| Summary: | Abstract Wildlife monitoring is of fundamental importance to establish baseline information, measure population changes, and extinction risk. Motion‐triggered camera traps are an increasingly popular tool for monitoring terrestrial species over large landscapes. Over the years, occupancy has become a robust and unbiased state variable to monitor species worldwide. However, the optimal sampling design required for robust estimations of occupancy is lacking for many species. Here, we estimated the optimum sampling design by varying the number of sites (50–400) and sampling days (10–25) for a range of mammal species using camera‐trap survey data from central India. We used power analysis and mean‐squared error and evaluated the hypothesis of how various species‐specific traits influence occupancy and detectability of the species. We found that mean‐squared error changed significantly with the number of sampling sites for rare species, whereas for species with moderate and high detection probability, the mean‐squared error changed significantly with the number of sampling occasions. Power increased with an increase in the number of sampling sites and occasions for all species, although the change was not significant for species with higher occupancies or detection probabilities. We found that body size was positively related to occupancy but did not influence detection probability significantly. No relationship was detected with social status or diet on occupancy or detection probability. Our results suggest a minimum of 50 sites for 15–20 d for common species and 100 sites for 20–30 d for rare, elusive species. Our results provide guidelines to managers and practitioners for effective allocation of cost and sampling effort for a wide variety of terrestrial mammals in camera‐trap surveys.
|
|---|
