University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Managing Vertebrate Invasive Species USDA National Wildlife Research Center Symposia

August 2007

INVASIVE SPECIES MANAGEMENT AND

RESEARCH USING GIS
Tracy Holcombe
Fort Collins Science Center, U.S. Geological Survey, Fort Collins, Colorado, USA

Thomas J. Stohlgren
Fort Collins Science Center, U.S. Geological Survey, Fort Collins, Colorado, USA

Catherine Jarnevich
Fort Collins Science Center, U.S. Geological Survey, Fort Collins, Colorado, USA

Follow this and additional works at: http://digitalcommons.unl.edu/nwrcinvasive

Part of the Environmental Indicators and Impact Assessment Commons

Holcombe, Tracy; Stohlgren, Thomas J.; and Jarnevich, Catherine, "INVASIVE SPECIES MANAGEMENT AND RESEARCH
USING GIS" (2007). Managing Vertebrate Invasive Species. 18.
http://digitalcommons.unl.edu/nwrcinvasive/18

This Article is brought to you for free and open access by the USDA National Wildlife Research Center Symposia at DigitalCommons@University of
Nebraska - Lincoln. It has been accepted for inclusion in Managing Vertebrate Invasive Species by an authorized administrator of
DigitalCommons@University of Nebraska - Lincoln.
INVASIVE SPECIES MANAGEMENT AND RESEARCH USING GIS

TRACY HOLCOMBE, THOMAS J. STOHLGREN, AND CATHERINE JARNEVICH, Fort Collins Science Center,
U.S. Geological Survey, Fort Collins, Colorado, USA

Abstract: Geographical Information Systems (GIS) are powerful tools in the field of invasive species
management. GIS can be used to create potential distribution maps for all manner of taxa, including plants,
animals, and diseases. GIS also performs well in the early detection and rapid assessment of invasive species.
Here, we used GIS applications to investigate species richness and invasion patterns in fish in the United
States (US) at the 6-digit Hydrologic Unit Code (HUC) level. We also created maps of potential spread of the
cane toad (Bufo marinus) in the southeastern US at the 8-digit HUC level using regression and environmental
envelope techniques. Equipped with this potential map, resource managers can target their field surveys to
areas most vulnerable to invasion. Advances in GIS technology, maps, data, and many of these techniques
can be found on websites such as the National Institute of Invasive Species Science (www.NIISS.org). Such
websites provide a forum for data sharing and analysis that is an invaluable service to the invasive species
community.

Key Words: buffer, early detection, environmental envelope, geographic information systems, GIS, invasive
species, regression, thiessen polygons.

Managing Vertebrate Invasive Species: Proceedings of

an International Symposium (G. W. Witmer, W.C. Pitt,
K.A. Fagerstone, Eds). USDA/APHIS/WS, National
Wildlife Research Center, Fort Collins, CO. 2007.

INTRODUCTION GIS can be a useful tool for monitoring invasive

Geographic Information Systems (GIS) and vertebrates, especially for early detection and rapid
Global Positioning Systems (GPS) provide a assessment. Species distributions are largely
mechanism to digitally pinpoint a location on earth, determined by the environment. A growing number
view the location on a map, and use the location of statistical models, called Species Environmental
and ancillary data in spatial analyses. Individuals Matching (SEM) models are being used to
are able to quickly and easily produce maps and determine current and potential distributions and
conduct spatial analyses that would otherwise be abundances of harmful invasive species (Stohlgren
difficult or possible to produce using a network of and Schnase 2006). SEM models relate observed
satellites, satellite receivers, and mapping software. species distributions to environmental (climatic,
GIS serves as a data storage and analysis device for topographic, edaphic) envelopes. Then, assuming
spatial data, making data easy to view and the same stable relationships, they project species
manipulate. spatial shifts (local, enrichment, or extinction) in
Health care, agriculture, and environmental response to envelope changes under current
industries are a few of many entities that have been conditions. These environmental envelopes,
positively influenced with the advent of GIS arranged along a gradient from proximal to distal
technologies. Large spatial databases can help predictors, may have direct or indirect effects on
companies track their hard goods and allow farmers species’ establishment and survival (Austin 2002).
to determine which areas of their fields need more SEM models are either created in a GIS or can be
fertilizer, eliminating the need to add fertilizer to displayed in GIS to give a visual representation of
the entire field. Ecological data often contains the environmental envelope and potential habitat or
spatial component. Where an animal spends its abundance.
time and the patterns of its movements can be An important consideration for invasive species
important clues to its biology. Biological management is that recent invaders may not have
information of non-native species provides insight filled all suitable habitats, while species naturalized
to explaining expanding distributions and provides long ago may have filled a larger proportion of
a watch list of spreading invasive species to suitable habitat. Defining where a species may
managers for early detection and rapid response. survive depends heavily on being able to determine
108
its existing or potential habitat. Technological While these resolutions provide a lot of information
advances in GIS software make these types of at a fine scale, this scale is not always necessary or
analyses more readily accessible to the general desired. In certain situations, GIS can be used to
public. Data such as elevation, vegetation type, and summarize data, often simplifying the data into the
climate information are now available for free on resolution of interest. The Spatial Analyst module
the internet, often paired with websites that allow of ArcGIS 9 (ESRI 2004) has a zonal statistics
persons to view and utilize the information. function that calculates raster layer summary
statistics for a large polygonal area like a county.
HOW GIS CAN BE USED This module will extract the average, minimum or
View Data maximum, and range for each polygon.
A very basic and effective way to use GIS is to Additionally, GIS can be used to extract the value
view data. Many datasets are very large and for a specific point in a DEM so that the entire
difficult to visualize as a table of numbers. When surface does not have to be stored. GIS functions
viewed spatially, these data often make more sense. like these make the retrieval of dependant data for
Stohlgren et al. (2006) combined native and non- models readily accessible.
native fish datasets from NatureServe and the
United States Geological Survey (USGS) Florida Field Data - Points, Lines, and Polygons
Integrated Science Center’s Non-Indigenous Spatial field data can be displayed and managed
Aquatic Species database to examine numbers of in a GIS. The data are stored in one of three
native and non-native species in each 6-digit HUC formats: points, lines, or polygons. Locations of
area (Figure 1). Without performing any statistical individual organisms are examples of points. These
analyses, they found a large majority of native fish are discrete one-dimensional places in space. A
in the US are centered throughout the mid-west and linear representation of interest is a line. Lines
south-central US. The non-native fish are found include rivers, transects, or roads. Polygons
primarily in the western and eastern US. These represent an area of interest, like a stand of trees or
patterns were ascertained without conducting any a lake.
statistical analysis, proving that displaying the data Point data types are an excellent medium for
in a spatial format can be a useful endeavor even recording presence or absence of a species because
without detailed analyses. they are discreet. Lines give similar amounts of
information, again lending themselves well to
Data Summary presence, absence, and additional attributes.
GIS can be helpful for summarizing large Polygons are unique because they cover an area
datasets for modeling habitat quality and which can contain additional information such as
distribution. Data layers, such as Digital Elevation abundance or percent cover. All these data types
Models (DEM), are often used in modeling because can be collected in the field using either paper maps
they provide a large amount of environmental or GPS devices to collect the data before
information. DEMs are available free via the downloading it to a computer.
internet, often at either 10- or 30-meter resolutions.

a. b.

Figure 1. Pattterns of (a.) native and (b.) non-native fish by 6-digit HUC drainage (Stohlgren et al. 2006)

109
Simple GIS models harmful invasive species. In short, we will be able
GIS can be used to create simple analyses such to better manage and assess risks associated with
as buffers and thiessen polygons. A buffer can be harmful invasive species because risk assessments
created around points, lines, or polygons. It is a require accurate modeling of current and potential
new polygon of specified distance from the original species distributions (Stohlgren and Schnase 2006).
feature. Any GIS program can create a buffer Numerous challenges exist in traditional SEM
around points, lines, or polygons that can be used or niche-based modeling for current and future
for various reasons, such as surrogates for habitat species distributions (see reviews by Pearson and
for poorly studied species. Buffers can also define Dawson 2003, Soberon and Peterson 2005, Elith et
potential habitat for species that have a very al. 2006, Guisan et al. 2006, Heikkinen et al. 2006,
specific distance they can be from a given feature, Hijmans and Graham 2006, Peterson 2006,
such as water. Buffers are a commonly used Beaumont et al. 2007). These challenges have not
transformation of spatial data. prevented scientists and resource managers from
Another analysis performed by GIS is the refining, testing, and using SEM models in their
creation of thiessen polygons, sometimes known as work. No two SEM models are identical, and each
voronoi polygons. Theissen polygons are created has advantages and disadvantages (Table 1).
around a group of points, one polygon for each
point. The polygons are created around each point Regression Models
in such a way that every location lies within the Logistic regression is a type of Generalized
polygon of the point to which it is nearest. The Linear Model (GLM) appropriate for data with a
easiest way to think about thiessen polygons is with binary distribution such as species presence or
fast food delivery areas. A fast food pizza chain absence (McCullagh and Nelder 1989). The output
would divide a city into thiessen polygons, only from logistic regression models can be taken from
delivering to customers that were closer to them statistical software and used in GIS to create a
than they were to the next restaurant. A wildlife visual representation of the model created. We
example would be, if there are twelve nests in an have done this with data obtained from the USGS
area, polygons are formed around those twelve Florida Integrated Science Center’s Non-
nests so that every place on the landscape falls into Indigenous Aquatic Species database on the
the polygon associated with the closest nest. This invasive cane toad (Bufo marinus). The cane toad
tool has many applications for studying territorial has become established in the US and invaded
animals. Nest location data could be used to several watersheds in Florida. We employed
generate thiessen polygons surrounding each nest. logistic regression with Systat 11.0 (SSI 2004)
The area surrounding each nest could be an using minimum temperature, minimum radiation,
estimate of territory range. Buffers and thiessen mean temperature, maximum temperature,
polygons are two of the many possible examples of maximum humidity, and maximum growing degree
simple operations that can be done using a GIS. days as predictor variables to determine how much
potential habitat exists for the cane toad in the
Statistical models south-eastern US. We constructed a step-wise
Statistical models use current species GLM, and only minimum temperature was selected
distribution data to try and predict potential habitat. as a significant variable. Results of the regression
Conceptually, the SEM models assume the fitted analysis had a high predictive power (McFadden’s
observational relationships to be an adequate Rho Squared = 0.92). When the results were
representation of the realized niche of a species imputed in a GIS, the map showed that the cane
under a stable equilibrium or quasi-equilibrium toad had invaded most of its suitable habitat in the
constraint. As such, the SEM model result is only a Florida area, with only a few un-invaded areas left
first approximation of future distributions of in high and medium habitat suitability areas (Figure
individual species (Pearson and Dawson 2003). 2). This result was a first approximation model.
SEM model results are also determined by other More data and ecological information on the cane
processes such as dispersal, adaptation, toad could produce better results in the future.
competition, succession, fire and grazing pressure
(Austin 2002). Still, an integrated model may The Environmental Envelope Model
contribute considerably to a robust early warning The Environmental Envelope Model (EEM,
system for decision makers to design more Jarnevich et al. 2007) was developed as a rapid
effective management and control strategies for assessment technique to estimate the potential
110
 Table 1. Commonly used species environmental matching models for predicting species distributions.
Model Citation Advantages Disadvantages
Maxent (Phillips et al. Presence only, nonlinear, Presence only (no
2006) nonparametric, not sensitive to consideration of
multicollinearity, provides absence data)
variables’ relative importance
(jackknifing), easy to run and
takes less time, becoming popular

Classification (Breiman et al. Non-parametric, Absence data needed

and Regression 1984) Presence/absence, easy to run and
Tree (CART) interpret

Boosted (Friedman 2001, Non-parametric, Absence data needed,

Regression Tree De'ath 2007) Presence/absence, limitations more statistical details
with spatial data

Logistic (McCullagh and Widely used, presence/absence Absence data needed,

Regression Nelder 1989) sensitive to
multicollinearity

Least square Most statistics Widely used, continuous response Needs continuous
regression software variable (e.g., species richness) response variable,
sensitive to
multicollinearity,
decision about
significance level (P
value?)

BIOCLIM (Busby 1991) Presence only, simple Presence only, does not
use absences, less
accurate than other
niche models

DOMAIN (Carpenter et al. Presence only, simple Presence only, does not
1993) use absences, less
accurate than other
niche models

ENFA (Env. (Hirzel et al. 2002) Presence only Presence only, does not
Niche Factor use absences
Analysis)
Envelope (Jarnevich et al. Presence only or absence only All environmental
2007) models can be run. factors are given equal
weighting.

distribution of a species given its present location chosen by the modeler that are relevant to the
and associated environmental attributes. It is species of interest or species growth in general, to
supported by ArcGIS 9x (ESRI 2004) and will be determine locations within the environmental
available on the National Institute of Invasive envelope where the species of interest may be able
Species Science (NIISS) website (www.NIISS.org). to become established. The minimum and
Envelope models use environmental variables, maximum of each independent variable are noted

111
Figure 2. Regression model of the cane toad showing low, medium, and high likelihood of suitable habitat in each 6-
digit HUC.

Figure 3. Envelope model of the cane toad showing the number of parameters in each 6-digit HUC that could
contain the species.

112
by the ArcGIS program for all of the locations that sophisticated and is open to the general public.
the species is present. These minimum and This may be the direction GIS software is heading
maximum values together become the "envelope" toward, reducing dependence on desktop GIS
in which the species can survive. For instance, if a software in the future.
species exists in only three counties and the
temperature in county A is 45° F, county B is 40° CONCLUSION
F, and county C is 43° F, then the temperature Advances in GIS technology have made it
envelope is 40 to 45° F. We would then compare become a useful tool for land managers and
the temperature for other counties to see if they fell academics alike. It is widely used as a tool to
within the range of potential habitats. The model perform basic functions such as displaying data and
can include several different environmental layers more complex functions like creating and
to determine suitable habitats. The output of the displaying SEM models. As we look to our
model informs how many of the input variables lie computers today and continue to look to the future
within the environmental envelope of the species. of GIS technologies, GIS is a tool that should, and
We conducted an EEM analysis on the cane could be used by many scientists and resource
toad to compare results generated from the managers alike.
regression model (Figure 3). We used
environmental data retrieved from the Daymet LITERATURE CITED
website (www.daymet.org) that was originally at 1 AUSTIN, M. P. 2002. Spatial prediction of species
km2 resolution for the dependant variables. We distribution: an interface between ecological theory
used zonal statistics in ArcGIS’s Spatial Analyst and statistical modelling. Ecological Modelling
(ESRI 2004) to summarize the data for the 6-digit 157:101-118.
HUCs. Variables included minimum radiation, BEAUMONT, L. J., A. J. PITMAN, M. PAULSEN, AND L.
minimum temperature, mean temperature, HUGHES. 2007. Where will species go?
maximum temperature, maximum humidity, and Incorporating new advances in climate modeling
into projections of species distributions. In G. C.
growing degree days. We used the same cane toad
Biology, editor. doi.10.1111/j.1365-
data from the regression model. The resulting map 2486.2007.01357.x.
showed that as distance increases from the BREIMAN, L., J. H. FRIEDMAN, R. A. OLSHEN, AND C. G.
peninsula of Florida, there are less environmental STONE. 1984. Classification and regression trees.
variables that fall within the cane toad’s Wadsworth International Group, Belmont,
environmental envelope. This trend supports the California, USA.
regression model that showed the cane toad did not BUSBY, J. R. 1991. BIOCLIM - A bioclimate analysis
have much more suitable habitat than what is and prediction system. Pages 64-68 in C. R. M. a.
already occupied. M. P. Austin, editor. Nature conservation: cost
effective biological surveys and data analysis.
CSIRO, Melbourne.
GIS on the Web
CARPENTER, G., A. N. GILLISON, AND J. WINTER. 1993.
Common issues confronting GIS users today Domain - a flexible modeling procedure for
include software and data availability and user mapping potential distributions of plants and
friendliness. GIS software is often expensive, animals. Biodiversity and Conservation 2:667-680.
making it difficult for many people to obtain. DE'ATH, G. 2007. Boosted trees for ecological modeling
Another subset of would-be GIS users have access and prediction. Ecology 88:243-251.
to software, but do not have the time required to ELITH, J., C. H. GRAHAM, R. P. ANDERSON, M. DUDIK, S.
learn to effectively and efficiently use the software. FERRIER, A. GUISAN, R. J. HIJMANS, F. HUETTMANN,
These issues are changing with the advances in GIS J. R. LEATHWICK, A. LEHMANN, J. LI, L. G.
technology. Many of the functions that are found LOHMANN, B. A. LOISELLE, G. MANION, C. MORITZ,
M. NAKAMURA, Y. NAKAZAWA, J. M. OVERTON, A.
in proprietary software can also be found on the
T. PETERSON, S. J. PHILLIPS, K. RICHARDSON, R.
internet. Much of the species distribution data used SCACHETTI-PEREIRA, R. E. SCHAPIRE, J. SOBERON,
in the examples in this paper were found and S. WILLIAMS, M. S. WISZ, AND N. E. ZIMMERMANN.
downloaded from the internet. Many websites, 2006. Novel methods improve prediction of species'
such as NIISS are encouraging an environment of distributions from occurrence data. Ecography
data sharing. The NIISS website includes an 29:129-151.
interface to upload data and a GIS interface to view
data graphically, create models, and print and save
final map products. The technology is very
113
ESRI. 2004. ArcGIS 9.1. in ESRI, Redlands, CA. MCCULLAGH, P., AND J. A. NELDER. 1989. Generalized
FRIEDMAN, J. H. 2001. Greedy function approximation: a linear models, 2nd edition. Chapman and Hall,
gradient boosting machine. Annals of Statistics London; New York.
29:1189-1232. PEARSON, R. G., AND T. P. DAWSON. 2003. Predicting
GUISAN, A., A. LEHMANN, S. FERRIER, M. AUSTIN, J. M. the impacts of climate change on the distribution of
C. OVERTON, R. ASPINALL, AND T. HASTIE. 2006. species: are bioclimate envelope models useful?
Making better biogeographical predictions of Global Ecology and Biogeography 12:361-371.
species' distributions. Journal of Applied Ecology PETERSON, A. T. 2006. Uses and requirements of
43:386-392. ecological niche models and related distributional
HEIKKINEN, R. K., M. LUOTO, M. B. ARAUJO, R. models. Biodiversity Informatics 3:59-72.
VIRKKALA, W. THUILLER, AND M. T. SYKES. 2006. PHILLIPS, S. J., R. P. ANDERSON, AND R. E. SCHAPIRE.
Methods and uncertainties in bioclimatic envelope 2006. Maximum entropy modeling of species
modelling under climate change. Progress in geographic distributions. Ecological Modelling
Physical Geography 30:751-777. 190:231-259.
HIJMANS, R. J., AND C. H. GRAHAM. 2006. The ability of SOBERON, J., AND A. T. PETERSON. 2005. Interpretation
climate envelope models to predict the effect of of models of fundamental ecological niche and
climate change on species distributions. Global species' distributional areas. Biodiversity
Change Biology 12:2272-2281. Informatics 2:1-10.
HIRZEL, A. H., J. HAUSSER, D. CHESSEL, AND N. PERRIN. SSI. 2004. SYSTAT 11.0. San Jose, California, USA.
2002. Ecological-niche factor analysis: How to STOHLGREN, T. J., D. BARNETT, C. FLATHER, P. FULLER,
compute habitat-suitability maps without absence B. PETERJOHN, J. KARTESZ, AND L. L. MASTER.
data? Ecology 83:2027-2036. 2006. Species richness and patterns of invasion in
JARNEVICH, C. S., D. T. BARNETT, T. J. STOHLGREN, AND plants, birds, and fishes in the United States.
J. KARTESZ. 2007. A simple framework for an Biological Invasions 8:427-447.
invasive species early warning system for counties. STOHLGREN, T. J., AND J. L. SCHNASE. 2006. Risk
Frontiers in Ecology and the Environment in review. analysis for biological hazards: What we need to
know about invasive species. Risk Analysis
26:163-173.

114