Assessment of Diurnal Wind Turbine Collision Risk for Grassland Birds on the Southern Great Plains

Wind energy is considered to be a positive contribution to wildlife conservation because of its potential to reduce the use of nonrenewable energy sources and associated environmental impacts (Pimentel et al. 2002; National Research Council 2007). However, there are concerns about potential direct and indirect impacts of wind energy facilities on wildlife populations, especially birds (Leddy et al. 1999; Osborn et al. 2000; Larsen and Clauson 2002; Pearce-Higgins et al. 2009). Direct impacts are primarily collision fatalities and indirect impacts include habitat loss, fragmentation and avoidance, and behavioral changes (Arnett et al. 2007; Kuvlesky et al. 2007; National Research Council 2007). Studies at wind energy facilities conducted postconstruction have found that wind turbines displace local avian communities, increase bird mortalities due to collision, and alter behaviors of birds (Barclay et al. 2007; Pearce-Higgins et al. 2009; Smallwood et al. 2009; Subramanian 2012).

Wind energy impact studies in the United States, primarily in California during the 1980s and 1990s, gained attention because of frequent collision and electrocution fatalities of raptors (National Research Council 2007; Smallwood and Karas 2009). At the time, little was known about the placement and design of wind energy facilities needed to reduce bird collisions and electrocution fatalities. Studies have since determined that turbines with lattice-style towers at high densities in areas with large avian populations lead to high rates of collision fatalities (National Research Council 2007; Smallwood and Karas 2009). However, some collision fatality remains a problem at wind energy facilities comprised of newer monopole towers at lower densities (National Research Council 2007; de Lucas et al. 2008; Marques et al. 2014). A major focus of most research on wind energy and birds has been on postconstruction collision fatalities. That information is useful for identifying bird species at greatest risk of mortality, but collision fatality observations can miss some species due to incomplete detection by observers and scavenger removal of carcasses (Smallwood 2007).

Preconstruction studies are useful for assessing bird collision risks at a site before placement of wind turbines. These assessments often examine species occurrence and density, animal movement through and within a site, and other animal behaviors (Arnett et al. 2007; Kuvlesky et al. 2007; Ferrer et al. 2012). However, the density of birds has not been found to predict their collision risks well (de Lucas et al. 2008; Ferrer et al. 2012). Instead, bird flight heights are indicative of collision risks (Barclay et al. 2007; Stantial and Cohen 2015), but knowledge of the distributions of bird flight heights is limited (Johnston et al. 2014). Radar works well for measuring bird flight heights, but it does not provide species-specific information (Gauthreaux and Livingston 2006; Fijn et al. 2015; Stantial and Cohen 2015). Species-specific flight heights have been visually binned into height classes, usually based on the rotor swept zone (RSZ) of wind turbines (i.e., below, within, or above the RSZ; Osborn et al. 1998; Hoover and Morrison 2005; Mabee et al. 2006; Johnston et al. 2014). Only recently have bird flight heights been measured using laser rangefinders (Stantial and Cohen 2015) or global positioning system tags with barometric pressure loggers (Cleasby et al. 2015).

Our study objective was to measure species-specific bird flight heights by using laser rangefinders to identify bird species that are at a greater risk of collision fatality due to the heights at which they fly. We determined their susceptibility to turbine blade collision by examining mean diurnal flight heights and the proportion of diurnal flight heights in the RSZ. We also compared seasonal flight heights to determine whether risk varies among seasons. These results, along with results from other bird studies, can be used to inform the placement of turbines at future wind energy facilities on the southern Great Plains and to help mitigate short-and long-term impacts on bird species.

We conducted research at two sites that were part of a planned wind energy facility in Gray and Donley counties, Texas (Figure 1). Both study sites were on the Llano Estacado Plateau or surrounding escarpments. Land use on the plateau was a mixture of irrigated cropland, cattle grazing, and oil and natural gas extraction (Smith 2003; Wulff 2010). Native land cover was primarily short-grass prairie and playa wetlands (i.e., shallow depressional recharge wetlands; Wulff 2010). The plateau is surrounded by relatively abrupt escarpments (hereafter breaks), ranging from 50 to 200 m in height, where primary land uses were cattle grazing and oil and natural gas extraction (Wulff 2010).

Research was conducted at the Gray County site from October 2008 to August 2009. The avian community was sampled on a 219-km² area during October 2008–February 2009 (Figure 1). We expanded the site to 303 km² during March–August 2009 because the planned wind energy facility increased in area (Figure 1). The uplands portion of the site (132 km² during October 2008–February 2009 and 170 km² during March–August 2009) was located on the plateau and consisted of mostly flat landscape that included irrigated cropland, pasture, playa wetlands, and Conservation Reserve Program land (i.e., environmentally sensitive private lands that are removed from agricultural production in exchange for annual rental payments; Haufler 2007) and other grasslands (Smith 2003). Common crops were corn, cotton, and winter wheat (Wulff 2010). The uplands portion of the Gray County site contained two cattle feedlots and a dairy operation. Some trees such as cottonwood Populus deltoides and elm Ulmus spp. were found around human structures.

The breaks habitat type (87 km² during October 2008–February 2009 and 133 km² during March–August 2009) was a broken landscape of gully washes and ravines, composed mostly of short-grass prairie. There were few water bodies limited to ephemeral creeks and water tanks for cattle. This area was also used for oil and natural gas extraction and has an extensive infrastructure of roads, oil wells, and other structures (Wulff 2010). Some trees, primarily cottonwood, were found within the breaks where deeper ravines held water. Prominent grasses included buffalo grass Buchloe dactyoids, blue grama Bouteloua gracilis, and other grama species Bouteloua spp. (National Resources Conservation Service 2006).

Data were gathered at the Donley County site (18.7 km²) during October 2008–February 2009. This site consisted of the breaks habitat type and was dominated by honey mesquite Prosopis glandulosa. Other trees or brush such as hackberry Celtis spp. and juniper Juniperus spp. occurred throughout the site on ridge tops and drainages, which were spring fed throughout the year. Primary grasses were buffalo grass and grama species (National Resources Conservation Service 2006). This site was used as rangeland with no oil or natural gas extraction.

We randomly selected 30 points spaced equal to or greater than 800 m apart (23 on the Gray County site and 7 on the Donley County site), and surveys were conducted from those points during October 2008–February 2009. For the expanded Gray County site, an additional 34 points (49 total points used; 8 of the original 23 points were removed due to land access issues) were randomly selected, and surveys were conducted from them during March–August 2009. Survey points were proportionally allocated over five cover types to ensure cover types were adequately represented in the sample (Wulff 2010): agriculture, breaks, plateau grasslands, playa wetlands, and prairie dog Cynomys ludovicianus towns. The plateau grassland cover type was broadly defined as grasslands located on the plateau that included Conservation Reserve Program, pasture, and other grasslands. Points were not placed within 400 m of cover edges to avoid overlap into other cover types. On the Gray County site, there were three highways (U.S. Highway 60, State Highway 152, and State Highway 273), and points were placed equal to or greater than 400 m from highways to avoid traffic noise. We used each random point (except the 10 points in playas or prairie dog towns due to the general size and shape of those features) as the start of an 800-m transect. Each transect was oriented along randomly selected compass bearings. Selected bearings were constrained so that transects remained within the study site and respective cover type and were spaced equal to or greater than 400 m apart.

Surveys occurred during four seasons with up to three samples per technique—point-count or line-transect—per season. Seasons were defined as winter (December–February), spring (March–May), summer (June–August), and autumn (September–November). We conducted both point-count and line-transect surveys from 0.25 h before sunrise until approximately 3 h after sunrise when diurnally active birds were most active and vocal (Diefenbach et al. 2003). We rotated the time of morning in which a point was monitored. Each point-count was conducted for 20 min, with surveys divided into two 10-min intervals. Surveys were not conducted in severe weather or if average wind speed was greater than 32 km/h (measured with a Kestrel 2000 pocket weather meter; Nielsen-Kellerman, Boothwyn, PA), because of reduced audibility and activity of birds (Diefenbach et al. 2003). Since no line-transects were established at survey points in playas and prairie dog towns, point-counts were conducted twice a month in those two cover types (Wulff 2010).

We measured distance (d) to the location on the ground below flying birds by using laser rangefinders (Monarch Gold Laser 1200; Nikon, Tokyo, Japan) and the angle of incline (θ) to the location of birds in the air by using clinometers (Suunto Clinometer, Vantaa, Finland). Measurements were recorded for the location where birds were first detected. We estimated flight height as tan θ·d plus the height of the observer. For groups of birds, measurements were recorded to the center of the group and a group was treated as a single sample. For our flight height analysis, we pooled flight height data between the two survey techniques and estimated range, mean, and standard deviation for each species by season. We used a SAS macro to conduct a power analysis to determine the appropriate sample size needed for two, three, and four treatments (i.e., seasons; Cary 1995; SAS Institute 2010). For this power analysis, we assumed effect size of 1.23, power (1 − β) of 0.80, and α = 0.05. For species in which the minimum sample was obtained, we compared seasonal means using an analysis of variance (ANOVA) or two-sample t-test (Dytham 2003). We conducted post hoc pairwise t-tests by using a Bonferroni adjustment (Dalgaard 2008). Seasonal data were pooled if sample size was inadequate.

Wind turbines planned for the study sites were designed to have a hub height of 78 m and a rotor diameter of 90 m (Jason Du Terroil, Iberdrola Renewables, Inc., personal communication). We used these measurements to characterize the RSZ. We added 2 m to the rotor diameter to define the RSZ (32–124 m) to allow for inaccuracies in flight height measurements. Bird species with mean flight heights within the RSZ were then identified as species of possible concern for blade collisions. These species were cross-referenced with Texas Parks and Wildlife Department (TPWD) species of greatest conservation need (Bender et al. 2005; TPWD 2012).

We also estimated the proportion of flight heights within the RSZ. We used χ² power analysis with an effect size of 0.5, power of 0.80, and α = 0.05 assumed to determine minimum sample size for two, three, and four treatments (i.e., seasons; Kabacoff 2008). For species in which the minimum sample was obtained, we estimated proportions on a seasonal basis and compared using a χ² test (Dalgaard 2008); otherwise, seasons were pooled. We considered species with greater than 25% of their flight heights within the RSZ as those at greatest risk of collision (exact binomial test). We performed statistical analyses in Program R (R Development Core Team 2012).

For 66 bird species, we obtained three or more observations of flight height (2,672 flight heights in total; Data S1, Supplemental Material). We obtained at least two observations of flight height for an additional 28 species (35 additional flight heights; Data S1, Supplemental Material). The species most commonly observed were red-winged blackbird Agelaius phoeniceus (n = 457), sandhill crane Grus canadensis (n = 278), mourning dove Zenaida macroura (n = 276), meadowlarks (both eastern meadowlark Sturnella magna and western meadowlark Sturnella neglecta; n = 240), horned lark Eremophila alpestris (n = 168), northern harrier Circus cyaneus (n = 149), and Canada goose Branta canadensis (n = 131). These seven species accounted for approximately 63% of our estimated flight heights.

Our power analysis for ANOVA and t-test suggested we needed 12 or more observations per season to detect differences in flight heights between two seasons, 14 or more observations per season to detect differences between three seasons, and 16 or more observations per season to detect differences between four seasons. We observed 14 species with sufficient sample sizes to detect seasonal differences (Table 1). We observed no differences among seasons for barn swallow Hirundo rustica (t = 1.56; df = 59; P = 0.125), Canada goose (t = 0.08; df = 129; P = 0.940), horned lark (F_3,164 = 1.59; P = 0.195), longspur Calcarius spp. (t = 0.21; df = 40; P = 0.837), mallard Anas platyrhynchos (t = 0.71; df = 40; P = 0.479), mourning dove (t = 1.58; df = 260; P = 0.115), northern harrier (F_2,146 = 0.70; P = 0.500), or red-winged blackbird (F_3,453 = 0.14; P = 0.939).

We found that the flight heights of six bird species varied among seasons. We observed common grackle Quiscalus quiscula flew higher in summer (68.6 ± 27.8 m; mean ± 95% CI) than spring (33.1 ± 12.6 m; t = 2.43; df = 59; P = 0.020). Great-tailed grackle Quiscalus mexicanus flew higher in summer (52.7 ± 19.0 m) than in spring (26.2 ± 9.5 m; t = 2.25; df = 46; P = 0.029). Killdeer Charadrius vociferous flew higher in summer (50.8 ± 12.1 m) than in spring (27.8 ± 10.8 m; t = 2.11; df = 71; P = 0.038). Therefore, common grackle, great-tailed grackle, and killdeer were at greater risk of collision in summer when their flight heights were within the RSZ than in spring when their flight heights were below the RSZ. Sandhill crane flew higher in winter (63.1 m ± 9.7) than in autumn (37.1 ± 5.1 m; t =5.13; df = 272; P < 0.001). Although sandhill crane flight heights during both seasons occurred within the RSZ, they flew lower during autumn than during winter. Western kingbird Tyrannus verticalis flew higher in summer (24.0 ± 4.6 m) than in spring (12.2 ± 4.7 m; t = 3.04; df = 63, P = 0.004). Also, meadowlarks flew higher in summer (17.0 ± 3.4 m) than in spring (11.2 ± 2.1 m), autumn (8.4 ± 1.6 m), or winter (7.2 ± 2.2 m; F_3,236 = 9.32; P < 0.001). Although mean flight heights varied among season for western kingbird and meadowlarks and were below the RSZ, the higher flights observed during summer put these species at greater risk of collision than during other seasons.

We found 28 (42.4%) of 66 species exhibited mean flight heights within the RSZ and mean flight heights of two species were above the RSZ. Fifteen of those species were wetland-associated species and 7 were raptor or vulture species. Six of these species were TPWD's species of greatest conservation need (Bender et al. 2005; TPWD 2012; Figure 2). Ferruginous hawk Buteo regalis was observed with a mean flight height of 60.7 m, Swainson's hawk Buteo swainsoni with a mean of 79.3 m, and Mississippi kite Ictinia mississippiensis with a mean of 159.6 m (Table 1). The bald eagle Haliaeetus leucocephalus, recently delisted from the US Endangered Species Act (ESA 1973, as amended) list, but still a species of concern in Texas and protected under the US Bald and Golden Eagle Protection Act (BGEPA 1940, as amended), was observed with a mean flight height of 57.2 m (Table 1). Northern pintail Anas acuta was observed with a mean flight height of 48.1 m and white-faced ibis Plegadis chihi with a mean flight height of 87.5 m (Table 1). We found that of the 28 species with mean flight heights in the RSZ, 8 species had their 95% confidence intervals (CIs) contained completely within the RSZ and are therefore thought to be at greater risk of turbine collision. These eight species were bald eagle (57.3 ± 21.91 m), Canada goose (92.8 ± 11.82 m), common grackle (47.6 ± 13.40 m), common nighthawk Chordeiles minor (74.4 ± 38.95 m), mallard (51.6 ± 13.31 m), sandhill crane (46.7 ± 4.99 m), snow goose (Chen caerulescens; 47.6 ± 17.75 m), and Swainson's hawk (79.3 ± 26.16 m; Figure 2).

Our power analysis for the χ² test suggested we needed 31 observations per season to detect differences in the proportion of flight heights within the RSZ between two seasons, 39 observations per season to detect differences between three seasons, and 44 observations per season to detect differences between four seasons. We had seven species with sufficient sample sizes to detect seasonal differences (Table 1). We observed no differences in the proportion of flight heights within the RSZ between autumn and winter for Canada goose (χ² = 0.06; df = 1; P = 0.814), horned lark (χ² = 1.78; df = 1; P = 0.182), or northern harrier (χ² = 1.29; df = 1; P = 0.256; Table 1). For mourning dove, we observed no differences between spring and summer (χ² = 2.96; df = 1; P = 0.086), and for red-winged blackbird, no differences were observed among spring, summer, autumn, and winter (χ² = 0.38; df = 3; P = 0.944; Table 1). We found the proportion of flight heights within the RSZ varied among seasons for two species. Meadowlark species flew more within the RSZ in summer than in spring or autumn (χ² = 8.36; df = 2; P = 0.015). Sandhill crane flew more within the RSZ in winter than in autumn (χ² = 3.83; df = 1; P = 0.050; Table 1).

We observed 14 species (21.2%) had greater than 25% of their flight heights within the RSZ, and we considered these species at greatest risk of collision (Table 1; Figure 3). We found that the array of species with greater than 25% of their flight heights within the RSZ was composed of 21.4% raptors and vultures, 50.0% wetland-associated species, and 28.6% passerines and other species (Figure 3). The 14 species were bald eagle (P = 0.016), Canada goose (P < 0.001), common grackle (P = 0.003), greater white-fronted goose (Anser albifrons; P < 0.001), great-tailed grackle (P = 0.018), mallard (P < 0.001), mourning dove (P = 0.009), northern pintail (P = 0.017), northern shoveler Anas clypeata (P = 0.020), red-winged blackbird (P = 0.009), sandhill crane during autumn (P < 0.001) and winter (P < 0.001), snow goose (P < 0.001), Swainson's hawk (P = 0.002), and turkey vulture Cathartes aura (P < 0.001; Table 1; Figure 3). Three of these species were identified as species of greatest conservation need by TPWD: bald eagle, northern pintail, and Swainson's hawk.

Ten birds identified as species of greatest conservation need by TPWD were observed, on average, below the RSZ (Table 1): American kestrel Falco sparverius; burrowing owl Athene cunicularia; Cassin's sparrow Peucaea cassinii; dickcissel Spiza americana; grasshopper sparrow Ammodramus savannarum; lark sparrow Chondestes grammacus; loggerhead shrike Lanius ludovicianus; longspur species, including some observations of McCown's longspur Rhynchophanes mccownii; northern harrier; and scissor-tailed flycatcher Tyrannus forficatus. However, five of these species were observed, at least some of the time, flying within the RSZ (i.e., see proportion in RSZ; Table 1). American kestrel was observed flying in the RSZ 31% of the time (95% CI = 15–51%), loggerhead shrike 13% of the time (95% CI = 0–53%), longspur species 20% of the time (95% CI = 10–35%), northern harrier 18% of the time (95% CI = 12–25%), and scissor-tailed flycatcher 18% of the time (95% CI = 5–40%).

Previous species-specific work on bird flight heights has been limited to categorizing observations as below, within, or above the RSZ (Osborn et al. 1998; Hoover and Morrison 2005; Mabee et al. 2006; Johnston et al. 2014) or limited to only a few species that can carry a sensor capable of measuring and monitoring flight heights (e.g., Cleasby et al. 2015). Work that categorizes observations as below, within, or above the RSZ has limited applicability because the RSZ changes as wind turbine technology evolves and from site to site (National Research Council 2007; Johnston et al. 2014). Our research is distinctive because it characterized distributions of diurnal flight heights for 66 bird species on the southern Great Plains (Table 1; Data S1, Supplemental Material), which allows application to wind turbines with different RSZs. Wind energy developers can use our data to help assess which bird species are at greatest risk of collision given their planned wind turbine specifications.

Most bird species were not at risk of collision with wind turbines due to their low diurnal flight heights (below the RSZ). However, some bird species did fly within the RSZ. Our results indicated that raptors and wetland-associated species are the avian groups at greatest risk of collision with wind turbines at our study sites due to their diurnal flight heights. We found that the flight heights of six bird species varied among seasons, indicating their risk of collision changed throughout the year. Flight heights tended to increase during summer for some passerines and killdeer, but sandhill crane exhibited risker flight heights during winter.

We found seven raptor species with mean flight heights within or above the RSZ (Table 1), and three had greater than 25% of their estimated flight heights within the RSZ (Figure 3). Although raptors and vultures made up only approximately 10% of our total observations, their abundance is likely not the best indicator of risk (de Lucas et al. 2008; Ferrer et al. 2012). Flight heights within or above the RSZ suggested that raptors and vultures may be a group at greater risk of collision with wind turbine blades than other species groups. At a wind energy facility in the Texas Panhandle, Miller (2008) found that approximately 8% of bird fatalities were raptors and approximately 36% were vultures.

Research into raptor collision fatalities has identified red-tailed hawk Buteo jamaicensis and American kestrel in many regions to have high collision potential, possibly due to their flight and hunting behaviors (Hoover and Morrison 2005; Arnett et al. 2007). Both species were observed exhibiting some flights in the RSZ during our study. Flight heights of red-tailed hawk were observed 47% of the time in the RSZ and American kestrel 31% of the time in the RSZ (Table 1). Hoover and Morrison (2005) found that red-tailed hawk uses the landscape and winds when hunting in a way that can lead to greater collision potential around wind turbines.

We found 15 species with mean flight heights within or above the RSZ were wetland-associated species (Table 1). Also, seven wetland-associated species had greater than 25% of their observed flight heights within the RSZ (Figure 3). Osborn et al. (1998) found that flight characteristics of both the raptor and waterfowl groups indicated they were at greatest risk of turbine collision. To lessen the risk of turbine collisions by wetland-associated birds, it may be prudent to avoid placement of wind turbines near playas and riparian systems. Mitigation of raptor fatality, however, may require additional considerations such as avoiding migratory routes, distance to nesting and perching habitat (i.e., trees in grassland-dominated communities), and prey availability (Hoover and Morrison 2005; Marques et al. 2014).

Most of the bird species (∼55%) that we observed flew, on average, below the RSZ (Table 1). However, 21% of bird species exhibited greater than 25% of their flight heights within the RSZ, indicating that they are at greater risk of collisions with turbines. Howe et al. (2002) found that birds flew below turbines in northeastern Wisconsin, with less than 14% of birds estimated within the RSZ (42–89 m). In the northern Great Plains, Osborn et al. (1998) also found that the majority of birds flew below the RSZ (21–51 m) and fewer (16–18%) birds flew in the RSZ; however, they observed that waterfowl and raptors were at greatest risk and passerines were at least risk of collision. We also observed few passerines in which their mean diurnal flight heights were within the RSZ (Table 1). However, migratory flights during nocturnal hours, which we were unable to observe, may be an important contribution to collision risk (e.g., Fijn et al. 2015).

Many bird species have been shown to modify their behavior around turbines. For example, Nicholson et al. (2005) reported that the majority of raptors and vultures (∼84%) avoided turbine blades by flying below or in adjacent valleys in the southeastern United States. In another area, northern harrier showed increased caution around a wind energy facility and avoided turbines (Smallwood et al. 2009). Western meadowlarks occasionally modified traveling behavior near turbines, but they were also recorded perching on turbines (National Research Council 2007; Smallwood et al. 2009). Smallwood et al. (2009) found that some species, such as American crow Corvus brachyrhynchos, cliff swallow Petrochelidon pyrrhonota, red-winged blackbird, and western meadowlark, will fly within 25 m of wind turbines. They also noted that some individuals and species of birds were less cautious around turbines when engaged in activities such as foraging and interacting with other birds (Smallwood et al. 2009). Similarly, in Minnesota, Osborn et al. (1998) reported that approximately 83% of the birds they observed modified their behavior by either flying above or below the RSZ (22–55 m). This observation suggested that birds can modify their behavior around wind energy facilities and turbine blades.

Bird collision risk with wind turbines can vary with species-specific behavior (Ferrer et al. 2012; Marques et al. 2014), and flight behaviors change due to foraging patterns, time of day, weather conditions, and migration activity (Johnston et al. 2014). For example, territorial displays, breeding displays, and migratory travel can lead to seasonally higher collision rates (National Research Council 2007; Smallwood et al. 2009). Smallwood et al. (2009) found that western meadowlark exhibited several high-risk behaviors near wind turbines during spring and summer, and these behaviors were correlated to seasonal flight heights within the RSZ. We also found meadowlark species typically flew at greater heights (17.0 ± 3.4 m) during summer than during other seasons (Figure 4). We found that six species flew at different heights during different seasons. For common grackle, great-tailed grackle, killdeer, meadowlark species, and western kingbird, we observed summer flight heights were greater than those in other seasons, putting these species at greater risk of collision during summer (Table 1; Figure 4). We found that sandhill crane had higher mean flight heights in winter than in autumn, perhaps as a response to changes in resource availability (Iverson et al. 1985; Table 1; Figure 3). Understanding seasonal patterns in flight heights provides wind energy facility managers with knowledge that is useful for developing mitigation strategies to reduce bird fatalities.

Estimates of diurnal flight heights provide an initial assessment of bird species that may be at risk of collision with turbines. However, detailed documentation of other behaviors that may expose birds to collision, such as nocturnal migrations, territorial displays, breeding displays, and foraging behaviors, are also needed (Osborn et al. 1998; Smallwood et al. 2009; Johnston et al. 2014; Marques et al. 2014). For example, at an offshore wind energy facility in the North Sea, Fijn et al. (2015) found that within in the RSZ, half of the bird echoes recorded by vertically mounted radar were at night. Other behaviors associated with breeding or foraging activities may be important as well. For example, some species such as the western meadowlark may have mean flight heights below the RSZ (Figure 4), but still exhibit other high-risk behaviors, such as perching on turbines or interacting with other birds near turbines that lead to collision fatalities (Smallwood et al. 2009). Collision fatalities from such high-risk behaviors have been observed at some currently operating wind energy facilities (National Research Council 2007). Osborn et al. (1998) noted that some species typically fly above the RSZ. We also observed cattle egret Bubulcus ibis and Mississippi kite, on average, flew above the RSZ (Table 1). These species are likely at greater risk of collision than species that flew below the RSZ because they must travel through the RSZ to reach those heights.

Identification of bird species at greater risk of wind turbine blade collision is important to help mitigate bird fatalities at wind energy facilities. As indicated by flight heights, we found raptor and wetland-associated species were at greatest risk of collision with wind turbines in the southern Great Plains. In prairie regions, we recommend avoiding wind turbine placement in locations with high concentrations of trees and shrubs that provide nesting and perching habitat for many raptor species. We recommend avoiding wind turbine placement in areas of high raptor prey densities where raptors may concentrate to feed (e.g., prairie dog towns). Also, we recommend avoiding placement of wind turbines near playa wetlands because many wetland-associated species, which were at greater risk of collision due to their flight heights, concentrate around them to feed, roost, and nest.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Diurnal bird flight heights observed during October 2008–August 2009 in Gray and Donley counties, Texas. All data used in the analysis of bird flight heights are contained in a tab-delimited text file (wulffdata.txt). This file contains seven columns and 2,707 records. The site column identifies which study site each observation was made. The species column identifies the bird species using four-letter alpha codes of Pyle and DeSante (2014), except MEAD was used for unidentified meadowlark species. The season column indicates in which season an observation was made. Seasons were defined as winter (December–February), spring (March–May), summer (June–August), and autumn (September–November). The month, day, and year columns indicate the date in which the observation was collected. The height column is the estimated flight height of each observation recorded in meters.

Found at DOI: http://dx.doi.org/10.3996/042015-JFWM-031.S1 (92 KB TXT).

We appreciate the cooperation of numerous private landowners who allowed access to lands. We thank D. Rankin, A. Berner, C. W. Boal, K. Boydston, A. Linehan, and J. Du Terroil for help with this project. We also thank M. Ellis and three anonymous reviewers for edits and suggestions. This is Texas Tech University, College of Agricultural Science and Natural Resources technical publication T-9-1281.

We thank Iberdrola Renewables, Inc.; Texas Parks and Wildlife Department; Texas Tech University; and the Bricker Foundation for sponsoring this study.

Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.