Estimating Decline and Distributional Change in Amphibians

Abstract

Concern over species declines has prompted researchers to use historical data as a basis for comparison with present-day information from the same sites to assess changes in presence/absence distributions. A review of the literature revealed that these resurveys typically lasted for 1 or 2 years, and many were based on museum records or other data relying on known historical presences. Using data on nine amphibian species from a set of 32 ponds at the E. S. George Reserve ( ESGR ) in Michigan, we evaluated the importance of the duration of a resurvey and the type of historical data used ( information on historical presences and absences vs. historical presences only ). We compared data we collected between 1996 and 2000 with information from the same ponds collected between 1967 and 1974. By systematically degrading the 1996–2000 data, we determined that a resurvey lasting 1 year would yield an estimated 45% decline in the number of presences, whereas a resurvey lasting 2 years would yield an estimated 28% decline. In contrast, a 5-year resurvey would yield an estimated 3% decline in the number of presences. In addition, when our historical data were limited to known presences in the past, even a 5-year resurvey yielded an estimated 30% decline in the number of presences. Our results suggest that estimates of decline and distributional change can be extremely sensitive to the duration of resurvey effort and the type of historical data used. The pattern we found in analyses of ESGR data is echoed in published studies in which multiple-year resurveys tended to yield smaller estimates of decline than single-year resurveys. Based on our findings, we suggest that future resurveys extend for long enough to estimate the value of additional data and that geographic scales of inference be chosen based on the amount and quality of historical information.