Academic journal article The American Midland Naturalist

Landscape Associations of Road-Killed Virginia Opossums (Didelphis Virginiana) in Central Massachusetts

Academic journal article The American Midland Naturalist

Landscape Associations of Road-Killed Virginia Opossums (Didelphis Virginiana) in Central Massachusetts

Article excerpt


Knowing the distribution of species at the landscape level can give insight into the proximate mechanisms determining the species' range on a regional scale. We used a survey of road-killed animals to investigate landscape features associated with the presence of Virginia opossums (Didelphis virginiana) in central Massachusetts. Volunteers noted road-killed opossums on their daily commutes through the Connecticut River Valley and surrounding towns in 2000 and 2002. We used a GIS to characterize both the locations of roadkills and random points according to elevation, land use, distance to open water, traffic speed and number of observers on the road and then used logistic regression to determine the association of roadkill sites with these variables. Dead opossums were found most often at low-elevation sites with less forest cover and more human development. Although the opossum usually is considered a habitat generalist found primarily in association with woodlands, opossums in central Massachusetts are not associated with woodlands and, instead, are most often found in and near urbanized areas.


Species' distributions often are mapped at a coarse resolution that does not consider local patterns of occurrence on the landscape. However, the distribution on the local landscape is particularly important for understanding the proximate interplay of factors that ultimately set the limits of a species' range. The Virginia opossum (Didelphis virginiana) is broadly distributed throughout southern New England and Ontario (Gardner and Sunquist, 2003; Fig. 1). However, modeling of the demography and winter physiology of Virginia opossums suggests that their populations should not persist north of central Massachusetts (Kanda and Fuller, 2004; Kanda, 2005a).

Anecdotal and camera survey evidence suggest that opossums do not occur uniformly throughout central Massachusetts. Opossums are commonly sighted in the developed regions of the Connecticut River Valley, but residents of the area believe opossums to be scarce in the primarily forested Pelham Hills that border the valley to the east (Fig. 2A). In addition, wildlife researchers trapping mesocarnivores in the forest around the Quabbin Reservoir east of the Pelham Hills did not capture opossums at the site (W. Healy, pers. comm.). A camera survey in conservation areas around the Connecticut River Valley found that opossums were associated with forest edge (generally within 300 m of a residence) rather than forest interior, and were not found above 120 m in elevation (M. Voight and T. Fuller, pers. obs.). Opossums also were found to be restricted to forest edges in a camera survey in the White Mountains of Vermont (Moruzzi et al., 2002; T. Moruzzi, pers. comm.). Based on these observations, we decided to identify landscape factors influencing the distribution of opossums in and near the Connecticut River Valley of central Massachusetts. To accomplish this goal, we indexed opossum abundance by identifying dead opossums on road transects throughout the valley and measured landscape variables associated with these sites.

Surveys of roadkills are a common tool for assessing the relative population size of medium-sized mammals. Counts often are used to index abundance (e.g., Eberhardt and Simmons, 1987; Rolley and Lehman, 1992; Gehrt, 2002) and its changes over time (e.g., case, 1978; Hicks, 1993; Gehrt et al., 2002). They also can be used to evaluate the landscape features associated with species' presence. A survey of mammal road carcasses in the Central Valley of California included the Virginia opossum, and indicated that opossum roadkills were equally likely in association with rural and urban development (Caro et al., 2000). In that study, Caro et al. (2000) used defined sections of road to evaluate a single predictor variable's effect on mammal abundance. Herein, we describe habitat use-availability analysis of multiple variables at roadkill sites compared with random sites along our transects (cf. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.