28th Annual Meeting and Symposium of the
Desert Tortoise Council, February 21-23, 2003
Abstracts

Monitoring the Desert Tortoise: Needs for the Present and Future

C.R. Tracy¹, K. E. Nussear¹, P.A. Medica², R.M. Marlow¹, and P.S. Corn^3
1Ecology, Evolution and Conservation Biology, and Biological Resources Research Center, University of Nevada, Reno, NV 89557
²United States Fish & Wildlife Service, 1510 North Decatur Blvd., Las Vegas, NV 89108
³Aldo Leopold Wilderness Research Institute, Missoula, MT 59807 USA

Transect sampling is currently the standard method for estimating population density of the federally listed desert tortoise. Estimates of population density from transect sampling requires accurate estimates of the probability that tortoises are available to be seen (G_o), and it requires accurate estimates of the detectability of tortoises (P_a) that are available to be seen. Each of these parameters depends upon other variables including daily and seasonal influences on tortoise behavior and activity, observer differences in different habitats, local and regional differences in physiognomy and vegetation, annual differences in rainfall and food resources, and more. Small inaccuracies in these parameters caused by different levels of the mechanistic determining variables make obtaining precise estimates of tortoise population density very difficult, Furthermore, estimating detectability of tortoises using "distance sampling" techniques requires too much of such a sparsely populated species. Additionally, statistical power analyses of density estimates of desert tortoise suggest that it would be very difficult to estimate population density under the best of circumstances. This is not an unusual circumstance as scientists monitoring bald eagles and other large species with ponderous population dynamics have faced similar challenges to obtaining density estimates and analyses of population trends. However, direct calibration of detectability, and sampling along transects in a way that creates replication may allow accurate and precise estimates of population density using the transect methods.

Even after we have good estimates of population density, those estimates need to be used to calculate population sizes. Furthermore, modern monitoring has additional elements other than estimates of population size. Monitoring needs to assemble data on (1) population size, (2) habitat for the focal species, and (3) threats to the species and ecosystems. Additionally, monitoring needs to be hypothesis based, and accumulate data on management actions. Thus, the efficacy of management actions such as highway fencing, removing domestic grazers, allowing further habitat fragmentation through urbanization, allowing increased population densities of feral dogs, etc. all need to be experimentally monitored through scientifically defensible protocols. Thus, an effective monitoring program will have more elements than we have currently in order to inform management of the species and to gather data necessary for future delisting when delisting criteria are all met.