ORIGINAL RESEARCH

published: 14 September 2020

doi: 10.3389/fmars.2020.575825

Species Distribution Modeling

Predicts Significant Declines in
Coralline Algae Populations Under
Projected Climate Change With
Implications for Conservation Policy
Cornelia Simon-Nutbrown 1,2* , Peter M. Hollingsworth 2 , Teresa F. Fernandes 3 ,
Lisa Kamphausen 4 , John M. Baxter 3,5 and Heidi L. Burdett 1
1
The Lyell Centre for Earth and Marine Science and Technology, Heriot-Watt University, Edinburgh, United Kingdom, 2 Royal
Botanic Garden Edinburgh, Edinburgh, United Kingdom, 3 Institute of Life and Earth Sciences, Heriot-Watt University,
Edited by: Edinburgh, United Kingdom, 4 Scottish Natural Heritage, Great Glen House, Inverness, United Kingdom, 5 School of Biology,
Jonathan Y. S. Leung, Faculty of Science, University of St Andrews, St Andrews, United Kingdom
University of Adelaide, Australia
Reviewed by: Anthropogenic climate change presents a major challenge to coastal ecosystems.
Cristina Pardo Carabias,
Mass population declines or geographic shifts in species ranges are expected to
University of A Coruña, Spain
Chiara Lombardi, occur, potentially leading to wide-scale ecosystem disruption or collapse. This is
Italian National Agency for New particularly important for habitat-forming species such as free-living non-geniculate
Technologies, Energy and Sustainable
Economic Development (ENEA), Italy
coralline algae that aggregate to form large, structurally complex reef-life ecosystems
*Correspondence:
with high associated biodiversity and carbon sequestration capability. Coralline algal
Cornelia Simon-Nutbrown beds have a worldwide distribution, but have recently experienced global declines
[email protected]
due to anthropogenic pressures and changing environmental conditions. However,
Specialty section: the environmental factors controlling coralline algal bed distribution remain poorly
This article was submitted to understood, limiting our ability to make adequate assessments of how populations
Global Change and the Future Ocean,
may change in the future. We constructed the first species distribution model for non-
a section of the journal
Frontiers in Marine Science geniculate coralline algae (focusing on maerl-forming species but including crustose
Received: 24 June 2020 coralline algae associated with coralline algal beds) and showed that bathymetry,
Accepted: 21 August 2020 temperature at the seabed and light availability are the primary environmental drivers of
Published: 14 September 2020
present-day non-geniculate coralline algae distribution. Our model also identifies suitable
Citation:
Simon-Nutbrown C, areas for species presence that currently lack records of occurrence. Large-scale spatial
Hollingsworth PM, Fernandes TF, declines in coralline algal distribution were observed under all IPCC Representative
Kamphausen L, Baxter JM and
Concentration Pathways (ranging from 38% decline under RCP 2.6 up to 84% decline
Burdett HL (2020) Species
Distribution Modeling Predicts under RCP 8.5), with the most rapid rate of decline up to 2050. Refuge populations
Significant Declines in Coralline Algae that may persist under projected climate change were also identified – informing priority
Populations Under Projected Climate
Change With Implications areas for future conservation efforts to maximize the long-term survival of this globally
for Conservation Policy. important ecosystem.
Front. Mar. Sci. 7:575825.
doi: 10.3389/fmars.2020.575825 Keywords: biodiversity, biogeography, climate change, Scotland, ecology, marine conservation, maerl, rhodolith

Frontiers in Marine Science | www.frontiersin.org 1 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

INTRODUCTION In Europe they are protected under the EU Habitats Directive

and the OSPAR Commission in the North Atlantic (under
Rising anthropogenic carbon dioxide (CO2 ) emissions are “OSPAR list of Threatened and/or Declining habitats”) (Hall-
causing rapid environmental change and under some scenarios, Spencer et al., 2008; European Commission, 2018), supported
the mean ocean surface temperature is projected to rise by up to by local implementation such as designation as Priority Marine
4.3◦ C by 2100 (IPCC, 2014). Geographic shifts or declines in the Features in Scotland (Scottish Government, 2018). Seven marine
distribution of species and ecosystems in response to this rapid features were included in the EU Species and Habitats Directive
environmental change have already been observed (e.g., Bellard some of which encompassed habitats where coralline algal beds
et al., 2012; Gallon et al., 2014; Martínez et al., 2018). This can occur (e.g., “Large shallow inlets and bays” and “Sandbanks which
result in significant ecosystem disruption (Hoegh-Guldberg et al., are slightly covered by seawater at all times”) which are listed
2007; Noisette et al., 2013a; Teagle and Smale, 2018), threatening under Annex I. Two maerl-forming species (L. coralloides and
the long-term survival of key ecosystem components and the P. calcareum) are additionally listed under Annex V (The Council
associated biodiversity and ecosystem services they provide. of the European Communities, 1992; European Commission,
Free-living non-geniculate red coralline algae are globally 2016; Riosmena-Rodriguez et al., 2016; Perry et al., 2017b).
distributed ecosystem engineers (Foster, 2001). They give Although the Habitats Directive has since been superseded into
rise to three-dimensional marine habitats, and these so- national law, it is the starting point from which most non-
called “maerl” or “rhodolith” beds have high associated geniculate coralline algal bed conservation and later European
biodiversity (Barbera et al., 2003; Peña et al., 2014) – legislation [e.g., the United Kingdom Biodiversity Action Plan
supporting many rare, endemic or commercially important (JNCC, 2016)] has evolved.
species (Grall and Hall-Spencer, 2003; Wilson et al., 2004; Habitat-forming ecosystem engineers, such as free-living
Nelson, 2009; Riosmena-Rodriguez et al., 2016) – and globally non-geniculate coralline algae, modify their environment in
significant roles in biogeochemical cycling, particularly long- such a way that benefits other species, for example, through
term carbon burial (Burdett et al., 2015; van der Heijden the provision of shelter and food, or by the reduction of
and Kamenos, 2015; Mao et al., 2020). Aggregations of free- environmental stress (Hastings et al., 2007; Peña et al., 2014;
living, non-geniculate coralline algae (maerl-beds), although Sheehan et al., 2015; Martínez et al., 2018). These ecosystems,
found globally, are particularly abundant in the North- therefore, are of high conservation concern and understanding
East Atlantic (from the Arctic Ocean to the wider North how they may react to changing environmental conditions is
Atlantic and the Canary Islands) (Pardo et al., 2014a), important as their loss or geographic shift could pose threats
with Galicia, Brittany, Norway, Ireland, and Scotland being to whole systems (Martínez et al., 2018; Teagle and Smale,
particular strongholds (Hall-Spencer et al., 2010). In this 2018; Cornwall et al., 2019). Projected changes in marine
North-East Atlantic region, beds most abundant up to a environmental conditions, particularly ocean warming (Martin
depth of ∼50 m (Pardo et al., 2014a). There are thought to and Gattuso, 2009; Noisette et al., 2013a; Qui-Minet et al., 2019)
be around 24 maerl-forming species found throughout the and acidification (Porzio et al., 2011; Noisette et al., 2013b;
North-East Atlantic but the most common are Phymatolithon Legrand et al., 2017; Martin and Hall-Spencer, 2017), are expected
calcareum, Lithothamnion coralloides, Lithothamnion glaciale, to have significant impacts on non-geniculate coralline algal
and Lithothamnion tophiforme (Pardo et al., 2014a). These distribution (Brodie et al., 2014). This is due to the consequential
species appear to follow a thermal cline, with L. tophiforme negative impacts on species-specific thermal niches, reduced
being limited to Arctic waters, L. glaciale occurring in Arctic calcification (Burdett et al., 2012a, 2018; Kamenos et al.,
and sub-Arctic areas, P. calcareum stretching from the sub- 2013; Qui-Minet et al., 2019) and potential overgrowth by
Arctic to tropical waters and L. coralloides occurring from the opportunistic non-calcifying algae, as has been observed during
Southern British Isles to the Canary Islands and throughout the coral reef degradation (Hughes et al., 2007).
Mediterranean (Hall-Spencer et al., 2010; Pardo et al., 2014a). Non-geniculate coralline algal beds form biogenic reefs, which
However, there are uncertainties regarding species identification provide a myriad of ecosystem services globally that are often
and hence species’ ranges, as morphological identification is overlooked and their potential underestimated (Bindoff et al.,
difficult and few widespread molecular studies have been 2019). These bed systems are biodiversity hotspots forming a
carried out to-date. Poor species identification additionally living layer over sediments that provides habitat and supports
makes identifying species-specific environmental requirements high numbers of species including many invertebrate phyla
challenging. Current evidence suggests that L. glaciale mostly (crustaceans, echinoderms, polychaetes, and mollusks), fleshy
occurs between 6–50 m (Perry et al., 2017a; Schoenrock et al., macroalgae, and various species which burrow either into
2018) while L. coralloides appears to be more restricted to individual free-living algae or the sediments below the bed
shallower depths, rarely being recorded below 30 m (Perry (Jackson et al., 2004; Hall-Spencer et al., 2008, 2010; Foster et al.,
et al., 2017a). P. calcareum is also more frequently recorded in 2013). When compared to adjacent habitat types it was found
shallower waters (above 30 m) (Perry et al., 2017a), but sporadic that coralline algal beds had nearly twice the species richness
deep-water observations are known [e.g., 80 m in the Pontian (Steller et al., 2003). Similarly, non-geniculate coralline algal
Archipelago, Mediterranean (Basso, 1998)]. beds support numerous commercial fish species, often forming
Currently, all coralline algae beds are listed as “Vulnerable” nursery habitats for economically important species such as
or “Endangered” on the IUCN Red List (Gubbay et al., 2016). cod (Gadus morhua) and hake (Melanogrammus aeglefinus)

Frontiers in Marine Science | www.frontiersin.org 2 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

(Kamenos et al., 2004; Hall-Spencer et al., 2006; Riosmena- justifications (Pearson, 2010; Elith et al., 2011; Merow et al.,
Rodriguez et al., 2016; Elliott et al., 2017). However, despite their 2013). MaxEnt v3.4.1 (Phillips et al., 2006) was run through
importance there is still a paucity of knowledge surrounding these the R v 3.6.1 (R Core Team, 2019) package DISMO v1.4
ecosystems and how a rapidly changing climate will affect them (Hijmans, 2017) to model the species niche and construct a
(Bindoff et al., 2019). probabilistic distribution map of non-geniculate coralline algal
Species distribution models (SDMs) or environmental niche beds around Scotland.
models (ENMs) are commonly used tools to estimate the
relationships between species and environmental or spatial Non-geniculate Coralline Algae
factors (Elith et al., 2011; Merow et al., 2013). SDMs represent an Occurrence Records
important addition to direct environmental change experiments Occurrence records for the study region were obtained from the
on non-geniculate coralline algae and their associated habitats Geodatabase of Marine Features in Scotland (GeMS) (Scottish
which are difficult to extrapolate to range-wide predictions Natural Heritage, 2019a) under the Open Government and
(Wernberg et al., 2012), and do not take into account the spatial Creative Commons License. Geographic locations and study area
variability of projected changes in environmental conditions. are provided in Supplementary Figure S1. The annually updated
SDMs are capable of modeling both small-scale local responses database includes observed, georeferenced occurrence records of
and range-wide changes, provided that datasets are available at Priority Marine Features in Scotland from 1970 to 2017 curated
a suitable scale. by Scottish Natural Heritage (SNH). “Maerl beds” are one of these
Here, we constructed the first SDM for coralline algal features. For this study, cut-offs were imposed on occurrence
beds, focusing on Scottish regional waters – a European data by removing entries from before 1990. This was done to
stronghold for this habitat (Hall-Spencer et al., 2008) and of increase reliability, as older records may have been subject to
conservation priority at national and international scales (Hall- different environmental conditions than recent records – for
Spencer et al., 2008; JNCC, 2016; European Commission, 2018; example due to construction of causeways changing tidal patterns
Scottish Government, 2018). This enabled us to identify: (1) the [known to have happened in several locations, for example the
primary environmental drivers of present-day coralline algal bed Vatersay causeway (1989) (Comhairle nan Eilean Siar, 2020b)
distribution, (2) previously unknown areas of occurrence, (3) or the Benbecula causeway (1983) (Comhairle nan Eilean Siar,
regional-scale impacts of projected climate change on coralline 2020a)]. Similarly, where there was evidence of a bed no longer
algal bed distribution, and (4) refuge populations tolerant supporting live individuals the occurrence point was removed
to projected climate change. Synthesizing this information prior to analysis.
allowed us to identify areas of priority conservation effort to Keyword searches in the GeMS database for “maerl” yielded
maximize the likelihood of long-term survival of this globally species-specific records of free-living coralline algae species
important marine habitat. (i.e., “L. glaciale” or “P. calcareum”) and records with no
species identification (“maerl beds”). All records pertaining to
aggregations of free-living non-geniculate coralline algae are
MATERIALS AND METHODS labeled as “maerl bed” in GeMS and therefore no other search
terms were needed. Within the study region species ranges
Study Area overlap and species identification in coralline algae without
Scotland is a European stronghold for non-geniculate coralline molecular analyses can be difficult and unreliable (Pardo et al.,
algae beds. Consequently, it was selected as the focus area 2014a,b). In order to maintain the maximum number of results
for distribution modeling in this study. The study area runs for the model, no attempt was therefore made to split these
from the Solway Firth in the southeast, to the northern coasts records up into species-specific groups, however; it is assumed
of the Shetland Isles, including the Outer Hebrides in the that the model will target the two main species in this region
west (Supplementary Figure S1). Within this region “maerl P. calcareum and L. glaciale. The database searching yielded 1430
beds” (i.e., aggregations of non-geniculate coralline algae) are occurrence records.
Priority Marine Features and protected under Scottish legislation Although all records in the GeMS database are of free-living,
(Scottish Government, 2018), ten marine protected areas around non-geniculate coralline algae, not enough is known about the
the Scottish coastline are designated for the protection of differentiation between this and crustose coralline algae (CCA)
coralline algal beds (referred to in legislation as “maerl beds”). to conclusively say the model will only predict the distribution
Despite this, the true extent of these systems in Scottish waters of free-living non-geniculate coralline algae, especially when
remains unknown, and even less is known about the factors that considering free-living specimens that are grown around a non-
govern their local distribution. algal core.
Observed occurrence records within the database which
Species Distribution Model were not the result of a targeted survey, e.g., SeaSearch data1 ,
Maximum entropy (MaxEnt) modeling was chosen to construct a have inherent amounts of observer and spatial bias (Veloz,
species distribution model (SDM) for maerl-forming species (i.e., 2009; Fithian et al., 2015). Similarly, some beds have multiple
non-geniculate coralline algae) as it is effective with presence- (but geographically close) location points, creating further
only data, is well regarded in the literature and can be highly
modified so all modeling decisions are based on biological 1
http://www.seasearch.org.uk/

Frontiers in Marine Science | www.frontiersin.org 3 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

spatial bias (Veloz, 2009). To eliminate the subsequent spatial- with pH and DA = 0.0757; without pH and DA = 0.108]
autocorrelation a 2-dimensional kernel density estimation of the (Supplementary Table S3). Average omission rates for the
data was created in R (R Core Team, 2019) using the package minimum training presence were higher for models without
“spatstat” v1.61-0 (Baddeley et al., 2019). This was then used in pH and DA compared to those with them (0.310 and 0.263,
the Java application OccurenceThinner (Verbruggen et al., 2013) respectively) thus also suggesting models were more over fit
to reduce geographical sampling bias leaving 314 occurrence without pH and DA (Pearson et al., 2007; Muscarella et al., 2014)
points in the model. (Supplementary Table S3). Model evaluation was conducted
in the R package ENMeval v0.3.0 (Muscarella et al., 2018).
Environmental Variables Model performance was assessed based on standard statistical
Environmental layers were accessed from bio-ORACLE v2 assessments including AUC, receiver operating characteristics
(Tyberghein et al., 2012). The datasets have a spatial resolution (ROC) and Akaike Information Criterion (AIC) (Muscarella
of five arc minutes (Tyberghein et al., 2012; Assis et al., 2018). et al., 2014; Steen, 2019).
Bathymetry data were accessed through Bio-ORACLE from
MARSPEC (Assis et al., 2018) and masked to the same resolution Spatial Distribution Model Settings
as the Bio-ORACLE datasets (Assis et al., 2018). ENMeval was further used to select model parameters that
Environmental variables were selected for their potential enabled the most efficient model performance (Supplementary
links to coralline algae distribution. Initially 21 datasets were Material). The optimal settings were found to be a regularization
considered (Supplementary Table S1). Variables were examined multiplier of one, and all feature classes. Models were clamped
for collinearity which may interfere with the performance of to training conditions to avoid extrapolation uncertainty. Once
the model (Júnior and Nóbrega, 2018). This was done by these settings were established the model was run in the R
calculating the Pearson correlation coefficient (r) between the package Dismo v1.4 (Hijmans, 2017). The contribution of each
datasets; datasets were considered collinear when −0.6 ≤ r ≥ 0.6 environmental variable was assessed by excluding it from the
(Assis et al., 2018; Júnior and Nóbrega, 2018). A high degree model and assessing the performance of the model without
of collinearity was observed between the 16 available variables. it (jack-knifing). Forty per cent of the occurrence data points
Where inclusion of maximum and minimum datasets created were randomly held back for model calibration and the model
collinearity with another dataset, the mean datasets were used was trained on the remaining 60% of the occurrence data.
instead. Following analysis for collinearity, the most ecologically The resultant model performed well under cross validation and
significant datasets were selected, according to literature evidence evaluation statistics (Supplementary Table S3). In the final
(e.g., Dutertre et al., 2014). These were: bathymetry (m), model, the AUC value was 0.959 ± 0.008 (mean ± SD), indicating
pH, mean diffuse attenuation coefficient (DA) (m−1 ), mean high predictive ability. Some areas of predicted distribution were
dissolved nitrate concentration (µmol L−1 ), minimum benthic not covered by presence records. Field validation of two positive
temperature (◦ C) (i.e., average minimum temperature recorded prediction sites were found to be true (Daliburgh, South Uist
at average seabed depth across the coastal shelf ranging from 0 [57.160913, −7.408005] and Kildonan [57.218810, −7.427824])
to 200 m depth), mean photo-active radiation (PAR) at seabed and one modeled “absence” site was confirmed to not have any
(E m−2 year−1 ), mean seabed salinity (PSS), maximum current free-living coralline algae or CCA present (Loch Eyenort, South
velocity at seabed (m s−1 ) and minimum current velocity at Uist [around 57.230275, −7.3514]), providing further evidence of
seabed (m s−1 ). model reliability.
The modeled benthic datasets used here contain some
inherent collinearity, however, this is negligible in the Bio- Threshold Selection
ORACLE datasets as in situ quality control has been carried out Implementation of thresholds was carried out in R to transform
by the Global Ocean Data Analyses Project (Assis et al., 2018) and logarithmic model predictions of probability of presence of
high agreement found between the interpolated benthic datasets non-geniculate coralline algae to binary presences and absences
and in situ data gathered (Assis et al., 2018). For pH and DA, (Supplementary Material).
benthic datasets were not available, so sea surface datasets were
used. Since these may not accurately represent benthic conditions Climate Projections
(where non-geniculate coralline algal beds are found), test models Climate projection datasets were sourced from Bio-ORACLE for
were run with and without these datasets and the probabilistic Representative Concentration Pathway (RCP) scenarios 2.6, 4.5,
distribution maps produced compared against known occurrence 6.0, and 8.5 for the years 2050 and 2100 (Assis et al., 2018).
data. The results of these test models showed that the overall These are all based on the IPCC Coupled Model Intercomparison
models were more statistically robust with the inclusion of Project (CMIP) 5 model for climate scenarios, ranging from
surface datasets. pH was consistently found to be the variable that a significant reduction in greenhouse gas emissions (RCP 2.6)
contributed fourth highest to the model (10.8% of the observed to a continuation at the current level (RCP 8.5) (Assis et al.,
distribution) and is known to be environmentally relevant to 2018). Where available, the environmental variables used to
non-geniculate coralline algal survival (McCoy and Kamenos, train the MaxEnt SDM were replaced with the projected future
2015), and so warranted inclusion in the model. Additionally, levels – this was possible for benthic temperature, salinity and
when run without pH and DA, overfitting (Pearson et al., 2007; current velocity (Assis et al., 2018) (Supplementary Table S2).
Muscarella et al., 2014) was observed [Area Under Curve (AUC) Projection datasets for other environmental variables (pH,

Frontiers in Marine Science | www.frontiersin.org 4 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

diffuse attenuation, nitrate concentration, photoactive radiation and the north of Lewis (Figure 2A). Apart from under RCP
at seabed and bathymetry), were not available, so present-day 2.6, suitable habitat in the Outer Hebrides was predicted to be
datasets were used. restricted to small areas in the north of Lewis by 2100 and was
completely lost under RCP 8.5 (Figure 3). By 2050, the Isle of
Skye (Inner Hebrides) – was predicted to no longer support non-
RESULTS geniculate coralline algae under RCPs 6.0 and 8.5 (Figures 2D,E).
Interestingly, the area east of Skye, which was deemed unsuitable
Present Day Coralline Algal Distribution for non-geniculate coralline algal beds under RCP 6.0 in 2050
The top three primary environmental drivers (total was retained under RCP 6.0 in 2100, but then lost under RCP
explanation = 85%) of non-geniculate coralline algal bed 8.5 (2100) (Figures 3D,E).
distribution were bathymetry (51.8%), minimum benthic Some areas were predicted to maintain non-geniculate
temperature (17.4%) and PAR at the seabed (15.8%). Sea surface coralline algal populations throughout the coming decades. In
pH accounted for an additional 10.8% of the association between the north of mainland Scotland (Loch Laxford and Ullapool/Loch
environmental variables and non-geniculate coralline algal Broom) and around Orkney and Shetland, high probabilities of
distribution. Coralline algal beds were predicted from the coralline algae occurrence (>0.7) were maintained by 2100, even
intertidal, down to 40 m depth, within a minimum temperature under RCP 8.5 (Figure 3 and Supplementary Figures S2, S3).
range of 2–9◦ C, a PAR range of 0–25 (E m−2 yr−1 ) and a pH A relatively high probability (>0.7) of persistence in 2100 under
range of 8.16–8.21. Predicted non-geniculate coralline algae all RCPs was also predicted for Loch Sunart (mainland Scotland)
habitat was spatially heterogeneous, characterized by a general and north Bute (Inner Hebrides) (Figure 3 and Supplementary
east/west divide. The most suitable areas for non-geniculate Figure S3). Under RCP 2.6, 6.0, a refuge population remained
coralline algal habitats were predominantly located in the coastal on the west of Islay by 2100 (Figures 3A–C) but there was
waters and sea lochs of the west coast of mainland Scotland, lower probability of it being maintained under RCP 8.5 (0.5–
around Orkney, and Shetland (north of mainland Scotland) and 0.7) (Figure 3 and Supplementary Figure S3). There is also a
around the southern islands of the Outer Hebrides (Figure 1). small patch of habitat near Carnigaan (south-west mainland) that
The east coast of Scotland is predominantly devoid of suitable retains a higher probability of occurrence even under RCP 8.5 in
areas for non-geniculate coralline algal habitats, except for the 2100 (>0.7) (Figure 3E and Supplementary Figure S3).
Moray Firth on the northeast mainland (Figure 1). The total area Similarly, under all RCPs the inner Moray Firth retains
suitable for non-geniculate coralline algal bed presence during several areas of suitable habitat by 2100 (>0.7) (Figure 3E and
present-day conditions was predicted to be 7130 km2 . Supplementary Figure S3).

Regional-Scale Distribution Under

DISCUSSION
Projected Environmental Change
Under all RCPs, a large decrease in suitable habitat is predicted Free-living non-geniculate coralline algal beds are a globally
with at least a 38% decline by 2050 (Table 1). The most distributed, ecologically important ecosystem (Pardo et al., 2014a;
immediate losses in suitable habitat are in southern areas of Riosmena-Rodriguez, 2016). Using regional-scale modeling,
mainland Scotland under all RCPs (Figures 2, 3). At the most bathymetry, water temperature and light availability were found
extreme, under RCP 8.5 a predicted 83.7% reduction in suitable to be the primary environmental drivers of non-geniculate
habitat for non-geniculate coralline algae was found in 2100 coralline algal bed distribution. Projected environmental change
compared to the present-day (Figure 3E and Table 1). The north- is predicted to have a severe negative impact on overall
west mainland coast of Scotland, patches around the Northern regional-scale coralline algal distribution in Scotland, although
Islands and the area around the inner Moray Firth remained spatial heterogeneity in the magnitude of environmental change
strongholds for non-geniculate coralline algae under all future with the current distributional range has allowed potential
climate projections (Figures 2, 3). No habitat that was previously future refugia to be identified – areas we propose for priority
unsuitable for non-geniculate coralline algal beds (present day) conservation status.
became suitable under any of the RCP scenarios. This regional-scale model highlights the need to understand
how small-scale environmental changes affect non-geniculate
Regional Variation in Distribution Under coralline algae and how beds may be adapted to very localized
Projected Climate Change conditions. Further investigation and understanding of both are
Some areas were predicted to be more negatively affected, in key to fully understanding the implications of climate change for
terms of coralline algal distribution, by projected environmental non-geniculate coralline algae populations globally.
change. This spatial variability in persistence was exacerbated
under higher-emission scenarios. The Outer Hebrides – currently Spatial Distributions Driven by
one of the most important regions for non-geniculate coralline Environmental Heterogeneity
algae – were predicted to experience a 29% decline in coralline We have identified a regional-scale east/west division in coralline
algal distribution by 2050 under RCP 2.6, leaving only small algal habitat. Coralline algae have not been extensively observed
patches remaining in the Sound of Barra, west of North Uist on the east coast, despite extensive surveys (both grab and

Frontiers in Marine Science | www.frontiersin.org 5 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

FIGURE 1 | Present-day modeled occurrence of coralline algae around Scotland as (A) probability of presence on a log likelihood scale ranging from 1 (dark red) to
0 (white), and (B) the probabilistic distribution with threshold of occurrence applied, where orange indicates presence while white (background) indicates absence.

TABLE 1 | Predicted coralline algal bed extent around Scotland in the present-day environmental conditions. Higher temperatures on the east coast,
and under Representative Concentration Pathway (RCP) emission scenarios in the
coupled with lower current velocities, are likely to be the reason
years 2050 and 2100.
the region is predominantly devoid of non-geniculate coralline
RCP Scenario Year Total predicted Percentage decline from algal bed populations. It is noteworthy that overall, the differences
area (km2 ) present-day (%) in environmental conditions between the east and west coasts of
Scotland are relatively minor, indicating how sensitive these non-
Present-day 7130
geniculate coralline algal populations are to their local conditions.
2.6 2050 4446 38
2100 2640 63
Fine sediments are known to be detrimental to coralline algal
4.5 2050 4029 44
survival due to smothering (Fabricius and De’ath, 2001). This
2100 2223 69
may explain why the only predicted east-coast population (inner
6.0 2050 3380 53
Moral Firth) is not backed up by records of non-geniculate
2100 2130 70
coralline algal presence – this area is comprised of patchy
8.5 2050 2778 61
gravel and sand (British Geological Survey, 2020). Where this
2100 1158 84
mixed sediment is present on the west coast, coralline algae are
commonly only found in areas of gravel (e.g., Sound of Barra,
Outer Hebrides). This may be further compounded by dredging
and shipping activities – extensive in Scotland (Fox et al., 2015) –
SCUBA) in the area (e.g., Dargie, 2001), and as the model leading to a resuspension and subsequent smothering of suitable
also identifies this area as predominantly unsuitable, with a bottom-types for coralline algal colonization. Unfortunately,
low probability of presence, this is unlikely to purely be due substrate-type could not be included in the model due to
to a lack of colonization opportunity in this area, but more to incompatibilities in the dataset resolution.

Frontiers in Marine Science | www.frontiersin.org 6 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

FIGURE 2 | Threshold maps of coralline algal bed distribution around Scotland in (A) the present-day and by 2050 under (B) RCP 2.6, (C) RCP 4.5, (D) RCP 6.0,
and (E) RCP 8.5. Orange indicates predicted presence of coralline algal beds.

Frontiers in Marine Science | www.frontiersin.org 7 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

FIGURE 3 | Threshold maps of coralline algal bed distribution around Scotland in (A) the present-day and by 2100 under (B) RCP 2.6, (C) RCP 4.5, (D) RCP 6.0,
and (E) RCP 8.5. Orange indicates predicted presence of coralline algal beds.

Frontiers in Marine Science | www.frontiersin.org 8 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

Looking forward, the ability to identify local areas within a et al., 2019) is at its northerly limit in Scotland, suggesting
regional space which have the potential to support non-geniculate a potential capacity to tolerate warming [although this has
coralline algae but where it is currently not present highlights not been supported by laboratory experimentation (e.g., Qui-
a potential opportunity for the translocation of populations Minet et al., 2019)]. Conversely, L. glaciale is at its southerly
as climate change progresses (National Species Reintroduction limit in Scotland (Pardo et al., 2014a) and is known to be
Forum, 2014). This is particularly pertinent given that no new prevalent across the Atlantic Arctic (Schoenrock et al., 2018).
areas of suitable habitat were identified under climate change Calcification and temperature in L. glaciale are known to
scenarios, in line with other previous distributional studies be negatively related (Kamenos and Law, 2010), suggesting
on other calcified marine organisms such as corals (Freeman projected warming over the coming century will negatively
et al., 2013). Reef restoration is of growing conservation interest, impact L. glaciale populations, supporting the habitat declines
particularly on tropical coral reefs (Suggett et al., 2019), but this seen here. Species-specific responses to projected environmental
is likely to be even more challenging for maerl-forming species change will thus determine the magnitude of coralline algal
due to their slow growth rates (Adey and McKibbin, 1970; Blake ecosystem change but input data at the species level was not
and Maggs, 2003; Brodie et al., 2014; Pardo et al., 2019), even if available for this study. However, it is clear from the results
environmentally suitable areas can be identified. presented here that coralline algal beds at the edges of their
range (be that geographically, or by depth) face severe threats
Bathymetry and Light Availability Limit over the coming century and should be the focus of immediate
conservation management.
Depth Distribution
Despite low-light adaption (Burdett et al., 2012b) and known
Future Drivers of Distributional Change
worldwide deep-water occurrences (Peña and Barbara, 2009;
Under all RCPs, there was a significant decrease in the amount
Hall-Spencer et al., 2010; Friedlander et al., 2014; Pardo et al.,
of suitable areas for non-geniculate coralline algal habitat, but
2014a), non-geniculate coralline algal survival was limited to
this was accompanied by high spatial variability. Warming
<40 m depth in this region of the NE Atlantic, with the
projections for our study region were characterized by latitudinal
majority found between 5 and 20 m depth. Mitigation of
trends in the magnitude of projected warming (higher in the
some effects of climate change may be possible for marine
south, lower in the north; Supplementary Figure S4) combined
organisms that can shift their habitats along a vertical gradient,
with localized areas of little warming compared to the overall
i.e., moving into deeper cooler waters (Jorda et al., 2020).
trend – this local-regional scale heterogeneity in warming
However, our results do not predict this kind of range shift
has also been projected globally (IPCC, 2014). Importantly,
to be available for non-geniculate coralline algae because of
variability in warming led to the persistence of refuge populations
light limitation. Increased Arctic melt-water introduction (Pan
throughout the study region. However, this may increase habitat
et al., 2019), may enable a downward expansion or movement
fragmentation, which can lead to inbreeding depression and
of non-geniculate coralline algae populations due to clearing
decreasing population health (Pavlova et al., 2017; Pardo et al.,
of the water column. If this was to occur, there would also
2019), ultimately weakening the likelihood of long-term survival
be knock-on implications for the biodiversity associated with
throughout the coming century and beyond.
this habitat, which may be depth limited or increased light
Clamping of the SDM to conditions only encountered
penetration may lead to macroalgae overgrowth. This could
in the training data mean that the results presented here
result in major shifts in community structure and subsequent
perhaps represent a “worst-case scenario” for the effects of
changes in ecosystems services, as have been seen and modeled
environmental change – it is assumed that environmental
in coastal fish and invertebrate communities of Australia
conditions beyond present-day variability will not be suitable
(Cheung et al., 2012).
for non-geniculate coralline algal habitat in the future. Present-
day latitudinal distributions, especially for P. calcareum, suggest
Temperature Exerts Control on this may not be representative, but the capacity for coralline
Latitudinal Extent algae to shift poleward is expected to be minimal because of
The second most important environmental driver of distribution their slow growth rates and limited dispersal abilities (Brodie
was minimum benthic temperature – known to be a key factor et al., 2014). This contrasts with other ecosystem engineers
in determining non-geniculate coralline algal species extent such as kelp, for which poleward shifts have already been
globally (McCoy and Kamenos, 2015). Where regional-scale recorded around the world (Smale, 2020). Likewise, there
non-geniculate coralline algal communities consist of species may be some opportunity for in situ adaptation to changing
with differing thermal envelopes, future distributional changes conditions, although there is little supporting evidence to
may become more complex. This is the case here: the two suggest this is likely to occur within the short-time frame
dominant free-living species – P. calcareum and L. glaciale – involved, and this is confounded by a general lack of
have different distributional ranges that overlap but show knowledge of population dynamics and the extent of sexual
different thermal tolerances; this may affect their survival under reproduction in non-geniculate coralline algal populations
projected climate change (especially warming). P. calcareum, (Pardo et al., 2019).
one of the dominant free-living non-geniculate coralline algal Future non-geniculate coralline algal bed distribution
species throughout Europe (Pardo et al., 2014a; Qui-Minet may also be affected by projected declines in ocean pH

Frontiers in Marine Science | www.frontiersin.org 9 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

(McCoy and Kamenos, 2015) – red coralline algae and their In our study region, although the majority of known
associated communities are generally considered to be highly non-geniculate coralline algal beds are located within the
sensitive to low pH (Martin and Gattuso, 2009; Burdett et al., Scottish MPA and SAC network (Supplementary Figure S6),
2012a, 2018; Kroeker et al., 2012; Kamenos et al., 2013, 2016; only ten of 217 are designated for free-living coralline algae
Qui-Minet et al., 2019). Thus, it would be expected that projected bed protection (as “maerl beds”) (Fetlar to Haroldswick,
declines in ocean pH would further limit the distributional extent Wyre and Rousay Sounds, Loch Laxford, Wester Ross, Loch
of non-geniculate coralline algae, perhaps counter-balancing Carron, Loch nam Madadh, Sound of Barra, Loch Sween,
the effects of clamping in the model. Unfortunately, a pH South Arran and Luce Bay, and Sands) (Supplementary
projection dataset at a suitable spatial resolution and extent was Figure S6). Even in these areas some permitted activities
not available for inclusion in the projection modeling and so (e.g., licensed fishing) may detrimentally affect this ecosystem,
could not be tested here. Scottish Natural Heritage (2019b). The placement of non-
geniculate coralline algae-specific MPAs does not cover full
Wider Ecosystem Implications present-day extent, and only 20% of the potential future
Reduced coralline algal habitat as a result of environmental refuge populations are currently covered (Supplementary
change will also have significant effects on associated species Figures S7, S8).
and biogeochemical cycling. This may result in large community Connectivity between refugia and neighboring habitats may
shifts (Brodie et al., 2014), as has been previously observed e.g., strengthen the future persistence of non-geniculate coralline
during the Little Ice Age/Medieval Warm Period transition (Mao algal beds, similar to the “spill over” of ecosystem services seen
et al., 2020). Coralline algal beds are a globally important store of in other MPA networks (Lynham et al., 2020). We therefore
blue carbon (van der Heijden and Kamenos, 2015; Porter et al., propose that the refuge areas (Loch Laxford, inner Moray
2020), with the potential to store the most carbon per unit area Firth, north of Arran, north of Islay, Mainland Orkney and
of all coastal ecosystems in the study area (Burrows et al., 2014, Mainland Shetland) should be of high conservation priority since
2017). The magnitude of stored carbon in coralline algal beds they are expected to offer persistently suitable habitat for the
is known to be positively related to water temperature (Mao coming decades – expected to be the period of most rapid
et al., 2020), suggesting areas of future coralline algal persistence distributional decline. In support of this, a current consultation
may also experience a proportional increase in carbon storage. by the Scottish Government and Marine Scotland, supported by
However, in areas of coralline algal bed loss, disturbance to the Scottish Natural Heritage, aims to develop fisheries management
system may release carbon that had been previously locked away measures to protect particularly sensitive benthic habitats and
for centuries if not millennia (Macreadie et al., 2019). Conversely, PMFs (including coralline algal beds) outside of MPAs (Marine
without disturbance, the persistence of a dead skeletal framework Scotland, 2018; Scottish Government, 2019).
may enable the continued accumulation of incoming organic
carbon and maintain sediment carbon stability, albeit without
photosynthetic carbon sequestration. SUMMARY AND CONCLUSION
The value of these non-geniculate coralline algal habitats,
along with other biogenic reef habitats, has been broadly Non-geniculate coralline algal beds are a globally important
overlooked to-date (Sunday et al., 2017; Bindoff et al., ecosystem for biodiversity, socio-economic benefit and carbon
2019). Globally, the value of these habitats as ecosystem storage. Scotland is a European regional stronghold for these
engineers, biodiversity hotspots and nursey habitats is still coralline algal beds, but here we demonstrate their high sensitivity
underestimated and under-protected despite the growing threat to projected climate change, and demonstrate the value of
they face due to increasingly changing climates (Bindoff et al., SDMs for understanding regional-scale environmental drivers
2019). Understanding, the environmental conditions needed to of their distribution. Projected climate change over the coming
maintain these habitats and the degree to which they are locally century is predicted to leave fragmented refugia that may be
adapted, emphasizes the need for better protection and further more vulnerable to long-term survival and therefore should be
study of these fragile but globally important systems (Kamenos a priority for conservation management for favorable conditions.
et al., 2008b; Sheehan et al., 2015; Sunday et al., 2017; Bindoff Applications of this SDM approach to free-living, non-geniculate
et al., 2019). coralline algae in other regions around the world will enable
a global network of distributional change to be determined,
Policy Implications facilitating international-scale conservation decisions and an
Few marine protected areas (MPAs) or special areas of assessment of the knock-on implications for the future of
conservation (SACs) around the world are specifically designated coralline algal ecosystem service provision.
for the protection of non-geniculate coralline algal beds (Barbera This paper highlights the need for a greater understanding of
et al., 2003), but the results from our study suggest their long- non-geniculate coralline algal bed habitats in order to prepare
term large-scale distribution is at risk. Although conservation for various climate futures. Increased survey activity in order
management is not able to directly protect against changing to better establish true presences and absences of beds, and
environmental conditions, it does lessen the impact of other high resolution data on environmental conditions would be
pressures, enhancing the potential for climate change-related appropriate places to start and allow for more detailed and in
mitigation and/or adaptation (Roberts et al., 2017). depth modeling to be carried out.

Frontiers in Marine Science | www.frontiersin.org 10 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

DATA AVAILABILITY STATEMENT on methods and provided detailed edits, and feedback
on the manuscript.
Publicly available datasets were analyzed in this study. This data
can be found here: Geodatabase of Marine Features in Scotland
(GeMS) (https://data.gov.uk/dataset/0e78afea-ac1e-4080-8758- FUNDING
980f2d5cff6d/gems-scottish-priority-marine-features-pmf). Bio-
ORACLE (https://www.bio-oracle.org/). This work was funded by a NERC iCASE studentship award to
HB, PH, JB, and TF (NE/R007233/1). The Royal Botanic Garden
Edinburgh acknowledges funding support from the Scottish
Government Rural and Environment Sciences and Analytical
AUTHOR CONTRIBUTIONS Services Division (RESAS). This is a contribution to the Scottish
Blue Carbon Forum.
HB, TF, JB, PH, and CS-N designed the experiment. LK and
CS-N identified and accessed the data. CS-N carried out
modeling and analysis, with help from HB, and wrote the SUPPLEMENTARY MATERIAL
manuscript. LK and JB provided information on conservation
management in Scotland and provided links to policy The Supplementary Material for this article can be found
organizations. All authors have confirmed and approved online at: https://www.frontiersin.org/articles/10.3389/fmars.
the submitted version of the manuscript, provided input 2020.575825/full#supplementary-material

REFERENCES Burdett, H. L., Hatton, A. D., and Kamenos, N. A. (2015). Coralline algae as
a globally significant pool of marine dimethylated sulfur. Glob. Biogeochem.
Adey, W. H., and McKibbin, D. L. (1970). Studies on the maerl species Cycles 29, 1845–1853. doi: 10.1002/2015GB005274
Phymatolithon calcareum (Pallas) nov. comb. and Lithothamnium coralloides Burdett, H. L., Hennige, S. J., Francis, F. T.-Y., and Kamenos, N. A. (2012b). The
Crouan in the Ria de Vigo. Bot. Mar. 13, 100–106. photosynthetic characteristics of red coralline algae, determined using pulse
Assis, J., Tyberghein, L., Bosch, S., Verbruggen, H., Serrão, E. A., and amplitude modulation (PAM) fluorometry. Bot. Mar. 55, 499–509. doi: 10.1515/
Clerck, O. D. (2018). Bio-ORACLE v2.0: extending marine data layers for bot-2012-0135
bioclimatic modelling. Glob. Ecol. Biogeogr. 27, 277–284. doi: 10.1111/geb. Burrows, M. T., Hughes, D. J., Austin, W. E. N., Smeaton, C., Hicks, N., Howe, J. A.,
12693 et al. (2017). Assessment of blue carbon resources in Scotland’s inshore marine
Baddeley, A., Turner, R., and Rubak, E. (2019). CRAN - Package Spatstat. protected area network. Scott. Nat. Herit. Comm. Rep. 957:283.
Available online at: https://cran.r-project.org/web/packages/spatstat/index. Burrows, M. T., Kamenos, N. A., Hughes, D. J., Stahl, H., Howe, J. A., and Tett,
html (accessed November 8, 2019). P. (2014). Assessment of carbon budgets and potential blue carbon stores
Barbera, C., Bordehore, C., Borg, J. A., Glémarec, M., Grall, J., Hall-Spencer, in Scotland’s coastal and marine environment. Scott. Nat. Herit. Comm. Rep.
J. M., et al. (2003). Conservation and management of northeast Atlantic and 761:90.
Mediterranean maerl beds. Aquat. Conserv. Mar. Freshw. Ecosyst. 13, S65–S76. Cheung, W. W. L., Meeuwig, J. J., Feng, M., Harvey, E., Lam, V. W. Y.,
doi: 10.1002/aqc.569 Langlois, T., et al. (2012). Climate-change induced tropicalisation of marine
Basso, D. (1998). Deep rhodolith distribution in the Pontian Islands, Italy: a model communities in Western Australia. Mar. Freshw. Res. 63, 415. doi: 10.1071/MF
for the paleoecology of a temperate sea. Palaeogeogr. Palaeoclimatol. Palaeoecol. 11205
137, 173–187. doi: 10.1016/s0031-0182(97)00099-0 Comhairle nan Eilean Siar (2020a). South Ford Causeway. Available online
Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., and Courchamp, F. (2012). at: https://www.cne-siar.gov.uk/roads-travel-and-parking/bridges-causeways-
Impacts of climate change on the future of biodiversity. Ecol. Lett. 15, 365–377. and-ferries/south-ford-causeway/ (accessed May 14, 2020).
doi: 10.1111/j.1461-0248.2011.01736.x Comhairle nan Eilean Siar (2020b). Vatersay Causeway. Vatersay Causew.
Bindoff, N., Cheung, W. W. L., Kairo, J. G., Arístegui, J., Guinder, V. A., Available online at: https://www.cne-siar.gov.uk/roads-travel-and-parking/
Hallberg, R., et al. (2019). Changing Ocean, Marine Ecosystems, and Dependent bridges-causeways-and-ferries/vatersay-causeway/ (accessed May 14, 2020).
Communities. . Available online at: https://www.ipcc.ch/srocc/chapter/chapter- Cornwall, C. E., Diaz-Pulido, G., and Comeau, S. (2019). Impacts of ocean
5/ (accessed August 3, 2020). warming on coralline Algal calcification: meta-analysis, knowledge gaps, and
Blake, C., and Maggs, C. (2003). Comparative growth rates and internal banding key recommendations for future research. Front. Mar. Sci. 6:186. doi: 10.3389/
periodicity of maerl species (Corallinales, Rhodophyta) from northern Europe. fmars.2019.00186
Phycologia 42, 606–612. doi: 10.2216/i0031-8884-42-6-606.1 Dargie, T. (2001). Sand Dune Vegetation Survey of Scotland: East Coast.
British Geological Survey (2020). British Geological Survey (BGS) Offshore Marine Volume 1: Main report. Available online at: https://www.nature.scot/sites/
Products. Available online at: https://www.bgs.ac.uk/data/services/offprodwms. default/files/2017-07/Publication%202001%20-%20SNH%20Research%2C%
html (accessed January 23, 2020). 20Survey%20and%20Monitoring%20Report%20179%20-%20Sand%20dune%
Brodie, J., Williamson, C. J., Smale, D. A., Kamenos, N. A., Mieszkowska, N., 20vegetation%20survey%20of%20Scotland%20-%20East%20Coast%20-
Santos, R., et al. (2014). The future of the northeast Atlantic benthic flora in %20Volume%201%20-%20Main%20report.pdf (accessed August 9, 2019).
a high CO2 world. Ecol. Evol. 4, 2787–2798. doi: 10.1002/ece3.1105 Dutertre, M., Grall, J., Ehrhold, A., and Hamon, D. (2014). Environmental factors
Burdett, H., Perna, G., McKay, L., Broomhead, G., and Kamenos, N. (2018). affecting maerl bed structure in Brittany (France). Eur. J. Phycol 50, 371–383.
Community-level sensitivity of a calcifying ecosystem to acute in situ doi: 10.1080/09670262.2015.1063698
CO2 enrichment. Mar. Ecol. Prog. Ser. 587, 73–80. doi: 10.3354/meps Elith, J., Phillips, S. J., Hastie, T., Dudík, M., Chee, Y. E., and Yates, C. J. (2011).
12421 A statistical explanation of MaxEnt for ecologists: statistical explanation of
Burdett, H. L., Aloisio, E., Calosi, P., Findlay, H. S., Widdicombe, S., Hatton, A. D., MaxEnt. Divers. Distrib. 17, 43–57. doi: 10.1111/j.1472-4642.2010.00725.x
et al. (2012a). The effect of chronic and acute low pH on the intracellular DMSP Elliott, S., Turrell, W., Heath, M., and Bailey, D. (2017). Juvenile gadoid habitat and
production and epithelial cell morphology of red coralline algae. Mar. Biol. Res. ontogenetic shift observations using stereo-video baited cameras. Mar. Ecol.
8, 756–763. doi: 10.1080/17451000.2012.676189 Prog. Ser. 568, 123–135. doi: 10.3354/meps12068

Frontiers in Marine Science | www.frontiersin.org 11 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

European Commission (2016). The Habitats Directive - Environment JNCC (2016). UK BAP Priority Habitats. Available online at: http://jncc.defra.gov.
- European Commission. EU Nat. Law. Available online at: http: uk/page-5718 (accessed November 10, 2018).
//ec.europa.eu/environment/nature/legislation/habitatsdirective/index_en.htm Jorda, G., Marbà, N., Bennett, S., Santana-Garcon, J., Agusti, S., and Duarte, C. M.
(accessed November 10, 2018). (2020). Ocean warming compresses the three-dimensional habitat of marine life
European Commission (2018). Natura 2000. Available online at: http://ec.europa. | Nature Ecology & Evolution. Nat. Ecol. Evol. 4, 109–114. doi: 10.1038/s41559-
eu/environment/nature/natura2000/index_en.htm (accessed November 10, 019-1058-0
2018). Júnior, P. D. M., and Nóbrega, C. C. (2018). Evaluating collinearity effects on
Fabricius, K., and De’ath, G. (2001). Environmental factors associated with the species distribution models: An approach based on virtual species simulation.
spatial distribution of crustose coralline algae on the great barrier reef. Coral PLoS One 13:e0202403. doi: 10.1371/journal.pone.0202403
Reefs 19, 303–309. doi: 10.1007/s003380000120 Kamenos, N., and Law, A. (2010). Temperature Controls on Coralline Algal
Fithian, W., Elith, J., Hastie, T., and Keith, D. A. (2015). Bias correction in species Skeletal Growth. J. Phycol. 46, 331–335. doi: 10.1111/j.1529-8817.2009.00
distribution models: pooling survey and collection data for multiple species. 780.x
Methods Ecol. Evol. 6, 424–438. doi: 10.1111/2041-210X.12242 Kamenos, N., Moore, G., and Hall-Spencer, J. M. (2004). Small-scale
Foster, M. S. (2001). Rhodoliths: between rocks and soft places. J. Phycol. 37, distribution of juvenile gadoids in shallow inshore waters; what role does
659–667. doi: 10.1046/j.1529-8817.2001.00195.x maerl play? ICES J. Mar. Sci. 61, 422–429. doi: 10.1016/j.icesjms.2004.
Foster, M. S., Amado Filho, G. M., Kamenos, N. A., Riosmena-Rodríguez, 02.004
R., Steller, D. L., Lang, M. A., et al. (2013). “Rhodoliths and Rhodolith Kamenos, N. A., Burdett, H. L., Aloisio, E., Findlay, H. S., Martin, S., Longbone,
Beds,” in Research and Discoveries: The Revolution of Science through Scuba, C., et al. (2013). Coralline algal structure is more sensitive to rate, rather than
(Washington: Smithsonian Contributions to the Marine Sciences), 143–155. the magnitude, of ocean acidification. Glob. Change Biol. 19, 3621–3628. doi:
Fox, C. J., Valcic, L., and Veszelovsxki, A. (2015). Evidence Gathering in Support of 10.1111/gcb.12351
Sustainable Scottish Inshore Fisheries: Work Package 4 Final Report A Pilot Study Kamenos, N. A., Perna, G., Gambi, M. C., Micheli, F., and Kroeker, K. J. (2016).
to Define the Footprint and Activities of Scottish Inshore Fisheries by Identifying Coralline algae in a naturally acidified ecosystem persist by maintaining control
Target Fisheries, Habitats and Associated Fish Stocks. Scotland: Marine Alliance of skeletal mineralogy and size. Proc. R. Soc. B Biol. Sci. 283:20161159. doi:
for Science and Technology Scotland. 10.1098/rspb.2016.1159
Freeman, L., Kleypas, J. A., and Miller, A. J. (2013). PLoS ONE 8:e82404. doi: Kamenos, N. A., Strong, S., Shenoy, D., Wilson, S., Hatton, A., and
10.1371/journal.pone.0082404 Moore, P. (2008b). Red coralline algae as a source of marine biogenic
Friedlander, A. M., Caselle, J. E., Ballesteros, E., Brown, E. K., Turchik, A., and Sala, dimethylsulphoniopropionate. Mar. Ecol. Prog. Ser. 372, 61–66. doi: 10.3354/
E. (2014). The real bounty: marine biodiversity in the pitcairn Islands. PLoS One meps07687
9:e100142. doi: 10.1371/journal.pone.0100142 Kroeker, K. J., Micheli, F., and Gambi, M. C. (2012). Ocean acidification causes
Gallon, R. K., Robuchon, M., Leroy, B., Le Gall, L., Valero, M., and Feunteun, E. ecosystem shifts via altered competitive interactions. Nat. Clim. Change 3,
(2014). Twenty years of observed and predicted changes in subtidal red seaweed 156–159. doi: 10.1038/nclimate1680
assemblages along a biogeographical transition zone: inferring potential causes Legrand, E., Riera, P., Lutier, M., Coudret, J., Grall, J., and Martin, S. (2017).
from environmental data. J. Biogeogr. 41, 2293–2306. doi: 10.1111/jbi.12380 Species interactions can shift the response of a maerl bed community to ocean
Grall, J., and Hall-Spencer, J. M. (2003). Problems facing maerl conservation in acidification and warming. Biogeosciences 14, 5359–5376. doi: 10.5194/bg-14-
Brittany. Aquat. Conserv. Mar. Freshw. Ecosyst. 13, S55–S64. doi: 10.1002/aqc. 5359-2017
568 Lynham, J., Nikolaev, A., Raynor, J., Vilela, T., and Villaseñor-Derbez, J. C. (2020).
Gubbay, S., Sanders, N., Haynes, T., Janssen, J. A. M., Airoldi, L., Battelli, C., et al. Impact of two of the world’s largest protected areas on longline fishery catch
(2016). European Red list of habitats. Luxembourg: Publications Office of the rates. Nat. Commun 11, 1–9.
European Union. Macreadie, P. I., Anton, A., Raven, J. A., Beaumont, N., Connolly, R. M., Friess,
Hall-Spencer, J., White, N., Gillespie, E., Gillham, K., and Foggo, A. (2006). Impact D. A., et al. (2019). The future of blue carbon science. Nat. Commun. 10, 1–13.
of fish farms on maerl beds in strongly tidal areas. Mar. Ecol. Prog. Ser. 326, 1–9. doi: 10.1038/s41467-019-11693-w
doi: 10.3354/meps326001 Mao, J., Burdett, H. L., McGill, R. A. R., Newton, J., Gulliver, P., and Kamenos, N. A.
Hall-Spencer, J. M., Kelly, J., and Maggs, C. A. (2008). Assessment of Maerl Beds (2020). Carbon burial over the last four millennia is regulated by both climatic
in the OSPAR Area and the Development of A Monitoring Program. Ireland: and land use change. Glob. Change Biol. 26, 2496-2504. doi: 10.1111/gcb.
Department of the Environment HaLG, Ireland. 15021
Hall-Spencer, J. M., Kelly, J., and Maggs, C. A. (2010). OSPAR Commission Marine Scotland. (2018). Improving Protection Given to Priority Marine
Background Document for Maerl beds. London: OSPAR Commission. Features Outside the Marine Protected Area Network. Available online
Hastings, A., Byers, J. E., Crooks, J. A., Cuddington, K., Jones, C. G., Lambrinos, at: https://www.gov.scot/publications/improving-protection-given-pmfs-
J. G., et al. (2007). Ecosystem engineering in space and time. Ecol. Lett. 10, outside-mpa-network/pages/8/ (accessed January 30, 2020).
153–164. doi: 10.1111/j.1461-0248.2006.00997.x Martin, S., and Gattuso, J.-P. (2009). Response of Mediterranean coralline algae to
Hijmans, R. (2017). Package “Dismo.”. Available online at: https://www. ocean acidification and elevated temperature. Glob. Change Biol. 15, 2089–2100.
rdocumentation.org/packages/dismo/versions/1.1-4 (accessed August 28, doi: 10.1111/j.1365-2486.2009.01874.x
2019). Martin, S., and Hall-Spencer, J. M. (2017). “Effects of Ocean Warming and
Hoegh-Guldberg, O., Mumby, P. J., Hooten, A. J., Steneck, R. S., Greenfield, Acidification on Rhodolith/Maërl Beds,” in Rhodolith/Maërl Beds: A Global
P., Gomez, E., et al. (2007). Coral reefs under rapid climate change Perspective Coastal Research Library, eds R. Riosmena-Rodriguez, W. Nelson,
and ocean acidification. Science 318, 1737–1742. doi: 10.1126/science.115 and J. Aguirre (Berlin: Springer).
2509 Martínez, B., Radford, B., Thomsen, M. S., Connell, S. D., Carreño, F., Bradshaw,
Hughes, T. P., Rodrigues, M. J., Bellwood, D. R., Ceccarelli, D., Hoegh-Guldberg, C. J. A., et al. (2018). Distribution models predict large contractions of habitat-
O., McCook, L., et al. (2007). Phase shifts. herbivory, and the resilience of forming seaweeds in response to ocean warming. Divers. Distrib. 24, 1350–1366.
coral reefs to climate change. Curr. Biol. 17, 360–365. doi: 10.1016/j.cub.2006. doi: 10.1111/ddi.12767
12.049 McCoy, S. J., and Kamenos, N. A. (2015). Coralline algae (Rhodophyta) in
IPCC (2014). Climate Change 2014: Synthesis Report. Contribution of Working a changing world: integrating ecological, physiological, and geochemical
Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel responses to global change. J. Phycol. 51, 6–24. doi: 10.1111/jpy.
on Climate Change. Available online at: https://www.ipcc.ch/report/ar5/syr/ 12262
(accessed February 3, 2020). Merow, C., Smith, M. J., and Silander, J. A. (2013). A practical guide to
Jackson, C. M., Kamenos, N. A., Moore, P. G., and Young, M. (2004). Meiofaunal MaxEnt for modeling species’ distributions: what it does, and why inputs and
bivalves in maerl and other substrata; their diversity and community structure. settings matter. Ecography 36, 1058–1069. doi: 10.1111/j.1600-0587.2013.07
Ophelia 58, 49–60. 872.x

Frontiers in Marine Science | www.frontiersin.org 12 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

Muscarella, R., Galante, P. J., Soley-Guardia, M., Boria, R. A., Kass, J. M., Uriarte, Porter, J. S., Austin, W. E. N., Burrows, M. T., Davies, G., Kamenos, N., Riegel,
M., et al. (2018). ENMeval: Automated Runs and Evaluations of Ecological Niche S., et al. (2020). Blue Carbon Audit of Scottish Waters. Available online
Models. Available online at: https://cran.r-project.org/web/packages/ENMeval/ at: https://www.researchgate.net/publication/338585272_Blue_Carbon_Audit_
index.html (accessed November 22, 2018). of_Orkney_Waters (accessed March 12, 2020).
Muscarella, R., Galante, P. J., Soley-Guardia, M., Boria, R. A., Kass, J. M., Uriarte, Porzio, L., Buia, M. C., and Hall-Spencer, J. M. (2011). Effects of ocean acidification
M., et al. (2014). ENMeval: an R package for conducting spatially independent on macroalgal communities. J. Exp. Mar. Biol. Ecol. 400, 278–287. doi: 10.1016/
evaluations and estimating optimal model complexity for Maxent ecological j.jembe.2011.02.011
niche models. Methods Ecol. Evol. 5, 1198–1205. doi: 10.1111/2041-210X.12261 Qui-Minet, Z. N., Coudret, J., Davoult, D., Grall, J., Mendez-Sandin, M., Cariou,
National Species Reintroduction Forum (2014). The Scottish Code for Conservation T., et al. (2019). Combined effects of global climate change and nutrient
Translocations. Available online at: https://www.nature.scot/sites/default/ enrichment on the physiology of three temperate maerl species. Ecol. Evol. 9,
files/Publication%202014%20-%20The%20Scottish%20Code%20for% 13787–13807. doi: 10.1002/ece3.5802
20Conservation%20Translocations.pdf (accessed March 23, 2020). R Core Team (2019). R: A Language and Environment for Statistical Computing.
Nelson, W. A. (2009). Calcified macroalgae - critical to coastal ecosystems and Vienna: R Foundation for Statistical Computing.
vulnerable to change: a review. Mar. Freshw. Res. 60:787. doi: 10.1071/MF Riosmena-Rodriguez, R. (ed.) (2016). “The Role of Rhodolith/Maërl Beds in Modern
08335 Oceans,” in Rhodolith/maerl Beds: A Global Perspective. New York, NY: Springer,
Noisette, F., Duong, G., Six, C., Davoult, D., and Martin, S. (2013a). Effects of 3–26.
elevated pCO2 on the metabolism of a temperate rhodolith Lithothamnion Riosmena-Rodriguez, R., Nelson, W. A., and Aguirre, J. (eds) (2016).
corallioides grown under different temperatures. J. Phycol. 49, 746–757. doi: Rhodolith/maerl Beds: A Global Perspective. New York, NY: Springer.
10.1111/jpy.12085 Roberts, C. M., O’Leary, B. C., McCauley, D. J., Cury, P. M., Duarte, C. M.,
Noisette, F., Egilsdottir, H., Davoult, D., and Martin, S. (2013b). Physiological Lubchenco, J., et al. (2017). Marine reserves can mitigate and promote
responses of three temperate coralline algae from contrasting habitats to near- adaptation to climate change. Proc. Natl. Acad. Sci. U.S.A. 114, 6167–6175.
future ocean acidification. J. Exp. Mar. Biol. Ecol. 448, 179–187. doi: 10.1016/j. doi: 10.1073/pnas.1701262114
jembe.2013.07.006 Schoenrock, K. M., Bacquet, M., Pearce, D., Rea, B. R., Schofield, J. E., Lea, J., et al.
Pan, B. J., Vernet, M., Reynolds, R. A., and Mitchell, B. G. (2019). The optical (2018). Influences of salinity on the physiology and distribution of the Arctic
and biological properties of glacial meltwater in an Antarctic fjord. PLoS One coralline algae, Lithothamnion glaciale (Corallinales, Rhodophyta). J. Phycol.
14:e0211107. doi: 10.1371/journal.pone.0211107 54, 690–702. doi: 10.1111/jpy.12774
Pardo, C., Guillemin, M.-L., Peña, V., Bárbara, I., Valero, M., and Barreiro, Scottish Government (2018). Priority Marine Features. Available online at:
R. (2019). Local coastal configuration rather than latitudinal gradient shape http://www2.gov.scot/Topics/marine/marine-environment/mpanetwork/PMF
clonal diversity and genetic structure of phymatolithon calcareum maerl beds (accessed January 11, 2019).
in North European Atlantic. Front. Mar. Sci. 6:149. doi: 10.3389/fmars.2019. Scottish Government (2019). Improving protection given to Priority Marine
00149 Features Outside the Marine Protected Area Network - Scottish Government
Pardo, C., Lopez, L., Peña, V., Hernández-Kantún, J., Gall, L. L., Bárbara, I., - Citizen Space. Available online at: https://consult.gov.scot/marine-scotland/
et al. (2014a). A Multilocus species delimitation reveals a striking number of priority-marine-features/ (accessed January 30, 2020).
species of coralline algae forming maerl in the OSPAR maritime area. PLoS One Scottish Natural Heritage (2019a). GeMS - Scottish Priority Marine Features
9:e104073. doi: 10.1371/journal.pone.0104073 (PMF). Available online at: https://gateway.snh.gov.uk/natural-spaces/dataset.
Pardo, C., Peña, V., Bárbara, I., Valero, M., and Barreiro, R. (2014b). Development jsp?dsid=GEMS-PMF (accessed August 28, 2019).
and multiplexing of the first microsatellite markers in a coralline red alga Scottish Natural Heritage (2019b). Loch nam Madadh SAC. Available online at:
(Phymatolithon calcareum, Rhodophyta). Phycologia 53, 474–479. doi: 10.2216/ https://sitelink.nature.scot/site/8301 (accessed December 18, 2018).
14-031.1 Sheehan, E. V., Bridger, D., and Attrill, M. J. (2015). The ecosystem service value
Pavlova, A., Beheregaray, L. B., Coleman, R., Gilligan, D., Harrisson, K. A., Ingram, of living versus dead biogenic reef. Estuar. Coast. Shelf Sci. 154, 248–254. doi:
B. A., et al. (2017). Severe consequences of habitat fragmentation on genetic 10.1016/j.ecss.2014.12.042
diversity of an endangered Australian freshwater fish: a call for assisted gene Smale, D. A. (2020). Impacts of ocean warming on kelp forest ecosystems. New
flow. Evol. Appl. 10, 531–550. doi: 10.1111/eva.12484 Phytol. 225, 1447–1454. doi: 10.1111/nph.16107
Pearson, R. G. (2010). Species’ Distribution Modeling for Conservation Educators Steen, B. (2019). Modelling hot spot areas for the invasive alien plant Elodea
and Practitioners. Lessons Conserv. 3, 54–89. nuttallii in the EU. Manag. Biol. Invasions 10, 151–170. doi: 10.3391/mbi.2019.
Pearson, R. G., Raxworthy, C. J., Nakamura, M., and Peterson, A. T. (2007). 10.1.10
Predicting species distributions from small numbers of occurrence records: Steller, D. L., Riosmena-Rodríguez, R., Foster, M. S., and Roberts, C. A. (2003).
a test case using cryptic geckos in Madagascar. J. Biogeogr. 34, 102–117. doi: Rhodolith bed diversity in the Gulf of California: the importance of rhodolith
10.1111/j.1365-2699.2006.01594.x structure and consequences of disturbance. Aquat. Conserv. Mar. Freshw.
Peña, V., and Barbara, I. (2009). Distribution of the Galician maerl beds and Ecosyst. 13, S5–S20. doi: 10.1002/aqc.564
their shape classes (Atlantic Iberian Peninsula): proposal of areas in future Suggett, D. J., Camp, E. F., Edmondson, J., Boström-Einarsson, L., Ramler, V., Lohr,
conservation actions. Cah. Biol. Mar. 50, 335–368. K., et al. (2019). Optimizing return-on-effort for coral nursery and outplanting
Peña, V., Barbara Criado, I., Grall, J., Maggs, C. A., and Hall-Spencer, J. (2014). The practices to aid restoration of the great barrier reef. Restor. Ecol. 27, 683–693.
diversity of seaweeds on maerl in the NE Atlantic. Mar. Biodivers. 44, 533–551. doi: 10.1111/rec.12916
doi: 10.1007/s12526-014-0214-7 Sunday, J. M., Fabricius, K. E., Kroeker, K. J., Anderson, K. M., Brown, N. E.,
Perry, F., Jackson, A., and Garrard, S. L. (2017a). “Lithothamnion glaciale Maerl,” Barry, J. P., et al. (2017). Ocean acidification can mediate biodiversity shifts
in Marine Life Information Network: Biology and Sensitivity Key Information by changing biogenic habitat. Nat. Clim. Change 7, 81–85. doi: 10.1038/
Reviews, eds H. Tyler-Walters and K. Hiscock (Plymouth: Marine Biological nclimate3161
Association of the United Kingdom). Teagle, H., and Smale, D. A. (2018). Climate-driven substitution of habitat-forming
Perry, F., Jackson, A., Garrard, S. L., and Hiscock, K. (2017b). “Phymatolithon species leads to reduced biodiversity within a temperate marine community.
calcareum Maerl,” in Marine Life Information Network: Biology and Sensitivity Divers. Distrib. 24, 1367–1380. doi: 10.1111/ddi.12775
Key Information Reviews, ed. H. Tyler-Walters (Plymouth: Marine Biological The Council of the European Communities (1992). Council Directive 92/43/EEC of
Association of the United Kingdom). 21 May 1992 on the Conservation of Natural Habitats and of Wild Fauna and
Phillips, S. J., Dudík, M., and Schapire, R. E. (2006). Maxent Software Flora. Available online at: https://eur-lex.europa.eu/legal-content/EN/TXT/
for Modeling Species Niches and Distributions. Available online at: http:// ?uri=CELEX:01992L0043-20070101 (accessed January 10, 2019).
biodiversityinformatics.amnh.org/open_source/maxent (accessed January 20, Tyberghein, L., Verbruggen, H., Pauly, K., Troupin, C., Mineur, F., and Clerck,
2019). O. D. (2012). Bio-ORACLE: a global environmental dataset for marine species

Frontiers in Marine Science | www.frontiersin.org 13 September 2020 | Volume 7 | Article 575825

Simon-Nutbrown et al. Climate Change and Coralline Algae

distribution modelling. Glob. Ecol. Biogeogr. 21, 272–281. doi: 10.1111/j.1466- Wilson, S., Blake, C., Berges, J. A., and Maggs, C. A. (2004). Environmental
8238.2011.00656.x tolerances of free-living coralline algae (maerl): implications for European
van der Heijden, L. H., and Kamenos, N. A. (2015). Reviews and syntheses: marine conservation. Biol. Conserv. 120, 279–289. doi: 10.1016/j.biocon.2004.
calculating the global contribution of coralline algae to total carbon burial. 03.001
Biogeosciences 12, 6429–6441. doi: 10.5194/bg-12-6429-2015
Veloz, S. D. (2009). Spatially autocorrelated sampling falsely inflates measures Conflict of Interest: The authors declare that the research was conducted in the
of accuracy for presence-only niche models. J. Biogeogr. 36, 2290–2299. doi: absence of any commercial or financial relationships that could be construed as a
10.1111/j.1365-2699.2009.02174.x potential conflict of interest.
Verbruggen, H., Tyberghein, L., Belton, G. S., Mineur, F., Jueterbock, A., Hoarau,
G., et al. (2013). Improving transferability of introduced species’ distribution Copyright © 2020 Simon-Nutbrown, Hollingsworth, Fernandes, Kamphausen,
models: new tools to forecast the spread of a highly invasive seaweed. PLoS One Baxter and Burdett. This is an open-access article distributed under the terms
8:e68337. doi: 10.1371/journal.pone.0068337 of the Creative Commons Attribution License (CC BY). The use, distribution or
Wernberg, T., Smale, D. A., and Thomsen, M. S. (2012). A decade of reproduction in other forums is permitted, provided the original author(s) and the
climate change experiments on marine organisms: procedures, patterns and copyright owner(s) are credited and that the original publication in this journal
problems. Glob. Change Biol. 18, 1491–1498. doi: 10.1111/j.1365-2486.2012. is cited, in accordance with accepted academic practice. No use, distribution or
02656.x reproduction is permitted which does not comply with these terms.

Frontiers in Marine Science | www.frontiersin.org 14 September 2020 | Volume 7 | Article 575825