부산대학교 도서관

Camera traps are an increasingly popular means to monitor wildlife populations. However, like other techniques for measuring populations, camera traps are subject to sources of error that may bias population estimates. Past studies accounting for detection error have failed to account for a simple but potentially widely pervasive source of environmental error: weather conditions. Using 5,108,416 photographs from 804 scent‐lured camera traps deployed in western Nebraska, USA, during spring and autumn of 2014 and 2015, we analyzed the relationship between weather conditions (barometric pressure, wind speed, precipitation, and temperature) and coyote (Canis latrans) detection probability. Using binomial generalized linear mixed‐effects models, we showed that detection probability was affected by all weather conditions examined. Weather effects on detection suggests that either weather alters coyote behavior or decreases trap efficacy. Detection probability also decreased over the exposure period, indicating that coyotes either avoided traps after initial exploration or that lure efficacy decreased over time. Our findings suggest that to achieve accurate population indices, camera‐trap studies need to incorporate effects of weather conditions and sampling duration into population models to account for detection bias in estimates. © 2020 The Wildlife Society. Weather conditions and exposure period of camera traps affected the detection probabilities of coyotes in western Nebraska, USA, potentially by altering their activity and movement patterns across the landscape. Our results suggest that detection errors extend beyond the functionality of equipment (here, cameras) and surveys need to account for the detection errors incurred by the behavior of target species to improve the accuracy of population estimates and subsequently management and conservation plans. [ABSTRACT FROM AUTHOR]