We initiated a multi-year, pre-construction study in mid-summer 2009 to investigate patterns of bat activity and evaluate the use of acoustic monitoring to predict mortality of bats at the proposed Resolute Wind Energy Project (RWEP) in east-central Wyoming. The primary objectives of this study were to: 1) determine levels and patternsof activity for three phonic groups of bats (high-frequency emitting bats, low-frequency emitting bats, and hoary bats) using the proposed wind facility prior to construction of turbines; 2) determine if bat activity can be predicted based on weather patterns; correlate bat activity with weather variables; and 3) combine results from this study with those from similar efforts to determine if indices of pre-construction bat activity can be used to predict post-construction bat fatalities at proposed wind facilities. We report results from two years of pre-construction data collection.
We recorded echolocation calls of bats with Anabat II zero-crossing ultrasonic detectors, programmed to record calls beginning ½-hour prior to sunset and ending ½-hour after sunrise each day of the study from 3 August–18 October 2009, and 2 June–30 September 2010. We assigned each bat pass to one of two phonic groups based on minimum frequency of the echolocation sequence; high frequency bats (≥33 kHz average minimum frequency; e.g., Myotis spp.) or low frequency bats (<33 kHz average minimum frequency; e.g., big brown, silver-haired, hoary bats). We also identified a third phonic group, hoary bats, a subset of the low frequency phonic group, because this species is vulnerable to wind-energy development and its echolocation sequences are relatively easy to distinguish among other low-frequency emitting bats. We used 5 meteorological (met) towers to position detector microphones at ~1.5 m and ~44 m above ground level (agl) to acoustically sample bat activity during this study.
In 2009, we recorded a total of 976 bat passes. We recorded 454 high frequency passes and 522 low frequency passes. Hoary bats comprised 22% (n = 114) of low frequency passes. In 2010, we recorded a total of 1,111 bat passes. We recorded 410 and 701 high frequency and low frequency passes, respectively. Hoary bats comprised 30% (n = 208) of low frequency passes.
Bat activity varied, by phonic groups, within and among nights. High frequency bats were most active 1–2 hours past sunset. Low frequency bat activity peaked during the middle of the night and hoary bats were most active within the first hour past sunset. Bat activity typically was highest between August and mid-September for all phonic groups. However, the timing and intensity of peak activity for each group differed between years.
Bat activity varied among phonic groups by height and among towers. We detected high frequency bats more often at 1.5 m agl with greatest activity recorded at tower 5042 in both 2009 and 2010. Low frequency bat activity was relatively consistent between heights and among towers for both years. We detected hoary bats more often at 44 m agl and recorded the greatest activity at towers 5032 and 5042 in 2009 and at tower 5032 in 2010. We recorded the fewest calls by any phonic group at tower 5034.
We modeled bat activity (passes/detector-night) in relation to tower location, temperature, and several measures of wind speed with 1) the probability of activity and 2) the estimated number of calls given that activity occurred. Tower location and temperature were consistently the most important factors in our models, accounting for ~5–29% of the variation in activity. However, location alone explained ~3–9.5% of the variation in activity. In general, we found the highest probability of activity and highest counts for each phonic group at towers on the western edge of the project. Both the probability of activity and estimated number of calls from each phonic group increased with increasing temperature. While some measure of wind was often important, it never explained more than an additional ~9% of the variation in activity. When included in the models, the effect of average wind speed on the probability of bat activity and estimated number of bat passes was always negative.
This study was conducted at a single proposed wind energy facility located on shrubland habitat in east-central Wyoming, and statistical inferences are limited to this site. However, we believe that our findings reflect patterns of bat activity on similar landscapes with comparable vegetation composition and topography in this region. Despite equipment malfunctions, we were able to quantify the spatial (vertical and horizontal) and temporal (seasonal and yearly) activity patterns of bats. These data may provide useful information for predicting when, where, and which bats may be most at risk of interactions with wind turbines at the RWEP. Moreover, specific timings and locations of peak activity may further refine the use of curtailment as a mitigation option.