Modeling connectivity for forest rodents

Carol E Rizkalla, Purdue University

Abstract

In fragmented landscapes, the likelihood that a species occupies a particular habitat patch is thought to be a function of both patch area and patch isolation. Measuring connectivity, the inverse of isolation, remains a central problem in ecology. I used experimental translocations of white-footed mice ( Peromyscus leucopus), eastern chipmunks (Tamias striatus ), and southern flying squirrels (Glaucomys volans) to empirically evaluate the ability of forest rodents to move through agricultural fields in west-central Indiana. All species demonstrated a strong motivation to find the forest and were capable of moving through the matrix successfully, however, I was unable to determine whether they would do so willingly. Building on a similar study conducted on eastern gray squirrels (Sciurus carolinensis ), North American red squirrels (Tamiasciurus hudsconicus ), and fox squirrels (Sciurus niger), I then evaluated the ability of ecologically scaled landscape indices, which combine a species' ecological profile with indices of patch area and connectivity, to explain patterns in patch occupancy among 5 of the forest rodent species in 35 landscapes in the upper Wabash River basin. These indices were not consistently important within the guild, and none of the models demonstrated high predictive ability. Alternatively, individual-based spatially explicit population models demonstrated the utility of immigration into patches as a measure of connectivity in explaining rodent abundance. Finally, I assessed the effect of projected land-use in 2020 on patch occupancy and connectivity of forest rodents. Occupancy models had low predictive ability but suggested that gray squirrels are most susceptible to land-use change. Furthermore, declines in connectivity did not correspond with the decline of forest.

Degree

Ph.D.

Advisors

Swihart, Purdue University.

Subject Area

Wildlife Conservation|Ecology|Forestry

COinS