
located in Lancaster County, Virginia, near the mouth of the Rappahannock River
(Fig. 1; Table 1). Although the site is no longer available for collecting due to
development, B. W. Blackwelder and T. M. Cronin collected oyster specimens in
1978 (VMNH 78BB79A, B; VMNH T8TC56) from a fossiliferous sandy silt unit
thought to represent the Shirley Formation (Fig. 2; Mirecki 1990; Mirecki et al.
1995). Ostracode assemblages at this site indicate a Pleistocene bottom temperature
between 12.5 and 15 ı C during winter and 27.5 ı C during summer months (Cronin
1979), both of which are warmer than modern conditions at the site by at least 1–
2 ı C (Massmann et al. 1952). Paleosalinity is thought to have ranged between open
sound (15–35 ppt) and estuarine (2–15 ppt) conditions (Cronin 1979).
Table 1 Location, stratigraphic unit, and geologic age of the five localities sampled for Pleis-
tocene oysters
Sample size Stratigraphic Pleistocene
State (35 mm) Latitude/longitude unit interval
Wailes Bluff MD 36 (36) 38.065560/76.365280ı Tabb Late
(WB)
Cherry Point VA 36 (36) 37.634184/76.412830ı Shirley Mid
(CP)
Holland Point VA 865 (611) 37.512088/76.432121ı Shirley Mid
(HP)
Stetson Pit (SP) NC 225 (225) 35.866291/76.293768ı Undetermined Late
Upper Lee Creek NC 85 (85) 35.324287/76.800213ı James City Early
(ULC)
Lower Lee NC 21 (21) 35.324287/76.800213ı Flanner Mid
Creek (LLC) Beach
The Lee Creek locality is divided into two sections: Upper Lee Creek (ULC) and Lower Lee Creek
(LLC), respectively
Fig. 2 Stratigraphic framework for four of the Pleistocene localities (WB Wailes Bluff, HP
Holland Point, CP Cherry Point, ULC Upper Lee Creek, LLC Lower Lee Creek). The stratigraphic
unit for Stetson Pit (SP) is undetermined
Holland Point (HP), the southernmost of the Chesapeake Bay sites, is located
on the Piankatank River (Fig. 1; Table 1). Amino acid racemization dating of C.
virginica and Mercenaria specimens suggests an age of approximately 195–243Ka
(MIS 7 or 9, J. Wehmiller, personal communication 2016). The unit in which the
oyster deposit occurs is thought to represent the Shirley Formation, which is middle
Pleistocene in age (Fig. 2; C. R. Berquist, personal communication 2016). The
exposed oyster deposit at Holland Point is laterally extensive (up to 25 m) and thick
(up to 3 m), containing thousands of oysters, many of which are articulated and
preserved in life position within a fine sandy, clayey silt matrix. Reconstructions of
both paleotemperature and salinity were accomplished as part of the current study.
The northernmost site in North Carolina is Stetson Pit (SP, Fig. 1, Table 1),
located at the mouth of the Albemarle Sound in Dare County, North Carolina,
and subsequently covered by landfill material. Bulk samples, collected in 1979
by B. W. Blackwelder (VMNH 79BB32(D)), yielded several C. virginica. This
shelly, sandy mud unit that was sampled has never been attributed definitively
to a specific stratigraphic unit (Miller 1982), but its position above a U-series-
dated coral suggests that it is late Pleistocene (late MIS 5a, J. Wehmiller personal
communication 2016) in age. Using assemblages of temperature-sensitive ostra-
codes, York et al. (1989) identified a high proportion of cryptophilic species,
suggesting relatively cooler temperatures compared to present-day Cape Hatteras,
North Carolina. These ostracode assemblages, and the presence of molluscan taxa
including Rangia cuneata, indicate a brackish paleosalinity (York et al. 1989).
The final site, Lee Creek (PCSD Phosphate Mine), is located on the Pamlico
River in Beaufort County, North Carolina. The site was sampled by L. W. Ward
in 1972 (VMNH 72LW8C, 72LW1B) and 1992 (VMNH 92LW60a) although it is
currently inaccessible (Fig. 1; Table 1). This site produced C. virginica specimens
at two stratigraphic horizons: Lower Lee Creek (LLC) within the James City
Formation (early Pleistocene) and Upper Lee Creek (ULC) within the Flanner
Beach Formation (middle Pleistocene) (Fig. 2; (Ward and Blackwelder 1987; Ward
and Bohaska 2008). The James City beds yielded a variety of mollusks in a
fossiliferous, medium-coarse grained quartz sand. Rare records of freshwater (i.e.,
Corbicula) and brackish (i.e., Rangia) taxa suggest that these beds represent an
offshore barrier bar influenced by migrating channels that intermittently opened and
closed, with more open-marine salinity conditions behind the back barrier (Ward
and Blackwelder 1987; Ward and Bohaska 2008). The Flanner Beach sediments
are characterized by a very fine sandy silt (Ward and Bohaska 2008), reflecting a
back barrier muddy estuary with corresponding brackish paleosalinity (Ward and
Bohaska 2008).
Field and Museum Sampling
Of the five sites studied, Holland Point is the only one still accessible for field
sampling. We sampled the exposed oyster deposit at Holland Point in July 2011,
in addition to describing the sedimentology and measuring the stratigraphic section.
Samples were collected in five columns spaced approximately 3 m apart along the
lateral extent of the deposit. We collected three bulk samples (spaced evenly apart
according to deposit thickness) from each column (n D 15 samples total), using a
0.5 m2 quadrat (Fig. 3).
Fig. 3 Sampling transects of the oyster deposit at Holland Point. Samples were collected in five
columns spaced approximately 3 m apart along the lateral extent of the deposit. Three bulk samples
(spaced evenly apart according to deposit thickness) were collected from each column (n D 15
samples total), using a 0.5 m2 quadrat
Bulk samples of C. virginica from the Virginia Museum of Natural History in

Martinsville, Virginia, were examined from Wailes Bluff (VMNH 71LW93), Cherry
Point (VMNH 78BB79A, B; 78TC56), Stetson Pit (VMNH 79BB32), and Lee
Creek (VMNH 72LW1B, 72LW8C, 92LW60A).
Oyster Size and Abundance Data
Samples from Holland Point were sieved using a 4 mm mesh size, sorted, and
molluscan whole specimens and hinge fragments were identified to the lowest
taxonomic level possible (Abbott 1974; Spencer and Campbell 1987).
For each museum and field sample, we used digital calipers to measure shell
height for all whole left valves of C. virginica. Shell height was measured as the
distance from the umbo to the ventral-most edge of the shell. Although this distance
is commonly referred to as “shell length,” it is more accurate to use the term “shell
height” (Galtsoff 1964). The field site at Holland Point yielded 865 whole left valves
of C. virginica for measurement. The number of museum specimens available for
each Pleistocene site varied from 21 to 225 whole left valves.
A subset of these left valves was randomly selected to be sawed in half, using
a diamond-tipped tile saw. The resulting bisected hinges were used to count shell
bands in cross-section. We recorded the number of thick, dark gray shell bands that
were continuous from the hinge to the outer shell layer (Fig. 4), to provide a proxy
for biological age (e.g., life span) of each specimen, following Harding and Mann
(2006) and Zimmt et al. 2016).
Fig. 4 Cross-section of bisected hinge from Holland Point, showing gray and white growth lines;
1–9 represent thick, gray shell bands that were continuous from the hinge to the outer shell layer
and were counted to provide an approximate biological age (i.e., life span) for oyster specimens
Reconstructing Paleotemperature and Salinity
In order to standardize for paleoenvironment, we compiled information on pale-

osalinity and paleotemperature for each Pleistocene locality, except for Holland
Point, from the literature.
To estimate paleosalinity at Holland Point, the raw abundance of all other
molluscan specimens identified to the species level was assessed. For bivalves,
whole shells and hinge fragments were sorted into left versus right valves and the
larger of the two values was used to represent raw abundance for each sample. For
gastropods, whole and fragmented specimens were sorted into apertures and apices
and the larger of the two values was used to represent abundance for each sample.
We compiled modern salinity tolerance ranges from the literature for those species
with more than one occurrence at Holland Point (Federighi 1931; Andrews 1953;
Menzel et al. 1966; Castagna and Chanley 1973; Buroker 1983; Zimmerman and
Pechenik 1991; Grabe et al. 1993; Griffin 2001; Hill 2004, Zachary and Haven 2004;
Wilson et al. 2005; Harding et al. 2010b; Cohen 2011).
To assess paleotemperature at Holland Point, one specimen of C. virginica was
assessed using clumped isotope analysis by G. A. Henkes (Johns Hopkins Univer-
sity). A sub-sample of shell material was collected from the hinge area using a
low-speed Dremel drill and analyzed following the methods of Henkes et al. (2013).
Additionally, two articulated C. virginica and one articulated M. mercenaria, all
preserved in life position from Holland Point, were assessed for paleotemperature
using sclerochronology under the supervision of G.S. Herbert (University of South
Florida). Sub-samples (one sample per every 0.78–1.75 mm) were collected from
each annual growth band of sectioned valves using a Dremel drill. Powdered
material was dissolved in 100% H3 PO4 at 25 ı C for 24 h. The resulting CO2 was
separated, focused, and analyzed on a Thermo Finnigan Delta V Advantage IRMS
in continuous flow mode coupled to a Gasbench II preparation device (Harke et al.
2015). Growing season paleotemperatures were calculated using the Craig (1965)
calcite-water equation for C. virginica and the Grossman and Ku (1981) aragonite-
water equation for M. mercenaria. The salinity range estimated from HP molluscan
occurrences was used to determine •18 Oseawater for these equations.
Modern and Colonial Data
Reconstructions of paleosalinity from the Pleistocene sites were used to identify

modern and colonial sites with C. virginica living in similar salinity conditions.
Data on shell height and growth rates in colonial and modern oysters were compiled
from the published literature and management agencies as described below.
In Maryland, colonial data on mean shell height and growth rates were compiled
from Miller (1986) and Catts et al. (1998) for four sites from similar salinity
regimes (15–25 ppt), including St. Mary’s City (sample size not published) and
Ashcomb’s Quarter (n D 99 shells). Data on modern shell height and growth rates
were provided by M. Tarnowski (Maryland Department of Natural Resources) for
six sites (n D 1176 shells) from Pocomoke and Tangier Sounds (15–25 ppt salinity
zone) sampled from 2013 to 2015.
In Virginia, colonial data on shell heights and growth rates were acquired from
Harding et al. (2008, 2010a) from the Jamestown Colony (n D 363 shells, salinity
15–30 ppt). Modern shell height and growth rate data were also compiled from
published sources (Harding et al. 2008; Sisson et al. 2011) for eight sites (n D 6916
shells) in the James River (collected from 2006 to 2008) and Lynnhaven River
(collected from 2005 to 2008), from the same salinity range.
We were unable to locate any colonial aged oyster data from North Carolina.
Data on shell height and growth rates for modern North Carolina oysters were
obtained from Puckett and Eggleston (2012), for six sites sampled from 2006 to
2008 (n D 5443 shells) in Pamlico Sound (15–30 ppt salinity zone).
3 Results
Paleoenvironmental Reconstruction of Holland Point

Paleotemperature
Clumped isotopic analysis of a single C. virginica valve from the Holland Point
fossil deposit produced a •13 Ccarb composition of 0.47 ˙ 0.02%PDB, •18 Ocarb of
2.08 ˙ 0.01%PDB, and 47carb of 0.725 ˙ 0.014% (Ghosh)(Ghosh et al. 2006;
Huntington et al. 2009). This 47 value corresponds to an oyster growth temperature
(i.e., averaged temperature at which the oyster was growing at this particular site,
at this particular time) of 14.1 ˙ 5 ı C. Given this paleotemperature and measured
oxygen isotope value, solving the Kim et al. (2007) equilibrium oxygen isotope
fractionation equation provides a •18 Owater value of 3.44% SMOW.
Isotopic sclerochronological analysis of one M. mercenaria and two C. virginica
shells from Holland Point yielded comparable results of •18 Owater D 3.40%
SMOW. Using the Craig (1965) and Grossman and Ku (1981) equations, respec-
tively, growing season paleotemperatures ranged from 5.3 to 20.5 ı C for C. virginica
and 10.9 to 20.9 ı C for M. mercenaria. This growth temperature is cooler than
the modern Piankatank River, which routinely reaches temperatures between 25
and 30 ı C during the summer growing months of modern C. virginica (Harding
et al. 2010b). Sclerochronological analyses of other species would be required to
determine whether C. virginica stopped growing in either the summer or winter
months at this site in the middle Pleistocene.
Paleosalinity
Almost 1500 whole and fragmented shells, representing 21 macroinvertebrate

species other than C. virginica, were identified in the bulk samples collected from
the Holland Point deposit. Raw abundance of these species ranged from 1 to 515
(mean D 55.62) per sample. Six taxa were excluded from the salinity analyses
because they could not be identified to the species level (undetermined crab, sponge
(likely Cliona), barnacle (likely Balanus)) or because data on salinity tolerance of
the modern representatives were not readily available (Trittia trivittata, Crepidula
convexa, and Melanella polita). When rare taxa, those with only a single occurrence,
are excluded, a paleosalinity range of 16–32 ppt encompasses all remaining species
(Fig. 5). This salinity is higher than that of the modern Piankatank River, which
fluctuates between 6 and 23 ppt (Harding et al. 2010b).
Fig. 5 Modern salinity tolerances of fossil macroinvertebrate species with abundance greater than
1 recorded at Holland Point (Virginia); shaded region (15–32 ppt) encompasses salinity range
within which all taxa could co-occur
Shell Height
Shell height was compared across all of the Pleistocene, colonial, and modern
localities using size frequency distributions and non-parametric statistical tests.
We excluded specimens 35 mm in shell height in order to avoid sampling spat
(Mann et al. 2009a). Kolmogorov-Smirnov tests for normality indicated that the
shell height data were non-normally distributed for most fossil (KS21–655 D 0.08–
0.22, p D 0.20–0.0001) and modern (KS24–5784 D 0.05–0.17, p D 0.20–0.0001)
samples. We therefore used a non-parametric test (Mann-Whitney U) to test for
differences in mean shell height among Pleistocene, colonial, and modern oysters
across three geographic regions: (1) upper Chesapeake Bay (Maryland), (2) lower
Chesapeake Bay (Virginia), and (3) North Carolina.
Starting with the Maryland portion of the Chesapeake Bay, late Pleistocene
oysters tend to be larger than colonial or modern oysters from similar salinity
regimes (15–25 ppt, Fig. 6, Table 2). Pleistocene oysters also exhibit more strongly
right-skewed size frequency distributions than modern oysters (Fig. 6).
Moving south, into the Virginia portion of the bay, middle Pleistocene (MP)
oysters were statistically significantly larger than both colonial and modern oysters
from similar salinity regimes (15–30 ppt, Fig. 7, Table 2). Pleistocene oysters
reached a maximum size of nearly 260 mm, in comparison to colonial (124 mm)
and modern (148 mm) specimens. The right tails of the size frequency distributions
for both the modern and colonial oysters appear truncated. These distributions are
missing the larger adults that contribute to the strongly right-skewed distribution of
the Pleistocene sample (Fig. 7).
In Albemarle and Pamlico Sounds in North Carolina, MP and LP oysters show
no significant difference in size or the shape of the size frequency distribution from
modern oysters from similar salinity regimes (15–30 ppt, Fig. 8, Table 2).
Growth Rate
Growth rates were plotted for Pleistocene, colonial, and modern oysters by plotting
biological age (i.e., life span) of each specimen versus shell height (Figs. 9 and
10). We excluded specimens 35 mm in shell height in order to avoid sampling
spat (Mann et al. 2009a). We calculated the slopes of these growth trajectories
(i.e., growth rates) and compared them using an analysis of covariance (ANCOVA,
modelled as linear). Growth rates were compared among Pleistocene, colonial,
and modern oysters across three geographic regions: (1) upper Chesapeake Bay
(Maryland), (2) lower Chesapeake Bay (Virginia), and (3) North Carolina.
In the upper Chesapeake Bay, growth trajectories in Pleistocene oysters extend
beyond 12 years (Fig. 9). In contrast, growth trajectories for colonial and modern
oysters rarely extend beyond 5 years (Fig. 9). Using ANCOVA to compare growth
rates in Maryland oysters between 0 and 5 years of age suggests that modern
oysters are growing significantly faster than Pleistocene or colonial oysters from
similar salinity regimes (15–30 ppt, Fig. 9, Table 3). Whereas modern oysters record
Fig. 6 Shell height (mm)

trends in late Pleistocene
(LP), colonial and modern
oysters from sites in the upper
Chesapeake Bay region
(Maryland) with salinities
ranging from 15 to 25 ppt. (a)
Mean shell height ˙ S.E.; (b)
Size frequency distributions
for Pleistocene and modern
oysters
Table 2 Mann-Whitney U tests comparing shell height (mm) in Pleistocene, colonial, and modern
oysters from Maryland, Virginia, and North Carolina
Maryland Virginia North Carolina
MP vs. LP – – Z225,21 D 2.93,
p D 0.003
MP vs. – Z647,364 D 4.48, p < 0.0001 –
colonial
MP vs. – Z647,6916 D 15.15, Z21,5443 D 1.05,
modern p < 0.0001 p < 0.29
LP vs. Z36,3 D 1.69, – –
colonial p D 0.09
LP vs. Z36,1176 D 6.44, – Z225,5443 D 3.76,
modern p < 0.0001 p < 0.0001
Colonial vs. Z3,1176 D 0.69, Z364,6916 D 9.56, p < 0.0001
modern p D 0.49
Statistically significant differences (p 0.05) highlighted in bold; all oyster comparisons from
similar salinity regimes; MP middle Pleistocene, LP late Pleistocene

trends in middle Pleistocene
(MP), colonial, and modern
oysters from sites in the lower
Chesapeake Bay region
(Virginia) with salinities
ranging from 15 to 30 ppt; (a)
for Pleistocene (MP),
colonial, and modern oysters
average growth rates of 14 mm/year (slope), colonial (slope D 11.78 mm/year), and
Pleistocene oysters grow significantly more slowly (slope D 8.51 mm/year).
Comparisons of growth trajectories among Pleistocene, colonial, and modern
oysters in the lower Chesapeake Bay yield similar results. Growth trajectories in
Pleistocene oysters extend beyond 20 years, while colonial and modern oysters
rarely live longer than 5 years (Fig. 10). Mean biological age (i.e., life span,
x axis of Fig. 10) ranges from 8.6 years at the Holland Point fossil locality to
2.38 years at the Jamestown colonial site to 2.59 years in modern bay sites with
salinities ranging from 15 to 30 ppt. Comparison of growth rates in Virginia via
ANCOVA (oysters aged 0–5 years) reveals that Pleistocene oysters have slower
growth rates than either colonial or modern oysters from similar salinity regimes

trends in middle Pleistocene
(MP), late Pleistocene (LP),
and modern oysters from
North Carolina with salinities
ranging from 15 to 30 ppt; (a)
for Pleistocene (MP and LP
combined) and modern
oysters
(15–30 ppt, Fig. 10, Table 3). Colonial (slope D 28.94 mm/year) and modern
oysters (slope D 19.73 mm/year) both have significantly greater growth rates than
Pleistocene (slope D 8.24 mm/year) oysters.
Interestingly, growth trajectories in North Carolina oysters do not extend beyond
4 years of age, for either Pleistocene or modern oysters. ANCOVA reveals that
growth rates in oysters 5 years are significantly steeper in modern, in comparison
to Pleistocene oysters, when salinity is controlled for (15–30 ppt, Fig. 11, Table
3). The slope of modern oysters (23.77 mm/year) is almost three times that of
Pleistocene oysters (8.87 mm/year).
Fig. 9 Growth rate trends in Pleistocene and modern oysters from sites in the upper Chesapeake
Bay region (Maryland) with salinities ranging from 15 to 30 ppt
Table 3 Results for ANCOVA comparing growth rates (shell height (mm) vs. biological age
(years)) in Pleistocene, colonial, and modern oysters from Maryland, Virginia, and North Carolina
Maryland Virginia North Carolina
Pleistocene vs. F2,11 D 3.64, p D 0.08 F2,33 D 9.93, p < 0001 –
colonial Not significant Colonial > Pleistocene
Pleistocene vs. F3.43 D 29.52, F2,1067 D 570, p < 0001 F2,875 D 472.53,
modern p < 0001 Modern > Pleistocene p < 0001
Modern > Pleistocene Modern > Pleistocene
Colonial vs. F2.37 D 104.18, F2,1057 D 571.02, –
modern p < 0001 p < 0001
Modern > Pleistocene Modern > Colonial
Growth rates modelled as linear for oysters 5 years; statistically significant differences (p 0.05)
highlighted in bold, all comparisons are controlled for salinity regime
Fig. 10 Growth rate trends in Pleistocene (middle and late Pleistocene combined), colonial, and
modern oysters from sites in the lower Chesapeake Bay region (Virginia) with salinities ranging
from 15 to 30 ppt
4 Discussion
Comparing Pleistocene to Modern Oysters
Fossil oysters from the Chesapeake Bay are 1.3–1.6 times larger than oysters
from either colonial or modern times (depending on location). This difference is
not the result of slower growth rates. In fact, colonial and modern oysters grew
approximately 3–4 times faster than Pleistocene oysters in the bay, even when
growth rates are only calculated for younger, faster-growing individuals (0–5 years).
The size difference is driven by the presence of significantly longer-lived (up to 3.6
times longer) adults in the Pleistocene assemblages. Growth trajectories in fossil
Chesapeake Bay oysters continued beyond 5 years of age (Figs. 9 and 10), but
colonial and modern bay populations rarely lived that long.
Fig. 11 Growth rate trends in Pleistocene (middle and late Pleistocene combined) and modern
oysters from sites in the Pamlico Sound region of North Carolina with salinities ranging from 15
to 30 ppt
The same pattern does not hold for North Carolina. While there is no significant
difference in size between middle Pleistocene and modern oysters, late Pleistocene
oysters are significantly smaller than modern ones. Although growth rates in modern
oysters in Pamlico Sound are 6–7 times faster than fossil oysters, the short life spans
of both fossil and modern oysters result in little to no difference in these population
parameters through time.
Environmental Controls on Oyster Size
A number of factors could be responsible for this significant decrease in oyster

size in the Chesapeake Bay from the Pleistocene to the modern. Sampling, for
both modern and fossil localities, was accomplished via bulk sampling, either by
hand (Pleistocene) or with the use of hydraulic patent tongs (modern). The open
dimensions of patent tongs make it possible to sample all of the oysters inhabiting
one square meter of bay bottom. Both sampling techniques are therefore random
with respect to size, suggesting that sampling bias is unlikely to be affecting these
trends.
The likelihood of fossil preservation varies according to shell size and thickness,
such that smaller mollusks are, in general, less likely to be preserved than larger
ones, due to processes such as fragmentation and dissolution (Cummins et al. 1987;
see reviews in Martin 1999; Kidwell 2013). In the field, these Pleistocene sites are
extremely well-preserved and qualitative inspection of the oyster shells reveals little
taphonomic damage (including few signs of dissolution or corrosion). The majority
of oysters at Holland Point and Stetson Pit (Miller 1982) were still articulated
and oriented in life position and the clayey-silty matrix shows no evidence of the
molds and casts that would be expected under dissolution regimes. The bimineralic
and relatively thick shell of oysters also makes it less likely that they would have
experienced dissolution, especially given the preservation of aragonitic material at
the same localities.
Environmental factors, including salinity, temperature, and nutrients, are closely
tied to oyster size in modern settings. Although oysters are tolerant of a wide range
of salinity conditions (0–40 ppt; Quast et al. 1988; Shumway 1996), optimal growth
and reproduction occur between 10 and 28 ppt, with larval recruitment only possible
above 6 ppt (Wilson et al. 2005). Salinity is generally negatively correlated to oyster
size in the modern Chesapeake Bay, not because of a direct link between salinity
and oyster size, but because of the high occurrence of disease, predators, and boring
sponges in fully marine conditions (Galtsoff 1964; Paynter and Burreson 1991;
VOSARA 2016). By explicitly limiting modern comparisons to those with similar
salinity regimes, we have controlled for the complicating effects of salinity on these
data.
Growth studies in the Chesapeake Bay found that modern Crassostrea virginica
grow between July and October, with growth stopping when mean temperatures
dropped below 10 ı C (Paynter and Dimichele 1990). Though environmental condi-
tions are similar during some spring months, the oysters exhibit no growth during
these periods (Paynter and Dimichele 1990). The Pleistocene sites sampled in this
study all represent interglacial intervals. Bottom water temperatures at these sites in
the Pleistocene range from substantially warmer (C10 ı C at Cherry Point, Virginia)
to substantially colder (10 ı C at Holland Point, Virginia) than the same locations
today (NDBC 2016). The colder paleotemperatures for Holland Point (Virginia) are
similar to those recorded today between Ocean City, Maryland and Lewes, Delaware
(NOAA NCEI 2016). The warmer paleotemperatures documented at Cherry Point
(Virginia) correspond to temperatures observed today near Wilmington, North
Carolina. If the changes in shell height were driven by temperature, we would
expect oysters from Pleistocene bay localities with warmer temperatures (i.e.,
Cherry Point and Wailes Bluff, Maryland) to differ from those from localities with
cooler temperatures (i.e., Holland Point, Virginia). In reality, there is no significant
difference in average or maximum shell height at these three localities.
Like most marine invertebrates, oysters grow larger under higher nutrient
conditions (Berg and Newell 1986; Rice and Rheault 1996). Although the complex
history of nutrient pollution in the Chesapeake Bay has been reconstructed in
detail for the last millennium (Cooper and Brush 1991; Cooper 1995; Zimmerman
and Canuel 2002; Kemp et al. 2005), considerably less is known about nutrients
along the Mid-Atlantic Coastal Plain during the Pleistocene. Krantz (1990) sampled
growth bands of middle Pliocene (Yorktown Formation) to early Pleistocene (James
City Formation) scallops from Virginia and found evidence of seasonal increases
in productivity (interpreted as spring phytoplankton blooms), but no record of
upwelling. This result suggests that local nutrient levels were most likely lower in
the Pleistocene than during historic and modern times. This increase in nutrients
may be driving the increase in oyster growth rates observed in the historic and
modern Chesapeake Bay (Kirby and Miller 2005; Harding et al. 2008; Mann et
al. 2009b), but it cannot explain the smaller shell sizes and shorter longevities. Until
proxy data for local productivity are available throughout the Pleistocene, it will
be difficult to quantify the effects of primary productivity on oyster growth in the
Mid-Atlantic region through time.
Human Factors Influencing Oyster Size
In addition to environmental factors, two anthropogenically driven factors are

affecting historical and modern oyster sizes: disease and overharvesting. Two
diseases, Dermo and MSX, exert a massive influence on oyster abundance in the
bay today. Mortality is higher with MSX although the virulence of Dermo appears
to have increased rapidly after the introduction of MSX (Burreson and Ragone
Calvo 1996; Carnegie and Burreson 2009). Today, MSX is thought to kill the
majority of oysters larger than 51 mm in high salinity regions of the bay. Disease-
resistant strains of C. virginica exist (Brown et al. 2005; Encomio et al. 2005;
Carnegie and Burreson 2011), but the majority of oyster growers prefer to grow
triploid (non-reproductively active) oysters. These oysters, which have three sets of
chromosomes, are just as susceptible to disease but tend to grow faster and reach
market size before dying.
Studies of the sustainability of Native American and colonial harvesting are just
beginning, but a compilation of oyster size across 28 archaeological sites spanning
3500 years of Native American harvest suggests that shell height remained relatively
stable throughout this interval (Rick et al. 2016). Although Native Americans
harvested oysters for thousands of years prior to European colonization, early
English settlers reported massive oyster reefs covering the bay and its tributaries
(Wharton 1957; Hargis and Haven 1999; Mann et al. 2009b; Rick et al. 2014).
Measures of Colonial Era oyster shell height and growth rates, using similar
techniques to those used here, suggest that Jamestown Colony oysters were
intermediate in size between Pleistocene and modern oysters (Harding et al. 2008,
2010a). Their growth rates were elevated, relative to both modern and Pleistocene
growth rates (Harding et al. 2008). Kirby and Miller (2005) observed the same
pattern in colonial oysters from the St. Mary’s and Patuxent Rivers (Maryland) and
argued that this pattern was the result of increased nutrient availability due to land
clearance. By 1860, oyster growth rates began to decrease, with the initiation of
harvesting by dredging, combined with hypoxia and harmful algal blooms (Kirby
and Miller 2005).
Today, oyster size and population density are at historic lows, in part due to the
culling of the larger tail of the size frequency distribution that is highlighted by
comparisons between Pleistocene and modern oyster sizes. Like disease, harvesting
disproportionately affects the larger, more reproductively active adult oysters
(Hutchings and Reynolds 2004). In fact, the average mortality of these larger oysters
exceeds 60% on natural oyster bars in the Piankatank (Harding et al. 2010b), James
(Mann et al. 2009a), and Great Wicomico (Southworth et al. 2010) Rivers.
Implications for Restoration
This study has documented a substantial decrease in oyster shell size in the
Chesapeake Bay from the Pleistocene to today. This difference in shell size is not
due to a decrease in growth rates or culling of smaller oysters. Instead, it is driven
by the culling of oysters at the larger end of the size frequency distribution in the
modern bay, due to both overharvesting and disease. The elimination of large adults
from the population is a common sign of overharvesting in many marine species
(Hutchings and Reynolds 2004; Berkeley et al. 2004; Birkeland and Dayton 2005;
Hsieh et al. 2006).
With that in mind, it is important to note that conservation efforts for Chesapeake
Bay oysters focus almost exclusively on early life stages—in particular larvae and
spat. These approaches include: (1) distribution of cultch on the bay floor to increase
the likelihood of larval settlement, (2) rearing and release of oyster larvae, and (3)
minimum size restriction on oysters harvested from both aquaculture (2 in.) and
natural beds (3 in.)(Kennedy 1989; Mann and Powell 2007; Kennedy et al. 2011;
Wilberg et al. 2011; Md. Code Regs. § 08.02.04.11; 4VA Admin Code 20-260-
30). Millions of dollars are spent on these approaches each year (Luckenbach et al.
1999; Mann and Powell 2007; Beck et al. 2011; zu Ermgassen et al. 2012) but, from
a statistical standpoint, very few of these oysters will actually settle and grow to
market size.
Efforts devoted to preserving adult oysters are minimal by comparison. Only a
small percentage of oyster habitat is protected from harvesting for the long term
(D sanctuaries, 9–25% or 9000 acres in Maryland, <2% or 200 acres in Virginia;
Schulte et al. 2009; MDDNR 2016; VMRC 2016a), which, in turn, protects both
the early and late stage oysters growing there. The majority of protected areas in
Virginia are closed to harvesting for only 1 to 3 years at a time (VMRC 2016a).
This approach means that adult oysters are afforded little to no protection in the
bay, despite the fact that they are the most reproductively important members of the
population (Mann et al. 2009a, b). Because older oysters put exponentially more
energy into reproduction and less into shell growth, lack of protection of adults
has a catastrophic effect on oyster population growth. Similar patterns have been
documented across a wide range of marine and freshwater fisheries (Berkeley et
al. 2004; Birkeland and Dayton 2005; Hsieh et al. 2006; Venturelli et al. 2009;
Arlinghaus et al. 2010). In the Chesapeake Bay, the overfishing of larger specimens
and culturing of triploid specimens also means that disease resistance is evolving
exceptionally slowly in oysters (Encomio et al. 2005). These problems are further
compounded by the fact that oysters are sequential protandric hermaphrodites and
that the ratio of females in a population increases with increasing age/size (Kennedy
1983; Heffernan et al. 1989; Harding et al. 2013). Oysters start their lives as male,
and, in the Chesapeake Bay, do not transition to female until they are at least 60 mm
long (approximately 1.6 years old; Harding et al. 2013). Thus, the culling of the
larger sizes of oysters also preferentially removes females from the population.
The exception to preferential removal is areas of the bay that have historically
avoided long-term harvesting because they are privately owned or excessively
polluted. For example, larger, more disease-resistant oysters have been documented
recently in locations such as Tangier Sound (Blankenship 1997; Encomio et al.
2005) and the Elizabeth River (Schulte et al. 2009; CBF 2016). This finding
suggests that Chesapeake Bay oysters have the potential, if not the opportunity, to
evolve disease resistance and grow to larger sizes that approach those seen in the
Pleistocene record.
The Pleistocene record of Chesapeake Bay oysters emphasizes the significant
extent to which colonial and modern populations have experienced age and size
truncation. Management solutions to age truncation in marine fisheries include: (1)
implementation of a maximum size or slot limit, (2) the establishment of marine
protected areas (MPA), or (3) significant lowering of exploitation rates (Berkeley et
al. 2004; Venturelli et al. 2009; Hixon et al. 2014). In many freshwater and some
marine species, fishery size restrictions include both a maximum and minimum (slot
limit), or a restriction based on reproductive stage. For example, management of
the blue crab (Callinectes sapidus) in the Chesapeake Bay includes: (1) an MPA
(Lipcius et al. 2003) and (2) fishing restrictions based on egg mass and molting
stage, in addition to overall size (MDDNR 2016; VMRC 2016b). The reproductive
stage of oysters would be difficult for waterpeople to determine because they
are broadcast spawners (Kennedy 1983; Heffernan et al. 1989). Maximum size
restrictions could, however, be enforced because minimum size restrictions are
already in effect for oysters collected in the bay.
MPAs are areas of oceans or lakes that are protected from human activity to
conserve natural or cultural resources. MPA approaches have proven useful across
a wide variety of fishery species (Russ 2002; Halpern 2003; Pelletier et al. 2005),
but effectiveness varies according to the size of the MPA and its duration (Claudet
et al. 2008; Edgar et al. 2014). Harvest moratoria are controversial but have been
suggested in the past by both ecological and economic managers to preserve oyster
resources in the bay (Schulte et al. 2009; Kasperski and Wieland 2010; Wilberg et
al. 2011).
If conservation funding were to shift toward protection of older, reproductively

more active oysters, it would represent a more efficient approach, as each adult
female produces an average of 2–115 million eggs per year, increasing with age
(Brooks 1996). This approach is not without its challenges, especially given the
importance of sustaining a culture supporting the local waterpeople (Paolisso 2007;
Paolisso and Dery 2010). But the benefits of shifting funding priorities to preserving
large, disease-resistant oysters make it worthwhile. The importance of prioritizing
adult over early life stages has been recognized for several other aquatic species
(Berkeley et al. 2004; Birkeland and Dayton 2005; Hsieh et al. 2006; Venturelli et
al. 2009).
A Role for Conservation Paleobiology
Oysters from the Mid-Atlantic states represent an excellent example of how fossil
data can inform conservation issues. Because humans have inhabited this region for
at least 13,000 years (Dent 1995; Lowery et al. 2010), and harvests were not tracked
until the 1870s (Lotze 2010; zu Ermgassen et al. 2012), a sizeable gap exists in our
understanding of how these ecosystems have changed over long timescales. By the
time marine scientists established monitoring of bay oysters in the 1940s, oyster
populations were already decimated by 75 years of dredging (Haven et al. 1978;
Rothschild et al. 1994; Rountree et al. 2007; zu Ermgassen et al. 2012). Given the
effects of introduced disease and increased sediment influx, the oyster beds that we
study today tell us very little about how healthy oyster reefs function, either in the
past or present. One could argue that bay managers have never seen a healthy oyster
reef.
Pleistocene fossil assemblages can serve as a baseline for Chesapeake Bay oyster
mitigation. These assemblages allow us to quantify body size, growth rates, and
other factors that can be used to shape modern mitigation efforts.
Combining historical (zu Ermgassen et al. 2012), archaeological (Rick and
Lockwood 2013; Rick et al. 2016), and paleobiological (this volume) approaches
makes it possible to bridge these timescales and to assess how oysters have
responded to various pressures, including harvesting, climate, and sea level change
in the past. Conservation paleobiology plays a particularly important role, as the
only approach to yield information on ecosystems before human settlement. In the
end, the fossil record provides us with a crucial baseline for mitigation, a glimpse
into the world of Mid-Atlantic oysters before humans.
5 Conclusion
In conclusion, Pleistocene oysters from the Chesapeake Bay region are larger, and
longer-lived than either colonial or modern oysters. This pattern is not the result of
environmental shifts in salinity or temperature. Instead, it is driven by culling of the
larger tail of the shell size frequency distribution in the modern bay, suggesting that
both human harvesting and disease are eliminating the larger, reproductively more
active female members of the population. Solutions to this size and age-skewed
population structure include: (1) decreasing harvest pressure, (2) establishing a
maximum size limit, and (3) designating MPAs specifically for oysters in the
Chesapeake Bay. The conservation paleobiologic approach applied in this study
emphasizes the impact that human activities have had on these ecosystem, at the
same time providing a baseline for future mitigation.
Acknowledgments We would like to thank Admiral Pete Roane and his family for access to
the study site at Holland Point, Virginia. Additionally, we would like to thank Dr. Roger Mann
(Virginia Institute of Marine Science) for information on the Holland Point site; Dr. Buck Ward
(Virginia Museum of Natural History) for field assistance and access to museum specimens;
students Chris Young, Sam Bonanni, and Eric Dale (The College of William and Mary); Mitch
Tarnowski (Maryland Department of Natural Resources) for Maryland oyster data; Dr. Gregory
Henkes (Johns Hopkins University) for clumped isotope analyses; Dr. Gregory Herbert (University
of South Florida) for help with sclerochronology lab work; Dr. Henry Miller (Historic St.
Mary’s City), Keith Doms (Newlin Grist Mill), and Drs. Torben Rick and Leslie Reeder-Myers
(Smithsonian Institution) for Maryland archaeological data; Dr. John Wehmiller (University of
Delaware) for amino acid racemization analyses; Rick Berquist (Virginia Division of Geology
and Mineral Resources) for designation of stratigraphic units; and Dr. Doug Jones (Florida
Museum of Natural History) for help with data interpretation. Comments from two reviewers,
Drs. Patricia Kelley (University of North Carolina at Wilmington) and Nicole Bonuso (California
State University at Fullerton) greatly improved the quality of the manuscript. Lastly, we would like
to thank the Ellen Stofan Fund from the College of William and Mary for financial support of this
research project.
References
Abbott RT (1974) American seashells. Litton Educational Publishing Inc., New York
Andrews JD (1953) Fouling organisms of the Chesapeake Bay. The Chesapeake Bay Institute of
the Johns Hopkins University Inshore Survey Program Interim Report XVII, 53-3
Andrews JD (1996) History of Perkinsus marinus, a pathogen of oysters in Chesapeake Bay
1950-1984. J Shellfish Res 15:13–16
Andrews JD, Wood JL (1967) Oyster mortality studies in Virginia. VI. History and distribution of
Minchinia nelsoni, a pathogen of oysters, in Virginia. Chesap Sci 8(1):1–13
Arlinghaus R, Matsumura S, Dieckmann U (2010) The conservation and fishery benefits of
protecting large pike (Esox lucius L.) by harvest regulations in recreational fishing. Biol
Conserv 143(6):1444–1459
Bartol IK, Mann R (1997) Small-scale settlement patterns of the oyster Crassostrea virginica on a
constructed intertidal reef. Bull Mar Sci 61:881–897
Bartol IK, Mann R (1999a) Small-scale patterns of recruitment on a constructed intertidal reef:
the role of spatial refugia. In: Luckenbach MW, Mann R, Wesson JA (eds) Oyster reef habitat
restoration: a synopsis of approaches. Virginia Institute of Marine Science, Gloucester Point,
pp 159–170
Bartol IK, Mann R (1999b) Growth and mortality of oysters (Crassostrea virginica) on constructed
intertidal reefs: effects of tidal height and substrate level. J Exp Mar Biol Ecol 237:157–184
Beck MW, Brumbaugh RD, Airoldi L et al (2011) Oyster reefs at risk and recommendations for
conservation, restoration, and management. Bioscience 61(2):107–116
Belknap DF (1979) Application of amino acid geochronology to stratigraphy of late Cenozoic

marine units of the Atlantic coastal plain. Dissertation, University of Delaware
Berg JA, Newell RI (1986) Temporal and spatial variations in the composition of seston available
to the suspension feeder Crassostrea virginica. Estuar Coast Shelf Sci 23(3):375–386
Berkeley SA, Hixon MA, Larson RJ et al (2004) Fisheries sustainability via protection of age
structure and spatial distribution of fish populations. Fisheries 29(8):23–32
Birkeland C, Dayton PK (2005) The importance in fishery management of leaving the big ones.
Trends Ecol Evol 20(7):356–358
Blake SF (1953) The Pleistocene fauna of Wailes Bluff and Langleys Bluff, Maryland. Smithsonian
Misc Coll 121:1–32
Blankenship K (1997) Oysters rescued from VA harvest may return favor. Chesapeake Bay J 7:9
Brooks WK (1996) The oyster. JHU Press, Baltimore
Brown BL, Butt AJ, Meritt D et al (2005) Evaluation of resistance to Dermo in eastern oyster
strains tested in Chesapeake Bay. Aquac Res 36(15):1544–1554
Buroker NE (1983) Population genetics of the American oyster Crassostrea virginica along the
Atlantic coast and Gulf of Mexico. Mar Biol 75:99–112
Burreson EM, Ragone Calvo LM (1996) Epizootiology of Perkinsus marinus disease of oysters in
Chesapeake Bay, with emphasis on data since 1985. Oceanogr Lit Rev 12(43):1265
Carnegie R, Burreson EM (2009) Status of the major oyster diseases in Virginia 2006–2008: a
summary of the annual oyster disease monitoring program. Virginia Institute of Marine Science
Report, Gloucester Point; p 20
Carnegie R, Burreson EM (2011) Declining impact of an introduced pathogen: Haplosporidium
nelsoni in the oyster Crassostrea virginica in Chesapeake Bay. Mar Ecol Prog Ser 432:1–15
Castagna M, Chanley P (1973) Salinity tolerance of some marine bivalves from inshore and
estuarine environments in Virginia waters on the western mid-Atlantic coast. Malacologia
12:47–96
Catts WP, Fiedel S, Kellogg DC et al. (1998) Phase III data recovery investigations at 18CV362,
Ashcomb Quarter (historic component) and Awapantop (prehistoric component), Solomons
Naval Recreation Center, Solomons, Calvert County, Maryland. Calvert soil Conservation
District (on file with Maryland Historical Trust, Crownsville, Maryland)
CBF (2016) Chesapeake Bay Foundation. http://www.cbf.org/document.doc?id=656. Accessed 2
Mar 2016
Claudet J, Osenberg CW, Benedetti-Cecchi L et al (2008) Marine reserves: size and age do matter.
Ecol Lett 11:481–489
Coen LD, Bolton-Warberg M, Stephen JA (2006) An examination of oyster reefs as a biologically-
critical estuarine ecosystems. Final report, grant R/ER-10, South Carolina Sea Grant Consor-
tium, p. 214
Coen LD, Brumbaugh RD, Bushek D et al (2007) As we see it: ecosystem services related to oyster
restoration. Mar Ecol Prog Ser 341:303–307
Cohen AN (2011) The exotics guide: non-native marine species of the North American Pacific
coast. Center for Research on Aquatic Bioinvasions, Richmond, California, and San Francisco
Estuary Institute, Oakland. http://www.exoticsguide.org/. Accessed 2 Mar 2016
Cooper SR (1995) Chesapeake Bay watershed historical land use: impact on water quality and
diatom communities. Ecol Appl 5:703–723
Cooper SR, Brush GS (1991) Long-term history of Chesapeake Bay anoxia. Science (Washington)
254(5034):992–996
Craig H (1965) Measurement of oxygen isotope paleotemperatures. In: Tongiorgi E (ed) Stable
isotopes in oceanographic studies and paleotemperatures. CNR Lab. Geol. Nucl., Pisa, pp
161–182
Cressmann KA, Posey MH, Mallin MA et al (2003) Effects of oyster reefs on water quality in a
tidal creek estuary. J Shellfish Res 22:753–762
Cronin TM (1979) Late Pleistocene marginal marine ostracodes from the southeastern Atlantic
coastal plain and their paleoenvironmental implications. Géog Phys Quatern 33(2):121–173
CTDOAG (2016a) Dermo disease, Connecticut Department of Agriculture. http://wwwctgov/doag/

lib/doag/aquaculture/dermopdf. Accessed 17 Mar 2016
CTDOAG (2016b) MSX disease, Connecticut Department of Agriculture. http://wwwctgov/doag/
lib/doag/aquaculture/msxpdf. Accessed 17 Mar 2016
Cummins H, Powell EN, Stanton RJ Jr et al (1987) The size-frequency distribution in palaeoe-
cology: effects of taphonomic processes during formation of molluscan death assemblages in
Texas bays. Palaeontology 29:495–518
Custer JF (1989) Prehistoric cultures of the Delmarva Peninsula. University of Delaware Press,
Newark
Dent RJ (1995) Chesapeake prehistory: old traditions, new directions. Plenum, New York
Edgar GJ, Stuart-Smith RD, Willis TJ et al (2014) Global conservation outcomes depend on marine
protected areas with five key features. Nature 506:216–220
Encomio VG, Stickler SM, Allen SK Jr et al (2005) Performance of “natural dermo-resistant”
oyster stocks-survival, disease, growth, condition and energy reserves. J Shellfish Res
24(1):143–155
Federighi H (1931) Salinity death-points of the oyster drill snail, Urosalpinx cinerea Say. Ecology
12:346–353
Galtsoff PS (1964) The American oyster. Fishery Bull Fish Wildlife Service 64:1–480
Ghosh P, Adkins J, Affek H et al (2006) 13 C–18 O bonds in carbonate minerals: a new kind of
paleothermometer. Geochim Cosmochim Acta 70(6):1439–1456
Grabe, SA, Courtney CM, Lin Z et al (1993) Environmental monitoring and assessment program-
estuaries west Indian province 1993 sampling, vol. 3, executive summary: a synoptic survey of
the benthic macroinvertebrates and demersal fishes of the Tampa Bay estuarine system. Tampa
Bay National Estuary Program Technical Publication 95–12
Grabowski JH, Brumbaugh R, Conrad R (2012) Economic valuation of ecosystem services
provided by oyster reefs. Bioscience 62:900–909
Griffin T (2001) Crepidula cf. plana. Smithsonian Marine Station at Fort Pierce. http://
www.sms.si.edu/irlspec/Crepid_cfplan.htm. Accessed 3 Mar 2016
Grossman EL, Ku T-L (1981) Aragonite-water isotopic paleotemperature scale based on benthic
foraminifer Hoeglundia elegans. Geol Soc Am Abstr Prog 13:464
Halpern BS (2003) The impact of marine reserves: do reserves work and does reserve size matter?
Ecol Appl 13:S117–S137
Harding JM, Mann R (2006) Age and growth of wild Suminoe (Crassostrea ariakensis, Fugita
1913) and Pacific (C. gigas, Thunberg 1973) oysters from Laizhou Bay, China. J Shellfish Res
25(1):73–82
Harding JM, Mann R, Southworth MJ (2008) Shell length-at-age relationships in James River,
Virginia, oysters (Crassostrea virginica) collected four centuries apart. J Shellfish Res
27(5):1109–1115
Harding JM, Spero HJ, Mann R et al (2010a) Reconstructing early 17th century estuarine drought
conditions from Jamestown oysters. Proc Natl Acad Sci U S A 107(23):10549–10554
Harding JM, Mann R, Southworth MJ et al (2010b) Management of the Piankatank River, Virginia
in support of oyster (Crassostrea virginica, Gmelin 1791) fishery repletion. J Shellfish Res
29(4):1–22
Harding JM, Powell EN, Mann R et al (2013) Variations in eastern oyster (Crassostrea virginica)
sex-ratios from three virginia estuaries: protandry, growth and demographics. J Mar Biol Assoc
U K 93(02):519–531
Hargis WJ Jr, Haven DS (1995) The precarious state of the Chesapeake public oyster resource. In:
Hill PL, Nelson S (eds) Toward a sustainable coastal watershed: the Chesapeake experiment,
Chesapeake Research Consortium publication no. 149. Solomons, Maryland, pp 559–584
Hargis WJ Jr, Haven DS (1999) Chesapeake oyster reefs, their importance, destruction and
guidelines for restoring them. In: Luckenbach MW, Mann R, Wesson JA (eds) Oyster reef
habitat restoration: a synopsis of approaches. Virginia Institute of Marine Science, Gloucester
Point, pp 329–358
Harke RM, Herbert GS, White NM et al (2015) Sclerochronology of Busycon sinistrum: late
prehistoric seasonality determination at St. Joseph Bay, Florida, USA. J Archaeol Sci 57:98–
108
Haven DS, Morales-Alamo R (1970) Filtration of particles from suspension by the American
Oyster Crassostrea virginica. Biol Bull 139:248–264
Haven DS, Hargis WJ Jr, Kendall PC (1978) The oyster industry of Virginia: its status, problems
and promise (A comprehensive study of the oyster industry in Virginia). Special Papers in
Marine Science, Virginia Institute of Marine Science, Gloucester Point (4)
Heffernan P, Walker R, Carr J (1989) Gametogeneic cycle of three marine bivalves in Wassaw
Sound, Georgia. II. Crassostrea virginica (Gmelin, 1791). J Shellfish Res 8:61–70
Henkes GA, Passey BH, Wanamaker AD Jr et al (2013) Carbonate clumped isotope compositions
of modern marine mollusk and brachiopod shells. Geochim Cosmochim Acta 106:307–325
Hill K (2004) Mercenaria, Smithsonian Marine Station at Fort Pierce. http://www.sms.si.edu/
irlspec/Mercen_mercen.htm. Accessed 3 Mar 2016
Hixon MA, Johnson DW, Sogard SM (2014) BOFFFFs: on the importance of conserving old-
growth age structure in fishery populations. ICES J Mar Sci 71(8):2171–2185
Hoellein TJ, Zarnoch CB, Grizzle RE (2015) Eastern oyster (Crassostrea virginica) filtration,
biodeposition, and sediment nitrogen cycling in two oyster reefs with contrasting water quality
in the Great Bay Estuary (New Hampshire, USA). Biogeochemistry 122:113–129
Horn J (2008) A land as God made it: Jamestown and the birth of America. Basic Books, New
York
Hsieh CH, Reiss CS, Hunter JR et al (2006) Fishing elevates variability in the abundance of
exploited species. Nature 443(7113):859–862
Huntington KW, Eiler JM, Affek HP et al (2009) Methods and limitations of ‘clumped’ CO2
isotope (D47) analysis by gas-source isotope ratio mass spectrometry. J Mass Spectrom
44:1318–1329
Hutchings JA, Reynolds JD (2004) Marine fish population collapses: consequences for recovery
and extinction risk. Bioscience 54(4):297–309
Kasperski S, Wieland R (2010) When is it optimal to delay harvesting? The role of ecological
services in the Northern Chesapeake Bay oyster fisheries. Mar Resour Econ 24:361
Kelso WM (2004) Jamestown rediscovery, vol 8. Association for the preservation of Virginia
antiquities, Charlottesville
Kemp WM, Boynton WR, Adolf J et al (2005) Eutrophication of Chesapeake Bay: historical trends
and ecological interactions. Mar Ecol Prog Ser 303(21):1–29
Kennedy VS (1983) Sex ratios in oysters, emphasizing Crassostrea virginica from Chesapeake
Bay, Maryland. Veliger 25:329–338
Kennedy VS (1989) The Chesapeake Bay oyster fishery: traditional management practices. In:
Caddy JF (ed) Marine invertebrate fisheries: their assessment and management. Wiley, New
York, pp 455–477
Kennedy VS, Breitburg DL, Christman MC et al (2011) Lessons learned from efforts to restore
oyster populations in Maryland and Virginia, 1990 to 2007. J Shellfish Res 30(3):719–731
Kidwell SM (2013) Time-averaging and fidelity of modern death assemblages: building a
taphonomic foundation for conservation palaeobiology. Palaeontology 56(3):487–522
Kim S, O’Neil JR, Hillaire-Marcel C et al (2007) Oxygen isotope fractionation between synthetic
aragonite and water: influence of temperature and Mg2C concentration. Geochim Cosmochim
Acta 71:4704–4715
Kirby MX, Miller HM (2005) Response of a benthic suspension feeder (Crassostrea virginica
Gmelin) to three centuries of anthropogenic eutrophication in Chesapeake Bay. Estuar Coast
Shelf Sci 62(4):679–689
Krantz DE (1990) Mollusk-isotope records of Plio-Pleistocene marine paleoclimate, US middle

Atlantic coastal plain. PALAIOS 5:317–335
Lipcius RN, Stockhausen WT, Seitz RD et al (2003) Spatial dynamics and value of a marine
protected area and corridor for the blue crab spawning stock in Chesapeake Bay. Bull Mar Sci
72(2):453–469
Lotze HK (2010) Historical reconstruction of human-induced changes in US estuaries. Oceanogr
Mar Biol Annu Rev 48:267–338
Lowery DL, O’Neal MA, Wah JS et al (2010) Late Pleistocene upland stratigraphy of the western
Delmarva Peninsula, USA. Quat Sci Rev 29(11):1472–1480
Luckenbach MW, Mann R, Wesson JA (1999) Oyster reef habitat restoration: a synopsis and
synthesis of approaches; proceedings from the symposium. Williamsburg, Virginia
Luckenbach MW, Coen LD, Ross PG Jr et al (2005) Oyster reef habitat restoration: relationships
between oyster abundance and community development based on two studies in Virginia and
South Carolina. J Coastal Res 40:64–78
Mann R (2000) Restoring oyster reef communities in the Chesapeake Bay: a commentary. J
Shellfish Res 19:335–340
Mann R, Evans DA (1998) Estimation of oyster, Crassostrea virginica, standing stock, larval
production and advective loss in relation to observed recruitment in the James River, Virginia.
J Shellfish Res 17:239–254
Mann R, Powell EN (2007) Why oyster restoration goals in the Chesapeake Bay are not and
probably cannot be achieved. J Shellfish Res 26(4):905–917
Mann R, Southworth M, Harding JH et al (2009a) Population studies of the native oyster,
Crassostrea virginica (Gmelin, 1791), in the James River, Virginia, USA. J Shellfish Res
28:193–220
Mann R, Harding JM, Southworth M (2009b) Reconstructing pre-colonial oyster demographics in
the Chesapeake Bay, USA. Estuar Coast Shelf Sci 85:217–222
Martin RE (1999) Taphonomy: a process approach. Cambridge University Press, Cambridge
Massmann WH, Ladd EC, McCutcheon HN (1952) Rappahannock River survey 1951: a biological
survey of the Rappahannock River part II. Virginia Fisheries Laboratory Special Scientific
Report 6
MDDNR (2016) Maryland Department of Natural Resources. http://dnr2.maryland.gov/fisheries/
Pages/oysters/sanctuaries.aspx. Accessed 1 Mar 2016
Menzel RW, Hulings NC, Hathaway RR (1966) Oyster abundance in Apalachicola Bay, Florida in
relation to biotic associations influenced by salinity and other factors. Gulf Res Rep 2:73–96
Meyer DL, Townsend EC (2000) Faunal utilization of created intertidal eastern oyster (Crassostrea
virginica) reefs in the southeastern United States. Estuaries 23:34–45
Miller W (1982) The paleoecologic history of late Pleistocene estuarine and marine fossil deposits
in Dare County, North Carolina. Southeast Geol 23:1–13
Miller HM (1986) Transforming a “splendid and delightsome land”: colonists and ecological
change in the Chesapeake 1607-1820. J Wash Acad Sci 76:173–187
Mirecki J (1990) Aminostratigraphy, geochronology and geochemistry of fossils from late
Cenozoic marine units in southeastern Virginia. Dissertation, University of Delaware
Mirecki JE, Wehmiller JF, Skinner AF (1995) Geochronology of quaternary coastal plain deposits,
southeastern Virginia, USA. J Coast Res 11:1135–1144
NDBC (2016) National Data Buoy Center. http://www.ndbc.noaa.gov/. Accessed 1 Mar 2016
NOAA NCEI (2016) NOAA National Centers for Environmental Information. http://
www.ncei.noaa.gov/. Accessed 1 Mar 2016
North EW, King DM, Xu J et al (2010) Linking optimization and ecological models in a decision
support tool for oyster restoration and management. Ecol Appl 20:851–866
Paolisso M (2007) Cultural models and cultural consensus of Chesapeake Bay blue crab and oyster
fisheries. NAPA Bullet 28(1):123–135
Paolisso M, Dery N (2010) A cultural model assessment of oyster restoration alternatives for the
Chesapeake Bay. Hum Organ 69(2):169–179
Paynter KT, Burreson EM (1991) Effects of Perkinsus marinus infection in the eastern oyster,
Crassostrea virginica: II Disease development and impact on growth rate at different salinities.
J Shellfish Res 10(2):425–431
Paynter KT, Dimichele L (1990) Growth of tray-cultured oysters (Crassostrea virginica Gmelin)
in Chesapeake Bay. Aquaculture 87:289–297
Pelletier D, García-Charton JA, Ferraris J, David G, Thébaud O, Letourneur Y et al (2005) Design-
ing indicators for assessing the effects of marine protected areas on coral reef ecosystems: a
multidisciplinary standpoint. Aquat Living Resour 18:15–33
Puckett BJ, Eggleston DB (2012) Oyster demographics in a network of no-take reserves:
recruitment, growth, survival, and density dependence. Mar Coast Fish 4(1):605–627
Quast WD, Johns MA, Pitts DE Jr et al (1988) Texas oyster fishery management plan. Fishery
management plan series number 1. Texas Parks and Wildlife Department, Coastal Fisheries
Branch, Austin, p 178
Rader EK, Evans NH (1993) Geologic map of Virginia: expanded explanation. Virginia Division
of Mineral Resources, Charlottesville, p 80
Reshetiloff K (2004) Chesapeake Bay: introduction to an ecosystem: Chesapeake Bay Program,
Annapolis, p 1–35
Rice MA, Rheault RB Jr (1996) Food-limited growth and condition index in the eastern oyster,
Crassostrea virginica (Gmelin, 1791), and the bay scallop, Argopecten irradians (Lamarck,
1819). J Shellfish Res 15(2):271–283
Rick TC, Lockwood R (2013) Integrating paleobiology, archeology, and history to inform
biological conservation. Conserv Biol 27(1):45–54
Rick TC, Reeder-Myers LA, Cox CJ et al (2014) Shell middens, cultural chronologies, and
coastal settlement on the Rhode River sub-estuary of Chesapeake Bay, Maryland, USA.
Geoarchaeology 29(5):371–388
Rick TC, Reeder-Myers LA, Hofman CA et al (2016) Millennial-scale sustainability of the
Chesapeake Bay native American oyster fishery. Proc Natl Acad Sci USA 13(23):6568–6573
Rothschild BJ, Ault JS, Goulletquer P et al (1994) Decline of the Chesapeake Bay oyster
population: a century of habitat destruction and overfishing. Mar Ecol Prog Ser 111(1–2):9–
39
Rountree H, Clark W, Mountford K (2007) John Smith’s Chesapeake voyages 1607–1609.
University of Virginia Press, Charlottesville
Russ GR (2002) Marine reserve as reef fishery management tools: yet another review. In: Sale PF
(ed) Coral reef fishes: dynamics and diversity in a complex ecosystem. Academic Press, San
Diego, pp 421–443
Schulte DM, Burke RP, Lipcius RN (2009) Unprecedented restoration of a native oyster metapop-
ulation. Science 325(5944):1124–1128
Shumway SE (1996) Natural environmental factors. In: Kennedy VS, Newell RIE, Eble AF (eds)
The eastern oyster Crassostrea virginica. Maryland Sea Grant College, University of Maryland,
College Park, pp 467–513
Sisson GMA, Kellogg ML, Luckenbach ML et al (2011) Assessment of oyster reefs in Lynnhaven
River as a Chesapeake Bay TDML best management practice. Virginia Institute of Marine
Science. Department of Physical Sciences, Gloucester Point
Southworth M, Mann R (1998) Oyster reef broodstock enhancement in the great Wicomico River,
Virginia. J Shellfish Res 17:1101–1114
Southworth M, Harding J, Mann R (2000) Management of oyster broodstock sanctuaries in the
great Wicomico and Piankatank rivers, Virginia: optimal use of cultch to maximize settlement.
Final report to Virginia Department of Environmental Quality, Coastal Resources Management
Program, p 12
Southworth M, Harding JM, Mann R et al (2010) Oyster (Crassostrea virginica Gmelin 1791)
population dynamics on public reefs in the great Wicomico River, Virginia, USA. J Shellfish
Res 29:271–290
Spencer RS, Campbell LD (1987) The fauna and paleoecology of the late Pleistocene marine
sediments of southeastern Virginia. Bull Am Paleontol 92(327):1–124
Thompson DE (1972) Paleoecology of the Pamlico formation. Dissertation, Rutgers University,

Saint Mary’s County
Thompson VD, Worth JE (2011) Dwellers by the sea: native American adaptations along the
southern coasts of eastern North America. J Archaeol Res 19:51–101
USGS (1998) Fact sheet, The Chesapeake Bay: geologic product of rising sea level.
http://pubsusgsgov/fs/fs102-98/. Accessed 28 Mar 2012
Venturelli PA, Shuter BJ, Murphy CA (2009) Evidence for harvest-induced maternal influences on
the reproductive rates of fish populations. Proc R Soc Lond B Biol Sci 276(1658):919–924
VMRC (2016a) Virginia Marine Resources Commission Oyster Sanctuary Areas. http://mrc.
virginia.gov/regulations/fr650.shtm. Accessed 1 Mar 2016
VMRC (2016b) Virginia Marine Resources Commission Commercial Fishing. http://mrc. vir-
ginia.gov/commercial.shtm. Accessed 1 Mar 2016
VOSARA (2016) Virginia Oyster Stock Assessment and Replenishment Archive.
http://cmap.vims.edu/VOSARA/VOSARA_Viewer/VOSARA.html. Accessed 8 Feb 2016
Ward LW, Blackwelder BW (1987) Late Pliocene and early Pleistocene Mollusca from the James
City and Chowan River formations at Lee Creek mine. Smithson Contrib Paleobiol 61:113–283
Ward LW, Bohaska DJ (2008) Synthesis of paleontological and stratigraphic investigations at
the Lee Creek mine, Aurora, NC (1958–2007). In: Geology and paleontology of the Lee
Creek Mine, North Carolina, IV Virginia Museum of Natural History Special Publication 14,
Smithsonian Institution Press, Washington, DC, pp 325–436
Waselkov GA (1982) Shellfish gathering and shell midden archaeology, Dissertation, University
of North Carolina
Wehmiller JF, Belknap DF (1982) Amino acid age estimates, quaternary Atlantic coastal
plain: comparison with U-series dates, biostratigraphy, and paleomagnetic control. Quat Res
18(3):311–336
Wesson JA, Mann R, Luckenbach M (1999) Oyster restoration efforts in Virginia. In: Luckenbach
MW, Mann R, Wesson JA (eds) Oyster reef habitat restoration: a synopsis of approaches.
Virginia Institute of Marine Science, Gloucester Point, pp 117–130
Wharton J (1957) The bounty of the Chesapeake. Fishing in Colonial Virginia. University of
Virginia Press, Charlottesville
Wilberg MJ, Livings ME, Barkman JS et al (2011) Overfishing, disease, habitat loss, and potential
extirpation of oysters in upper Chesapeake Bay. Mar Ecol Prog Ser 436:131–144
Wilson C, Scotto L, Scarpa J et al (2005) Survey of water quality, oyster reproduction and oyster
health status in the St Lucie Estuary. J Shellfish Res 24:157–165
York LL, Wehmiller JF, Cronin TM et al (1989) Stetson pit, Dare County, North Carolina: an
integrated chronologic, faunal, and floral record of subsurface coastal quaternary sediments.
Palaeogeogr Palaeoclimatol Palaeoecol 72:115–132
Zachary A, Haven DS (2004) Survival and activity of the oyster drill Urosalpinx cinerea under
conditions of fluctuating salinity. Mar Biol 22:45–52
Zimmerman AR, Canuel EA (2002) Sediment geochemical records of eutrophication in the
mesohaline Chesapeake Bay. Limnol Oceanogr 47(4):1084–1093
Zimmerman KM, Pechenik JA (1991) How do temperature and salinity affect relative rates of
growth, morphological differentiation, and time to metamorphic competence in larvae of the
marine gastropod Crepidula plana? Biol Bull 180:372–386
Zimmt JB, Lockwood R, Andrus CFT, Herbert GS (2016) Revisiting growth increment counting
as a method for biologically aging Crassostrea virginica from the U.S. Mid-Atlantic. Geol Soc
Am Abstr Prog 48:212
zu Ermgassen PS, Spalding MD, Blake B et al (2012) Historical ecology with real numbers: past
and present extent and biomass of an imperiled estuarine habitat. Proc R Soc Lond B Biol Sci
279(1742):3393–3400
Coral Reefs in Crisis: The Reliability
of Deep-Time Food Web Reconstructions
as Analogs for the Present
Peter D. Roopnarine and Ashley A. Dineen
Abstract Ongoing anthropogenic alterations of the biosphere have shifted empha-

sis in conservation biology from individual species to entire ecosystems. Modern
measures of ecosystem change, however, lack the extended temporal scales neces-
sary to forecast future change under increasingly stressful environmental conditions.
Accordingly, the assessment and reconstruction of ecosystem dynamics during
previous intervals of environmental stress and climate change in deep time has gar-
nered increasing attention. The nature of the fossil record, though, raises questions
about the difficulty of reconstructing paleocommunity and paleoecosystem-level
dynamics. In this study, we assess the reliability of such reconstructions by
simulating the fossilization of a highly threatened and disturbed modern ecosystem,
a Caribbean coral reef. Using a high-resolution coral reef food web from Jamaica,
we compare system structures of the modern and simulated fossil reefs, including
guild richness and evenness, trophic level distribution, predator dietary breadth,
food chain lengths, and modularity. Results indicate that despite the loss of species,
guilds, and trophospecies interactions, particularly zooplankton and other soft-
bodied organisms, the overall guild diversity, structure, and modularity of the
reef ecosystem remained intact. These results have important implications for the
integrity of fossil food web studies and coral reef conservation, demonstrating that
fossil reef communities can be used to understand reef community dynamics during
past regimes of environmental change.
1 Introduction
Recent years have witnessed a transition in the emphasis of conservation biology

from an initial concern for individual species at risk to habitat preservation and
restoration, and most recently to a focus on entire biological communities and
P. D. Roopnarine () · A. A. Dineen

Invertebrate Zoology and Geology, Institute for Biodiversity Science and Sustainability,
California Academy of Sciences, San Francisco, CA, USA
e-mail: [email protected]; [email protected]

106 P. D. Roopnarine and A. A. Dineen
ecosystems (Franklin 1993; Schwartz et al. 2000; Meffe et al. 2012; Cardinale
et al. 2012; Hewitt et al. 2016). The shift reflects two major developments in
conservation biology and community ecology, the first being an acknowledgement
that species are parts of larger, integrated systems, themselves part of an integrated
biosphere, and that those systems provide services vital to human societies (Mace
et al. 2012). The nascent field of conservation paleobiology has already incorporated
the transition, using multi-taxon approaches as the bases for forecasting models of
risk to entire biotas, and paleocommunity models as tools to examine ecosystem
dynamics under conditions which lie outside of documented ecological experiences,
such as mass extinction (Villeger et al. 2011; Barnosky et al. 2012).
The second development is the growing understanding of communities and
ecosystems as complex and rich dynamical structures which often have profound
impacts on individual species. Communities have typically been viewed as stable
and somewhat static systems (i.e., species interactions are in balanced equilibria)
and that stability promotes complexity, but these notions have been replaced by
the idea that communities are capable of transitions among multiple, alternative
regimes and dynamic equilibria encompassing a range of community parameters
(Holling 1973; Scheffer et al. 2001; Knowlton 2004). Most recently, it has been
suggested that community persistence itself might act as an agent of long-term
selection, with functional structures and interactions appearing repeatedly within
ecosystems over geological time (Roopnarine and Angielczyk 2015). Therefore,
attempts to understand and conserve modern communities, or the most recent con-
temporary states of long-lasting communities, must account for the dynamic nature
of those communities on multiple timescales and under different environmental
circumstances.
The management of modern marine ecosystems undergoing current and future
anthropogenically driven change must be informed by how similar systems have
responded to environmental variations in the past. This can be fulfilled at the present
time with theoretical and experimental studies, in addition to the examination of
the historical and geological records. The latter have the added advantages that the
range of environmental conditions to which ecosystems have been subjected in the
past is far broader than those within the realm of current scientific and societal
experience, and that the possibility of observing alternative dynamic ecosystem
regimes is increased when longer intervals of time are considered. For example,
a study by Pandolfi and Jackson (2006) showed that despite sea level and climate
variability, coral reef communities in Barbados were stable in composition for at
least 95,000 years during the Pleistocene. Examining and comparing today’s reefs
with those of the Pleistocene showed that recent human impacts have resulted in a
coral reef structure different from anything seen in the last 220,000 years. Without
the establishment of baselines for what non-anthropogenically altered reefs look
like, we would not know how altered and degraded today’s marine ecosystems are
in comparison (Knowlton and Jackson 2008).
As the future of oceanic ecosystems is still very much uncertain, deep-time
studies provide our best proxy for what we can anticipate in the Earth’s near future
(Harnik et al. 2012). For example, major questions in marine conservation and
Coral Reefs in Crisis: The Reliability of Deep-Time Food Web Reconstructions. . . 107
global change biology center on how marine ecosystems will respond to environ-
mental stresses and/or large disturbances, and what makes some communities more
vulnerable to perturbation than others (Seddon et al. 2014). These concerns are
driven by the degraded state of ecosystems in today’s oceans, resulting in the decline
of species at an alarming rate and unprecedented magnitude (Cardinale et al. 2012).
Coral reef ecosystems are of particular interest because one quarter of all marine
species may be found in these threatened marine communities (McClanahan 2002;
Hoegh-Guldberg et al. 2007; Carpenter et al. 2008; De’ath et al. 2012).
Preserving the Past
One of the foundations of modern paleontology is understanding the impacts of

preservation on interpretations of the fossil record, and the ability to use that infor-
mation to reconstruct the past (Kidwell and Bosence 1991; Kidwell 2001). Current
paleontological studies rarely ignore factors of bias, such as selective preservation
among taxa, biased preservation of parts of taxa, and outcrop dimensions. Those
concerns must be extended to conservation-based paleobiological approaches. A
major complication, however, is that in addition to biases of preservation and
discovery, the information conveyed by integrated systems may also be biased by
the manner in which integration is preserved. For example, the extent to which prey
richness and abundances are preserved surely influences the interpretation of the
dynamics of molluscan drilling predation, and differences of generation times bias
relative abundances (Leighton 2002; Vermeij and Leighton 2003). It is therefore
important that we analyze the ways in which fossilization potentially biases our
interpretation of community and ecosystem function and ecology. In this chapter, we
explore the ways in which the retention and loss of data affect our reconstructions
of attributes important to measuring system dynamics. Using a high-resolution food
web of a modern Jamaican coral reef ecosystem (Roopnarine and Hertog 2013),
we simulate its fossilization and document subsequent changes of taxon richness
and interspecific interactions. We then measure how those changes alter important
quantitative measures of food web structure and function, specifically predator
dietary breadth, the web’s food chain lengths and trophic level distribution, guild
richness and evenness, and the modularity of the system. Finally, we assess the
reliability with which we could reconstruct those features of a paleocommunity
by assuming that our starting point is the fossilized reef system, then constructing
a guild-level representation of the fossil community and contrasting the implied
trophic structure to that of a similarly resolved modern reef.
Endangered Coral Reefs
Threats to modern coral reef ecosystems bear several similarities to conditions

from intervals in the Earth’s past, e.g., increasing CO2 , ocean acidification, and
oceanic temperatures. Coral reefs may thus be one of the best proxies we have
for predicting future ecological changes and biodiversity loss in the oceans. The
ocean is a large sink for anthropogenic carbon dioxide today (30% of total), and
a predicted increase of CO2 concentrations in the coming century is expected to
adversely affect marine organisms in a multitude of ways, particularly by decreasing
biocalcification (Kleypas et al. 1999; Sabine et al. 2004). Evidence already exists for
such a scenario, which is apparent in decreasing calcification rates of individual
species and communities, especially in coral reef environments (Albright et al.
2016). In addition, global sea surface temperatures (SST) have risen in the past
century (0.4–0.8 ı C), with warming predicted to accelerate in the near future
(Joachimski et al. 2012). Increasing temperatures have had a large effect on the
marine realm, affecting ocean circulation, benthic and planktonic diversity and
abundance, productivity, and overall invertebrate physiology (Walther et al. 2002;
Parmesan 2006). Coral reefs are extremely sensitive to changes in temperature and
pH, frequently expelling their zooxanthellae (photosynthetic algal symbionts) when
physiologically stressed, resulting in coral bleaching (Hoegh-Guldberg et al. 2007).
During the writing of this chapter (March–April, 2016), Australia’s Great Barrier
Reef system is experiencing unprecedented bleaching, with approximately 95% of
reefs in the ecosystem being affected (A.C. of Excellence Coral Reef Studies 2016).
However, corals are not the only organisms under threat in reef ecosystems. Other
invertebrates and vertebrates are also in decline due to overfishing, degradation of
their coral habitat, and pollution (Nyström et al. 2000; Jackson et al. 2001; Bellwood
et al. 2004; Pandolfi et al. 2005; Hardt 2009; Cramer et al. 2015).
Caribbean reefs in particular have suffered significant losses, with reports of
considerable reduction (80%) of coral reef cover since the 1970s in addition to
frequent human-induced degradation (Gardner et al. 2003; Mora 2008). Further-
more, in the early 1980s, a massive disease-induced die-off of the urchin, Diadema
antillarum, resulted in a macroalgal bloom that persists to this day (Lessios
et al. 1984; Andres and Witman 1995; Mumby et al. 2006). The less biodiverse
algal-dominated state of Jamaica’s reefs in particular is exacerbated by historical
overexploitation of herbivorous fish (Hardt 2009). It appears, however, that Diadema
has been functionally replaced to some extent by parrotfishes, highlighting the
potential importance of functional redundancy in coral reef and other ecosystems
(Andres and Witman 1995; Mumby et al. 2007; Hardt 2009; Nash et al. 2016).
Regardless, local increases of parrotfish in various areas of the Caribbean have not
reversed the algal-dominated state of the reefs, although studies have indicated that
when parrotfish are able to escape overfishing inside marine reserves, they were
able to increase grazing intensity and reduce macroalgal cover (Mumby et al. 2006).
Recent reports do show a slight recovery of Diadema and a subsequent increase in
coral recruitment and survivorship, though not nearly as rapidly as expected, with
current populations at only 11.62% of their premortem density (Idjadi et al. 2010;
Lessios 2016). Jamaican reefs have come to represent an unhealthy reef system,
frequently disturbed by anthropogenic (e.g., overfishing and pollution) and non-
anthropogenic (i.e., disease, hurricanes) stresses (Hughes 1994).
Today, climate change and other human-induced disturbances are progressing at
an unparalleled and alarming rate, increasing the need for studies that use novel
analogs and proxies from Earth’s past. Data from the fossil record give us a
long-term historical perspective from which we can test the influence of extreme
environmental conditions on ecological dynamics and community structure. Fossil
food web studies are particularly needed due to recent studies indicating that
functional extinction of apex predators, large herbivores, or ecosystem engineers
in coastal ecosystems may occur several decades to centuries after the onset of
ecosystem degradation, resulting in potential collapse of trophic webs (Jackson
et al. 2001; Lotze et al. 2006). Thus, deep-time studies have much to contribute
to evaluating potential losses of biodiversity, stability, and sustainability in marine
ecosystems due to current and future climate change. For example, Aronson
et al. (2007) determined that Antarctic benthic food web structure was established
41 million years ago when the climate was much cooler, resulting in predators
(i.e., sharks, crabs) being pushed from Antarctic waters. An increase in current
temperatures as a result of climate change may result in the the invasion of such
durophagous predators, profoundly affecting benthic food web structure. As such,
paleoecological perspectives are vital to modern conservation strategies in order
to establish how we can buffer anthropogenic and climate-related stress in fragile
modern ecosystems currently and in the future.
2 Fossilizing a Coral Reef
The Jamaican coral reef food web describes species interactions from coral reef and
adjacent seagrass habitats within Jamaica’s marine geopolitical territory, including
the offshore Pedro Bank (Fig. 1). The food web model is an amalgamation of data
drawn from multiple specific localities. Differences between localities, however,
are expected to be ephemeral and changing; this is supported by the significant
compositional overlap both among localities within Jamaica, as well as with similar
food webs constructed for the neighboring Cayman Islands and Cuba (Roopnarine
and Hertog 2013). This northern Caribbean region represents a common regional
species pool, and the Jamaican dataset is thus a sample of that pool, integrated over
the spatial variation present among the Jamaican localities. Taxon composition was
determined from published compilations and reports, including Fishbase (FishBase
2016), and the REEF (Reef Environmental Education Foundation Volunteer Survey
Project Database 2011) survey database compiled up to 2011. Details of sources and
methods used to determine species interactions are given in Roopnarine and Hertog
(2013), and the complete data are archived in the DRYAD database (Roopnarine and
Hertog 2012). Together, the resources represent approximately 50 years of data, a
Fig. 1 Map of the Greater Antilles, showing the regions covered by the coral reef data (adapted
from Roopnarine and Hertog (2013)). Data treated in this study are from Jamaican geopolitical
areas, outlined in red
temporal resolution which is likely much finer than any available in the fossil record.
Comparison to sub-fossil and archaeological data from Jamaica, however, suggest
that compositionally the data would be congruent with fossil data time-averaged
on at least a millennial scale (Hardt 2009). The complete dataset documents the
interactions of 749 species in the northern Caribbean, ranging from single-celled
protists to multicellular macrophytes and metazoans, of which 728 have records
of occurrence in Jamaica. Multiple species were collapsed into trophospecies
when they shared exactly the same prey and predators (i.e., had exactly the same
interactions in the food web). This process resulted in 265 trophospecies, of which
249 are present in Jamaica, with a total of 4105 inter-trophospecific predator–prey
interactions.
Fossilization of the community was simulated with a simple binary filter at the
genus level. The occurrence of the genus to which each taxon in the food web is
assigned was checked for occurrence in the fossil record using the Paleobiology
Database (downloaded September, 2014) and Sepkoski’s Compendium of Fossil
Marine Genera (Sepkoski et al. 2002) and was considered fossilized if the genus
has a documented fossil record. The presumption is based on the premise that
characteristics which promote fossilization, such as morphology, life habits, and
habitat, can be extended to all members of the genus. If the genus does not have a
documented fossil record, then the food web taxon was eliminated from the dataset.
This simulated fossilization resulted in a reduced dataset comprising 433 species
aggregated into 172 trophospecies, and 1737 inter-trophospecific interactions.

We tested whether the likelihood of species fossilization is uniform among the
trophospecies, suspecting that biases would exist because of the usual vagaries
of fossilization, including biases against soft-bodied taxa, small body size, easily
disarticulated skeletons, and depositional environment. The expected number of
species fossilized in a trophospecies is estimated simply as the species richness of
the trophospecies times the overall fraction of preservation for all the trophospecies,
E.preservation/ D .species richness/ 433=728 (1)
Thus, the probability that the number of species lost (i.e., not fossilized) from a
single trophospecies is consistent with uniform probabilities of non-preservation
among all trophospecies is given by the hypergeometric probability
1
L SL S
P.observed/ D (2)
lX s X lX sX
where S is food web species richness, L is the number of species lost during
simulated fossilization, sX is the richness of trophospecies X, and lX is species loss
from X during fossilization. A hypothesis of uniform, unbiased levels of preser-
vation could not be rejected for 130 trophospecies. The remaining trophospecies
had either improbably low or high levels of preservation (Fig. 2), and all can be
explained by body composition and body size. For example, trophospecies with
unexpectedly poor preservation include nanno-zooplankton (p = 0.0006), epibenthic
sponges (p D 1:02 108 ), micro-zooplankton trophospecies such as cyclopoid
copepods (p D 0:0003), gorgonians (p D 8:91 106 ), and sphenopid and zoanthid
anthozoans (p D 8:91 106 ). In contrast, trophospecies with unexpectedly good
preservation include mixotrophic scleractinian corals (p D 3:95 1015 ) and soft-
sediment dwelling, infaunal suspension feeding bivalves (p D 0:0001).
A loss of biotic interactions is expected to accompany the loss of taxa with
lower probabilities of preservation. We examined the distribution of lost interactions
among the trophospecies, expecting that the loss would be a function of the
number of interactions of a trophospecies; that is, more connected trophospecies
would lose a proportionally greater number of interactions. Biased preservation of
species within some trophospecies, however, would also render other trophospecies
more poorly connected, or disconnected, than expected based on their numbers of
interactions alone. We therefore examined incoming, that is, predatory interactions
for each trophospecies using a hypergeometric probability similar to that explained
above for taxon preservation. Given that 1737 trophospecies interactions out of
4105 were preserved, the expected number of interactions (prey) retained by a
trophospecies is estimated as
E.preservation/ D .interactions/ 1737=4105 (3)

Fig. 2 Expected and observed losses of taxa (left) and interactions (right). Values after simulated
fossilization are plotted against the modern values. The expected values, based on an assumption
of a uniform probability of preservation among all trophospecies, is plotted as a line. Grey regions
represent 1.96 standard deviations around the expected values (see Appendix 1)
The probability of observing the number actually lost is then

1
L IL I
P.observed/ D (4)
lX iX lX iX
where I is the total number of interactions in the food web, iX is the number of
incoming interactions, or prey, of trophospecies X, and L is the total number of
interactions lost. Trophospecies that lost more incoming interactions (i.e., predatory
interactions) than expected include scleractinian corals (p D 0:003 for both
mixotrophic and fully heterotrophic trophospecies), the butterflyfish Chaetodon
capistratus (p D 0:004), and a multi-taxon trophospecies including the fish blue
chromis (Chromis cyanea), brown chromis (C. multilineata), and royal gramma
(Gramma loreto) (p D 0:003). Zooplankton constitute a major or sometimes
total portion of the diet of all these taxa, and the significantly low preservation
probabilities of zooplankton trophospecies result in their predators being either
poorly connected or completely disconnected from the fossil food web. In contrast,
a number of fish trophospecies retain more interactions than expected, primarily
because benthic invertebrates with hard body parts, and hence high probabilities
of preservation dominate their diets. This set of consumers includes the seabream
(Diplodus caudimaculatus) (p D 0:00007), the pufferfish (Sphoeroides spengleri)
(p D 3:1 107 ), and the squirrelfish (Sargocentron vexillarium) (p D 0:00004).
The reconstruction of taxon level trophic properties is obviously affected by the
differential probabilities of preservation of taxa, based primarily on body com-
position and body size, and consequently the preservation of trophic interactions.
The low preservation of a major group such as the copepod zooplankton not only
generates a negative bias against the inferred dietary breadths of their predators,
but also creates a positive bias toward taxa whose prey have exceptionally high
probabilities of preservation. These results are not surprising given what we know
about the vagaries and biases of fossil preservation, but the relevant question here is
what are the biases created when viewing the community as an integrated system,
and not merely as a collection of taxa and interactions. We therefore examined
several system-level properties, including the distribution of dietary breadths (“in-
degree” distribution), food chain lengths, trophic levels, modularity, and guild
structure and diversity.
Dietary Breadth
The distribution of dietary breadths is the distribution of the number of prey

species per consumer species, or in the case of aggregated food webs such as
this, the number of trophospecies preyed upon by each consumer trophospecies.
This distribution is commonly referred to in the food web literature as an in-
degree distribution, where degree refers to the number of interactions per species,
trophospecies, or trophic guild. Surveys of these distributions (Williams and
Martinez 2000; Dunne et al. 2002; Roopnarine 2009) show that the overwhelming
majority are “decay” distributions, where the density of the distribution is biased
toward low dietary breadths, meaning that there are more taxa with specialized
diets in the web than there are species with generalist diets. The precise nature
of a web’s distribution is unknowable unless all interactions have been recorded,
but estimates from a variety of webs and models suggest that the distributions may
be exponential, power law, or a mixture of decay distributions (Williams 2010).
One interesting feature is that the tails of these distributions are long (hyperbolic)
and occupied by generalist species with broad diets. Topological analyses of food
web networks suggest that the presence of such highly connected species provides
community robustness against the cascading effects of random extinctions because
those taxa are less likely to lose all their connections and thus provide some
insurance to the community (Dunne et al. 2002). That conclusion must be tempered,
however, by a corresponding weakness of many of those links because a general
theoretical conclusion is that strong interactions destabilize species interactions and
hence communities in general (May 1973; McCann et al. 1998; McCann 2000).
Furthermore, empirical measures of interaction strengths between modern species
show that the majority of those interactions are relatively weak (Paine 1980).
Reconstructions of Lagerstätten food webs, with presumably high probabilities of
preservation, have also yielded hyperbolic distributions (Dunne et al. 2008). Models
of paleocommunity food web dynamics have used hyperbolic decay distributions,
that is, fat-tailed distributions, but the majority of interactions in the community
models are typically weak (Roopnarine et al. 2007; Mitchell et al. 2012; Roopnarine
and Angielczyk 2015).
Fig. 3 Trophic in-degree, or dietary breadth, distributions. The number of predatory species with
a specific number of prey or resource species (y-axis) is plotted against that specific number of prey
(x-axis). Both the modern (open circles) and fossil (open squares) distributions are fit significantly
by logarithmic power functions
Given the above observations of taxon and interaction loss during fossilization,
however, can paleocommunities be reliable sources for the reconstruction of in-
degree distributions? A comparison of the modern and simulated fossil distributions
for the coral reef community shows that, at least in this case, a reconstruction based
on the paleocommunity would be very accurate (Fig. 3). The modern distribution
can be described best with a power function, ln.p/ D 3:48 0:78Œln.r/ where r is
the number of prey trophospecies per consumer trophospecies, that is consumer in-
degree, and p is the number of consumer trophospecies of that degree (F.1; 54/ D
130:74, r2 D 0:702, p < 0:0001). The fossil distribution is essentially identical,
ln.p/ D 3:05 0:78Œln.r/, (F.1; 37/ D 95:41, r2 D 0:713, p < 0:0001), and
statistically indistinguishable (Student’s t D 0:01, p D 0:990). Thus, despite the
loss of taxa and interactions, including the structurally and functionally important
zooplankton trophospecies, one can predict the number of interactions per fossil
trophospecies with a high degree of confidence.
Trophic Chains and Levels
Trophic level is an important characteristic in both species and system con-

servation efforts (Pace et al. 1999; Sergio et al. 2005; Becker and Beissinger
2006). It is generally accepted that high trophic level species, that is, predators
with superior capabilities and few predators of their own, are often at greater
risk from anthropogenic actions. The reasons for this vary between marine and
terrestrial habitats. In the former, high trophic level species are often targeted for
harvesting because of high biomass, while terrestrially they are viewed as threats to
domestic livestock or competitors for resources, and are often sensitive to habitat
alteration or destruction. The impact of overfishing on oceanic predators has been
devastating, and populations throughout the Caribbean are in decline (Jackson et al.
2001; Hardt 2009). Scleractinian reefs, however, originated and evolved during a
series of changing predatory clades, including Mesozoic marine reptiles, and the
radiations of modern chondrichthyans, teleosts, and teleost clades specialized to
reef environments (Wood 1998). The potential richness and complexity of coral reef
systems might therefore vary with the relative proportions of different life history
strategies present, and different suites of trophic levels would represent alternative,
yet persistent ecosystem regimes.
Trophic level is broadly understood to describe, in some manner, the position of a
species within a food web, if species are arranged hierarchically from basal, primary
producers, up to apex predators. Species positions in the food web determine the
flow of energy to their populations, and the feasibility of their persistence depends
on a productive basal component, sufficient to support all those populations further
up the food chains (Williams and Martinez 2004). Energy is lost thermodynamically
along each step of a food chain, however, because of respiration and the inefficiency
of energy assimilation from consumed material, and this loss is thought to constrain
the lengths of food chains (Yodzis 1984). Food chain length may also be constrained
by the likelihood of decreasing dynamical stability as the length of a chain grows
(Pimm and Kitching 1987), and by numerous factors specific to a particular food
web, its composition and environmental context (Post 2002).
All these limitations would restrict the diversity and complexity of predation
within a community, but there is variability driven by organismal variation. Yodzis,
thinking in terms of webs rather than chains, suggested that increased productivity
would make both increased chain length and predator diversity feasible (Yodzis
1989). This in turn means that any predator toward the top of a food chain would
have to be a super-generalist, spreading its effort over a greater diversity of prey, in
order to garner sufficient energy. This effort could be focused on other predators near
the top, or by omnivory, where the predator would feed at multiple levels of the chain
thus circumventing some of the thermodynamic loss. Finally Pimm (1982), noting
the decreasing trend of efficiency of energy conversion moving from ectothermic
invertebrates, to ectothermic vertebrates, to endothermic vertebrates, suggested
that invertebrates should support longer chains (Pimm 1982). The generally larger
population sizes of smaller species, and greater rates of invertebrate population
growth, would also increase the likelihood of dynamical stability (Post 2002).
Variability of complexity is driven in modern Caribbean reefs largely by
anthropogenic factors (Jackson et al. 2001; Bascompte et al. 2005), but over longer
timescales this variability could be a function of clade diversity dynamics and
macroevolutionary trends. Identification of any such trends, and the establishment
of reliable historical baselines for modern reefs, depends on whether trophic chains
and levels can be reconstructed from fossil and sub-fossil data. As such, here we
introduce a method for calculating trophic position from food web network data,
network trophic level, and use it to compare the distribution of trophic levels in
the modern and fossilized reefs. The network trophic level (ntl) of a species or
trophospecies is the average shortest distance of its prey species to a primary
producer. Primary producers are assigned an ntl of 1.0, and the ntl of a consumer
species i is calculated as
S
1X
ntli D 2 C aij lj (5)
ri jD1
where ri is the number of prey species of i, aij D 1 if i preys on species j, and zero
otherwise, and lj is the shortest path length of j to a primary producer. Ntl differs
from the prey-averaged trophic level of Williams and Martinez (2004) by a factor
of one for consumers. Trophic level is measured in various ways, ranging from
simplistic integer values corresponding to discrete categories such as “primary,”
“secondary,” “tertiary” consumer, and so forth, to inferences of the number of steps
to a consumer as measured by stable isotopic composition of consumer tissues. A
common measure used for fish is fractional trophic level (ftl). Ftl is based on the
proportion of specific prey species in a consumer’s diet and is a weighted average of
the prey ftl values. Romanuk et al. (2011) presented a global database of empirically
measured and inferred ftl values for fish. Using that database, Roopnarine (2014)
showed that ftl and ntl are correlated significantly (Fig. 4), concluding that ntl is a
reliable measure of trophic level, and that a significant proportion of trophic level
variance is based only on position in the food web.
Consumer ntl ranges from 2 to 5.5 in the Jamaican reef food web, with a mean
value of 2.89 which increases to 3.29 if primary consumers (ntl = 2) are excluded.
Almost all trophospecies with ntl values between 4 and 5 are vertebrates, primarily
predatory fish with a broad range of body sizes. Carnivorous ophiuroids is the single
invertebrate trophospecies in this range, the included species being both benthic
deposit feeders and polychaete carnivores (e.g., Ophiocoma echinata). All trophic
levels of five and above, however, are occupied by invertebrate trophospecies,
specifically corallivorous polychaetes and gastropods, e.g., Hermodice carunculata
and Coralliophila caribbaea (ntl = 5.0), and gastropod predators of polychaetes,
e.g., Conus regius (ntl = 5.5). These very high trophic levels are the result of very
long food chains which extend the phytoplankton–zooplankton food chain, but
which are also very simple, meaning that the prey breadths of trophospecies along
the chain tend to be low. The high ntl invertebrate taxa are also not apex predators,
but instead are subject to predation by trophospecies that feed at multiple trophic
levels and along multiple food chains. For example, C. regius is preyed upon by a
trophospecies of carnivorous crustaceans, including Penaeus duorarum, with a ntl
of 3.18. The assessment of trophic level is therefore complicated by the branching
topology of a food web. Ntl, as perhaps with other trophic level measures such as
Fig. 4 The relationship between network trophic level and fractional trophic level fit with a
reduced major axis function
stable isotopic composition (Post 2002), is thus not a linear ordination of taxa among
multiple food chains. They are, however, reliable measures of the average distances
of consumers from the productive base of a web. The ntl measures reported here
are consistent with other observations that invertebrates tend to occupy longer food
chains than vertebrates (May 1983; Yodzis 1984; Pimm et al. 1991). Moreover,
although high ntl vertebrate predators frequently prey on invertebrates along long
food chains, those vertebrates also feed in much shorter food chains and thus
again have lower ntl values. For example, the top predator Caribbean reef shark
Carcharhinus perezi is of ntl 3.86. The fact that the most powerful predators in
a food web network will not be the furthest removed from primary producers is
a warning against the common practice of simplifying food web structures into
discrete, or nearly discrete trophic levels.
The ntl distribution of the modern reef is significantly higher than the simulated
fossil reef (Fig. 5; Kruskal-Wallis, 2 D 30:70, p D 0:0001). Maximum ntl in the
modern reef is 5.5, but only 4 in the fossil reef. The reduction could be the result of
a loss of high ntl trophospecies, but there is no significant difference between the ntl
distributions of those trophospecies that are preserved and those that have no fossil
representation (Student’s t D 1:612, p D 0:108). The difference of ntl distribution
is in fact attributable to the reduction of ntl of preserved trophospecies. Many of
those trophospecies are specialized predators with poorly preserved prey, such as
carnivorous gastropod predators of gorgonians or polychaetes, or zooplanktivorous
fish. Those consumers’ ntl values have collapsed to two, implying incorrectly that
they are primary consumers. Other trophospecies ntl values are reduced because
6
Network trophic level
Network trophic level

5
5
4
4
3
3
2
2
0 .5 1 1.5 2 0 .5 1 1.5 2
Density Density
Fig. 5 Network trophic level (ntl) distributions for the modern (left) and simulated fossil (right)
food webs. The fossil distribution is truncated, with a maximum ntl value of 4 in contrast to the
modern web’s 5.5. Primary producers (ntl = 1) are excluded from the plots
they have no preserved prey and their ntl values are 1.0. Thus, the poor preservation
of key trophospecies such as zooplankton and soft-bodied taxa has a significant
impact on estimates of fossil trophic level, and the impact will always lead to an
underestimation of trophic level and food chain lengths. There is essentially no
way in which data can be recovered to address this issue, and the evolution and
historical baselines of trophospecies trophic level in coral reef ecosystems, and
indeed all paleoecosystems, are obscured by biases of preservation. It is conceivable,
however, that information is retained, and can be inferred, at higher levels of system
organization that might be less subject to bias, and we turn to those in the following
sections.
Modularity
A module is a subset of nodes within a network which have more interactions

with each other, and fewer with other nodes, than would be expected if inter-
actions occurred at random (Newman 2006). Obvious examples occur in social
networks where persons within a family or circle of friends may represent a
module within the larger society. A module within a food web would therefore
comprise species sharing more interactions with each other than they do with
other species in the community, a condition also referred to in the food web
literature as compartmentalization. In offering this definition, we distinguish this
usage of “module” and “modularity” from an alternative use which refers to pairs
or trios of interacting species without regard to their other interspecific interactions
(e.g., Kondoh (2008)). We also discount trivial compartmentalizations that result in
discrete, non-overlapping sets of interactions, for example, as might occur across

strong habitat boundaries. Such compartments are in effect independent food webs.
We focus instead on compartments embedded in, and sharing interactions with the
rest of the network.
Interest in food web modularity stems from May’s theoretical work (May 1973)
on the relationship between the local stability of a community (i.e., its ability
to return to a static equilibrium after a minor perturbation) and the community’s
richness, connectance (the density of interspecific interactions), and average inter-
action strength. May noted that for random networks or food webs, the probability
of stability decreases with an increase of any of those community parameters,
calling into question the long-held hypothesis of a positive relationship between
community complexity and stability. May also pointed toward possible “solutions”
to this seeming paradox, including a hypothesis that many food webs might consist
of compartments, or modules, which would increase the probability of stability. We
suggest here that in addition to stable dynamics, modular structure of a food web
would indicate ways in which the energy supplied to a community is partitioned, and
the extent to which the community could then be viewed as energetically integrated
or compartmentalized.
Results of subsequent searches for food web modularity have been mostly
equivocal (Dunne et al. 2005), possibly because of the relatively low resolution
of many current food web datasets. Another probable cause is the difficulty of
identifying modules in highly resolved, complex food webs. In a community
of S species, the number of modules could range from 1 (all species within a
single module) to S (each species represents a separate module). All modular
arrangements of size between these two extremes would represent all possible
combinations of partitioning schemes of S, and that number grows very rapidly with
increasing S! There are no objective and exact analytical approaches short of an
exhaustive search of all those possible combinations of species clusters. Heuristic
approaches do exist, however, and here we applied the modularity algorithm due to
Newman (2004), which is an optimization of the property Q, where Q = (fraction
of interactions within modules) (expected fraction of such interactions). We used
a fast approximation of Newman’s algorithm developed by Blondel et al. (2008),
as implemented in the network visualization software Gephi (Bastian et al. 2009),
to generate repeated modularity measures on the same network. We further tested
the null hypothesis that the community network is indistinguishable from a random
network of equal species richness and number of interactions, by comparing the
food web to such equivalent random networks using the netcarto program (Guimerá
and Amaral 2005, 2005a; Guimerà et al. 2007).
Both the netcarto and Blondel et al. implementations support the sub-division of
the food web into four modules (Fig. 6), a relative modularity value of 0.282, and
with the netcarto randomizations failing to support the null hypothesis of no unique
modularity (z-test; mean relative modularity of 1000 random networks = 0.139,
sd = 0.002, p < 0:0001). The coral reef community thus comprises four modules or
sub-communities, and that substructure is highly unlikely to arise by any random
assembly of species and interactions. The collapse of 249 trophospecies into
Fig. 6 Modularity of the modern (left) and simulated fossil (right) food webs. Numbers indicating
each module are explained in the text, and correspond to the same module between the webs.
Visually, modules are identifiable by the density of within-module interactions which are shown on
the periphery, in contrast to between-module interactions which cross the interior of the circularly
arranged webs
four modules is striking, but do the modules make any sense ecologically? Our
expectation that the food web would be partitioned among different conduits of
energy flow is partially correct, but factors of habitat and modes of life also play a
role. Figure 7 illustrates summaries of the modules, presented as food chains and
featuring examples of the trophospecies within each module.
Module 0 (modules are numbered here 0, 1, 2, and 3), composed of 34
trophospecies, is a basal energetic module in the sense that it includes all the
macrophytic trophospecies, including macroalgae and seagrasses. It also includes
the sponges and herbivorous fish as well as low trophic level omnivores that
consume macroalgae and benthic sponges, such as the parrotfish Scarus iserti.
The highest trophic level trophospecies in the module includes piscivores that
specialize on benthic grazing fish, e.g., the tiger grouper Mycteroperca tigris, and
generalist predatory piscivores, including the apex predator Caribbean reef shark
Carcharhinus perezi.
Module 1, comprising 66 trophospecies, is dominated by benthic food chains,
including benthic foraminifera and metazoan deposit feeders, and grazers such as
lytechinid echinoids. Food chains are extended through the module primarily by
omnivorous and carnivorous benthic macrocrustaceans, e.g., pagurid crabs, and
both benthic invertebrate predators and teleost predators of those taxa, including
gonodactylid stomatopods. The food chains are again capped within the module by
high trophic level piscivores such as the king mackerel Scomberomorus cavalla.
Module 2 does not have a notable primary producer base, containing only
encrusting coralline algae, but the module size is nevertheless substantial, com-
prising 74 trophospecies. Much of this module’s diversity is dominated by benthic
macroinvertebrates, including herbivorous, omnivorous, and carnivorous grazers,
Fig. 7 Representative food chains within each of the modern community’s modules. From left
to right, trophospecies, or representative species, selected from each module are: M0—bacteria,
the sheet macroalga Ulva, epibenthic basket sponge Xetospongia, parrotfish Scarus iserti, tiger
grouper Mycteroperca tigris, and Caribbean reef shark Carcharhinus perezi; M1—foraminifera
Globigerina and Archaias, the sea urchin Lytechinus, hermit crab Pagurus brevidactylus, seabream
Archosargus rhomboidalis, mantis shrimp Gonodactylus, margate Haemulon album, mutton snap-
per Lutjanus analis, and king mackerel Scomberomorus cavalla; M2—coralline algae Hydrolithon,
queen conch Strombus gigas, hairy crab Pilumnus marshi, reef octopus Octopus briareus,
purplemouth moray Gymnothorax vicinus, and red hind Epinephelus guttatus; M3—epiphytic
diatoms, copepods (Podon sp. shown), staghorn coral Acropora cervicornis, reef butterflyfish
Chaetodon capistratus, and Atlantic spadefish Chaetodipterus faber
such as the queen conch Strombus gigas, the hairy crab Pilumnus marshi, and the
reef octopus Octopus briareus, respectively. Higher trophic level predators in the
module are primarily fish specialized to predation in the coral habitat, such as moray
eels and their predators.
Module 3, a plankton-based module, is the largest module with 75 trophospecies

and includes the major phytoplankton and zooplankton trophospecies, as well as
major benthic and pelagic zooplanktivores, such as corals and the Atlantic spadefish
Chaetodipterus faber. There are fewer high trophic level predators in the module,
and its trophospecies richness is a function of the great diversity of trophic strategies
employed by zooplankton and zooplanktivores.
Although the community is modular, and the modules are highly interpretable
in ecological terms, the modules are united by high trophic level predators, such
as the Caribbean reef shark. This predator, though assigned to a single module,
preys on species in all modules. An examination of other large shark species in the
northern Caribbean region, as documented in similar food webs from the Cayman
Islands and Cuba (Roopnarine and Hertog 2013), reveals similar broad, generalist
predation. However, most of those species have been exterminated on local scales
by overfishing and are rare in regional reef systems today. Prior to their extirpation,
modularity of the Jamaican reef would have been weaker because of greater
high trophic level, cross-module predation. Therefore, in spite of the theoretical
expectations of a positive role for modularity, measurements based on modern,
anthropogenically altered communities might result in overestimates of natural
modularity. This possibility could be addressed with historical and paleontological
records, but only if modularity is preserved in spite of the losses of taxa and
interactions. We therefore repeated the modularity analyses using the simulated
fossilized community.
Modularity analysis of the fossilized food web yielded four modules, and a mod-
ularity measure of 0.287. Comparison to 1000 random networks failed to support
a null hypothesis of insignificant modularity (z-test; mean relative modularity of
random networks = 0.169, sd = 0.002, p < 0:0001). The number of modules equals
that for the modern food web, and the modularities are very similar. The important
question though is how comparable are the modules in the two webs? To answer this,
we compared the membership of preserved trophospecies in both modern and fossil
modules, examining the distribution of fossil modules within each modern module
(Fig. 8). There is a single dominant fossil module occurring in each modern module,
suggesting significantly that those fossil modules are equivalent to the modern
modules (Chi-square test, 2 D 273:606, p < 0:0001). Accepting this hypothesis
results in only 13 fossil trophospecies not corresponding to their presumed modern
modules, a correct classification rate of 92.4%. Remarkably, despite the loss of
trophospecies and interactions, the modular structure of the modern reef is expected
to be preserved in fossil reef communities.
3 Guild Structure and Diversity
Food web reconstructions often aggregate species into groups on the basis of
presumed ecological similarities, such as “benthic macroinvertebrates,” and “salt
marsh plants,” etc. Aggregation is perhaps justified when more precise data are
0 20 40 60 80 100 0 1
0 1 2 3 0 1 2 3
2 3
0 20 40 60 80 100
0 1 2 3 0 1 2 3
Fig. 8 The distribution of modules identified in the simulated fossil food web, within each module
of the modern food web. The statistically dominant occurrence of a single fossil module in each
modern module supports the equivalency of the webs’ modular structures. Each modern module is
plotted separately, as indicated at the top of each plot. Modules in the fossil food web are shown on
x-axes, numbered 0–3, to indicate significant statistical equivalence to their modern counterparts
not available, or more precise data are unsuitable for the analytical models being
employed. The extent to which aggregation obscures or distorts food web structure
and dynamics is an unsettled argument, but one solution would be to simply increase
the precision with which food web networks are constructed. A version of this
solution employs the trophospecies, or “trophic species” approach, in which species
aggregates comprise taxa with exactly the same trophic interactions (Martinez
1992). Trophospecies simplify a food web without sacrificing precision and is the
approach taken in the construction of the modern Caribbean coral reef community.
The trophospecies approach has also been applied to the fossil record in at least
three instances, including reconstructions of Cambrian marine food webs of the
Burgess Shale and Chengjiang faunas (Dunne et al. 2008), and an Eocene terrestrial
food web of the Messel Shale (Dunne et al. 2014). All these reconstructions take
advantage of fossil Lagerstätte, where fossil preservation is exquisite, probabilities
of preservation are high, and some traces of trophic interactions are available.
Furthermore, although taxon preservation is almost certainly incomplete, the Messel
Shale food web includes 700 taxa, and with the current coral reef food web
represents one of the largest food web reconstructions available. Thus, in spite of
incomplete preservation, fossil food webs are among the best food web reconstruc-
tions currently available. There are three drawbacks to the Lagerstätten approach
however, and those are that one is limited to the time, place, and community type of
the Lagerstätten. Additionally, Cambrian marine food webs are unlikely to offer
much insight into the dynamics of coral reef ecosystems under conditions of a
rapidly changing ocean, as the majority of Cambrian taxa are not extant.
Another solution to the problem of aggregation is to estimate precise interspecific
interactions even when data on interactions are limited. Simply put, this approach
models the precision required for a species-level food web by estimating the number
of interactions per species and the species which interact, while constraining those
estimates on the basis of known or inferred ecologies. This approach has been
developed most extensively in analyses of Paleozoic and Mesozoic terrestrial food
webs around the end-Permian and end-Cretaceous mass extinctions (Mitchell et al.
2012; Roopnarine et al. 2007; Roopnarine and Angielczyk 2015). Those analyses
used very well preserved biotas from the Karoo Basin of South Africa and western
North America, respectively, but their preservation falls below Lagerstätten quality.
Roopnarine and co-workers addressed this problem by assigning taxa to trophic
guilds, which are aggregations based on common body size, habitat, and potentially
overlapping prey and predators. Coral reef examples would include the placement
of all carcharhinid sharks into a guild, or all epibenthic sponges into a guild. Trophic
interactions between guilds lack the precision of interspecific interactions, but they
are in actuality sets of interspecific interactions, and each interspecific interaction
belongs to a single interguild interaction. Given a community of species, there is a
finite number of food webs that can be constructed, but many of them would not be
consistent with the ecologies and interactions of the community. Guild-level food
webs, also termed metanetworks, limit this number to a subset of species-level food
webs that are consistent with ecological reality. The dynamic terrestrial models of
Roopnarine et al. (2007) and Roopnarine and Angielczyk (2015) sample species-
level food webs from the metanetwork subset.
An unresolved question is how well guilds and metanetworks actually capture
the functional structure of a community. We address this here by first identifying
the guild structure of the reef, and then taking the perspective that one is required
to reconstruct the guild structure given only those taxa that would be preserved in
the simulated fossil reef. The reliability of the fossil metanetwork is then evaluated
by comparing its diversity and evenness to the modern metanetwork, and examining
whether the reef’s distribution of trophic levels can be reconstructed from the fossil
metanetwork.
Identifying Guilds in a Food Web
The purpose of a trophic guild is to aggregate species with overlapping prey and
predators. Trophospecies are therefore a type of guild, one in which members
share exactly the same prey and predators. Membership can be extended, while
maintaining exact overlap, to species with different numbers of prey and predators,
but where the interactions of some of the species are subsets of others, i.e., the
interactions of one species are nested within those of another. For example, prey
of the porcupine fish Chilomycterus antennatus is a subset of the trophically
more general triggerfish, Balistes vetula. A requirement of exact overlap, however,
becomes problematic under two conditions. First, increasingly precise data are liable
to uncover small differences between species, thereby separating species which
were formerly assigned to the same trophospecies. Second, if most interactions are
indeed relatively weak, then how much weight should be assigned to interactions
that are not shared between species, versus those that are shared? In both cases,
species can be ecologically similar, and yet their aggregation in a food web be
ambiguous. The questions could potentially be answered for modern communities
by ever-increasing empirical work although these are very difficult data to obtain.
Furthermore, the questions simply cannot be answered for paleocommunities. Here,
we propose a heuristic solution where we use the overlap of interactions among
species to recognize guilds, but limit the expected overlap according to the limits of
fossil preservation.
The procedure begins with a pairwise measure of interaction overlap between
all species in the food web. Overlap is measured separately for both types of
interactions, predator–prey and prey–predator, as
rij
oD (6)
min.ri ; rj /
where ri is the number of prey (predators) of species i, and rij is the number of prey
(predators) that i and j have in common. The value of o is thus zero if the species
have no prey in common, and 1 if they have all prey in common or one of the prey
sets is nested completely within the other. Predator overlap is measured similarly,
counting number or predators for ri and rij , such that o equals zero when the species
have no predators in common. Each set of measures yields a S S matrix of overlap
measures, where S is the species richness of the community. The two matrices are
then combined for each pair of species as the simple product of prey and predator
overlap, producing a single matrix of overlap indices, O. The resulting elements
of O then range from 0, where either prey or predator interactions, or both, fail to
overlap between the two species, to 1, where overlap or nestedness is complete.
Within the O matrix will be groups of species which overlap among themselves
more strongly than they do with other species, in effect forming modules. Guilds
can therefore be identified by examining the modularity of the O matrix and would
be equivalent to O’s modules.
Given the concerns expressed above regarding the precision of overlap, we
examined the matrix’s modularity, and hence guild composition, at multiple levels of
overlap by applying thresholds, where values below a threshold would be excluded
from guild recognition. We proceeded from a threshold value of 0.1, where species
sharing an overlap greater than 0.1 could potentially be assigned to the same guild,
to 1, where species within a guild would have perfectly nested sets of interactions.
200 0.8
150 0.6
No. of guilds
Modularity
100 0.4
50 0.2
0 0
0 0.2 0.4 0.6 0.8 1
Overlap index
Fig. 9 The number of guilds, or modules, recognized as the overlap or nestedness of interactions
between species. The extent of overlap, or threshold, is indicated on the x axis, and the
corresponding number of guilds (filled circles), and strength of the modularity (open squares),
are indicated on the left and y axes, respectively
Guild inclusivity thus decreases as the threshold increases, and the number of guilds
should therefore increase. These expectations are indeed met, as the number of
guilds in the coral reef community, and the strength of overlap as measured by
Newman’s modularity index, increase as the threshold increases, with modularity
being at a maximum when the threshold is 0.9 (Fig. 9). The number of guilds
increases from 10 at a threshold of 0.1 to 149 at a threshold of 1. The 728 species in
the Jamaican coral reef food web, represented by 249 trophospecies, may therefore
be nested within 149 guilds. This number is comparable to work by Bambach
(Bambach et al. 2007), in which 118 modes of life were found to exist for recent
marine metazoans. Given the greater resolution available in our data, 118 most likely
represents a lower limit of the number of guilds.
4 Reconstructing the Community
Imagine that our starting point is not a modern coral reef, but a fossil coral reef
instead. Can an understanding of how community information is transformed during
fossilization be combined with fossil taxon richness to reconstruct a paleocom-
munity food web? The easiest reconstruction is at the guild level, where species
are aggregated into sets of overlapping interactions. Not all the 149 guilds of the
modern reef would be preserved, however, and many interactions are also lost. The
losses cannot be accounted for in many cases, for example, the absence of a guild of
corallivorous polychaetes, because not only do those species lack a fossil record, but
it is doubtful that their existence could otherwise be inferred from preserved taxa.
There are two rich and highly connected groups, however, whose existence can be
inferred even though they are absent from the fossil record: epibenthic sponges and
zooplankton. The sponges do not have a taxonomic record, but their presence is
recorded as large numbers of disarticulated spicules. The sponges are a key link
between pelagic microorganisms on which they prey and guilds of spongivorous
macroinvertebrates and vertebrates. Major groups within the zooplankton guilds
are also absent from the fossil record, e.g, copepods, or have no records which
include coral reef-dwelling members, e.g., scyphozoans. There are large numbers
of preserved zooplanktivorous species and guilds, however, such as corals and
many other cnidarians and zooplanktivorous fish. We could therefore insert guilds
representing the missing groups into the set of preserved guilds with defensible
confidence and connect them to appropriate prey and predator guilds; but two
problems must be addressed.
First, whereas the epibenthic sponge trophospecies is assigned to a single guild
out of 149 in the modern food web, micro-zooplankton are distributed among six
different guilds, and macro-zooplankton are assigned to two guilds, one of which
also contains micro-zooplankton. Given the absence of zooplankton from our fossil
data, we would have no idea how many zooplankton guilds should be inserted
into the reconstructed community. The best and most conservative answer would
be to use a threshold of overlap at which a single zooplankton guild first emerges,
thereby aggregating all zooplankton into a single guild, yet distinguishing them from
other guilds. We therefore examined guild structure at each threshold as described
above (Fig. 9), and the zooplankton first emerge as a single guild at the 0.6 level
as thresholds increase from 0.1 to 1. There are 42 guilds at this threshold (see
Appendix 2, Table 1), and we consider it to be the best resolution that could be
inferred from our knowledge of fossil taxa, in the absence of modern data. Thirty
three of the 42 guilds include fossilizable taxa.
Second, how many species should be assigned to the sponge and zooplankton
guilds? Deriving sponge richness from spicule diversity is notoriously difficult and
unreliable (Jones 1984), and the zooplankton essentially have no fossil record in
the coral reef system. We addressed this problem by referring to a similar situation
in terrestrial paleocommunities with insect faunas. Mitchell et al. (2012), in their
reconstructions of Late Cretaceous North American communities, estimated insect
richness based on a positive relationship between the insect richness of well-
preserved faunas spanning the Phanerozoic, and the richness of vertebrate predation
on those faunas. The relationship is consistent over this time interval. The logic
of the relationship is that predator and prey richnesses may often be related in
communities over evolutionary time, since a greater diversity of prey could support
a greater diversity of predators, both energetically and by reducing competition
among the predators. Similarly, the presence of predators generally supports a
greater diversity of prey species. We therefore examined this relationship for a
zooplankton food chain at the 0.6 threshold. Guilds on this food chain include “phy-
toplankton,” “nanno-zooplankton,” “foraminifera,” and other heterotrophic protists,
“zooplankton,” “gorgonians,” and “pelagic planktivores.” The set of observations
Fig. 10 The relationship between predator guild richness and prey guild richness for guilds along
the major “zooplankton” food chain. The line indicates a fitted reduced major axis function. Prey
guilds are indicated at each corresponding data point
consists of the richness of a prey guild, and the sum of predators supported by that
guild, weighted by number of prey guilds linked to each predatory guild. That is,
predator richness
G
X aij jGj j
jpredatorsj D (7)
jD1
rj
where jGj j is richness of a guild of predators linked to prey guild i, and rj is

the number of guilds preyed upon by predatory guild j. The relationship was
fit using reduced major axis (RMA) regression both because neither variable
can be considered independent, and neither is measured without error (Fig. 10).
Furthermore, a RMA function is symmetrical, meaning that the function is invariant
to the choice of which variable is treated as the predictor. The RMA function
jpreyj D 0:275jpredatorsj C 10:362 (8)
is significant (r2 D 0:97, F D 270:468, p D 0:00005) and predicts a zooplankton

guild richness of 45 species. This is in excellent agreement with the actual richness
of zooplankton species (44).
A similar function could not be constructed successfully for the sponges because
the sponge guild-level food chain at 42 guild resolution includes two highly
aggregated guilds, the “hard benthic macroinvertebrates” and “large vertebrate

macrophyte and invertebrate grazers” which confound any more specific numerical
relationships between prey and predator richnesses. The overall fraction of species
preserved in the simulated paleocommunity is 0.59 (433 out of 728 species). Sponge
richness was therefore estimated using the fraction of overall preservation (0.59),
yielding (0.59 53) = 31 species, which is greater than the 16 species (all endolithic)
expected to be fossilized, but closer to the original richness of 53 species.
Diversity and Evenness
In order to explore the similarities and/or differences between the modern and fossil
coral reef food webs, we adopted methods commonly employed in paleoecology,
calculating the Simpson Index of Diversity (1-D) and the Shannon Index (H0 ) to
examine the richness and distribution of species across guilds (Shannon 1948;
Simpson 1949). The Simpson Index of Diversity (1-D) is a common metric for
quantifying taxonomic diversity and abundance and is used here to estimate the
taxonomic richness of guilds in the modern and fossil coral reef communities.
Hence, the higher the value of the Simpsons Index of Diversity, the greater the
taxonomic diversity of the guilds. The Shannon Index (H0 ) is used to calculate
the evenness (E D H0 =H0max ) of the distribution of species across guilds, with a
maximum value of 1.0 indicating a perfectly even distribution and a minimum value
of 0 indicating a highly uneven distribution.
In applying these metrics to the modern Jamaican coral reef community,
consisting of 728 species within 42 guilds, we found that the community had
very high taxonomic diversity, with a Simpson Index of 0.92. Meanwhile, the
fossilized coral reef community, consisting of 33 guilds and 433 species, had a
very similar taxonomic diversity, with a Simpson Index of 0.88. The modern coral
reef community was also fairly even in its distribution of species across guilds,
with an evenness (E) value of 0.75; the fossil reef community was comparable with
E D 0:72. In addition, as previously discussed, one can safely assume via fossil
evidence (e.g., sponge spicules) that a zooplankton and a sponge guild are likely
to be present. By replenishing those two guilds (34 guilds present, 493 species) and
recalculating the diversity metrics, we found that the community now had a Simpson
Index (1-D) of 0.90 and an evenness (E) value of 0.74. As such, it is apparent that
by restoring these unfossilized taxa to the fossil community, the taxonomic diversity
values become almost identical to that of the modern community.
However, while it is reassuring that these two diversity metrics are comparable
between the modern and fossilized communities, we have the foresight to be aware
that ten guilds that appear in the modern coral reef community are not preserved
in the fossilized community. Thus, it would appear that caution must be taken in
relying on these metrics to quantify guild structure; other metrics and analyses,
such as in-degree distribution, modularity, and trophic levels are needed in order
to fully characterise food web structure. Nonetheless, these metrics show that
both the modern and fossil coral reef communities had high taxonomic richness
though potentially lacking some redundancy within guilds. These results agree
with previous hypotheses indicating that coral reef ecosystems, despite containing
a high diversity of species and guilds, are extremely vulnerable to environmental
and anthropogenic perturbations given the potential for the loss of key ecosystem
processes and decreased resilience due to an uneven distribution of species across
guilds (Bellwood et al. 2003; Micheli and Halpern 2005; Guillemot et al. 2011;
Brandl and Bellwood 2014; Nash et al. 2016). Low redundancy within guilds, as
implied here, would further exacerbate this vulnerability.
Simulated Food Webs
A primary goal of paleo-food web reconstruction has been the modelling of

paleocommunity and ecosystem dynamics (Roopnarine et al. 2007; Mitchell et al.
2012; Roopnarine and Angielczyk 2015; Angielczyk et al. 2005; Roopnarine 2006,
2009). Those models require the simulation of food webs that are consistent with
paleocommunity structure, as described by resolvable guild-level networks. The
reconstruction of the coral reef at the guild level presented here shows that guild
structure is preserved during the simulated fossilization, within limits set by the
absence of data on at least two significant groups, the zooplankton and epibenthic
sponges. Furthermore, the data can be improved with reliable estimates of species
for those missing guilds, with analyses of diversity and evenness supporting the
improvement. A concluding step in this exercise is then the derivation of a species-
level food web from the preserved guild structure. We did this by assigning
in-degrees (number of prey) to each consumer species with random draws from
the in-degree equation derived above (section “Dietary Breadth”), constrained to the
total richness of the guilds upon which the species could potentially prey. The actual
prey species are then assigned randomly to the consumer (see Roopnarine (2009) for
further details). The resulting simulated food web consists of 398 connected species,
with 965 interactions. The latter number is far below the 4105 inter-trophospecific
interactions of the modern community, but recall that the simulated food web
is based on an aggregation into 44 guilds. A second food web was simulated,
but in this case using a mixed exponential-power law in-degree distribution used
commonly in previous paleo-food web dynamics studies (Roopnarine et al. 2007;
Mitchell et al. 2012; Roopnarine and Angielczyk 2015; Angielczyk et al. 2005;
Roopnarine 2006). This distribution takes the form P.r/ D er=" , where r is the
in-degree of the consumer, " D e.1/ ln.M/=./ , M is the total number of prey
potentially available to a consumer, and , the power law exponent, is 2.5. The
value of the exponent is the mid-point of a range previously explored for this
distribution when applied to Permian-Triassic paleocommunities from the Karoo
Basin of South Africa (Roopnarine et al. 2007). Distributions within the range were
found to construct species-level food webs with linkage densities and connectances
comparable to those reported for modern food webs (Dunne et al. 2002; Roopnarine
2009). Applying the distribution to the system of fossil coral reef guilds here yields
a web with a greater density of interactions, with 8924 interspecific interactions.
It is conceivable that, despite the disparities of interactions among the webs,

they could nevertheless yield similar dynamical properties. Such similarity would
be possible, however, only if the patterns of interaction are similar among the webs
(Roopnarine and Angielczyk 2015). This was checked for the species-level fossil
and modern webs by comparing their network trophic level (ntl) distributions. Recall
that those distributions differed significantly between the modern and simulated
fossil webs (Fig. 5). The question addressed here is whether the reconstructions,
based upon the reconstruction of guild structure, and restoration of unfossilized
zooplankton and epibenthic sponges mitigate any of that loss of structure. The
ntl distribution of the simulated web based on the observed in-degree power law
distribution differs significantly from the modern web (ANOVA; df = 2, 478; F D
30:25, p < 0:0001), and a majority of taxa are below ntl 3.0 (Fig. 11). The ntl
distribution of the web based on the mixed exponential-power law distribution,
however, is statistically indistinguishable from the modern web (Scheffe’s multiple
comparison test, p D 1:0) and reconstructs both the mean (3.29 in both cases)
a
0 .2 .4 .6 .8
b
5
Density
0
c
1.5
1
.5
0
2 3 4 5 6
Graphs by tag ntl
Fig. 11 Observed and reconstructed network trophic level (ntl) distributions for all secondary
and greater consumers. (a) Observed distribution measured from the modern food web. (b) The
ntl distribution based on a species-level food web reconstructed from the fossil guild network,
and using the power law in-degree (dietary breadth) distribution of the modern food web. (c) A
similar reconstruction, but made using a mixed exponential-power law in-degree function. The
power law-based reconstruction in b fails to recreate the trophic level distribution, yielding a highly
truncated and short-chained food web. The mixed distribution produces a food web with trophic
level distribution statistically indistinguishable from the modern food web
and maximum values (5.5 and 6, respectively). The mixed exponential-power law
distribution predicts denser interactions than does the power law, in effect being
“fatter-tailed.” This feature compensates for the loss of interactions due to the loss
of taxa and obviously reconstructs the hierarchical structure of the modern food
web with significant fidelity. We predict that dynamic analyses of fossil coral reef
communities based on this type of reconstruction would be comparable to analyses
of modern communities.
5 Summary
The ongoing, unprecedented, global, anthropogenically driven degradation of nat-

ural systems has led to a need for conservation strategies that account for whole
ecosystem structures. Coral reefs have received much attention in regard to the
effects of anthropogenic and non-anthropogenic stressors on ecosystem functioning,
as recent surveys have indicated a global decline in reef diversity, coral cover, and
overall functioning (Carpenter et al. 2008; Veron et al. 2009). Caribbean reefs,
in particular, have suffered considerable historical damage and overexploitation,
potentially since initial colonization of islands and coastal areas by humans (Hardt
2009). Studies of modern coral reef functioning and diversity cannot capture
the shifted baselines of species composition, population sizes, and community
structure although these are likely required to forecast what we can expect in a
future of continuing environmental decline. Deep-time studies have the potential to
compliment modern studies by providing insight into how prior coral reef systems
have responded to similar environmental perturbations, or perturbations of similar
magnitude, in Earth’s past (Greenstein and Pandolfi 1997).
Here, we tested the plausibility for such studies by simulating the fossilization
of a modern Jamaican coral reef food web and determining how realistically a
food web could be recreated from the fossil record (Roopnarine and Hertog 2013).
The initial modern Jamaican food web consisted of 728 species, which were then
collapsed into 249 trophospecies (i.e., species that share exactly the same prey
and predators) with a total of 4105 inter-trophospecific predator–prey interactions.
Simulated fossilization, or removal of genera without a documented fossil record,
resulted in a community comprising 433 species, 172 trophospecies, and 1737 inter-
trophospecific interactions. As expected, the loss of trophospecies and interspecific
interactions resulted in a significant bias, both against poorly preserved species, and
in favor of those species that had better than average fossil records. Particularly sig-
nificant was the almost complete loss of zooplankton trophospecies, which resulted
in their predators being either poorly connected or completely disconnected from the
fossil food web. While the absence of zooplankton in the simulated fossil data places
a negative bias against zooplanktivorous species in terms of biotic interactions, the
importance of taxa whose prey have exceptionally high probabilities of preservation,
e.g. durophagous crustaceans and fish, may be overestimated.
A positive consequence of the latter bias, coupled with the preservation of more
than half of the species in the food web, is that several important features are retained
by the simulated fossil community, namely the distribution of dietary breadths
among consumers, and the modularity or compartmentalization of the community.
The distribution of dietary breadths, or in-degree distribution, describes the number
of resource species per consumer species. Compilations based on a variety of
modern food webs suggest that those distributions are generally of a decay type
and hyperbolic. This means that more species are specialists, consuming relatively
few species, and fewer species are generalists; yet the hyperbolic nature of the
distributions also implies that generalist species occur at frequencies greater than
would be expected were dietary breadths distributed normally. The coral reef meets
these expectations, being fit significantly with a power function. The distribution
for the reduced, simulated fossilized web is statistically indistinguishable from the
modern web’s distribution. Nevertheless, the loss of interactions does result in a
significant alteration of the trophic level structure of the community. Many species
have lower trophic levels in the fossil community, particularly those involved in
very long food chains that include taxa with low probabilities of preservation. The
result is that the distribution of trophic levels in the simulated fossil community is
significantly truncated, and there is no possibility of recreating trophic levels based
on taxon composition alone.
Modularity has been proposed to be important to the stability or resistance of
food webs to perturbations, yet the search for modules in food webs has yielded
equivocal results. Here, we demonstrated that the Jamaican coral reef food web
is indeed modular. Modularity is most distinct toward the bases of food chains,
and the four modules identified in the community are based primarily on the
differential utilization of basal and low trophic level resources. High trophic level
predators, however, such as the Caribbean reef shark, feed across modules, both
uniting modules via top-down effects, and reducing the modularity of the system.
Furthermore, many of those predatory species are involved in very strong biotic
interactions (Bascompte et al. 2005). The apparent modularity of the modern
community might therefore be an anthropogenically amplified effect, because many
high trophic level predators have been either extirpated from reef communities
throughout the Caribbean, or are now present in very low numbers. The fossilized
food web retains the modularity of the modern web, perhaps because it too
lacks many higher trophic level species, though for preservational reasons. This
emphasizes the caution which should be exercised in the analysis of modern, post-
disturbance communities, and it is useful to speculate that assessments of reef
community modularity based on historical and sub-fossil records (Hardt 2009) could
yield significantly different results. It is also worth considering, in this context,
whether the introduction of the invasive lionfish, Pterois volitans, into the Caribbean
has restored some of the effects of predation now lacking in those communities.
We tested our ability to recreate a coral reef food web from fossil data only,
by assuming the simulated fossil reef as a starting point. Following procedures
developed for terrestrial paleocommunities (Roopnarine et al. 2007), trophic guild
structure was used as the basis to aggregate species into biotically interacting units.
We used a heuristic approach to first identify trophic guilds in the modern food
web. If we limit our ability to recognize guilds according to the limits of fossil
preservation, we find a total of 42 guilds, whereas the maximum number identified
in the modern community is 149. Nevertheless, Simpson (1-D) and Shannon (H0 )
Indexes for modern and fossil guild diversity were almost identical despite a loss of
guilds in the fossil community.
Finally, species-level food webs were reconstructed from the fossil and guild data
using stochastic techniques described in (Roopnarine 2009). One reconstruction,
based on the in-degree distribution of the modern web, fails to recreate the trophic
level structure, and hence hierarchical arrangement of the modern community.
In contrast, a second reconstruction, based on a model in-degree distribution
developed in previous studies (Roopnarine et al. 2007), creates a trophic level and
hierarchical structure statistically indistinguishable from the modern community.
The greater success of the latter in-degree distribution is a consequence of its
overestimation of species dietary breadths, thereby compensating for losses incurred
during fossilization. The important implication of this successful recreation is that
coral reef communities can indeed be recreated from fossil data and subjected
to the types of dynamic analyses essential for forecasting the behavior of those
communities under broad magnitudes of environmental and biotic disturbance.
This emphasizes the validity and usefulness of paleoecological data to the field of
conservation biology. The creation of deep-time paleocommunity food webs has the
ability to enrich and advance our current knowledge of how natural systems behave,
especially in response to future environmental changes.
Acknowledgements We wish to thank Courtney Chin, Chrissy Garcia, and Rhiannon Roopnarine
for the food chain illustrations and other graphical assistance. We also thank the CAS Paleoecology
Discussion Group, including Allen Weik, for many useful discussions which contributed to this
chapter. Finally, thanks to Kenneth Angielczyk and Carol Tang for asking difficult questions.
Appendix 1
Hypergeometric Variance
The hypergeometric distribution describes the probability given a population with

a featured subset of observations, and a sample from the population, that a random
selection of individuals from the population will include a certain fraction of the
sample comprising individuals with the featured characteristic. In this case, the
population consists of 728 species (N), of which 433 are featured as fossilized (K).
A sample is the richness of a trophospecies (n), of which a subset are observed as
fossilized (k). The expected fossilization or hypergeometric mean value is
O Dn K
E.k/ (9)
N
Fig. 12 Hypergeometric variance given a total population of 728 species, and trophospecies of
increasing size
The variance of the expectation is given as
K NK Nn
2 .k/ D n (10)
N N N1
This value grows initially as n because the number of unique ways in which k objects
may be selected also grows. The value declines, however, as n ! N (Fig. 12).
Appendix 2
See Table 1.
Table 1 List of guilds recovered at the 0.6 threshold of interaction overlap

Guilds Modern Fossil
Bacteria – –
Phytoplankton 29 15
Nanno-zooplankton 18 3
Macrophytes, diatoms 28 14
Sponges 53 31
Corals 70 64
Micro-detritivores 19 18
Corallivorous polychaetes 2 0
Zooplankton 44 45
Gorgonians 15 0
Benthic carnivores 22 0
Eucidarid echinoids 1 1
Hard benthic macroinvertebrates 123 90
Cypraeids 3 3
Sinum 1 1
Charonia 1 1
Epiphyte-grazing gastropods 3 3
Bryozoans 25 20
Endolithic polychaetes 3 3
Soft benthic macroinvertebrates 80 75
Polychaete-consuming gastropods 6 6
Molluscivorous crustaceans 1 1
Macroinvertebrate predators I 60 25
Stomatopods I 2 2
Large grazers of macroalgae and invertebrates 79 47
Macroinvertebrate predators II 4 4
Octopus 1 0
Macroinvertebrate predators III 11 2
Stomatopods II 3 3
Hemiramphus brasiliensis 1 0
Trachinotus goodei 1 0
Pelagic piscivores I 8 8
Mycteroperca bonaci 1 1
Gymnothorax moringa 1 0
Lutjanus analis 1 0
Scomberomorus cavalla 1 1
Green sea turtle 1 1
Hawksbill turtle 1 1
Loggerhead turtle 1 1
Blacktip shark 1 1
Oceanic whitetip shark 1 1
Caribbean reef shark 1 1
Modern and fossil guild richnesses are shown. Guilds containing single taxa have
those taxa listed by taxonomic or unique common name. Fossil guild richnesses with
an asterix were estimated
References
A.C. of Excellence Coral Reef Studies (2016) Coral Bleaching Taskforce documents most severe
bleaching on record. http://www.coralcoe.org.au/media-releases/coral-bleaching-taskforce-
documents-most-severe-bleaching-on-record
Albright R, Caldeira L, Hosfelt J, Kwiatkowski L, Maclaren JK, Mason BM, Nebuchina Y,
Ninokawa A, Pongratz J, Ricke KL et al (2016) Reversal of ocean acidification enhances net
coral reef calcification. Nature 531(7594):362
Andres N, Witman J (1995) Trends in community structure on a Jamaican reef. Oceanogr Lit Rev
10(42):889
Angielczyk KD, Roopnarine PD, Wang SC (2005) Modeling the role of primary productivity
disruption in end-permian extinctions, Karoo Basin, South Africa. The nonmarine Permian.
N M Mus Nat Hist Sci Bull 30:16–23
Aronson RB, Thatje S, Clarke A, Peck LS, Blake DB, Wilga CD, Seibel BA (2007) Climate change
and invasibility of the Antarctic benthos. Annu Rev Ecol Evol Syst 38:129–154
Bambach RK, Bush AM, Erwin DH (2007) Autecology and the filling of ecospace: key metazoan
radiations. Palaeontology 50(1):1
Bascompte J, Melián CJ, Sala E (2005) Interaction strength combinations and the overfishing of a
marine food web. Proc Natl Acad Sci U S A 102(15):5443
Bastian M, Heymann S, Jacomy M et al (2009) Gephi: an open source software for exploring and
manipulating networks. ICWSM 8:361
Barnosky AD, Hadly EA, Bascompte J, Berlow EL, Brown JH, Fortelius M, Getz WM, Harte J,
Hastings A, Marquet PA et al (2012) Approaching a state shift in Earth’s biosphere. Nature
486(7401):52
Becker BH, Beissinger SR (2006) Centennial decline in the trophic level of an endangered seabird
after fisheries decline. Conserv Biol 20(2):470
Bellwood DR, Hoey AS, Choat JH (2003) Limited functional redundancy in high diversity systems:
resilience and ecosystem function on coral reefs. Ecol Lett 6(4):281
Bellwood DR, Hughes TP, Folke C, Nyström M (2004) Confronting the coral reef crisis. Nature
429(6994):827
Blondel V, Guillaume J, Lambiotte R, Lefebvre E (2008) Fast unfolding of communities in large
networks. J Stat Mech Theor Exp 2008:10008
Brandl SJ, Bellwood DR (2014) Individual-based analyses reveal limited functional overlap in a
coral reef fish community. J Anim Ecol 83(3):661
Cardinale BJ, Duffy JE, Gonzalez A, Hooper DU, Perrings C, Venail P, Narwani A, Mace GM,
Tilman D, Wardle DA et al (2012) Biodiversity loss and its impact on humanity. Nature
486(7401):59
Carpenter KE, Abrar M, Aeby G, Aronson RB, Banks S, Bruckner A, Chiriboga A, Cortés J,
Delbeek JC, DeVantier L et al (2008) One-third of reef-building corals face elevated extinction
risk from climate change and local impacts. Science 321(5888):560
Cramer KL, Leonard-Pingel JS, Rodríguez F, Jackson JB (2015) Molluscan subfossil assemblages
reveal the long-term deterioration of coral reef environments in Caribbean Panama. Mar Pollut
Bull 96(1):176
De’ath G, Fabricius KE, Sweatman H, Puotinen M (2012) The 27–year decline of coral cover on
the Great Barrier Reef and its causes. Proc Natl Acad Sci U S A 109(44):17995
Dunne JA, Williams RJ, Martinez ND (2002) Food-web structure and network theory: the role of
connectance and size. Proc Natl Acad Sci U S A 99(20):12917
Dunne JA, Williams RJ, Martinez ND (2002) Network structure and biodiversity loss in food webs:
robustness increases with connectance. Ecol Lett 5(4):558
Dunne JA, Brose U, Williams RJ, Martinez ND (2005) Modeling food web dynamics: complexity-
stability implications. In: Aquatic food webs: an ecosystem approach. Oxford University Press,
Oxford, pp 117–129
Dunne JA, Williams RJ, Martinez ND, Wood RA, Erwin DH (2008) Compilation and network
analyses of Cambrian food webs. PLoS Biol 6(4):e102
Dunne JA, Labandeira CC, Williams RJ (2014) Highly resolved early Eocene food webs show
development of modern trophic structure after the end-Cretaceous extinction. Proc R Soc Lond
B: Biol Sci 281(1782):20133280
FishBase (2016) (World Wide Web electronic publication). www.fishbase.org
Franklin JF (1993) Preserving biodiversity: species, ecosystems, or landscapes? Ecol Appl
3(2):202
Gardner TA, Côté IM, Gill JA, Grant A, Watkinson AR (2003) Long-term region-wide declines in
Caribbean corals. Science 301(5635):958
Greenstein BJ, Pandolfi JM (1997) Preservation of community structure in modern reef coral life
and death assemblages of the Florida Keys: implications for the Quaternary fossil record of
coral reefs. Bull Mar Sci 61(2):431
Guillemot N, Kulbicki M, Chabanet P, Vigliola L (2011) Functional redundancy patterns reveal
non-random assembly rules in a species-rich marine assemblage. PloS One 6(10):e26735
Guimerá R, Amaral LAN (2005) Functional cartography of complex metabolic networks. Nature
433(7028):895
Guimerá R, Amaral LAN (2005) Cartography of complex networks: modules and universal roles.
J Stat Mech Theor Exp 2005(02):P02001
Guimerà R, Sales-Pardo M, Amaral LAN (2007) Module identification in bipartite and directed
networks. Phys Rev E 76(3):036102
Hardt MJ (2009) Lessons from the past: the collapse of Jamaican coral reefs. Fish Fish 10(2):143
Harnik PG, Lotze HK, Anderson SC, Finkel ZV, Finnegan S, Lindberg DR, Liow LH, Lockwood
R, McClain CR, McGuire JL et al (2012) Extinctions in ancient and modern seas. Trends Ecol
Evol 27(11):608
Hewitt JE, Ellis JI, Thrush SF (2016) Multiple stressors, nonlinear effects and the implications of
climate change impacts on marine coastal ecosystems. Glob Chang Biol 22(8):2665–2675
Hoegh-Guldberg O, Mumby PJ, Hooten AJ, Steneck RS, Greenfield P, Gomez E, Harvell CD, Sale
PF, Edwards AJ, Caldeira K et al (2007) Coral reefs under rapid climate change and ocean
acidification. Science 318(5857):1737
Holling CS (1973) Resilience and stability of ecological systems. Annu Rev Ecol Syst 4:1–23
Hughes TP (1994) Catastrophes, phase shifts, and large-scale degradation of a Caribbean coral
reef. Science 265(5178):1547
Idjadi J, Haring R, Precht W (2010) Recovery of the sea urchin Diadema antillarum promotes
scleractinian coral growth and survivorship on shallow Jamaican reefs. Mar Ecol Prog Ser
403:91
Jackson JB, Kirby MX, Berger WH, Bjorndal KA, Botsford LW, Bourque BJ, Bradbury RH, Cooke
R, Erlandson J, Estes JA et al (2001) Historical overfishing and the recent collapse of coastal
ecosystems. Science 293(5530):629
Joachimski MM, Lai X, Shen S, Jiang H, Luo G, Chen B, Chen J, Sun Y (2012) Climate warming
in the latest Permian and the Permian–Triassic mass extinction. Geology 40(3):195
Jones WC (1984) Spicule dimensions as taxonomic criteria in the identification of haplosclerid
sponges from the shores of Anglesey. Zool J Linnean Soc 80(2–3):239
Kidwell SM (2001) Preservation of species abundance in marine death assemblages. Science
294(5544):1091
Kidwell SM, Bosence DWJ (1991) Taphonomy: releasing the data locked in the fossil record. In:
Taphonomy and time-averaging of marine shelly faunas. Plenum, New York, pp 115–209
Kleypas JA, Buddemeier RW, Archer D, Gattuso JP, Langdon C, Opdyke BN (1999) Geochemical
consequences of increased atmospheric carbon dioxide on coral reefs. Science 284(5411):118
Knowlton N (2004) Multiple “stable” states and the conservation of marine ecosystems. Prog.
Oceanogr. 60(2):387
Knowlton N, Jackson JB (2008) Shifting baselines, local impacts, and global change on coral reefs.
PLoS Biol 6(2):e54
Kondoh M (2008) Building trophic modules into a persistent food web. Proc Natl Acad Sci U S A
105(43):16631
Leighton LR (2002) Inferring predation intensity in the marine fossil record. Paleobiology
28(03):328
Lessios H (2016) The Great Diadema antillarum Die-Off: 30 Years Later. Annu Rev Mar Sci 8:267
Lessios HA, Robertson D, Cubit J (1984) Spread of Diadema mass mortality through the
Caribbean. Science 226(4672):335
Lotze HK, Lenihan HS, Bourque BJ, Bradbury RH, Cooke RG, Kay MC, Kidwell SM, Kirby MX,
Peterson CH, Jackson JB (2006) Depletion, degradation, and recovery potential of estuaries
and coastal seas. Science 312(5781):1806
Mace GM, Norris K, Fitter AH (2012) Biodiversity and ecosystem services: a multilayered
relationship. Trends Ecol Evol 27(1):19
Martinez ND (1992) Constant connectance in community food webs. Am Nat 139:1208–1218
May RM (1973) Stability and complexity in model ecosystems, vol 6. Princeton University Press,
Princeton
May RM (1983) Ecology: the structure of food webs. Nature 301:566
McCann KS (2000) The diversity–stability debate. Nature 405(6783):228
McCann K, Hastings A, Huxel GR (1998) Weak trophic interactions and the balance of nature.
Nature 395(6704):794
McClanahan TR (2002) The near future of coral reefs. Environ Conserv 29(04):460
Meffe G, Nielsen L, Knight RL, Schenborn D (2012) Ecosystem management: adaptive,
community-based conservation. Island Press, Washington
Micheli F, Halpern BS (2005) Low functional redundancy in coastal marine assemblages. Ecol
Lett 8(4):391
Mitchell JS, Roopnarine PD, Angielczyk KD (2012) Late Cretaceous restructuring of terrestrial
communities facilitated the end-Cretaceous mass extinction in North America. Proc Natl Acad
Sci U S A 109(46):18857
Mora C (2008) A clear human footprint in the coral reefs of the Caribbean. Proc R Soc Lond B:
Biol Sci 275(1636):767
Mumby PJ, Hedley JD, Zychaluk K, Harborne AR, Blackwell PG (2006) Revisiting the catas-
trophic die-off of the urchin Diadema antillarum on Caribbean coral reefs: fresh insights on
resilience from a simulation model. Ecol Model 196(1):131
Mumby PJ, Dahlgren CP, Harborne AR, Kappel CV, Micheli F, Brumbaugh DR, Holmes KE,
Mendes JM, Broad K, Sanchirico JN et al (2006) Fishing, trophic cascades, and the process of
grazing on coral reefs. Science 311(5757):98
Mumby PJ, Hastings A, Edwards HJ (2007) Thresholds and the resilience of Caribbean coral reefs.
Nature 450(7166):98
Nash KL, Graham NA, Jennings S, Wilson SK, Bellwood DR (2016) Herbivore cross-scale
redundancy supports response diversity and promotes coral reef resilience. J Appl Ecol 53:646–
655
Newman ME (2004) Fast algorithm for detecting community structure in networks. Phys Rev E
69(6):066133
Newman ME (2006) Modularity and community structure in networks. Proc Natl Acad Sci U S A
103(23):8577
Nyström M, Folke C, Moberg F (2000) Coral reef disturbance and resilience in a human-dominated
environment. Trends Ecol Evol 15(10):413
Pace ML, Cole JJ, Carpenter SR, Kitchell JF (1999) Trophic cascades revealed in diverse
ecosystems. Trends Ecol Evol 14(12):483
Paine RT (1980) Food webs: linkage, interaction strength and community infrastructure. J Anim
Ecol 49(3):667
Pandolfi JM, Jackson JB (2006) Ecological persistence interrupted in Caribbean coral reefs. Ecol
Lett 9(7):818
Pandolfi JM, Jackson JBC, Baron N, Bradbury R et al (2005) Are US coral reefs on the slippery
slope to slime? Science 307(5716):1725
Parmesan C (2006) Ecological and evolutionary responses to recent climate change. Annu Rev
Ecol Evol Syst 37:637–669
Pimm SL (1982) Food webs. Springer, New York
Pimm SL, Kitching R (1987) The determinants of food chain lengths. Oikos 50:302–307
Pimm SL, Lawton JH, Cohen JE (1991) Food web patterns and their consequences. Nature
350(6320):669
Post DM (2002) The long and short of food-chain length. Trends Ecol Evol 17(6):269
Reef Environmental Education Foundation Volunteer Survey Project Database (2011) (World Wide
Web electronic publication). www.REEF.org
Romanuk TN, Hayward A, Hutchings JA (2011) Trophic level scales positively with body size in
fishes. Glob Ecol Biogeogr 20(2):231
Roopnarine PD (2006) Extinction cascades and catastrophe in ancient food webs. Paleobiology
32(01):1
Roopnarine PD (2009) Conservation paleobiology. Paleontological society papers, vol 15,
GP. Dietl, KW. Flessa (eds). The Paleontological Society, Tokyo pp 195–220
Roopnarine PD (2009) Ecological modeling of paleocommunity food webs. Paleontol Soc Pap
15:196
Roopnarine PD (2014) Humans are apex predators. Proc Natl Acad Sci U S A 111(9):E796
Roopnarine PD, Angielczyk KD (2015) Community stability and selective extinction during the
Permian-Triassic mass extinction. Science 350(6256):90
Roopnarine P, Hertog R (2012) Data from: detailed food web networks of three greater antillean
coral reef systems: the Cayman Islands, Cuba, and Jamaica. doi:10.5061/dryad.c213h. http://
dx.doi.org/10.5061/dryad.c213h
Roopnarine PD, Hertog R (2012) Detailed food web networks of three greater Antillean coral reef
systems: the Cayman Islands, Cuba, and Jamaica. Dataset Papers in Science 2013
Roopnarine PD, Angielczyk KD, Wang SC, Hertog R (2007) Trophic network models
explain instability of early triassic terrestrial communities. Proc R Soc Lond B: Biol Sci
274(1622):2077
Sabine CL, Feely RA, Gruber N, Key RM, Lee K, Bullister JL, Wanninkhof R, Wong C, Wallace
DW, Tilbrook B et al (2004) The oceanic sink for anthropogenic CO2. Science 305(5682):367
Schwartz MW, Brigham C, Hoeksema J, Lyons K, Mills M, Van Mantgem P (2000) Linking bio-
diversity to ecosystem function: implications for conservation ecology. Oecologia 122(3):297
Scheffer M, Carpenter S, Foley JA, Folke C, Walker B (2001) Catastrophic shifts in ecosystems.
Nature 413(6856):591
Seddon AW, Mackay AW, Baker AG, Birks HJB, Breman E, Buck CE, Ellis EC, Froyd CA, Gill JL,
Gillson L et al (2014) Looking forward through the past: identification of 50 priority research
questions in palaeoecology. J Ecol 102(1):256
Sepkoski JJ Jr (2002) A compendium of fossil marine animal genera. Bull Am Paleontol
363(16):1–560
Sergio F, Newton I, Marchesi L (2005) Conservation: top predators and biodiversity. Nature
436(7048):192
Shannon CE (1948) A mathematical theory of communication. Bell Syst Tech J 27:379
Simpson EH (1949) Measurement of diversity. Nature 183:688
Vermeij GJ, Leighton LR (2003) Does global diversity mean anything? Paleobiology 29(01):3
Villeger S, Novack-Gottshall PM, Mouillot D (2011) The multidimensionality of the niche reveals
functional diversity changes in benthic marine biotas across geological time. Ecol Lett 14(6),
561
Veron J, Hoegh-Guldberg O, Lenton T, Lough J, Obura D, Pearce-Kelly P, Sheppard C, Spalding
M, Stafford-Smith M, Rogers A (2009) The coral reef crisis: The critical importance of
<350 ppm CO2. Mar Pollut Bull 58(10):1428
Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee TJ, Fromentin JM, Hoegh-Guldberg
O, Bairlein F (2002) Ecological responses to recent climate change. Nature 416(6879):389
Williams RJ (2010) Simple MaxEnt models explain food web degree distributions. Theor Ecol
3(1):45
Williams RJ, Martinez ND (2000) Simple rules yield complex food webs. Nature 404(6774):180
Williams RJ, Martinez ND (2004) Limits to trophic levels and omnivory in complex food webs:
theory and data. Am Nat 163(3):458
Wood R (1998) The ecological evolution of reefs. Annu Rev Ecol Syst 29:179–206
Yodzis P (1984) Energy flow and the vertical structure of real ecosystems. Oecologia 65(1):86
Yodzis P (1989) Introduction to theoretical ecology. Harper & Row, New York
Exploring the Species–Area Relationship
Within a Paleontological Context,
and the Implications for Modern
Conservation Biology
Matthew J. Pruden and Lindsey R. Leighton
Abstract Human-driven habitat loss and fragmentation is one of the largest threats
to modern biodiversity. The relationship between species and area is one of the
oldest patterns observed in ecology; models based on the species–area relationship
(SAR) have been used to estimate both local and global extinction rates. A critical
difficulty for modern conservation studies is that they are forced to predict the
future; empirical testing of these predictions puts at risk the very species the
predictions are designed to protect. To help mitigate this problem, we explore
the use of the species–area relationship using the paleontological record, which
provides empirical data from both before and after species loss. Using species
presence/absence data, and a contiguous minimal area grid system, species–area
curves (SACs) were constructed for one biofacies in each phase of a transgressive–
regressive package. A regression observed for the Pliocene Etchegoin Formation of
the San Joaquin Basin (California) inland sea, and the effect of sea-level fall on the
perched molluscan fauna, was used as an analogy for modern habitat loss. Using
the equation of the species–area curve of the transgressive interval, we predicted the
number of species to go extinct in the subsequent regression. To examine extinction
severity, we compared these predictions to the data observed for the regressive
curve. The results showed that the regressive curve had fewer species than the
prediction based on the transgressive curve, suggesting that species loss was more
severe than predicted. The results demonstrate an example of how the species–area
relationship can be used in the fossil record, and that the paleontological record
could be used to provide relevant information for modern conservation problems.
Keywords California · Kettleman Hills · Pliocene · Sea level · Habitat loss ·

Extinction
M. J. Pruden () · L. R. Leighton

Canada

144 M. J. Pruden and L. R. Leighton
1 Introduction
The Earth is on the brink of a sixth mass extinction, whereby species extinction
rates are estimated to be 1000–10,000 times that of the background rate (Pimm et al.
1995; Millennium Ecosystem Assessment 2005), and the leading cause is human-
driven habitat loss (Leakey and Lewin 1996; Millennium Ecosystem Assessment
2005; Barnosky et al. 2011). Preventing the loss of species due to human habitat
destruction is one of the top priorities for conservation biologists. Ecologists have
been building models to quantify and predict the extent of species loss for decades
(MacArthur and Wilson 1967; Levins 1969; Richter-Dyn and Goel 1972; Strebel
1985). Ecologists and conservation biologists have frequently used species–area
relationships (SARs) to predict the number of species at risk due to current habitat
destruction (National Research Council 1980; Ehrlich and Ehrlich 1981; WCMC
1992). The SAR is a positive logarithmic relationship between species and area
(Rosenzweig 1995). For example, Wilson (1989) used deforestation rates, in terms
of area lost, of the world’s rain forests to predict that 0.2–0.3% of the global
species would go extinct per year, or approximately 125,000 species over a 10-
year period. Thomas et al. (2004) used SARs to predict that 15–35% of species
will be extinct by 2050. A more recent study predicted that 35% of tropical African
species, 20% of tropical Asian species, and 15% of tropical American species are
set to be extinct by 2050 (Millennium Ecosystem Assessment 2005). However, most
modern ecological studies attempting to predict extinction are necessarily limited to
models, which can only be tested empirically by the passage of time—at which
point, vulnerable species may already have been lost. The goal of this study is to
review the use of SARs by studying them in a paleontological context and to explore
the use of the paleontological record in studying modern conservation models. This
study constructs species–area curves (SACs) for marine molluscs from the Pliocene
Kettleman Hills in California; and then uses the curves to compare predicted and
observed extinctions due to habitat loss, in order to assess the relationship between
species–area curves and extinction severity.
The relationship between species and area is one of the oldest ecological patterns,
first described more than 100 years ago by Watson in 1859, or possibly even
earlier in 1835 (Dony 1963; Williams 1964). The basic idea of the species–area
relationship is that as area increases, so does the number of species (Rosenzweig
1995). Among proposed mechanisms for the SAR, Williams (1943) introduced
the habitat diversity hypothesis; larger areas contain a greater array of habitats
and therefore can support more species. The disturbance hypothesis claims that
populations on smaller “islands” are more susceptible to disturbances because
smaller areas are likely to have fewer refuges from the disturbances, suggesting
that as area decreases, so should the number of species due to the increase in local
extinction (Williams 1943). These explanations for the species–area relationship
are similar to the well-known island biogeography model (MacArthur and Wilson
1967), island being a loose term used to describe geographically isolated areas that
Exploring the Species–Area Relationship Within a Paleontological Context: : : 145
still allow for immigration and emigration, such as actual islands, lakes, mountain
tops, forest fragments, and inland seas.
There are two fundamental (and mathematically equivalent) equations of SAR.
The power-function (1) S D cAz is the equation of the logarithmic curve within
arithmetic space, whereby S is the number of species, A is the area, c describes the
slope of the graph, and z is a constant. More commonly, researchers compare the
species–area relationship in log space; the equivalent equation is (2) log S D z log
A C log c, which is the equation of the straight line within a log-log space. In eq.
(2), z describes the slope of the log-log relationship, and log c is the y-intercept
(Rosenzweig 1995). Researchers have been trying to find the canonical z values,
particular values with consistent biological meaning, for decades. Preston (1962)
considered values between 0.17 and 0.33, while MacArthur and Wilson (1967)
accepted values between 0.20 and 0.35, to fall within the canonical range. Connor
and McCoy (1979) examined 100 SACs and found 45% of the z values fell between
0.20 and 0.40. There may be a relationship between island (or region) size and z
values. Smaller islands have z values near 0.35, whereas the mainland tends to have z
values near 0.15 (Connor and McCoy 2001). In addition, Sugihara (1980) suggested
that the z value provides information on the processes behind the generation of the
community structure.
Species-area data can be sampled using a contiguous gridded plot system.
Contiguous grids are made up of fixed area cells, which share a border with an
adjacent cell. As cells are added, species richness is recorded for progressively
larger regions, while always including the previously recorded smaller subsets of
area (Scheiner 2003). The species-area data are then used to construct the log-log
equation and the c and z values are calculated. Once c and z are determined for a
large area, it is possible to predict the number of species for any smaller area within
the large area; this approach can be used to assess species loss owing to habitat loss.
Two concerns regarding the application of SARs are specific to modern ecolog-
ical studies. Firstly, as suggested above, empirical data can be collected only after
the extinction event has occurred, at which point vulnerable species have been lost.
Secondly, in an analysis of the effect of habitat destruction on metapopulations of
two species, Nee and May (1992) made two key observations: there was a time-
delay on extinction due to habitat loss, and that the first species lost would be
the best competitor. To explain these observations, Tilman et al. (1994) coined the
term “extinction debt”: a time-delayed, deterministic extinction of species due to
habitat loss. In Tilman et al.’s model, after a habitat has been reduced, species
start to go extinct in order from the best competitor to the poorest. Competitive and
dispersal abilities were assumed to be trade-offs; thus, the weakest competitor would
have the best dispersal, and would maintain its local population longer through
immigration from remaining habitats (Tilman et al. 1994; Ovaskainen and Hanski
2001). However, the basic model also suggested that it may take 40–500 years
before the extinction of the top competitor. Because it may take several generations
for the species to go extinct, they are in “debt.” If extinction debt does exist, then it is
a major problem for modern conservation efforts—it would be difficult to convince
policy-makers that extinctions based on models are reliable, and often impossible
to provide the relevant empirical data within a single human lifetime. This study
proposes a solution to this difficult conundrum: we aim to illustrate the use of
SAR models in predicting extinction severity by using paleontological data. The
paleontological record has extensive species presence/absence data from before, as
well as after, an extinction event. This allows for the SAR model to be tested using
empirical data without risking the loss of modern species. Also, the fossil record
spans timescales of thousands to millions of years, which bypasses any lag period
or extinction debt, so that the full extent of the extinction is evident.
While the idea of using the fossil record to provide needed empirical data
for conservation is definitely appealing, it is also important that paleontological
examples are meaningful. A major difference between past extinctions and the
modern biodiversity crisis is that the primary cause of modern habitat loss is due to
human expansion (Leakey and Lewin 1996). For example, a growing requirement
for land has led to an increase in deforestation worldwide, with an estimated
1% of the world’s rain forests being destroyed each year (Wilson 1989). Due to
the majority of the world’s species residing within the world’s rain forests, the
destruction of the rain forests has been used to predict the approximate global
extinction rate due to habitat loss (Primack 2006). As the rain forest’s (or any
other habitat’s) area is reduced or fragmented, the habitat maintains an edge and
an interior environment. The edge is the transitional boundary between different
landscapes, such as between a forest interior and a prairie. While the edge may
become more extensive with progressive habitat loss or deterioration, the interior
of the forest maintains a stable environment (Bannerman 1998). The pattern of
extinction from this sort of progressive habitat loss may be very different from that
of some of the major extinctions observed in the fossil record. Whereas many of the
major mass extinctions may have been caused by widespread habitat deterioration
(e.g., a bolide striking the planet), modern habit fragmentation due to human
interference results in a decrease in habitat while some habitat still remains viable.
In essence, there is a “moving front” of habitat loss and ultimately species loss.
Conservation paleobiology thus needs to find appropriate analogs in the fossil record
for this sort of habitat loss.
One such analog for modern habitat loss would be regressions observed in an
inland sea connected to the open ocean. As the sea-level drops, the total area within
the inland sea may decrease, and facies will migrate oceanwards. Even though the
edge of the sea and related nearshore habitat may now comprise a greater proportion
of the total remaining area, and so more of the region is vulnerable to disturbances
and salinity fluctuations from freshwater input, the interior may continue to remain
stable; previous facies may persist but in a new location. We propose to build
species–area curves from the transgressive and regressive phases of a T-R cycle,
as a means of assessing how habitat loss—in terms of loss of area—affects species
richness.
An idealized transgressive–regressive package would provide a plausible analog
for modern habitat loss, as well as provide an opportunity to construct species–
area curves to examine related species loss. Unfortunately, not all transgressive–
regressive (T-R) intervals are ideal. Although the idea that regressions in a basin
will reduce area seems intuitive, in reality not all regressions are associated with
area decrease. Holland (2012) used measurements of modern topography, and sim-
ulations of siliciclastic deposition (Holland and Christie 2013), to demonstrate that
transgression and regressions may not necessarily correlate with habitat increase
and decrease, respectively. Thus, if one wishes to use T-R cycles as an analog for
habitat loss, then regardless of the environmental interpretation, it is essential to
demonstrate that the habitable area has actually decreased through time.
In addition, a regression would alter the environment/facies at any one locality (to
a shallower habitat), probably resulting in an associated change in the community.
A perceived loss in diversity from the transgressive to regressive phase might simply
be a function of habitat and community change, rather than being due to a loss
in area. Comparisons of area and diversity between transgressions and regressions
need to control for environment, which would usually entail sampling from the
“new” geographic positions of the focal environment, i.e., the regressive samples
would need to be not only stratigraphically higher but also lateral to the transgressive
samples, in keeping with Walther’s Law. By controlling for environment, any
changes in diversity can be attributed to changes in area alone.
There has been previous work using SAR models in the paleontological record
(e.g., Flessa and Sepkoski 1978; McKinney 1998; Barnosky et al. 2005). Although
not strictly speaking an SAR study, Johnson (1974) examined the impact that
regressions have had on marine faunal extinctions, in particular, the impact of
regressions on the numerous Paleozoic epicontinental seas and their related perched
faunas. Perched faunas are faunas living in epicontinental or other similar basins;
in other words, they are “perched” on the continent (Johnson 1974). These faunas
are thus restricted to a geographic location with very little immediate immigration
or emigration in comparison with faunas on open coastlines. Johnson (1974) noted
that such faunas were potentially vulnerable to regressions, particularly as shallow,
gently sloped, basins would experience major habitat loss during a rapid drop in
sea level. However, he focused on general trends at a global scale, spanning longer
chronological time frames, usually millions of years, than those of today’s rapid
habitat loss. Valentine and Jablonksi 1991) studied the effect of sea-level change on
Californian Pleistocene molluscs along the Californian coast. They calculated the
probable size of the Pleistocene coast, and used predetermined, estimated z values
to predict the effect of the fall in sea level during the Pleistocene, and compared their
prediction with observed species diversity in the modern. They found that despite
the sea-level fall and related loss in area, there was no significant change in species
diversity, except they noted that “perched” faunas were more vulnerable.
The present study focuses on the Etchegoin Formation in the Kettleman Hills
of the San Joaquin Basin, California, USA. The Kettleman Hills has an extensive
molluscan fossil record, which was recorded by Woodring et al. (1940). The
fossils represent a perched fauna that lived in a Pliocene inland sea and are
phylogenetically and morphologically similar to today’s molluscs. Stanton and
Dodd (1970) remarked that the faunal lists by zone were representative of the
stratigraphic assemblages and could be used for general faunal time trends. Stanley
and Campbell (1981) suggested that Pliocene diversity on the open Pacific coast
was largely unaffected by Pliocene cooling, but a study by Bowersox (2005), based
largely on the Woodring et al. (1940) dataset, demonstrated that the San Joaquin
Basin taxa experienced extinction rates nearly double that of the open Pacific.
This result is consistent with Johnson’s (1974) and Valentine and Jablonksi (1991)
predictions that perched faunas would be more vulnerable to extinction, possibly as
a consequence of habitat loss. The San Joaquin Basin inland sea experienced several
transgressive–regressive cycles (T-R cycles) throughout the Pliocene. We focus on
one specific T-R cycle within the upper Etchegoin Formation. The combination of
richly fossiliferous units representing a perched fauna in an inland sea with a T-R
cycle makes this setting highly suitable for a study of habitat loss, extinction, and
the species–area relationship.
2 Geological Setting
The Kettleman Hills are located within the San Joaquin basin in California, USA.
During the Neogene, the San Joaquin basin was an inland sea connected to the
Pacific Ocean through a narrow strait west of the present day city of Coalinga
(Fig. 1) (Stanton and Dodd 1970; Bowersox 2005). The Kettleman Hills include
three northwest trending anticlinal domes, referred to as the North, Middle, and
South Domes. There are four formations: the Jacalitos and Etchegoin Formations
(Miocene-Lower Pliocene), San Joaquin Formation (Upper Pliocene), and Talure
Formation (Pleistocene; Woodring et al. 1940). The present study focuses on the
uppermost Etchegoin Formation. This part of the formation consists predominantly
of marine deposits, primarily brown, muddy to clean sandstone, with local lenses
of conglomerate and silt; and some freshwater beds near the top of the formation.
This interval is estimated to be late Early Pliocene in age, approximately 4.2–
4.0 mya (Bowersox 2005). Within this interval, Woodring et al. (1940) described
five faunal zones based on the most abundant taxa: the Patinopecten, Macoma,
Siphonalia, upper Pseudocardium, and Littorina zones. Subsequently, Adegoke
(1969) included all of these five zones within his faunal zone 12. It should be noted
that the stratigraphic nomenclature and even which beds are a part of the Etchegoin
Formation has changed several times since its original description (Anderson 1905).
Consequently, our focal interval, identified as uppermost Etchegoin by Loomis
(1990) and Bowersox (2005), was considered lower Etchegoin in the earlier work
by Woodring et al. (1940).
Each of Woodring et al.’s five zones experienced a local-scale T-R cycle (Stanton
and Dodd 1970) within an overall regressive trend (Stanton and Dodd 1970;
Bowersox 2005). As the basin was only open to the ocean through the Priest Valley
Strait in the northwest during deposition of the Etchegoin Formation (Loomis 1990),
facies migrated south (basinwards) during transgressions, and migrated north during
regressions (Stanton and Dodd 1970; Perkins 1987; Bowersox 2005). Facies in the
southern portion of the basin are typically non-marine; the basin in general became
increasingly brackish and eventually non-marine through time (Stanton and Dodd
1970; Perkins 1987; Bowersox 2005).
Fig. 1 Pliocene paleogeographic map of the San Joaquin basin California, modified from Stanton
and Dodd (1970)
Peak molluscan diversity in the San Joaquin Basin in the Pliocene occurred in the
three lower zones (Patinopecten, Macoma, Siphonalia) of the uppermost Etchegoin,
when the northern part of the basin was still of normal-marine salinity (Bowersox
2005). These units and associated facies are best exposed in a roughly 200-meter
thick interval on and adjacent to the north dome of the Kettleman Hills (Bowersox
2005). The Kettleman Hills do not constitute the entirety of the San Joaquin Basin
but they do account for the vast majority of localities in the uppermost Etchegoin,
and lateral migration of facies in response to local sea-level change is evident.
Woodring et al. (1940) named the Patinopecten zone after the circular pectin
found throughout much of the zone. The rocks are predominantly brownish sand
silt and silty sand. There are some outcrops of sandstone and conglomerate with
scour discontinuity at the base of the overlying Macoma zone. This horizon has
been used to separate the Patinopecten from the Macoma zones, but the sandstone
conglomerate is not recognized in all areas where the Etchegoin Formation outcrops,
and sedimentation appears continuous between the two zones over much of the
region. The Macoma zone consists of a brown to light-grey tuffaceous sandstone
and was originally defined by the abundance of the bent-nose clam, Macoma, and
the absence of Patinopecten.
Stanton and Dodd (1970) used sedimentology, faunal assemblages, and isotopic
analyses from each faunal zone within the Etchegoin Formation to compare the
zones to the modern San Francisco Bay, classifying the zones as either open marine,
outer bay, middle bay, inner bay, or freshwater. In this sense, the zones, while
originally meant to be stratigraphic, are biofacies indicative of paleoenvironment.
As such, the zones, while representing relative stratigraphic position within any
given section, are actually time transgressive: the “zones” or facies also appear
horizontally, with the more oceanward/normal-marine facies appearing to the north
of the more basinward/brackish-water facies. Stanton and Dodd (1970) interpreted
the Patinopecten zone as an outer-to-middle bay setting, while the Macoma zone
consisted of a mosaic of assemblages showing various depositional environments,
but primarily in a middle bay setting. Nonetheless, examination of Woodring et al.’s
(1940) original data indicate that the mollusc communities within the two zones
are very similar; 80% of the species in the Macoma community were also present
in the Patinopecten community. While the definition of the stratigraphic Macoma
zone indicates the absence of Patinopecten, the outer-to-middle bay communities
during both of the Patinopecten and Macoma zones are very diverse and typically
include high abundances of the bivalves Patinopecten, Chione, and to a lesser extent,
Macoma. This Patinopecten-Chione-Macoma (PCM) community was widespread
in the North Dome area during both the Patinopecten and Macoma zones. Although
this community is indicative of deeper conditions relative to the other San Joaquin
Basin communities, it should be noted that all of these communities are nearshore,
shallow-marine communities. Stanton and Dodd (1970) estimated that even the
outer-to-middle bay was probably no more than 50 m deep at most, and it may
have been less than 15 m deep (Bowersox 2005).
Stanton and Dodd (1970) interpreted the oldest zone, the Patinopecten, to
be transgressive, with the middle bay community moving southwards (Stanton
and Dodd 1970; Perkins 1987); subsequently, the middle bay community moved
northward during the more regressive Macoma zone. Thus, the two lowest zones
of Woodring et al. (1940) in the uppermost Etchegoin Formation contain highly
fossiliferous, well-preserved, and diverse samples from a community/biofacies that
migrated laterally with local transgressions and regressions. It is plausible that the
area occupied by the community also varied with the transgression and regression,
and is therefore an appropriate setting to examine the relationship between species,
area, and extinction.
3 Methods
We used Pliocene fossil data and outcrop localities within the Etchegoin Formation
from Woodring et al.’s (1940) geological survey of the Kettleman Hills (Fig. 1).
Using the public land survey system (divides the land into a grid system) and
distances from the section lines provided by Woodring et al. (1940), the localities
were plotted onto 1:24,000 scale topographic maps of the North Dome. The plots
were transferred onto a digital topographic map of the area using the image editing
software GIMP 2.0. The physical map was used first in order to keep the scale
constant while plotting the localities. Contiguous minimal area cells of 1 km2 were
used to calculate the total area represented by the localities. Contiguous minimal
area grids permit the calculation of the number of species as the geographic area
increases. Contiguous grids are a series of fixed area cells, whereby each cell shares
a common border with the adjacent cell (Scheiner 2003; for examples, see Carey
et al. 2006; Rybicki and Hanski 2013). The term “minimal” refers to using as
few square cells as possible to measure each area of interest. As the number of
cells increases, the contiguous area increases; for each larger area, the preceding
area is nested within. The number of species for each progressively larger area
was recorded, and the information was used to construct species–area curves by
graphing the number of species versus area. Log-log plots were used to determine
the equations for the SAR curves and the c and z values. The process was performed
for each of the two lower faunal zones in the uppermost Etchegoin Formation.
The strength of the SAR was assessed by calculating the coefficient of determina-
tion (denoted R2 ) of each graph. R2 signifies how much of the variation in the y-axis
(number of species) is influenced by the x-axis (area). The R2 value is expressed
as a number between 0 and 1; the greater the number, the stronger the relationship.
Using the equations defining the SAR of the transgressive Patinopecten zone, we
predicted the number of species to go locally extinct due to the loss in habitat
caused by the subsequent regression in the Macoma zone. Then we compared the
expected values for the reduced area with those actually observed in the regression.
We examined the extinction severity by comparing how well the regression’s curve
matches the transgression’s curve. If the regression’s curve generally matches the
transgressive curve, and the predicted and observed paleodiversities are similar, that
would suggest the SAR model accurately predicted the extinction. If the regression’s
curve is above the transgression’s curve (its observed paleodiversity is greater
at equivalent areas), then this suggests that the extinction was less severe than
predicted, or that there may be an extinction debt. Finally, if the regression’s curve
is lower than the transgression’s curve (its observed paleodiversity is less than
predicted), then the extinction was more severe (Fig. 2). As the paired transgressive
and regressive curves do not necessarily have observed data at equivalent areas, we
compared the observed species counts at the areas examined on the regressive curve
with the predicted species counts from the equivalent areas on the transgressive
curve.
4 Results
The total area for the regressive-phase Macoma zone was less than that of the
transgressive-phase Patinopecten zone, thus validating an examination of species
loss relative to habitat loss. Both of the species–area curves showed logarithmic
relationships, with very large R2 values and significant slopes (Table 1). The c and z
Fig. 2 A hypothetic SAC. The middle line represents a standard logarithmic SAC (transgressive
curve). The top line (squares) represents a scenario of an extinction debt, and the bottom line
(triangles) represents a scenario in which the extinction is more severe than predicted (regressive
curves). The vertical dotted line represents the extent of area lost
Table 1 R2 , c, and z values Faunal zone R2 value c value z value

of the Etchegoin Formations
faunal zones Patinopecten 0.941 0.175 1.176
Macoma 0.818 0.269 1.632
values for the transgressive and regressive SAR curves are presented in Table 1. The
slope of the regressive Macoma zone curve was less than that of the transgressive
Patinopecten zone curve, and the observed species richness at equivalent areas were
much less than those predicted using the SAC, not only at the maximum area
recorded for the regression, but also for smaller areas as well (Fig. 3).
5 Discussion
The large R2 values signify that there is a strong logarithmic relationship between
the number of species present and area size. This result demonstrates that SAR
curves using paleontological data are similar to modern examples of such curves.
The observed paleodiversity for the regression’s species–area curve was much
less than the diversity predicted using the equation defined from the preceding
transgressive zone. This suggests that the local extinction in the Pliocene inland sea
during this point in Etchegoin time was quite severe. While the long-term regression,
closing of the basin, and a related decrease in salinity eventually led to a larger
extinction at the end of the Etchegoin (Bowersox 2005), the local loss in species
Fig. 3 SACs of the Patinopecten zone (transgression, black circles, solid line) and the Macoma
zone (regression, hollow triangles, dashed line). The vertical dotted line represents the extent of
area lost. Note, the graphs were plotted in arithmetic space for clarity, but the c and z values were
calculated from log-log space
during Macoma zone time detected herein was based on observations from one
community, the PCM community, which was interpreted as living in an outer-to-
middle bay, normal-marine, environment. This environment migrated northward
but changes in salinity or depth probably were not the culprit. Climate cooling
might still have affected this system but the decrease in habitat area may also have
contributed to species loss. During the subsequent transgressive Siphonalia zone,
the PCM community and associated outer-to-middle bay environment re-expanded
southward, and species richness again increased, only to be lost completely for the
remainder of Etchegoin deposition during the severe regression that followed the
Siphonalia zone. This pattern is consistent with a species-area model. That said, as
the transgressive curve serves as a prediction—a null model—the more severe than
expected loss of species indicates that habitat loss alone was insufficient to explain
the species loss. What is causing the difference? We consider three possibilities.
He and Hubbell (2011) examined the use of SARs in predicting extinctions
by comparing them to endemic–area relationships (EAR). Endemic species exist
only in a single geographical area (although the size of this area varies among
species); in other words, the species’ habitat is a single, continuous area (Green and
Ostling 2003). The EAR predicts a species will go extinct only when its total single
geographic habitat is destroyed, while the SAR assumes all species are randomly
distributed, and does not distinguish between endemic and widespread species.
If the species are randomly distributed, SAR and EAR curves will be the same.
However, most species are non-randomly distributed: many species are gregarious
and form “clumped” distributions (Sasaki 1997; Leighton and Schneider 2004). He
and Hubbell (2011) built SAR and EAR curves for 279 species of passerine birds
from the United States of America’s lower 48 states. They noticed that the z values
from the SAR curves were much larger than the z values from the EAR curves,
i.e., the slopes of the SAR curves were steeper and species richness values were
smaller than those from the EAR curves at comparable areas (See Fig. 2 from He
and Hubbell 2011). So for non-randomly distributed species, the SAR and EAR
curves cannot be used to infer one another. He and Hubbell argue that this difference
between the curves is due to the fact that when generating an SAR curve, the area
required to find the first species is much smaller than the area that must be lost for
the last species to go extinct. Therefore, it requires a greater loss of area to lose a
species than it does to add a new species. The SAR model does not take this into
account and assumes the area effect on the addition or loss of a species is the same.
Based on this argument, SAR curves will always overestimate the extinction debt.
However, the results obtained in this study contradict He and Hubbell’s (2011)
claim. This may be due to the fact that the local extinction in the Kettleman
Hills occurred over an extended period of time, almost certainly longer than any
extinction lag effect (i.e., long enough to exceed any extinction debt). However,
this process does not explain the sampling error He and Hubbell mentioned, that
SARs do not account for clumped distributions of species and instead assumes that
distributions are random. This concern may be overstated. First, He and Hubbell
(2011) argued that SAR curves will always be steeper than EAR curves, due to the
clumped nature of species distributions. The steepness of the SAR slope will project
a lower diversity than that of the EAR for any given area when area is small, thus
exaggerating the projected loss of species if habitat was reduced to that area. Yet, as
the species counts for larger regions (equivalent to gamma diversity) of SAR curves
will be the same as those of EAR curves, the slope of an SAR curve, relative to the
related EAR curve, is largely determined by the diversity within the first few sample
plots, i.e., the smallest plots. If the species distribution is extremely clumped, and the
first sample is taken at a point of lowest possible alpha diversity, then the SAR curve
may be very steep, and will likely overestimate extinction debt. However, if the first
sample is taken at a point of maximum alpha diversity, within a concentration of
clumps, the SAR and EAR curves may look very similar, if not identical. The SAR
curve will have a gentler slope and would estimate extinction more accurately. He
and Hubbell (2011) have thus identified an important but easily fixable concern;
workers building SAR curves must use care in choosing those samples representing
the smallest areas on the curve.
An additional factor mitigating the He and Hubbell (2011) concern, for pale-
ontological or historical studies in particular, is the process of time-averaging.
Just as a function of the normal dispersal of juveniles, the spatial distribution of
species from one generation to the next is not identical, varying even (or especially)
on smaller scales. Most rock units, beds, or even unlithified sediments are time-
averaged, in that even though the depositional environment of the bed did not change
significantly, the fauna preserved within the bed represent multiple time-slices, and
therefore multiple communities and multiple generations and spatial distributions
of species (Olszewski 1999). This effect may be enhanced further by potential post-
depositional reworking, and temporal mixing of sediments in some beds. The rock
or sediment thus represents an average of time but given the spatial variation of
species among generations, the rock or sediments will typically accumulate diversity
through multiple time-slices. Time-averaging is sometimes a major concern because
over a sufficiently long enough period of time, there is an increasing risk of mixing
taxa that never actually lived together. However, such conditions are usually readily
identifiable by the beds’ stratigraphic context (Kidwell and Holland 2002) and
by the faunal assemblages within them (e.g., taxa with equivalent preservation
potential are preserved differently in the same bed). In many cases, including the
use of SAR curves, time-averaging is a benefit to paleontologist and ecologist alike
because it reduces noise (Walker and Bambach 1971; Olszewski 1999), including
very localized spatial variation. A time-averaged sample will be less vulnerable to
clumping of species because such spatial variation is softened by the averaging of
this variation through time within the bed (Leighton and Schneider 2004).
For many fossil beds, the diversity within a time-averaged bed, at least for
taxa such as molluscs that have good preservation potential, is potentially the
same or greater than if one did a count on a modern plot (or a single time-
slice). Studies by Kidwell and Tomašových (2009, 2013) comparing live versus
dead molluscan assemblages have demonstrated that the fidelity of species counts
for a given assemblage is quite strong, and that dead assemblages have higher
alpha diversity than live assemblages. Tyler and Kowalewski (2016), in a similar
live versus dead study, also noted that many dead samples had greater diversity
than the live counterpart, presumably due to the time-averaging effect described
above. So when collecting species-area data, the diversity from the first, smallest,
sample is often greater than that of a modern sample, or single time-slice, from the
equivalent area. As Kidwell and Tomašových (2009) noted, the beta diversity of a
time-slice is effectively being converted to alpha diversity of the paleontological
sample. However, eventually as area increases, this effect becomes diminished.
When sampling the bed from multiple localities, there will be an area at which
maximum diversity is reached, and the curve will flatten out, just as it does in the
modern. At these larger scales, species clumping is irrelevant to diversity counts of
both modern and fossil samples. If the diversities of the larger areas are the same
for both fossil and modern curves, and the diversities of the smaller areas used in
the fossil curves are greater (due to time-averaging), then the slope of the fossil
curve becomes shallower, essentially negating He and Hubbell’s problem. Several
other studies have argued against He and Hubbell’s claim that SAR’s will always
overestimate extinctions (Brooks 2011; Fattorini and Borges 2012; Pereira et al.
2012; Thomas and Williamson 2012), but in any event, this problem seems to be
less of a concern for a paleontological study, and overestimation of the extinction
clearly was not the problem for the present study in the Etchegoin Formation.
Holland (2013) studied the effect of habitat loss on the Pleistocene extinctions
on the California coast using numerical simulations to measure relaxation times,
the time required for a population to reach equilibrium after an extinction event or
change in environment (Diamond 1972). One equilibrium state is when a population
equals zero, so there is a strong potential relationship between equilibrium and lag
times, or extinction debt. Based on these simulations, Holland suggested several

reasons why predicted and observed patterns of species loss may differ. One reason
is species longevity: If a species has a low mortality rate, or has a high likelihood of
colonizing vacated sites, that species could persist during times of decreased habitat
area, which could increase relaxation time. The inverse is true as well; if a species
has a high mortality rate or is a poor colonizer, relaxation time could decrease.
Variation in speciation rates also may influence relaxation time; after habitat had
been reduced, an increase in speciation would lengthen the time required for the
biodiversity to fall to equilibrium (Holland 2013), while similarly, a decrease in
speciation might shorten the time. These factors could potentially have played a
role in the difference between the Etchegoin’s predicted and observed species loss.
One additional possible explanation for the apparently more severe local loss of
species during the Macoma zone is that the SAR model does not directly account
for biological interactions, and the effect of the loss of one species on other species.
There are reports outlining ecosystem collapses in the modern, sometimes caused
by the extinction of keystone species, a species crucial to maintaining diversity
and stability in its ecosystem (Paine 1966, 1969; Woodruff 1989; Rohlf 1991;
Mills et al. 1993). Typical examples of such keystone species include top predators
(Paine 1969) and biofacilitators, such as reef-forming corals (Bruno et al. 2003).
An extinction cascade occurs when the extinction of one species leads to the
extinction of one or more species, until eventually the ecosystem reaches a tipping
point in which too many species have been lost, in turn leading to a trophic level
collapse (Ricklefs and Miller 1999). There is still debate as to whether increased
food web complexity increases or decreases stability. Initial models (May 1973;
Pimm 1979; Plotnick and McKinney 1993) suggested that more complex food webs
were more vulnerable to collapse, precisely because of the strong interconnectivity
and interdependence among species. More recently, both model (Roopnarine 2006)
and empirical (Fox and Olsen 2000; Leighton and Schneider 2008) studies have
suggested that high diversity within a given trophic level might stabilize the trophic
web and provide a buffer against an extinction cascade. If a single species is lost,
its competitors, which are likely to be functionally similar, will compensate for the
loss by filling the relevant niche. Regardless of the effects, both good and bad, of
increasing connectivity among species in a foodweb, there is some threshold of
species loss at which an extinction cascade occurs (Roopnarine 2006). While it
is well beyond the scope of the present study to solve this problem, the greater
than expected species loss observed in the present study could be an example of
an ecosystem that exceeded the threshold of species loss necessary to trigger an
extinction cascade. If this is the case, then the loss of habitat in the Macoma zone
may have been a trigger, rather than the immediate cause. The combination of
using SAR curves to predict severity with a detailed examination of which species
(including possible keystones) are lost through time is a promising avenue for future
research.
The results presented herein are not meant to be the definitive answer to diversity
and extinction patterns in this interval of the uppermost Etchegoin Formation of
the Kettleman Hills. Rather, the goal was simply to demonstrate that an SAR-based
approach could be applied to the fossil record and to show that such research could
potentially provide information that is relevant to modern conservation problems.
Furthermore, predictions from species–area curves could potentially serve as null
hypotheses; if species loss was significantly greater or less than that predicted,
then we need to understand why. An important point of the present study is that
although the results do not support nor disclaim the existence of extinction debt
(nor could they, given the time frame of the study), there is still a relationship
between habitat loss and species extinction. Even if there is a lag period, and it may
take 40–500 years for extinction to occur, the species affected are still doomed to
die, if there is no intervention. This raises an ugly problem for conservationists:
convincing policy-makers to consider truly long-term solutions to conservation
problems. Hopefully, more paleontological studies showing long-term effects of
habitat loss will continue to provide much-needed empirical data. We also need
to consider the real possibility that just as modern climate change and habitat
loss are accelerating at rates that are exceptionally fast when compared to past
rates, the unfortunate result may be that what we have previously considered to be
long-term negative consequences on ecosystems may instead occur during a single
human generation. For example, the rate of species extinction has already increased
substantially in just a few decades (Wilson 2003).
The purpose of this study was to provide a worked example to review the
use of the SAR model by studying it in a paleontological context, rather than a
definitive case study. As such there are some caveats that need to be addressed
for future research. Firstly, this was a meta-analysis using data originally collected
for mapping and structural interpretation, rather than for a detailed paleoecological
study. Naturally, future research should focus on collecting fossil data specifically
for the purpose of ecological studies using SAR curves. Secondly, this study used
only one T-R package containing two faunal zones, as these units were the only such
zones in the section that were immediately stratigraphically adjacent, i.e., there were
no major temporal gaps within the transgressive–regressive cycle, and they were the
only such zones that captured the same community/biofacies. While the results are
of interest, one result provides only one example. Further research from other times
and regions is needed to determine if the extinction severity patterns found in this
study are unique only to those zones, or if there is a more general pattern throughout
the paleontological record. We hope that this study encourages other workers to
explore these possibilities.
6 Conclusion
1. A local extinction within the Californian Pliocene inland sea was significantly
more severe than predicted using species–area curves. These results contradict
an earlier study which claims that the species–area relationship will always
overestimate extinctions.
2. If extinction debt does exist, and it takes 40–500 years for the extinction event
to begin, it would be practically impossible to collect empirical data to confirm
extinction debt within a single human lifetime. The paleontological record spans
a coarser timescale of thousands to millions of years, essentially bypassing any
possible lag period, and so provides an opportunity to test extinction severity and
related models with empirical data.
3. Modern ecological studies are necessarily limited to using predictive models of
species extinction, which can only be empirically tested through the passage of
time, at which point species may be irrevocably lost. The paleontological record
provides species presence/absence data from before as well as after an extinction
event, which allows for the models to be tested with empirical data without the
loss of modern species.
4. Many extinctions studied in the paleontological record were caused by catas-
trophic events, such as a bolide impact or large scale volcanic activity, which
lead to widespread habitat deterioration. Modern human-driven habitat loss is
different and acts as a “moving front,” whereby habitat decreases while the
remaining habitat maintains a stable environment. Thus, there needs to be a more
appropriate analog in the fossil record. Regressions affecting perched faunas may
be an excellent analog for modern habitat loss, and further exploration of such
systems in the fossil record has considerable potential to test SAR models and
modern extinction predictions.
Acknowledgments The authors would like to thank Steve Holland and Alycia Stigall for their
thorough and extremely thoughtful reviews, which greatly improved the manuscript. We thank
Kristina Barclay for reading an earlier draft of the manuscript. We also wish to thank Carrie Tyler
and Chris Schneider for inviting this contribution. Funding for this research included an NSERC-
USRA award to the lead author, and an NSERC Discovery grant to the second author.
References
Adegoke OS (1969) Stratigraphy and paleontology of the marine Neogene formations of the
Coalinga Region, California. University of California, Publications in Geological Sciences,
Berkeley. 80, p 241
Anderson FM (1905) A stratigraphic study in the Mount Diablo region of California. Proceedings
of the California Academy of Sciences, third Series 2:155–248
Bannerman S (1998) Biodiversity and interior habitats: the need to minimize edge effects.
Extension note/British Columbia. Ministry of Forests, Research Program, Victoria
Barnosky AD, Carrasco MA, Davis EB (2005) The impact of the species-area relationship on
estimates of paleodiversity. PLoS Biol 3:1356–1361
Barnosky AD, Matzke N, Tomiya S et al (2011) Has the Earth’s sixth mass extinction already
arrived? Nature 471:51–57
Bowersox JR (2005) Reassessment of extinction patterns of Pliocene molluscs from California
and environmental forcing of extinction in the San Joaquin Basin. Palaeogeogr Palaeoclimatol
Palaeoecol 221:55–82
Brooks TM (2011) Extinctions: consider all species. Nature 474:284
Bruno JF, Stachowitz JJ, Bertness MD (2003) Inclusion of facilitation into ecological theory.
Carey S, Harte J, del Moral R (2006) Effect of community assembly and primary succession on
the species-area relationship and disturbed ecosystems. Ecography 29:866–872
Connor EF, McCoy ED (1979) The statistics and biology of the species-area relationship. Am Nat
113:791–833
Connor EF, McCoy ED (2001) Species-area relationships. In: Levin SA (ed) Encyclopedia of
biodiversity, 5 p 397–412
Diamond JM (1972) Biogeographic kinetics: estimation of relaxation times for avifaunas of
southwest Pacific islands. Proc Natl Acad Sci USA 69(11):3199–3203
Dony JG (1963) The expectation of plant records from prescribed areas. Watsonia 5:377–385
Ehrlich PR, Ehrlich AH (1981) Extinction: the causes and consequences of the disappearance of
species. Random House, New York
Fattorini S, Borges PAV (2012) Species–area relationships underestimate extinction rates. Acta
Oecol 40:27–30
Flessa KW, Sepkoski JJ (1978) On the relationship between Phanerozoic diversity and changes in
habitable area. Paleobiology 4:359–366
Fox JW, Olsen E (2000) Food web structure and the strength of transient indirect effects. Oikos
90:219–226
Green JL, Ostling A (2003) Endemics-area relationships: the influence of species dominance and
spatial aggregation. Ecology 84:3090–3097
He F, Hubbell SP (2011) Species-area relationships always overestimate extinction rates from
habitat loss. Nature 473:368–371
Holland SM (2012) Sea level change and the area of shallow-marine habitat: implications for
marine biodiversity. Paleobiology 38:205–217
Holland SM (2013) Relaxation time and the problem of the Pleistocene. Diversity 5(2):276–292
Holland SM, Christie M (2013) Changes in area of shallow siliciclastic marine habitat in response
to sediment deposition: implications for onshore-offshore paleobiologic patterns. Paleobiology
39:511–524
Johnson JG (1974) Extinction of perched faunas. Geology 2(10):479–482
Kidwell SM, Holland SM (2002) The quality of the fossil record: implications for evolutionary
analyses. Annu Rev Ecol Syst 33:561–588
Kidwell SM, Tomašových A (2009) Fidelity of variation in species composition and diversity
portioning death assemblages: time-averaging transfers diversity from beta to alpha levels.
Paleobiology 34(1):94–118
and conservation biology. Trends Ecol Evol 44:539–563
Leakey R, Lewin R (1996) The sixth mass extinction: patterns of life and the future of humankind.
Doubleday, New York
Leighton LR, Schneider CL (2004) Neighbor proximity analysis, a technique for assessing spatial
patterns in the fossil record. PALAIOS 19(4):396–407
Leighton LR, Schneider CL (2008) Taxon characteristics that promote survivorship through the
Permian-Triassic interval: transition from the Paleozoic to the Mesozoic brachiopod fauna.
Paleobiology 34(1):65–79
Levins R (1969) Some demographic and genetic consequences of environmental heterogeneity for
biological control. Bull Entomol Soc Am 15:237–240
Loomis KB (1990) Depositional environments and sedimentary history of the Etchegoin formation,
west-central San Joaquin Valley, California. In: Kuespert JG, Reid SA (eds) Structure, stratig-
raphy and hydrocarbon occurrences of the San Joaquin Basin, California. Society of economic
Paleontologists and Mineralogists and American Association of Petroleum Geologists Pacific
Sections, book 64 and guidebook 65, p 231–246
MacArthur RH, Wilson EO (1967) The theory of island biogeography. Princeton University Press,
Princeton
May RM (1973) Stability and complexity in model ecosystems. Princeton University Press,
Princeton
McKinney ML (1998) On predicting biotic homogenization: species-area patterns in marine biota.
Glob Ecol Biogeogr 7(3):2897–2301
Millennium Ecosystem Assessment (2005) Ecosystems and human well-being: biodiversity
synthesis. World Resources Institute, Washington
Mills SL, Soulé ME, Doak DF (1993) The keystone-species concept in ecology and conservation.
Bioscience 43:219–224
National Research Council (1980) Research priorities in tropical biology. National Academy of
Sciences, Washington
Nee S, May RM (1992) Dynamics of metapopulations: habitat destruction and competitive
coexistence. J Anim Ecol 61:37–40
Olszewski T (1999) Taking advantage of time-averaging. Paleobiology 25:226–238
Ovaskainen O, Hanski I (2001) Spatially structured metapopulation models: global and local
assessment of metapopulation capacity. Theor Popul Biol 60(4):281–302
Paine RT (1966) Food web complexity and species diversity. Am Nat 100:65–75
Paine RT (1969) A note on trophic complexity and community stability. Am Nat 103(929):91–93
Pereira HM, Borda-de-Água L, Martins IS (2012) Geometry and scale in species-area relation-
ships. Nature 482:E4–E5
Perkins JA (1987) Provenance of the Upper Miocene and Pliocene Etchegoin Formation: Implica-
tions for Paleogeography of the Late Miocene of Central California. United States Geological
Survey Open-File Report, p 87–167
Pimm SL (1979) Complexity and stability: another look at MacArthur’s original hypothesis. Oikos
33:351–357
Pimm SL, Russell GJ, Gittleman JL et al (1995) The future of biodiversity. Science 269:347–350
Plotnick RE, McKinney ML (1993) Ecosystem organization and extinction dynamics. PALAIOS
8:202–212
Preston FW (1962) The canonical distribution of commonness and rarity. Ecology 43:185–215
Primack RB (2006) Essentials of conservation biology, 4th edn. Sinauer Associates, Sunderland
Richter-Dyn N, Goel NS (1972) On the extinction of a colonizing species. Theor Popul Biol
3(4):406–433
Ricklefs RE, Miller GL (1999) Ecology, 4th edn. W.H. Freeman and Company, New York
Rohlf DJ (1991) Six biological reasons why the endangered species act doesn’t work and what to
do about it. Conserv Biol 5:273–282
Roopnarine PD (2006) Extinction cascades and catastrophe in ancient food webs. Paleobiology
32:1–19
Rosenzweig ML (1995) Species diversity in space and time. Cambridge University Press,
Cambridge
Rybicki J, Hanski I (2013) Species-area relationships and extinctions caused by habitat loss and
fragmentation. Ecol Lett 16:27–38
Sasaki A (1997) Clumped distribution by neighborhood competition. J Theor Biol 186(4):415–430
Scheiner SM (2003) Six types of species-area curves. Global Ecol and Biogeogr 12:441–447
Stanley SM, Campbell LD (1981) Neogene mass extinction of Western Atlantic molluscs. Nature
293:457–459
Stanton RJ Jr, Dodd JR (1970) Paleoecological techniques: comparison of faunal and geochemical
analyses of Pliocene paleoenvironments, Kettleman Hills, California. J Paleontol 44:1092–
1211
Strebel DE (1985) Environmental fluctuations and extinction single species. Theor Popul Biol
27:1–26
Sugihara G (1980) Minimal community structure: an explanation of species abundance patterns.
Am Nat 116:770–787
Thomas CD, Cameron A, Green RE et al (2004) Extinction risk from climate change. Nature
427:145–148
Thomas CD, Williamson M (2012) Extinction and climate change. Nature 482:E4–E5
Tilman D, May RM, Lehman CL, Nowak MA (1994) Habitat destruction and the extinction debt.
Nature 371:65–66
Tyler CL, Kowalewski M (2016) Quantifying the fidelity of the fossil record across multiple higher
taxa in nearshore marine macro-invertebrate assemblages (North Carolina USA). Geol Soc
America Abstr Prog 48(7)
Valentine JW, Jablonksi D (1991) Biotic effects of sea level change: the Pleistocene test. J Geophys
Res 96:6873–6878
Walker KR, Bambach RK (1971) The significance of fossil assemblages from fine-grained
sediments: time-averaged communities. Geol Soc America Abstr Prog 3:783–784
WCMC (1992) Global biodiversity: status of the Earth’s living resources. WCMC
Williams CB (1943) Area and number of species. Nature 152:264–267
Williams CB (1964) Patterns in the balance of nature. Academic Press, London
Wilson EO (1989) Threats to biodiversity. Sci Am 261:108–116
Wilson EO (2003) The future of life, 1st edn. Random House, New York
Woodring WP, Stewart R, Richards RW (1940) Geology of the Kettleman Hills oilfield, California;
stratigraphy, paleontology, and structure. U. S. Geological Survey, Prof. Paper 195 p 170
Woodruff DS (1989) The problems of conserving genes and species. In: Western D, Pearl MC (eds)
Conservation for the twenty-first century. Oxford University Press, New York, pp 76–88
Marine Refugia Past, Present,
and Future: Lessons from Ancient
Geologic Crises for Modern Marine
Ecosystem Conservation
Chris L. Schneider
Abstract Refugia are one means of species survivorship during a global crisis.
As the Earth is facing a major crisis in the marine biosphere, the study of
refugia through past extinctions and other global crises is relevant to creating
and maintaining effective marine reserves (including marine protected areas and
other formally established havens for conservation). A synthesis of previous studies
identifies the following properties common to most definitions of a refugium: (1)
During a global crisis, a species can persist in a refugium, which can include a
range shift, habitat shift, or migration or contraction to an isolated geographic area.
Subsets of isolated geographic refugia include life history refugia (areas necessary
for breeding), cryptic refugia (small areas, must remain connected for populations
to remain viable), and harvest refugia (defined from the modern literature to escape
overfishing pressure). (2) In the refugium, the habitat may remain stressed but is
sufficiently habitable for the species to maintain sufficient albeit small populations
(relative to pre-crisis population size) over many generations. (3) After the crisis
ends, the species emerges from the refugium and expands during the recovery
interval. Otherwise, the refugium will become a refugial trap in which the species
remains a relict population or ultimately becomes extinct.
The present understanding of refugia from the geologic past comes from
three sources, namely fossil data, phylogeographic reconstructions, and species
distribution models, the latter two being more common for studies across the last
glacial maximum. The synthesis herein suggests several important factors when
considering the future of marine reserves. Because climate change is an ongoing
process, the present refugia of marine reserves may not be sufficient for the future
survival of marine species. Short-term refugia of some present marine reserves
may deteriorate because of further climate change and have to be abandoned for
new long-term options as new habitats become available. Cryptic refugia of small
reserves must remain connected in terms of species’ dispersal and exchange, but
C. L. Schneider ()
Canada

164 C. L. Schneider
must also be flexible, in that cryptic refugia naturally are sometimes ephemeral
because of habitat heterogeneity through time. Finally, habitats in marine reserves
must be of sufficiently low stress to maintain viable populations, but should
frequently be re-evaluated to avoid becoming refugial traps in the future.
Keywords Extinction survivorship · Climate change · Range shift · Habitat

shift · Refuge · Species distribution · Conservation paleobiology
1 Introduction
Through all of the global biotic crises in Earth’s past, life has found a way to
survive, recover, and ultimately, flourish once again. One means of survival through
past global crises has been the retreat of taxa into refugia, which are areas where
species, lineages, and ecosystems persisted, sheltered from the fatal environmental
devastation that culminated in a global mass extinction. Because extinction-causing
stresses are present in modern oceans, refugia may again become an important factor
in the survival of marine taxa. And because human intervention may be necessary
for establishing refugia—in the form of marine reserves, herein including marine
protected areas and other places formally established for conservation purposes—
modern conservation efforts may be strengthened through a better understanding of
why and how ancient refugia were successful through global crises.
Accepting that the Earth’s biosphere is facing a mass extinction rivaling those
of the geologic past, it is critical that if ecosystems are to survive—and be able
to provide the support and services needed by the global human population—then
actions to protect ecosystems need to be taken immediately and especially with
forethought to the future. In the face of impending climate change that impacts
ocean water in terms of warming, oxygenation, acidification, circulation, and other
factors, marine reserves as they currently exist may be insufficient to protect systems
from a multitude of stressors. The design and establishment of future marine
reserves, along with the management of current systems, must take into account
future habitat changes while providing refugia in a deteriorating ocean environment.
Models of future changes are hypotheses, not certainties; basing future marine
reserves on any one model could result in failure (Makino et al. 2015). The answer
to which areas might remain habitable becomes an uncertain prediction, especially
without a baseline reference of what constitutes a successful refugium. Meanwhile,
the application of the refugial concept to management and design of marine reserves
is not a novel idea; increasingly, research is focusing on the possibility that marine
reserves can be viable refugia. Although this body of work is producing critical
information for conservation efforts, it often lacks perspectives that can only be
provided by the fossilized past, specifically how a refugium facilitated survival
through a biosphere crisis, from its inception, and very importantly, through its
persistence to its successful end.
Although the extensiveness, and even value, of ancient refugia is debated (e.g.,
refugia create relict populations; Rickards and Wright 2002), refugia through
Marine Refugia Past, Present, and Future: Lessons from Ancient Geologic. . . 165
ancient mass extinctions have been shown to exist and contribute to global biosphere
recovery (e.g., pre-recovery graptolite diversification in Late Ordovician refugia;
Xu et al. 2005). Since the defining work of Vermeij (1986), the understanding of
refugia has greatly increased with the recovery of additional fossil and geological
evidence, along with a rapidly expanding body of phylogeographic studies across
the last glacial maximum and the incipience of species distribution models in
paleoecological niche reconstructions. Furthermore, investigations into ancient
refugia increasingly emphasize the need to understand why those refugia were
successful, sometimes also identifying factors relevant to modern marine biosphere
survival. Therefore, a synthesis of investigations into refugia is timely, emphasizing
the contributions from the studies of ancient refugia to modern conservation and the
prospect for future refugial investigation.
The terms “refuge” and “refugium” are ubiquitous across ecology and pale-
oecology. Aside from a means of survivorship through environmental crises and
mass extinctions, the terms have also been used for ecological phenomena, such as
predation refuges or diurnal refuges. Although these are appropriate uses of the term
“refuge” and also ensure survivorship, I focus below on the mechanism that allows
populations to survive through global environmental perturbations, through multiple
generations, and usually, over geologic time scales. Here, I use the Latinized
“refugium” (pl., “refugia”) to indicate the strategy of extinction survivorship.
2 Defining Refugium
Refugia, as identified across scientific study, are fairly diverse in definitions and
types. In general, during an environmental crisis, a species contracts its range to a
smaller geographic area, shifts its range to a new region, or takes advantage of a
new, but unperturbed, habitat. Then, a species spends time in the refugium, where it
can evolve genomes that define an endemic population or, given time, speciate into
one or more new taxa, sometimes adapting irrevocably to the specific conditions of
the refugium. The refugial interval of survival can last from several generations to
millions of years, depending on the severity and duration of the global crisis. Finally,
after environmental recovery in extra-refugial habitats, the species expands out of
its refugium. Thus, the refugial survival strategy has three stages: entry into the
refugium with the onset of environmental perturbation; time spent in the refugium
while the global environmental crisis continues elsewhere; and re-expansion out of
the refugium with the bettering of adverse environmental conditions.
Other types of refugia, such as physiological or temporal refugia, have been
defined in the modern marine literature and depend neither on habitat nor on
location. These are outside the scope of this review as they pertain to individuals
or single generations. Below, I use the perspective of a single hypothetical species
in order to expound on what ideally happens during its time in a refugium. In the
survival of a species, the consensus of most definitions of refugia considers the
following aspects:
166 C. L. Schneider
A Species Must Have a Range Contraction, Range Shift,

or Migration in Order to Escape the Onset of Global
Environmental Degradation That Would Otherwise
Cause Extinction of That Species
One general consensus for defining a refugium is that a species must have altered its
range in order to escape adverse environmental conditions. A refugium is a place—
either isolated and restricted in size (such as the Iberian Peninsula refugium at the
last glacial maximum for species that are presently more northerly) or a shift in
habitat (such as a deep water refugium to avoid lethal conditions in shallow water) or
geographic location (such as a southward shift of North Atlantic temperate species
during glacial maxima). The refugial area is generally understood to be smaller in
size than that of the original species’ range (Maggs et al. 2008; Ashcroft 2010;
Zonneveld et al. 2010a). Furthermore, these refugia are areas within species’ ranges
that remain habitable (“in situ” refugia; Ashcroft 2010; Gavin et al. 2014; “internal”
refugia; Shoo et al. 2013) or are acceptable habitats outside of the original range
to which species migrate (“ex situ” refugia; Ashcroft 2010; Gavin et al. 2014;
“external” refugia; Shoo et al. 2013). Although it may seem obvious, a refugium
must be accessible to the species that needs it—the location cannot be a refugium,
no matter how pristine the environment, if organisms cannot gain entrance.
Ideally, to identify a refugium, multiple species should inhabit the area in
question, each of which has gone through a range contraction into their refugial
populations (Vermeij 1986). As a result of the intense study of demographics of
species across the last glacial maximum, many of the same refugia have been
identified for species of vastly different life modes and trophic levels. For instance,
North Atlantic refugia in the English Channel and Iberian Peninsula coast have been
identified for macroalgae (Hoarau et al. 2007; Neiva et al. 2014; Assis et al. 2016),
invertebrates (Jolly et al. 2006; Campo et al. 2010; Albaina et al. 2012), and fishes
(Chevolot et al. 2006; Almada et al. 2012; Mateus et al. 2016). In the ancient fossil
record, faunas within refugia through mass extinctions were often more diverse than
extra-refugial faunas (e.g., Ordovician brachiopods; Rassmussen and Harper 2011).
One factor in determining which species has a greater chance of refugial survival
may be original range size. A species with a broad distribution before an extinction-
causing event may be more likely to survive (Kiessling and Aberhan 2007), at
least in the short term in the geologic sense (Clapham et al. 2013). In terms
of refugial survival, a large range could increase the chances of refugia forming
within the species’ distribution. However, other studies found no relationship
between range size and survivorship, indicating that other factors may be more
important in determining survivorship (e.g., Permo-Triassic brachiopods; Leighton
and Schneider 2008; Cretaceous molluscs; Myers et al. 2013).
Refugia, as identified in the literature, comprise several different strategies: range
and habitat shifts, isolated geographic refugia, cryptic refugia, and life history
refugia. In addition, modern anthropogenic pressures on marine species have also
created a type of refugium unique to this interval of geologic history, the harvest
refugium. Certainly, species may rely on more than one type of refugium for survival
through a biotic crisis, and hybrid refugial types may occur. Likewise, marine
reserves may be targeted as a specific type of refugium, or may serve as multiple
types of refugia, depending on the species escaping the environmental crisis. The
studies covered in this synthesis are summarized in tables in the appendix, and the
types of refugia are briefly reviewed below.
Range Shifts
The range shift of species is one of the earliest hypothesized methods of survival
during a global environmental crisis. For example, Packard (1886) recognized the
southward shift of North American species away from the continental ice sheet and
used the term “refugia” to denote places where species survived. Darwin (1859)
suggested an ecosystem-wide survival mechanism during the ice age, in that species
from the same habitat shifted their ranges together with changing climate, thus
keeping their communities intact.
In a range shift, changes in the range of a species parallel the changes in the
environment. Throughout the range shift, the species maintains its preferred habitat
(Fig. 1a, b). A shift in range can be necessary with warming or cooling ocean
temperatures or other climate-related changes. For instance, if sea level drops, the
species moves with sea level, following its perfect depth range on the reduced shelf
area (e.g., contraction to the shelf edge during sea level lowstands of glacial maxima;
Holterhoff 1996). Or, if climate warms, a species tracks that depth range and habitat
to higher latitudes as the optimal temperature moves away from the equators (e.g.,
thermal refugia in higher latitudes for shallow-water tropical corals during the last
interglacial of Australia; Greenstein and Pandolfi 2008).
As with any refugial strategy, a species escapes lethal conditions; otherwise,
it would remain in its original range. A range shift by a species is a series of
responses by individuals, such that there is no single, coordinated mass exodus into
a refugium during a crisis (Bennett and Provan 2008). Thus, the range shift is a
gradual process, with organisms propagating along the habitable leading edge of
their ranges and increasing die-off along the trailing edge. In other words, while
environmental deterioration occurs in part of a species range, newly opened habitat
along another margin of its range allows for local population expansion. Thus, the
geographic centroid of the population moves through space and time, tracking its
optimal habitat during environmental change. However, this simplified model does
not consider potential novel stresses in the newly opened habitats, such as the lack
of suitable rocky substrate and increased thermal stress for rockweed with warming
climate (Hoarau et al. 2007) or a decrease in the aragonite saturation state as tropical
corals shift away from equatorial warming (Descombes et al. 2015), which could
alter the process of the range shift.
Range shifts during times of global stress generally coincide with demographic
contractions, compared to expansion of ranges in more amenable intervals. For
example, as climate cooled during glaciation, temperate and tropical taxa shifted
168 C. L. Schneider
a b c
sea level
previous
1
sea level
d e 3
sea level
f g
1
2
Fig. 1 Types of refugia. (a) Range shift: As the population moves, the centroid of the population
distribution shifts. (b) Range shift: The preferred habitat is maintained; in this example, with sea
level rise, the population moves upslope, following the migration of its preferred habitat. (c) Range
shift in a fluctuating climate, as in a temperate species during glacial–interglacial cycles: (1) range
during one climate regime, such as the interglacial range of a northern temperate species; (2) range
during the other climate regime, such as the glacial range of a northern temperate species that
has shifted south; (3) the “climatic refugium” of Neiva et al. (2014), a stable area of demographic
overlap between the two range extremes, where the environment is always habitable and is always
occupied by the species. (d) Habitat shift: As the population moves, the centroid of the population
distribution shifts. (e) Habitat shift: The preferred habitat is abandoned for a new habitat; in this
example, sea level is static and the population moves upslope, abandoning the deeper water habitat.
(f) Isolated geographic refugia: the population (1) contracts to one or more smaller areas within
its original range or (2) migrates to one or more areas outside of its original range. (g) Cryptic
refugia: The population either reduces to rare individuals, or reduces to small populated pockets;
both types can be within or outside of the original range, and both types constitute a connected
metapopulation
toward the equator, often into smaller populations (Ashcroft 2010). Meanwhile,
the colder global climate was optimal for northern polar species, which expanded
their ranges southward (Almada et al. 2012). Conversely, interglacials are times of
thermal stress for polar species. Their present ranges are expected to contract even
further poleward with warming climate (Stewart et al. 2010; Ashcroft 2010).
Range shifts between interglacial and glacial periods can result in overlapping
areas of potentially continuous habitation (Fig. 1c). Neiva et al. (2014) referred
to such areas as “climate refugia” and recognized this type of refugium for the
intertidal seaweed Pelvetia canaliculata between the Iberian Peninsula and Ireland.
Similar types of long-term stable populations across glacial–interglacial changes
may occur in long-lived reefs of the Indo-Pacific region (Vermeij 1986). In these
types of refugia, although the geographic area is maintained, many species will still
need to migrate upslope or downslope with changing sea level.
In general, marine reserves have not yet been established specifically as refugia
for species that will shift their ranges because of climate change effects on the
ocean environment. Range shift refugia in the form of marine reserves would have
to anticipate future demographics of species. For example, populations of tropical
species would expand away from the tropics in a warming ocean. Marine reserves
currently located in temperate zones and within potential expansion areas may serve
as future refugia for tropical species. Likewise, areas similar to the climate refugia
of Neiva et al. (2014) may also be adequate to accommodate range shifts, in a
location where the marine reserve would overlap the present-day poleward edge of
a tropical species’ range plus its potential future expansion into currently temperate
habitats. Furthermore, a warming global climate will be accompanied by sea level
rise, so species that currently inhabit shallow water (e.g., zooxanthellate corals,
rocky intertidal algae) will follow their preferred ranges as they move upslope, in
addition to any latitudinal shifts in distribution.
Habitat Shifts
Similar to range shifts, during a habitat shift, a taxon disappears from part or all of
its range, but instead of moving to a new geographic location and maintaining its
preferred habitat, a species invades a new habitat (with or without a range shift)
(Fig. 1d, e). Although not ideal, the new habitat is survivable compared to the
catastrophe in the species’ old habitat. Most commonly, habitat shifts in the marine
realm involve a change in water depth to find new, tolerable conditions.
Shifts to deeper water habitats are a common means of escaping adverse
conditions in shallow water and have been documented across geologic time.
Several Late Cambrian trilobite clades survived in a deeper water environment when
shallow-water habitats became lethal (Westrop 1989). During hypoxic events, the
goniatite Kokenia sp. likely survived in deeper marine habitats during the Early
Devonian (House 1996). Algeo et al. (2010) reported a deep water refugium for
radiolarians across the Late Permian mass extinction when the oxygen minimum
zone extended over continental shelves. A move to deep water and proximity to
upwelling has been proposed as a survival mechanism for Early Jurassic reefs
to escape global warming (Kiessling et al. 2009). Slope through abyssal habitats
were once thought to be the main havens for Antarctic shelf species during glacial
maxima (Barnes et al. 2006) although this idea is now in contention because of the
discovery of cryptic species complexes replacing what were previously thought to
be relict populations from prior refugia (Allcock and Strugnell 2012).
Throughout the Phanerozoic, shallow-water refugia were frequently exploited
during times of increased oxygen stress in deep water. Particularly relevant during
times of global ocean stratification, habitat shifts to shallow water became a critical
survival strategy. Some examples of the move to shallow-water refugia to escape
170 C. L. Schneider
adverse conditions in deeper environments include Ordovician brachiopods (Huang

et al. 2013), Devonian ostracods (Lethiers and Cassier 1999), and Cretaceous
hydrothermal vent faunas (Jacobs and Lindenberg 1998).
A special type of shallow-water refugium during the Early Triassic—the “habit-
able zone”—was proposed by Beatty et al. (2008, see also Zonneveld et al. 2010a,
b, Clapham et al. 2013). In this case, the most habitable area was capped near the
base of breaking fair-weather waves, where water was sufficiently oxygenated, and
where wave energy and turbidity were tolerable (Beatty et al. 2008). The base of the
habitable zone extended to the upper limits of anoxia, where storm wave influence
intermittently oxygenated the water. Wide shelves (Zonneveld et al. 2010a) and
embayments therein (Beatty et al. 2008) protected the habitable zone refugium from
the upwelling of deep, anoxic water.
Other types of habitat shifts include movement into radically new environments.
As environmental conditions deteriorate, those species that are more flexible in their
environmental tolerance, or are more physiologically plastic, can take advantage
of unexploited habitat. For instance, amphipods moved into brackish water glacial
lakes of the last glacial maximum along the European coast and from there expanded
into freshwater rivers (Krebes et al. 2010). Open marine benthos moved into a
frequently perturbed delta system in a humid region to escape fatal warming and
oxygen stress of Late Permian and Early Triassic oceans, as recorded in rocks from
this interval in France (Gall and Grauvogel-Stamm 2005).
Like range shifts, marine reserves do not currently target future habitat shifts of
species. Some species may have experienced habitat shifts in response to adverse
conditions (e.g., cold-water kelp survival near the bottom of their photic range in
the Galapagos, Graham et al. 2007) or already overlap a marginally acceptable
habitat within their present range, such as the potential for preferred survival of
mesophotic coral populations in a warming ocean that could disperse to shallow-
water populations (i.e., “deep reef refugia,” Bongaerts et al. 2010). These examples
may suggest the types of habitat shift refugia for future consideration.
Isolated Geographic Refugia
Contraction of ranges to smaller, isolated geographic areas is the most commonly

cited refugial strategy in both empirical and theoretical studies. In this type of
refugium, a taxon is restricted to one or more locations, each of which is much
smaller than the original range (Fig. 1f). During the last glacial maximum, many
spatially restricted refugia have been identified in all parts of the shallow ocean,
from peri-glacial refugia in the North Atlantic (e.g., Maggs et al. 2008), tropical
refugia in the Indo-Pacific (e.g., Vermeij 1986), and refugia in unglaciated areas
along the Antarctic shelf (e.g., Graham and Smith 2012). Geographic refugia
through ancient mass extinctions tend to be larger regions than those of the last
glacial maximum, possibly in part because of the coarse geographic and taxonomic
scale of many studies of the ancient past (e.g., regional survivorship of brachiopods
across the Late Ordovician mass extinction, Rassmussen and Harper 2011).
Isolated geographic refugia exist primarily because of the inherent heterogeneity

of nature. Global deterioration that causes a major crisis in the biosphere tends to be
variable across space and time. For instance, spatio-temporal variability in oxygen
stress of Permo-Triassic oceans resulted in some areas remaining habitable, while
others were fully anoxic (Algeo et al. 2010). Glacial drawdowns of sea level caused
major shifts in habitat, such as the reduction of rocky shorelines to a few areas
(e.g., North Atlantic refugia of areas along the coast of Ireland, the Hurd Deep, and
the Iberian coastline, among others; Maggs et al. 2008) and population isolation
resulting from barriers created by a drop in sea level (e.g., Sunda Shelf; Ludt and
Rocha 2015).
Isolated geographic refugia sometimes seem to arise from a “luck of the draw.”
When environmental conditions deteriorate within a species’ range, and if one or
more areas remain amenable within the range, then those subsets of the original
population have the chance of survival “when the world closes in” (Rickards and
Wright 2002, p. 3). For instance, Antarctic glaciers at the last glacial maximum
reached the shelf edge, save for small, sometimes ephemeral refugia on the outer
shelf (Barnes and Kulinski 2010). Organisms that happened to occupy these
isolated, unglaciated refugia survived.
Lucky survivors in refugia do not necessarily preclude other taxa from migrating
into these open spaces, but certainly competition for resources (particularly food
and space) in a geographically restricted refugium could be very tight. For example,
incumbent survivors have the advantage of being already established in, and
adapted to, the refugium, as interpreted for Patagonian molluscs in nearshore deltas
following the Cretaceous-Paleogene mass extinction (Aberhan and Kiessling 2014).
One contentious location of isolated geographic refugia is that of sea mounts.
Sea mounts, submerged volcanoes, and other conical islands may increase habitable
area in a deteriorating ocean system, particularly in protection from acidification
(Jacobs and Lindenberg 1998; Rowden et al. 2010; Tittensor et al. 2010), deep
water anoxia (Galil and Zibrowius 1998; Rowden et al. 2010), and global eustatic
regressions (Jablonski and Flessa 1986). Sea mounts may have provided refugia
to Jurassic corals (Stanley and Beauvais 1994) although corals may have survived
better elsewhere (Kiessling et al. 2009). Based on ostracods across the Paleocene-
Eocene Thermal Maximum, Yamaguchi and Norris (2015) found that extinction
could be greater and recovery longer on these geographically isolated areas than
on continents, mainly because of the lengthy “waiting time” for dispersal from
distal refugia. Jablonski and Flessa (1986) pointed out that sea mounts and newly
exposed islands could act as refugia when sea level fell below continental shelf
edges. Conversely, because of the rising sea level predicted for the future, drowned
islands, sea mounts, and volcanoes may decrease the availability of these types of
isolated geographic refugia.
Most marine reserves are isolated geographic refugia. They are spatially, and
sometimes depth, defined, and the species therein acquire partial to full protection
from anthropogenic activities. Two recently expanded or established, large marine
protected areas include the US Papahānaumokuākea Marine National Monument,
which is over 935,000 km2 of habitat protected from commercial fishing (Obama
172 C. L. Schneider
2014; The White House 2016) and the 1.55 million km2 Ross Sea Region Marine
Protected Area, which includes no-take and restricted use areas (U.S. Department
of State 2016). The smallest reserves are under 1 km2 and have been shown to have
positive effects on fishery stocks (Roberts and Hawkins 1997).
Life History Refugia
Life history refugia refer to those that are necessary only during one portion of a
species’ life cycle, usually during breeding or for nursery habitats. These types of
refugia are mainly used by mobile organisms that can move into the refugium during
the necessary interval, and then leave when the habitat is no longer needed. In the
fossil record, particularly in demographic reconstructions of last glacial maximum
refugia, these are a subset of isolated geographic refugia, in that they are locales that
are briefly used, either once (as nurseries) or intermittently (as rookeries).
Breeding refugia are critical for marine animals that need to rear their young on
land. Seals, sea lions, marine birds, and sea turtles used breeding refugia through the
last glacial maximum (Harlin-Cognato et al. 2006; Pinsky et al. 2010; Naro-Maciel
et al. 2014; Younger et al. 2016). Red king crabs, which are a deep water taxon little
affected by last glacial maximum climate and sea level changes during their adult
stage, needed shallow-water nursery refugia for juveniles, some which were located
near a few unglaciated coastal islands of North America and along northeastern Asia
(Grant and Cheng 2012).
A shift to a refugium may also result from intolerance of extra-refugial conditions
during a life stage. For coccolithophores across the Paleocene-Eocene Thermal
Maximum, fatal conditions in the tropics caused them to shift to temperate waters,
possibly to avoid stress during one life stage (Gibbs et al. 2016).
Other species may survive in an environmentally poor refugium during a life
stage, particularly as spores, propagules, or seeds that could remain in stasis for
some period of time (Ashcroft 2010). A dinoflagellate taxon is currently surviving
as cysts in the refugium of the warm Indo-Pacific, awaiting an increase in sea surface
temperatures greater than that of a normal interglacial period (Mertens et al. 2014).
Many marine protected areas dually serve as isolated geographic and life history
refugia. For example, several reserves in New Zealand are key breeding habitats for
marine mammals and birds and thus, serve as breeding refugia, such as the Pohatu
and Auckland Islands, Motu Maha Marine Reserves (New Zealand Department of
Conservation 2016). Protection of breeding habitat is critical for migratory marine
animals, such as whales. Along some of the Hawaiian coastlines, the Hawaiian
Islands Humpback Whale National Marine Sanctuary was established to protect the
overwintering habitat and breeding of humpback whales (National Ocean Service
2016). Likewise, many of the reserves along the west coast of Mexico and southern
California also protect overwintering and breeding whales, but also protect other
migratory and far-ranging species (Commission for Environmental Cooperation
2011).
Life history refugia in marine reserves often result from protecting the entire
ecosystem. For example, marine reserves along the California coast host high
populations of the red sea urchin, under whose spines juvenile red abalones find
protection from predators (Rogers-Bennett and Pearse 2001). Similarly, a marine
reserve for king scallops became more effective with the increase in algal cover,
which increased habitat complexity and served as a nursery for juvenile scallops
(Howarth et al. 2015).
Cryptic Refugia
Cryptic refugia are populations that comprise rare and sparsely scattered individuals
that are still able to breed (Stewart and Lister 2001) or are a subset of isolated
geographic refugia in which micro-refugial populations that are not reproductively
viable unto themselves are connected as meta-populations (Bennett and Provan
2008) (Fig. 1g). These types of refugia are most commonly defined for terrestrial
species (Stewart and Lister 2001; Bennett and Provan 2008), but have applicability
to marine refugia. Cryptic refugia are likely difficult to locate in the fossil record
(Gavin et al. 2014).
A possible special case of cryptic refugia in the fossil record is that of disaster
taxa. Disaster taxa are those that proliferate during and immediately following mass
extinctions; where other taxa are in great distress, these taxa thrive (Schubert and
Bottjer 1995; Kauffman and Harries 1996). With post-extinction environmental
recovery, these taxa are again suppressed into rarity within ecosystems or find a
refugium in high-stress habitats (Schubert and Bottjer 1995; Kauffman and Harries
1996). Thus, “classic” disaster taxa—such as the brachiopod Lingularia (Schubert
and Bottjer 1995; Rodland and Bottjer 2001; although see a re-assessment of Early
Triassic Lingularia, Zonneveld et al. 2010a, b)—existed in cryptic refugia and other
refugia, awaiting times of global environmental catastrophe to once again flourish.
A marine reserve in which one or more species cannot successfully maintain its
population is a cryptic refugium for that species. The importance of networks that
connect multiple reserves cannot be understated, as most marine reserves are too
small to maintain viable population levels and genetic diversity of many species
therein. Connectivity inherently occurs outside of reserves, because propagules,
larvae, juveniles, and migrants must pass from one reserve to another, which
can include passing through openly exploited areas (Steneck et al. 2009). While
within-reserve benefits increase with the size of the protected area, connectivity
between reserves can be improved by maximizing the number of habitats captured
by multiple reserve areas and by spacing reserves at appropriate distances for
population exchange of target species, which is often over tens to hundreds of
kilometers (Almany et al. 2009). Additionally, a network of marine reserves is
insurance against the collapse and loss of one or two protected habitats, which
Almany et al. (2009) referred to as “risk spreading.”
Another type of potential cryptic refugia is the “artificial marine micro-reserve,”
defined by Garcia-Gomez et al. (2011, 2014). These are small, commercial, and
174 C. L. Schneider
urban infrastructures that provide areas of colonization. An example they provided

is that of the ferruginous limpet, Patella ferruginea that is the most endangered
invertebrate of the Mediterranean. Although it is rare in most locations, it is
highly abundant on man-made structures in the Strait of Gibraltar. Similarly, Inger
et al. (2009) suggested that, with careful planning to negate harmful environmental
impacts, renewable energy infrastructures built in the ocean may provide “de-facto”
marine reserves by providing localized harvest refugia.
Harvest Refugia
Essentially havens from anthropogenic fishing pressures, harvest refugia protect

some populations of exploited taxa from potential extinction caused by overfishing.
In particular, these refugia commonly shelter source populations that disperse to
fished populations as well as maintain the refugial population (e.g., an increase in
Anadara sp. clam populations outside of the marine reserve; Gell and Roberts 2003)
or serve as insurance against fatal depletion of harvested taxa (e.g., increase in reef
fish sizes and populations in a no-take marine reserve near South Africa; Maggs
et al. 2013). Although harvest refugia have been studied only on the order of years to
decades so far, these refugia have considerable importance for the long-term future
of fisheries and harvested species.
Before marine reserves, areas that were inaccessible to fishermen remained
unexploited and thus, were natural harvest refugia that supported fished populations
(Johannes 1978; Gell and Roberts 2003). However, with constant and exponential
improvement in resource-gathering technology, naturally occurring harvest refugia
are becoming rarer, as areas previously difficult to access become accessible (Gell
and Roberts 2003). So, artificially created harvest reserves in the form of marine
reserves take on the huge role of supporting populations both within and outside of
reserves (Baskett and Barnett 2016).
Most marine reserves are active harvest refugia in that they protect a population
of one or more exploited species. Inherent to the harvest refugium of the marine
reserve is the requirement not only to maintain intra-reserve population levels, but
also to support extra-reserve fisheries and other needs of the local habitat. For exam-
ple, the San Juan and Upright Channels Sea Cucumber and Sea Urchin Commercial
Harvest Exclusion Zones ensure that a dense population of sea urchins, including
large individuals, is present to enhance recruitment and breeding, both within
the reserve boundaries and in populations outside of the reserves (Washington
Department of Fish and Wildlife 2016). Because sea urchins have an impact on kelp
forests and sea otter populations, as well as being a commercial resource, population
maintenance became a necessary focus in this region. Likewise, many marine
reserves in New Zealand are critical harvest refugia, in which species disperse
to extra-refugial areas and sometimes into baited traps and fishing lines outside
the boundaries of the reserves, thereby supporting local fisheries and recreational
fishing (Ballantine 2014).
Studies have repeatedly shown that with age, the marine reserve becomes a
more efficient harvest refugium. In the marine reserves of the Philippines, those
that had been in place for 10 years are more maintain high yields for harvested
fish species which had positive spillover effects (i.e., recruitment of fish into the
extra-reserve habitat). A similar study in marine reserves of Europe also found
a correlation between the increase of biomass and abundance in marine reserves
that were older than 10 years, among other factors (Edgar et al. 2014). Benefits
of restocking harvested species outside of refugia can occur in as little as 6 years,
but timing varies with trophic level of the target fish (da Silva et al. 2015). Coleman
et al. (2015) found that changes in targeted species’ abundance and diversity took 5–
20 years to manifest, and indirect effects on ecosystems took even longer. Thus, time
and continuous stability of protection for target species are important for creating
harvest refugia.
The Environmental Conditions of a Refugium Are Sufficiently

Habitable Such That the Species’ Population Remains Viable
During Its Time in the Refugium
Note that this portion of the definition of refugium contains the qualification
“sufficiently habitable”; conditions in the refugium need not be pristine, but merely
sufficient so that the population can reproduce and recruit. This means that an ideal
refugium must have a reduced, yet bearable, level of stress within its boundaries and
be isolated from the fatal conditions of the extra-refugial environment. Environmen-
tal stress can be countered by beneficial factors typical of local habitats (e.g., large
amplitude internal waves providing intermittent cooling of reefs in otherwise fatally
warm temperatures; Buerger et al. 2015; Wall et al. 2015; Schmidt et al. 2016), or
by the addition of one or more ameliorating factors that boost resistance to stress
(e.g., increased nutrients; Vermeij 1986; Graham et al. 2007).
Many ancient refugia were not stress free. The Early Triassic habitable zone and
similar environments were stressed by low oxygen and high temperatures in addition
to turbidity and shifting sediment (Beatty et al. 2008; Zonneveld et al. 2010a;
Clapham et al. 2013; Chen et al. 2015; Song et al. 2015; Zhao et al. 2015). Also
across the Late Permian mass extinction, ostracods and other taxa had sufficient
oxygen and food within microbial mats, which helped them to survive heat stress in
the refugial habitat (Forel et al. 2013; Posenato 2009). The Triassic deltaic cryptic
refugia preserved in France lacked the lethal anoxia of the open ocean, but were
frequently perturbed by shifting sediment and other stresses inherent to deltaic
environments (Gall and Grauvogel-Stamm 2005).
The populations of a species existing on the margin of its range distribution,
which usually inhabit more stressful conditions than the main demographic range,
may be essentially pre-adapted to life within a refugium (Kauffman and Harries
1996; Rickards and Wright 2002). Refugial habitats are rarely, if ever, matched to
176 C. L. Schneider
species’ optimal environmental preferences. Thus, individuals on the margins of

a geographic range already tolerate less than optimal conditions, and thus might
survive better within refugia that are somewhat stressed.
Likewise, marine reserves are not necessarily stress free. In a changing ocean,
conditions that are not particularly stressful for some species can preclude others
that would otherwise be present (Ashcroft, 2010). Thus, if marine reserves, or
reserve networks, capture a range of habitats, more species are likely be protected
(Ballantine and Langlois 2008; Almany et al. 2009; Ballantine 2014). Alternatively,
marine reserves established within the climate refugia of Neiva et al. (2014)
may protect species that have already adapted to, and persisted through, glacial–
interglacial climatic and eustatic fluctuations, and thus may be tolerant to major
climate change (Gavin et al. 2014).
In fact, the presence of non-lethal levels of stress in the habitat may be
beneficial for some species. Organisms that are adapted to climatically stable
habitats are likely more susceptible to perturbations (Ashcroft 2010), but those
that are exposed to survivable levels of stress (e.g., zooxanthellate corals and sub-
lethally but potentially bleaching warm water masses) and are given sufficient time
to recover may adapt to stressful conditions, and thus be positive contributors to
species conservation (Chollett et al. 2014). Similarly, species or ecosystems that
display resilience and a rapid recovery between perturbations would be appropriate
locations for the refugia of marine reserves (Berger et al. 2011).
A Species’ Population Is Smaller in the Refugium Than Its

Pre-environmental Perturbation Size
A reduced population is rarely an explicit component of any definition of refugium,

but usually is assumed or inferred. While in the reduced geographic area of refugia,
or in the less optimal habitat such as a shift to deeper or shallower environments,
resources and space may be limited or more difficult to acquire. Refugial conditions
therefore cause populations to adjust for the limiting conditions of the refugium.
Explicit data on population size in fossil communities is nearly impossible,
but studies of demographics through the last glacial maximum show that many
populations of intertidal species went through declines. Available habitats were
limited by a series of factors: few peri-glacial ice-free zones (e.g., peri-glacial
refugia in the North Atlantic; Jolly et al. 2006; Maggs et al. 2008; Almada et al.
2012; Krakau et al. 2012), isolated rocky shorelines at the glacio-eustatic lowstand
(e.g., Hurd Deep in the English Channel; Provan et al. 2005; Hoarau et al. 2007;
Campo et al. 2010; Doellman et al. 2011; Li et al. 2015), and sufficient nurseries
for juveniles (e.g., islands along western North America; Harlin-Cognato et al.
2006; Pinsky et al. 2010; Grant and Cheng 2012) are three common constraints
for last glacial maximum intertidal refugia. Within the limited areas of North
Atlantic refugia, populations were constrained to be smaller than their pre-refugial

counterparts.
Likewise, with a drop in sea level, shelf area is reduced and species are
squeezed into much smaller areas. Although no data exist for population sizes, the
constrained areas of refugia around the Antarctic continent would suggest that taxa
sequestered in glacial refugia had smaller populations. Similarly, with the glaciation
of coastlines, rookery space became constrained; Younger et al. (2016) suggested
that the populations of some sea mammals and birds were reduced to bottlenecks of
a few thousand breeding females.
A possible alternative to reducing population size is to reduce the size of
individuals, or the Lilliput effect (Urbanek 1993; also see Harries and Knorr 2009
for alternate explanations). When resources are decreased and other environmental
factors are less than favorable, a reduction in body size compensates for the
limitations of the habitat. Thus, a species can maintain a viable population via
an overall reduction of body size, if conditions in the refugium cannot otherwise
support the species. A body size reduction requires sufficient time in the refugium
for the species to adapt in this manner to limited resources. For example, Early
Triassic trace fossils were smaller than their pre-extinction and post-recovery
counterparts (Beatty et al. 2008; Knaust 2010; Zonneveld et al. 2010a). Although
these traces are not body fossils, the smaller size of the traces suggests the Lilliput
effect on the animal creating the trace.
As mentioned above for harvest refugia, target species rebound in abundance and
biomass with time after the establishment of a marine reserve, especially in areas
given high levels of protection or that are designated as no-take. However, recovered
species in marine reserves must not only maintain the population inside the reserve,
they must also recruit and contribute to the exploited or disturbed population outside
of the reserve boundaries.
The Species Remains in the Refugium for Many Generations
Taxa inhabited refugia through the durations of global crises, which are highly
variable in duration, depending on magnitude and type of crisis. Time spans range
from a few thousand years across the last glacial maximum to several million years
during prolonged anoxia and warm climate of the Early Triassic. Thus, for any given
species, a multitude of generations passed inside the haven of the refugium. The
longer a species inhabited in a refugium, the greater the likelihood of adaptation
to the unique habitat of the refugium. With sufficient time, a species can evolve
and even diversify within the refugium (Harries et al. 1996). Likewise, time spent
in a refugium can lead to evolutionary consequences such as genetic drift, founder
effects, and population bottlenecks.
Population bottlenecks in refugia are best understood from the last glacial
maximum because of the ability to reconstruct demographic properties (Bennett
and Provan 2008; Stewart et al. 2010; Allcock and Strugnell 2012). Bottlenecks
178 C. L. Schneider
caused founder effects in refugial populations, the results of which are still seen
in molecular studies of phylogeography across the last glacial maximum (Ludt
and Rocha 2015). For instance, extreme reductions in population size and genetic
diversity have been reconstructed from molecular data for the intertidal rockweed
Palmaria palmata along Atlantic Canada (Li et al. 2015). In another example,
Antarctic shelf populations of many taxa were reduced to small populations that
caused a decline in genetic diversity because of founder effects (Allcock and
Strugnell 2012; Gonzalez-Wevar et al. 2013). Likewise, populations of various
penguin species were diminished at the last glacial maximum to only 2000 breeding
females, based on fossil and phylogeographic data (Younger et al. 2016).
Whether or not genetic bottlenecks occurred, time spent in isolated refugia across
the last glacial maximum often resulted in genetic drift. Even now, endemic mito-
chondrial and nuclear haplotypes remain unique to refugia-proximal populations,
regardless of mixing of refugial populations from the last glacial maximum through
secondary contact during interglacial expansion (e.g., review and re-analysis of eight
North Atlantic intertidal taxa; Maggs et al. 2008). Potential examples are numerous,
and refugial-proximal endemic populations have been discovered worldwide. One
area where the discovery of endemic haplotypes helped to reveal a specific refugium
during the last glacial maximum is that of the Hurd Deep with its rock walls and
deep marine lake, where populations of everything from intertidal seaweeds (Provan
et al. 2005; Hoarau et al. 2007; Maggs et al. 2008; Li et al. 2015; Assis et al. 2016)
to rays (Chevolot et al. 2006) survived in isolation.
Longer time spans spent in the unique conditions of a refugium can result in
allopatric speciation (Vermeij 1986; Ludt and Rocha 2015). Theoretically, adaptive
radiations can begin in a refugium, where species have time to originate and
specialize in the tight, competitive confines (Schubert and Bottjer 1995). However,
species can adapt to the specific environmental constraints of a refugium that is
not necessarily free of stress, just more survivable than the outside world. Harries
et al. (1996) called these evolved populations “endemic refugia species” and pointed
out that these species might take longer to expand into their previous ranges, if
at all. Species that essentially become evolutionarily “stuck” in refugia can also
become relict species, which are those that do not re-expand or otherwise contribute
significantly to post-crisis recovery.
Graptolites survival across Paleozoic mass extinctions provides two examples
of in-refugial diversification, one contributing to post-crisis recovery, and the other
becoming a short-lived relict fauna. In the recovery example, graptolite survivors
of the Late Ordovician extinction in South China diversified within a refugium that
was recognized as being important to post-extinction expansion (Xu et al. 2005).
In the Late Silurian, a series of trough-like refugia in Australia and led to relict
graptolite taxa that never contributed to global recovery and ultimately went extinct
(Rickards and Wright 2002). In both cases, the reason for survival in refugia was
identified—isolation from poor extra-refugial conditions and favorable conditions
within the refugium—but the reasons for post-refugial proliferation or demise were
not identified.
Evolutionary stasis can also occur in a refugium, at least over a relatively

short interval. This was seen in one species of intertidal gastropod, Acanthinucella
spirata, along the west coast of North America, in which substantial morphological
change occurred during expansion out of the refugium rather than in the refugium
itself. Although genetic diversity decreased as the species spread northward along
the coast, the gastropod evolved a shell shape distinct from southern populations
and fossil conspecifics (Hellberg et al. 2001).
As seen in the examples provided for harvest refugia, research in marine reserves
emphasize the increasing positive impacts on protected species, diversity, and
ecosystems with age of the reserve. Given the decadal scales for species and
ecosystems to respond to protection (Ballantine 2014; Coleman et al. 2015), the
effects of newly established marine reserves can only be predicted and anticipated.
After the Environmental Crisis Ends, the Species Recovers

by Inhabiting Newly Re-opened Habitats, Either Through
Population Expansion or Through Adaptive Radiation;
Otherwise, the Refugium Became a Trap
The end product of a refugium is not only survival through a global crisis, but
ultimately a contribution to global ecological recovery. After survival through
a global crisis, refugial populations should ideally expand to repopulate newly
available habitat (Holterhoff 1996; Bennett and Provan 2008; Ashcroft 2010). Thus,
a post-refugium taxon has a wider geographic distribution and larger population.
Vermeij (1986) stressed the importance of understanding the full range of a species
before the onset of a global crisis; equally, it is critical to understand the pattern of
re-expansion—or not—after the environmental crisis has ended.
A taxon can re-expand into its former range, expand to greater than its former
range, or can take over an entirely new habitat. How the recovery occurs after time
spent in a refugium depends in part on the degree of adaptation that occurred during
its residence and in part on extra-refugial conditions—environmental and biotic—
after the global crisis has ended. In ancient refugia through mass extinction events,
time spent in refugia can be on the order of millions of years, allowing plenty of
time for evolution and diversification (Harries et al. 1996); thus, the taxon or taxa
exiting the refugium can be quite different from the ones that were seeking a safe
haven.
Thus, conditions within the refugium must be sufficient that a taxon can be
ready to expand after a crisis has ended (Vermeij 1986), plus conditions in the
recovered environment must be amenable to species expansion. A species, or a
population, that cannot recover from its time in a refugium cannot contribute to
global rediversification; thus, it becomes a relict species or relict population (or
“stranded population” of Kauffman and Harries 1996) and the habitat was not a
refugium, but a refugial trap. And, if a taxon becomes extinct while in a purported
180 C. L. Schneider
refugium, even though that taxon might thrive briefly during a global crisis, then
that place was also a refugial trap.
Refugial traps are known from geologic time. Over million-year time spans,
refugial traps initially appear to be adequate, but eventually, refugial conditions
decline and the taxa within either become relicts or cease to exist. For example,
the Carboniferous Tindouf Basin appeared to be a prime refugium for taxa across
the Serpukhovian extinction event: isolation from oceanic currents and tectonic
adjustments to the basin maintained a habitable water depth (Cozar et al. 2014).
However, eventually tectonism led to uplift and exposure, and the basin refugium
(and populations therein) no longer existed (Cozar et al. 2014).
In another refugial trap, brachiopod taxa that survived the Late Permian mass
extinction persisted into the Early Triassic within habitable zone refugia, but shifting
anoxic and fatally thermal water masses eventually terminated the refugium and
its inhabitants (Clapham et al. 2013). In this case, sudden and lethal environmental
perturbations eliminated refugial habitats, causing the extinction of brachiopod taxa,
a problem that Clapham et al. (2013) warn is a caution for modern conservation
efforts.
Relict populations exist as remnants of more recent refugia. A population of
rockweed is trapped on the northwest Iberian Peninsula by oceanographic barriers
to the north and south and is threatened by frequent fatal perturbations of heat stress
(Hoarau et al. 2007). Other relict intertidal seaweed populations are trapped by
winter sea ice and temperatures along the Canadian Atlantic coast (Li et al. 2015).
A thriving, but relict, pre-Pleistocene gastropod fauna remains in its refugium in
the Gulf of Venezuela, isolated from the rest of the Caribbean by upwelling (Petuch
1981).
Sometimes the reason for the development of a trap from a seemingly good
refugium is unknown. The best example is that of the Late Silurian (Wenlock-
Ludlow) graptolites that survived in a trough-and-rise system currently in rocks of
New South Wales through the mass extinction at the end of the Wenlock (Rickards
and Wright 2002). The taxa became a relict fauna, never contributing to the post-
extinction recovery (Rickards and Wright 2002). The refugium seemed an adequate
habitat, isolated from less habitable conditions in the rest of the ocean by local
tectonism, but the reason for the refugium ultimately becoming a trap remains to be
known.
However, there is one caveat; relict species may not indeed be “dead-end”
taxa, but may remain in refugia, awaiting an environmental change to provide
an appropriate habitat. For instance, disaster taxa presumably inhabit refugia in
suppressed populations during “normal” environmental conditions (Kauffman and
Harries 1996). Polar taxa withdraw to refugia during interglacial periods, in sync
with the expansion of temperate and tropical taxa from glacial refugia (Ashcroft
2010; Stewart et al. 2010; Almada et al. 2012). The dinoflagellate Dapsilidinium
pastielsii is extinct globally except in a warm area of the Indo-Pacific, presumably
its current refugium, awaiting warmer sea surface temperatures than Pleistocene
interglacials allow (Mertens et al. 2014). Thus, before defining a refugium, it is
important to be certain of how a species responded, or will respond, when the need
for the refugium has passed.
Currently, all marine reserves are in use as necessary refugia. Nowhere have
conditions ameliorated sufficiently that a reserve is no longer necessary, and so the
crises that caused the need for marine reserves have not yet ended.
3 Identifying Ancient Refugia
Refugia in deep time are, at best, difficult to identify. More material is not preserved,
eroded, or undiscovered than exists in fossil collections. With a spatially, temporally,
and preservationally heterogeneous fossil record—combined with highly variable
sampling worldwide—identifying undisputable refugia is a challenge. Locating a
refugium also depends in part on its geographic extent which influences the chances
of encountering the location; in part on the geographic scale of the study; and in
part on the quality of preservation of the stratigraphic interval and the fossils within
(Westermann 2001).
Many reports of refugia include paleoenvironmental information to support
the interpretation of the existence of a refugium. Although ultimately a refugial
trap, Cozar et al. (2014) recognized the survival of a reef community through the
Serpukhovian-Bashkirian extinction in the Tindouf Basin of Morocco and Algeria,
not only through the persistence of fauna that went extinct elsewhere, but also
through isolation caused by ocean currents and regional tectonism. Paleoenviron-
mental and paleogeographic analysis of the habitable zone of Western Canada
in the early Triassic explained not only the survival, but also the diversity of
shallow-water ichnofauna (Beatty et al. 2008; Zonneveld et al. 2010a, b). In more
recent time, interpretations of refugia based on molecular reconstruction of species’
demographics through the last glacial maximum are often upheld by independent
data that indicate deglaciated areas (e.g., potential Antarctic refugia revealed by
geophysical analysis of the sea floor; Graham and Smith 2012).
Tracing the space–time relationships of a lineage through a global crisis would
be ideal in identifying a refugium. Unfortunately, such data are very rare; because
of sampling or stratigraphic absences, lineages usually disappear from the known
rock record through a mass extinction, either to their ultimate demise or to reappear
later. The terms “Lazarus effect” and “Lazarus taxa” were coined by Flessa and
Jablonski (1983) and further defined by Jablonski (1986) for those taxa that were
present before a mass extinction event, then disappeared through the mass extinction
interval, to reappear intact later in the fossil record. Stanley and Beauvais (1994)
appropriately quip that Lazarus taxa “mysteriously leapfrog over large intervals of
geologic time.” Lazarus taxa have been described as an extreme form of extirpation,
in which the refugia are unknown (Westermann 2001).
Thus, Lazarus taxa and their lengthy disappearance from the rock record indicate
the likely presence, somewhere in the world, of undiscovered or unpreserved refugia
(Vermeij 1986; Donovan 1989; Erwin and Hua-Zhang 1996). The duration of the
182 C. L. Schneider
“Lazarus” phase of the taxon (while it is absent from the rock record) corresponds
to the length of time of the global environmental crisis (Wignall and Benton 1999).
Locating a refugium for a Lazarus taxon is problematic because of decreased
area covered by the taxon during demographic contractions compounded by the
inherently incomplete nature of the fossil record. Rarely are refugia for Lazarus
taxa discovered, such as the Late Ordovician refugium for graptolites located in
present-day south China (Xu et al. 2005).
However, the refugial survivorship of Lazarus taxa is in contention (see discus-
sion in Fara 2001); they also could artificially result from their overly theorized
existence in evolutionary, phylogenetic, or taxonomic studies (Westermann 2001)
or simply be overlooked during sampling because of rarity or size (Erwin and Hua-
Zhang 1996; Hautmann and Nützel 2005; Hautmann et al. 2011). Lazarus taxa may
also have reduced to population sizes below the preserved resolution of the fossil
record (Erwin and Hua-Zhang 1996), such that they would occur as “cryptic refugia”
species, those that survive as sparsely distributed and rare individuals across a
landscape (Stewart and Lister 2001; Bennett and Provan 2008). The possibility that
Lazarus taxa are simply unsampled led Payne (2005) to suggest that they are in
fact “Waldo taxa”; that they are present all along, but easily overlooked and require
more effort in searching, similar to finding the character Waldo among a cartoon
menagerie of people and items.
Aside from the debate of whether Lazarus taxa are actual and represent unfound
refugia, there are currently three methods for identifying ancient refugia: analysis
of fossil data, phylogeographic reconstructions, and species distribution models.
A more thorough review, from the perspective of refugia during the last glacial
maximum, was provided by Gavin et al. (2014), but the methods are briefly revisited
here:
Fossil Data
The presence of abundant fossils and diverse taxa in a particular area during a global
crisis interval might indicate the presence of a refugium. As mentioned earlier,
independent paleoenvironmental analysis would provide further support for the
existence of a refugium. Furthermore, if an appropriate refugial paleoenvironment
was present, it should harbor several taxa, not just one (Vermeij 1986).
The benefit of interpreting refugia from fossil and rock data is the direct exami-
nation of the actual location and properties of the refuge, relative to conditions in the
rest of the world. Why taxa survived in this location, and under what constraints—
ecological, demographic, and environmental—are directly identifiable. Also, fossil
data can provide information about ecosystem response, not just surviving taxa,
to survivorship in refugia. For instance, Holterhoff (1996) identified one crinoid
biofacies that migrated as an intact community through Pennsylvanian-Permian
glacio-eustatic fluctuations in the Anadarko Basin of west Texas, but other refugial
faunas during sea level lowstands were mixes of taxa from different highstand
communities.
Fossil data across the last glacial maximum can also be integrated with molecular
analyses and radiocarbon dating for deeper understanding of refugial existence.
Only fossils can confirm that a specific region is a refugium; molecular studies can
only estimate the location of a nearby refugium (Ludt and Rocha 2015). Fossils
confirmed the refugial location of Nucella lima gastropods along the northeast
Pacific coast during glacial maxima (Cox et al. 2014). Subfossils themselves
not only indicate refugial locations, but also can provide the molecular data for
phylogeographic studies (e.g., Pinsky et al. 2010).
The obvious caveat of fossil data is the incomplete nature of the fossil record.
Preservation of organisms is dependent on the habitat preferences, life mode,
taphonomy, and discovery of individual specimens. Transport and reworking of
skeletal material can also affect the interpretation of a refugium by taking specimens
out of a refugium or concentrating specimens into a high diversity, high abundance
assemblage. Of course, the usual taphonomic analyses—comparative taphonomy,
abrasion, etc.—can identify potential preservational issues.
One problem with the study of fossils is the resolution of paleontological
research. Spatial, stratigraphic, and taxonomic resolution can be coarse (e.g., provin-
cial, stage, and generic levels), which typically averages out high-resolution data
like exact refugial boundaries and detailed information about population history.
However, the coarse resolution of ancient refugial studies has one advantage: to
observe the “big picture” reasons behind the formation and maintenance of refugia,
like tectonism and oceanographic changes (e.g., refugia of South China carbonate
platforms across the Late Ordovician mass extinction; Huang et al. 2013; Xu et al.
2005). Likewise, the study of long-term refugia can trace the responses of taxa
during their isolation, whether refugial evolution and diversification or in becoming
relict faunas.
Phylogeographic Studies
Constrained mostly to interpretations based on molecular data of refugia of the

recent geological past, phylogeographic studies identify ancient demographic pat-
terns, particularly population constrictions that can help identify refugial remnants.
Most studies are based on mitochondrial or chloroplast DNA, but recent advances
in molecular analysis has allowed the inclusion of nuclear DNA in demographic
reconstructions.
Whichever method is used, the patterns of diversity and endemism interpreted
from genetic data enable the identification of demographic contractions that are
followed by later expansion and re-mixing. Populations that are remnants of
previous refugia usually have high diversity, but also frequently have endemic or
“private” haplotypes (Bennett and Provan 2008). Populations that have re-expanded
out of refugia that existed during the last glacial maximum are lower in haplotype
diversity and lack genetic endemism (Bennett and Provan 2008). Secondary contact,
the mixing of populations from two or more refugia, results in high genetic diversity,
184 C. L. Schneider
but like unmixed expansion populations, lack the private haplotypes of refugial
populations (Bennett and Provan 2008).
The benefits of phylogeographic studies are the identification of historical
migration patterns, population sizes, and the effects of natural selection (Bennett and
Provan 2008; Gavin et al. 2014). Comparative studies across taxa can identify shared
refugia during the last glacial maximum (Maggs et al. 2008; Gavin et al. 2014).
Furthermore, integration of molecular clock interpretations can help determine
timing of contraction into refugia and expansion into secondary contact zones
(Gavin et al. 2014).
Phylogeographic analyses are limited in that they cannot identify the actual
location of refugia (Bennett and Provan 2008; Ludt and Rocha 2015); these areas
have long since been drowned by rising sea level, and the remnant population
bearing the refugial molecular signal has accordingly adjusted its habitat. However,
with the comparison to independent environmental data—such as which coastal
areas that remained unglaciated (e.g., unglaciated coastal areas and islands of
Antarctica; Allcock and Strugnell 2012; Gonzalez-Wevar et al. 2013; Chown et al.
2015) or the paleo-presence of adequate rocky shorelines for encrusting biota (e.g.,
Hurd Deep in the English Channel; Provan et al. 2005; Hoarau et al. 2007; Maggs
et al. 2008; Doellman et al. 2011)—it is possible to interpret the proximity of
modern populations to their refugia during the last glacial maximum.
Species Distribution Models
This theoretical method involves the comparative analysis of the factors that deter-
mine the range of a species with environmental data or models, in order to predict its
occurrences through space and time. Strengths of using species distribution models
are numerous: they can be used to reconstruct past (e.g., intertidal species and
environments through the last glacial maximum; Waltari and Hickerson 2013) and
identify potential future (e.g., the three possible future refugia of Japanese reefs;
Makino et al. 2015) refugia with respect to climate change. Thus, they create testable
hypotheses of the existence of past refugia and of possible key conservation areas
(Maggs et al. 2008; Gavin et al. 2014). Additionally, species distribution models
are relatively inexpensive compared to molecular analyses and field-intensive fossil
studies.
A major disadvantage of species distribution models is that results are only as
good as the information input into the model. Climate change projections often are
too coarse to have meaningful applications and produce results that are too broad for
accurately identifying possible past or future refugia (Chollett and Mumby 2013;
Gavin et al. 2014; although see van Hooidonk et al. 2015). Furthermore, models
assume that species–environment or species–climate relationships are static and
predicable through time; in other words, species distribution models rely on the
assumption of niche stability of the species and do not accommodate population-
scale processes (Gavin et al. 2014).
Another problem with species distribution models, particularly in identifying

potential future refugia, is the choice of climate model(s) used in the analysis.
Makino et al. (2015) pointed out that, out of multiple climate models available,
each provides a different result, and caution that the results based on the input of
various climate models do not produce a “no regrets” overlapping area for an ideal
refugium that covers all projections.
4 Lessons from the Past for Identifying Future Refugia
Some of the environmental crises facing the marine biosphere today are similar to
those of the ancient past: warming sea surface temperatures, ocean acidification,
sea level rise, and related effects are some of the stresses previously experienced
by the marine biosphere. Presently, these “old hat” stressors are compounded by
anthropogenic-related problems; pressures like overharvesting effects on trophic
systems and ongoing pollution have never previously been experienced on modern-
day scales. The outcome of this unique set of old and new stressors acting on modern
marine ecosystems is difficult to quantify, let alone predict. However, the need for
viable refugia is clear if global biodiversity is to be maintained.
Worldwide, marine reserves improve short-term survival during environmental
perturbations and have the potential to become long-term refugia through a changing
ocean system. In the short term, marine reserves have the potential to buffer
organisms from the negative impacts of environmental perturbations (Baskett and
Barnett 2016). Over longer time intervals, marine reserves can act as refugia, and
thus have an inherent conservation value by giving species and ecosystems the
ability to persist through climate change and other global crises (Ashcroft 2010).
Thus, conservation biologists have a heavy task set for them in managing marine
reserves; in addition to protection from immediate stressors, there exists a need to
identify where, how, and why future refugia will exist in a changing world.
Refugia from the cold sea surface temperatures and lower sea level of the last
glacial maximum are not directly applicable to modern environmental crises, even
though studies of population contractions into refugia are most numerous for this
period in geologic history. However, past survival through glacial and interglacial
periods has proven the resilience of species, which may mean that taxa are more
adaptable to climate change than given credit (Gavin et al. 2014), especially if
provided with adequate refugia and sufficient time to adapt.
The rock record before recent glacial–interglacial fluctuations has multiple rele-
vant examples of how Earth has fared during global warming, ocean acidification,
sea level fluctuations, and biosphere catastrophes. Although these studies are not
as numerous as those across the last glacial maximum, they provide examples of
how life survived mass extinctions and other global catastrophes. More importantly,
some properties of refugia recur in space and time, regardless of the type of
environmental stressors, the taxa that find protection, or the scale of the study.
The benefit of studying refugia in the ancient past is that the entire story of a
refugium can be told: the contraction or move of a taxon into a refugium, its time
186 C. L. Schneider
spent in the refugium, and its ultimate extra-refugial population expansion (or its
demise to a relict population or ultimate extinction, in the case of a refugial trap).
The dilemma of modern conservation is that changes in taxon ranges have only
begun and that refugia from future environmental deterioration are only beginning
to be investigated. The exact locations of future, successful marine refugia, and the
final outcome in terms of timing and nature of the post-refugial recovery are yet
unknown. The fossil record contains multiple intervals of global crisis in which
organisms retreated to refugia, survived, and ultimately recovered, and from these,
there are several lessons:
As the Marine Environment Continues to Change, Refugia May

Need to Shift
Sometimes one refugium, or one type of refugium, is not appropriate for survival
through a crisis. What was an appropriate refugium for initial conditions during the
onset of a global crisis may turn into a trap as conditions deteriorate. Thus, initial
refugia may no longer be adequate for survival through the later conditions of the
crisis. In these instances, refugial survival calls for a succession of range or habitat
shifts or a migration between refugia.
As the global climate warms and the oceans transform, the changing needs of
taxa may be considered in terms of short-term and long-term refugia. At first, short-
term refugia may be needed to support populations through the initial perturbations
caused by climate change (e.g., temporary thermal refugia in areas of delayed
thermal stress; van Hooidonk et al. 2015). Short-term refugia are areas in the
ranges of species where stress levels are not yet fatal, sub-populations are adapted
to slightly higher stress thresholds, or recovery intervals are sufficiently lengthy
between extreme perturbations. Eventually, with ongoing climate change, short-
term refugia will cease to exist as perturbations become too frequent or stress
becomes too great, and taxa will need to migrate to long-term refugia (Ashcroft
2010).
An example of potential refugial shifts in coral reefs already exists. Currently,
short-term refugia from heat stress for coral reefs are in places where stresses
are decreased, such as those within pathways of large-amplitude internal waves
(Buerger et al. 2015; Wall et al. 2015; Schmidt et al. 2016), or where corals have
adapted to have a slightly higher thermal tolerance, such as in gulfs within the
Red Sea (Fine et al. 2013) or in current shadows (McClanahan et al. 2007). While
some corals and associated organisms persist in these short-term refugia, reefs in
general are shifting their ranges northward (Beger et al. 2014; Descombes et al.
2015; examples of coral range shifts include the Caribbean; Precht and Aronson,
2004; Japan; Yamano et al. 2011; Australia; Baird et al. 2012). Meanwhile, as sea
levels continue to rise, reefs can be expected to climb upslope in sync with rising
sea level, facing new consequences from changes in nearshore factors like terrestrial
runoff (e.g., coral reef shifts during interglacial periods; Pandolfi 1999; Greenstein
and Pandolfi 2008). Reefs will also have to contend with changing pollution and
fishing pressures with continuous shore-ward migration. Therefore, future long-term

reef refugia will need to be protected from anthropogenic-caused stresses, upslope
from present-day reef habitats, and in higher latitudes. Finally, habitats that mitigate
some of the effects of ocean acidification may become adequate refugia, such as
shallow-water sea mount habitats suggested by Tittensor et al. (2010).
Conservation strategies must foresee changes in the refugial needs of targeted
taxa or ecosystems. Changing refugial needs are already in consideration by some
researchers: short-term and long-term refugia must be accommodated by current
and future location and management of marine reserves. Currently, most of the
emphasis is on short-term refugia, with or without the realization that many existing
conservation areas and populations adapted to currently higher stress-thresholds
may be only temporary survival strategies. Thus, all marine reserves as they
currently exist support short-term survival, but as environmental conditions change,
many taxa and ecosystems may need to move to long-term refugia as they become
available.
However, one caution: when considering climate change and shifting refugia, the
concept of “plan for the worst, hope for the best” may not be the best strategy for a
successful conservation outcome. One study undertaken by Makino et al. (2015)
compared the results of three separate models of greenhouse gas trajectories in
predicting the future distribution of coral reefs around Japan. Each of the three
results differed with little to no overlap in future coral distributions; thus, there was
no area to establish what they called a “no regrets” marine reserve that would assure
conservation in all three future scenarios. So, implementing a plan required by one
climate change scenario could fail, if the actual, final environment differed from
predicted.
Refugial Size and Connectivity Can Enhance Survivorship,

But Can Also Have Evolutionary Consequences
Size and connectivity of refugia is already understood by conservation biologists in

considering marine reserves and proved repeatedly in the geologic past. Obviously,
a taxon confined to a single, small refugium (or reserve) is at greater risk of being
eliminated by a fatal perturbation than if that taxon is spread across a large area
(Almany et al. 2009). If a marine reserve is sufficiently large, it is comparable to
a refugium that is adequate to support the entire species for the length of time
needed. Multiple small marine reserves are equivalent to cryptic refugia, because
in many cases, they must rely on connectivity between reserves or with individuals
in unprotected areas.
For taxa with dispersal limitations or specific habitat needs, connectivity
becomes difficult and populations are at risk of become fragmented and isolated.
From the geologic past, population isolation of these species has been caused by
a multitude of factors, such as larval duration (e.g., Nucella lima gastropods of
the north Pacific; Cox et al. 2014), dispersal (e.g., female dispersal patterns of
thornback rays; Chevolot et al. 2006), oceanographic conditions (e.g., Alboran
188 C. L. Schneider
Sea isolation during the Messinian salinity crisis; Perrin and Bosellini 2013;
equatorial split in Azorean barnacle populations, Quintiero et al., 2015), and habitat
constraints (e.g., Antarctic intertidal limpet Nacella concinna unable to utilize
deeper refugia; Gonzalez-Wevar et al. 2013). Ultimately, what is expressed in
phylogeographic studies are only those populations that survived. Populations
that were insufficient in abundance or genetic diversity, or experienced other
constraints such as unsurvivable habitat conditions ultimately vanished, and thus are
unrecoverable by genetic analyses. However, of those isolated populations that did
survive, molecular analyses frequently highlight the consequences of geographic
and demographic isolation, namely genetic drift, founder effects, and bottlenecks.
Isolated geographic refugia across the last glacial maximum resulted in many
endemic refugial populations worldwide. Northern temperate and polar taxa, espe-
cially those requiring specific habitats (e.g., rocky intertidal zones; Maggs et al.
2008) and those needing ice-free rookeries (e.g., northern fur seals; Pinsky et al.
2010), commonly display endemism in the genomes that originated from refu-
gial populations. Tropical taxa likewise show the genetic signatures of isolated
populations (e.g., north and south refugia for Great Barrier Reef corals across
the last glacial maximum; Van Oppen et al. 2011), as do populations isolated
in the Southern Ocean (e.g., crinoids; Hemery et al. 2012). At greater extremes,
bottlenecks appear in some populations. Among the most severe are bottlenecks
across the last glacial maximum, in some marine mammals and birds, in which
only a few thousand breeding females survived the decline in habitat availability
and quality (Younger et al. 2016). What is important about these and other studies
of refugia from the last glacial maximum are that founder effects and bottlenecks
within refugia occurred only within a few thousand years and under conditions
that were free of harvesting pressure and pollutants. Thus, isolation within refugia
(whether from natural causes or from fishing pressures outside of a refugium) can
lead to evolutionary consequences within a geologically very short time.
Cryptic refugia may mitigate the evolutionary problems of isolated populations.
By nature, cryptic refugia are connected in space and time. Reproduction, recruit-
ment, and dispersal within the metapopulation cross refugial boundaries, while the
distribution of individuals across the generations shifts as microhabitats emerge or
disappear (Stewart and Lister 2001; Bennett and Provan 2008). Furthermore, cryptic
refugia can be ephemeral. As conditions in a heterogeneous environment shift, new
refugia are created as others disappear. The example of Triassic deltaic refugia of
Gall and Grauvogel-Stamm (2005) for marine taxa is apt: as sediments and currents
shifted in the delta system, old habitats were destroyed and others opened. Likewise,
shifting glacial ice opened and closed ephemeral refugia on the Antarctic shelf
which taxa had to colonize or abandon accordingly (Graham and Smith 2012). It is
important to remember that cryptic refugia occur on the order of very short geologic
time spans, which means that on the order of human life spans, cryptic refugia may
appear fairly stable, such as individual reefs that have survived for decades in a
warming habitat (e.g., temperature refugia on the Great Barrier Reef; Ban et al.
2012).
Thus, the geological record of refugia proves two important points. (1) Genetic
diversity and connectivity must be maintained within and across refugia, a factor
well understood by conservation biologists. (2) Cryptic refugia—those areas sup-

porting small populations that are not viable unto themselves, but contribute to a
greater network of connectivity—must be able to disappear and appear elsewhere,
if circumstances warrant. As with the realization that refugia may need to shift
in response to a changing climate, in heterogeneously perturbed areas, cryptic
refugia may have to be allowed to disappear and new refugia identified as they
become established in new microhabitats. Thus, planning and management must be
sufficiently flexible to accommodate shifting habitats.
However, modern refugia will be fully unlike deep-time refugia in one respect:
marine reserves must also be sources for the populations harvested outside of
protected areas. Thus, reserves must function beyond the support of sufficient
populations within their borders, and act as “super-refugia” serving both the
demographic needs of the reserves and the resources required by a growing human
population. This factor may require refugia that are larger, in area and connectivity,
than a simple survival-through-crisis would require.
Conditions Inside Refugia May Not Necessarily Remain

Pristine, But Will Need to Be of Sufficiently Lower Magnitude
of Total Stress to Maintain Viable Populations
The persistence of a species is affected by the quality of the refugial habitat

(Ashcroft 2010). As stated earlier, Vermeij (1986) pointed out that a good refugium
does not need to be free of stress, but should have tolerable stress levels, which can
include an offsetting or balancing ameliorating factor (such as increased nutrient
supply). As shown in earlier examples, many refugia in the geologic past were not
pristine, unaffected sanctuaries from the stresses that caused mass extinctions. More
often than not, the refugial paleoenvironment was less than perfect.
Although it is beyond the goal of this manuscript to review the effectiveness of
no-take and less protected reserves, studies increasingly show that partial protection,
when compared to no-take marine reserves, may not be sufficient to support
populations or ecosystems (e.g., Kelaher et al. 2014; Nakin and McQuaid 2014;
Toth et al. 2014; Sciberras et al. 2015). In modern ecosystems, there is some
evidence that, by eliminating harvest pressure, no-take marine reserves increase
the resilience of populations to environmental perturbations (e.g., coral survival in
temperature refugia within marine reserves on the Great Barrier Reef; Ban et al.
2012). Thus, decreasing a major stressor or introducing a major benefit may create
greater tolerance to lesser stresses. However, there also is evidence that the presence
of one stress can increase susceptibility to another (e.g., ocean acidification may
increase susceptibility to thermal bleaching in corals; Anthony et al. 2008; Fine
et al. 2013). Likewise, if environmental perturbations occur within refugia, even if
they are not fatal, organisms need sufficient time between stressful events to recover
or may become too vulnerable (Connell et al. 1997; Coumou and Rahmstorf 2012;
Wenger et al. 2016).
190 C. L. Schneider
The geologic record of refugia demonstrates the broad range of tolerance and
adaptation by survivor species to stressful refugial conditions. In some refugia,
stressors were similar to, but reduced from, conditions of the extra-refugial envi-
ronment. One previously discussed example of reduced, but tolerable, stress in
a refugium was the escape of Early Triassic taxa from lethally hot and anoxic
surface water through survival the nearshore habitable zone, which had tolerable,
albeit warm, temperatures and perhaps also had a non-lethal level of oxygen stress
(Beatty et al. 2008; Zonneveld et al. 2010a, b; Clapham et al. 2013; Song et al.
2015; Zhao et al. 2015). Other refugia were very different in the nature stress to be
tolerated, as in the shifting sediments of Triassic deltaic refugia, which experienced
frequent disturbance from shifting sediment and water currents, but was protected
from anoxia and lethal thermal stress in the open ocean (Gall and Grauvogel-Stamm
2005). Sometimes, ameliorating conditions in refugia allowed the species to tolerate
stress. For example, oxygen and food production by Late Permian microbial mats
allowed ostracods to exist in hot surface water (Forel et al. 2013) and the availability
of rocky shores made up for cold, peri-glacial waters for macroalgae across the
last glacial maximum (Olsen et al. 2010). Thus, there are a multitude of potential
ameliorating factors that might enhance a species’ odds of survival in a refugium
where conditions are less than pristine. Which ameliorating factors are beneficial to
survivorship may depend on responses at the species, and even population, level,
and may not be generalizable across taxa.
There has been some suggestion that species that have persisted in refugia
through both interglacial and glacial periods may have a high level of tolerance
to future climate change (Gavin et al. 2014). However, if climate change follows
extreme trajectories because of anthropogenic enhancement (i.e., warming greater
than past interglacial levels and ocean acidification), places that have been long-
term hotspots of diversity, such as Indo-Pacific reefs (Vermeij 1986; Pellissier et al.
2014), may have to be abandoned for new refugia (Descombes et al. 2015).
Beware the Refugial Trap
Refugial traps can seem, initially, to be excellent refugia. Only as the refugium
deteriorates, or when taxa do not expand from the refugium, is the location realized
as a trap. The importance for conservation biology is understanding and predicting
what constitutes a trap rather than a successful refugium.
The commonality across all refugial traps is that, initially, the location appears
to be a suitable refugium. In refugia that have become traps since the last glacial
maximum, Pleistocene populations found the needed sanctuary from environmental
deterioration, but since then, it is the modern populations that are in distress. For
example, some intertidal seaweed populations of Atlantic Canada and the Iberian
Peninsula were unable to expand from their refugia because of geographic barriers
that occurred with rising sea levels or warming oceans (Hoarau et al. 2007; Li et al.
2015), resulting in one Iberian Peninsula isolated, relict population of Fucus serra-
tus rockweed population currently facing acute thermal distress (Hoarau et al. 2007).
Likewise, the Bahaman Caribbean blacknose shark population remains isolated by
oceanographic barriers since its last glacial maximum refugium (Portnoy et al.
2014). As these studies show, initially successful refugia can lead to later isolation
and potentially, populations in distress. In these cases, sea level rise and changes in
ocean currents are two of the factors that contributed to population isolation.
Refugial traps from the ancient geologic past remained viable refugia for longer
periods of time than those of the Pleistocene and the circumstances that terminated
those refugia also tended to be of larger magnitude. Tectonism and eventual drying
of an ocean basin were the demise of ecosystems in the Carboniferous Tindouf Basin
of Morocco (Cozar et al. 2014) and Anadarko Basin of the USA (Holterhoff 1996).
In these cases, the gradual transformation of refugia into traps is apparent in the
tectonic uplift and basin fill. More relevant to modern refugia are the spatially and
temporally shifting, lethally hot and anoxic water masses that eventually devastated
holdover brachiopod faunas in Early Triassic habitable zones (Clapham et al. 2013).
Thus, refugial traps often occur where isolation of populations is inevitable or
where conditions are in danger of fatal levels of deterioration. This may be an
obvious statement, but the key to understanding whether the refugium will become
a trap is in adequately predicting future conditions, such as oceanographic changes
with sea level rises, temperature and salinity fluctuations of water masses, plumes
of polluted water, heterogeneity of acidification, etc. One factor is the recognition
of dead-end basins, which are those areas that are at risk of becoming fatal traps
because there is literally no escape route for the organisms therein from lethal
conditions. For example, in the Red Sea, although short-term refugia are present
because of the higher tolerance of many taxa to salinity and thermal fluctuations
(Fine et al. 2013), the basin may eventually become too warm, and thus become
inhabitable for many of its endemic taxa in the future. If taxa cannot escape
deteriorating conditions in the enclosed basin of the Red Sea, short-term refugia
within the basin will become traps
For most of the fossil examples, the actual changes that turned seemingly
adequate refugia into traps were geologically gradual. From a human time–space
perspective, the development of a trap out of a hospitable environment could be
so gradual that the transformation could be too subtle to be easily noticed during
a single career of conservation research. In this case, marine reserves may require
continuous observation and comparison to their baselines to ensure that the reserves
are not becoming traps. However, refugia can also become traps very rapidly, such as
shifting anoxic water masses that terminated brachiopod faunas in the Early Triassic
habitable zone (Clapham et al. 2013), which may each have occurred suddenly, from
the perspective of geological time.
5 Future Directions for Investigating Ancient Refugia
Since the seminal work of Vermeij (1986), paleobiologists have increasingly

found potential refugia that allowed species to persist during crises in the marine
biosphere. Molecular studies on modern populations have increased the recognition
of marine refugia through the last glacial maximum by an order of magnitude,
particularly in species that underwent population bottlenecks and founder effects.
192 C. L. Schneider
Most importantly, research has focused on the “why” of refugia, rather than simply
identifying the locales of survivorship; it is the reasons for refugial survival that are
pertinent to modern conservation biology.
The fossil and geological record of refugia is a baseline for understanding one
way of how the marine biosphere survives a global crisis. Although lacking the
level of detail of modern ecological investigation, paleontological study of refugia
nonetheless can impart critical information relevant to conservation paleobiology.
The properties of taxa that thrived in refugia, why a particular refugium formed,
how a refugium and the species therein changed through time, and what caused the
successful recovery or fatal trap at the end of the global crisis, are among some of
the questions that paleobiology can answer. These answers, in turn, are relevant to
understanding what may constitute a successful marine reserve.
For conservation biologists, the synthesis of ancient refugia herein is meant to
provide information, as it stands, to help refine current and future marine reserves
for best efficacy. For paleobiologists, the purpose of this synthesis is a baseline
from which future directions of refugial studies can proceed. This requires not
only further research and confirmation of presently hypothesized refugia discussed
above, but also the effort to discover and investigate other refugia. Intervals that
are characterized by environmental and ecological degradation similar to modern
systems may be most appropriate, such as consequences of ocean warming across
the Permo-Triassic mass extinction and the Paleocene-Eocene Thermal Maximum,
and times of reef demise, such as the Frasnian-Famennian biotic crisis.
6 Conclusions
Throughout geologic history, life repeatedly has survived global crises of environ-
mental deterioration and mass extinction. Life meanwhile will survive the current
global crisis, but the important issue to humanity is that species and ecosystems need
to persist in a way that allows humanity to not only survive, but thrive. This means
that humanity needs a marine biosphere that provides the same, and likely greater,
level of services as has always been expected. Thus, it is beneficial to understand
how and why organisms survived ancient global crises, and of importance to marine
reserves to know how refugia played a role in that survivorship.
As reviewed above, refugia are areas where, during a global crisis, species
contract into isolated geographic areas, undergo range and habitat shifts, or find
locations for certain life stages, such as where populations can successfully breed.
While in the refugium, conditions are not necessarily stress free, but the total sum
of detrimental and beneficial environmental parameters in the refugium is sufficient
that populations can remain viable. Furthermore, a single refugium may not persist
through the entire crisis of global climate change; distinct short-term and long-
term refugia may be necessary for some taxa. The ultimate goal of refugia is the
contribution of surviving species to global recovery after a crisis has ended. Species
will need to re-expand into previous, greater, or new ranges, thus re-populating
extra-refugial environments.
Conservation biologists can learn several lessons from ancient extinction sur-
vivorship, applicable to the future of marine reserves as refugia. In particular, with
a future of ongoing climate change, refugial strategies need to be flexible. Some
marine reserves, as they currently exist, may serve as short-term refugia that are
adequate for present environmental conditions, but as perturbations become more
frequent and intense, these areas may become too deteriorated for the species
currently inhabiting them. Short-term refugia will need to be abandoned for long-
term refugia, when they become available and as the new stable state of the oceans
is reached. These refugia may be located at higher latitudes or upslope with sea
level rise relative to the locations of present-day marine reserves. Furthermore, the
locations of short-term refugia, whether as reserves or as habitats recognized as less
perturbed than others, may also need to be flexible in space and time and therefore,
allowed to move elsewhere, especially in areas where environmental deterioration
is heterogeneous.
As in the geologic past, environmental and ecological conditions in current
marine reserves may not remain pristine, but if they are stressed, they will need
to be adequate to support populations that can easily continue to reproduce and
recruit. This not only includes reduced levels of stress and tolerable effects of
combined stressors—such as decreased pollution, buffering from acidification, and
less frequent negative perturbations—but also the amelioration of stress by other
favorable conditions, such as the location of a refugium in areas that have high
nutrient levels (e.g., kelp in upwelling zones; Graham et al. 2007).
Types, sizes, and population connectivity of refugia matter. Multiple cryptic
refugia, similar to smaller marine reserves, can remain connected through inter-
refugial areas to support the continuance of one or more species. Furthermore,
multiple refugia are potential insurance against the risk of extinction: if one or a few
refugia are lost, so long as there are sufficient populations, the species can continue
(Ballantine and Langlois 2008; Ballantine 2014). Otherwise, a single refugium that
is sufficiently large can also insure a low extinction risk, provided that it covers an
adequate selection of habitats and allows space for future range shifts.
“No regrets” marine reserves—those that cover all possible future scenarios of
environmental deterioration—will be difficult to identify, if not impossible (Makino
et al. 2015). Areas of “last resort” where conditions have not yet deteriorated
completely and populations continue to survive must be protected (Chollett and
Mumby 2013) as short-term refugia to allow time for the identifying and planning
for future, long-term refugia. However, it is important to realize that many of these
short-term refugia will also disappear (e.g., Indo-Pacific reef biodiversity hotspots;
Descombes et al. 2015), and that efforts will need to shift to identifying and
establishing new, long-term refugia before the demise of short-term protected areas.
One important lesson from conservation biology to paleobiologists is where
paleontology can provide information relevant to modern marine crises. Pale-
ontology and geology have the unique perspective of following a previous and
devastating global crisis from start through finish and finally to recovery. Refugia
and other extinction survivorship strategies are important to assisting in modern
biosphere survival; as Vermeij (1986) suggested, understanding the how and why of
survivorship of taxa through mass extinctions is of equal importance as studying the
extinction itself.
194 C. L. Schneider
Whatever the type, a refugium (or multiple refugia) needs to last through the
entire crisis. Understanding the length of time that a refugium may be needed
is critical. Refugia across the last glacial maximum existed on the order of
thousands of years; those across mass extinctions were utilized by survivors and
their descendent taxa for millions of years. Modern marine ecosystems are facing
stressors that have been experienced by, and survived by, ancient life, but the modern
marine system has the unique factors of human-caused environmental crises and the
potential for significant human intervention. As the ultimate ecosystem engineer,
humanity has the potential to not only create and maintain adequate refugia, it
also can undo human-caused environmental degradation. So, a refugium (marine
reserve) must last as long as it needs to, which is dependent on how fast and how
well society can “fix” the global environmental crisis. As Kauffman and Erwin
(1995) compared the modern crisis to ancient mass extinctions, if unfixed, recovery
from this newest mass extinction at its greatest could take up to 2–10 million years,
unchecked. Thus, the best long-term solutions need to go beyond identifying and
establishing refugia and need to reverse the human-caused crisis.
Acknowledgements I am grateful to my reviewers S. Donovan and J.P. Zonneveld who provided

advice which greatly improved the manuscript. In particular, they helped improve the flow and
relevance of the manuscript in addition to cuing me to additional studies. I also thank L. Leighton
and the Leighton Lab at the University of Alberta (D. Molinaro, K. Barclay, S. Mendonca, E.
Garcia, M. Provan, A. Rocca, and C. Berendt) who either participated in a discussion about
refugia or commented on an early draft of the manuscript. I also wish to acknowledge, thank,
and remember my late friend, S. Needlepaws, whose presence during future manuscripts will be
sorely missed.
Appendix
A summary of studies cited in the manuscript.

Types of refugia:
R D Range shift
H D Habitat shift
H (D) D Habitat shift to deep water
H (S) D Habitat shift to shallow water
I D Isolated geographic refugia
I (M) D Sea mount refugia
C D Cryptic refugia
L D Life history refugia
LGM D last glacial maximum
LIG D last interglacial
OMZ D oxygen minimum zone
N D north
E D east
S D south
W D west
Reference Type Age, where Taxa Location Interpretation
Westrop (1989) H (D) Late Cambrian, North Trilobites Upper continental Slope environments expanded
America slope during sea level rise
*
Rickards and Wright I Late Ordovician, New South Graptolites Trough-and-rise A series of deep troughs were
(2002) Wales tectonic system isolated from extinction-causing
stress in the ocean
Xu et al. (2005) I Late Ordovician, South Graptolites Yangtze Platform Isolated from open ocean; preserved
China the preferred habitats of graptolites
Rassmussen and Harper I Late Ordovician, various Brachiopods Isolated regions Biodiversity hotspots between first
(2011) and second phases, Late Ordovician
extinction
Huang et al. (2013) H (D) Late Ordovician, South Dicoelosia sp. Shallow water Deep water taxon moved to shallow
China, Norway brachiopods water to avoid “abnormal”
conditions
Lethiers and Cassier H (S) Late Devonian, worldwide Ostracods Shallow water Dysoxic-tolerant ostracods spread
(1999) into near-littoral environments
House (1996) H (D) Late Devonian, worldwide Kokenia sp. goniatite Deeper water Hypothetical; escape hypoxic
events in shallow water
*
Cozar et al. (2014) I Mississippian- Foraminifera, corals, West Tindouf Basin Position in paleocurrents, tectonism,
Pennsylvanian, N brachiopods, algae and emerging land masses isolated
Marine Refugia Past, Present, and Future: Lessons from Ancient Geologic. . .
Africa the basin

**
Holterhoff (1996) I, R Pennsylvanian-Permian, SW Crinoids, including Anadarko Basin Contraction and expansion during
USA one biofacies glacio-eustatic changes; basin
ultimately dried up
195
196
Reference Type Age, where Taxa Location Interpretation

Forel et al. (2013) H Late Permian, equatorial Ostracods, others Microbial mats Microbial mats provide oxygen,
unlimited food supply, but remain in
hot environments
*
Farabegoli et al. (2007) I Late Permian, Italy Benthic fauna Isolated region Few million years survival of
holdover fauna after decline
elsewhere
Waterhouse and Shi I? Late Permian—Early Invertebrates Region, perhaps Southern hemisphere Gondwana
(2010) Triassic, Gondwana with refugia region provided refugia from anoxia
in the northern hemisphere
Schubert and Bottjer ? Early Triassic, W USA Invertebrates Unknown refugia Hypothetical existence of refugia
(1995) from poor environmental conditions
Gall and I, C Early-Middle Triassic, Invertebrates and Delta Holdover, Lazarus, and progenitor
Grauvogel-Stamm (2005) France others taxa protected from anoxia;
fragmented habitats
Beatty et al. (2008) H (S) Early Triassic, W Canada Trace fossils Habitable zone Survival in habitable zone between
wave base and anoxic deeper water
in embayments along coast
*
Posenato (2009) H Early Triassic, Dolomites Orbecoelia, Microbial mats Survival for a few thousand years
Strepto-rhynchus after decline elswhere in less
brachiopods stressed, nearshore habitats
Algeo et al. (2010) H (D) Early Triassic, Japan Radiolarians Deeper water Warm water, low oxygen-tolerant
taxa moved to deep ocean to escape
upwelling OMZ on shelf
Zonneveld et al. (2010a, H (S) Early Triassic, W Canada Trace fossils Habitable zone Survival in narrow habitable zone
b) between wave stressed proximal
and anoxic distal environments
C. L. Schneider
Reference Type Age, where Taxa Location Why refugium exists
Shen et al. (2013) H (S) Early Triassic, S China Acritarchs Shallow water Moved to shallow shelf during
times of stress
*
Clapham et al. (2013) H (S) Early Triassic, W USA, Brachiopods Habitable zone in Moved to littoral zone, but
Italy, S China carbonate platform eventually overcome by shifting
anoxic water masses
Chen et al. (2015) H (S) Early Triassic, S China Brachiopods Mid-water refugia Heat, hypoxia-tolerant species
zone escaped lethally hot surface and
anoxic deep water
Zhao et al. (2015) H (S) Early Triassic, S China Trace fossils Shallow water Refugium between lethally hot
surface water and anoxic deep water
Stanley and Beauvais I (M) Early Jurassic, North Reef corals Volcanic island Long-lived survivors of Karnian and
(1994) America, Peru end Triassic extinction events in
accreted terrains
Westermann (2001) I Early Jurassic, New Zealand Ammonites Contraction to Contraction from E Pacific to New
region Zealand, later re-expanded to E
Pacific
Kiessling et al. (2009) I Early Jurassic, Europe Reef corals Continental shelf Did not need volcanic islands for
refugia; continental shelf
environment was sufficient
*
Yin and Fursich (2008) I Early Jurassic, Tibet Ammonites, bivalves Isolated region Several taxa survived in Tibet
longer than elsewhere
Jacobs and Lindenberg I Cretaceous, Pacific Hydrothermal vent Submarine With high spreading rates, vent
(1998) faunas volcanoes faunas escaped deep water anoxia
on volcanoes
Gibbs et al. (2016) H PETM, worldwide Coccolithophores Higher latitudes Heat-intolerant during one life
cycle, but tolerated pH changes at
higher latitudes
197
198

*
Perrin and Bosellini I Miocene, Alboran Sea Corals Range contraction Initially escaped onset of Messinian
(2013) salinity crisis; finally only Porites
and microbialites remained
Calvo et al. (2015) I Miocene, Atlantic, E Reef-building Range shift Contraction to areas with normal
Mediterranean vermetid gastropods marine conditions
Petuch (1981) I Pliocene-present, Gulf of Molluscs Isolated region Area isolated by ocean currents and
Venezuela upwelling; colder temperatures than
the rest of Caribbean
Vermeij (1986) I Plio-Pleistocene, Tropical shallow Diversity hotspot in Nutrient-rich and comparatively low
Indo-Pacific marine species Indo-Pacific magnitude of changes in
temperature
Greenstein and Pandolfi R LIG, Australia Reef corals Higher latitudes “Temperature refugia” in higher
(2008) latitudes away from equatorial
warming
Barnes et al. (2006) R, H LGM, Antarctic shelf Benthos Continental slope Speciation down slope; modern
eurybathy among taxa
Barnes and Kulinski I LGM, Antarctic shelf Bryozoans Isolated areas No deep sea refugia, so must be in
(2010) situ ice-free areas on the shelf
Allcock and Strugnell I, H LGM, Peri-Antarctic areas Marine taxa Isolated shelf areas, Populations in ice-free shelf refugia
(2012) deep sea went through bottlenecks; deep sea
taxa didn’t
Graham and Smith (2012) I LGM, Alexander Island, W – Isolated areas Geophysical study; ephemeral and
Antarctic Peninsula one permanent ice-free areas on the
outer shelf
Hemery et al. (2012) I LGM, various shelf and Promachocrinus Isolated populations Development of genetic endemism
islands, Antarctica kerguelensis crinoid of seven populations
C. L. Schneider
Gonzalez-Wevar et al. I LGM, S Georgia Island Nacella concinna Isolated area Development of genetic endemism
(2013) limpet of S Georgia Island population
Younger et al. (2016) I (L) LGM, Antarctic coastal Penguins, petrels, Rookeries Ice free breeding habitats and
areas elephant seals nearby foraging grounds limited;
many genetic bottlenecks
Provan et al. (2005) I LGM, N Atlantic Palmaria palmate Isolated areas Development of genetic endemism
seaweed in Hurd Deep, NW Atlantic, and
Ireland populations
Harlin-Cognato et al. I LGM, N Atlantic, N Pacific Eumetopias jubatus Isolated rookeries Contraction and expansion across
(2006) Stellar’s sea lion glacial cycles with decline and
increase in rookery availability
Chevolot et al. (2006) I LGM, NE Atlantic Raja clavata, Isolated areas LGM populations centered in Iberia,
Thornback rays the Mediterranean, and Hurd Deep
Jolly et al. (2006) I LGM, NE Atlantic Two spp. polychaete Isolated areas Shared history of vicariant events
tubeworms between the two taxa, including
northern peri-glacial refugia
Hoarau et al. (2007) I LGM, NE Atlantic Fucus serratus Isolated areas Expansion from Ireland, less so
rockweed Hurd Deep populations; relict
Iberian Peninsula population
Maggs et al. (2008) I LGM, N Atlantic Eight intertidal species Isolated areas Nine rocky shoreline refugial
populations, including four
peri-glacial northern areas
Campo et al. (2010) I, R LGM, NW Atlantic Pollicipes pollicipes Isolated areas and Two refugial populations in Iberia
stalked barnacle southerly shift and Brittany; large N Africa
population; expansion during LIG
Olsen et al. (2010) I LGM, N Atlantic Ascophyllum nodosum Isolated areas Refugial populations from Atlantic
seaweed Canada, Brittany rocky intertidal
areas
199
200

Krebes et al. (2010) H LGM, NE Atlantic, Europe Gammarus duebeni Brackish coastal Refugia in brackish coastal glacial
amphipod outwash outwash lakes to avoid competition,
later adapted to rivers
Doellman et al. (2011) I LGM, N Atlantic Littorina saxatilis Isolated areas Many local European refugia; two
gastropod northern Atlantic and two Atlantic
N American refugia
Almada et al. (2012) I LGM, N Atlantic Taurulus bubalis Isolated areas Coldwater species; multiple refugia
longspined bullhead in Atlantic and north sea, including
peri-glacial refugia
Krakau et al. (2012) I, R LGM, N Atlantic Cerastoderma edule Isolated area and Northern peri-glacial refugium and
cockle range shift large LGM population British Isles
to northwest Africa
Waltari and Hickerson I LGM, N Atlantic Six intertidal Isolated areas Spp distribution models identified
(2013) invertebrates several European and North
American refugia
Naro-Maciel et al. (2014) I LGM, Central Atlantic Chelonia mydas green Isolated rookeries Distinct N and S populations, each
sea turtle with multiple small refugial
populations across the LGM
Neiva et al. (2014) I, R LGM, NE Atlantic Pelvetia canaliculata Isolated area and Spp distribution models: Morocco
fucoid seaweed range shift south to Celtic Sea; phylogeography:
refugia in Iberia, Great Britain
Quintiero et al. (2015) I, R LGM, Azores to Cabo Verde Megabalanus azoricus Isolated areas Stable southern refugia during
islands, Atlantic barnacle glacial maxima in the islands
Li et al. (2015) I LGM, N Atlantic Palmaria palmata red Isolated areas Multiple European endemic refugial
macroalgae populations; two relict refugial
populations Atlantic Canada
C. L. Schneider
Assis et al. (2016) I LGM, NE Atlantic Laminaria Isolated areas Niche modelling; several LGM
hyperborean kelp refugia; southern Iberian
populations will disappear with
warming
Mateus et al. (2016) I LGM, Iberia European lampreys Isolated area Long term isolation of LGM
refugial Iberian populations; later
expansion northward
Albaina et al. (2012) I, R LGM, Adriatic Sea, Iberia Nassarius nitidus Isolated areas Refugia in the Adriatic Sea and
whelk around the Iberian Peninsula;
current expansion front at North Sea
Portnoy et al. (2014) I, R LGM, S Gulf of Mexico Blacknose shark Isolated area Five modern populations from one
refugial population in southern Gulf
of Mexico
Hellberg et al. (2001) I, R LGM, southern California Acanthinucella spirata Isolated area, range Contraction of population to an area
gastropod shift north of Los Angeles
Pinsky et al. (2010) L, R LGM, NE Pacific Callorhinus ursinus Isolated rookeries High disperser that utilized many
northern fur seals available sub-Arctic rookeries, so no
endemic populations
Grant and Cheng (2012) L LGM, N Pacific Paralithodes Isolated nurseries Juvenile nurseries near Japan,
camptschaticus crab Kodiak Island, Haida Gwaii
Cox et al. (2014) I, R LGM, N Pacific Nucella lima gastropod Isolated areas Split between E, W refugia 317 ka;
E Pacific population bottleneck
during LGM
Van Oppen et al. (2011) I, R LGM, Great Barrier Reef Acropora millepora Isolated areas Northern and southern refugial
coral populations during glacial maxima

Ludt et al. (2012) I LGM, Hawaii Two spp. lagoonal and Isolated areas Lagoonal populations fragmented
two spp. reef slope with sea level drop; reef slope
wrass population remained connected
Pellissier et al. (2014) I LGM, Indro-Australian Reef fish Reefs Stable reef habitats during past
Archipelago climate changes influence modern
201
reef fish distribution

202 C. L. Schneider
References
Aberhan M, Kiessling W (2014) Rebuilding biodiversity of Patagonian marine molluscs after the
end-Cretaceous mass extinction. PLoS One 9:e102629
Albaina N, Olsen JL, Couceiro L et al (2012) Recent history of the European Nassarius nitidus
(Gastropoda): phylogeographic evidence of glacial refugia and colonization pathways. Mar
Biol 159:187–1884
Algeo TJ, Hinnov L, Moser J et al (2010) Changes in productivity and redox conditions in the
Panthalassic Ocean during the latest Permian. Geology 38:187–190
Allcock AL, Strugnell JM (2012) Southern Ocean diversity: new paradigms from molecular
ecology. Trends Ecol Evol 27:520–528
Almada VC, Almada F, Francisco SM et al (2012) Unexpected high genetic diversity at the extreme
northern geographic limit of Taurulus bubalis (Euphrasen, 1786). PLoS One 7:e44404
Almany GR, Connolly SR, Heath DD et al (2009) Connectivity, biodiversity conservation and the
design of marine reserve networks for coral reefs. Coral Reefs 28:339–351
Anthony KRN, Kline DI, Diaz-Pulido G et al (2008) Ocean acidification causes bleaching and
productivity loss in coral reef builders. Proc Natl Acad Sci USA 105:17442–17446
Ashcroft MB (2010) Identifying refugia from climate change. J Biogeogr 37:1407–1413
Assis J, Lucas AV, Barbara I et al (2016) Future climate change is predicted to shift long-
term persistence zones in the cold-temperate kelp Laminaria hyperborea. Mar Environ Res
113:174–182
Baird AH, Sommer B, Madin JS (2012) Pole-ward range expansion of Acropora spp. along the
east coast of Australia. Coral Reefs 31:1063–1063
Ballantine B (2014) Fifty years on: lessons from marine reserves in New Zealand and principles
for a worldwide network. Biol Conserv 176:297–307
Ballantine WJ, Langlois J (2008) Marine reserves: the need for systems. Hydrobiologia 606:35–44
Ban NC, Pressey RL, Weeks S (2012) Conservation objectives and sea-surface temperature
anomalies in the Great Barrier Reef. Conserv Biol 26:799–809
Barnes DKA, Kulinski P (2010) Bryozoans of the Weddell Sea continental shelf, slope and abyss:
did marine life colonize the Antarctic shelf from deep water, outlying islands or in situ refugia
following glaciations? J Biogeogr 37:1648–1656
Barnes DKA, Hodgson DA, Convey P et al (2006) Incursion and excursion of Antarctic biota: past,
present and future. Glob Ecol Biogeogr 15:121–142
Baskett ML, Barnett LAK (2016) The ecological and evolutionary consequences of marine
reserves. Annu Rev Ecol Evol Syst 46:49–73
Beatty TW, Zonneveld J-P, Henderson CM (2008) Anomalously diverse Early Triassic ichnofossil
assemblages in northwest Pangea: a case for a shallow-marine habitable zone. Geology
36:771–774
Beger M, Sommer B, Harrison PL et al (2014) Conserving potential coral reef refuges at high
latitudes. Divers Distrib 20:245–257
Bennett KD, Provan J (2008) What do we mean by ‘refugia’? Quat Sci Rev 27:2449–2455
Berger M, Babcock R, Booth DJ et al (2011) Research challenges to improve the management
and conservation of subtropical reefs to tackle climate change threats (Findings of a workshop
conducted in Coffs Harbour, Australia on 13 September 2010). Ecol Manag Restor 12:e7–e10
Bongaerts P, Ridgway T, Sampayo EM et al (2010) Assessing the ‘deep reef refugia’ hypothesis:
focus on Caribbean reefs. Coral Reefs 29:309–327
Buerger P, Schmidt GM, Wall M et al (2015) Temperature tolerance of the coral Porites
lutea exposed to simulated large amplitude internal waves (LAIW). J Exp Mar Biol Ecol
471:232–239
Calvo M, Alda F, Oliverio M et al (2015) Surviving the Messinian Salinity Crisis? Divergence
patterns in the genus Dendropoma (Gastropoda: Vermetidae) in the Mediterranean Sea. Mol
Phylogenet Evol 91:17–26
Campo D, Molares J, Garcia L et al (2010) Phylogeography of the European stalked barnacle

(Pollicipes pollicipes): identification of glacial refugia. Mar Biol 157:147–156
Chen J, Tong J, Song H et al (2015) Recovery pattern of brachiopods after the Permian–Triassic
crisis in South China. Palaeogeogr Palaeoecol 433:91–105
Chevolot M, Hoarau G, Rijnsdorp AD (2006) Phylogeography and population structure of
thornback rays (Raja clavata L., Rajidae). Mol Ecol 15:3693–3705
Chollett I, Mumby PJ (2013) Reefs of last resort: locating and assessing thermal refugia in the
wider Caribbean. Biol Conserv 167:179–186
Chollett I, Enriquez S, Mumby PJ (2014) Redefining thermal regimes to design reserves for coral
reefs in the face of climate change. PLoS One 9:e110634
Chown SL, Clarke A, Fraser CI et al (2015) The changing form of Antarctic biodiversity. Nature
522:431–438
Clapham ME, Fraiser ML, Marenco PJ et al (2013) Taxonomic composition and environmental
distribution of post-extinction rhynchonelliform brachiopod faunas: constraints on short-term
survival and the role of anoxia in the end-Permian mass extinction. Palaeogeogr Palaeoecol
374:284–292
Commission for Environmental Cooperation (2011) North American marine protected areas
network. http://www2.cec.org/nampan/. Accessed 05 Nov 2016
Connell JH, Hughes TP, Wallace CC (1997) A 30-year study of coral abundance, recruitment, and
disturbance at several scales in space and time. Ecol Monogr 67:461–488
Coumou D, Rahmstorf S (2012) A decade of weather extremes. Nat Clim Chang 2:491–496
Cox LN, Zaslavskaya NI, Marko PB (2014) Phylogeography and trans-Pacific divergence of the
rocky shore gastropod Nucella lima. J Biogeogr 41:615–627
Cozar P, Vachard D, Somerville ID et al (2014) The Tindouf Basin, a marine refuge during
the Serpukhovian (Carboniferous) mass extinction in the northwestern Gondwana platform.
Palaeogeogr Palaeoecol 394:12–28
da Silva IM, Hill N, Shimadzu H et al (2015) Spillover effects of a community-managed marine
reserve. PLoS One 10:e0111774
Darwin C (1859) On the origin of species by means of natural selection, or the preservation of
favoured races in the struggle for life. John Murray, London
Descombes P, Wisz MS, Leprieur F et al (2015) Forecasted coral reef decline in marine biodiversity
hotspots under climate change. Glob Chang Biol 21:2479–2487
Doellman MM, Trussell GC, Grahame JW et al (2011) Phylogeographic analysis reveals a deep
lineage split within North Atlantic Littorina saxatilis. Proc Biol Sci 278:3175–3183
Donovan SK (1989) Palaeontological criteria for the recognition of mass extinction. In: Donovan
SK (ed) Mass extinctions: processes and evidence. Bellhaven Press, London, pp 19–36
Edgar GJ, Stuart-Smith RD, Willis TJ et al (2014) Global conservation outcomes depend on marine
protected areas with five key features. Nature 506:216–220
Erwin DH, Hua-Zhang PH (1996) Recoveries and radiations: gastropods after the Permo-Triassic
mass extinction. In: Hart MB (ed) Biotic recovery from mass extinction events. Geological
Society Special Publication, London, pp 223–229
Fara E (2001) What are Lazarus taxa? Geol J 36:291–303
Farabegoli E, Perri MC, Posenato R (2007) Environmental and biotic changes across the Permian–
Triassic boundary in western Tethys: The Bulla parastratotype, Italy. Glob Planet Chang
55:109–135
Fine M, Gildor H, Genin A (2013) A coral reef refuge in the Red Sea. Glob Chang Biol 19:
3640–3647
Flessa KW, Jablonski D (1983) Extinction is here to stay. Paleobiology 9:315–321
Forel M-B, Crasquin S, Kershaw S et al (2013) In the aftermath of the end-Permian extinction: the
microbialite refuge? Terra Nova 25:137–143
Galil BS, Zibrowius H (1998) First benthos samples from Eratosthenes Seamount, eastern
Mediterranean. Senckenberg Marit 28:111–121
Gall J-C, Grauvogel-Stamm L (2005) The early Middle Triassic ‘Grès àVoltzia’ Formation of
eastern France: a model of environmental refugium. Comptes Rendus Palevol 4:637–652
204 C. L. Schneider
Garcia-Gomez JC, Lopez-Fe CM, Espinosa F et al (2011) Marine artificial micro-reserves: a

possibility for the conservation of endangered species living on artificial substrata. Mar Ecol
32:6–14
Garcia-Gomez JC, Guerra-Garcia JM, Espinosa F et al (2014) Artificial Marine Micro-Reserves
Networks (AMMRNs): an innovative approach to conserve marine littoral biodiversity and
protect endangered species. Mar Ecol 36:259–277
Gavin DG, Fitzpatrick M, Gugger PF et al (2014) Climate refugia: joint inference from fossil
records, species distribution models and phylogeography. New Phytol 204:37–54
Gell FR, Roberts CM (2003) Benefits beyond boundaries: the fishery effects of marine reserves.
Gibbs SJ, Bown PR, Ridgwell A et al (2016) Ocean warming, not acidification, controlled
coccolithophore response during past greenhouse climate change. Geology 44:60–62
Gonzalez-Wevar C, Saucede T, Morley SA et al (2013) Extinction and recolonization of maritime
Antarctica in the limpet Nacella concinna (Strebel, 1908) during the last glacial cycle: toward a
model of Quaternary biogeography in shallow Antarctic invertebrates. Mol Ecol 22:5221–5236
Graham AGC, Smith JA (2012) Palaeoglaciology of the Alexander Island ice cap, western
Antarctic Peninsula, reconstructed from marine geophysical and core data. Quat Sci Rev
35:63–81
Graham MH, Kinlan BP, Druehl LD et al (2007) Deep-water kelp refugia as potential hotspots of
tropical marine diversity and productivity. Proc Natl Acad Sci USA 104:16576–16580
Grant WS, Cheng W (2012) Incorporating deep and shallow components of genetic structure into
the management of Alaskan red king crab. Evol Appl 5:820–837
Greenstein BJ, Pandolfi JM (2008) Escaping the heat: range shifts of reef coral taxa in coastal
Western Australia. Glob Chang Biol 14:513–528
Harlin-Cognato A, Bickham JW, Loughlin TR et al (2006) Glacial refugia and the phylogeography
of Steller’s sea lion (Eumatopias jubatus) in the North Pacific. J Evol Biol 19:955–969
Harries PJ, Knorr PO (2009) What does the ‘Lilliput Effect’ mean? Palaeogeogr Palaeoecol
284:4–10
Harries PJ, Kauffman EG, Hansen TA (1996) Models for biotic survival following mass extinction.
In: Hart MB (ed) Biotic recovery from mass extinction events. The Geological Society, London,
pp 41–60
Hautmann M, Nützel A (2005) First record of a heterodont bivalve (Mollusca) from the Early Tri-
assic: palaeoecological significance and implications for the ‘Lazarus problem’. Palaeontology
48:1131–1138
Hautmann M, Bucher H, Bruhwiler T et al (2011) An unusually diverse mollusc fauna from the
earliest Triassic of South China and its implications for benthic recovery after the end-Permian
biotic crisis. Geobios 44:71–85
Hellberg ME, Balch DP, Roy K (2001) Climate-driven range expansion and morphological
evolution in a marine gastropod. Science 292:1707–1710
Hemery LG, Eleaume M, Roussel V et al (2012) Comprehensive sampling reveals circumpolarity
and sympatry in seven mitochondrial lineages of the Southern Ocean crinoid species Proma-
chocrinus kerguelensis (Echinodermata). Mol Ecol 21:2502–2518
Hoarau G, Coyer JA, Veldsink JH et al (2007) Glacial refugia and recolonization pathways in the
brown seaweed Fucus serratus. Mol Ecol 16:3606–3616
Holterhoff PF (1996) Crinoid biofacies in Upper Carboniferous cyclothems, midcontinent North
America: faunal tracking and the role of regional processes in biofacies recurrence. Palaeogeogr
Palaeoecol 127:47–81
House MR (1996) Juvenile goniatite survival strategies following Devonian extinction events.
In: Hart MB (ed) Biotic recovery from mass extinction events. Geological Society Special
Publication, London, pp 163–185
Howarth LM, Roberts CM, Hawkins JP et al (2015) Effects of ecosystem protection on scallop
populations within a community-led temperate marine reserve. Mar Biol 16:823–840
Huang B, Rong J-Y, Harper DAT (2013) A new survivor species of Dicoelosia (Brachiopoda) from
Rhuddanian (Silurian) shallower-water biofacies in south China. J Paleontol 87:232–242
Inger R, Attrill MJ, Bearhop S et al (2009) Marine renewable energy: potential benefits to
biodiversity? An urgent call for research. J Appl Ecol 46:1145–1153
Jablonski D (1986) Causes and consequences of mass extinctions: a comparative approach. In:
Elliot DK (ed) Dynamics of extinction. Wiley and Sons, New York, pp 183–229
Jablonski D, Flessa KW (1986) The taxonomic structure of shallow-water faunas: implications for
Phanerozoic extinctions. Malacologia 27:43–66
Jacobs DK, Lindenberg DR (1998) Oxygen and evolutionary patterns in the sea: onshore/offshore
trends and recent recruitment of deep-sea faunas. Proc Natl Acad Sci U S A 95:9396–9401
Johannes RE (1978) Traditional marine conservation methods in Oceania and their demise. Annu
Rev Ecol Syst 9:349–364
Jolly MT, Viard F, Gentil F et al (2006) Comparative phylogeography of two coastal polychaete
tubeworms in the Northeast Atlantic supports shared history and vicariant events. Mol Ecol
15:1841–1855
Kauffman EG, Erwin DH (1995) Surviving mass extinctions. Geotimes 14:14–17
Kauffman EG, Harries PJ (1996) The importance of crisis progenitors in recovery from mass
extinction. In: Hart MB (ed) Biotic recovery from mass extinction events. Geological Society
Special Publication, London, pp 15–39
Kelaher BP, Coleman MA, Broad A et al (2014) Changes in fish assemblages following the
establishment of a network of no-take marine reserves and partially-protected areas. PLoS One
9:e85825
Kiessling W, Aberhan M (2007) Geographical distribution and extinction risk: lessons from
Triassic–Jurassic marine benthic organisms. J Biogeogr 34:1473–1489
Kiessling W, Roniewicz E, Villier L et al (2009) An early Hettangian coral reef in southern France:
implications for the end-Triassic reef crisis. PALAIOS 24:657–671
Knaust D (2010) The end-Permian mass extinction and its aftermath on an equatorial carbonate
platform: insights from ichnology. Terra Nova 22:195–202
Krakau M, Jacobsen S, Jensen KT et al (2012) The cockle Cerastoderma edule at Northeast
Atlantic shores: genetic signatures of glacial refugia. Mar Biol 159:221–230
Krebes L, Blank M, Jurss K et al (2010) Glacial-driven vicariance in the amphipod Gammarus
duebeni. Mol Phylogenet Evol 54:372–385
Leighton LR, Schneider CL (2008) Taxon characteristics that promote survivorship through
the Permian-Basal Triassic interval: implications for the change in evolutionary faunas.
Paleobiology 34:78–92
Lethiers F, Cassier J-G (1999) Autopsie d’une extinction biologique. Un exemple: la crise de al
limite Frasnien-Famennien (364 Ma). Earth Planet Sci 329:303–315
Li J, Hu Z, Duan D (2015) Genetic data from the red alga Palmaria palmata reveal a mid-
Pleistocene deep genetic split in the North Atlantic. J Biogeogr 42:902–913
Ludt WB, Rocha LA (2015) Shifting seas: the impacts of Pleistocene sea-level fluctuations on the
evolution of tropical marine taxa. J Biogeogr 42:25–38
Ludt WB, Bernal MA, Bowen BW et al (2012) Living in the past: phylogeography and population
histories of Indo-Pacific wrasses (genus Halichoeres) in shallow lagoons versus outer reef
slopes. PLoS One 7:e38042
Maggs CA, Castilho R, Foltz D et al (2008) Evaluating signatures of glacial refugia for North
Atlantic benthic marine taxa. Ecology 89:S108–S122
Maggs CA, Mann BQ, Cowley PD (2013) Contribution of a large no-take zone to the management
of vulnerable reef fishes in the South-West Indian Ocean. Fish Res 144:38–47
Makino A, Klein CJ, Possingham HP et al (2015) The effect of applying alternate IPCC climate
scenarios to marine reserve design for range changing species. Conserv Lett 8:320–328
Mateus CS, Almeida PR, Mesquita N et al (2016) European lampreys: new insights on postglacial
colonization, gene flow and speciation. PLoS One 11:e0148107
McClanahan TR, Ateweberhan M, Muhando CA et al (2007) Effects of climate and seawater
temperature variation on coral bleaching and mortality. Ecol Monogr 77:503–525
Mertens KN, Takano Y, Head MJ et al (2014) Living fossils in the Indo-Pacific warm pool: a refuge
for thermophilic dinoflagellates during glaciations. Geology 42:531–534
206 C. L. Schneider
Myers C, MacKenzie RA III, Lieberman BS (2013) Greenhouse biogeography: the relationship

of geographic range to invasion and extinction in the Cretaceous Western Interior Seaway.
Paleobiology 39:135–148
Nakin MDV, McQuaid CD (2014) Marine reserve effects on population density and size structure
of commonly and rarely exploited limpets in South Africa. Afr J Mar Sci 36:303–311
Naro-Maciel E, Reid BN, Alter SE et al (2014) From refugia to rookeries: phylogeography of
Atlantic green turtles. J Exp Mar Biol Ecol 461:306–316
National Ocean Service (2016) Hawaiian islands humpback whale national marine sanctuary. http:/
/hawaiihumpbackwhale.noaa.gov. Accessed 05 Nov 2016
Neiva J, Assis J, Fernandes F et al (2014) Species distribution models and mitochondrial DNA
phylogeography suggest an extensive biogeographical shift in the high-intertidal seaweed
Pelvetia canaliculata. J Biogeogr 6:1137–1148
New Zealand Department of Conservation (2016) Marine reserves A-Z. http://www.doc.govt.nz/
marinereserves. Accessed 04 Nov 2016
Obama B (2014) Proclamation 9173—Pacific remote islands national marine monument expan-
sion. Fed Regist 79:58645–58653
Olsen JL, Zechman FW, Hoarau G et al (2010) The phylogeographic architecture of the fucoid
seaweed Ascophyllum nodosum: an intertidal ‘marine tree’ and survivor of more than one
glacial–interglacial cycle. J Biogeogr 37:842–856
Packard AS (1886) Geological extinction and some of its apparent causes. Am Nat 20:29–40
Pandolfi JM (1999) Response of Pleistocene coral reefs to environmental change over long
temporal scales. Am Zool 39:113–130
Payne JL (2005) Evolutionary dynamics of gastropod size across the end-Permian extinction and
through the Triassic recovery interval. Paleobiology 31:269–290
Pellissier L, Leprieur F, Parravinci V et al (2014) Quaternary coral reef refugia preserved fish
diversity. Science 344:1016–1019
Perrin C, Bosellini FR (2013) The Late Miocene coldspot of z-coral diversity in the Mediterranean:
patterns and causes. Comptes Rendus Palevol 12:245–255
Petuch EJ (1981) A relict Neogene caenogastropod fauna from northern South America. Mala-
cologia 20:307–347
Pinsky ML, Newsome SD, Dickerson BR et al (2010) Dispersal provided resilience to range
collapse in a marine mammal: insights from the past to inform conservation biology. Mol Ecol
19:2418–2429
Portnoy DS, Hollenbeck CM, Belcher CN et al (2014) Contemporary population structure
and post-glacial genetic demography in a migratory marine species, the blacknose shark,
Carcharhinus acronotus. Mol Biol 23:5480–5495
Posenato R (2009) Survival patterns of macrobenthic marine assemblages during the end-Permian
mass extinction in the western Tethys (Dolomites, Italy). Palaeogeogr Palaeoecol 280:150–167
Precht WF, Aronson RB (2004) Climate flickers and range shifts of reef corals. Front Ecol Environ
2:307–314
Provan J, Wattier RA, Maggs CA (2005) Phylogeographic analysis of the red seaweed Palmaria
palmata reveals a Pleistocene marine glacial refugium in the English Channel. Mol Ecol
14:793–803
Quintiero J, Manent P, Perez-Dieguez L et al (2015) Phylogeography of a marine insular endemic
in the Atlantic Macaronesia: The Azorean barnacle, Megabalanus azoricus (Pilsbry, 1916).
PLoS One 10:e0124707
Rassmussen C, Harper DAT (2011) Interrogation of distributional data for the End Ordovician
crisis interval: where did disaster strike? Geol J 46:478–500
Rickards RB, Wright AJ (2002) Lazarus taxa, refugia and relict faunas: evidence from graptolites.
J Geol Soc Lond 159:1–4
Roberts CM, Hawkins JP (1997) How small can a marine reserve be and still be effective? Coral
Reefs 16:150
Rodland DL, Bottjer DJ (2001) Biotic recovery from the end-Permian mass extinction: behavior
of the inarticulate brachiopod Lingula as a disaster taxon. PALAIOS 16:95–101
Rogers-Bennett L, Pearse JS (2001) Indirect benefits of marine protected areas for juvenile
abalone. Conserv Biol 15:642–647
Rowden AA, Dower JF, Schlacher TA (2010) Paradigms in seamount ecology: fact, fiction and
future. Mar Ecol 31:226–241
Schmidt GM, Wall M, Taylor M et al (2016) Large-amplitude internal waves sustain coral health
during thermal stress. Coral Reefs 35:869–881
Schubert JK, Bottjer DJ (1995) Aftermath of the Permian-Triassic mass extinction event: paleoe-
cology of Lower Triassic carbonates in the western USA. Palaeogeogr Palaeoecol 116:1–39
Sciberras M, Jenkins SR, Mant R et al (2015) Evaluating the relative conservation value of fully
and partially protected marine areas. Environ Evid 16:58–77
Shen J, Lei Y, Algeo TJ et al (2013) Volcanic effects on microplankton during the Permian-Triassic
transition (Shangsi and Xinmin, south China). PALAIOS 28:552–567
Shoo LP, Hoffman AA, Garnett S et al (2013) Making decisions to conserve species under climate
change. Clim Chang 119:239–246
Song H, Yang L, Tong J et al (2015) Recovery dynamics of foraminifers and algae following the
Permian-Triassic extinction in Qingyan, South China. Geobios 48:71–83
Stanley GD Jr, Beauvais L (1994) Corals from an early Jurassic coral reef in British Columbia:
refuge on an oceanic island reef. Lethaia 27:35–47
Steneck RS, Paris CB, Arnold SN et al (2009) Thinking and managing outside the box: coalescing
connectivity networks to build region-wide resilience in coral reef ecosystems. Coral Reefs
28:367–378
Stewart JR, Lister AM (2001) Cryptic northern refugia and the origins of the modern biota. Trends
Ecol Evol 16:608–613
Stewart JR, Lister AM, Barnes I et al (2010) Refugia revisited: individualistic responses of species
in space and time. Proc Biol Sci 277:661–671
Tittensor DP, Baco AR, Hall-Spencer JM et al (2010) Seamounts as refugia from ocean acidifica-
tion for cold-water stony corals. Mar Ecol 31:212–225
Toth LT, van Woesik R, Murdoch TJT et al (2014) Do no-take reserves benefit Florida’s corals?
14 years of change and stasis in the Florida Keys National Marine Sanctuary. Coral Reefs
33:565–577
U.S. Department of State (2016) On the new marine protected area in Antarctica’s Ross Sea. https:/
/www.state.gov/secretary/remarks/2016/10/263763.htm. Accessed 04 Nov 2016
Urbanek A (1993) Biotic crises in the history of the Upper Silurian graptoloids: a palaeobiological
model. Hist Biol 7:29–50
van Hooidonk R, Maynard JA, Liu Y et al (2015) Downscaled projections of Caribbean coral
bleaching that can inform conservation planning. Glob Chang Biol 21:3389–3401
Van Oppen MJH, Peplow LM, Kininmonth S et al (2011) Historical and contemporary factors
shape the population genetic structure of the broadcast spawning coral, Acropora millepora, on
the Great Barrier Reef. Mol Ecol 20:4899–4914
Vermeij GJ (1986) Survival during biotic rises: the properties and evolutionary significance of
refuges. In: Elliott DK (ed) Dynamics of extinction. Wiley, New York, pp 231–246
Wall M, Putchim L, Schmidt GM et al (2015) Large-amplitude internal waves benefit corals during
thermal stress. Proc Biol Sci 282:1–9
Waltari E, Hickerson MJ (2013) Late Pleistocene species distribution modelling of North Atlantic
intertidal invertebrates. J Biogeogr 40:249–260
Washington Department of Fish and Wildlife (2016) Marine protected areas within Puget Sound.
http://wdfw.wa.gov/fishing/mpa/sanjuan_upright_channels.html. Accessed 04 Nov 2016
Waterhouse JB, Shi GR (2010) Evolution in a cold climate. Palaeogeogr Palaeoecol 298:17–30
Wenger AS, Williamson DH, daSilva ET et al (2016) Effects of reduced water quality on coral
reefs in and out of no-take marine reserves. Conserv Biol 30:142–153
Westermann GEG (2001) Modes of extinction, pseudo-extinction and distribution in Middle
Jurassic ammonites: terminology. Can J Earth Sci 38:187–195
208 C. L. Schneider
Westrop SR (1989) Trilobite mass extinction near the Cambrian-Ordovician boundary in North
America. In: Donovan SK (ed) Mass extinctions: processes and evidence. Columbia University
Press, New York, pp 89–103
White House (2016) Fact sheet: President Obama to create the world’s largest marine
protected area. https://www.whitehouse.gov/the-press-office/2016/08/26/fact-sheet-president-
obama-create-worlds-largest-marine-protected-area. Accessed 04 Nov 2016
Wignall PB, Benton MJ (1999) Lazarus taxa and fossil abundance at times of biotic crisis. J Geol
Soc Lond 156:453–456
Xu C, Melchin MJ, Sheets HD et al (2005) Patterns and processes of latest Ordovician graptolite
extinction and recovery based on data from south China. J Paleontol 79:842–861
Yamaguchi T, Norris RD (2015) No place to retreat: heavy extinction and delayed recovery on a
Pacific guyot during the Paleocene–Eocene Thermal Maximum. Geology 43:443–446
Yamano H, Sugihara K, Nomura K (2011) Rapid poleward range expansion of tropical reef corals
in response to rising sea surface temperatures. Geophys Res Lett 38:L04601
Yin J, Fursich FT (2008) Dispersal events of Triassic-Jurassic boundary faunas, and paleoenviron-
ment of Tibetan Himalaya. Sci China Ser D Earth Sci 52:1993–2000
Younger JL, Emmerson LM, Miller KJ (2016) The influence of historical climate changes on
Southern Ocean marine predator populations: a comparative analysis. Glob Chang Biol 22:474–
493
Zhao X, Tong J, Yao H et al (2015) Early Triassic trace fossils from the Three Gorges area of South
China: implications for the recovery of benthic ecosystems following the Permian–Triassic
extinction. Palaeogeogr Palaeoecol 429:100–116
Zonneveld J-P, Gingras MK, Beatty TW (2010a) Diverse ichnofossil assemblages following the P–
T mass extinction, Lower Triassic, Alberta and British Columbia, Canada: evidence for shallow
marine refugia on the northwestern coast of Pangaea. PALAIOS 25:368–392
Zonneveld J-P, McNaughton RB, Utting J et al (2010b) Sedimentology and ichnology of the Lower
Triassic Montney Formation in the Pedigree-Ring/Border-Kahntah River area, northwestern
Alberta and northeastern British Columbia. Bull Can Petrol Geol 58:115–140
Training Tomorrow’s Conservation
Paleobiologists
Patricia H. Kelley, Gregory P. Dietl, and Christy C. Visaggi
Abstract Conservation paleobiology (CPB) is committed to the mission of apply-

ing geohistorical records to the conservation and restoration of biodiversity and
ecosystem services. Fulfilling this mission remains challenging because most
conservation paleobiologists have not been trained to translate their science into
management practice and policy. Ongoing discussion among conservation biologists
provides lessons applicable to training tomorrow’s conservation paleobiologists. We
offer six recommendations for more effective training in conservation paleobiology:
(1) integrate CPB into truly cross-disciplinary conservation curricula; (2) promote
a problem-solving and policy-oriented approach to CPB education; (3) implement
hands-on experience in real-world settings; (4) promote informal opportunities for
cross-disciplinary interaction; (5) establish connections among stakeholders; and (6)
modify faculty incentive systems to reward CPB activities. Institutional constraints
make implementation of these recommendations challenging, and development of
CPB programs will involve tradeoffs. Despite these challenges, the approach we
propose will better prepare tomorrow’s conservation paleobiologists to function
effectively in the conservation world.
Keywords Conservation training · Cross-disciplinary curricula · Interpersonal

skills · Problem-solving · Policy-oriented education · Real-world experience
Patricia H. Kelley and Gregory P. Dietl contributed equally.

P. H. Kelley ()
Department of Earth and Ocean Sciences, University of North Carolina Wilmington, Wilmington,
NC, USA
G. P. Dietl ()
C. C. Visaggi
Department of Geosciences, Georgia State University, Atlanta, GA, USA

210 P. H. Kelley et al.
1 Business As Usual Is Not Enough
The past decade has witnessed workshops defining and charting the course of the
emerging field of conservation paleobiology (CPB), sessions and presentations on
CPB-related topics at professional meetings, an expanding technical and popular
literature (Dietl 2016), and even university classes on the new field. Some of us now
consider ourselves conservation paleobiologists and have begun to train students
to follow in our footsteps. But is it enough to follow in our footsteps, or must
tomorrow’s conservation paleobiologists blaze a new path? What does it really mean
to be a conservation paleobiologist?
Obviously, it means more than being a paleobiologist; otherwise, the modifier
would be unnecessary. By adding “conservation” to our moniker, we are committing
to a mission—a value-laden framework for our work (Dietl 2016). This mission-
driven approach is implicit in definitions adopted for CPB. For instance, a 2011
National Science Foundation (NSF)-funded Conservation Paleobiology Workshop
(2012, p. 1; see also Dietl et al. 2015) defined CPB as “the application of geohis-
torical records to the conservation and restoration of biodiversity and ecosystem
services.” Although conservation paleobiologists have done a good job extracting
and interpreting geohistorical records, as exemplified by papers in this volume,
the application to the conservation and restoration of biodiversity and ecosystem
services remains challenging.
Most of us who claim to be conservation paleobiologists typically conduct the
research we would have done anyway; at best, we comment vaguely (usually in the
introduction and conclusions sections of our papers) that the work has implications
for conservation (e.g., “we studied the Pliocene and the Pliocene was warm, so
our results have implications for how biota are likely to respond in a warmer
Anthropocene world”). Often the conservation implications of our work are poorly
articulated and, even if stated, are only read by our fellow conservation paleobi-
ologists. This ineffectiveness at translating our results into management practice
and policy is not surprising because the majority of conservation paleobiologists
are traditionally trained paleontologists who are recruits to the new field from our
own specialized research disciplines within paleontology. As academic scientists,
we have not been trained to interpret geohistorical data in a management context.
Our abilities in management and policy, if any, are self-taught or acquired by trial
and error, or better yet trial and success (see Flessa 2017).
The need for cross-disciplinary (sensu Ciannelli et al. 2014)1 training of con-
servation professionals is recognized widely (e.g., Jacobson 1990; Jacobson and
Robinson 1990; Kainer et al. 2006; Moslemi et al. 2009). The field of CPB is
often touted as being cross-disciplinary (e.g., Conservation Paleobiology Workshop
1
Cross-disciplinary work may include multidisciplinary, interdisciplinary, or transdisciplinary
collaboration (Rosenfield 1992; Ciannelli et al. 2014) in which workers from multiple disciplines
address a common problem with no (multidisciplinary), some (interdisciplinary), or substantial
(transdisciplinary) integration of disciplines.
Training Tomorrow’s Conservation Paleobiologists 211
2012), but few of us were trained in a cross-disciplinary manner and our collabo-
rations with non-paleobiologists (beyond the occasional biologist) are limited. CPB
is dominated by academic scientists “from strongly single-disciplined, specialized
backgrounds” (as described by Jacobson 1990, p. 433, for conservation biologists)
who have little experience outside the university and are ill-prepared to train
students for the current job market in conservation. Consequently, this conventional
emphasis on disciplinary research in CPB training leaves many students poorly
prepared to contribute to the field of conservation. This problem is particularly
acute because, as noted by Noss (1999), the vast majority of jobs for conservation
biologists (and presumably future conservation paleobiologists) are not and will
not be in universities, but in governmental land-managing and regulatory agencies,
conservation groups (e.g., NGOs), consulting firms, research institutes, and industry.
Indeed, Lucas et al. (2017) noted that only 10% of conservation jobs in the current
market are in academia. We cannot assume that our students will pursue research-
based academic careers as we did. This situation is true for the natural sciences in
general; the number of PhD scientists entering academic positions declined and of
those taking positions in business and industry increased markedly (from 19 to 32%)
from 2004 to 2014 (Kurilla 2016, citing data from the NSF). Doctoral students are
trained for research careers, but employers seek candidates with more than research
experience (Campbell et al. 2005; Perez 2005; Blickley et al. 2012; Kurilla 2016).
Fortunately (or perhaps unfortunately), we are not alone in our concerns about the
lack of adequate training for future conservation (paleo)biology practitioners. The
issue of inadequate training of conservation biologists to meet real-world challenges
is long-standing (e.g., Jacobson 1990; Noss 1999; Clark 2001; Perez 2005; Muir
and Schwartz 2009). In this essay, we draw lessons from this ongoing discourse
in conservation biology that are applicable to training tomorrow’s conservation
paleobiologists.
2 A Call to Action
Brewer (2001, p. 1203), in what might be characterized as a “call to action,” asked

“What do we want conservation biologists in the twenty-first century to know and
be able to do?” Similarly, a good place to start in determining how we might
train tomorrow’s conservation paleobiologists is by considering what a conservation
paleobiologist should know and be able to do.
Our students need to be trained to function effectively in the real world of
conservation. Such training should enable students to acquire both the appropriate
scientific knowledge and the ability to apply it. Traditional academic training
focuses on acquisition of knowledge and, especially at the PhD level, development
of discipline-focused research skills. However, we do an abysmal job of preparing
students to function outside narrowly defined disciplinary research topics, even
though the job market in conservation “increasingly requires broadly trained,
versatile professionals” (Noss 1999, p. 118). A decades-long lament in conservation
biology (Jacobson and McDuff 1998, p. 433) is that most academic curricula
“remain rigidly departmental with little opportunity for interdisciplinary training”
(Noss 1999, p. 118); depth of knowledge is emphasized at the graduate level at the
expense of breadth, even though solving complex environmental problems requires
working closely with colleagues from many disciplines (Jacobson 1990, Jacobson
and Robinson 1990, Ciannelli et al. 2014).
Jacobson and McDuff (1998, p. 263) warned that, in conservation biology “we
may, in fact, be training idiot savants—individuals skilled in certain areas : : : but
largely inept in other aspects of the field.” In the case of CPB, our “idiot savants”
may be well trained in the technical paleontological methods and theories applicable
to conservation, but they lack the ability to translate the results of their research into
action. This gap between research and implementation is recognized increasingly
in conservation biology, in which “conservation assessments are rarely translated
into actions that actually conserve nature” (Knight et al. 2008, p. 610; see also
Jarvis et al. 2015). In general, the basic sciences do not value development of
essential interpersonal skills in policy processes (Clark 2001), communications
(e.g., the ability to resolve conflict or build trust in relationships), and leadership,
including the ability to influence others (Jacobson and McDuff 1998). None of
these abilities (or “soft skills”) is viewed traditionally as part of the professional
identity of a paleontologist, and yet they are essential if conservation paleobiologists
are to function effectively in the conservation world. Translating Jacobson and
McDuff’s comments into CPB terminology, a graduate advisor would never send
a paleontology student into the field without training in measuring sections,
describing lithology, and collecting samples, as appropriate to the project. Yet we
are currently sending the first generation of conservation paleobiologists into the
conservation field “to confront diverse stakeholders without any training on how
to assess and address” real-world conservation issues (Jacobson and McDuff 1998,
p. 265). This situation must change if CPB is to have any real-world impact. The
next generation of conservation paleobiologists must have truly cross-disciplinary
training, including training in the human dimensions of conservation issues, if
they are to help solve conservation problems. To achieve this goal, CPB training
must adapt in order to produce graduates with the skillset required to navigate the
conservation landscape.
3 Bridging the Gap
We offer six specific recommendations, none of which is original to us, but all of
which would prepare students better to be effective conservation paleobiologists,
i.e., who can bridge the widening research-implementation gap in conservation
(Knight et al. 2008; Jarvis et al. 2015). Recommendations are curricular, pro-
grammatic, and/or institutional: (1) integrate CPB into truly cross-disciplinary
conservation curricula; (2) promote a problem-solving and policy-oriented approach
to CPB education; (3) implement hands-on experience in real-world settings;
(4) promote informal opportunities for cross-disciplinary interaction; (5) establish

connections among stakeholders; and (6) modify faculty incentive systems to reward
CPB activities.
Recommendation 1
Most universities have established programs relevant to conservation. The Society

for Conservation Biology2 reports nearly 600 academic programs worldwide specif-
ically in conservation biology. Of those, 60 are graduate programs. Such programs
typically are housed within traditional biology departments or wildlife and natural
resources programs (Jacobson 1990). Many schools without formal conservation
biology programs offer programs in environmental studies that provide some of
the same training. Thus, CPB programs need not “reinvent the wheel”; making
connections with existing programs on campus is an ideal way to develop cross-
curricular CPB programs.
Jacobson (1990) highlighted 16 exemplary conservation biology programs,
the majority of which were cross-disciplinary and included non-biology faculty.
She provided a model for cross-disciplinary conservation biology education, in
which biology represented one of five general areas. Additional areas considered
fundamental to conservation biology education included the physical environment
(chemistry, geology, physics, geography), social environment (economics, political
science, sociology, anthropology, philosophy), applied management sciences (e.g.,
wildlife, forestry, fisheries), and the implementational environment (e.g., planning,
education, law, communication). Typically, the highlighted programs require a
few core courses, with electives allowing students to emphasize areas of interest
(Jacobson 1990).
We recommend working with existing programs on campus to design
a cross-disciplinary CPB option or track, which could be housed within a
geology/geoscience degree program, a conservation biology program, or an
environmental studies program as appropriate to the culture of the institution.
The CPB track would include the necessary geology courses (e.g., paleoecology,
stratigraphy, field and/or research methods) but would also benefit from including
courses in such areas as conservation biology, environmental policy, economics,
environmental law, and environmental philosophy. Another option would be to
implement a cross-disciplinary conservation certificate program, rather than a
degree program, as has been done successfully at University of Florida (the
“Tropical Conservation and Development” program3 ; Kainer et al. 2006). Even
if formal implementation of a CPB track within a degree or certificate program
2
https://conbio.org/professional-development/academic-programs.
3
http://www.tcd.ufl.edu/.
is not possible, students with CPB interests should be advised to take courses
within those areas to obtain the cross-disciplinary training needed to function as
conservation professionals (see also Recommendation 2).
Recommendation 2
Education of future conservation paleobiologists should not only be cross dis-

ciplinary but also policy oriented. Jacobson and McDuff (1998) and Muir and
Schwartz (2009) argued for inclusion of more of the “human dimension” in
conservation training. A survey of conservation alumni employed in the academic
and non-academic sectors revealed that, whereas academicians emphasized research
skills, the practitioners focused on decision-making, policy, and management skills
(Muir and Schwartz 2009). Clark (2001) also recognized that conservation biology
curricula need a more policy-oriented approach. Likewise, CPB will only fulfill the
first part of its name if tomorrow’s conservation paleobiologists receive training
in policy making. Critical thinking skills beyond those involved in conducting
scientific research are needed; students need to be able to apply these skills to
policy problems (Clark 2001). For instance, CPB research can inform decision-
making on a host of policy issues, including establishing conservation priorities
(e.g., regarding populations, species, regions), selecting biological reserves, and
restoring ecosystem function by substituting extant for extinct species in “rewilding”
projects (Conservation Paleobiology Workshop 2012). However, for the results of
CPB research to be used most effectively, future practitioners need to understand
not only the technicalities of the decision-making process and the political context
(Flessa 2017) but also the psychology of the human interactions involved. The
results of CPB research are rarely translatable directly into policy. In addition, the
best solution to a conservation issue from a scientific standpoint may not be the best
solution from a societal perspective. Goals and values of policy makers and their
constituents play a role that may conflict with recommendations based solely on
scientific principles (Kroll 2007; Dietl and Flessa 2018). The scientific principles,
however, must be conveyed in a manner accessible to policy makers and to the
general public (Brewer 2001), including via appropriate communication through the
media (Flessa 2017), e.g., news media, social media, and public-accessible videos.
Policy-oriented training might be accomplished in formal course work. Most
institutions offer a wide range of courses that could contribute to a policy-oriented
approach to conservation education. For instance, at the University of North Car-
olina Wilmington4 (UNCW), relevant graduate-level courses include “Coastal and
Environmental Science and Policy,” “Decision Making and Negotiation,” “Strategic
Communication: Analysis of Persuasion and Cultural Principles and Techniques,”
“Advanced Environmental Law and Policy,” “Conservation and Culture,” “Envi-
4
http://www.uncw.edu/gradschool/index.html.
ronmental Management,” and “Sociology of Organizations,” among many others.

Judicious choice of electives by future conservation paleobiologists could help fill a
major gap in their education.
Recommendation 3
Education needs to go beyond course work to provide real-world experiences

for future conservation paleobiologists. Meffe (1998, p. 260) noted that “Without
practical experience [students] could provide only theoretical ‘knowledge’ of an
experience-based activity. Medical schools train their students in real hospitals with
actual patients. Auto mechanics learn on real transmissions and clutches. Why
should conservation biologists not be similarly trained in the actual complexities
and difficulties of real-world issues?” We pose the same question for conservation
paleobiologists.
Case studies are used by some conservation programs (Jacobson and Robinson
1990; Clark 2001) to provide real-world context. By giving students the opportunity
to participate in devising solutions to a particular environmental issue and inter-
acting with a stakeholder (e.g., Ciannelli et al. 2014), this approach can provide
practical experience in the real world (i.e., experiential learning). To a degree,
service-learning activities embedded within content courses (Hansen and Fortner
2016) also can provide such experiences. Conservation-related service learning,
involving hands-on community service, allows students to engage in local issues.
For instance, at Indiana UniversityPurdue University Indianapolis5 (IUPUI),
students undertake environmental stewardship activities at sites in central Indi-
ana, including community work days focused on ecosystem restoration (Tedesco
and Salazar 2006). Such place-based education (Gruenewald 2003; Sobel 2004)
capitalizes on the personal attachment that learners have for their surroundings
to teach concepts across disciplines; it increases student engagement, academic
success, and retention (Tedesco and Salazar 2006; Semken and Butler Freeman
2008; Semken et al. 2009) and provides skills valued by employers, including
communication, project management, collaboration, and technical skills (Shriberg
and Harris 2012). Such activities would provide useful additions to CPB-focused
courses at the undergraduate or graduate level.
The Smithsonian-Mason School of Conservation,6 in which students can enroll
for a semester (similar to study abroad enrollment), combines course work with
a practicum experience at the Smithsonian Conservation Biology Institute. Other
approaches incorporate policy and management components into field or research
experiences (Ort et al. 2006; Kelley and Dietl 2012). For example, the San Juan
River field course operated by the Environmental Sciences Program at Northern
5
http://csl.iupui.edu/.
6
http://smconservation.gmu.edu/.
Arizona University7 combines classroom work with a cross-disciplinary (geology,

biology, and chemistry) field experience with in-depth discussion of related policy
and land management issues (Ort et al. 2006). In the NSF-funded “Research
Experiences for Undergraduates in Biodiversity Conservation” summer program8
at UNCW (Kelley and Dietl 2012) in which the three of us were involved,
geology, biology, archeology, and environmental science students worked in cross-
disciplinary teams to compare the modern marine ecosystem in the Carolinas with
communities of the past, using the Plio-Pleistocene fossil record, the archeological
record of shell middens, and live-dead analysis in modern environments. We
discussed current environmental issues, their policy implications, and how conserva-
tion paleobiology might contribute to their resolution, for instance, by determining
baselines and the natural range of variability for use in restoration ecology.
Although such experiences are valuable, more immersive experiences in real-
world situations are needed. The conservation education literature emphasizes that
internships are necessary to provide job experience outside the university (Jacobson
1990; Jacobson and Robinson 1990; Jacobson and McDuff 1998; Soulé and Press
1998; White et al. 2000; Clark 2001; Kainer et al. 2006; Kroll 2007; Moslemi et
al. 2009; see also Conservation Paleobiology Workshop 2012 and Flessa 2017).
Undergraduate and graduate students who participate in summer or semester-long
internships with government agencies (e.g., the National Parks Service Geoscientist-
in-the-Park internship), industry, and mainstream environmental NGOs (e.g., The
Nature Conservancy, World Wildlife Fund, Oceana) are far better prepared to
function in the conservation profession (Conservation Paleobiology Workshop
2012). Internships foster development of skills valued by employers, including
teamwork, problem solving, communication, fundraising and project management,
policy analysis, and negotiation (Blickley et al. 2012). They allow students to
observe the decision-making process (Kroll 2007) and interact with those making
the decisions as well as with those affected by the decisions (Clark 2001). Therefore,
we feel strongly that an internship or similar sustained, immersive, real-world
experience is essential to a future conservation paleobiologist’s training, even if the
internship does not have an explicit paleontology dimension.
Recommendation 4
We recognize that institutional and other barriers exist to implementing our previous
three recommendations (see below), but less formalized program elements can also
contribute to CPB training. The Tropical Conservation and Development certificate
program at University of Florida includes core courses, but much of the training
occurs in “alternative learning spaces” (Kainer et al. 2006, p. 9). In addition to
7
https://nau.edu/cefns/natsci/seses/sce/.
8
http://www.nsf.gov/awardsearch/showAward?AWD_ID=0755109&HistoricalAwards=false.
internships, these activities include workshops, seminars, retreats, and visits by

professionals from outside the university. Many of these program elements are
shared by the Biogeochemistry and Environmental Biocomplexity Program9 at
Cornell University (Moslemi et al. 2009, p. 516). Such activities are not difficult
to implement and can be done within the framework of existing university structure.
When students go beyond the minimum expectations of course and
thesis/dissertation work, they acquire non-academic skills and experiences that
are valued in the job market. Thus, involving students in organizing such
seminars and workshops allows them to develop and demonstrate skills in project
management, networking, and communication that can prepare them for non-
academic conservation careers (Blickley et al. 2012). Seminar series that cross
departmental boundaries foster sharing of knowledge among disciplines and cross-
disciplinary collaboration. Such venues also can provide opportunities for students
to interact with local conservation leaders. Workshops, short courses, and training
academies (Kainer et al. 2006; Moslemi et al. 2009; Ciannelli et al. 2014) have also
been used to foster interaction across disciplines and students’ abilities to work in
cross-disciplinary teams. Less formal activities, such as social events (e.g., taking
an outside speaker for drinks or dinner), retreats, and reading groups, can also
promote a feeling of community among students involved in conservation programs
(Moslemi et al. 2009). For instance, at Cornell University, CREST10 —the “Cornell
Roundtable on Environmental Studies Topics”—brings together faculty, graduate
students, and community members from the humanities, arts, social sciences,
natural and physical sciences to discuss environmental issues.
Recommendation 5
Campbell et al. (2005) have argued that improvement of doctoral education will
require establishing connections among stakeholders, including those organizations
that fund and hire doctoral students. Students also need to be made aware of the
range of employment opportunities outside academia and the skill sets needed
in such careers; seminar series that include speakers from outside academia, or
meetings with alumni, are useful in this regard (Campbell et al. 2005). Panels
composed of alumni have been used successfully at our institutions to provide such
guidance.
Partnerships with practitioners can be enhanced by including them as outside
members on thesis/dissertation committees (Jacobson and McDuff 1998; Blickley
et al. 2012), where they can provide input regarding real-world implications of
CPB research. When students address research questions of interest to an external
organization, partnership with the organization can “provide practical experience
9
http://www.geo.cornell.edu/biogeo/beb/.
10
https://blogs.cornell.edu/crest/.
with the institutional structure, culture, and work processes” (Blickley et al. 2012, p.
32) that will better prepare the student for a career in conservation, although from the
perspective of the NGO or government agency the product of such collaborations is
more likely to be solving a conservation problem rather than publishing a scientific
paper (Boyer et al. 2017). Including practitioners as adjunct faculty or on program
advisory boards or visiting committees would also strengthen ties between academic
and non-academic stakeholders (Jacobson and McDuff 1998). Indeed, the process
of establishing (and subsequently reviewing) a formal CPB program would require
input from practitioners.
Brewer (2002) stressed the importance of partnerships with the community, for
instance, with local schools or in citizen science. Environmental service learning
at IUPUI, described above, is made possible by collaboration with 30 community
partners, including the Indianapolis Department of Parks and Recreation Land
Stewardship Office (Tedesco and Salazar 2006). Such partnerships are facilitated
by the campus-wide IUPUI Center for Service and Learning, with a mission
to promote civic engagement. Conservation paleobiology students could benefit
from community partnerships as exemplified by recent efforts in community-
based planning and research initiatives. In such projects, a researcher works with
the community (e.g., NGOs, neighborhood councils, local schools) and students
(undergraduate and graduate students, and often K-12) to set goals, collect data,
and discuss implications; perspectives of all participants are included and valued.
Students involved in such projects (e.g., mapping invasive plant species in an urban
forest; Hawthorne et al. 2015) gain experience in teamwork and policy processes in
a real-world context, addressing conservation issues that are important to the local
community.
Recommendation 6
Implementing any of the previous recommendations will require a commitment of

time and energy on the part of CPB faculty. Likewise, the mentoring necessary to
prepare students for careers (Campbell et al. 2005) in CPB will require effort beyond
that normally involved in advising graduate students within a single discipline.
Thus, our final recommendation is that faculty incentive systems will need to
be modified to promote cooperation among departments and to reward individual
faculty for real-world problem solving at local, national, and international levels.
Such cultural changes cannot be accomplished by the small number of CPB faculty
alone; faculty in other disciplines who are engaged in applied community work will
need to be enlisted as allies in influencing institutional culture.
The feasibility of implementing this strategy will vary among institutions.
University structure can either hinder or facilitate the work we have described.
A precedent for more applied community or regional engagement exists in the
mission of some land-grant institutions. For instance, Florida Gulf Coast University,
with a mission that “promotes and practices environmental sustainability, : : :
nurtures community partnerships, values public service, [and] encourages civic

responsibility,”11 is actively involved in environmental management and restoration
in southwest Florida. Conservation paleobiology plays an essential role in the
university’s involvement in these activities, with faculty working across disciplines
and with government agencies (e.g., Army Corps of Engineers, South Florida Water
Management District, USGS, US Fish & Wildlife, Everglades National Park) and
NGOs, including the Conservancy of Southwest Florida, the Audubon Society,
Caloosahatchee River Watch, and the Southwest Florida Watershed Council (Boyer
et al. 2017).
Some institutions also already have cross-disciplinary institutes or centers that
could be the locus of CPB activities. The existence of such units within a university
means that systems are in place to facilitate communication and cooperation among
disciplines. For instance, the Atkinson Center for a Sustainable Future12 is a cross-
disciplinary center at Cornell University (of which Dietl is a faculty fellow) focusing
on economic development, energy, and the environment; it involves faculty from
the natural sciences, social sciences, agriculture, medicine, engineering, arts, and
humanities. The Center supports students and postdoctoral scholars through grants
and fellowships (including some current CPB projects) and provides internships at
environmental organizations as well as engagement with policy makers and other
external stakeholders. For instance, the NatureNet Science Fellows Program13 is
offered in collaboration with the Nature Conservancy, and the Atkinson Postdoctoral
Fellows in Sustainability14 must have a co-advisor external to the university (e.g.,
Environmental Defense Fund). The Center’s activities thus align well with recom-
mendations made by the Conservation Paleobiology Workshop (2012) concerning
partnerships with stakeholders, postdoctoral opportunities, and integration with the
social sciences. When such cross-disciplinary units already exist within a university,
involvement by faculty in cross-disciplinary work is more likely to be rewarded by
the institution.
Collaborative research tends to be more costly in terms of time, effort, and
infrastructure (National Research Council 2005). The availability of funding thus
has been a major factor in developing cross-disciplinary conservation programs;
the Pew Charitable Trust’s initiative on “Integrated Approaches to Training in
Conservation and Sustainable Development” provided a significant impetus to
such programs (Jacobson 1990). Integrative Graduate Education and Research
Traineeship (IGERT) grants from the National Science Foundation (now the
NSF Research Traineeship program15 ) have supported successful interdisciplinary
training programs (National Research Council 2005; Morse et al. 2007; Moslemi
et al. 2009); the Tropical Conservation and Development program at University
11
http://www.fgcu.edu/info/mission.asp.
12
https://www.acsf.cornell.edu/index.php.
13
http://www.nature.org/science-in-action/naturenet-science-fellowship.xml.
14
http://www.atkinson.cornell.edu/grants/postdoc/.
15
https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=505015.
of Florida has been supported by grants from the MacArthur Foundation, Ford
Foundation, and state of Florida, among others (Kainer et al. 2006). Likewise,
service learning at IUPUI is supported by both corporate and civic sponsors
(Tedesco and Salazar 2006). Research-funding agencies are also making cross-
disciplinary work more attractive financially (National Research Council 2005). For
instance, the “Dynamics of Coupled Natural and Human Systems” program16 at the
NSF is cross-directorate (Biological Sciences; Geosciences; and Social, Behavioral,
and Economic Sciences), supporting research on the interactions of human and
natural systems by teams with expertise in the natural, social and behavioral
sciences. In our experience, university administrators tend to be more enthusiastic
about initiatives that can attract external funding. Thus, continuing to work with
agencies such as the NSF (Conservation Paleobiology Workshop 2012) to enhance
funding opportunities for CPB work should remain a priority.
4 Okay, But : : :
We recognize that the approach we have recommended is not without challenges.

Our recommendations that CPB training be cross disciplinary, with an emphasis on
real-world experience, raise a variety of issues.
Cross-disciplinary programs suffer from institutional constraints (Jacobson and
Robinson 1990). Tenure and promotion systems, as well as vehicles for program
evaluation, occur within a departmental context. Rewards and incentives for faculty
to improve performance, to the extent that they exist, are distributed primarily at
the departmental level, and reward systems tend to value products of individual
scholarship more than team efforts (National Research Council 2005). Cross-
disciplinary research, and faculty participation in cross-disciplinary programs, can
be difficult to evaluate (Klein 2008); disciplines vary in their paradigms, terminol-
ogy, research methods, philosophical approaches (Jacobson and Robinson 1990;
Morse et al. 2007), and opportunities for publication and funding. Most faculty
are uncomfortable venturing beyond a narrow range of research topics (whether
it be taphonomy or brachiopod systematics or drilling predation). We prefer to
stay within our comfort zone and avoid taking risks, especially when the rewards
are uncertain. Soulé and Press (1998) also noted that applied research, especially
when used for advocacy, is often seen as subjective or biased (see also Flessa 2017;
Dietl and Flessa 2018). Such work may suffer when subject to peer review for
publication, funding, tenure, and promotion. For instance, candidates for tenure and
promotion may find that their institution classifies such applied research as service
or instruction (if students are involved) rather than as a research contribution.
Departments are often placed in adversarial roles with one another, as they
compete for limited resources (Jacobson and Robinson 1990). With state support
16
http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=13681.
for higher education declining, at public universities even modest departmental

contributions to cross-disciplinary initiatives (e.g., to bring in a speaker relevant to
multiple departments) require a sacrifice of resources. Even cooperation by offering
cross-departmental co-taught courses can be problematic in terms of sharing
costs, faculty salaries, and allocation of credit for the students enrolled, despite
general acknowledgment that students may benefit from observing how faculty with
different expertise work together (e.g., Kuban 2016). Joint appointments of faculty
in more than one department can help alleviate such problems but may be difficult to
negotiate. Thus, departments may have few incentives to cooperate with one another
(Kelley, personal experience as department chair). Programs that depend on the
good will of participating departments for resources and faculty may suffer (thus
Soulé and Press 1998 argued for autonomy of environmental studies programs).
In addition, allocation of resources is often based on generation of student credit
hours by a department. Courses that do not attract a minimum number of students
are threatened with cancellation. As a result, department chairs may discourage
faculty from advising students to take courses outside their home department.
Even in the absence of these institutional barriers to cross-disciplinary education,
attitudinal barriers may inhibit implementation of our recommendations. “Hard”
sciences may be seen as superior in some way to “softer” disciplines (social
sciences, humanities), which may discourage natural science faculty from advising
students to take courses representing the “human dimension” of conservation.
In addition, philosophical disagreements occur regarding the appropriate balance
between depth and breadth of training—a common concern in conservation biology
(Jacobson 1990; Jacobson and Robinson 1990; Kainer et al. 2006; Morse et al. 2007;
Moslemi et al. 2009). Soulé and Press (1998) argued that environmental studies
programs are plagued by a problem of “hyper-diverse shallow curricula” (p. 397),
resulting in “multidisciplinary illiteracy” (p. 402) and a lack of quality and rigor.
Thus, a key question for CPB is whether breadth of training can be increased without
sacrificing depth of training in the core discipline of paleontology.
Undoubtedly, tradeoffs will be involved in developing programs in CPB (see
Muir and Schwartz 2009). Conservation paleobiologists need to be trained as
paleobiologists; the core science, including training in how to do paleobiological
research, must remain. In addition, extending time to degree is undesirable (Kainer
et al. 2006; Blickley et al. 2012). Time to degree is one of the metrics used to
evaluate program productivity (Kelley, personal experience as department chair);
programs that take longer to complete are at a disadvantage in terms of recruitment
and cost (to students and to the institution).
Muir and Schwartz (2009) recognized that training cannot be added to conserva-
tion biology requirements without cutting other aspects of training and concluded
that not all skills need to be taught to all students. We may need to rethink current
training programs for CPB and not automatically funnel our students into PhD
programs focused only on research skills. The need for conservation paleobiologists
with PhDs will remain, in order to train students, but we should also be preparing
students to enter the workforce at the MS level. Indeed, MS programs may be more
amenable to implementing such aspects as internships. For example, at UNCW
the MS Geoscience degree17 has both a thesis and a non-thesis option. Students
in the non-thesis option have the opportunity to do a final project, an internship,
or both, and the non-thesis option requires 33 credit hours as compared to 30
credit hours in the thesis option. Similarly, at Georgia State University the MS
in Geosciences18 non-thesis option usually incorporates a project, often with a
partner organization, and many are related to sustainability or conservation. The
greater flexibility of the non-thesis option could be an excellent venue for taking
a broader range of courses (our Recommendations 1 and 2) and providing real-
world experience (Recommendation 3). Such students may not become experts in
taphonomy, brachiopod systematics, or even drilling predation, but they would be
better prepared for conservation careers. Practice-oriented degree programs (e.g.,
the Master of Professional Studies at Cornell,19 which requires a problem-solving
project) serve a similar function.
5 In the Meantime : : :
Programmatic changes (Recommendations 1, 2, and 3) and modifications to insti-

tutional attitudes and cultures (Recommendation 6) will require time to implement.
Although informal program elements described in Recommendations 4 and 5 can be
introduced more quickly, is there more that students currently in the CPB pipeline
can do to develop the skills needed for a conservation career?
Current students who aspire to a career in conservation paleobiology would do
well to follow the recommendations of Blickley et al. (2012). Their analysis of
advertisements for conservation science positions and interviews of conservation
professionals revealed that skills desired for new hires vary from governmental to
nonprofit to the private job sector (see also Lucas et al. 2017). Government and
nonprofit sectors showed preference for interpersonal, networking, oral communi-
cation, program leadership, and project-management skills; nonprofits also valued
fundraising skills. Technical, disciplinary, written communication, and field skills
were priorities for the private sector. Because student time is limited for develop-
ment of skills beyond those needed for academic success in the discipline, a student
who is able to identify a particular career goal can target acquisition of the relevant
skills. CPB students should assess strategically whether their program of study can
provide these skills. If the answer is “no,” students should take charge of their future
by making their own opportunities (Perez 2005; Blickley et al. 2012): mentoring
undergraduates in research, planning and participating in K-12 outreach activities,
seeking an internship or volunteering with an organization in the desired job
sector, asking a practitioner from an outside organization to be a thesis/dissertation
17
http://uncw.edu//msgeoscience/index.html.
18
http://geosciences.gsu.edu/grad-programs/m-s-degree-in-geosciences/.
19
https://cals.cornell.edu/mps.
committee member, meeting with alumni, planning departmental events, organizing

a seminar or asking a faculty member to offer a course or an independent study
on professional skills (e.g., Kelley co-developed a “Professional Skills Practicum”
course at the University of North Dakota; the Network of Conservation Educators
and Practitioners at the American Museum of Natural History20 offers open-access
teaching modules that include communication skills). The possibilities are limited
only by student energy, time, and creativity; advisors can suggest opportunities and
help students assess feasibility of ideas and achieve a balance between completing
program requirements and acquiring non-academic skills.
6 A Bright Future
We are convinced that, if we are clever, conservation paleobiologists can find

ways to collaborate with existing cross-disciplinary programs in the conservation
sciences. Working within the framework of existing cross-disciplinary units holds
particular promise. Admittedly this approach is not without obstacles, including
convincing other conservation workers of the applicability of the fossil record to
biodiversity conservation (Durham and Dietl 2015; Flessa 2017; Smith et al. 2018).
Regardless of issues in overcoming institutional hurdles to formalizing CPB pro-
grams, implementing any of the informal approaches proposed in Recommendations
4 and 5 will better prepare tomorrow’s conservation paleobiologists to function
effectively in the conservation world.
We stand at the brink of exciting changes and progress in CPB education. If we
continue along our present trajectory, CPB will be increasingly ineffective in its
primary mission. However, if we help our students blaze a new path, the payoff is
clear: we will do a better job of training our CPB students, which will produce a
more effective workforce doing conservation on the ground.
Acknowledgments We thank Carrie Tyler and Chris Schneider for the opportunity to contribute
to this volume. Karl Flessa and Frank Forcino provided helpful reviews of the manuscript.
References
Blickley JL, Deiner K, Garbach K et al (2012) Graduate student’s guide to necessary skills for
nonacademic conservation careers. Conserv Biol 27:24–34
Boyer AG, Brenner M, Burney D et al (2017) Conservation paleobiology roundtable: from promise
to application. In: Dietl GP, Flessa KW (eds) Conservation paleobiology: science and practice.
University of Chicago Press, Chicago, pp 291–302
20
http://www.amnh.org/our-research/center-for-biodiversity-conservation/capacity-development/
network-of-conservation-educators-and-practitioners-ncep/.
Brewer C (2001) Cultivating conservation literacy: “trickle-down” education is not enough.

Conserv Biol 15:1203–1205
Brewer C (2002) Outreach and partnership programs for conservation education where endangered
species conservation and research occur. Conserv Biol 16:4–6
Campbell SP, Fuller AK, Patrick DAG (2005) Looking beyond research in doctoral education.
Front Ecol Environ 3:153–160
Ciannelli L, Hunsicker M, Beaudreau A et al (2014) Transdisciplinary graduate educa-
tion in marine resource science and management. ICES J Mar Sci 71(5):1047–1051.
https://doi.org/10.1093/icesjms/fsu067
Clark TW (2001) Developing policy-oriented curricula for conservation biology: professional and
leadership education in the public interest. Conserv Biol 15:31–39
Conservation Paleobiology Workshop (2012) Conservation paleobiology: opportunities for the
earth sciences. Report to the Division of Earth Sciences, National Science Foundation.
Dietl GP, Kidwell SM, Burney DA et al (2015) Conservation paleobiology: leveraging knowledge
Durham SR, Dietl GP (2015) Perspectives on geohistorical data among oyster restoration
professionals in the United States. J Shellfish Res 34:227–239
pp 283–289
Gruenewald DA (2003) Foundations of place: a multidisciplinary framework for place-conscious
education. Am Educ Res J 40:619–654
Hansen SS, Fortner S (2016) Geoscience service-learning literature themes. Document prepared
for the National Academy of Sciences (NAS) Service-Learning in Undergraduate Geo-
sciences Workshop, 20–21 April 2016. http://sites.nationalacademies.org/cs/groups/dbassesite/
documents/webpage/dbasse_171831.pdf
Hawthorne TL, Elmore V, Strong A et al (2015) Mapping non-native invasive species and
accessibility in an urban forest: a case study of participatory mapping and citizen science in
Atlanta, Georgia. Appl Geogr 56:187–198
Jacobson SK (1990) Graduate education in conservation biology. Conserv Biol 4:431–440
Jacobson SK, McDuff MD (1998) Training idiot savants: the lack of human dimension in
conservation biology. Conserv Biol 12:263–267
Jacobson SK, Robinson JG (1990) Training the new conservationist: cross-disciplinary education
in the 1990s. Environ Conserv 17:319–327
Jarvis RM, Borrelle SB, Breen BB et al (2015) Conservation, mismatch and the research-
implementation gap. Pac Conserv Biol 21:105–107
Kainer KA, Schmink M, Covert H et al (2006) A graduate education framework for tropical
conservation and development. Conserv Biol 20:3–13
Kelley PH, Dietl GP (2012) The “Research Experiences for Undergraduates in Biodiversity
Conservation” program: training the first generation of conservation paleobiologists. Geol Soc
Am Abstr Programs 44:76
Klein JT (2008) Evaluation of interdisciplinary and transdisciplinary research: a literature review.
Am J Prev Med 35:S116–S123
Knight AT, Cowling RM, Rouget M et al (2008) Knowing but not doing: selecting priority
conservation areas and the research–implementation gap. Conserv Biol 22:610–617
Kroll AJ (2007) Integrating professional skills in wildlife student education. J Wildl Manag
71:226–230
Kuban A (2016) Transdisciplinary collaboration: an introduction. In The Trenches 6:1–2
Kurilla B (2016) The evolution of the 21st-century scientist. Am Sci Macroscope. August 12
Lucas J, Gora E, Alonso A (2017) A view of the global conservation job market and how to succeed
in it. Conserv Biol 31:1223–1231
Meffe GK (1998) Softening the boundaries. Conserv Biol 12:259–260
Morse W, Nielsen-Pincus M, Force J et al (2007) Bridges and barriers to developing and conducting
interdisciplinary graduate-student team research. Ecol Soc 12:1–14
Moslemi JM, Capps KA, Johnson MS et al (2009) Training tomorrow’s problem solvers: an
integrative approach to graduate education. Bioscience 59:514–521
Muir MJ, Schwartz MW (2009) Academic research training for a nonacademic workplace: a case
study of graduate student alumni who work in conservation. Conserv Biol 23:1357–1368
National Research Council (2005) The geological record of ecological dynamics: understanding
the biotic effects of future environmental change. National Academy Press, Washington DC
Noss R (1999) Is there a special conservation biology? Ecography 22:113–122
Ort MH, Anderson DE, Ostergren DM (2006) Integrating policy and land management issues into
a natural sciences education: teaching environmental sciences on the lower San Juan River,
Utah. J Geosci Educ 54:116–122
Perez HE (2005) What students can do to improve graduate education in conservation biology.
Conserv Biol 19:2033–2035
Rosenfield P (1992) The potential of transdisciplinary research for sustaining and extending
linkages between the health and social sciences. Soc Sci Med 35:1343–1357
Semken S, Butler Freeman C (2008) Sense of place in the practice and assessment of place-based
science teaching. Sci Educ 92:1042–1057
Semken S, Butler Freeman C, Watts NB et al (2009) Factors that influence sense of place as a
learning outcome and assessment measure of place-based geoscience teaching. Electron J Sci
Educ 13:1–25
Shriberg M, Harris K (2012) Building sustainability change management and leadership skills in
students: lessons learned from “Sustainability and the Campus” at the University of Michigan.
J Environ Stud Sci 2:154–164
Sobel D (2004) Place-based education. Connecting classrooms and communities. The Orion
Society, Great Barrington
Soulé ME, Press D (1998) What is environmental studies? Bioscience 48:397–405
Tedesco LP, Salazar KA (2006) Using environmental service-learning in an urban environment to
address water quality issues. J Geosci Educ 54:123
White R, Fleischner TL, Trombulak SC (2000) The status of undergraduate education in
conservation biology. Education Committee of the Society for Conservation Biology Report
A Conceptual Map of Conservation
Paleobiology: Visualizing a Discipline
Carrie L. Tyler
Abstract Disciplinary boundaries frame the basic questions and central issues of
research, providing the context for the evolution of prevailing theories or paradigm
shifts. This chapter aims to outline the development and scope of conservation
paleobiology using bibliometrics. Publication records relating to research on con-
servation paleobiology were downloaded from Web of Science to generate two
datasets, one aimed at producing a more conservative representation of conservation
paleobiology, and the other more expansive. Bibliographic maps were created to
provide insight into the development and structure of the discipline for both charac-
terizations of conservation paleobiology research (conservative versus expansive).
Bibliographic maps indicated that individual researchers working on conservation
paleobiology specialize in several fields. Regardless of how conservation paleobi-
ology is defined, research involving both paleontology and conservation appears
to be highly multidisciplinary, including at least three main research domains
broadly categorized as: (1) environmental history and conservation archeobiology,
(2) genetics and evolutionary biology, and (3) ecology. Furthermore, paleontological
publications did not form a distinct cluster, but rather were integrated within
conservation science. This supports the proposition that, in practice, conservation
paleobiology is a field of study within conservation science, and not a sub-discipline
within paleontology. Analyses also revealed emerging research fronts in several
topics and confirmed the need for long-term data that pre-dates human activities.
Keywords Science mapping · Conservation science · Paleoecology ·

Bibliometrics · Discipline structure
C. L. Tyler ()
Department of Geology and Environmental Earth Science, Miami University, Oxford, OH, USA

228 C. L. Tyler
1 Determining the Current State and Structure

of Conservation Paleobiology
Conservation paleobiology is a young and emerging discipline, rapidly developing

in response to the urgent need for long-term data in conservation science, in the face
of climate change and biodiversity loss (Smith et al. 2018). However, conservation
paleobiology has yet to rally together to formalize goals, or to establish whether
the goals of conservation paleobiology are consistent with, or independent of those
of conservation science (expanding our understanding of the role of biodiversity in
ecosystem health and functioning, and the conservation of biodiversity). Meine et al.
(2006) asked “Is [the field of conservation biology] performing its core function—
providing reliable and useful scientific information on biological diversity and its
conservation—in the most effective manner possible? Is that information making a
difference on the ground?”. Given that the Earth system is facing unprecedented
environmental pressures and reduction in biodiversity, it is worth considering how
conservation paleobiology can similarly aim to make the most difference. The
creation and organization of a common conceptual framework is vital to identify
central issues within the field, particularly in young disciplines (Torraco 2005;
van der Have and Rubalcaba 2016), as disciplinary boundaries frame the basic
questions and central issues of research, providing the context for the evolution
of prevailing theories, or paradigm shifts. Given the complexity and size of the
scientific literature, personal knowledge and experience are no longer sufficient for
decision-making, or understanding and identifying such research trends.
Scientific contributions form a dynamic and self-organizing system of knowl-
edge, and research subfields share a common knowledge base in the form of article
references. A research field can, therefore, be defined as a network of publications
covering a set of research questions and methodologies, referring to overlapping
literature (Van Den Besselaar and Heimeriks 2006). Journals with the same research
fields are expected to have aggregated citation patterns, and inter-journal citations
can be used to map journals with similar citing patterns (Van Den Besselaar and
Heimeriks 2006). These patterns can be analyzed to produce a visualization of the
bibliometric landscape of the field. Through network modeling and visualization,
the intellectual landscape can be explored to reveal current discipline structure
and dynamics, by identifying prominent scholars, seminal articles, collaborators,
and prevalent research areas. In other words, depicting the current set of questions
that the scientific community is addressing. Emergent research trends and frontiers
can be identified, or pre-eminent individuals and publications in these cutting-
edge research areas. Bibliometrics can be used to assess such shifts in research
focus using citation data (Small 1977), and networks of citations can be traced to
explore the history and evolution of article chains on particular topics. Bibliomet-
rics is the quantitative analysis of publications and their citations and employed
globally to quantitatively evaluate research performance by scientists, university
administrators, government labs, funding agencies (including the National Science
Foundation), and policymakers. These analyses can be advantageous, gathering
A Conceptual Map of Conservation Paleobiology: Visualizing a Discipline 229
objective information necessary for decision-making, and providing a comprehen-

sive perspective on research activity, as citation data provide an objective and
quantitative indicator for evaluating research performance. Weighted measures, such
as papers per researcher or citations per paper, remove characteristics such as the
place of production, or reputation, that can color human perceptions of quality. For
example, when thinking of “the best,” it is not hard to think automatically of the
biggest producers such as individuals, labs, and universities. But these locations may
not be the source of the most impactful work. Bibliometric analyses can therefore
be used to identify top performers, balancing human perceptions of reputation.
Here bibliometrics was employed to map conservation paleobiology, in an effort
to characterize the structure and dynamics of the discipline.
2 Mapping a Discipline
Creating a science map is the first step towards exploring and understanding
scientific frontiers in conservation paleobiology. Here bibliometric networks were
constructed to create graphic representations of conservation paleobiology to:
(1) conceptualize intellectual structure and dynamics and (2) visualize scientific
advancement and emerging research topics.
Bibliometric data were downloaded from Web of Science (WoS) which contains
over 90 million records covering a wide range of natural science disciplines as far
back as the 1900s and includes thorough coverage of records within the fields of
conservation and paleontology. Despite known inaccuracies with assigning articles
to WoS categories, the overall network structures observed when using WoS are
robust to changes in classifications, degree of aggregation using journals rather than
subject categories, and over time periods studied herein (Rafols and Leydesdorff
2009). WoS also allows expansion of searches beyond the Web of Science™ Core
Collection’s nine indexes, to simultaneously search 15 different content sets, by
employing the “All Databases” search option. Although the latter does not yield
data for all record fields, excluding several types of bibliometric analyses. WoS
was searched by “Topic,” which searches the title, abstract, author keywords, and
®
Keywords Plus fields within a record. This ensured that the searched phrases
are likely to be a significant component of the publication and not merely briefly
mentioned in the body of the text. To conservatively capture records fulfilling
the narrowest definition of conservation paleobiology, the words “conservation
paleobiology” were searched in “All Databases.” However, the search term “paleo-
biology” may not accurately capture the full breadth of the discipline, given that
conservation paleobiology includes studies using many types of paleontological
data, such as paleogeography or paleoecology. Therefore, to more broadly visualize
the conservation paleobiology literature as a whole, a second search was conducted
using the WoS “Core Collection” and the terms “conservation paleo*.” The addition
of a wildcard following “paleo” expanded the topic search to include all records
using variations with the prefix “paleo.” This increased inclusiveness, capturing any
230 C. L. Tyler
studies combining research on paleontology and conservation, but not specifically

defined using the term “paleobiology.” Use of the “Core Collection” searched
records in the nine most relevant databases and expanded available record fields,
facilitating additional bibliometric analyses.
VOSViewer was used to construct and visualize bibliometric networks
(www.vosviewer.com; Van Eck and Waltman 2010, 2014), which uses an algorithm
analogous to multidimensional scaling to position data in a multidimensional vector
space and minimize stress (Van Eck et al. 2010; Leydesdorff and Rafols 2012).
Data downloaded from citation databases such as WoS can be input directly into
VOSViewer to construct relationships between citations, keywords, documents,
and authors in the form of a science map (Boyack and Klavans 2010; Small
et al. 2014). Two types of mapping approaches were applied here: visualizations
based on bibliographic data (co-authorship and co-citation) and text data (co-
occurrence). Highly connected portions within networks, or modules, were grouped
by VOSViewer. Modularity is assigned based on the degree to which nodes in the
network can be divided into groups. Nodes within the same group are connected
more densely than nodes between different groups. Modularity can thus be used
to identify the integration and strength of relationships between groups of authors,
documents, topics, and fields of research.
Here bibliometric networks were constructed using a variety of approaches,
relying on different units of analyses necessary for the examination of both
discipline structure and dynamics and emerging research areas. What is the current
state of the field of marine conservation paleoecology, and what have been our major
contributions to date? As the application of a combination of bibliometric analyses
can provide a deeper understanding of discipline structure and dynamics (Yan and
Ding 2012; Chang et al. 2015), conservation paleobiology was explored using three
types of bibliometric networks: co-authorship, text co-occurrence, and author and
document co-citation.
Bibliographic Co-Authorship Visualizations
To map discipline structure and dynamics, it is important to first examine the

individual participants and researchers that form the discipline. Bibliographic
co-authorship analyses are widely used to investigate the structure of research
fields (White and Griffith 1981; Ramos-Rodríguez and Ruíz-Navarro 2004; Eom
2008; Raasch et al. 2013), as significant collaborations are expected to result in
co-authored publications. Co-authorship is thus viewed as evidence of scientific
collaboration and interactions among scientists and research teams (Melin and
Persson 1996). Although not all collaborations result in co-authored publications,
the number of collaborations that do not produce co-authored publications is
typically considered negligible (Melin and Persson 1998).
Documents that have more than one author are considered co-authored, and links
between co-authors form networks. The strength of the relationship is based on
the number of co-authored documents. In all co-authorship analyses, full counting
was employed, meaning that each occurrence has equal weight. For example, if
a publication had three co-authors, it was assigned to each author with a weight
of one. Authors with fewer than five documents were excluded. Temporal overlays
were then applied to identify shifts in productivity over time and to determine which
authors and research groups are currently contributing most to the field.
Text Co-Occurrence Visualizations
Co-occurrence analysis is a well-established method in bibliometrics (Callon

et al. 1983), employing the co-occurrence of keywords and mapping proximity
between keywords in scientific documents (Boyack and Klavans 2010). The more
frequently words co-occur, the stronger the relationships between them, as they
reside in similar research sub-fields. Emerging topics can therefore be identified
by examining the frequency of the use of specific terms and the body of literature
that scientists are actively citing (de Solla Price 1965). By detecting and visualizing
research fronts, we can ask how conservation paleobiology started, recognize critical
paths in its evolution, and identify state-of-the-art research. Temporal overlays were
applied to maps produced by text co-occurrence analyses, to identify fast-growing,
emerging areas of research.
VOSViewer extracted terms from titles and abstracts, terms are defined as
a sequence of nouns and adjectives ending with a noun. Binary counting was
employed, counting the presence or absence of a word and not the number of times
a word occurs in a document. A relevance score was calculated for each word,
and using these scores, the 60% most relevant terms were selected for mapping.
VOSViewer used the extracted terms to make a map, where distances between two
terms indicates number of co-occurrences of the terms. The smaller the distance
between two terms, the larger the number of co-occurrences of the terms.
Bibliographic Co-Citation Visualizations
Citation analysis employs one of the most crucial indicators of scholarship: citations
(Small 1977). Author co-citation analysis records indirect interactions between
scientists through the scientific literature, by identifying and counting the number of
times author pairs are cited among a set of articles (White and Griffith 1981; McCain
1990; White and McCain 1998). In other words, co-citation occurs when two
authors are cited by a third author, regardless of which articles are being cited (Small
1973; White and Griffith 1981; White and McCain 1998). When more authors cite a
pair of authors, higher co-citation strength results (Small 1973), i.e., the more cited
authors two documents have in common, the greater the co-citation strength. This
is often used as an indication of similarity in subject specialties between authors
(White and McCain 1998). Co-citation measures the association between concepts
represented by highly cited papers (Marshakova 1973; Small 1973) and can be
232 C. L. Tyler
used to identify areas of common interest. Altogether, these authors constitute the
“invisible college,” bringing to light documents in the research network that refer to
each other without being linked by recognizable organizational ties (de Solla Price
1965; Crane 1972; Lievrouw 1989; Gmür 2003). Authors of highly cited papers in
co-citation analysis are thought to constitute the leading scientists in a discipline
(White and McCain 1998). Author co-citation analysis can therefore also be used to
identify highly cited, and presumably influential researchers, and their specialties.
Journal co-citation was also employed to identify the proportions of documents
appearing in journals assigned to multiple subject categories and the prevalence of
boundary crossing co-citations. Large numbers of publications in interdisciplinary
journals would indicate that authors consider their research relevant for audiences
in multiple disciplines.
As only the WoS Core Collections data contain full records, necessary for
co-citation analyses, all co-citation analyses were performed using the more
conservative “Core Collection” dataset (see below). Co-citation analyses were
applied using authors, and sources (journals) as the units of analyses. The full-
counting method was applied assigning each occurrence equal weight. Disciplinary
affiliations of authors were assigned based on authors’ institutional web pages to
view the structure and membership of the science.
Bibliographic Coupling Visualizations
Bibliographic coupling can be used to represent a disciplines knowledge base

(Garfield et al. 2003). When two documents cite the same reference, they are biblio-
graphically coupled, with higher values indicating a strong relationship between the
two documents (Kessler 1963). Bibliometric coupling measures similarity between
papers and the number of references two papers have common (Kessler 1963).
To determine which journals currently publish high impact and/or cutting-edge
conservation paleobiology research, journal names in the references were used as
the unit of analysis, and a temporal overlay was applied. Similar to co-citation
analysis, full records were required, restricting analysis to the more conservative
“Core Collection” dataset (see below). The full-counting method was applied
assigning each occurrence equal weight.
3 Bibliometric Networks
WoS searches resulted in two datasets consisting of bibliographic records (Table 1):
The topic search for the phrase “conservation paleo*” in the WoS “Core Collection”
(CP-CC) yielded 833 records, and “conservation paleobiology” in “All Databases”
(CPb-AD) yielded 3418 records. Although number of records varies dramatically
between these datasets, they record similar publication trends, and the number of
records produced per year in both increases substantially from 1990 to 2000 to
present (Fig. 1).
Table 1 Bibliographic WoS database Topic search terms Records

datasets
Core collection Conservation paleo* 833
All databases Conservation paleobiology 3418
Number of records downloaded from WoS for each
database and search term. Topic searches included the
®
title, abstract, author keywords, and Keywords Plus fields
within a record. Records downloaded from WoS on Octo-
ber 21st, 2016
a b
100 150 200 250 300

80
number of publications
60
40
20
50
0
0
1980 1990 2000 2010 1940 1960 1980 2000

Year Year
Fig. 1 Publication trends through time. Number of publications per year from first usage to 2016.
Panels plotting data generated using two search terms and either WoS “Core Collection” or “All
Databases”. (a) Number of records per year resulting from search terms “conservation paleo*” in
WoS “Core Collection”; (b) Number of records per year resulting from search terms “conservation
paleobiology,” in WoS “All Databases”
Bibliographic Co-Authorship Networks
Bibliographic co-authorship networks were generated using both datasets, however,

as the CP-CC dataset did not yield a sufficiently populated map for meaningful
interpretation (only 14 authors met the minimum requirement of five documents1 ,
creating 14 isolated nodes), only the CPb-AD network is presented here.
The CPb-AD dataset contained 10,318 authors, of which 69 met the minimum
threshold of five documents. Twenty-three clusters of authors were identified, of
which ten reveal strong relationships between authors, representing working groups,
frequent collaborations between authors, or shared lines of inquiry (Fig. 2). The
most influential groups of authors were identified using number of citations as an
estimate of research impact. Four groups including authors with average citation
rates above 40 are apparent in Fig. 2, while the remainder of authors has an average
1
Authors ordered by number of citations, from highest to lowest: S. T. Jackson, D. R. Foster, K.
J. Willis, C. Saiz-Jimenez, J. M. Gonzalez, M. C. Portillo, J. C. Svenning, V. Rull, J. Salse, T.
Vegas-Vilarrubia, G. P. Dietl, E. Montoya, J. M. Pandolfi, and J. L. McGuire.
234 C. L. Tyler
Fig. 2 Citation rates of research groups and frequent collaborators for CPb-AD. Visualization of
CPb-AD bibliographic co-authorship analysis. Clusters generated by strong relationships between
authors, indicating research groups and frequent collaborations. Nodes were weighted by number
of citations, with larger nodes and warmer colors signifying higher numbers of citations. Colors
correspond to the scale on the bottom right, numbers on scale are average number of citations
citation rate of 20. The largest group of related authors consisted of a cluster of 11
authors, the next largest of nine authors, and clusters three and four both consisting
of eight authors (Fig. 3a–d). Clusters of authors are highly interdisciplinary
(Table 2). When nodes were weighted by documents, and viewed with a temporal
overlay, individuals within groups that made their most significant contributions in
the previous decade could be distinguished from those actively publishing (Fig. 4).
All clusters containing more than one author include at least one individual with
on average more recent publications (after 2013), and three of the four largest
research groups contained members with average publication dates prior to 2010
(Figs. 3 and 4).
Fig. 3 Largest research groups and most frequently cited contributors for CPb-AD. The four
largest research groups identified in the CPb-AD bibliographic co-authorship analysis (from
Fig. 2). (a) The largest research group consisting of 11 collaborating authors, (b) second largest
cluster consisting of 9 authors, (c, d) tied for third position, both consisting of eight authors. Nodes
were weighted by number of citations, with larger nodes and warmer colors signifying higher
numbers of citations. Colors correspond to the scale on the bottom right, numbers on scale are
average number of citations. When compared with Fig. 4, it is apparent that the second largest
cluster (b) was the only cluster lacking an author whose average contributions predate 2010
Text Co-Occurrence Networks
VOSViewer extracted 69,407 terms from the CPb-AD dataset, of which 1523 met
the minimum threshold of 10 occurrences. Relevance scores were then calculated
for the 914 terms making up the 60% most relevant terms. The resulting terms list
was carefully inspected and terms not related to research topics, such as geographic
locations or species names were removed (the following are examples of removed
terms: “usage,” “English,” “book,” “page,” “chapter,” “bibliography,” “illustra-
tion,” “Balkan peninsula,” “Americas,” “southern Brazil”). The text co-occurrence
analysis identified three modules (Fig. 5). All three modules included references
to conservation and anthropogenic impacts in various forms, however, modules
236 C. L. Tyler
Table 2 Disciplinary structure of the largest research groups

Group Author Discipline Documents Av. year Av. citations
1 Barber, RF Statistics 5 2015 0.20
Burney, DA Paleontology 6 2011 18.33
Crowley, BE Anthropology 5 2014 4.40
Dietl, GP Paleontology 8 2013 11.38
Godfrey, LR Anthropology 5 2008 8.00
Jackson, ST Botany 6 2012 84.67
Kidwell, SM Paleontology 7 2013 23.29
Minckley, TA Geography 6 2013 5.17
Smith, FA Biology 7 2014 15.29
Svenning, JC Biology 8 2015 16.50
Tomasovych, A Paleontology 7 2014 5.71
2 Barnes, I Paleontology 6 2013 29.50
Doadrio, I Genetics 5 2012 16.80
Gilbert, MPT Genetics 6 2013 84.00
Hofreiter, M Evolutionary biology 7 2012 38.86
Hooghiemstra, H Palynology 5 2012 6.80
Kitchener, AC Biology 5 2011 2.60
Orlando, L Genetics 5 2012 99.20
Shapiro, B Genetics 6 2014 81.67
Willerslev, E Genetics 7 2012 82.86
3 Birks, HJB Biology 5 2013 17.40
Feurdean, A Biology 7 2013 18.14
Gillson, L Biology 10 2013 11.80
Lamentowicz, M Paleontology 5 2011 24.00
Sayer, CD Geography 5 2013 16.20
Williams, JW Geography 6 2012 14.17
Willis, KJ Ecology 7 2009 26.43
4 Budd, AF Paleontology 5 2005 13.80
Harnik, PG Paleontology 6 2012 24.33
Liow, LH Ecology and evolution 7 2013 24.00
Lockwood, R Paleontology 5 2014 21.60
Mcguire, JL Biology 6 2014 103.17
Pandolfi, JM Historical ecology 9 2012 18.00
Rick, TC Archaeology 6 2013 9.50
Simpson, C Paleobiology 8 2013 17.88
Authors in the four largest research groups (Figs. 2 and 3) and their disciplines. Authors classified
into generalized disciplines based on research interests. Documents—number of publications,
average publication year, and average citation—average scores of the documents by author. Each
working group contains collaborators from a minimum of four disciplines, indicating widespread
interdisciplinary exchange and integration
Fig. 4 Publication rates of research groups and frequent collaborators for CPb-AD. Visualization
of CPb-AD bibliographic co-authorship analysis as in Fig. 2. Nodes were weighted by number of
documents, with larger nodes signifying higher numbers of documents. Warmer colors indicate
more recent publications, and colors correspond to the scale on the bottom right where numbers
on the scale are average publication year
appeared to define boundaries grouping relatively more cohesive sub-disciplines

with greater intellectual overlap (Fig. 5a). Module one consisted of 134 items and
included terms describing environmental history and conservation archeobiology
such as “anthropogenic activity,” “anthropogenic disturbance,” “anthropogenic
impact,” “baseline,” “death assemblage,” “drought,” “disturbance,” “European
settlement,” “fire,” “little ice age,” “macrofossil,” “microfossil,” “palaeoecology,”
“palynology,” “pollen,” and “sea level rise.” Module two consisted of 118 phrases
and was dominated by references to genetics and evolutionary biology such
as “allele,” “base pair,” “biodiversity hotspot,” “conservation concern,” “disper-
sal,” “divergence,” “DNA,” “endemic,” 12 phrases beginning with “gene” (e.g.,
gene flow, genetic drift), “invasive species,” “molecular clock,” “niche,” “pop-
ulation,” and “refuge,” with no direct references to paleontology. Module three
consisted of 57 terms describing ecology and paleontology including “biodiversity
238 C. L. Tyler
Fig. 5 Most commonly occurring terms in the CPb-AD dataset. Size of nodes indicates occurrence frequency, with words associated with larger nodes
occurring more frequently. (a) Colors differentiate the three modules identified by VOSViewer, based on strength of inter-group relationships. Modules could
be classified into three field of research: environmental history and conservation archeobiology (red, module one), genetics and evolutionary biology (green,
module two), and ecology and paleontology (blue, module three). Paleontological terms occurred in both the historical ecology module (red) and ecology and
paleontology (blue), but not in genetics and evolutionary biology. (b) Citation frequency overlay, warmer colors represent more highly cited topics. Scale on
bottom left represents average number of citations. Commonly occurring terms (larger nodes and corresponding text) did not have the highest average citation
frequencies, and common terms plotted with cooler colors. (c) Temporal overlay with warmer colors denoting more recent publications, colors correspond to
the scale on the bottom left indicating average publication year
A Conceptual Map of Conservation Paleobiology: Visualizing a Discipline
239
240 C. L. Tyler
loss,” “biotic interaction,” “conservation biology,” “conservation paleobiology,”

“extinction risk,” “geographic range,” “morphology,” “paleontology,” “paleobiol-
ogy,” “paleoecology,” “paleoenvironment”, “predation,” “recovery,” “reintroduc-
tion,” “taphonomy,” and “trace fossil.”
Colors signifying average citation frequency scores for the documents in which
a term occurs were overlaid on the text co-occurrence map to identify major areas
of research (Fig. 5b). The average publication year of documents in which common
terms occur are prior to, or during, 2012 for the ten most common terms in each
module (Table 3). Similarly, average publication years of which documents in which
Table 3 Ten most frequently occurring terms by module within the CPb-AD dataset
Module Terms Occurrences Av. year Av. citations
1 Sediment 304 2008 14.39
Erosion 206 2009 11.53
Pollen 180 2008 13.99
Soil 146 2005 13.64
Human activity 143 2008 18.41
Fire 132 2010 17.11
Deposition 113 2008 10.38
Human impact 107 2010 21.24
Land use 98 2007 14.32
Charcoal 87 2010 12.46
2 Population 614 2010 13.97
Divergence 173 2012 15.29
Genetic diversity 155 2012 15.54
Diversification 141 2011 16.16
Refugia 131 2011 25.25
Phylogeography 116 2012 15.48
Evolutionary history 110 2012 10.99
Dispersal 108 2010 18.21
Genetic structure 106 2012 11.08
Gene flow 103 2011 14.50
3 Morphology 142 2008 13.68
Biogeography 98 2008 16.49
Trait 71 2011 18.39
Taxonomy 67 2009 7.72
Mass extinction 58 2010 28.43
Recovery 51 2010 22.41
Conservation biology 48 2009 11.54
Paleoecology 47 2008 16.51
Biology 45 2006 10.56
Paleontology 41 2001 11.20
The ten most common terms in each module and the frequency with which they occur (Fig. 5).
Average publication year and average citation are average scores of the documents in which the
term occurs. Module numbers correspond with the colors in Fig. 5a (see figure caption)
a term occurs were overlaid on the text co-occurrence map to identify emerging
research fronts (Fig. 5c). Terms within genetics and evolutionary biology were
overwhelmingly more recent, likely reflecting technological advances in genetics.
Although terms with more recent average publication scores did not have high
occurrence frequencies (Table 4). The origin of conservation paleobiology in 2006
and its subsequent rise is apparent in the temporal overlay, as the average publi-
cation year for terms such as “paleoenvironment” (2005), “palaeontology” (2006),
“taphonomy” (2007), “microfossil” (2007), “paleoecology” (2008), “paleoecologi-
cal data” (2008), “paleobiology” (2009), “paleoecological record” (2009), “death
Table 4 Ten most recently occurring terms by module within the CPb-AD dataset
Module Terms Av. year Occurrences Av. citation
1 Anthropocene 2015 17 18.24
Ecosystem service 2013 19 13.42
Charcoal record 2012 17 12.94
Palaeoecological data 2012 17 20.77
Grain size 2012 18 8.72
Anthropogenic impact 2012 25 9.84
Baseline 2012 39 16.31
Proxy 2012 71 11.37
Palaeoecological record 2012 15 10.47
Anthropogenic disturbance 2012 24 5.75
2 Haplotype network 2014 11 2.18
High genetic diversity 2014 11 7.36
Ecological niche model 2014 16 6.69
Species distribution model 2014 18 9.11
Lineage diversification 2014 12 6.67
Nuclear gene 2014 19 7.63
Cryptic diversity 2013 15 6.47
CP-DNA 2013 11 12.00
Distribution range 2013 22 9.09
Cryptic species 2013 45 7.42
3 Biodiversity loss 2013 19 26.42
Evolutionary process 2013 28 19.96
Extinction risk 2012 38 13.47
Geographical range 2012 22 11.55
Conservation biologist 2012 10 9.70
Body size 2012 29 14.55
Molecular study 2012 16 11.00
Reintroduction 2012 13 9.46
Biotic response 2012 13 12.77
Biodiversity conservation 2012 20 33.30
The ten most recent terms in each module based on average publication year of the documents in
which the term occurs, and the frequency with which they occur (Fig. 5). Occurrences are listed for
each term and average citation scores (average scores of the documents in which the term occurs).
Module numbers correspond with the colors in Fig. 5a (see figure caption)
242 C. L. Tyler
assemblage” (2011), “fossil evidence” (2011), and “conservation paleobiology”

(2011) appear over time.
VOSViewer extracted 26,856 terms from the CP-CC dataset, of which 442 met
the minimum threshold of ten occurrences. Relevance scores were then calculated
for the 265 terms making up the 60% most relevant terms. The resulting terms list
was carefully inspected, and terms not related to research topics, such as geographic
locations or species names were removed. The text co-occurrence analysis identified
three modules (Fig. 6). Similar to the larger CPb-AD dataset, references to
conservation and anthropogenic impacts occur in all modules; however, modules
are grouped by sub-disciplines with greater intellectual overlap (Fig. 6a). Module
one consisted of 42 items and included terms describing paleontology, conser-
vation archeobiology, and environmental history such as “archaeology,” “condi-
tion,” “core,” “deposit,” “fossil,” “human impact,” “paleontology,” “sediment,” and
“soil.” Module two consisted of 41 phrases and was dominated by references to
genetics and evolutionary biology such as “biogeography,” “biology,” “dispersal,”
“divergence,” “DNA,” “gene,” “gene flow,” “genetic diversity,” “genetic varia-
tion,” “haplotype,” “mtDNA,” and “phylogeography.” Module three consisted of
22 terms describing climate science and ecology including “biotic response,”
“climate change,” “climatic condition,” “conservation status,” “extinction risk,”
“expansion,” “fragmentation,” “geographic range,” “global warming,” “last glacial
maximum,” “migration,” “species distribution,” and “temperature.”
A color overlay was applied to visualize average citation frequency scores
to identify major areas of research (Fig. 6b). The average publication year of
documents in which common terms occur was prior to, or during, 2013 for the
ten most common terms in each module (Table 5). The average publication year
overlay identified emerging research fronts (Fig. 6c). Terms within a combination of
genetics and ecology were overwhelmingly more recent, many of which pertained to
ecological response to climate change. Terms with more recent average publication
scores, again, did not have high occurrence frequencies (Table 6). The origin of con-
servation paleobiology in 2006 and its subsequent rise were apparent in the temporal
overlay, and terms such as “paleobiology” (average publication year 2011), “fossil”
(2010), “paleontology” (2009), “paleoecology” (2009), and “paleoecological data”
(2006) appeared over time.
I
Fig. 6 Most commonly occurring terms in the CP-CC dataset. Size of nodes indicates occurrence
frequency, with words associated with larger nodes occurring more frequently. (a) Colors differ-
entiate the three modules identified by VOSViewer, based on strength of inter-group relationships.
Modules could be classified into three field of research: paleontology, conservation archeobiology
and environmental history (red, module one), genetics and evolutionary biology (green, module
two), and climate science and ecology (blue, module three). Paleontological terms occur in the
conservation biology module. (b) Citation frequency overlay, warmer colors represent more highly
cited topics. Scale on bottom left represents average number of citations. Commonly occurring
terms (larger nodes and corresponding text) did not have the highest average citation frequencies,
and common terms plotted with cooler colors. (c) Temporal overlay with warmer colors denoting
more recent publications, colors correspond to the scale on the bottom left indicating average
publication year
244 C. L. Tyler
Table 5 Ten most frequently occurring terms by module within the CP-CC dataset
Module Terms Occurrences Av. year Av. citations
1 Condition 145 2009 19.70
Ecosystem 94 2010 32.98
Assemblage 85 2009 29.61
Fossil 61 2010 14.90
Sediment 57 2010 13.11
Water 50 2009 25.24
Nature 48 2010 24.96
Soil 44 2008 21.30
Deposit 43 2009 15.21
Resource 42 2009 14.67
2 Population 173 2009 26.60
Evolution 102 2008 26.16
Sequence 83 2008 22.37
Lineage 67 2009 34.37
Gene 58 2010 19.83
Divergence 48 2008 44.42
Clade 47 2010 29.17
Biogeography 41 2009 38.44
Morphology 40 2008 23.68
3 Response 86 2010 41.95
Climate change 78 2011 35.04
Expansion 58 2011 16.88
Temperature 49 2008 38.04
Richness 41 2010 22.34
Last glacial maximum 29 2011 22.66
Threat 28 2011 39.61
Precipitation 26 2009 22.08
Climatic change 24 2008 46.79
Risk 23 2010 23.57
The ten most common terms in each module and the frequency with which they occur (Fig. 6).
Average publication year and average citation are average scores of the documents in which the
term occurs. Module numbers correspond with the colors in Fig. 6a (see figure caption)
Bibliographic Co-Citation Networks
Bibliographic co-citation networks were created using the WoS CP-CC data for
cited authors, and cited sources. In the author co-citation network, out of 28,839
cited first authors, 167 met the minimum threshold of 25 documents per author.
Isolated nodes which did not share any citations in common with other papers in the
dataset were excluded in the visualization for enhanced comprehensibility, which
shows relationships between 153 of the authors. Four modules were present (Fig. 7),
Table 6 Ten most recently occurring terms by module within the CP-CC dataset
Module Terms Av. year Occurrences Av. citation
1 Paleobiology 2011 17 15.90
Discipline 2011 19 19.64
Proxy 2011 17 15.72
Ancient DNA 2011 17 23.36
Human impact 2011 18 24.70
Protected area 2011 25 6.23
Ecosystem 2010 39 32.98
Fossil 2010 71 14.90
Restoration 2010 15 18.41
Nature 2010 24 24.96
2 Genetic structure 2012 11 15.81
Historical biogeography 2011 11 12.90
Background 2011 16 9.81
Phylogeography 2011 18 20.29
Evolutionary history 2011 12 14.17
Refugia 2011 19 26.48
Genetic variation 2011 15 12.77
mtDNA 2010 22 59.42
Gene 2010 45 19.83
3 Extinction risk 2013 19 8.00
Conservation status 2012 28 3.50
Conservation planning 2012 38 25.56
Geographic distribution 2011 22 8.35
Last glacial maximum 2011 10 22.66
Global warming 2011 29 26.47
Species distribution 2011 16 33.07
Climate change 2011 13 35.04
Body size 2011 13 49.44
Persistence 2011 20 39.76
The ten most recent terms in each module based on average publication year of the documents in
which the term occurs, and the frequency with which they occur (Fig. 6). Occurrences are listed
for each term, and average citation scores (average scores of the documents in which the term
occurs). Module numbers correspond with the colors in Fig. 6a (see figure caption)
broadly representing the following research domains based on the disciplinary

affiliations of authors: genetics and evolutionary biology (red), ecology (green),
paleontology (blue), and conservation ecology and paleobiology (yellow). Multiple
highly cited and influential authors were present in all four specialties (Fig. 7):
(1) genetics and evolutionary biology—J. C. Avise (70 citations), C. Moritz (67
citations), A. J. Drummond (63 cites); (2) ecology—K. J. Willis (155 citations),
V. Rull (134 citations), S. T. Jackson (96 cites); (3) paleontology—S. M. Kidwell
(104 citations), D. Jablonski (85 citations), J. B. C. Jackson (71 cites); and (4)
246 C. L. Tyler
Fig. 7 The invisible college of conservation paleobiology crosses research domains. Visualization
of bibliographic co-citation author analysis of the CP-CC dataset. Links indicate relationships
between authors (nodes) that have published a minimum of 25 documents cited in the CP-CC
dataset. Greater line width corresponds to stronger relationships. Invisible colleges are based on
documents in a research network that refer to each other without being linked by recognizable
organizational ties. Authors’ last names and first initials are shown over nodes. Size of nodes
represents relative number of citations, with larger nodes identifying greater numbers of citations.
Colors denote four distinct modules representing disciplinary affiliations of authors with the fields
of genetics and evolutionary biology (red), ecology (green), paleontology (blue), and conservation
ecology and conservation paleobiology (yellow)
conservation ecology and paleobiology—D. K. Grayson (83 cites), A. D. Barnosky

(64 cites), G. P. Dietl (58 cites).
Bibliometric Coupling Networks
In the network visualization of bibliographic coupling using sources (Fig. 9a),

modules did not clearly define journal disciplines, mainly due to the influence
of high impact interdisciplinary journals such as PNAS and Science. The five
journals with the largest numbers of citations included Science with 1762 citations,
PNAS with 1142, Quaternary Science Review with 547, Biodiversity and Conser-
vation with 461, and Molecular Phylogenetics and Evolution with 453 (Fig. 9a).
Fig. 8 The journal landscape of conservation paleobiology. Visualization of the bibliographic co-
citation journal sources analysis using the CP-CC dataset. Links indicate relationships between
journal sources (nodes) that have a minimum of 20 citations of the sources. Greater line width
corresponds to stronger relationships. Size of nodes represents relative number of citations, with
larger nodes identifying greater numbers of citations. Colors denote seven modules, loosely
approximating journal disciplines such as earth sciences (red), paleontology (pink), and evolu-
tionary biology (green). The highest impact interdisciplinary journals plotted in the center of the
network and had a greater number of links with multiple modules
Journals more recently publishing articles on conservation paleobiology (Fig. 9b)

included Ecology and Evolution (average publication year 2015), PLOS One (2013),
Ecography (2013), Global Change Biology (2012), and Quaternary International
(2012). More highly cited journals had older average years of publication, and
the average publication years for the top five most highly cited journals ranged
from 2008 to 2010 (e.g., Science average is 2009). The largest numbers of articles
(Fig. 9c) were published in PNAS and PLOS One (26 documents), followed
by Molecular Phylogenetics and Evolution and Quaternary International (17),
Biological Conservation (15), and Quaternary Science Reviews (13). Of these,
PLOS One had the most recent average publication year.
Out of 15,946 journal sources in co-citation network, 312 journals met the
minimum requirement of 20 citations of a source. Seven modules were present
(Fig. 8), which loosely approximated journal disciplines such as earth sciences
(red), paleontology (pink), and evolutionary biology (green). Although dominated
by geological journals, the largest module (in red), for example, also included
248 C. L. Tyler
journals from ecology, paleontology, and conservation such as Ecology, Human

Ecology, and Climatic Change. The top ten journals with the highest citations in
conservation paleobiology were: Science (1602 citations), Nature (1289), PNAS
(1047), Molecular Ecology (744), Journal of Biogeography (643), TREE (582),
Palaeogeography Palaeoclimatology Palaeoecology (577), Conservation Biology
(524), and Paleobiology (508). These high impact interdisciplinary journals, plotted
towards the center of the network, and had a greater number of links to multiple
modules.
4 The Intellectual Landscape
The approximate origin of the concept of conservation paleobiology appears to

have been around 2006, and its subsequent rise was visible in the rapid increase in
references including the phrase (Fig. 1), and in the temporal overlays (Figs. 5 and 6).
After this phase of rapid ramping-up, publications on conservation paleobiology
increased exponentially peaking at 92 in 2015 in the CP-CC dataset, and 336 in 2010
the CP-AD dataset (Fig. 1). These findings are consistent with other assessments
of the growth and development of conservation paleobiology (Tyler and Schneider
2018).
Twenty-three interconnected groupings including researchers pursuing similar
lines of inquiry, intellectual collaborations, or research “teams” are evident in the co-
authorship network, many of which are interdisciplinary (Fig. 2). Interdisciplinary
co-authorship indicates a dependence on methods and knowledge from other
scientific domains (Moody 2004), which is to be expected in a rapidly growing
new discipline, where ideas, publications, and core-literature have not yet reached
a critical mass. Disciplines bring researchers with similar interests together, and
J
Fig. 9 Trending journals for cutting-edge research in conservation paleobiology. Bibliographic
coupling visualization of sources. Link width indicates strength of relationships. (a) Colors differ-
entiate the three modules identified by VOSViewer, based on strength of inter-group relationships.
Modules do not clearly define journal disciplines, mainly due to the influence of high impact
interdisciplinary journals such as PNAS and Science. Node sizes indicate numbers of citations, the
largest of which is Science with 1762 citations, PNAS with 1142, Quaternary Science Review with
547, Biodiversity and Conservation with 461, and Molecular Phylogenetics and Evolution with
453. (b) Temporal overlay, warmer colors represent more recent average publication years (scale on
bottom left). Size of nodes indicates number of citations. Journals more recently publishing articles
on conservation paleobiology plotted with warmer colors include Ecology and Evolution (average
publication year 2015), PlosOne (2013), Ecography (2013), Global Change Biology (2012), and
Quaternary International (2012). (c) Node sizes indicate number of documents published with
temporal overlay (colors correspond to the scale on the bottom left) indicating average publication
year. The largest numbers of articles were published in PNAS and PlosOne (26 documents),
followed by Molecular Phylogenetics and Evolution and Quaternary International (17), Biological
Conservation (15), and Quaternary Science Reviews (13)
250 C. L. Tyler
citations may be motivated by a combination of intellectual and interpersonal ties

between authors (White et al. 2004). Groups are therefore likely the result of a
combination of factors in addition to direct collaboration, including intellectual
ties with former advisees, and academic social networks. Three of the four largest
schools of thought contained at least one highly influential researcher, and one more
“senior” researcher with contributions averaging in 2010.
Researchers working on the topic of conservation paleobiology were clearly
not constrained to paleontology, and disciplinary affiliations of authors included
genetics and evolutionary biology, ecology, paleontology, and conservation ecology
and paleobiology (Fig. 7). Furthermore, authors working predominantly on conser-
vation included both paleontologists and conservation scientists, while authors with
more traditional paleontological contributions formed a distinct intellectual cluster
(Fig. 7), likely predominantly contributing foundational science (e.g., advances in
taphonomy and the fidelity of the fossil record) facilitating the expansion of research
in conservation paleobiology.
Applications of conservation paleobiology have been broad (Figs. 5, 6 and 7),
and research on the topics of conservation and paleontology has been conducted
by scholars from a wide range of disciplines, with significant intellectual exchange
between sub-disciplines. Conservation research conducted by paleontologists did
not form a distinct sub-discipline (i.e., conservation research within paleontology),
but was substantively coherent and interrelated with environmental history and
conservation archeobiology (Figs. 5 and 6), within conservation science. These
findings indicate a strong unity and inter-relatedness of content despite intellectual
sub-communities, suggesting that conservation science is one research field with
at least three major research themes: (1) environmental history and conservation
archeobiology, (2) genetics and evolutionary biology, and (3) ecology. Regardless
of the dataset used, conservation paleobiology did not form a distinct module in
either dataset, and content is integrated into environmental history and conserva-
tion archaeobiology. Increased interdisciplinary exchange and integrating problem
solving by scholars from different disciplines could have many potential benefits,
as poorly understood problems benefit from the greater complexity afforded by
interdisciplinary approaches, often yielding superior results (Birnbaum 1981).
Although not currently a sub-discipline, it is possible that over time as literature
accrues, paleontological research on conservation paleobiology may become an
independent self-sustaining structure.
Although many perceive conservation paleobiology to fall under the sensu stricto
domain of paleontology, given that conservation paleobiology frequently draws
from expertise and methods in conservation sciences, these results suggest that when
multiple types of paleontological data are considered, conservation paleobiology
may in practice be a sub-discipline within conservation science. These results
also indicate that conservation paleobiology is not a closed circuit, i.e., we are
not only talking to ourselves. If conservation paleobiology identifies with the
values and ethical norms steering conservation science (Soule 1985; Noss 1999;
Meine et al. 2006), and we accept conservation paleobiology as a field of research
within conservation science (Dietl 2016), perhaps our goals should similarly be
to forge relationships with economists, educators, environmental policy makers,

and conservation scientists (Savarese 2018; Dietl and Flessa 2018). Recognizing
conservation paleobiology as a sub-discipline of conservation science would allow
us to hook into pre-existing infrastructure and professional societies. For example,
attending professional conferences outside of paleontology, such as the “Society for
Conservation Biology” meeting, as suggested by Savarese (2018). Intra-disciplinary
research is typically favored by individual experience and expertise, communication
skills, career incentives and rewards, public outlets, and institutional organizational
structures (Klein and Porter 1990). Attempts to integrate with conservation science
may therefore be less than straightforward for researchers already trained and
networked within paleontology, and typically housed in geosciences departments.
However, simply recognizing the intellectual landscape of the field could be ben-
eficial, and appreciating the breadth of the conservation community could prompt
researchers to actively seek out exposure to topics and tools across disciplines.
While authors publishing cutting-edge research are likely to target well-
established high impact interdisciplinary journals such as Science, or PNAS,
most articles were published in journals within the authors’ specialty. Given the
interdisciplinary nature of the science, authors should carefully consider the target
audience, a large part of which lies outside of their specific discipline. Authors may
wish to submit their research to interdisciplinary or conservation science journals
to reach the broader community working on conservation paleobiology, or journals
publishing an increasing number of conservation paleobiology papers.
5 Emerging Frontiers
Thorough reviews of potential research areas to which conservation paleobiology

has contributed are provided elsewhere (Willis et al. 2010; Dietl et al. 2011, 2015;
Dietl and Flessa 2011; Fordham et al. 2016). Here instead, I encourage researchers
to consider potential paleontological contributions within the discipline framework
identified by bibliometrics. Paleontological and historical terms appeared within
various sub-disciplines (Tables 3, 4, 5 and 6). On average, terms within genetics
and evolutionary biology were overwhelmingly more recent, likely reflecting
technological advances and breakthroughs (Figs. 5 and 6). Although terms with
more recent average publication scores did not have high occurrence frequencies
(Tables 3 and 5), this is likely a product of the novelty of the topic. The newest
research themes and topics have not been present long enough to accumulate
high occurrences, nor are they likely to be widely recognized by the scientific
community and extensively published upon. Some of these emerging topics can
be expected to dissipate, but significant research areas should persist and grow.
The terms “paleobiology,” “fossil,” “Anthropocene,” “palaeoecological record,”
and “last glacial maximum” are included among the most recent topics, consistent
with the rise of this emerging sub-discipline, and the increasing awareness in
conservation science of the need for long-term data that pre-dates human activities
252 C. L. Tyler
(Smith et al. 2018). However, an increased awareness and expressed need for long-
term or paleontological data does not necessarily indicate conservation paleobiology
in practice. Furthermore, terms relating to explicit policy-oriented research were
less common (e.g., “conservation status,” “conservation planning,” “protected
area”). Several trending topics are apparent (Tables 4 and 6), and while many
related to broad problems that need to be addressed (e.g., “climate change”),
several topics likely to represented emerging frontiers to which paleobiology has
much to contribute, including extinction risk, refugia (e.g., see Schneider 2018),
identifying evolutionary relationships between populations, species and genetic
diversity, ecosystem services, establishing baselines (e.g., Kusnerik et al. 2018), and
biotic responses to perturbation (e.g., Roopnarine and Dineen 2018).
6 Conclusions
We have lacked a common organizing framework to identify central questions

in the field, which is particularly important for the emergence of new fields
of research (Torraco 2005). Here bibliometrics was used to promote a clearer
understanding of the intellectual structure of the field and its current development, to
remove some of the ambiguity surrounding the scope and meaning of conservation
paleobiology. The conservation paleobiology community is arguably at a critical
juncture, either moving forward predominantly within paleontology and geology,
or actively identifying with conservation science (Dietl 2016). Results showed
that conservation paleobiology is a relatively young emerging field of research
that is highly interdisciplinary. Paleontological research did not form a cohesive
sub-discipline, but was assimilated within other historical approaches to conserva-
tion. Although research communities appear clustered into several sub-disciplines,
research is highly integrated across communities. Interdisciplinary collaborations
and approaches will therefore likely play an increasingly important role in the
future, as innovate and impactful research is likely to be interdisciplinary, and many
of the challenges we face today require a broader approach.
Acknowledgments Thank you to Chris Schneider and Lindsey Leighton for reviewing this
manuscript, and to Anthony Giuffre for thoughtful discussions, which helped shape this contri-
bution.
References
Birnbaum PH (1981) Contingencies for interdisciplinary research: matching research questions

with research organizations. Manage Sci 27:1279–1293
Boyack KW, Klavans R (2010) Co-citation analysis, bibliographic coupling, and direct citation:
which citation approach represents the research front most accurately? J Am Soc Inf Sci
Technol 61:2389–2404
Callon M, Courtial J-P, Turner WA, Bauin S (1983) From translations to problematic networks: an
introduction to co-word analysis. Soc Sci Inf 22:191–235
Chang Y-W, Huang M-H, Lin C-W (2015) Evolution of research subjects in library and information
science based on keyword, bibliographical coupling, and co-citation analyses. Scientometrics
105:2071–2087
Crane D (1972) Invisible colleges: diffusion of knowledge in scientific communication. University
of Chicago Press, Chicago
Dietl G (2016) Brave new world of conservation paleobiology. Front Ecol Evol 4:21
Evol 26:30–37
Dietl GP, Kidwell SM, Brenner M et al (2011) Conservation paleobiology: opportunities for earth
science. Ithaca, New York
Eom S (2008) All author cocitation analysis and first author cocitation analysis: a comparative
empirical investigation. J Informetr 2:53–64
Fordham DA, Akçakaya HR, Alroy J et al (2016) Predicting and mitigating future biodiversity loss
using long-term ecological proxies. Nat Clim Chang 6:909–916
Garfield E, Pudovkin AI, Istomin VS (2003) Why do we need algorithmic historiography? J Am
Soc Inf Sci Technol 54:400–412
Gmür M (2003) Co-citation analysis and the search for invisible colleges: a methodological
evaluation. Scientometrics 57:27–57
van der Have RP, Rubalcaba L (2016) Social innovation research: an emerging area of innovation
studies? Res Policy 45:1923–1935
Kessler MM (1963) Bibliographic coupling between scientific papers. Am Doc 14:10–25
Klein JT, Porter AL (1990) Preconditions for interdisciplinary research. In: Birnbaum-More PH,
Rossini FA, Baldwin DR (eds) International research of management: studies in interdisci-
plinary methods from business, government, and academia. Oxford University Press, New
York, pp 11–19
Kusnerik KM, Lockwood R, Grant AN (2018) Using the fossil record to establish a baseline and
recommendations for oyster mitigation in the mid-Atlantic U.S. In: Tyler CL, Schneider CL
Leydesdorff L, Rafols I (2012) Interactive overlays: a new method for generating global journal
maps from web-of-science data. J Inf Secur 6:318–332
Lievrouw LA (1989) The invisible college reconsidered. Commun Res 16:615–628
Marshakova IV (1973) System of document connections based on references Sci Technical Inf Ser.
VINITI 3–8
McCain KW (1990) Mapping authors in intellectual space: a technical overview. J Am Soc Inf Sci
41:433–443
Meine C, Soule M, Noss RF (2006) “A mission-driven discipline”: the growth of conservation
biology. Conserv Biol 20:631–651
Melin G, Persson O (1996) Studying research collaboration using co-authorships. Scientometrics
36:363–377
Melin G, Persson O (1998) Hotel cosmopolitan: a bibliometric study of collaboration at some
European universities. J Am Soc Inf Sci 49:43–48
Moody J (2004) The structure of a social science collaboration network: disciplinary cohesion
from 1963 to 1999. Am Sociol Rev 69:213–238
Noss R (1999) Is there a special conservation biology? Ecography (Cop) 22:113–122
Raasch C, Lee V, Spaeth S, Herstatt C (2013) The rise and fall of interdisciplinary research: the
case of open source innovation. Res Policy 42:1138–1151
254 C. L. Tyler
Rafols I, Leydesdorff L (2009) Content-based and algorithmic classifications of journals: per-

spectives on the dynamics of scientific communication and indexer effects. J Am Soc Inf Sci
Technol 60:1823–1835
Ramos-Rodríguez A-R, Ruíz-Navarro J (2004) Changes in the intellectual structure of strategic
management research: a bibliometric study of the strategic management journal, 1980–2000.
Strateg Manag J 25:981–1004
Roopnarine PD, Dineen AA (2018) Coral reefs in crisis: the reliability of deep-time food
web reconstructions as analogs for the present. In: Tyler CL, Schneider CL (eds) Marine
Savarese M (2018) Effectively connecting conservation paleobiological research to environmental
management: examples from Greater Everglades’ restoration of southwest Florida. In: Tyler
Schneider CL (2018) Marine refugia, past, present, and future: lessons from ancient geologic
crises for modern marine ecosystem conservation. In: Tyler CL, Schneider CL (eds) Marine
Small H (1973) Co-citation in the scientific literature: a new measure of the relationship between
two documents. J Am Soc Inf Sci 24:265–269
Small HG (1977) A co-citation model of a scientific specialty: a longitudinal study of collagen
research. Soc Stud Sci 7:139–166
Small H, Boyack KW, Klavans R (2014) Identifying emerging topics in science and technology.
Res Policy 43:1450–1467
de Solla Price DJ (1965) Networks of scientific papers. Science 149(3683):510–515
Soule ME (1985) What is conservation biology? Bioscience 35:727–734
Torraco RJ (2005) Writing integrative literature reviews: guidelines and examples. Hum Resour
Dev Rev 4:356–367
Tyler CL, Schneider CL (2018) An overview of conservation paleobiology. In: Tyler CL, Schneider
CL (eds) Marine conservation paleobiology. Springer, Cham, pp 1–10
Van Den Besselaar P, Heimeriks G (2006) Mapping research topics using word-reference co-
occurrences: a method and an exploratory case study. Budapest Sci 68:377–393
Van Eck NJ, Waltman L (2010) Software survey: VOSviewer, a computer program for bibliometric
mapping. Scientometrics 84:523–538
Van Eck NJ, Waltman L (2014) Visualizing bibliometric networks. In: Ding Y, Rousseau R,
Wolfram D (eds) Measuring scholarly impact: methods and practice. Springer, New York, pp
285–320
Van Eck NJ, Waltman L, Dekker R, van den Berg J (2010) A comparison of two techniques for
bibliometric mapping: multidimensional scaling and VOS. J Assoc Inf Sci Technol 61:2405–
2416
White HD, Griffith BC (1981) Author cocitation: a literature measure of intellectual structure. J
Am Soc Inf Sci 32:163–171
White HD, McCain KW (1998) Visualizing a discipline: an author co-citation analysis of
information science. J Am Soc Inf Sci 49:327–355
White HD, Wellman B, Nazer N (2004) Does citation reflect social structure? Longitudinal
evidence from the “Globenet” interdisciplinary research group. J Am Soc Inf Sci Technol
55:111–126
Willis KJ, Bailey RM, Bhagwat SA, Birks HJB (2010) Biodiversity baselines, thresholds and
resilience: testing predictions and assumptions using palaeoecological data. Trends Ecol Evol
25:583–591
Yan E, Ding Y (2012) Scholarly network similarities: how bibliographic coupling networks,
citation networks, cocitation networks, topical networks, coauthorship networks, and coword
networks relate to each other. J Am Soc Inf Sci Technol 63:1313–1326
Index
A C
Acanthinucella spirata, 179 Caloosahatchee River, Southwest Florida
Anadara sp., 174 geographic locations, 64
Analysis of covariance (ANCOVA), 85, water management of, 65–66
87–89 Carboniferous Tindouf Basin, Morocco, 180,
Aquaculture, 77, 94 191
Artificial marine micro-reserve, 173 Chesapeake Bay oysters
Association of Environmental Studies and dermo disease, 76–77
Sciences (AESS), 61 Pleistocene, colonial and modern oysters
Atkinson Center for a Sustainable Future, 219 (see Pleistocene fossil oysters,
Atkinson Postdoctoral Fellows in Mid-Atlantic U.S.)
Sustainability, 219 Climate change
anthropogenic enhancement, 190
glacial–interglacial climatic and eustatic
B fluctuations, 176
Bibliographic co-authorship visualizations, greenhouse gas trajectories, 187
230–231 impacts on ocean water, 164
Bibliographic co-citation networks, 245–246 in marine reserves, 169
Bibliographic co-citation visualizations, short-term refugia, 186
231–232 in species distribution models, 184
Bibliographic coupling visualizations, 232 Climate refugia, 168, 169, 176
Bibliometric coupling networks, 246–249 Clumped isotopic analysis, 82, 83
Bibliometrics networks Coastal and Estuarine Research Federation
bibliographic co-authorship networks, (CERF), 61
233–235 Comprehensive Everglades Restoration Plan
bibliographic co-citation networks, (CERP), 66
244–246 Conservation archeobiology, 5, 6
coupling networks, 246–249 Conservation biology, 2
datasets, 232–233 and community ecology, 105
text co-occurrence networks dynamical structures, 106
average citation frequency, 240 geohistorical data, 4–5
CPb-AD dataset, 235, 240–244 conservation archeobiology, 5, 6
modules, 235–240 conservation paleobiology (see
Binary counting, 231 Conservation paleobiology)
Biodiversity loss, 14 environmental history, 5–6

Topics in Geobiology 47, https://doi.org/10.1007/978-3-319-73795-9
256 Index
Conservation biology (cont.) long-term data, marine conservation (see

habitat preservation and restoration, 105 Long-term data)
integrated biosphere, 106 National Science Foundation, 210
Conservation, definition, 7 near-time vs. deep-time studies, 5–7
Conservation paleobiologists objectives of, 8
NSF-funded Conservation Paleobiology partnerships with, 219
Workshop, 210 role in university’s involvement, 219
Society for Conservation Biology, 213 science advocacy, 12–13
specific recommendations, 212–220 science communication
advertisements and interviews, 222 own values and biases, 13
alternative learning spaces, 216–217 pure scientist to honest broker,
collaborative research, 219–220 14–18
conservation biology programs, 213 stakeholders, 13
in conservation paleobiology, 210 value biodiversity, 13
cross-disciplinary institutes/centers, 219 as sub-discipline within conservation
cross-disciplinary work, 210–214 science, 3, 7
faculty incentive systems, 218 Conservation science, 2, 3, 228, 250–251
improvement of doctoral education, 217 Conservation training, in human dimension,
issues in, 220–222 214
partnerships with practitioners, Coral reefs, food web reconstructions
217–218 anthropogenic and non-anthropogenic
policy oriented training, 214–215 stressors, 132
practical experience, 215–216 biocalcification, 108
students mentoring, 222–223 Caribbean reefs, 108
traditional academic training, 211 community reconstruction
Conservation paleobiology (CPB), 56, 106, diversity and evenness, 129–130
192 epibenthic sponge trophospecies, 127
academicize, 12, 20 fossilization with fossil taxon, 126
bibliometric networks hard benthic macroinvertebrates, 129
co-authorship networks, 233–235 large vertebrate macrophyte and
co-citation networks, 244–246 invertebrate grazers, 129
coupling networks, 246–249 predator and prey guild richness,
datasets, 232–233 128–129
text co-occurrence networks, 235–244 RMA function, 128
career in, 222 simulated food webs, 130–132
community partnerships, 218 terrestrial paleocommunities with insect
cross-disciplinary, 210–211 faunas, 127
current state and structure, 228–229 0.6 threshold of interaction overlap,
decision making policy, 214 127, 136
definition, 2 zooplanktivorous species, 127, 132
discipline mapping coral bleaching, 108
bibliographic co-authorship fossil food web studies, 109
visualizations, 230–231 fossilizing (see Fossilization, coral reefs)
bibliographic co-citation visualizations, guild structure and diversity
231–232 aggregation, 122–124
bibliographic coupling visualizations, carcharhinid sharks, 124
232 guild-level food webs, 124
bibliometric data, 229–230 guilds identification, 124–126
text co-occurrence visualizations, 231 in-degree distribution, 134
emerging frontiers, 251–252 inferred ecologies, 124
facts vs. feelings, 18 Lagerstätten approach, 123–124
fear factor, overcoming, 19–20 metanetworks, 124
growth and development of, 3, 4 stochastic techniques, 134
intellectual landscape, 249–251 trophospecies, 123
Index 257
hypergeometric variance, 134–135 Environmental management and restoration,

Jamaican reefs, 108–109 56
ocean, anthropogenic carbon dioxide, 108 community-engaged scholarship, 62
SST, 108 community services, recognition of, 63
terrestrial paleocommunities, 133 environmental problem, identification of,
zooxanthellae, 108 58
Cornell Roundtable on Environmental Studies environmental science and restoration
Topics (CREST), 217 conferences, 61
Cornell University, 216 Greater Everglades’ restoration (see
CPB, see Conservation paleobiology Greater Everglades’ restoration)
Crassostrea virginica, 76 liaison functions, 60
paleosalinity, 83 management collaboratives, participation
paleotemperature in, 60
clumped isotope analysis, 82, 83 oral presentation to management
isotopic sclerochronological analysis, community, 61
82–84 partnerships and collaborative teams,
sample collection, 77, 80–81 development of, 59
shell height, 81, 82 peer-reviewed journal articles, 61
stratigraphic unit and geologic age, 79, 80 students, training of, 62
Cretaceous hydrothermal vent faunas, 170 technical reports, 61
Cretaceous-Paleogene mass extinction, 171 Environmental perturbation, 165
Cross-disciplinary curricula Environmental stressors, 33–36, 44–45, 49
Atkinson Center for a Sustainable Future, Excess of objectivity, 15
219 Extinction-causing stress, 164
availability of fund, 219 Extinction survivorship, 165, 193
certificate program, 214
of conservation professionals, 210–211
institutional constraints, 220 F
San Juan River field course, 215 Florida Gulf Coast University (FGCU), 57–58,
in specific recommendations, 212, 213 218
Cryptic refugia, 173–174 Ford Foundation, 220
Fossilization, coral reefs
amalgamation, 109
D biased species preservation, 111
Dapsilidinium pastielsii, 180 binary filter, 110
Deep-time food web reconstructions dietary breadths, 113–114, 133
coral reef (see Coral reefs, food web DRYAD database, 109
reconstructions) expected and observed losses of taxa, 111,
environmental stresses, 107 112
preservating past, 107 Greater Antilles map, 109, 110
Dermo disease, 76–77 hypergeometric probability, 111
Devonian ostracods, 170 loss of biotic interactions, 111
modularity
analysis, 122
E Caribbean reef shark, 122
Early Jurassic reefs, 169 compartmentalization, 118
Early Triassic habitable zone, 175 fossil modules, distribution of, 122, 123
Eastern oyster, see Crassostrea virginica historical and paleontological records,
Ecological Society of America listserv 122
(ECOLOG-L), 26, 46 historical and sub-fossil records, 133
Ecosystem monitoring, 2 long-held hypothesis, 119
Endemic–area relationships (EAR), 153 May’s theoretical work, 119
Environmental conservation, 63 modern and simulated fossil food webs,
Environmental history, 5–6 119–120
258 Index
Fossilization, coral reefs (cont.) late Cambrian trilobite clades, 169

Module 0, 121 refugia type, 167, 168
Module 1, 120 shallow-water refugia, 169–170
Module 2, 120–121 wide shelves and embayments, 170
Module 3, 122 Harvest refugia, 174–175
Newman’s algorithm, 119 Hawaiian Islands Humpback Whale National
representative food chains, 120, 121 Marine Sanctuary, 172
stability/resistance, 133 Historical and geological records, 106
simulated fossilization, 132 Hyper-diverse shallow curricula, 221
taxon composition, 109
taxon level trophic properties, 112
trophic levels I
characteristic, 114 Iberian Peninsula coast, 166, 168
food chain length, 115, 118 Indiana University-Purdue University
fractional trophic level, 116, 117 Indianapolis (IUPUI), 215
habitat alteration/destruction, 115 Informative advocacy, 15–16
measurement, 116 Integrative Graduate Education and Research
network trophic level, 116–118 Traineeship (IGERT), 219
phytoplankton–zooplankton food chain, Intellectual landscape, in conservation
116 paleobiology, 249–251
potential richness and complexity, 115 Interpersonal skills, 212
predator diversity, 115 Isolated geographic refugia, 170–172
Scleractinian reefs, 115 Isotopic sclerochronological analysis, 82–84
trophospecies, 110–112 IUPUI Center for Service and Learning, 218
zooplankton, 112
K
G Kettleman Hills
Georgia State University, 222 Etchegoin Formations, 148, 150
Greater Everglades’ restoration, 63 Jacalitos Formations, 148
Caloosahatchee River Macoma zone, 149–150
geographic locations, 64 Patinopecten zone, 149–150
water management of, 65–66 PCM community, 150
environmental agencies/NGOs, 57 Pliocene paleogeographic map, 148, 149
historic and current flow patterns, 63–64 San Joaquin Formation, 148, 149
Picayune Strand restoration project Talure Formation, 148
CERP, funds, 66 Kokenia sp., 169
conservation paleobiology, 67 Kolmogorov-Smirnov tests, 85
geographic locations, 64
management collaborative, 67
oyster reef development, 68–71 L
project development teams, 67 Late Ordovician extinction, 178
restoration design, 70, 71 Late Permian mass extinction, 169
roads and canal systems, environmental Lazarus effect, 181
problem, 66–67 Lazarus taxa, 181–182
salinity targets, 68 Lingularia, 173
Greenhouse-gas emissions, 15 Long-term data (LTD)
Green technology, 15 definitions, 24
neontological and geohistorical data, 24–25
survey, marine conservation biologists
H conservation barriers, 38–40, 45, 50
Habitat shift conservation goals and approaches,
change in water depth, 169 29–30, 43, 47–48
habitable zone, 170 conservation paleobiologists, 40–41
Index 259
conservation-related roles, 27, 28, short-term survival, during environmental

42, 47 perturbations, 185
data types, 32, 33, 50 size and connectivity of, 187–189
environmental stressors, 33–36, 44–45, species population, 176–177
49 tolerable stress levels, 189–190
fields of study, 26–27, 42, 46–47 Mid-Atlantic oysters
gender/ethnicity, 26, 27, 42 colonial and modern oysters
level of biological organization, 27, 28, growth rates, 83
43 vs. Pleistocene oysters (see Pleistocene
level of education, 26, 27, 42 fossil oysters, Mid-Atlantic U.S.)
LTD sources, 30–32, 43–44, 49 shell height, 83
population selection and distribution, conservation paleobiology, role of, 96
26, 45–46 human harvest, record of, 76
reference conditions/baselines, temporal management approaches, 77
and spatial data, 36–38, 45, 48–50 restoration, implications for, 94–96
respondent workplace, 26, 27, 42 Modern marine ecosystem conservation,
selections of timescales, 30, 31, 43 marine refugia, see Marine refugia
years of working experience, 27, 28, 42 Modern marine ecosystem management, 106
Long-Term Ecological Research Network, 2 Motu Maha Marine Reserves, 172
LTD, see Long-term data
N
M National Association of Marine Laboratories
MacArthur Foundation, 220 (NAML), 26, 46
Mann-Whitney U tests, 85, 86 National Science Foundation (NSF)-funded
Marine ecosystems, 2 Conservation Paleobiology
Marine protected areas (MPA), 95 Workshop, 210
Marine refugia Nature Conservancy, 219
ancient refugia, identifying of NatureNet Science Fellows Program, 219
fossil data, 182–183, 192 Network trophic level (ntl), 116–118
geological record, 192 Non-governmental organizations (NGOs), 57,
heterogeneous fossil record and 59
sampling, 181 North Atlantic refugia, in English Channel,
Lazarus effect and Lazarus taxa, 166
181–182 Nucella lima, 183
phylogeographic studies, 183–184
space–time relationships, 181
species distribution models, 184–185 O
cryptic refugia, 173–174 Ordovician brachiopods, 170
duration and time span, 177–179 Oxygenation, 164
environmental conditions of refugium, Oyster reefs
175–176 Blackwater and Faka Union Bay, 68–71
in global biosphere recovery, 165 habitat destruction, 76, 77
habitat shifts, 169–170 harvesting, 76
harvest refugia, 174–175 increased sediment input, 76, 77
isolated geographic refugia, 170–172 in Maryland, Virginia, and North Carolina
life history, 172–173 (see Mid-Atlantic oysters)
range contraction, 166 parasitic diseases, 76–77
range shifts, 167–169 Ten Thousand Islands, 68, 69
refugial traps, 190–191
refugium, definition of, 165
re-opened habitats, 179–181 P
rock record, 185 Paleocene-Eocene Thermal Maximum, 172
short-term and long-term refugia, 186–187 Paleoecology, 229, 230, 240–242
260 Index
Palmaria palmata, 178 stratigraphic unit and geologic age,

Park grass experiment, 2 77–79
Patella ferruginea, 174 Plio-Pleistocene fossil record, 216
Patinopecten-Chione-Macoma (PCM) Pohatu and Auckland Islands, 172
community, 150 Policy-oriented education, 212, 214
Pelvetia canaliculata, 168 Problem-solving approach
Permo-Triassic oceans, 171 in CPB education, 212
Pew Charitable Trust, 219 practice-oriented degree programs, 222
Picayune Strand restoration project Project Delivery Team (PDT), 60
CERP, 66 Project Implementation Report (PIR), 60
conservation paleobiology, 67
geographic locations, 64
management collaborative, 67 R
oyster reef development, 68–71 Range shift(s)
project development teams, 67 climate refugia, 168, 169
restoration design, 70, 71 hypothesized methods of survival, 167
roads and canal systems, environmental Real-world experience
problem, 66–67 cross-disciplinary programs, 220
salinity targets, 68 for future conservation paleobiologists, 215
Planetary-scale tipping point, 15 Reduced major axis (RMA) regression, 128
Pleistocene fossil oysters, Mid-Atlantic U.S. Refuge, 165, 182
Cherry Point Refugium, definition of, 165
field and museum sampling, 81 Requests for proposals (RFPs), 59
location, 78, 79 Rock record, 185
stratigraphic unit and geologic age, 78, Ross Sea Region Marine Protected Area, 172
79
vs. colonial and modern oysters
environmental control on oyster size, S
91–93 SAR, see Species–area relationship
growth rates, 85, 87–91 Science advocacy, 12–13
human factors on oyster size, 93–94 Science communication
shell height, 85–88 own values and biases, 13
Holland Point pure scientist to honest broker, 14–18
field and museum sampling, 80–81 stakeholders, 13
location, 78, 79 value biodiversity, 13
oyster size and abundance data, 81, 82 Science mapping
paleosalinity, 82, 84 bibliographic co-authorship visualizations,
paleotemperature, 82–84 230–231
stratigraphic unit and geologic age, bibliographic co-citation visualizations,
79–80 231–232
Lee Creek bibliographic coupling visualizations, 232
field and museum sampling, 81 bibliometric data, 229–230
location, 78–80 in conservation paleobiology, 229–232
stratigraphic unit and geologic age, 79, text co-occurrence visualizations, 231
80 VOSViewer, 230
Stetson Pit Sea surface temperatures (SST), 108
field and museum sampling, 81 Serpukhovian-Bashkirian extinction, 181
location, 78–80 Serpukhovian extinction event, 180
stratigraphic unit and geologic age, 79, Shallow-water tropical corals, 167
80 Smithsonian-Mason School of Conservation,
Wailes Bluff 215
field and museum sampling, 81 Society for Conservation Biology, 213
location, 77–79 Society for Ecological Restoration (SEC), 61
Index 261
South Florida Water Management District SACs, 152, 153

(SFWMD), 65 sea-level drops, 146–147
Species–area curve (SACs), 152, 153 time-averaging, 154–155
Species–area relationship (SAR), transgressive curve, 153
paleontological context T-R cycles, 146–148
application, 145 Species distribution models
biofacies, 157 disadvantage in, 184
climate cooling, 153 for identifying ancient refugia, 182
contiguous gridded plot system, 145 problem with, 185
disturbance hypothesis, 144
EAR curves, 153–154
equilibrium, 155–156 T
equivalent equation, 145 Tectonism, 191
extinction cascade, 156 Text co-occurrence networks
extinction debt, 145–146 average citation frequency, 240
fossil record, 146 CPb-AD dataset, 235, 240–244
habitat diversity hypothesis, 144 modules, 235–240
human-driven habitat loss, 144 Text co-occurrence visualizations, 231
human expansion, 146 Transgressive–regressive cycles (T-R) cycles,
island biogeography model, 144 146–148
Kettleman Hills Tropical Conservation and Development
Etchegoin Formations, 148, 150 program, 216, 219
Jacalitos Formations, 148
Macoma zone, 149–150
Patinopecten zone, 149–150 U
PCM community, 150 University of Florida, 213, 216
Pliocene paleogeographic map, 148, University of North Carolina Wilmington
149 (UNCW), 214
San Joaquin Formation, 148, 149 US Papahānaumokuākea Marine National
Talure Formation, 148 Monument, 171
meta-analysis, 157
methods, 150–151
non-randomly distributed species, 154 V
paleozoic epicontinental seas, 147 Virginia Museum of Natural History (VMNH),
perched faunas, 147 80, 81
Pliocene diversity, 147 VOSViewer, 230, 231
power-function, 145
rain forests, destruction of, 146
relaxation time, 156 W
results, 151–153 Web of Science (WoS), 229–230

(Topics in Geobiology 47) Carrie L. Tyler, Chris L. Schneider - Marine Conservation Paleobiology-Springer International Publishing (2018)

Uploaded by

Copyright:

Available Formats

(Topics in Geobiology 47) Carrie L. Tyler, Chris L. Schneider - Marine Conservation Paleobiology-Springer International Publishing (2018)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Topics in Geobiology 47) Carrie L. Tyler, Chris L. Schneider - Marine Conservation Paleobiology-Springer International Publishing (2018)

Uploaded by

Copyright:

Available Formats

Topics in Geobiology 47

More information about this series at http://www.springer.com/series/6623

Library of Congress Control Number: 2018935181

© Springer International Publishing AG, part of Springer Nature 2018

Printed on acid-free paper

sustainable populations for harvested resources. However, conservation biology is

Oxford, OH, USA Carrie L. Tyler

An Overview of Conservation Paleobiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Training Tomorrow’s Conservation Paleobiologists . . . . . . . . . . . . . . . . . . . . . . . . . 209

Gregory P. Dietl Paleontological Research Institution, Ithaca, NY, USA

Peter D. Roopnarine Invertebrate Zoology and Geology, Institute for Biodiversity

Carrie L. Tyler and Chris L. Schneider

Abstract The field of conservation paleobiology was formally established in

Keywords Conservation science · Conservation biology · Geohistorical data ·

© Springer International Publishing AG, part of Springer Nature 2018 1

1 Defining and Establishing Conservation Paleobiology

The importance of records documenting previous ecosystem states is increasingly

Meine et al. 2006)—Dietl (2016) suggested that it may be beneficial to approach

Table 1 Timeline for the growth and development of conservation paleobiology

2 Data in Conservation Paleobiology

Past 2 million yrs

Past 10,000 yrs

Past 500 yrs

placed on human–environment interactions (Balée 2006; Szabó and Hédl 2011)

Conservation is defined in the Merriam-Webster’s Dictionary as “a careful preser-

Numerous principal objectives of conservation paleobiology have been proposed

Gregory P. Dietl and Karl W. Flessa

Abstract Conservation paleobiology is an applied science. Academically based

Keywords Conservation biology · Conservation science · Objectivity · Science

© Springer International Publishing AG, part of Springer Nature 2018 11

2 To Advocate, or Not to Advocate

3 Speaking Honestly to Power

What does it mean to be a responsible advocate? For a conservation paleobiologist—

4 From Pure Scientist to Honest Broker

makes it difficult to exclude a wide range of future outcomes. In such contexts,

Informative advocacy attempts to expand the range of options, recognizing that

But drought happened. Southwestern water agencies wanted to see if a dormant

Decisions about environmental policies and their implementation are rarely—if

6 Overcoming the Fear Factor

7 Later Is Too Late

In the coming years, conservation paleobiologists will (and should) continue to

Baron N (2016) So you want to change the world? Nature 540:517–519

Jansen A. Smith, Stephen R. Durham, and Gregory P. Dietl

Abstract Marine conservation biologists increasingly recognize the value of long-

© Springer International Publishing AG, part of Springer Nature 2018 23

Keywords Conservation barriers · Conservation paleobiology ·

1 What is “Long Term”?

governmental affiliations. In response to the distribution of respondents from the

3 Survey Responses and What They Mean for Conservation

A total of 90 marine conservation researchers and professionals completed the

protected areas), Technology (n D 46, 27%; e.g., satellite tagging), Mathematics

distribution of timescales on which LTD sources were considered by respondents

respondents indicated that data on timescales of decades or shorter would suffice

Table 1 Respondent ranks 1 2 3 4 5 Unsure Mean

conservation paleobiologists have an immediate and unmistakable opportunity to

for human-dominated ecosystems”—as a solution to this problematic mindset.

The application of LTD to conservation practice is challenging and the predominant

Another barrier to the application of LTD in marine conservation that was

4 Takeaways for Conservation Paleobiologists

Our survey results suggest conservation paleobiologists have much work to do