Estimating bird/bat fatality rates and using stable-hydrogen isotopes to assess hoary bat migratory patterns associated with a wind-energy facility in western Oklahoma

Abstract

Offsetting a portion of the energy deficit of the United States with emission-free, renewable energy is driving a rapid increase in the number of wind-power installations across the country. Although wind energy has environmental benefits, the associated turbine-related fatalities have generated concern for the impact on bird and bat populations. Estimating the number of fatalities is key to determining the population impact, but fatality patterns routinely violate assumptions of the methods commonly used to estimate fatality rates, namely non-uniform death rates and/or searcher-induced perturbations on steady-state processes. Therefore, four new mathematical models were developed to reduce bias in fatality estimates, as were new methods using exponential-decay functions to estimate carcass-removal rates. Compared to the new models, earlier methods can result in estimated fatality-rate errors ranging from -100.0 to 65.7%. The degree of error is related to the carcass-removal rate, search interval, and carcass-detection probabilities. In general, error is likely to increase with carcass size. Estimates of bat-fatality rates typically will be less biased than those for large-sized birds. Examples illustrating the application of the models are given, as well as conditions in which each is appropriately applied.

Approximately one-third of all United States wind energy is generated in the south-central portion of the country, yet little effort has been devoted to determining fatality risk in this region. Bird and bat fatality rates were assessed for Blue Canyon II Wind Power Project (the study site), an array of 84 1.8-MW turbines in Caddo and Kiowa counties, Oklahoma. Weekly fatality searches were conducted over a period of 8 months (Mar.-Jun. and Aug.-Nov.) during 2006-2008 at 50 wind turbines. For the study site, we estimate 97.7 bird fatalities/year (1.16/turbine). Almost half of detected bird fatalities were of turkey vultures (Cathartes aura, 47.8%), with others involving 15-17 species. Bat fatalities involved 8 species, with hoary bats (Lasiurus cinereus, 60.8%) and eastern red bats (L. borealis, 21.3%) being the majority; fatalities of these species were male biased. Overall, bat fatalities rates were estimated to be 641.3 bats/year (7.63/turbine). Bird fatalities were evenly spread among survey months, with bat fatalities concentrated during late summer and fall migration (66% occurring within 3-week interval in late Aug. and early Sep.). Our derived mathematical models demonstrate that such migratory events exert little bias in fatality estimates. Bird fatalities at the study site were lower than found at most facilities nationwide and not likely to be biologically significant. Bat fatality rates were moderate and in the range found at other facilities, yet cumulative impacts are of concern.

While estimating the number of fatalities at individual wind-energy facilities is important, mitigation strategies are needed to reduce fatality risk. Hoary bats (Lasiurus cinereus) are among the most susceptible species and our limited knowledge of bat migration hinders our ability to (1) determine the reason for their susceptibility, (2) implement mitigation strategies, and (3) assess long-term population impacts. Bat carcasses collected from the study site were also used to assess migratory patterns of the hoary bat with stable-hydrogen isotopes, a technique that can connect hydrogen-isotope ratios of an animal's keratin-based tissues (i.e., fur and claws) to the geographic region in which it was grown. Thus, large-scale movements of individuals can be tracked to identify locations of summer and/or winter residency of migratory species. The reliability of stable-isotope methods is influenced by patterns and timing of tissue growth, which is unknown for many species. To circumvent potential bias of fur samples, techniques were developed to appropriately sample claw tissue to assess migratory patterns.

Claws are composed of multiple layers and our analysis indicates only unguis tissue should be used in isotope analysis. Otherwise, assignments of geographic locations will be biased to latitudes further north than where tissues were actually grown. Recently grown claw tissue of fall migrants can be used to infer location of summer residency and fall-migratory origins. The claw tip may be useful for inferring locations of winter residency, but further research is needed to confirm this relationship. Furthermore, congruency of δDfur (i.e., the hydrogen-isotope ratio of fur) and δDclaw (i.e., the hydrogen-isotope ratio of claws) in female hoary bats suggests molt occurs at location of summer residency. However, in male hoary bats, δDclaw was significantly more depleted than δDfur, indicating males do not molt at location of summer residency; thus, claws may be more reliable for assigning migratory origins.

Hoary bats have a wide-spread distribution across the entire contiguous United States; thus, it is unknown from where bats that are killed at wind-energy installations may have migrated. Using claw keratin, hydrogen-isotope ratios revealed all female and 87.8% of male hoary bats recovered at the wind farm were not summer residents, but migrants from other geographic locations. The represented locations of summer residency were from a broad range of latitudes extending from Oklahoma into northern Canada (δDclaw from -43.6 / to -153.0 /). The change in δD from the claw root to the claw tip within the same individual appears to be an indirect measure of migratory distance, as individuals from more northern localities had the greatest change in δD. Individuals with summer residency at greater latitudes migrate greater distances than those from lower latitudes, suggesting the species exhibits a differential migration. These differences were not explained by sex as there was no difference in latitudinal distributions of males and females. There was also a significant relationship between the arrival date at the study site and the geographic region from which individuals likely migrated. Individuals with summer residency in greater latitudes arrive later than individuals from lower latitudes.

While the cumulative, nation-wide impact of wind energy is of concern, our studies highlight characteristics that are at least best-case scenarios for minimizing population impacts. Unlike some bird fatalities, bat fatalities are not likely to suffer from substantially underestimated fatality rates and the biased removal of males will have less impact on the population's reproductive potential. Furthermore, individuals killed are representatives of many geographic regions, which likely disperse the risk among many populations and may minimize localized population reductions. We highlight the use of wind farm fatalities and claw keratin to address migratory patterns that, in turn, may provide insights to the development of mitigation strategies.