The Feasibility of Gray Wolf Reintroduction to the Grand Canyon Ecoregion

Paul G. Sneed Environmental Studies , Master of Arts Program, Prescott College, 4906 Box Canyon Road, Billings, MT 59101; (406) 245-9117; (fax) (406) 245-0787; [email protected]

Abstract As part of a regional conservation planning initiative, this study is being undertaken to determine the biophysical and socioeconomic feasibility of reestablishing a top carnivore, the gray wolf (Canis lupus), in the Grand Canyon Ecoregion (GCE). The GCE is a roughly 1.5 million km2 area located on the southern Colorado Plateau. The last remaining gray wolves were probably eradicated in the 1920s and 1930s. Because of an interest in restoring extirpated native species to this ecoregion, and the desire to increase the size of the gray wolf metapopulation in the Southwest, there is need for an objective and spatially explicit landscape-scale model of potential gray wolf habitat. The first phase of this conservation GIS analysis involves utilizing six habitat characteristics or factors—vegetation cover, surface water availability, prey density, human population density, road density, and land ownership—to identify and describe potential reintro- duction sites in the Arizona section of the Grand Canyon Ecoregion. Initial results show that there are at least two localities in northern Arizona suitable for reintroduction of around 100 wolves. This paper is a preliminary report on observations, results, and some recommendations deriving from the feasibility study.

Introduction Crumbo 1998). This paper, how- Regardless of inaccuracies in the Conservation biologists have shown ever, will focus on current, prelimi- historical record, a partial picture of that large or top carnivores are often nary results from research done on where wolves occurred prior to their keystone species whose removal a limited number of factors in the extermination in the Southwest can jeopardizes the maintenance of eco- northern Arizona section of the still be pieced together. These records logical integrity in large-scale ecosys- ecoregion (see Figure 1). show that at least small populations tems (Soule and Noss 1998; Terborgh of wolves were found throughout the et al. 1999). Therefore, conservation Historic occurrence and woodlands and forests of northern planners interested in restoring and taxonomic position Arizona (Brown 1984). For example, protecting large ecosystems or To accurately reconstruct the his- from these records we know that there ecoregions emphasize recovery of top toric distribution of gray wolves in were at least 30 wolves on or near the predators. The primary goal of this the GCE is challenging for a vari- North Kaibab because of the number study—which is supported by the ety of reasons. Nineteenth century reported killed between 1907 and Grand Canyon Wildlands Council, writers often accidentally or pur- 1926 (Russo 1964). Brown (1984) Defenders of Wildlife, and Prescott posefully misidentified coyotes claimed that "the last wolf in this part College—is to determine the capabil- (Canis latrans), wolves, and wolf- of northern Arizona was taken on the ity and suitability (together, the fea- dog hybrids (Gipson et al. 1998). Paria Plateau about 1928", but a sibility) of reintroducing one top car- Wolf hunters and trappers some- former Civilian Conservation Corps nivore, the gray wolf (Canis lupus), times exaggerated the number of worker recently reported that he saw to the Grand Canyon Ecoregion wolves in an area to enhance their wolves on three different occasions (GCE) (Figure 1). job security and occasionally mis- in 1935 on the North Rim of Grand Ultimately, this study will ad- represented where a wolf was killed Canyon National Park (GCNP) dress 26 factors or aspects, grouped in order to claim a local bounty. (Leslie, personal communication). into two dimensions—biophysical Furthermore, the widespread use of Moreover, "as recently as March 3, and human—that are expected to af- poisons meant that many animals, 1948, assistant chief ranger A.L. fect the feasibility of wolf recovery including wolves, were dispatched Brown reported wolf tracks in fresh in the entire GCE (Sneed and without any record of their death. snow in the area of Bright Angel

Vol. 18 No. 4 2001 Endangered Species UPDATE 153 contemporary molecular ge- to adequate food supplies, security netics (Wayne et al. 1992). from human disturbance and perse- Adopting this approach, cution are important factors affect- Nowak (1995) lumps the two ing the suitability of a landscape for previously identified GCE wolf recovery. At this stage in the wolf subspecies, (C. l. research, three critical human di- mogollonensis and C. l. mension aspects are considered: youngii), in with the geographi- human population density, road cally widespread subspecies C. density, and land status. l. nubilus. Nowak (1995) also affirms the validity of a truly Biophysical factors Southwestern subspecies, the Several reintroduction studies (e.g., Mexican wolf (C. l. baileyii), Mladenoff et al. 1995) suggest that which may have occupied or gray wolves, at least those living Figure 1. Grand Canyon ecoregion dispersed into the southeast- south of the Arctic, tend to prefer for- ern part of the GCE. ested landscapes. Historically, in the Southwest, wolves were most com- Point" on the North Rim of GCNP Habitat capability and monly found associated with wood- (Hoffmeister 1971). Finally, the last suitability mapping lands and montane forests (Groebner wolf inhabiting the Mogollon Rim Restoration of viable large carnivore et al. 1995; FWS 1996). When ob- area in the southern part of the GCE populations is probably among served elsewhere, such as in grass- was reportedly taken in 1942 society's greatest challenges, requir- lands, they were probably simply (Hoffmeister 1986). Clearly, gray ing extraordinary innovation and co- passing through as they moved be- wolves occurred within the Grand operative management on an tween their preferred habitat of for- Canyon Ecoregion well into the twen- ecoregional scale (Paquet and Hack- ested highlands. Figure 2 maps the tieth century, although their exact man 1995). Furthermore, solutions distribution of these two vegetation numbers and range will probably to large predator conservation are types, as well as others such as never be known with certainty. economic, sociological and political shrublands and grasslands. This fig- Due to the taxonomic splitting (human dimension issues), as well as ure plainly illustrates a broad band of approach of the time, Young and biological and ecological (biophysi- forestlands-woodlands extending Goldman (1944) identified 23 sub- cal factors) (Clark et al. 1996). The north-south from the Kaibab Pla- species of North American gray feasibility of wolf recovery depends teau, through the Flagstaff area to wolves (based on skull measure- on the capability and suitability of the Mogollon Rim, interrupted only ments, pelage color, and size) and habitat for sustaining wolf popula- by the Grand Canyon and urbanized mapped their geographic distribution. tions. Although many factors can and areas such as Flagstaff. Other ar- Two of these 23 nominal subspe- should be considered, the ultimate de- eas of woodland/forest vegetation cies—C.l. mogollonensis (the Ari- terminants are ungulate prey and hu- types are found in isolated moun- zona wolf), C.l. youngii (the Great man impact (Fuller et al. 1992) or, put tain areas of the Arizona Strip as Basin or Intermountain wolf), and, another way, sustenance and security. well as the Hualapai and Navajo possibly, C.l. baileyii (the Mexican Course screen landscape-scale Indian Reservations. wolf)—inhabited the Grand Can- habitat mapping for the Arizona por- Because wolves require large yon Ecoregion (Brown 1984; FWS tion of the GCE (see Figure 1) has amounts of water to aid digestion 1996). Development of similar been done following other similar (Lopez 1978; Mech 1970), several classification schemes continued studies (Mladenoff et al. 1995; studies of wolves in the Southwest into the 1970s (e.g., Hall and Quinby et al. 1999; Ratti et al. 1999; (Groebner et al. 1995; FWS 1996) Kelson 1959) until some taxono- Wydeven et al. 1998). Various bio- and elsewhere (Quinby et al. 1999) mists began questioning the split- physical factors can be considered in have suggested that the availability ting tradition of wolf taxonomy. evaluating the capability of habitat to of free water is an important deter- Modern lumping systems of wolf support wolves, but this study focuses minant of gray wolf abundance and taxonomy are based on multivariate on vegetation cover, surface water distribution. Figure 3 illustrates the statistical analysis of large sample availability, and, most importantly, distribution of currently mapped sizes and confirmed by the results of ungulate prey abundance. In addition lakes, springs, and streams in the Ari-

154 Endangered Species UPDATE Vol. 18 No. 4 2001 man population den- sity. Studies (e.g., Mladenoff et al. 1995; Ratti et al. 1999) have shown that lands with a hu- man population den- sity greater than 12 to 13 persons per square kilometer will not be suitable wolf habitat. The Figure 2. Vegetation cover Figure 3. Surface water map displayed in Figure 6 indicates zona portion of the Grand Canyon teau enjoys a very high density of that most of the Arizona section of Ecoregion. Although the digital data mule deer (eight to 13 animals per the Grand Canyon Ecoregion has available is very incomplete, this fig- km2), while the remainder of the area population densities less than 13 per- ure shows that there are more than has an adequate density of three to sons per km2. Except for the Flag- enough sources of surface water on eight deer per km2. The Coconino staff-Sedona urban zone, the entire the Kaibab Plateau, in the Flagstaff Plateau around Flagstaff also supports north-south corridor from the Kaibab area, and along the Mogollon Rim. quite dense populations (three to eight Plateau to the eastern, slightly urban- This conclusion is supported by the per km2). Furthermore, similar den- ized, part of the Mogollon Rim has a observed presence of relatively high sities probably exist on parts of the human population density of less than numbers of other large predators (e.g., Hualapai and Navajo reservations in four people per km2. Not surprisingly, mountain lions (Felis concolor) and the GCE, but no data is readily avail- this same corridor has high prey spe- prey species in these locales. able to confirm this supposition. cies densities and seems capable of Clearly, one of the most impor- Even if the current mule deer popu- supporting wolves. tant determinants of suitable wolf lation density is one-half of what Wolves are usually not threatened habitat is the abundance of ungulate these figures indicate, as some Ari- by roads, except when they are struck prey species. The primary prey spe- zona Game and Fish Department per- by motor vehicles (Mech 1977). cies for wolves in the Grand Canyon sonnel suggest (e.g., Goodwin, per- Nonetheless, roads can provide ac- Ecoregion are mule deer (Odocoileus sonal communication), there are still cess to generally undisturbed areas hemionus), followed in order of im- more than sufficient deer densities to where humans may harass or kill portance by elk (Cervus elaphus), support gray wolves. Figure 5 shows wolves. Studies of road density and pronghorn (Antilocarpa americana), that elk densities, while somewhat wolf distribution relationships by and bighorn sheep (Ovis canadensis). lower on average than mule deer, are Thiel (1985) and Mech et al. (1988) Information about abundance (den- quite high (i.e. two to three animals suggest a road density threshold value sity) and distribution of these wild- per km2) around Flagstaff and south- of between 0.6 and 0.8 kilometers of life species in the Arizona part of the east along the Mogollon Rim. Of road per square kilometer of area. ecoregion was obtained from the Ari- course, elk also average three times Higher road density values generally zona Game and Fish Department. the biomass of deer. Figures 4 and 5 result in unsuccessful breeding at- Figures 4 and 5 display the approxi- combined map an adequate ungulate tempts. Mladenoff et al. (1995), us- mate density of mule deer and elk prey base extending north-south from ing radio collar data on recolonizing populations in Arizona GCE. the Kaibab Plateau, through the Flag- wolves in northern Wisconsin, dis- Other reintroduction studies (e.g., staff area, and southeast along the covered that road density and fractal FWS 1996) indicate that a density of Mogollon Rim. dimension—reflecting the degree of approximately two to six deer per km2 habitat fragmentation (often the re- would be required to support a Mexi- Human dimensions sult of road building)—were the most can wolf population and, presumably, An important determinant of habitat important predictors of favorable similar numbers would be adequate suitability for gray wolves and other wolf habitat. Figure 7 shows that for wolves in the GCE. Figure 4 dem- large carnivores such as grizzly bears most of the north-south corridor, ex- onstrates that much of the Kaibab Pla- (Merrill et al. 1999) seems to be hu- tending from the Kaibab Plateau

Vol. 18 No. 4 2001 Endangered Species UPDATE 155 through to the Mogollon Rim south- The landscape-scale east of Flagstaff, has road densities habitat mapping, included in higher than 0.68 km per km2, but gen- this progress report, is admit- erally lower than 1.4 km per km2. tedly somewhat incomplete Road density in many parts of the at this stage in the research. GCE is somewhat higher than recom- Nonetheless, the mapped mended in other studies, but most of variables of both biophysical the numerous roads in the ecoregion and human dimensions point are tertiary or unimproved roads strongly towards the prob- that could be eliminated on public ability that at least two ar- lands with a vigorous road-closing eas—the Kaibab Plateau and program. Furthermore, the low hu- much of the Mogollon Rim man density numbers (Figure 6) south and east of Flag- Figure 4. Mule deer density might indicate that these areas are staff—are capable of sup- favorable wolf recovery habitat de- porting viable wolf popula- spite the existence of relatively high tions and suitable for rein- unimproved road densities. troduction of gray wolves Favorable land status, defined here as lands in public ownership and, Projected wolf densities especially, designated protected ar- Assuming that wolf reintro- eas, can help make a landscape suit- duction is feasible, it is rea- able for gray wolf reintroduction sonable to ask how many (Southern Rockies Ecosystem wolves might the Arizona Project, 1998). Identifying, describ- portion of the Grand Canyon ing, and mapping proposed and des- Ecoregion support. Utilizing ignated wilderness areas and other ar- the existing deer and elk den- eas designed to protect ecological sity distribution maps (Fig- Figure 5. Elk density processes or wildlife, such as the ures 4 and 5), and following Grand Staircase/Escalante National Fuller (1989), very prelimi- Monument and the Grand Canyon nary calculations of predicted Game Preserve in the Kaibab Forest wolf density were done us- (Miller 1996) is especially important. ing these equations: Figure 8 maps distribution of public lands, both state and federal, exclu- W = 3.4 + 3.7D sive of Indian reservations. This re- veals that a wide band of federal pub- and lic lands (including large tracts of pro- tected areas) runs north-south from W = 3.4 + 3.7(3E) the Kaibab Plateau through the Flag- staff area and southeast along the where W is predicted wolf Mogollon Plateau (again, corre- density (per 1000 km2), D is Figure 6. Human density sponding with the distribution of estimated mule deer density important biophysical factors). (per km2), and 3E is estimated State lands, even though currently elk density (per km2) times a relative suggest that to reintroduce at least 100 interspersed in a "checkerboard" biomass value (elk biomass is 3 x 1 gray wolves into the Arizona portion fashion (see Figure 8) with private deer). Both low and high ungulate den- of the Grand Canyon Ecoregion and federal lands, could be consoli- sity estimates were utilized in calculat- would be feasible. dated through land trades and pur- ing a range of predicted wolf numbers chases to create wildlife corridors shown in Table 1. Potential stock for wolf between federal public lands such Although this facet of the study reintroduction as the Coconino and Kaibab Na- is far from finished, historical records When the time comes to make a de- tional Forests. and initial carrying capacity research cision about reintroducing gray

156 Endangered Species UPDATE Vol. 18 No. 4 2001 tundra) and experienced in tial reintroduction sites in the Arizona hunting the types of ungu- section of the Grand Canyon late species that are most Ecoregion. Initial results show that abundant in the GCE. there are at least two localities in Alternatively, captive or northern Arizona available for rein- wild-bred Mexican wolves troduction of around 100 wolves. (C.l. baileyii) could be uti- Source stock for wolf recovery in the lized for reintroduction. GCE could come from existing large Given the difficulties expe- C. l. nubilus populations in the Great rienced with captive bred Lakes and/or a recovered C. l. baileyii stock in the current Mexican population in the Southwest. Clearly, wolf recovery effort, how- the future extension of wolf recovery Figure 7. Road denisty ever, it seems best to wait for into northern Arizona and other parts the availability of surplus of the GCE will have to be done un- wild-raised stock GCE (Par- der the legal mandate of the ESA sons, personal communica- and will most likely be sponsored tion). Also, when the Mexi- by a federal agency such as the Fish can wolf population reaches and Wildlife Service and/or Na- a viable size in the wild, dis- tional Park Service. persers from eastern Arizona Further investigation and study and western New Mexico will continue to refine the habitat ca- will likely attempt to colo- pability and suitability analyses, as nize the southeastern part of well as to help determine the most ap- the GCE. Thus, this recov- propriate subspecies for wolf reintro- ery opportunity in the GCE duction in the ecoregion. In the end, could help extend the geo- however, the most important consid- Figure 8. Land status graphic range and eration is how to best assist nature in metapopulation of the cur- restoring gray wolves to the Grand wolves to the Grand Canyon rently recovering, but still endan- Canyon Ecoregion and thereby help Ecoregion, we should "…consider gered, Mexican wolf. in the national effort to conserve behavioral or demographic factors to this magnificent and ecologically be more important than maintenance Conclusions essential carnivore. of the genetic purity of putative wolf The first phase of this landscape-scale subspecies…" (Wayne et al. 1992). analysis involved utilizing six factors Acknowledgements If Nowak's (1995) recent revision of of the biophysical and human dimen- This project was primarily funded by wolf taxonomy is accepted, it seems sions to identify and describe poten- the Grand Canyon Wildlands Coun- biologically appropriate that stock for reintroduction could be taken from anywhere in the historic range of C. Table 1. Minimum and maximum number of wolves predicted to oc- l. nubilus. While finding areas of cupy Arizona part of Grand Canyon Ecoregion based on combined surplus wolf populations with habi- mule deer and elk densities and areas. tat exactly comparable to the GCE will be difficult, regions such as the L)ocation A)rea (square kms Wolves (pop. range Great Lakes, currently supporting C.l. nubilus populations, do exhibit analo- gous forested ecosystems (albeit dif- N2orth Kaibab 1247 35 to 6 ferent forest types) and have similar ungulate prey species (i.e., deer and South Colorado 35849 80 to 12 elk). Wild wolves translocated from Plateau the Great Lakes region, for ex- ample, would at least be habituated T1otal Arizona GCE 5732 115 to 18 to forest habitats (as opposed to

Vol. 18 No. 4 2001 Endangered Species UPDATE 157 cil, Defenders of Wildlife, and bana, Illinois. Final draft: feasibility study on the reintro- Prescott College. Although I am Hoffmeister, D.F. 1986. Mammals of Ari- duction of gray wolves to the Olympic Pen- zona. The University of Arizona Press: insula. Unpublished report prepared for the solely responsible for any errors or Tucson, AZ. United States Fish and Wildlife Service, shortcomings, I could not have made Lopez, B.H. 1978. Of wolves and men. Lacey, Washington. the progress I have without the help Charles Scribner's Sons, New York, New Russo, J.P. 1964. The Kaibab North deer of these people: Kelly Burke, Larry York. herd – Its history, problems and manage- Stevens, and Kim Crumbo of the Mech, L.D. 1970. The wolf: the ecology and ment: A research and management study. behavior of an endangered species. Uni- Wildlife Bulletin No. 7. Arizona Game and Grand Canyon Wildlands Council; versity of Minnesota Press, Minneapolis, Fish Department: Phoenix, Arizona. Melissa Grigione and Dan Smith, Minnesota. Sneed, P. and K. Crumbo. 1998. The rein- Defenders of Wildlife; Mike Mech, L.D. 1977. Productivity, mortality, troduction of gray wolves to the Grand Siedman, Phoenix Zoo; Mike Senn, and population trends of wolves in north- Canyon Ecoregion: A proposed feasibility eastern Minnesota. Journal of Mammology study. Unpublished research proposal. Rick Miller, John Goodwin, Ray Lee, 58:559-574. Grand Canyon Wildlands, Flagstaff, Ari- Matt Alderson, and Amber Munoz, Mech, L.D., S.H. Fritts, G.L. Raddel, and W.J. zona. Arizona Game and Fish Department; Paul. 1988. Wolf distribution and road Soule, M. and R. Noss. 1998. Rewilding and Elaine Leslie, Grand Canyon Na- density in Minnesota. Wildlife Society biodiversity: Complementary goals for con- tional Park; Rick Moore and Aimee Bulletin 16:85-87. tinental conservation. Wild Earth 8(3): 18- Merrill, T., D.J. Mattson, R.G. Wright, and 28. Porcaro, Grand Canyon Trust; Dave H.B. Quigley. 1999. Defining landscapes Southern Rockies Ecosystem Project. 1998. Hanna and Craig Martinsen, Prescott suitable for restoration of grizzly bears Wolf recovery in the Greater Southern College; and, Jim Riedman, Arizona Ursus arctos in Idaho. Biological Conser- Rockies: an initial suitability analysis. Land Resources Information System. vation 87:231-248. Unpublished in-house report, Nederland, Miller, R. 1996. History of the intent and Colorado. management of the Grand Canyon Game Terbourgh, J., J.A. Estes, P. Paquet, K. Ralls, Literature cited Preserve. Unpublished report prepared for D. Boyd, B. Miller and R. Noss. 1999. The Brown, D.E., ed. 1984. The wolf in the south- Arizona Game and Fish Department, Flag- role of top carnivores in regulating terres- west: the making of an endangered species. staff, Arizona. trial ecosystems. Pp 39-64 in Soule, M. The University of Arizona Press, Tucson, Mladenoff, D.J., T.A. Sickley, R.G. Haight, and J. Terbourgh, ed. Continental conser- Arizona. and A.P. Wydeven. 1995. A regional land- vation: design and management principles Clark, T.W., A.P. Curlee, and R.P. Reading. scape analysis and prediction of favorable for long-term, regional conservation net- 1996. Crafting effective solutions to the gray wolf habitat in the northern Great works. Island Press: Washington, D.C. large carnivore conservation problem. Lakes region. Conservation Biology 9:279- Thiel, R.P. 1985. Relationship between road Conservation Biology 10:940-948. 294. densities and wolf habitat suitability in Fuller, T.K. 1989. Population dynamics of Nowak, R.M. 1995. Another look at wolf Wisconsin. American Midland Naturalist wolves in north-central Minnesota. Wild- taxonomy. Pp 375-398 in L.N. Carbyn, 113:404-407. life Monographs 105:1-41. S.H. Fritts, and D.R. Seip, eds. Ecology United States Fish and Wildlife Service. 1996. Fuller, T.K., G.L. Raddle, M.S. Lenarz, and and conservation of wolves in a changing Final environmental impact statement— G.B. Joselyn. 1992. A history and current world. Occasional Publication No. 35, Reintroduction of the Mexican wolf within estimate of wolf distribution and numbers Canadian Circumpolar Institute, University its historic range in the southwestern United in Minnesota. Wildlife Society Bulletin of Alberta, Edmonton, Alberta. States. U.S. Fish and Wildlife Service, 20:42-55. Paquet, P.C. and A. Hackman. 1995. Large Albuquerque, New Mexico. Gipson, P.S., W.B. Ballard, and R.M. Nowak. carnivore conservation in the Rocky Moun- Wayne, R.K., N. Lehman, M.W. Allard, and 1998. Famous North American wolves and tains. World Wildlife Fund Canada, R.L. Honeycutt. 1992. Mitochondrial the credibility of early wildlife literature. Toronto, Ontario. DNA variability of the gray wolf: genetic Wildlife Society Bulletin 26:808-816. Quinby, P., S. Trombulak, T. Lee, J. Lane, M. consequences of population decline and Groebner, D.J., A.L. Girmendonk, and T.B. Henry, R. Long, and P. MacKay. 1999. habitat fragmentation. Conservation Biol- Johnson. 1995. A proposed cooperative Opportunities for wildlife habitat connec- ogy 6:559-569 reintroduction plan for the Mexican wolf tivity between Algonquin Park, Ontario and Wydeven, A.P., T.K. Fuller, W. Weber, and K. in Arizona. Technical Report 56, the Adirondack Park, New York. Unpub- MacDonald. 1998. The potential for wolf Arizona Game and Fish Department, Phoe- lished report prepared for The Greater recovery in the northeastern United States nix, Arizona. Laurentian Wildlands Project, South via dispersal from southeastern Canada. Hall, E.R. and K.R. Kelson. 1959. The mam- Burlington, Vermont. Wildlife Society Bulletin 26:776-784. mals of North America. Vol. 2. The Ronald Ratti, J.T., M. Weinstein, J.M. Scott, P. Young, S.P. and E.A. Goldman. 1944. The Press, New York, N.Y. Avsharian, A. Gillesberg, C.A. Miller, wolves of North America. American Wild- Hoffmeister, D.F. 1971. Mammals of Grand M.M. Szepanski, and L.K. Bomar. 1999. life Institute, Washington, D.C. Canyon. University of Illinois Press, Ur-

158 Endangered Species UPDATE Vol. 18 No. 4 2001