Translator Disclaimer
2 December 2020 Estimating bat fatality at a Texas wind energy facility: implications transcending the United States–Mexico border
Author Affiliations +

Wind energy development causes bat fatalities. Despite emphasis on understanding and reducing these impacts, few data are available for the southwest region of the United States and northern Mexico. We monitored bat fatalities for a full year (March 2017–March 2018) at a wind energy facility in south Texas near the United States–Mexico border. We established search plots of 100-m radius at eight randomly selected turbines (of 255) and searched the roads and pads at an additional 92 turbines. We conducted weekly searches from spring through fall and bimonthly during winter. We used GenEst (Generalized Mortality Estimator) to estimate bat fatalities corrected for searcher efficiency, carcass removal, and density-weighted proportion of area searched. We found 205 bats during standardized searches, the majority of which were Brazilian free-tailed bats (Tadarida brasiliensis, 76%). The corrected fatality estimates were 16 bats/megawatt/year (95% confidence interval [CI]: 12 – 30 bats/megawatt/year) across all species. Species composition at our site is similar to that of northern Mexico, an area of expanding wind energy development with no published studies.

Impacts from burning fossil fuels, including climate change (Schlesinger and Mitchell 1987; Inkley et al. 2004), are stimulating development of renewable energy alternatives (Bernstein et al. 2006; Baños et al. 2011; Intergovernmental Panel on Climate Change 2011; Manzano-Agugliaro et al. 2013). In response, wind energy development is expanding around the globe (Bernstein et al. 2006). Wind turbines can, however, affect wildlife adversely, particularly bats (Hall and Richards 1972; Arnett and Baerwald 2013). Wind turbines currently are one of the largest sources of mass bat mortality worldwide (O'Shea et al. 2016). Although detailed information on population sizes is lacking for most bat species, bats are long-lived and slow to reproduce (Austad and Fischer 1991; Barclay and Harder 2003). Thus, populations may be unable to recover from large-scale sustained fatalities, such as those caused by to wind turbines (Frick et al. 2017).

These concerns have stimulated enumeration of bat fatality rates related to wind turbines in North America at varying temporal and spatial scales (Kunz et al. 2007; Arnett et al. 2008; Arnett and Baerwald 2013; Hayes 2013). Arnett and Baerwald (2013) estimated that cumulative bat fatalities ranged from approximately 650,000 to 1.3 million in the United States and Canada between 2000 and 2011, with an additional 196,000 to 396,000 estimated bat fatalities in 2012. Nonetheless, estimates of cumulative bat fatality rates based on currently available studies are unlikely to be accurate because methodologies and estimators varied among studies and the data used were not representative of all regions. Some areas with extensive wind energy development remain data-deficient in that respect (Arnett et al. 2008; Huso and Dalthorp 2014).

Of particular concern is the lack of data from the southwest United States, especially Texas. Texas leads the United States in installed wind energy capacity (Smallwood 2013; Huso and Dalthorp 2014a), has the highest bat diversity in the United States (Ammerman et al. 2012), yet has few publicly available or reliable estimates of wind energy impacts on bat populations (Smallwood 2013). Because species composition (Hall 1981) and fatality patterns are likely to vary strongly among regions, cumulative continental impacts and continental patterns of fatality cannot be assessed and appropriate regional strategies to reduce impact cannot be selected or implemented (Hein and Schirmacher 2016). Here we provide estimates of bat fatality from a wind energy facility in south Texas and provide additional insights on potential impacts of wind turbines on bats in northern Mexico, a region with similar species composition as Texas (Ammerman et al. 2012; Ceballos 2014).

Materials and Methods

Study site.—We monitored bat fatalities at the Los Vientos III (26°28′56″N, 98°35′59″W, elevation 135 m), IV (26°32′46″N, 98°41′24″W, elevation 155 m), and V (26°23′8″N, 98°34′35″W, elevation 98 m), wind energy facilities near Rio Grande City, Starr County, Texas (Fig. 1). The site is located in the Texas-Tamaulipan Thornscrub Level IV ecoregion (Griffith et al. 2007) and, according to the 2011 National Land Cover Data, the habitat is comprised primarily of scrub/shrub (39%), cultivated crops (25%), pasture/hay (24%), developed open space (5%), grassland/herbaceous (5%), and developed, low intensity (2%) lands (Homer et al. 2015).

The three facilities are adjacent to one another and are indistinguishable at a landscape level, with the same turbine model and single operations and maintenance facility, and thus were treated as a single facility for this study (hereafter Los Vientos). Los Vientos encompasses approximately 227 km2 and consists of 255 Vestas V-110, 2-megawatt turbines. All turbines had a nacelle height of 95 m and rotor diameter of 110 m, with a rotor-swept area of 9,503 m2. Commercial operations began in April 2015 for some turbines and commenced for all turbines in August 2016.

Fig. 1.

The Los Vientos wind energy facility in Starr County, Texas. Point locations on the map show the spatial distribution of all searched turbines during post-construction bat fatality monitoring from 24 March 2017, through 23 March 2018. Triangles represent turbines with full plots up to a maximum 100-m radius (2,054 m2) and circles represent turbines at which only roads and pads were searched to a maximum 100-m radius. Gray lines represent political boundaries.


Carcass monitoring and correcting fatality estimates.—We randomly selected 100 of the 255 turbines for fatality searches. At a subset of eight randomly selected turbines, we established circular plots, centered on the turbine with an approximate search radius of 100 m. Some plots contained obstacles such as trees, leading plot radii to be as small as 60 m. We established parallel linear transects across each plot spaced at 5 m. Transects ranged in length from 200 to 62 m. We walked all transects and searched 2.5 m on both sides