Species Richness and Seasonality of Bat Occupancy on Northwestern National Wildlife Refuges

Bats are among the most diverse fauna of North America, with 45 species found within the continental United States and Canada, and numbers within colonies reaching into the millions (Hammerson et al. 2017). Bat conservation is critical for ecosystem integrity; however, bats are among the most overlooked but economically important nondomesticated species in North America (Boyles et al. 2011; Kunz et al. 2011). Indeed, many basic aspects of bat biology are unknown because of difficulties in studying small, nocturnal, and far-dispersing species, including foraging practices, roost selection, seasonal distributions, and effects of management actions (Findley 1993; Fenton 1997; Pierson 1998; Cryan 2003; Williams et al. 2006), particularly for solitary and crevice-dwelling species (Barclay et al. 1988). North American bat populations face growing threats (see review by O'Shea et al. 2016), including a novel fungal pathogen, white-nose syndrome (WNS; Blehert et al. 2009; O'Regan et al. 2015); changes in land use and increasing energy production (Arnett et al. 2008; Arnett and Baerwald 2013; Russo and Ancillotto 2015; Jung and Threlfall 2016); and a changing climate (Humphries et al. 2002; Jones et al. 2009; Adams 2010; Sherwin et al. 2013). As a result of recent large die-offs associated with the emergence of WNS (Cryan et al. 2010), North American bats are currently facing an unprecedented array of threats from both disease and anthropogenic stressors (Dixon et al. 2013). As of 2015, 18–31% of the 45 North American species are known to be at some measure of risk (Hammerson et al. 2017). Once common bat species are now faced with extirpation (Frick et al. 2010, 2015; Russell et al. 2015) and, subsequently, land managers have increased focus on bat conservation efforts (U.S. Fish and Wildlife Service 2011). However, despite advances in monitoring technologies, evaluating the impacts of these threats on populations and across ranges is limited by challenges associated with studying such cryptic and dispersed species (O'Shea and Bogan 2003; Hayes et al. 2009; Weller et al. 2009; Meyer 2015).

During the summer, most bat species in the United States form colonies dominated by females (i.e., maternity colonies; Dixon et al. 2013). Female bats generally exhibit fidelity to maternity colonies, at which they congregate to raise offspring (Arnold 2007; Dixon 2011). Many bat species then make annual migrations between summer and winter ground (Fleming and Eby 2003); however, migration may be the most poorly understood aspect of bat biology (McGuire et al. 2012). Some species appear to travel only a few hundred kilometers from summer roosts to hibernaculum, whereas others may move much longer distances of > 1,000 km (McGuire et al. 2012). Few studies have detailed seasonal movements of many species of North American bats despite evidence of migration (Cryan 2003; Popa-Lisseanu and Voigt 2009; Norquay et al. 2013). Information on where many bat species overwinter, including whether or not individuals remain active during winter months or hibernate (Cryan et al. 2014), is lacking. Further, flexibility in migratory behaviors and partial migration appears to be common in bats (see Fleming and Eby 2003; Popa-Lisseanu and Voigt 2009). Much of what is known about seasonal distributions comes from observations recorded by different researchers working in limited areas during different seasons, although when compiled, this information can reveal general trends (Cryan 2003). In addition, documenting the phenology of bat migration (i.e., timing of seasonal arrivals and departures) can provide important information on migratory routes (Strelkov 1969). Information on timing of emergence from hibernacula and movement to maternity colonies is also important for monitoring distribution and impact on WNS, which was recently discovered on the West Coast (Washington Department of Fish and Wildlife 2018).

Public lands established for conservation, such as National Wildlife Refuges and national parks, have great potential to support resource protection, disease surveillance, and population monitoring efforts for bats (Rodhouse et al. 2016). However, to plan and implement effective monitoring programs, seasonal distribution and species composition information is needed across the larger network, including on which land management units do bats overwinter; when do species susceptible to WNS emerge or arrive on the landscape in spring; and the locations of hibernacula and maternity colonies that can then be monitored for WNS and species declines. Acoustic monitoring equipment has reduced many of the issues associated with traditional capture and visual methods for collecting species occurrence data (Williams et al. 2006), and has proven to be more effective and less invasive in detecting the presence of many species (Kalko et al. 1996; O'Farrell and Gannon 1999). In particular, acoustic equipment can be deployed for extended time periods and at multiple sites simultaneously. Our objectives in this study were to use passive acoustic monitoring to assess the current assemblage of bat species, identify overwinter locations and activity, and document seasonal timing of species occurrence.

Bats were sampled on six National Wildlife Refuges (Refuge) in the northwestern United States. Two Refuges were in the Northern Rockies (Kootenai [Idaho] and Little Pend Oreille [Washington]); three Refuges were in the Columbia Basin portion of Washington (Columbia, McNary, and Toppenish); and one Refuge was in the northern Great Basin (Sheldon [Nevada]; Figure 1). Detectors were placed next to water features on all Refuges but the type of feature and landscape surrounding each Refuge varied. At the Columbia Basin sites, habitats were varied. Habitat on Columbia Refuge consisted of wetlands, lakes, and streams surrounded by sagebrush steppe Artemisia spp. and invasive annual grasses (i.e., Bromus tectorum). The Refuge was surrounded by agricultural lands and the detector was placed next to a small, permanent stream, flanked on either side by low (3–6 m) basalt cliffs. McNary Refuge was characterized by open water sloughs along the Columbia River, bordered by agricultural and suburban landscapes. The detector was placed next to a permanent slough. Toppenish Refuge was dominated by seasonal wetland and bounded to the south by sagebrush steppe land cover and agricultural land on the other sides. The detector was placed along an irrigation ditch that was flanked by woody riparian vegetation. Within the Northern Rockies sites, Kootenai Refuge supported seasonal wetlands, surrounded by agricultural land in a broad valley. The land cover surrounding the valley was dominated by mixed-coniferous forest (western red cedar Thuja plicata, Ponderosa pine Pinus ponderosa, and Douglas fir Pseudotsuga menziesii), and the detector was placed next to a small pond. Little Pend Oreille Refuge was dominated by Ponderosa pine forest and the detector was placed along a stream within a herbaceous meadow. In the Great Basin, Sheldon Refuge was dominated and surrounded by sagebrush steppe habitats. The detector was placed next to a thermal perennial creek at the head of a deep, narrow gorge. Among the Refuges, elevations ranged from 540 to 1,090 m in the Northern Rockies sites, 105 to 250 m in the Columbia Basin sites, and 1,326 to 2,183 m at the Great Basin site.

Between 2011 and 2016, we collected bat call data from stationary passive acoustic monitoring sites; the majority of the data was collected in 2014–2015. We remotely collected call data using an ultrasound detector (Pettersson ultrasound detector, model D 500x; Pettersson Elektronik AB, Uppsala, Sweden) using either the detector's internal microphone or an attached external microphone mounted approximately 1–3 m above the ground. We standardized detector settings across the project and were as follows: frequency = 500 kHz; pretrigger = OFF; file length = 5 s; input gain = 80; trigger level = 120; interval = 0. We scrubbed collected calls to remove files containing miscellaneous noise using SonoBat (version 3.1) and automatically processed them using SonoBat (version 3.1.1, Eastern Washington or Great Basin filters) with the SonoBatch feature. We standardized SonoBatch default settings as follows: max. no. of calls to consider per file = 8; acceptable call quality = 0.80; acceptable quality to tally passes = 0.20; decision threshold = 0.90. At Sheldon Refuge, the duration of individual trapping sessions were affected by site accessibility, battery life, and equipment failures. For example, factors such as heavy snow periodically limited site access and unanticipated cold temperatures reduced battery life. At the other study sites, the detectors were solar powered and operated nightly, although equipment failures led to occasional interruptions. As such, because the trapping effort was not uniform across sites or time, we grouped calls into bimonthly time periods for analysis. We defined the winter period as 15 November–15 March (Neubaum et al. 2006); we classified bats as overwintering or potentially overwintering if detected during that time period. We defined movement associated with seasonal dispersal as arrival in March through April (O'Farrell and Studier 1975), although early March is a known period of overlap between hibernation and transition periods (Neubaum et al. 2017). We defined departure in the fall as between August (Fellers and Pierson 2002; O'Farrell and Studier 1980; McGuire et al. 2012) and the first half of November.

Methods for identifying species followed O'Farrell et al. (1999). We defined a sequence as a series of individual vocalizations (calls > 0.5 ms) with more than two individual calls by a single bat as it passed within range of the detector (O'Farrell and Gannon 1999). We discarded sequences that were unidentifiable because of poor quality or lack of distinctive characteristics. We used only calls of sufficient quality (discriminant probability ≥ 0.90) and minimum number of call pulses accepted by SonoBat (for LACI, no. accpt > 1; all other species, no. accpt. >3) in the analysis. To discriminate species by echolocation calls, SonoBat uses a probabilistic process whereby multiple quantitative parameters of a call are compared with the same parameters of reference calls from known species. Discriminant probability is the indicator of how far the parameters measured from an unknown bat call fall within the known data for that species (SonoBat, version 3.1.1). We selected a discriminant probability of ≥ 0.90 as the default value suggested by the software's developers (SonoBat, version 3.1.1).

A subset of calls (n = 6,623) underwent a secondary manual vetting process and we classified each as to confidence of correct species identification: 1) probable: high-quality, archetypical sequence for the species; 2) likely: insufficient qualitative or quantitative data (bandwidth or quality-call pulses), but suggestive for the species indicated; and 3) possible: insufficient components for a high degree of confidence, but identify species not included in the probable or likely categories (J.D. Tyburec, Bat Survey Solutions, personal communication). We vetted at least one call per species per detector per month and we vetted all calls from species outside of their expected range. We used calls vetted as probable and likely to evaluate species presence by time frame. We also used individual bats identified through capture efforts from a recent mist-netting effort on Sheldon Refuge (2009–2010; G.H. Collins, unpublished data) to further evaluate the likelihood of overall species occurrence for that study site.

Between August 2011 and November 2016, we collected 210,444 bat calls during 3,691 trap nights (x̄ = 42.79 calls/trap night; Table 1; Data S1, Supplemental Material). Of those, 55.0% (n = 115,855) were of sufficient quality to be autoclassified (≥ 90% probability) to species by the software. We detected a total of 16 species: pallid bat Antrozous pallidus, Townsend's big eared bat Corynorhinus townsendii, big brown bat Eptesicus fuscus, spotted bat Euderma maculatum, western red bat Lasiurus blossevillii, hoary bat Lasiurus cinereus, silver-haired bat Lasionycteris noctivagans, California myotis Myotis californicus, western small-footed myotis Myotis ciliolabrum, long-eared myotis Myotis evotis, little brown myotis Myotis lucifugus, fringed myotis Myotis thysanodes, long-legged myotis Myotis volans, Yuma myotis Myotis yumanensis, western pipistrelle Pipistrellus hysperus, and Brazilian free-tailed bat Tadarida brasiliensis (Table 2). We secondarily confirmed all species as present by manual vetting of acoustic files, with the exception of the Brazilian free-tailed bat occurrence on Sheldon Refuge. For that species, we at best classified the vetted calls as “possible” and we captured no individuals during recent mist-netting efforts on Sheldon Refuge; therefore, more survey work is needed to confirm their presence.

Overall bat species richness varied across the study sites. The most diverse site was Sheldon Refuge in the Great Basin (n = 15 species detected) and the least was McNary Refuge in the Columbia Basin (n = 6 species detected; Figure 2). Species richness also varied across time: in general, we detected the highest number of species during the summer months (June–August) across all sites (Figure 2). We detected all 16 bat species expected to occur across the larger area (Harvey et al. 1999). Within the Great Basin, we detected 14 of 15 possible species (fringed myotis was expected but not found). We did detect western red bat at Sheldon Refuge, which was outside of the predicted range for that species (Harvey et al. 1999). For the Columbia Basin, we detected 12 of 14 expected species; missing were spotted and long-legged bats. In the Northern Rockies, we detected 11 of 11 expected species. We additionally detected western pipistrelle at Little Pend Oreille Refuge during this study, which was outside of the species expected range (Harvey et al. 1999). We also detected this species at the other Northern Rockies site, Kootenai Refuge, during other acoustic sampling efforts outside of this study (J.K. Barnett, unpublished files).

We found overwintering bats on Refuges across all three geographic areas (Table 3). Only one species, western pipistrelle, was documented as active year round on more than one Refuge, Columbia and Sheldon. On Sheldon Refuge, we also documented silver-haired bats as active year round. We additionally classified six species as likely overwintering because we detected them in early March, a period of overlap between hibernation and transition periods (Neubaum et al. 2017), but we did not consistently detect them during the winter months: silver-haired bat (Little Pend Oreille); Yuma myotis (Toppenish); western small-footed myotis (Sheldon, Toppenish); and pallid, spotted, and hoary bats (Sheldon; Table 2). Other incidental detections during the winter period, generally only one or two calls, included big brown bat (Columbia), Yuma myotis (Little Pend Oreille), little brown myotis (McNary), and California myotis and western pipistrelle (Toppenish).

We selected four species common across the geographic areas to evaluate for seasonal occurrence across the larger landscape: hoary bat, silver-haired bat, big brown bat, and little brown myotis. First detections for nonoverwintering species ranged between early March to late April and last detections ranged between early October and early November. Both the activity level and number of species detected generally peaked in July to September (Figure 3). For arrival and departure of the migratory tree bats, hoary and silver-haired bats were the most protracted across time and space. Silver-haired bats were active in the Great Basin throughout winter, in the Northern Rockies in February, but we did not detected them in the Columbia Basin until March (Figure 3). We did not find big brown bats to overwinter in our study area; they arrived in late March and April, and then departed the Northern Rockies 6 wk earlier than the other geographic regions. In contrast, little brown myotis exhibited a compressed arrival pattern, with activity increasing rapidly over a 2-wk period in late April. In fact, across the entire study area, we detected one little brown bat in late March, seven in early April, and 106 in late April. Detections stayed relative high throughout the summer and then decreased rapidly in early October. The species was absent from all areas from late October to early March (Figure 3).

We evaluated three additional species that we did not commonly detect on all six Refuges but are considered potentially vulnerable to WNS—western small-footed myotis, long-eared myotis, and Yuma myotis—for seasonal activity after hibernation. With a lack of large hibernacula in the west, the best time to test for WNS is in the spring as the bats return to maternal colonies (Washington Department of Fish and Wildlife 2018). In the Northern Rockies, Yuma myotis became active starting in early March. The timing for both Yuma and western small-footed myotis was also early March in the Columbia Basin. In the Great Basin, western small-footed myotis became active in late March, but long-eared myotis and Yuma myotis both weren't found to be active until early May (Figure 4).

Passive acoustic monitoring is a useful tool in determining and comparing seasonality and species richness of bats across large geographic areas, and can be used to evaluate changes over the long term. The detectors worked well to monitor for bat activity in remote areas, could be used simultaneously, and were able to operate for long periods of time. However, site accessibility for maintenance needs to be a consideration when planning a deployment, particularly during winter in areas of high snow. The ability to close the gap between observed and expected species is an important conservation action needed within protected-area networks (Rodhouse et al. 2016). We detected a total of 16 species across six Refuges, although the observed vs. expected species varied by geographic area. Species richness varied between Refuge and throughout seasons, but was consistently highest during the summer months. Only two Refuges documented the same species of bat (western pipistrelle) as consistently active during winter. In the Great Basin, we also detected silver-haired bats as active during the winter, and we identified an additional six species as potentially overwintering within the different geographic areas.

Two main factors appear to determine distribution and habitat use by bats: roost availability and prey abundance (Fenton 1997), and roost sites are often a limiting resource (Hayes and Wiles 2013). The diversity of bat species that we documented at Sheldon Refuge was likely due to the diversity of habitats and abundant roosting opportunities (Williams et al. 2006). For example, we detected 65% of the bat species known to occur in Nevada (n = 23 species; Bradley et al. 2006) at our single location on Sheldon Refuge, the Refuge with the highest species richness. That area of northwest Nevada met several of the criteria to score as high for conservation value for bats, including high species richness (> 4 species), isolated with none to few disturbances, and high opportunities for public land conservation (Ball 2002). In contrast, species richness was lowest at McNary Refuge. Low richness at McNary Refuge may be due in part to lack of rock features offering crevice roosts. Rock features were widely available in the general vicinity of the other Refuges and immediately adjacent to the detector at Columbia Refuge. Toppenish and Columbia Refuges also likely have higher availability and diversity of roost structures. Refuges in the Northern Rockies offered potential roosts of numerous types, including rock crevices, trees, and buildings. Overall, the Refuges were utilized by all species expected in the Northwest. However, the lack of spotted bats at Columbia Refuge was surprising, given the extensive cliff habitat on the Refuge and known spotted bat populations within 100 km (Hayes and Wiles 2013). Spotted bat populations are highly localized and associated with suitable roosting cliffs, usually > 30 m high in Washington (Hayes and Wiles 2013). Perhaps cliff features on Columbia Refuge are of insufficient height to support the species. Previous acoustic inventories on Columbia Refuge have also failed to document spotted bat (J.K. Barnett, unpublished files).

All hibernating bat species periodically arouse and some individuals leave the hibernaculum, but it is believed that they do not fly long distances (Lausen and Barclay 2006). Detection of echolocations calls in winter suggests that bats hibernate nearby. We consistently detected calls on Sheldon and Columbia Refuges but no bats were detected on the other four Refuges during the middle of winter. There is limited information on hibernacula for many western bat species and few winter congregations are known. Western species may winter in undiscovered hibernation sites in rock crevices (Hayes and Wiles 2013). Rock cliffs were immediately adjacent to the detectors on Columbia and Sheldon Refuges, where western pipistrelle was detected in winter. The other four Refuges may not provide suitable roosts for hibernation. For species detected in early March, but not earlier in winter, it is likely they were hibernating in the vicinity (Hayes and Wiles 2013). We defined winter as November 15 to March 15, but March 1 to 15 is concurrently classified as a transient period, when bats begin to move from hibernacula to summer roosts (Neubaum et al. 2017). In all cases we also classified species detected in early March as seasonal and consistently detected on the respective Refuges throughout the summer, suggesting there could also be maternity colonies in the general area.

Among migrating bats, hoary bats in particular are considered to be long-distance migrators, and we found additional evidence for latitudinal migration during our study, as described by others (Cryan 2003, Johnson et al. 2011). Hoary bats were active in the Great Basin area during fall, with activity peaking in September and October. We then detected the species in the Columbia Basin in early March and in the Northern Rockies in late March, suggesting northward movement of the species during the migration period. The reverse occurred in fall, with the last detections in the Northern Rockies occurring in late October, early November in the Columbia Basin, and late November to the south in the Great Basin. In the Great Basin, hoary bat activity levels peaked in fall, which may represent a pulse of animals migrating through. However, a small number of detections of hoary bats on Sheldon Refuge in February also suggests that at least some individuals may remain to overwinter in the area. Silver-haired bats also make long migrations (Johnson et al. 2011), although both migratory and winter populations are known to occur in Washington (Hayes and Wiles 2013). We detected the species throughout the winter on Sheldon Refuge and in late winter (March) on Little Pend Oreille Refuge. Silver-haired bats have also been detected in northeastern Washington (Spokane County) during winter (N. Williams, personal communication). We did not detect silver-haired bats in the Columbia Basin until March, but nightly activity at McNary Refuge was extremely high from late March through early May (average bat pass per night = 174.6; range 1–679; J.K. Barnett, unpublished files).

Arrival time of little brown myotis on the landscape is of particular interest because of the species' susceptibility to WNS, which was detected in Washington after our study was concluded (www.whitenosesyndrome.org/static-spread-map). Of the seven species we investigated for seasonal movements, little brown myotis had the most consistent and latest arrival (late March to early April) followed by an abrupt and consistent departure (early October) across all three geographic regions. Even on Kootenai Refuge, where the detector was placed within 400 m of a known maternity colony, little brown myotis arrival was sudden and occurred in April. A noticeable peak of activity occurred on Sheldon Refuge in late August, potentially indicating a swarming site. Swarming is associated with breeding behaviors and marked by when bats congregate around roosting sites in the fall; such sites potentially offer a known place to mark bats to find hibernacula. In the Columbia Basin, western small-footed myotis and Yuma myotis were detected 2 to 4 wk earlier than little brown myotis. Yuma myotis have been confirmed with WNS in Washington (Washington Department of Fish and Wildlife 2018), although the susceptibility of western small-footed and long-eared myotis to WNS is unknown. Few hibernacula for these species have been located, so WNS surveillance activities may be conducted at maternity sites (Washington Department of Fish and Wildlife 2018). Further, long-term acoustic monitoring such as that used in our study has the potential to capture changes in species composition or seasonality over time, which might indicate the presence of WNS, particularly if there is high overwinter mortality of a susceptible species (Ford et al. 2011).

Results of this study document the importance of National Wildlife Refuges to a wide range of bat species, including supporting migratory tree bats and both wintering and breeding populations of multiple Myotis species. The major threats to North American bats (i.e., WNS, energy development, and disturbance) primarily affect bats when they are hibernating or migrating (Hammerson et al. 2017). Myotis species are particularly vulnerable to WNS, with eastern populations experiencing significant mortality (Frick et al. 2010). It has been also hypothesized that migratory bat populations will decline with climate warming because of increased water deficiencies and resulting water stress (Popa-Lisseanu and Voigt 2009). Renewable energy interests such as wind energy also have the potential to affect migratory bat populations, as large numbers of bats and birds are killed by collisions with wind turbines (Kunz et al. 2007; Winhold et al. 2008). Understanding the seasonal movements of bats has become even more urgent with the emergence of WNS, and documenting overwintering areas is an important first step in assessing their vulnerability (Norquay et al. 2013; Cryan et al. 2014). As with national parks (Rodhouse et al. 2016), National Wildlife Refuges have the ability to promote long-term viability of bat populations and contribute to monitoring programs (Dixon et al. 2013). Because of the large scale of bat seasonal movements, information gained from individual units will be more limited when compared with a broader monitoring network (Rodhouse et al. 2016). Further, efforts at landscape monitoring across units would allow for prioritized conservation and resource efforts, where land management units with high species richness, rare or underrepresented species, or critical habitats could receive additional focus (Rodhouse et al. 2016).

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. Number of classified bat calls collected, by date and species, on six national wildlife Refuges in the northwestern United States between 2011 and 2016. Refuges included Columbia (CMB, Washington), Kootenai (KTN, Idaho), Little Pend Oreille (LPO, Washington), McNary (MCN, Washington), Sheldon (SAR, Nevada), and Toppenish (TPN, Washington). Species included ANPA = pallid bat Antrozous pallidus; COTO = Townsend's big-eared bat Corynorhinus townsendii; EPFU = big brown bat Eptesicus fuscus; EUMA Euderma maculatum = spotted bat; LABL = western red bat Lasiurus blossevillii; LACI = hoary bat Lasiurus cinereus; LANO = silver-haired bat Lasionycteris noctivagans; MYCA = California myotis Myotis californicus; MYCI = western small-footed myotis Myotis ciliolabrum; MYEV = long-eared myotis Myotis evotis; MYLU = little brown myotis Myotis lucifugus; MYTH = fringed myotis Myotis thysanodes; MYVO = long-legged myotis Myotis volans; MYYU = Yuma myotis Myotis yumanensis; and PAHE = Western pipistrelle Pipistrellus hysperus.

Found at DOI: https://doi.org/10.3996/032019-JFWM-019.S1 (3.43 MB PDF).

Reference S1. O'Shea TJ, Bogan MA. 2003. Monitoring trends in bat population of the United States and Territories: problems and prospects. U.S. Geologic Survey, Biological Research Discipline, Information and Technical Report USGS/BRD/ITR-2003-0003.

Found at DOI: https://doi.org/10.3996/032019-JFWM-019.S2 (9.54 MB PDF); also available at https://pubs.usgs.gov/itr/2003/0003/report.pdf.

Reference S2. U.S. Fish and Wildlife Service. 2011. A national plan for assisting states, federal agencies, and tribes in managing white-nose syndrome in bats. Hadley, Massachusetts: U.S. Fish and Wildlife Service.

Found at DOI: https://doi.org/10.3996/032019-JFWM-019.S3 (752 KB PDF); also available at https://s3.us-west-2.amazonaws.com/prod-is-cms-assets/wns/prod/b0634260-77d3-11e8-b37b-4f3513704a5e-white-nose_syndrome_national_plan_may_2011.pdf.

Reference S3. Washington Department of Fish and Wildlife. 2018. White-nose syndrome strategy plan. Olympia: Washington Department of Fish and Wildlife, Wildlife Diversity Division.

Found at DOI: https://doi.org/10.3996/032019-JFWM-019.S4 (4.26 MB PDF); also available at https://www.whitenosesyndrome.org/response-plans/washington-department-of-fish-and-wildlife-strategy-plan-april-2018.

Funding was provided by the U.S. Fish and Wildlife Service. Call vetting was conducted by J.K.B. (USFWS) and J. D. Tyburec (Bat Survey Solutions, LLC). We thank D. Ellis, M. Munts, G. Warrick, S. McFall, K. Lotz, and J. Lucas, who helped with the detectors. M. Gregg and J. Kasbohm reviewed early drafts of the manuscript. We appreciate the critical insights of the editors and reviewers.

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