Spatial and Temporal Distributions of Moose-Vehicle Collisions in Newfoundland

In Newfoundland, 5,422 moose-vehicle collisions (MVCs) occurred between 1988 and 1994, resulting in 14 human and approximately 4,800 moose (Alces alces) fatalities. We examined daily, seasonal, and spatial distribution patterns of MVCs and used log-linear modeling to assess effects of darkness, posted speed limits, road condition, vehicle occupants, and road alignment on severity of human injury resulting from such collisions. Seventy-five percent of all MVCs occurred between dusk and dawn. We found no differences among diurnal patterns of MVCs involving moose of different sex and age groups. Seasonally, 70°/O of MVCs occurred between June and October and peak accident periods differed somewhat among moose age classes (calf, yearling, adult). However, we found no seasonal differences between males and females. Spatially, MVCs were dependenton moose densities and traffic volume, with greater probability of MVCs in areas of high or low (but not moderate) moose densities and high traffic flow. Risk of severe human injury was 2.0 times greater at highway speeds ranging from 80 to 100 km/hour and 2.1 times greater at night. Human injuries were more severe for MVCs occurring when road conditions were dry and when passengers were present. We found no relationship between road alignment and injury severity, although 79% of MVCs occurred on straight sections of highway. The predominant influence of darkness on MVCs may result from low light conditions and increased movement of moose at night. The high occurrence of MVCs under dry road conditions and on straight sections of road was surprising and may indicate inadequate driver attention to potential road hazards. We suggest that a long-term driver education program may be the only viable mitigation effort available to reduce number of MVCs. This content downloaded from 157.55.39.175 on Sun, 02 Oct 2016 05:55:23 UTC All use subject to http://about.jstor.org/terms 282 Wikilife Society BglleEn 2001, 29(1):281-291 MVCs and this did not include costs for operative treatment or follow-up visits. Additionally, Oosenbrug et al. (1991) estimated annual losses of $ 1,000,000 (Canadian) in vehicle damage; $600,000 in moose meat; and approximately $200,000 in wildlife outfitting, adventure tourism, and other local industries in Newfoundland. Moose-vehicle collisions are not random events in either time or place. Peak accident periods have been identified for most areas where MVCs occur, although seasonal patterns are not consistent among regions. In Newfoundland, MVCs occur throughout the year with the most occurring during summer (Oosenbrug 1986). In southern Sweden (Lavsund and Sandegren 1991) and Maine (K. Morris, Maine Department of Inland Fisheries and Wildlife, personal communication) the greatest number of MVCs occurs during periods of seasonal transition (spring, fall), when animals are more vagile;Alaska (Del Frate and Spraker 1991), British Columbia (Child et al.1991), and northern Sweden (Lavsund and Sandegren 1991) report most MVCs in winter. Despite this seasonal dissimilarity, diurnal patterns are consistent. Moose-vehicle collisions are more likely to occur during dusk and dawn and at night than during the day for all areas (Sweden [Almkvist et al.1980, Lavsund and Sandegren 1991], Newfoundland [Oosenbrug 1986], British Columbia [Child et al.1991]). Moose-vehicle collisions are clustered spatially (Almkvist et al. 1980, Damas and Smith 1982, Poll 1989, Child et al.1991). Sites of spatial aggregations have been targeted for a variety of mitigative measures, such as highway reflectors (Schafer and Penland 1985, Ingebrigtsen and Ludwig 1986,Waring et al.1991), animal reflectors (Farrell et al.1996), artificial lights (Pojar et al.1975, Reed and Woodward 1981), fences (Falk et al. 1978, McDonald 1991, Brown and Ross 1994), and predator scents aolicocur and LeMay 1993; W. Bradford, Parks Canada, personal communication;T. Joyce and S. Mahoney, Newfoundland and Labrador Department of Forest Resources and Agrifoods [NLFRA], unpublished data), but usually with little success and often without examining the possible factors contributing to these accidents. The only thorough published examinations of accident site characteristics were by Bashore et al. (1985) in Pennsylvania and Poll (1989) in Kootenay National Park, where they examined numerous habitat and highway variables thought to influence the occurrence of deer-vehicle collisions. We describe the temporal and spatial clustering of MVCs in Newfoundland from 1988 to 1994 and relate the rate and severity of human injury to time of accident, road conditions, road alignment, vehicle speed (via posted speed limits), number of vehicle occupants, and the sex and age of moose struck. Our objective was to develop measures to reduce number of MVCs and severity of human injuries.