Abstract
Addressing risk to potentially vulnerable bat species from wind energy development has been identified as a high priority by the wind industry and other conservation stakeholders. Improving understanding of the large-scale ecological and geographical factors associated with potential collision risk is particularly important as many decisions regarding risk avoidance are made during the siting and design phases of a wind energy facility, and the identification of potentially higher or lower risk areas for bats can inform these early-stage decision processes as well as provide guidance for state or regional-level planning by agencies or conservation organizations. However, few studies have been conducted to explore the relationships among landscape-level factors and risk to bats from wind energy operations. This study examined fatality rates of three species of migratory tree-roosting bats commonly observed as fatalities at operational wind energy facilities - hoary bat (Lasiurus cinereus), silver-haired bat (Lasionycteris noctivagans) and eastern red bat (Lasiurus borealis) - in relation to landscape-scale features at varying scales in the midwestern and northeastern regions of the United States. A multistage process including ensemble learning (random forests) and predictive modeling (generalized linear models) was used to explore associations between bat fatality rates, based on data collected during post-construction fatality monitoring studies at individual wind energy facilities throughout the two regions, and various landscape metrics calculated at the local, 2.5-kilometer (km), 5-km and 25-km, scales. Findings indicated that landscape structure at the broadest scale examined was most strongly associated with fatality rates, and revealed both similarities and differences between the two regions. In the midwestern region, a positive association between fatality rates and the proportion of developed land occurring within 25-km facility buffers was observed for all three of the target species, a pattern that was also observed for hoary bat and silver-haired bat in the northeastern region. In the midwestern region a negative relationship with road density was also observed at the 25-km scale whereas at the turbine area (i.e., local, facility-level) scale fatality rates of the three target species tended to increase with road density. Hoary and eastern red bat fatality rates were also higher in the midwest region when small disaggregated patches of open, noncultivated habitat as opposed to clumped, larger patches occurred within and adjacent to facilities. Finally, silver-haired bats were observed as fatalities at higher rates in the midwestern region when more turbines occurred on the broader landscape. Wetland structure was also associated with fatality rates in both regions. For example, in the northeastern region, fatality rates for hoary, and eastern red bats were highest when facilities were located in landscapes characterized by wetland complexes comprising large and small wetland patches. The landscape patterns revealed in this study and others can better inform future research and siting decisions and feed into an adaptive learning process that will, over time, reduce uncertainty and lead to an improved understanding of factors associated with bat collision risk at wind facilities. It is anticipated that this enhanced understanding will further assist in the development of more accurate tools for assessing this risk and lead to the identification of scientifically-informed options for avoiding, minimizing, and mitigating risk to bats.
A two page Results Summary for this report is available here.