Estimates of population size for a rare and cryptic species: The importance of spatial clustering of individuals

Estimating population size is a critical component of the current management strategy for the Mojave desert tortoise (Gopherus agassizii). Estimates are used to evaluate the status of the population as well as the effectiveness of conservations actions. Monitoring populations based solely on estimates of population size can be effective for many species. However, individuals of species such as the desert tortoise can be very cryptic, are rarely active out of their burrows and are sparsely distributed in space. These characteristics may impair our ability to estimate population size accurately and precisely. Additional difficulties can evolve from the assumption that desert tortoises are randomly distributed in space. The goals of this study were to determine the extent to which desert tortoises are clustered in the landscape, and to assess the effects of spatial clustering on estimates of the population sizes of  the desert tortoise. We analyzed a combination of empirical data from transect sampling of tortoise presence as well as simulated data of the dispersion of individuals in space. Our analyses indicate that spatial clustering occurs at both a regional scale and a local scale. Further, we assessed the magnitude of errors in estimates of population size resulting from (a) densities of clusters and (b) densities of individuals within clusters. Without accounting for population clustering, the estimates of population sizes for natural distributions of desert tortoises likely include very misleading errors. Thus, the precision required to assess short- and long-term trends for desert tortoises may not be attainable using estimates of population size. Future efforts for monitoring desert tortoise populations may require additional metrics aside from population densities.