Tree Selection and Landscape Analysis of Eastern Red Bat Day Roosts

Authors

  • DANA L. LIMPERT,

    Corresponding author
    1. Maryland Cooperative Fish and Wildlife Research Unit, University of Maryland Eastern Shore, 1120 Trigg Hall, Princess Anne, MD 21853, USA
    Search for more papers by this author
    • Maryland Department of Natural Resources, 909 Wye Mills Road, Wye Mills, MD 21679, USA

  • DIXIE L. BIRCH,

    1. Maryland Cooperative Fish and Wildlife Research Unit, University of Maryland Eastern Shore, 1120 Trigg Hall, Princess Anne, MD 21853, USA
    Search for more papers by this author
    • Blackwater National Wildlife Refuge, 2145 Key Wallace Drive, Cambridge, MD 21613, USA

  • MICHAEL S. SCOTT,

    1. Salisbury University, 1101 Camden Avenue, Salisbury, MD 21801, USA
    Search for more papers by this author
  • MELISSA ANDRE,

    1. Maryland Cooperative Fish and Wildlife Research Unit, University of Maryland Eastern Shore, 1120 Trigg Hall, Princess Anne, MD 21853, USA
    Search for more papers by this author
    • Colorado State University, Fort Collins, CO 80526, USA

  • ERIN GILLAM

    1. Maryland Cooperative Fish and Wildlife Research Unit, University of Maryland Eastern Shore, 1120 Trigg Hall, Princess Anne, MD 21853, USA
    Search for more papers by this author
    • University of Tennessee, Knoxville, TN 37996, USA


E-mail: dlimpert@dnr.state.md.us

Abstract

ABSTRACT Declining bat populations and increasing demands on forest resources have prompted researchers to investigate tree roost selection of forest bats. Few studies, however, have investigated different spatial scales and landscape pattern as criteria for selection of tree roosts. In 1999 and 2000, we radiotracked 23 eastern red bats (Lasiurus borealis) to 64 day roosts. Using univariate and multivariate comparisons, we tested roost tree variables with random tree data at 3 circular spatial scales: roost tree, plot, and landscape. We found 15 variables that were entered in a stepwise discriminant analysis to best differentiate between the roost and random samples; 11 (73.3%) were landscape variables measured with a geographic information system. On average (x̄ ± SE), red bats roosted in deciduous trees (42.0 ± 2.1 cm dbh) that were located in plots with more (3.1 ± 0.1 m2) basal area, higher (84.0 ± 1.3) percentage of canopy closure, and lower (27.2 ± 2.2) percentage of groundcover than random plots. At the landscape scale (by percent magnitude), red bat buffers (1,000-m-radius circle) had significantly less development (81.6%), less feeding operations (70.4%), more deciduous (52.9%) and pine forest (63.8%), and fewer local roads (5.4%) but more trails (94.1%), open water (61.4%), wetland areas (80.4%), and stream areas (63.1%) than random buffers. Red bat roost trees were significantly closer (χ2 = 22.0088, df = 1, P < 0.001) to trails (106.2 ± 13.3 m) than to streams (279.4 ± 28.5 m). Our results suggest that red bats in our study area select roosts in mature riparian forests near trails, open water, and wetlands. The high percentage of landscape values in the discriminant analysis lends support to using landscape metrics as an investigative technique of resource selection. We recommend that managers consider landscape factors when protecting red bat day-roost habitat.

Ancillary