Where are Endangered Species Found in the United States?

By: Jon Paul Rodriguez, W. Mark Roberts and Andy Dobson

Public and scientific opinion is divided over the current Endangered Species Act (ESA). Many private landowners and real-estate developers see it as a major impediment to economic progress; conversely, many conservation biologists and natural resource managers see it as an only partially effective legal mechanism for protecting biodiversity. Both camps would like to see the ESA modified before its pending reauthorization by the current Congress, and both present arguments for strengthening or weakening the "civil rights" of the estimated one million resident species of the United States.

In an ongoing collaboration between Princeton University and the Environmental Defense Fund, we have been examining the geographical distribution of endangered species in the United States. Our ultimate goal is to provide a systematic analysis of the available information on endangered species distributions. We hope that this may reduce some of the antagonism surrounding the reauthorization of the ESA; at present, subjective and contentious opinions reign in the face of inadequate scientific understanding of threats to biodiversity. Furthermore, resources available for conservation are limited, and must be allocated where they will be most effective. We believe that maximizing the number of endangered species represented in a comprehensive network of protected areas in the U.S. is an effective way to maximize efficiency of conservation efforts. This principle underlies the analyses we describe below.

In the present paper, we draw extensively on our recently published results (Dobson et al. 1997) to answer three basic questions.

(1) Are endangered species distributed randomly across the United States, or are there "hot spots" with particularly high numbers of endangered species?

(2) If hot spots do exist, are they the same for different taxonomic groups?

(3) Do certain taxonomic groups of endangered organisms indicate the presence of endangered species from other groups?

We used data provided in the Endangered Species by County Database of the Office of Pesticide Programs, United States Environmental Protection Agency (Washington, D.C.). This database lists the counties of occurrence of all plants and animals protected under the ESA in the 50 states, as well as all species, subspecies, and populations proposed for protection by this statute as of August 1995. The database includes 924 species distributed in 2,858 counties throughout the U.S. For comparative analyses, we divided the data into seven taxonomic groupings: plants, mollusks (which combines clams and snails), arthropods (arachnids, crustaceans and insects), fish, herptiles (amphibians and reptiles), birds, and mammals (Table 1).

The patterns in the distribution of endangered animals and plants in the U.S. are clearly non-random (Figures 1 and 2). Hawaii, southern California, the southeastern coastal counties (mainly in Florida and Georgia), and southwestern Appalachia emerge as hot spots, with large numbers of endangered species (see also Flather et al. 1994). Upon examining the association between the density of endangered species in each state with the intensity of human activities, climate, topology, and vegetative cover of each state, we found that the overall density of endangered species is strongly correlated with one anthropogenic and one climatic variable: the value of agricultural output and either average temperature or rainfall (stepwise multiple linear regression, r²=0.80, p<0.01). Agricultural activities best predict densities of endangered plants (r²=0.61, p<0.01), mammals (r²=0.68, p<0.01), birds (r²=0.64, p<0.01), and reptiles (r²=0.46, p<0.05), groups which together make up close to three-quarters of the entire data set (Dobson et al. 1997).

The selection of representative sites for the creation of a comprehensive network of protected areas for biodiversity conservation is currently an area of active research (e.g., Csuti et al. 1997; Pressey et al. 1993; 1996). One current approach bases its site-selection criteria on the principle of complementarity (Vane-Wright et al. 1991), which selects areas as follows: from a list of sites, we select the site with the largest number of species. Then, we remove all the species found in that site from consideration in the remaining sites, and re-rank the sites according to the number of species not already sampled. Next, we select the site with the largest number of remaining species, and repeat this process until all species have been accounted for. The list of sites obtained in this fashion is the complementary subset, and minimizes the total area needed to sample all species at least once.

We applied this procedure to determine the complementary county subsets (CCS) for the endangered species in the U.S., where the number of species in each county (our approximation of a "site") was the basis for the analysis. The results of our complementarity analysis can be found in Table 1. Plants have the largest CCS (136 counties, 9.61 percent of the U.S. landmass), while herptiles have the smallest (28 counties, 0.97 percent of the U.S. landmass).

There are at least two reasons to be cautious about this procedure. First, the fact that a species has been recorded in a county does not mean that it occupies the county's entire area. Second, the presence of a species in a county does not mean that the county contains an entire viable population of that species. Our approach is biased in favor of range-restricted endemics, such as many of the plants in the database, and against widespread species, like the grizzly bear (Ursus arctos horribilis) and bald eagle (Haliaeetus leucocephalus). In the context of our analysis, the selection of a county implies that it possesses great value to the conservation of endangered species at the level of the whole country. This does not mean, however, that by creating one protected area within the boundaries of the chosen counties the conservation problem is solved. In the case of a range-restricted species, the protected area would have to be in the right place within the county (i.e., the range of the threatened species); in the case of a widespread species, the protected area would likely have to cross the boundaries of single countiesÑand perhaps state boundaries as well. We must add, however, that in the case of a widespread species, as the number of counties in a CCS increases, so will the likelihood that more than one population of the species will be included in the CCS.

The endangered species hot spots depicted in Figure 1 probably represent the interaction between centers of endemism (e.g., clams in southwestern Appalachia [Banarescu 1992], or plants in Florida [Gentry 1986]) and anthropogenic activities (e.g., urbanization and agricultural development). Because different kinds of organisms have different habitat requirements and evolutionary origins, we should not expect them to share centers of endemism. At the same time, human activities may affect different taxa differently. Consequently, hot spots for different groups overlap in only a few areas. Aside from these locations, the key areas for most groups overlap only weakly, which suggests that endangered species hot spots for one group do not necessarily correspond with those of other groups (Dobson et al. 1997).

The extent of hot spot overlap for different taxa is a critical issue in assessing how suitable one group of endangered species might be as an indicator of others. Indicators would be useful in cases where knowledge about endangered species is poor (such as spiders and insects) and another, better-known group might be used as a surrogate for those species.

As there are many more species of endangered plants than of other groups, and plants' distributions tend to be restricted to one or two counties, the CCS for plants is large. Not surprisingly, then, the 136 counties necessary to sample all plants also include more species from other taxa than do the complementary subsets of any other groups. The plant CCS contains 73 percent of all other taxa combined, 94 percent of the birds, 76 percent of the mammals and 74 percent of the herptiles (Dobson et al. 1997). This observation might lead to the conclusion that saving all of the plants should be the main objective of a national conservation strategy, as many other taxa will also be included.

There are two problems with this approach. First, only 4.1 percent of the U.S. landmass is currently protected from all human activities (World Conservation Monitoring Centre 1992), and preservation of endangered plants by setting aside their CCS would more than double this areaÑa politically difficult task. Second, the plant CCS fails to include 45 percent of the fish, 46 percent of the arthropods, and 61 percent of the mollusks (Dobson et al. 1997).

We propose that the power of each complementary subset be used as a measure of how well a particular group indicates the presence of endangered species diversity in all other taxa. Power is calculated by dividing the number of endangered species from other groups present in the complementary county subset by the number of such species in an equivalent area of randomly selected counties. To do this, we used a bootstrapping algorithm that accumulated counties at random until their total area matched or just exceeded that of the complementary subset. Repeating this process 200 times allowed us to assign error bounds to our estimates (Table 1).

Fish and plants emerge as the worst indicators of number of species among other endangered taxa. For fish this is probably because there are not that many other threatened taxa within bodies of water. In the case of the CCS for plants, it would seem that it is hard to sample more species than those already included in their complementary county subset simply because the CCS is so large. Birds and herptiles, on the other hand, seem to be the best indicators for other endangered taxa. Their CCS are 4.00 and 3.26 times better, respectively, than randomly selected county sets of equivalent size.

The main point of this analysis is to demonstrate that a comprehensive, well-designed portfolio of protected areas should be able to include populations of most endangered species in the U.S., if these areas are located in the right places. Table 1 suggests that the critical minimum area to represent all currently endangered species is somewhere between 2 and 10 percent of the total U.S. landmass. Once again, we must emphasize that this might not assure their long-term conservation, but it is a reasonably good starting point. The next step is to evaluate the current system of protected areas and determine which endangered species are present in them and which are not. This will help prioritize areas in need of immediate protection, and the analysis would need to be updated as the number of species listed under the ESA changes.

Recent government studies indicate that over half of the species on the federal endangered species list have more than 80 percent of their habitat on non-federal lands. Of these species, more than 80 percent are in the southern U.S. and in Hawaii. In contrast, a charismatic minorityÑsuch as grizzly bears, bald eagles and Northern spotted owls (Strix occidentalis occidentalis)Ñare found in the north (mainly northwest of the Rockies), with approximately 50 percent of their ranges on federal lands.

This variability in geographic distribution throws important light on discussions about the cost of the ESA both to the federal government and to private landowners. Recent studies have shown that the primary threats to endangered species on federal lands are recreation, grazing of domestic livestock, and timber and mineral extraction (Losos et al. 1995; Wilcove et al. 1996). Many of these activities are subsidized by the federal government, and as they tend to jeopardize the health of the ecosystem where many endangered and threatened species live, the federal government then has to foot the bill for activities that might reduce the risk of future extinctions. The government finds itself in the ironic economic position of subsidizing the activities that threaten species, while also paying for activities aimed at protecting them (Losos et al. 1995). In many cases, the intensity with which endangered species need be protected would be reduced and the costs would be at least revenue neutral if grazing and recreation rights were competitively priced on public lands, and subsidies for timber and mineral extraction were reduced.

The situation in the south is more subtle and would benefit from a more direct reform of the ESA. More than 90 percent of the listed endangered species in the U.S. are found in Hawaii, California, Florida, Texas and southwestern Appalachia, and more than 95 percent of the range of these species is on private lands. In these regions this puts constraints on development of that land, so many developers feel frustrated by the ESA's powers to restrict modification of habitat for endangered species. Furthermore, the presence of potentially endangered species on private land means that many species remain unprotected, lest attempts to list them lead to their "sudden disappearance" from areas which would then be protected from further development.

However, in these areas many other private landowners manage their lands in a way that either directly or indirectly encourages the presence of endangered species. The main threat to this habitat usually occurs when owners die and estate taxes must be paid. In many cases, the estate is subdivided and sold, or logged if the area is forested, in order to pay estate taxes. This suggests that tax reforms and easements would be a potentially valuable way to reward private landowners who manage their land in such a way that it protects endangered species.

Thus, two possible directions in which the ESA could move in the future are toward (1) tax relief for owners of private lands whose activities allow endangered species to persist on their land, and (2) competitive pricing of access to natural resources on federal lands. These legal mechanisms, combined with continued scientific analyses of existing and potential protected areas and species distributions, will help traditionally combatant groups move toward constructive solutions in the twenty-first century.

We are grateful to L. Turner and M. Hood at the Environmental Protection Agency for providing us with the raw data for this analysis. K. M. Clark dramatically improved a previous version of the manuscript. The work was made possible by a grant to the Environmental Defense Fund from the Charles Stewart Mott Foundation.

Banarescu, P. 1992. Zoogeography of fresh waters, vol. 2, Distribution and dispersal of fresh water animals in North America and Eurasia. AULA-Verlag, Wiesbaden, Germany.

Csuti, B., S. Polasky, P. H. Williams, R. L. Pressey, J. D. Camm, M. Kershaw, A. R. Kiester, B. Downs, R. Hamilton, M. Huso, and K. Sahr. 1997. A comparison of reserve selection algorithms using data on terrestrial vertebrates in Oregon. Biological Conservation 80:83-97.

Dobson, A. P., J. P. Rodriguez, W. M. Roberts, and D. S. Wilcove. 1997. Geographical distribution of endangered species in the United States. Science 275: 550-553.

Flather, C. H., L. A. Joyce, and C. A. Bloomgarden. 1994. Species endangerment patterns in the United States. Fort Collins, Colorado, USA, U.S. Department of Agriculture, Forest Service.

Gentry, A. H. 1986. Endemism in tropical versus temperate plant communities. Conservation Biology. The Science of Scarcity and Diversity. M. E. Soulé. Sinauer Associates, Inc., Sunderland, Massachusetts. 153-181.

Losos, E., J. Hayes, A. Phillips, D. Wilcove, and C. Alkire. 1995. Taxpayer-subsidized extraction harms species. BioScience 45:446-455.

Pressey, R. L., C. J. Humphries, C. R. Margules, R. I. Vane-Wright, and P. H. Williams. 1993. Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution 8:124-128.

Pressey, R. L., H. P. Possingham, and C. R. Margules. 1996. Optimality in reserve selection algorithms: when does it matter and how much? Biological Conservation 76:259-267.

Vane-Wright, R. I., C. J. Humphries, and P. H. Williams. 1991. What to protect? - Systematics and the agony of choice. Biological Conservation 55:235-254.

Wilcove, D . S., M. G. Bean, R. Bonnie, and M. McMillan. 1996. Rebuilding the ark: Toward a more effective Endangered Species Act for private lands. Environmental Defense Fund, Washington, D.C.

World Conservation Monitoring Centre. 1992. Global Biodiversity: Status of the Earth's Living Resources. Chapman & Hall, London.

Jon Paul Rodriguez, W. Mark Roberts and Andy Dobson are at the Ecology and Evolutionary Biology Department, Princeton University, Princeton, NJ 08544-1003.