Evaluating capture-recapture population and density estimation of tigers in a population with known parameters

• Published in: Animal conservation Vol. 13; no. 1; pp. 94 - 103
• Main Authors: Sharma, R.K, Jhala, Y, Qureshi, Q, Vattakaven, J, Gopal, R, Nayak, K
• Format: Journal Article
• Language: English
• Published: Oxford, UK Oxford, UK : Blackwell Publishing Ltd 2010; Blackwell Publishing Ltd
• Subjects: 1/2 MMDM; abundance; bias and precision; camera density; Cameras; Conservation; effective trapping area; Endangered species; home-range radius; Kanha Tiger Reserve India; Panthera tigris; Population density; Sampling; sampling cost; spatial likelihood density estimation; Telemetry; Trapping; Traps; India
• ISSN: 1367-9430; 1469-1795
• DOI: 10.1111/j.1469-1795.2009.00305.x

Conservation strategies for endangered species require accurate and precise estimates of abundance. Unfortunately, obtaining unbiased estimates can be difficult due to inappropriate estimator models and study design. We evaluate population-density estimators for tigers Panthera tigris in Kanha Tiger Reserve, India, using camera traps in conjunction with telemetry (n=6) in a known minimum population of 14 tigers. An effort of 462 trap nights over 42 days yielded 44 photographs of 12 adult tigers. Using closed population estimators, the best-fit model (program capture) accounted for individual heterogeneity (Mh). The least biased and precise population estimate ( [graphic removed] (SE) [ [graphic removed] ]) was obtained by the Mh Jackknife 1 (JK1) [14 (1.89)] in program care-2. Tiger density ( [graphic removed] (SE) [ [graphic removed] ]) per 100 km² was estimated at 13 (2.08) when the effective trapping area was estimated using the half mean maximum distance moved (1/2 MMDM), 8.1 (2.08), using the home-range radius, 7.8 (1.59), with the full MMDM and 8.0 (3.0) with the spatial likelihood method in program density 4.1. The actual density of collared tigers (3.27 per 100 km²) was closely estimated by home-range radius at 3.9 (0.76), full MMDM at 3.48 (0.81) and spatial likelihood at 3.78 (1.54), but overestimated by 1/2 MMDM at 6 (0.81) tigers per 100 km². Sampling costs (Rs. 450 per camera day) increased linearly with camera density, while the precision of population estimates leveled off at 25 cameras per 100 km². At simulated low tiger densities, a camera density of 50 per 100 km² with an effort of 8 trap nights km⁻² provided 95% confidence coverage, but estimates lacked precision.