USING PUBLIC SIGHTING INFORMATION TO INVESTIGATE COYOTE USE OF URBAN HABITAT

Conventional approaches to studying large mammal habitat selection using radio telemetry may not be feasible in urban environments because of public opposition to trapping, risk of catching nontarget species, restrictions of observer movements, and limited spatial scales. Therefore, I evaluated use of public sightings for determining habitat selection of urban coyotes (Canis latrans) in Seattle, Washington. I used telephone interviews to document coyote sightings from 1988 to 1989. To examine biases in public sightings, I used radio telemetry to investigate habitat selection of coyotes in 1990. Both methods of data collection showed that coyotes were associated (P < 0.05) with forest habitat. Telemetry data revealed that coyotes moved shorter distances (P < 0.001) and remained closer (P = 0.004) to forest habitat during daylight than during night. Public sightings were farther from forest habitat (P < 0.05) than were telemetry data collected during daylight and, to a lesser degree, during night. Public sightings occurred mostly during daylight (93%, P < 0.001) and during summer (63%, P < 0.001), and were biased toward habitats where people were concentrated and coyotes easily seen. However, bootstrap sampling showed that the association between coyotes and forest habitat was unchanged (P greater-than-or-equal-to 0.08) by replacing 20% of actual public sightings with random locations, which suggested that small amounts of random error did not obscure the most important habitat association. In addition, the association between coyotes and forest habitat was the same (P > 0.8) for bootstrap samples of 20, 60, and 108 public sightings, which suggested that small datasets may reveal important habitat associations. Usefulness of public sighting information will depend on identifying sources of bias and research objectives.