Bioregional monitoring design and occupancy estimation for two Sierra Nevadan amphibian taxa

Abstract

Land-management agencies need quantitative, statistically rigorous monitoring data, often at large spatial and temporal scales, to support resource-management decisions. Monitoring designs typically must accommodate multiple ecological, logistical, political, and economic objectives and constraints. We present a long-term bioregional monitoring program to assess the status and change in populations of the federally listed candidate frog species, Yosemite toad (Anaxyrus [Bufo] canorus) and mountain yellow-legged frog (Rana muscosa/sierrae complex), on USDA Forest Service lands in the Sierra Nevada, California. The program takes advantage of advances in survey design and analysis to: 1) collect data at a metapopulation scale (i.e., small basins), 2) provide occupancy data on ≥2 species with overlapping ranges with the same field-monitoring protocols, 3) provide occupancy estimates applicable to the entire range of each species in the study region, 4) incorporate information from historical occupancy records, and 5) link the survey design to an existing survey design. We estimated occupancy assuming imperfect detection by extending existing procedures for maximum likelihood estimation to incorporate the unequal probability of selection used in the survey design. From 2002 to 2009, we estimate that the Yosemite toad used 0.25 ± 0.01 (SE), 0.86 ± 0.04, and 0.86 ± 0.03 of basins over its range, with historical presence, and with presence since 1990, respectively, and the mountain yellow-legged frog used 0.04 ± 0.01, 0.43 ± 0.04, and 0.47 ± 0.04 of basins over its range, with historical presence, and with presence since 1990, respectively. Survey date and snow pack affected detection of the Yosemite toad but not of the mountain yellow-legged frog. Monitoring costs were reduced by using a complex survey design with panels that required generalizing existing methods for estimating occupancy under imperfect detection.