Long-term Acoustic Assessment of Bats at Battle Creek, South Dakota for 2013-2015 Prepared for: Bureau of Land Management Prepared by: Dan Bachen, Braden Burkholder, Alexis McEwan, Shannon Hilty, Scott Blum, and Bryce Maxell Montana Natural Heritage Program A cooperative program of the Montana State Library and the University of Montana June 2020 Suggested Citation: Bachen, D. A., B. O. Burkholder, A. L. McEwan, S. L. Hilty, S. A. Blum, and B. A. Maxell. 2020. Long-term acoustic assessment of bats at Battle Creek, South Dakota for 2013-2015. Report to Bureau of Land Management. Montana Natural Heritage Program, Helena, Montana. 19 pp. Acknowledgments This project was conducted with funding from the Bureau of Land Management and would not have been possible without the support of this agency and its staff. Staff at Wildlife Acoustics assisted with questions regarding the SM2Bat+ ultrasonic detector and SMX-US microphone and Kaleidoscope Pro software. Joe and Nick Szewczak provided SonoBat 4.1 software, feedback on its use, and the 2011 Humboldt State University Bat Lab’s echolocation call characteristic summaries for western and eastern U.S. bats that we used to develop the call characteristic summary for Montana bats (Appendices 6 & 7 in Maxell 2015). Bowen Deng provided technical support while processing acoustic data on Montana Tech’s high performance computing cluster. At the Montana Natural Heritage Program, Darlene Patzer assisted with grant administration and Dave Ratz assisted with downloading of weather station data from the Mesowest application programming interface. Battle Creek, South Dakota - i Table of Contents List of Tables ................................................................................................................................................ iii List of Figures ............................................................................................................................................... iv Background ................................................................................................................................................... 1 Methods ........................................................................................................................................................ 3 Bat Detector Deployment ......................................................................................................................... 3 Data Management & Call Analyses ........................................................................................................... 3 Weather Station Data ............................................................................................................................... 4 Solar and Lunar Data ................................................................................................................................. 4 Results ........................................................................................................................................................... 5 General Patterns of Bat Activity................................................................................................................ 6 Timing of Bat Activity ................................................................................................................................ 7 Temperature and Bat Activity ................................................................................................................... 8 Wind Speed and Bat Activity ..................................................................................................................... 9 Barometric Pressure and Bat Activity ..................................................................................................... 10 Precipitation and Bat Activity ................................................................................................................. 11 Moonlight & Bat Activity ......................................................................................................................... 12 Species Activity Patterns ......................................................................................................................... 13 Discussion.................................................................................................................................................... 15 Listed Species Conservation .................................................................................................................... 15 White-nose Syndrome ............................................................................................................................ 15 Wind Energy Development ..................................................................................................................... 16 Management Recommendations ............................................................................................................... 18 Literature Cited ........................................................................................................................................... 19 Battle Creek, South Dakota - ii List of Tables Table 1. Species hand confirmed at the Battle Creek detector, by season. Species only observed previously and newly detected within the local area (50.0 km) are noted. ................................................. 5 Table 2. The number of months each bat species was confirmed by hand analysis of calls identified by automated software, the number of months reviewed, and the respective successful classification rate; only active season data are shown. ............................................................................................................ 13 Table 3. Management considerations for species detected within 50.0 km of the Battle Creek detector. Species presence is summarized by season and include this and any previous efforts. ............................ 17 Battle Creek, South Dakota - iii List of Figures Figure 1. Placement of the Battle Creek detector. ....................................................................................... 1 Figure 2. Photo of the detector deployment site at Battle Creek. ............................................................... 2 Figure 3. Total monthly bat passes recorded at the Battle Creek acoustic monitoring station. Months marked with an asterisk should be interpreted with caution as those data may not represent valid trend due to data collection for only part of that month, equipment malfunction or other issues. ..................... 6 Figure 4. Average nightly activity of bats recorded at the Battle Creek acoustic monitoring station across the active season. ......................................................................................................................................... 7 Figure 5. Average bat pass temperatures (red line) and average background temperatures (black line) across the year at the Battle Creek detector. ............................................................................................... 8 Figure 6. A comparison of average temperatures during bat passes (red) and average hourly background temperatures (blue) recorded at the INBS2 station, located 13.1 km to the west-northwest of the detector. Where the bars showing passes exceed hours, bat activity is higher than expected for this temperature bin. ........................................................................................................................................... 8 Figure 7. The cumulative sum of wind speeds recorded at the INBS2 station during bat passes. The speed at which 95% of all activity occurs at or below is highlighted in red. ........................................................... 9 Figure 8. A comparison of background windspeeds recorded at the INBS2 station (blue) and those recorded during bat passes (red). Where the bars showing passes exceed hours, bat activity is higher than expected for this wind speed bin. ........................................................................................................ 9 Figure 9. Hourly changes in background barometric pressure at the KEFC station (blue) compared to changes in pressure when bat passes were recorded (red) at the Battle Creek detector. Where the bars showing passes exceed hours, bat activity is higher than expected for this pressure change bin. ........... 10 Figure 10. A comparison of hours with and without precipitation for bat passes (red) and all nighttime hours (blue) during the active season as recorded at the INBS2 station, located 13.1 km to the west- northwest of the detector. Where the bars showing passes exceed hours, bat activity is higher than expected for this precipitation bin. ............................................................................................................ 11 Figure 11. Percent of bat passes (red) and background hours (blue) at various moon illumination categories (0% = no illumination and 100% = full moon) and with the moon above and below the horizon. Where the bars showing passes exceed hours, bat activity is higher than expected for this moon horizon/illumination bin. ............................................................................................................................ 12 Figure 12. Bat passes through the deployment period identified to species using SonoBat 4.1. Note that these species identification are only suggestions and should only be used to assess general trends for species for which the classifier works well. ................................................................................................ 14 Battle Creek, South Dakota - iv Background We established a long-term, ultrasonic acoustic monitoring site on BLM lands in northwestern South Dakota (Figure 1), following deployment and maintenance protocols in Maxell (2015). The detector was placed adjacent to a small lentic waterbody (Figure 2). The surrounding area was not forested and not rugged (flat topography). Figure 1. Placement of the Battle Creek detector. A SM2Bat+ detector with a SMX-US microphone was deployed on 23 Oct 2013 and decommissioned on 27 Jun 2015, for a total of 613 nights deployed. Throughout the recording period the detector functioned well, recording on 612 nights or 99.8% of the time. In total this unit collected data over 21 months and did not meet our minimum of 2 years of deployment for analysis of long-term trends. Battle Creek, South Dakota - 1 Figure 2. Photo of the detector deployment site at Battle Creek. Battle Creek, South Dakota - 2 Methods Bat Detector Deployment Across the acoustic network, detectors were placed at locations to maximize species diversity and bat activity through placement near features important for bats such as roosts, foraging areas, and waterbodies suitable for drinking. We assessed potential sites based on: (1) open water for as much of the year as possible; (2) rock outcrops and trees that might be used as roosts by bats; (3) southern solar exposure that would allow a solar panel to charge a battery even during the winter; (4) year-round accessibility; and (5) a low likelihood of vandalism. At all sites, a detector/recorder unit and microphone were deployed. The microphones at all operational sites in 2015 were upgraded to SMX-U1 microphones (Wildlife Acoustics Inc., Maynard, MA). The detector/recorder was deployed, monitored, and maintained with the equipment, supplies, settings, and protocols listed in Montana’s Bat and White- Nose Syndrome Surveillance Plan and Protocols 2012- 2016 (Maxell 2015). Many aspects of the equipment and site selections influenced the detection of a bat echolocation call and the quality of the resulting recording. These included sensitivity of the individual microphone, temperature, humidity, wind speed, and frequency, amplitude, distance, and directionality of echolocation calls emitted by bats (Parsons and Szewczak 2009, Agranat 2014). The energy of sounds spreading in all directions diminishes by one fourth for every doubling of distance because the surface area of a sphere is related to the square of its radius. Furthermore, higher frequency sounds are diminished over shorter distances because of atmospheric absorption (Parsons and Szewczak 2009, Agranat 2014). Testing of the SMX-US microphone used through June 2015 across the acoustic network indicated that bats emitting frequencies in the range of 20 kHz should be detected at distances of 24 to 33 meters from the microphone while those emitting frequencies in the range of 40 kHz should be detected at distances of 18 to 22 meters (Agranat 2014). These distances are the radii of the relevant spheres of detection around microphones when they are at full sensitivity. However, we know that sensitivity varied over time by an unknown magnitude because some precipitation and freezing events permanently reduced the sensitivity. In 2015 the microphones at active detectors were upgraded to the SMX-U1 microphone, which increased the quality of recorded calls and reduced the effect of adverse weather on microphone sensitivity over time. Due to this change in hardware, comparisons between data collected before and after June 2015 should be made with caution as the different models of microphone may affect the number of calls and species detected. Where applicable, individual reports for each unique equipment configuration were produced to minimize any interpretation errors. Data Management & Call Analyses Acoustic file recordings, in both original WAC and processed WAV formats, are stored in the Montana Bat Call Library which is housed on a series of 20-40 terabyte Drobo 5D storage arrays at the Montana State Library as well as a secondary offsite location to protect against catastrophic loss. Acoustic analysis results, temperature files, weather station data, and solar and lunar data were all processed and combined within SQL database tables in accordance with the general workflow pattern for data management and analysis outlined in the text and in Appendices 8-10 of Maxell (2015). Bat call sequences were analyzed with the goal of definitively identifying individual species presence by month and individual species’ minimum temperatures of activity in accordance with the Echolocation Call Battle Creek, South Dakota - 3 Characteristics of Montana Bats and Montana Bat Call Identification materials in Appendices 6 and 7 of Montana’s Bat and White-Nose Syndrome Surveillance Plan and Protocols 2012- 2016 (Maxell 2015). Weather Station Data Weather station data were downloaded using the Mesowest application programming interface (API) as outlined in Appendix 9 of Maxell (2015). Temperature, wind speed, solar, and precipitation data were downloaded from weather stations across the regions. Distance from the detector to the station varied by site and data type. All data from weather stations were averaged by hour and associated with all call sequences recorded within this hour bin for use in our analyses. Solar and Lunar Data Solar and lunar data were calculated for all hours of detector deployment using the Python package ephem (3.7.6.0), which uses well established numeric routines to produce high precision astronomy computations (see Appendix 10 of Maxell 2015). The underlying code produces results nearly identical to data available from the U.S. Naval Observatory (Astronomical Applications Department). Precise times for sunrise, sunset, moonrise, moonset, and percent illumination at the detector were calculated based on latitude, longitude, and date. It should be noted that local topography is not incorporated into any of these calculations. Therefore, the exact timing of these events on the ground may differ slightly from those produced by this model but should typically be within a few minutes unless local terrain differs greatly from the modeled horizon (e.g. if the site is at the bottom of a canyon). Battle Creek, South Dakota - 4