Crossing Paths: Gray Wolves and Highways in the Minnesota-Wisconsin Border Region
File(s)
Date
1999-12Author
Frair, Jacqueline L.
Publisher
College of Natural Resources, University of Wisconsin-Stevens Point
Metadata
Show full item recordAbstract
Highways may threaten wolf (Canis lupus) habitat connectivity in the Great Lakes Region. Highways may further define the useable amount and arrangement of habitat within areas identified as suitable habitat. The two papers of this thesis explored wolf habitat with respect to roads by: (1) inferring wolf habitat connectivity across a major highway from predictions of crossing habitat, and (2) identifying how highways influenced suitable habitat choices and within-territory movements. Track searches, radio-telemetry, and observations identified 62 crossings of 4 major highways in northwestern Wisconsin between March 1991 and February 1999. Selected crossings (n=33) were visited to identify local crossing site characteristics. Compared to paired random sites wolves were equally likely to cross through forested and open landscapes (p = 0.192). In their immediate crossing environment, wolves favored factors related to visibility and ease of movement - less visual obscurity (p = 0.003) and less deciduous canopy cover (p = 0.038). Among the variables not significantly correlated with crossing sites were margin slope, maximum distance visible, shrub cover, and the presence/absence of trails or fences nearby. Satellite imagery data were used to map habitat composition and pattern within 25-400 ha of crossing sites. Patch density, an index to human-induced fragmentation, was the most significant and consistent indicator of crossing habitat across 5 landscape sampling resolutions (p < 0.005). Crossing landscapes had fewer patches, less open water, less developed land, and more forested or unforested wetlands than the available landscape matrix. Generally predictive ability increased with increased sampling area indicating that wolves perceived and reacted to landscape pattern at larger-spatial scales. Some differences between highways were noted, e.g., traffic volume and resident wolf activity in neighboring habitat, but the model captured elements of habitat selection specific to individual highways and common to all. Disperser and resident wolf crossings were equally represented by the model. Along U.S.H. 53 (a recently widened highway which bisects the primary wolf dispersal corridor in WI) the model mapped 68% of the road-adjacent habitat to have moderate-high crossing potential indicating a high degree of habitat connectivity. Highway crossing mitigations, e.g., widened medians and a highway underpass, were located where crossing potential was at least moderate. Only 20% (14 km) of the highway had high crossing potential of which 2 areas did not have any structural mitigations to aid crossings. The model was proved valid by 19 crossings reserved from analyses. Highway design may prove critical for continued connectivity as recreational and commercial use of the landscape increases in combination with increased traffic volume and vehicle speeds. Along the Minnesota-Wisconsin border, individual wolves from 4-12 wolf packs were radio-monitored between 1992-96. Resident wolf locations (n=3,448) were compared to unused locations (n=3,535)- outside of known wolf territories -with respect to the amount of major highways, minor highways, and non-highway public roads found within 200-ha of each location. Univariate logistic regression models indicated that highways did not strongly influence suitable habitat choices (less than 22% accuracy for unused habitats) but the density of non-highway public roads did (74% overall accuracy). Overall road density was still the best predictor of suitable habitat (77% overall accuracy). No improvement in classification accuracy was achieved by segregating highways from non-highway public roads. Males and pups demonstrated less tolerance for roads than females and older wolves (p < 0.001 ). All wolves combined demonstrated less tolerance for roads during the breeding and nomadic months (p < 0.001). These results indicated that not all habitats were equally available to wolves at all times - a critical consideration when estimating the total amount of useable wolf habitat in a given area. Sequential radio-locations (12-48 hours apart) were connected to investigate how often wolves crossed roads within their annual home ranges. Although highways were not a major factor in population-level habitat selection they strongly influenced within-territory movements (p < 0.001) which in tum affects the geographic arrangement of wolf territories in the landscape. All territories contained some nonhighway public roads, and wolves were either indifferent or attracted towards such roads. In contrast only one-half of the territories contained a major highway. Minor highways were strongly avoided in regular movements (p < 0.00 I). Potential highway tolerance limits for this population were identified as 0.09 km/km2 of major highways and 0.15 km/km2 of minor highways within territories. Considering the amount and arrangement of highways in the landscape could further help to define the subset of "suitable" wolf habitat (based on overall road densities) which is truly "useable" habitat to the population.
Subject
gray wolf
connectivity
habitat selection
highways
Minnesota
movements
predictive model
roads
Wisconsin