Twenty-Third Annual Meeting and Symposium of the
Desert Tortoise Council, April 3-5, 1998
Abstracts

The use of distance sampling for estimating desert tortoise populations is briefly reviewed and forms a foundation for the current study. Data for this study was obtained from the eight 9 km² study plots in the southern Mojave Desert established at Marine Corps Air Ground Combat Center and Joshua Tree National Park in 1995. An experimental design was developed to sample live tortoises, carcasses, burrows, and scats using distance sampling as the population (in the statistical sense) estimator. Tortoise, burrow, and scat densities were estimated at nested multiple-scales to produce an unbiased desert tortoise distribution/density surface for the landscape. This analysis and protocol was presented at the Desert Tortoise Council Symposium in 1997. The experimental design and resulting database provided an opportunity to directly compare distance sampling and strip transects of varying widths for estimating burrow and scat densities on landscape scales at plots that varied enormously in desert tortoise abundance. The tested hypothesis was: Tortoise sign density estimates using distance sampling should be independent of and indeed similar at any surveyed band width; while wide strip transects should underestimate tortoise sign density, because all sign within the transect are not detected by the surveyor. However, as strip transect widths approach the detection function parameter calculated from distance sampling, the two methods should yield comparable results. The optimal strip transect width for estimating scat densities should be much narrower than the corresponding transect band for burrows, because scats are much more difficult to observe than burrows. The data strongly and unequivocally support the posed hypothesis. Distance sampling is time intensive because the perpendicular distance between each located object and an established line transect must be accurately measured. The data from my protocol suggest that distance sampling need only be used on actual tortoises. Tortoise sign densities can be adequately estimated from strip transects whose optimal widths are separately calculated for burrows and scats from an initial pilot study using distance sampling. This pilot study is necessary, because optimal strip transect width is habitat and to some extent year specific. For example, in a highly productive year the increase in vegetation and litter would require narrower strip transects for burrows and scats.