Using Citizen Science to Map Human-Coyote Interaction in Suburban New York, USA

The expansion of coyotes (Canis latrans) into the northeastern United States is a major challenge to wildlife professionals, especially in suburban and urban areas where reports of human coyote interaction (HCI) are on the rise. To assist wildlife professionals in identifying potential hot spots of interaction and homeowners in evaluating their risk of a backyard encounter, we used the techniques of citizen science to build a landscape model of HCI for suburban residential properties in Westchester County, New York, USA. We distributed surveys via school children (kindergarten to grade 12) as part of a voluntary class assignment, to maximize the number of homeowners participating in our study and to provide learning experiences for students. Of 6,000 surveys distributed to schools, >1,500 students interviewed their parents on whether a coyote had been seen or heard on their property from 2003 to 2006. Although surveys could not be distributed randomly owing to the participatory process of individual schools, we did receive responses from across Westchester County, representing the spectrum from the most rural to the most urban towns. Homeowners who encountered (i.e., seen or heard) a coyote on their property were on average 50% closer to forest, 36% closer to grassland, and 66% farther from medium- to high-intensity development, complementing existing knowledge on urban coyote habitat use. Our model seemed robust in predicting an independent set of coyote observations (r = 0.88). Based on this model, we generated a map describing the probability of HCI that can be used by both wildlife professionals and homeowners. Regarding the former, state wildlife agencies could more precisely target education campaigns on how to live with coyotes where the possibility of HCI was greatest. Homeowners, in turn, could evaluate their own risk and modify behaviors that would make their property less attractive to coyotes. Furthermore, in creating a descriptive model of HCI from citizen-generated data, we demonstrated how citizen science can be a useful exploratory tool, generating a wealth of data over a large geographic area in a short period, especially when the inquest is appropriate to stakeholder participation in data collection.