Abstract
Bat fatalities have been reported at wind facilities worldwide and expansion of wind energy raises concern over potential cumulative impacts on bat populations, particularly when several species are known or suspected to be in decline. Strategies to avoid or minimize adverse impacts of wind development on bats begins with assessing risk prior to construction of a site, but once a facility is built and operational, mitigation options are currently limited to curtailment of operations during predictable high risk periods when the greatest number of fatalities occur. State and federal agencies, natural resource managers and wind energy developers may benefit from pre-construction assessments of bat presence and activity, but ideally these assessments are most valuable if they also can be used to accurately predict risk of fatalities.
To establish whether pre-construction bat activity relates to post-construction fatality, we synthesized available data from 94 pre-construction bat activity and 75 post-construction bat fatality studies at wind energy facilities across 4 regions in the U.S. and Canada. We present summaries by broad geographic regions of bat activity, measured by acoustic detectors, and fatality rates measured from carcass searches. We also related pre-construction activity and post-construction fatality rates from 12 sites with adequate (i.e., both pre-construction acoustic and post-construction fatality) data to determine whether bat acoustic data gathered prior to construction can be used to predict fatality.
Pre- and post-construction data varied considerably both within and among regions, with the exception of the Great Plains, which showed both low and relatively precise means for activity and fatality. Bat fatality in the Basin-Desert was, on average, the lowest and most precise among, but differed substantially from the higher more variable activity in the Region. We observed a lower mean activity level in the Midwest Region compared to fatality. Although the Midwest had a slightly greater mean fatality rate than the Eastern Forest Region, both regions showed similar high and variable fatality estimates compared to the other regions. Activity in the Eastern Forest Region was, on average, the highest, and the most variable among the regions.
Based on 12 sites with paired data, the regression line was not significantly different from zero, but did suggest a positive relationship (F1,10 = 4.06; p = 0.072). Only a small portion of the variation in fatalities was explained by activity (adj. R2 = 21.8%). The 95% prediction intervals for this relationship indicate that, given current available data, acoustic data gathered prior to construction cannot accurately predict bat fatality.
Currently, it remains unclear whether pre-construction acoustic data are able to adequately predict post-construction fatality. Given the number of facilities in operation across the U.S. and Canada within the 4 regions presented here, presumably more data exist, but are as of yet unavailable, to aid in this assessment. Considerable resources are being expended in studying bat activity and fatality at wind energy facilities, yet our ability to assess risk of a proposed site remains tenuous, at least based on data available for this synthesis. Regardless, acoustic surveys still provide valuable data useful for understanding the timing and conditions under which bats are more or less active at a site, particularly for regions in which wind development is relatively new. Modeling bat activity or species presence using acoustic detectors as a function of time (i.e., night, season, or year) and meteorological conditions can provide powerful insight to predict when bats are most at risk, and which strategies are best suited to minimize fatalities while maximizing power production.