Premium
Evaluating Abundance Estimate Precision and the Assumptions of a Count‐Based Index for Small Mammals
Author(s) -
WIEWEL ANDREW S.,
ADAMS AMY A. YACKEL,
RODDA GORDON H.
Publication year - 2009
Publication title -
the journal of wildlife management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.94
H-Index - 111
eISSN - 1937-2817
pISSN - 0022-541X
DOI - 10.2193/2008-180
Subject(s) - abundance (ecology) , trapping , sampling (signal processing) , mark and recapture , statistics , range (aeronautics) , population , catch per unit effort , ecology , confidence interval , environmental science , fishery , biology , mathematics , demography , computer science , materials science , filter (signal processing) , sociology , composite material , computer vision
ABSTRACT Conservation and management of small mammals requires reliable knowledge of population size. We investigated precision of mark‐recapture and removal abundance estimates generated from live‐trapping and snap‐trapping data collected at sites on Guam ( n = 7), Rota ( n = 4), Saipan ( n = 5), and Tinian ( n = 3), in the Mariana Islands. We also evaluated a common index, captures per unit effort (CPUE), as a predictor of abundance. In addition, we evaluated cost and time associated with implementing live‐trapping and snap‐trapping and compared species‐specific capture rates of selected live‐ and snap‐traps. For all species, mark‐recapture estimates were consistently more precise than removal estimates based on coefficients of variation and 95% confidence intervals. The predictive utility of CPUE was poor but improved with increasing sampling duration. Nonetheless, modeling of sampling data revealed that underlying assumptions critical to application of an index of abundance, such as constant capture probability across space, time, and individuals, were not met. Although snap‐trapping was cheaper and faster than live‐trapping, the time difference was negligible when site preparation time was considered. Rattus diardii spp. captures were greatest in Haguruma live‐traps (Standard Trading Co., Honolulu, HI) and Victor snap‐traps (Woodstream Corporation, Lititz, PA), whereas Suncus murinus and Mus musculus captures were greatest in Sherman live‐traps (H. B. Sherman Traps, Inc., Tallahassee, FL) and Museum Special snap‐traps (Woodstream Corporation). Although snap‐trapping and CPUE may have utility after validation against more rigorous methods, validation should occur across the full range of study conditions. Resources required for this level of validation would likely be better allocated towards implementing rigorous and robust methods.