Where does a deer cross a road? : road and landcover characteristics affecting deer crossing and mortality across the U.S. 93 corridor on the Flathead Indian Reservation, Montana

Loading...
Thumbnail Image

Date

2007

Journal Title

Journal ISSN

Volume Title

Publisher

Montana State University - Bozeman, College of Letters & Science

Abstract

Reducing deer vehicle collisions (DVCs) on highways is an issue facing highway planners and wildlife managers. In western Montana, federal, state and tribal governments intend to reduce DVCs along a 51-mile stretch of US Highway 93 (US 93) on the Flathead Indian Reservation using on site reconstruction and highway engineering. This project was part of pre-construction wildlife monitoring that forms the baseline for evaluating effectiveness of mitigation measures associated with the US 93 reconstruction project. Road and landcover variables were recorded at randomly sampled locations along US 93. DVC site selections were made based on Montana Department of Transportation maintenance and Montana Highway Patrol accident reports from 1998- 2003. Observed crossing areas were based on 32 sand tracking beds (each 100 m long) placed randomly in three key areas along the route. Variables collected included habitats types, topography, and rural residential developments and anthropogenic effects. A geographic information system was used to determine proportion of landcover variables at three spatial scales centered on the highway, encompassing 0.16 km (0.5 mi), 0.32 km (1 mi), 0.64 km (2 mi).
Local scale habitat characteristics were collected in the field including presence of specific variables directly adjacent and up to 100 meters from the pavement edge. A priori models were developed to determine useful variables for predicting deer crossing and collision locations. Akaike's Information Criterion for small sample size (AICc) was used to rank best overall models and determine variables with a greater ability to predict crossing or collision occurrence. The results showed that landcover variables could be used to predict crossing or kill location. Top predictors included a positive correlation to forest cover, distance to the nearest city, and low intensity residential development. Negative correlations were found for distance to nearest water and population density. Results of this project will be used for comparison to post-construction movement patterns. This project is unique in that data are available from another related concurrent project providing OCA data, which allows us to look at both observable crossing rates as well as collision rates, in relation to landscape and road variables.

Description

Keywords

Citation

Copyright (c) 2002-2022, LYRASIS. All rights reserved.