Bulletin 9: Applications

home

APPLICATIONS

Place Prioritization for Texas Using GAP Data: The Use of
Biodiversity and Environmental Surrogates Within Socioeconomic Constraints

SAHOTRA SARKAR¹, NICK C. PARKER², JUSTIN GARSON¹, ANSHU AGGARWAL¹, AND SHERI HASKELL²

1Biodiversity and Biocultural Conservation Laboratory, University of Texas, Austin

2U. S. Geological Survey-Biological Resources Division, Texas Cooperative Fish and Wildlife Research Unit, Texas
Tech University, Lubbock

Introduction

The Texas Gap Analysis Program (TX-GAP) (Gonzalez-Rebeles et al. 1997; Parker et al. 1998)
has developed a Geographical Information System (GIS) containing biogeographic and localized
socioeconomic data for the entire state of Texas. These data have been used to populate a set of
1,183 hexagons produced as part of the U.S. Environmental Protection Agency’s Monitoring and
Analysis Program (EMAP). Ignoring local topographic details, these hexagons each have an
area of 649 sq. km. The attributes recorded for each cell include:

environmental attributes such as area, average temperature, average maximum and minimum
temperatures, average rainfall, average snowfall, elevation, aspect, dominant soil type, and
acreage of each of 46 soil classes;
biodiversity attributes such as dominant vegetation type, acreage of each of 38 vegetation
types, and presence/absence of 655 animal species; and
socioeconomic attributes including human population density.

This data set was used to perform three tasks that are central to conservation planning (Margules
and Pressey 2000):

Subsets of the environmental and biodiversity attributes were used to prioritize the hexagons
and establish potentially reserved sets of hexagons so as to achieve a required level of
representation of the attributes in a reserve network. This is called the “place prioritization
problem,” and the goal is to achieve the targeted representation as efficiently as possible (that is, with the least required number of hexagons).
Attempts were made to use small subsets of attributes, called “estimator-surrogates,” to
obtain the same set of prioritized places as would be obtained if larger sets (the so-called
“true surrogates” for general biodiversity) were used. This is called the “surrogacy
problem,” and the goal is to obtain a small set of readily available estimator-surrogates (for
instance, through remote sensing) for conservation planning.
Prioritization results were overlaid on socioeconomic data to explore the potential for
conflicts between land conversion for human use and biodiversity. This is the beginning of more general attempts to optimize the place prioritization process in the presence of multiple objectives such as biodiversity conservation and economic development (Faith 1995).

All place prioritization algorithms used here are based primarily on rarity and complementarity
(Margules et al. 1988); some also incorporate adjacency requirements (where adjacent cells are
preferred over nonadjacent ones). These have all been implemented in the ResNet software
package developed at the University of Texas at Austin in collaboration with the Commonwealth
Scientific and Industrial Research Organization, Australia (Aggarwal et al. 2000). The
algorithms are initialized using: a) a set of pre-existing reserved cells; b) the cell with the rarest
attributes; or c) the cell richest in attributes. The cells correspond to the hexagons of the GAP
data set. The iterative procedure uses rarity, followed by complementarity (choosing the cell
with the most representatives that are not yet adequately represented in the reserved set of cells),
and then adjacency (if imposed). Further ties are broken randomly (by lexical order). The
algorithms terminate if a predetermined level of representation of attributes in the selected cells
is satisfied or cost and area constraints are exceeded. All these algorithms were run on the Texas
data.

The surrogacy problem was investigated in the following way: A place prioritization was carried
out, using a set of putative estimator-surrogates, and then compared to one carried out using the
targeted “true” surrogates. The estimator-surrogate set used consisted of soil types, vegetation
classes, average high and low temperature and average precipitation; the “true” surrogates were
the vertebrate species for which there were distributional data. For a preliminary socioeconomic
analysis, the most densely populated hexagons of Texas, as projected for the future, were
overlaid on the hexagons that emerged with the highest priority from the ResNet runs using
terrestrial vertebrate data. Human population density gives an estimate of risk for vertebrate
populations; this has obvious implications in conservation planning.

Results

A total of 121 hexagons were selected when vertebrate data were used, and the target of
representation was set at 10% of all presence records (Figure 1). The selection procedure was
initialized by rarity.

When an environmental parameter set of vegetation classes, soil types, average high and low
temperatures, and average precipitation was used for the estimator-surrogates, and terrestrial
vertebrate species were targeted as “true” surrogates, a surrogacy graph was produced (Figure 2).
By the time each estimator-surrogate has been represented at least once, 89.5% of the vertebrate
species have been represented at least once.

An overlay of the 100 most populated hexagons, as projected for 2020, on the 100 hexagons that
receive top priority using the vertebrate data was developed (Figure 3). There was an overlap for
26% of the hexagons.

Discussion

The results presented here should be regarded as preliminary and tentative and are indicative of
the way in which conservation planning should proceed rather than an actual guideline for policy
decisions. The most serious deficiency, which affects all the results, is that the hexagons are too
large, and each hexagon is ecologically too heterogeneous, to serve as a unit for site-specific
conservation decisions. A much smaller scale must be used, usually less than a tenth of the
hexagon area, when land acquisition and use decisions are made by conservation agencies
(Sarakinos et al. 2001). Ideally, this analysis should be repeated at those scales prior to making
an acquisition. However, the analysis conducted at the EMAP scale has immediate value. The
size of the hexagons (649 km 2 ) is about 1/3 of the size of a county in Texas. In a state where 96-
97% of the land is privately owned, a tool that identifies areas 1/3 the size of a county is a
valuable aid in landscape conservation. Fine detail would be needed to support acquisition or
other site-specific conservation measures; however, this broader-scale analysis does not target
individual landowners and thereby may minimize premature landowner opposition or speculative
purchases.

The use of environmental estimator-surrogates is fraught with danger since full representation of
these surrogates does not guarantee a full representation of the targeted attributes (Figure 2).
Much more work needs to be done to see if there even exists a small set of readily available
estimator-surrogates for this purpose. If not, there is no way to avoid the conclusion that
systematic conservation planning will require detailed survey data that are hard to obtain.

Using vertebrate species as true surrogates for biodiversity, 26% of the hexagons with the
highest conservation priority are at high risk because of the impact of the projected human
population increase (Figure 3). This is one conclusion that is not affected by the size of the
hexagons since this increased human population density will almost certainly have negative
impacts on biodiversity, particularly terrestrial vertebrates, at scales smaller than , and up to , the
size of the hexagons.

Product Availability

The ResNet software package can be downloaded free of charge along with a manual from the
Web site of the Biodiversity and Biocultural Conservation Laboratory at the University of Texas-
Austin (http://uts.cc.utexas.edu/~consbio/Cons/Labframeset.html).

Acknowledgements

We thank Clint Boal, U.S. Geological Survey-Biological Resources Division, and Robert
Bradley, Department of Biological Sciences, Texas Tech University, for their critical review of
this manuscript. The Texas Cooperative Fish and Wildlife Research Unit is jointly sponsored by
Texas Parks and Wildlife, Texas Tech University, the U.S. Geological Survey, and the Wildlife
Management Institute.

Literature Cited

Aggarwal, A., J. Garson, C.R. Margules, A.O. Nicholls, and S. Sarkar. 2000. The ResNet Manual, Version 1.1. Biodiversity and Biocultural Conservation Laboratory, University of
Texas, Austin.

Faith, D.P. 1995. Biodiversity and regional sustainability analysis. CSIRO Division of Wildlife
and Ecology, Lyneham.

Gonzales-Rebeles, C., N.C. Parker, R.W. Sims, Y. Lan, and M. Cano. 1997. Spectrum software
for Gap Analysis projects in Texas and Mexico. Gap Analysis Bulletin 6:24-27. USGS/BRD Gap Analysis Program, Moscow, Idaho.

Margules, C.R., A.O. Nicholls, and R.L. Pressey. 1988. Selecting networks of reserves to
maximize biological diversity. Biological Conservation 43:63 -76.

Margules, C.R., and R.L. Pressey. 2000. Systematic conservation planning. Nature 405: 243 - 253.

Parker, N.C., R.J. Baker, R.D. Bradley, C. Jones, R.R. Monk, D.J. Schmidly, R.W. Sims, and F.D. Yancey, II. 1998. A partnership providing field data for Gap Analysis: Texas Tech
Museum and Texas GAP. Pages 37-39 in E.S. Brackney and M.D. Jennings, editors. Gap
Analysis Bulletin No. 7. USGS/BRD/Gap Analysis Program, Moscow, Idaho.

Sarakinos, H., A.O. Nicholls, A. Tubert, A. Aggarwal, C.R. Margules, and S. Sarkar. 2001. Area prioritization for biodiversity conservation in Québec on the basis of species
distributions: a preliminary analysis. Biodiversity and Conservation , in press.

Figure 1. These are the hexagons selected if the target for each vertebrate species is set at 10%
of the records. Therefore, preservation of these sites would preserve 10% of the sites containing
the highest diversity of terrestrial vertebrate species. Collectively, these sites contain
representatives of all terrestrial vertebrates in Texas.

Figure 2. The percentage of terrestrial vertebrate species in Texas protected when places are
selected using estimator-surrogate set consisting of soil types, vegetation classes, average high
and low temperature and average precipitation. The x-axis shows the percentage of the
estimator-surrogates protected at each stage.

Figure 3. Identification of the top 100 of the hexagons (gray) containing the highest diversity of
terrestrial vertebrate species in Texas and 100 hexagons (black dot) with the human population at
the greatest density. The overlap is 26%.