Evaluating heterogeneity of sex-specific capture probability and precision in camera-trap population estimates of tigers

Abstract

Population estimates of tigers (Panthers tigris) are affected by movement patterns of the sexes. We evaluated sex-specific heterogeneity to estimate the tiger population at Ranthambhore Tiger Reserve, western Rajasthan, India. We also compared population attributes obtained using a likelihood-based spatially explicit capture–recapture (SECR) method with traditional approaches that use alternate information on animal space used to estimate an effective sampled area around the traps. We obtained 122 tiger photo-captures of 27 individuals (14 F and 13 M) from 165 camera-trapping sites in a 215-km2 area over 3,300 trap-nights, for 20 sampling occasions in 2010. The capture probability of males was about 2 times greater than that of female tigers, but overall detection probability was high, with total population estimate of about 30.2 tigers (SE = 3.7). Male and female tiger density estimates/100 km2 were 2.8 (SE = 0.8) and 4.1 (SE = 1.1), respectively, and overall density estimates of 6.1 (SE = 1.2) individuals/100 km2 were obtained using the SECR method. Non-spatial models using the full mean maximum distance moved (MMDM) to estimate an effective sampled area underestimated the density by 9% compared with the SECR model as a baseline, and 1/2MMDM models overestimated the density by 25% as compared with results from the SECR model. We conclude that to assess the reliability of abundance estimates, heterogeneity in sex-specific capture probabilities should be considered. We recommend using an SECR model because these methods appear to be minimally biased for estimating an effective sampled area and can be applied consistently among studies of various taxa. © 2014 The Wildlife Society.