Bryan Chruszcz

Spatial patterns and factors influencing small vertebrate fauna road-kill aggregations

We examined the spatial patterns and factors influencing small terrestrial vertebrate road-kill aggregations in the Bow River Valley, Alberta, Canada. We surveyed roads varying in traffic volume, configuration and adjacent landscape attributes for road-kills between 1997 and 2000. The spatial pattern of road-kills was described using neighbour K statistics. We investigated the importance of road-kills at three taxonomic levels using logistic regression. Mammal and bird road-kill indices were consistently higher on a low volume parkway than on the high-speed, high volume Trans-Canada highway (TCH). Birds were more vulnerable to collisions than mammals on the TCH. Road-kill aggregations were nonrandomly distributed. Parkway road-kills were aggregated on small scales and characterized by low clustering intensities compared to the TCH. Road-kills were less likely to occur on raised sections of road. Road-kills tended to occur close to vegetative cover and far from wildlife passages or culverts. Our findings reveal how two distinct road types can have different effects in terms of vertebrate mortality and their spatial pattern. We recommend a series of mitigation measures for existing roads or future road planning projects.

Uncontrolled field performance of Televilt GPS-Simplex™ collars on grizzly bears in western and northern Canada

Abstract Commercially available telemetry collars for wildlife that employ Global Positioning System (GPS) devices generally have the ability to gather a large volume of precise location data. We appraised the performance of 12-channel Televilt GPS-Simplex™ collars (Televilt/TVP Positioning AB, Lindesberg, Sweden) deployed across western and northern Canada on grizzly bears (Ursus arctos). Of 71 collar deployments between 2000 and 2002, 38 were retrieved and performed according to their programmed schedule, 20 were retrieved and had some degree of failure, and 13 experienced catastrophic failures and were not retrieved. In addition to these deployments, 10 collars failed predeployment. GPS collar fix success rates were greater for the retrieved collars from the Northwest Territories than for the 4 study areas in British Columbia and Alberta (F4,50 = 10.82, P < 0.001); thus, the latter areas were grouped for further analyses. Collar fix success rates in the British Columbia and Alberta study areas differed between the retrieved collars that functioned normally (x̄ = 65%, SE = 2.3, n = 28) and collars retrieved with failure events (x̄ = 56%, SE = 4.3, n = 17; t43 = 2.09, P = 0.043). Fix success rates were lower the longer collars were in the field (rs = −0.35, n = 45, P = 0.020). Locations from the GPS collars had a mean dilution of precision of <4 for 2D and 3D locations and thus had a good degree of precision. We were satisfied with the volume and quality of the location data; however, we advise other researchers that significant time and money may be lost troubleshooting problems with the Televilt Simplex system. Other recommendations for future and current users are considered.