Automated Analysis of Wildlife-Vehicle Conflict Hotspots Using Carcass and Collision Data

States increasingly maintain databases of wildlife-vehicle conflict (WVC), including locations of carcasses and crashes involving animals. Once these data are collected, a common and expensive barrier before they can be used in safety and environmental planning is identification of “hotspots” of incidents (here defined as locations of high-rates and/or statistically-significant clusters). In this project, we developed a web-based analytical environment that state DOTs can use to automate certain analyses of WVC hotspots in order to inform planning to improve driver and wildlife safety. Specifically, we coordinated with staff from 3 states: Idaho, Nevada, and Maine, and used data that we had for California to develop the pilot automated hotspots analysis tool. We tested several methods for representing both density and statistically-significant clustering of WVC events. These were implemented in the statistical package R and driven by scripts that can operate in the project web-system. Testing was conducted using California data while other states prepared/delivered their data. We developed hotspots analyses for California, Maine, Idaho, and Nevada based upon their input. We prepared a web-system to operate the tool (https://roadecology.ucdavis.edu/hotspots), that DOT staff from 13 states used so far with their own data.