Abstract
This study describes a three part investigation of bird and bat mortality at 31 wind turbines in northern Kewaunee County, Wisconsin between 1998 and 2001. Construction of the towers was completed during the summer of 1999 by Wisconsin Public Service Corporation (WPS) and Madison Gas and Electric Company (MGE). The 14 WPS turbines are configured in three rows within 1.5 km of one another, while the MGE turbines are located in two irregular clusters approximately 3.5 km apart.
Point surveys for diurnal birds were conducted by several observers within two adjacent areas of approximately 75km2. One area (Turbine Area) encompassed the 31 wind turbines, while the other (Reference Area) served as a "control" with similar topography and land u se. Field observers recorded 165 bird species in the entire study area. More than 60% of all individuals belonged to 5 species: Ring-billed Gull, European Starling, Red-winged Blackbird, Canada Goose, and House Sparrow. Notable species included declining or uncommon grassland birds such as Eastern Meadowlark (23rd most abundant), Bobolink (28th most abundant), Northern Harrier (47th most abundant) and Upland Sandpiper (52nd most abundant), and 25 bird species listed as endangered, threatened, or special concern by the Wisconsin Department of Natural Resources. Two federally endangered/threatened raptors, Peregrine Falcon and Bald Eagle, were recorded during the surveys, but neither species was resident in the immediate project area.
Average numbers of species per point count were highest during summer, while average numbers of individuals were highest during autumn. During winter, large flocks of Lapland Longspurs and Snow Buntings were frequently recorded, although overall bird numbers generally were very low.
Neither the numbers of species nor the numbers of individuals recorded during short (3 minute) point counts differed significantly between the Turbine Area and Reference Area. During longer (30 minute) counts, the numbers of species (but not total numbers of individuals) were significantly higher in the Reference Area. Species composition was very similar in the two areas, although water birds tended to be more abundant in the Reference Area, which was located closer to the shores of Green Bay.
Comparison of diurnal birds before and after construction showed no significant difference in the average numbers of species. Numbers of individuals, however, were greater before construction. This change was mainly due to a decline in the abundance of gulls, which were likely influenced by conditions outside of the study areas and unrelated to presence of the wind turbines (e.g., changes in the conditions of nesting islands or roosting sites.).
Most diurnal birds were recorded at altitudes below the sweep area of the wind turbines. Fewer than 14% of the birds encountered were estimated at heights between 42-89 m, the range defined by the lower and upper reaches of the wind turbine blades. During spring, the percentage of birds observed in the sweep area was higher than at other times of the year, a result that was consistent over both years when spring samples are available. Differences in flight altitude between the Turbine Area and Reference Area were not significant overall, although the percentage of birds in the sweep area was highest in the Reference Area during 4 of the 6 seasons for which data are available.
Acoustic surveys of nocturnal migrants by William Evans led to the identification of 10,364 individuals representing at least 35 species or species groups. Major movements of migrants were highly irregular. Highest numbers of birds were recorded during May and from mid-August through late September. Much of the migration occurred during a relatively small number of nights. Results indicate that migrants flying over the wind turbines are no more numerous (and in some cases significantly lower) than at other stations in the area; numbers of nocturnal migrants were highest by far at a site located near Lake Michigan. Comparisons between sites cannot be made with confidence, however, because other factors such as background noise affect the numbers of interpretable calls. The most frequently recorded calls were made by warblers, including two species complexes and the regionally abundant Ovenbird and American Redstart. Cape May Warbler, a species of special concern in Wisconsin, was the 7th most frequently recorded bird.
Altitude of nocturnal migrants was evaluated for 7 nights with the highest frequency of calls. Most birds flew above the sweep area of the turbine blades; based on time delays in recordings from two different microphones, approximately 20%-22% of the calls were made by birds flying within the sweep area of the turbines. The distributions of calls were highly variable yet were not clearly related to cloud cover or storms. Variations in numbers of birds flying overhead during migration are consistent with the episodic nature of bird mortality at communications towers; if any significant bird mortality occurs at wind turbines, we might expect it to be similarly episodic based on these results. No large episodes of mortality were recorded during this study period, however.
During more than 1200 hr of field investigation (equally distributed among turbines) 25 bird carcasses were recovered, 13 at the WPS turbines and 12 at the MGE turbines. Two listed species were found, an immature Loggerhead Shrike (state endangered) along a road near one of the turbines and a Grasshopper Sparrow, a species of special concern in Wisconsin. The shrike was probably killed by a motor vehicle collision. Mortality was seasonal, with all but 4 carcasses appearing during the migration periods of April-May and August-October.
Bat mortality at the wind turbines was nearly 3 times higher than bird mortality (72 vs. 25 specimens). Nearly all carcasses were found between mid-August and mid-September, indicating a highly seasonal pattern. All but 7 of the specimens belonged to 3 migratory species (Hoary Bat, Red Bat, and Silver-haired Bat).
The spatial distribution of bird and bat carcasses suggested that the sampling area did not cover the entire area in which turbine-caused mortalities might be found. Collections along the access roads were used to adjust for this bias. Predator/scavenger removal experiments and observer efficiency experiments also were conducted to account for specimens that were overlooked. In fields where vegetation height was low (less than about 0.25 m) observer efficiency ranged from 20-72%. Predator/scavenger removal followed a fairly consistent probability of about .16 per night. By 20-22 days, all of the planted carcasses were gone.
Adjusting for a larger sampling area and the bias of searching inefficiency, we estimated that the number of carcasses recovered by observers represented only about 25% of all fatal collisions. This leads to actual mortality estimates of 1.29 birds / turbine / year and 4.26 bats / turbine / year. The bird estimate is slightly lower than a national estimate of 2.19 birds based on meta-analysis (Erickson et al. 2001), while the bat estimate is similar to results from a preliminary analysis of 3 wind turbines in Tennessee (Erickson et al. 2002).
Compared with other sources of human-caused bird mortality, the annual numbers of deaths caused by the Kewaunee County wind turbines are negligible, assuming that 1999-2000 and 2000-2001 were typical years. A catastrophic mortality episode during spring or autumn migration periods is not beyond the realm of possibility, of course, especially if mortality events at communications towers serve as a guide. Because the wind turbines are lower than communications towers and wires are not used for support, the probability of such a catastrophe is nevertheless low compared with the probability of catastrophes at tall communication towers.
The significance of bat mortality is less clear. Proximity of riparian forest might help explain the relatively high rates of bat mortality observed in our study compared with other recent investigations. Estimated mortality rates in Kewaunee County are similar to results from a predominantly forested landscape in Tennessee. As with birds, this level of mortality might be negligible compared with annual mortality from other sources, including human-caused mortality. Bats tend to be longer-lived and have lower reproductive rates than songbirds, however, so the effects of human-caused mortality might be correspondingly greater for bats than for birds. Nevertheless, Erickson et al. (2002) have argued that reported rates of bat mortality at wind turbines represent only a very small fraction of migratory or local bat numbers.