Improving inferences from coyote (Canis latrans) surveys across Illinois using integrated state-space models

Abstract Anthropogenic influences on landscapes changed wildlife communities immensely over the last century. One notable example of human-derived change is the expansion of coyote ( Canis latrans ) populations following the extirpation of apex carnivores and the inception of industrial agriculture in North America. In Illinois, USA, the Department of Natural Resources estimates relative abundance of coyotes using a variety of survey types, including harvest, participatory science, and observational measures. To examine how different survey methods tracked long-term changes in the state coyote population, we used four datasets of coyote relative abundance in state-space models, deriving estimates of abundance and population growth rates from 1980 to 2023. We compared estimates from individual datasets, as well as estimates from one model that integrated all datasets concurrently. Of the four datasets, only the spotlight survey indicated long-term positive population growth of coyotes, despite the spotlight survey also having the largest observational error. The integrated model showed long-term positive population growth and had high precision compared to the individual models. Individual counties showed high variation, illustrating declines and increases during different decades – fine-scale patterns likely reflect differences in human-caused mortality, prey availability, and habitat quality over time. We support the use of integrated models to combine inferences from relative abundance data, and we encourage the examination of observational biases that may overshadow long-term population changes. Further studies of fine-scale demographic change in coyote populations are needed.