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.

中文翻译：

---

天气和暴露时间会影响相机陷阱中的土狼检测

相机陷阱是监视野生动植物种群的一种越来越流行的手段。但是，像其他用于测量人口的技术一样，相机陷阱也容易受到误差源的影响，而误差可能会使人口估计值产生偏差。过去关于检测错误的研究未能说明一个简单但潜在的普遍性环境错误来源：天气状况。利用2014年春季和秋季以及2015年春季和秋季在美国内布拉斯加州西部部署的804个带有气味的照相机陷阱中的5,108,416张照片，我们分析了天气条件（气压，风速，降水和温度）与土狼（Canis latrans）的检测概率。使用二项式广义线性混合效应模型，我们表明检测概率受所检查的所有天气条件的影响。天气对检测的影响表明，天气会改变土狼的行为或降低诱捕器的功效。在暴露期间，检测概率也有所降低，这表明土狼在初次探索后避免了陷阱，或者诱饵效力随时间降低。我们的发现表明，要获得准确的人口指数，相机诱集研究需要将天气条件和采样持续时间的影响纳入人口模型中，以解决估计中的检测偏差。©2020野生动物协会。