Interspecific variation in lower temperature thresholds of an assemblage of wintering bats

Abstract

Winter presents environmental and energetic challenges for temperate insectivorous bats as colder temperatures increase metabolic rates while simultaneously reducing resource availability. While bats in northern regions typically hibernate or migrate to circumvent these adverse conditions, there is growing evidence of winter bat activity as weather permits. Bats at lower latitudes may experience shorter, milder winters, increasing opportunities for activity. To better understand the relationship between ambient temperature and winter bat activity, we deployed acoustic detectors in central Louisiana and eastern Texas and examined data at 3 levels of biological organization: overall bat activity, species richness, and species-specific activity. Across 1,576 detector-nights, we recorded 37,435 bat passes. Bats responded positively to warmer temperatures but the temperature threshold for winter activity varied among species, ranging from 7.2 to 15.6 °C. Consequently, observed species richness increased at warmer ambient temperatures. With activity linked to environmental conditions in a species-specific manner, different subsets of the winter bat assemblage may be active from night to night. Additionally, our study adds to a rather limited body of literature of winter bat activity and provides a baseline for future studies as white-nose syndrome and climate change affect North American bat populations.

El invierno ocasiona dificultades ambientales y energéticas a murciélagos insectívoros de zonas templadas puesto que las temperaturas bajas hacen que aumente la tasa metabólica al mismo tiempo que produce la escasez de recursos naturales. Si bien los murciélagos de regiones norteñas típicamente hibernan o migran para eludir estas condiciones adversas, hay evidencias que apuntan a que existen actividades de murciélagos durante el invierno siempre y cuando se cumplan ciertas condiciones ambientales. A menores latitudes los murciélagos cuentan con inviernos más cortos y moderados, lo que favorece al aumento de actividades nocturnas. Para comprender la relación que existe entre la temperatura ambiental y la actividad de los murciélagos durante el invierno, hemos instalado detectores acústicos en el centro del estado de Louisiana y al Este de Texas y hemos examinados los datos en tres niveles de organización biológica: actividad total, riqueza de especies, y actividad específica por especies. De 1.576 noches de muestreo hemos registrado 37.435 pases de murciélagos. Estos, responden positivamente a las temperaturas más cálidas, sin embargo la temperatura umbral para actividades de invierno variaron de 7.2 a 15.6 °C en las distintas especies de murciélagos. Como consecuencia, la riqueza de especies observada aumentó con el aumento de la temperatura ambiental. Debido a que la actividad de murciélagos está relacionada a condiciones ambientales que son específicas para cada especie, distintos grupos de murciélagos del ensamblaje de invierno pueden activarse diariamente. El presente estudio contribuye al limitado número de literatura científica acerca de las actividades de invierno de los murciélagos y sirve de referencia para estudios posteriores como el síndrome de la nariz blanca y los cambios climáticos que afectan a las poblaciones de murciélagos de Norteamérica.

For temperate endotherms, winter is an energetically challenging period when cold temperatures increase thermoregulatory costs and reduce food availability (Marchand 2013). Thermoregulatory challenges are particularly pronounced for small-bodied endotherms with high mass-specific metabolic rates (Bennett and Harvey 1987; Spaargaren 1994). Accordingly, species that are able to remain active throughout winter tend to be large-bodied (e.g., American Bison, Bison bison, that can move snow to continue foraging on ground plants; Fortin et al. 2002) or have reliable access to food (e.g., Canada Jay, Perisoreus canadensis, that cache food to sustain themselves throughout winter; Sutton et al. 2021). For species where thermoregulatory costs are too high or food is not reliably available, there are 2 general winter strategies (Auteri 2022): species can escape to more favorable conditions through long-distance movements (e.g., migration; Dingle 2014) or engage in seasonal torpor and hibernate until favorable conditions return (Geiser 2013).

The effectiveness and availability of different overwintering strategies depend on physiological and morphological constraints, resource availability, and geography (Marchand 2013). Migration is a strategy that is restricted to highly mobile species that tend to be either large-bodied (e.g., ungulates) or volant (e.g., birds). Conversely, hibernation is generally restricted to small-bodied species (Geiser 2004) that cannot otherwise travel sufficient distances to escape winter. Whether migration or hibernation, overwintering strategies represent a balance of costs and benefits that might vary by resource availability (e.g., increased food resources for hibernating Eastern Chipmunks, Tamias striatus, reduced the expression of torpor; Humphries et al. 2003a) or geography (e.g., birds overwintering at lower latitudes having lower body mass; Nolan and Ketterson 1983). As such, it is often informative to consider cases that do not fall neatly into predefined categories. Understanding intermediate strategies and exceptions to the rule can provide a broader or more generalized understanding of biological phenomena.

As small-bodied heterotherms, temperate insectivorous bats are capable of hibernating, but their ability to fly also allows them to travel long distances. There is extensive research on hibernation strategies in temperate bats, where individuals persist on accumulated fat stores until warmer temperatures and reliable food resources return (Humphries et al. 2003b; Willis 2017). Other species migrate long distances, thereby limiting exposure to unfavorable winter conditions (e.g., Cryan 2003; Fleming and Eby 2003). However, there is mounting evidence to suggest that bats may be active to varying degrees during winter. As small-bodied insectivores, it is not possible for bats to remain homeothermic and survive through winter when insect availability is greatly reduced. But, in regions with milder or more variable weather, heterothermic bats might use torpor through times of inclement weather and become intermittently active during periods when conditions are favorable (Boyles et al. 2006). Despite a growing list of anecdotal records (Boyles et al. 2006), winter bat activity is generally perceived as rare. At more northern latitudes, some species are intermittently active throughout winter (e.g., Schwab 2014; White et al. 2014; Klüg-Baerwald et al. 2016) while a growing body of literature has indicated a higher diversity of winter-active bats at southern latitudes (e.g., Grider et al. 2016; Bernard and McCracken 2017; Stevens et al. 2020; Jackson et al. 2022; Kunberger and Long 2022) where winters are warmer and shorter.

Although the knowledge base for overall winter activity is growing, there has been little consideration of species-specific differences in responses to winter conditions. All temperate bat species can use torpor to escape adverse weather conditions, but morphological and ecological characteristics might determine the conditions under which each species is active. Interspecific variation in winter activity is difficult to assess at northern latitudes because overall activity is low and there are generally few species present during winter (Lausen and Barclay 2006; Falxa 2007; Schwab 2014; White et al. 2014; Klüg-Baerwald et al. 2016; Lemen et al. 2016; Johnson et al. 2017; Reynolds et al. 2017). Conversely, at southern latitudes, there is greater species richness (Stevens and Willig 2002), including overwintering migrants that arrive from more northern summer locations (Cryan 2003; Fraser et al. 2012). Thus, temperate insectivorous bats overwintering in areas such as the southeastern United States provide a study system in which a variety of species may adopt intermediate overwintering strategies and do not strictly migrate, hibernate, or remain continuously active.

The combination of migration, hibernation, and winter activity may also be considered from a community ecology perspective. The winter bat assemblage in temperate regions consists of resident species, individuals of partially migratory species that remain overwinter, and individuals of migratory species that arrive during winter from northern breeding areas. Although individuals of each of those species may be present in the region, species that hibernate throughout the winter do not contribute to the assemblage of active bats. From the perspective of the assemblage of active bats, a species that is fully migratory (all individuals depart in winter) and a species that exclusively hibernates throughout the winter (all individuals enter hibernation in autumn and do not emerge until after winter is over) are functionally equivalent, and equivalently absent. However, species that may be intermittently active throughout winter may or may not be included in the assemblage of active bats depending on whether they are active over the timeframe of the specific question under investigation. Moreover, if there is predictable interspecific variability in response to environmental conditions (e.g., winter temperatures), then the active assemblage structure may vary on a nightly basis with implications for interpreting bat community data collected during winter.

To better understand how bat species are affected by ambient temperature during winter, we investigated an assemblage of insectivorous bats overwintering in Louisiana and eastern Texas, United States. In our study region, there were periods of cold that precluded any flying insect activity (Mellanby 1939), interspersed by relatively mild conditions that could have permitted foraging. We hypothesized that activity patterns of overwintering insectivorous bats would be related to variation in ambient temperature. Specifically, we predicted that overall bat activity would increase with temperature as has been found in previous studies of winter activity (e.g., Klüg-Baerwald et al. 2016). However, we further predicted that species would have different temperature thresholds for activity. Smaller-bodied species lose heat to the environment more rapidly and may be less likely to be active on colder nights, cavity-roosting species are more sheltered from the weather than species that roost in foliage, and species with geographic ranges that extend far north of our study site may be better adapted to colder environmental conditions than species near the northern limit of their distribution. Finally, we considered how different temperature thresholds among species might contribute to temporally variable structure of the bat assemblage. We predicted that if species exhibited varying temperature thresholds for activity, then observed species richness—and not just total bat activity—would increase with temperature.

Materials and methods

Study area and site selection.

We used acoustic monitoring to document bat activity in managed pine (Pinus spp.) forests of the South Central Plain ecoregion of central Louisiana and eastern Texas (Fig. 1). Most stands were comprised of planted Loblolly Pine (Pinus taeda), managed on an approximate 25- to 30-year rotation. Typical silviculture included site preparation after harvest of the existing stand, planting pine seedlings, herbaceous weed control, and thinning. Hardwood forests dominated by oak (Quercus spp.), Sweet Gum (Liquidambar styraciflua), American Beech (Fagus grandifolia), and Bald Cypress (Taxodium distichum) typically occurred along riparian corridors and in lowland areas. Hardwood trees seldom occurred in pine-dominated forest stands. We concentrated our efforts in areas with the most contiguous forest represented by the broadest range of stand ages, and deployed equipment in diverse forest types. Our study occurred in 3 study sites—2 in central Louisiana (Jackson/Bienville Parishes and Winn Parish; note that parishes are jurisdictional regions equivalent to counties) and 1 in eastern Texas (Newton County). All fieldwork methods were approved by the Texas Tech University Institutional Animal Care and Use Committee (protocol 17092-12).

Acoustic detector deployment and monitoring.

Each week, from 30 December 2017 to 26 March 2018 and from 31 December 2018 to 11 March 2019, we deployed 12 acoustic detectors (SM4BAT-FS with SMM-U1 microphone; Wildlife Acoustics, Inc., Maynard, Maryland) at 72 distinct locations of varying stand-age classes and cover types (n = 24 in each study site). We placed detectors in areas with limited understory cover to minimize clutter calls and thereby produce higher-quality echolocation recordings for more reliable identification (Britzke et al. 2013). We placed microphones 3 m above ground, pointed down at a 45° angle to permit water to run off (Weller and Zabel 2002). We set detector parameters to reduce influence of background noise and increase likelihood of recording bat calls—Gain 12 dB, 16k High Filter off, Sample Rate 256 kHz, Minimum Trigger Duration 1.5 ms, No Maximum Trigger Duration, Minimum Trigger Frequency 16 kHz, Trigger Level 12 dB, Trigger Window 3 s, and Maximum File Length 15 s. We set detectors to record from 30 min before sunset to 30 min after sunrise and deployed them for a minimum of 7 days before being rotated to the next study site, as weather permitted. Over the 2 winters, we recorded bat activity at each of the 72 locations for a minimum of 21 days.

To obtain local temperatures, we suspended an iButton (model DS1921G, iButtonLink, LLC, Whitewater, Wisconsin) from the base mounting flange of each detector and recorded ambient temperature at 10-min intervals. We used iButton temperature recordings to identify the temperature at sunset and the temperature at the time of each recording, to the nearest 10 min. This allowed us to characterize ambient temperatures at the time and place that bats were recorded.

Acoustic analysis.

To acquire species-level identifications, we considered the twelve species known to occur in the South Central Plain ecoregion: Corynorhinus rafinesquii (Rafinesque’s Big-eared Bat), Eptesicus fuscus (Big Brown Bat), Lasiurus borealis (Eastern Red Bat), Lasiurus cinereus (Hoary Bat), Lasiurus intermedius (Northern Yellow Bat), Lasiurus seminolus (Seminole Bat), Lasionycteris noctivagans (Silver-haired Bat), Myotis austroriparius (Southeastern Myotis), Myotis septentrionalis (Northern Long-eared Bat), Nycticeius humeralis (Evening Bat), Perimyotis subflavus (Tricolored Bat), and Tadarida brasiliensis (Brazilian Free-tailed Bat). Of these species, L. borealis and L. seminolus cannot be acoustically differentiated and were subsequently grouped for analysis. We used 2 automated identification programs to analyze recordings: SonoBat (v4.2.2 southeast region pack SE[C20170529]; SonoBat, Arcata, California) and Kaleidoscope Pro (v5.1.9i; Wildlife Acoustics Inc., Maynard, Maryland). In our analyses of species-specific patterns, we included only recordings for which there was agreement on species identification between the 2 programs. We further manually vetted identifications to remove obvious classification errors such as noise files identified as L. cinereus or feeding buzzes that were misidentified as Myotis spp. but we did not manually override species identifications from the software. Thus, we conducted species-level analyses with a data set that included only files identified to species, whereas analysis of overall bat activity included all files with confirmed echolocation calls whether identified to species or not.

Statistical analyses.

We quantified bat activity as the number of passes (sequence of calls recorded as a bat flew past the microphone; Fraser et al. 2020). To evaluate effects of temperature on overall bat activity, we determined the mean sunset temperatures recorded across all active detectors. We then used a breakpoint regression (Muggeo 2008) to identify the mean sunset temperature at which an increase in overall activity was apparent. We conducted all remaining analyses using the subset of passes that were identified to species.

When identifying species-specific lower temperature thresholds, we did not consider species that represented <1% of the classified data set as there were insufficient data to properly characterize their response to temperature. For the remaining species, the temperature distribution was likely biased by sample size. The most common species had more data points to characterize the tails of the temperature distribution. To avoid this bias, we conducted an iterative process to calculate the 10th percentile of the temperature distribution of each species based on a common sample size. The selection of the 10th percentile was arbitrary, but we selected this threshold to account for individual bats that may have been active at colder than typical temperatures for reasons other than the general temperature preference of the species (e.g., activity that may have been caused by disturbance or individuals that were active at cold temperatures not typical of the species). This approach provided a standardized method to characterize the lower tail of the distribution of temperatures at which each species was active. We randomly selected 100 passes which represented approximately one-third of the data set for the species with the fewest recordings. We then calculated the 10th percentile of the temperatures at which those passes were recorded and repeated this process 30 times for each species to obtain 30 estimates of the minimum temperature threshold for each species. We then compared temperature thresholds across species with ANOVA and Tukey’s post hoc test.

Although the possible number of species in our data set is limited (n = 11 possible species), we conducted preliminary analyses to test for possible ecological or morphological correlates that might affect temperature thresholds. To test for an effect of body size on temperature threshold, we used linear regression to compare the estimated temperature threshold with mean forearm length for each species (taken from Ammerman et al. 2012). Our sample did not permit statistical analysis of roosting differences, but we examined the temperature threshold of the 2 foliage-roosting species in our data set (L. borealis/L. seminolus, L. cinereus) as compared to the other species that roost in cavities. Similarly, our data set did not permit statistical analysis based on geographic distribution of the species, but we examined the temperature threshold for 2 of our species (T. brasiliensis, M. austroriparius) that have a notably more southern distribution than the other species.

Finally, we evaluated the relationship between species richness and temperature. We used linear regression to compare the species richness identified across all detectors on a night of recording with the mean sunset temperature across those detectors. This approach avoided issues of undersampling at any given detector and removed potential biases associated with habitat selection across detectors. We conducted all statistical analyses in R (v4.2.2; R Core Team 2022). We assessed breakpoint regressions using R package “segmented” (Muggeo 2008), and used R package “multcomp” (Hothorn et al. 2008) for post hoc tests. We report all results as mean ± standard error.

Results

During winters of 2017 – 2018 and 2018 – 2019, we recorded in central Louisiana and eastern Texas over 132 recording nights, resulting in 1,576 detector-nights, and 37,435 bat passes. We documented all 11 species/groups expected to occur in the region (Table 1) and we detected bats on 127 of 132 recording nights. Nights without bat activity were colder than nights with activity (t-test: t_5.3 = 6.2, P = 0.001). Mean sunset temperature on nights without bat activity across all 12 detectors was 3.2 ± 1.4 °C (range 0.6 to 8.0 °C) compared with 12.7 ± 0.6 °C (range −3.3 to 26.0 °C) on nights when at least 1 pass was recorded.

Bat activity increased at warmer temperatures, but the relationship between bat activity and sunset temperature was better explained by a breakpoint regression model than a simple linear regression (F_2,128 = 44.9, P < 0.0001). The breakpoint model indicated a sharp increase in activity above 17.4 ± 0.6 °C (Fig. 2a; below breakpoint slope = 1.4, F_1,128 = 192.8, P < 0.0001; above breakpoint slope = 14.3, F_1,128 = 89.8, P < 0.0001). Of the 11 species/groups, 5 were detected at below-freezing temperatures (Table 1), but activity at these temperatures was rare with only 9 passes identified to species. Four species were excluded from species-specific analyses, each representing <1% (n < 140 passes) of the total number of identified passes (C. rafinesquii, L. intermedius, L. noctivagans, and M. septentrionalis). Of the 7 remaining species (including the L. borealis/L. seminolus group), the lower temperature threshold varied among species (Fig. 3; F_6,203 = 296.5, P < 0.001). Temperature thresholds were similar for L. cinereus and N. humeralis (Tukey’s post hoc t = 0.24, P > 0.99), but all other pairwise species comparisons indicated distinct temperature thresholds (Fig. 3; Table 1). Acknowledging a limited sample size of 7 species/groups, temperature thresholds did not vary with forearm length (F_1,5 = 0.05, P = 0.83).

We observed the maximum possible species richness (11 species/groups) on 11 nights. All but one of these nights were in March (1 night in mid-January), and all with mean sunset temperature > 14 °C (range 14.9 to 25.5 °C). Consistent with the observation that temperature thresholds varied among species, species richness was positively related to sunset temperature (r² = 0.63, F_1,130 = 223.6, P < 0.001; Fig. 2b).

Discussion

We recorded bat activity throughout the winter, and although mean sunset temperature in our study region was considerably warmer (12.3 ± 0.6 °C, n = 132) than those reported in previous acoustic surveys of winter activity (e.g., Alberta: −7.3 °C; Klüg-Baerwald et al. 2016; Tennessee, 4 to 11 °C; Bernard and McCracken 2017), we regularly recorded bats on colder nights. All species were detected at temperatures < 10 °C with 4 species active at subfreezing temperatures (Table 1). Although bat activity at subfreezing temperatures has been previously reported (O’Farrell and Bradley 1970; Lausen and Barclay 2006; Schwab 2014; Klüg-Baerwald et al. 2016), foraging in these circumstances appears to be rare (summarized in Boyles et al. 2006; see also Bernard et al. 2021). Foraging in subfreezing temperatures is unlikely as insect activity is greatly reduced (Mellanby 1939). In our study region, periods of freezing temperatures did not persist long enough for waterbodies to freeze over, making drinking a plausible explanation for activity. Roost-switching is also a likely alternative as some species are known to change roost type when faced with colder ambient temperatures, i.e., L. borealis (Saugey et al. 1998; Mormann and Robbins 2007); L. seminolus (Hein et al. 2008); L. noctivagans (Perry et al. 2010). However, these movements appear to be less frequent in freezing conditions (Hein et al. 2008).

The lower temperature threshold for activity varied among species (Table 1; Fig. 3) but there was no obvious morphological, behavioral, or ecological factors underlying these differences based on our data. For example, foliage-roosting bats (L. cinereus, L. borealis/L. seminolus) have longer, more dense fur and occupy roosts that are more exposed to ambient conditions but were not the species with the lowest temperature thresholds. Conversely, some of the cavity-roosting species in our region had higher temperature thresholds, and others had lower temperature thresholds. More detailed field study of the actual winter roosts used in our study region—and how those vary with changing weather conditions—is required before conducting a more robust assessment of the influence of roosting ecology.

Temperature response was also not apparently related to body size, where larger bats with relatively smaller surface area to volume ratio and lower mass-specific metabolic rates might be expected to be more tolerant of colder temperatures. The temperature threshold for larger bats such as E. fuscus and L. cinereus was much higher than for M. austroriparius, one of the smallest species. In fact, after removing rare species from our analysis, E. fuscus had the highest temperature threshold of activity (15.5 °C). As future studies are conducted in other regions and collect data on additional species, it will be possible to conduct more robust analyses of possible influences of body size, although it will be important to account for potential intraspecific regional variation (see below).

Contrary to expectations, the lowest temperature thresholds were observed for T. brasiliensis and M. austroriparius (8.3 °C and 7.2 °C, respectively), the species with the most southern distribution in our study (Jones and Manning 1989; Wilkins 1989). The geographic range of T. brasiliensis is primarily Neotropical but the local subspecies T. brasiliensis cynocephala occurs in the region year-round (LaVal 1973) and has been rapidly expanding northward, potentially in response to climate change (McCracken et al. 2018). In the northern parts of the range, M. austroriparius hibernate throughout winter but in more southern regions generally remains active throughout winter (Jones and Manning 1989), consistent with observations in our study. Species with a historically southern distribution may not be adapted for extended periods of hibernation. In contrast with the southern M. austroriparius that had a notably lower temperature threshold and relatively high winter activity, P. subflavus (similar body size compared to M. austroriparius) has a distribution that extends far to the north of our study sites. As may be expected for a species that regularly occurs in temperate regions with longer and harsher winters, P. subflavus had the second highest observed temperature threshold, and winter activity was rare. During culvert surveys in our study regions, it was not uncommon to observe active M. austroriparius but P. subflavus were typically torpid and likely to have been hibernating for extended periods as indicated by extensive condensation on their fur (our unpublished observations; Stevens et al. 2017).

Geographic distribution is also interesting to consider for widespread species in the context of regional variation. Of the species in our study region, E. fuscus is one of the most widely distributed and had the highest temperature threshold (15.5 °C). Although we documented E. fuscus activity at temperatures as low as −4.5 °C, the high-temperature threshold indicates that winter activity at low temperatures was rare in our study region. This observation is in contrast with observations of winter activity at more northern latitudes (Lausen and Barclay 2006; White et al. 2014; Klüg-Baerwald et al. 2016). Metabolic responses to cold exposure may vary among populations, whether through local adaptation or other mechanisms and southern populations may be less cold-tolerant than their northern counterparts. Data from E. fuscus in Montana (47° north latitude) indicate that torpid metabolic rate does not vary across a wide range of temperatures (McGuire et al. 2021), but when exposed to colder temperatures (≤2 °C) big brown bats from southern populations (Alabama, 31° north latitude) experienced more elevated torpid metabolic rates compared to their counterparts from more northern populations (Michigan 46° north latitude) experimentally exposed to the same temperature conditions (Dunbar and Brigham 2010). Eptesicus fuscus is a sedentary species, which may allow for differentiation among populations, unlike some of the migratory species for which populations from widespread regions may overlap in winter.

Variation in winter temperature thresholds likely exists in all winter bat assemblages, but previous studies may have been limited in the abilities to characterize interspecific differences by low species richness typical of more northern latitudes (Lausen and Barclay 2006; Falxa 2007; Schwab 2014; White et al. 2014; Klüg-Baerwald et al. 2016; Lemen et al. 2016; Johnson et al. 2017; Reynolds et al. 2017). In Louisiana and east Texas, there were as many as 11 species active on a given night, with a positive relationship between observed species richness and mean sunset temperature during winter (Fig. 2b). Driven by variation in temperature thresholds among species (Fig. 3), the assemblage of active bats varies on a nightly basis, and likely varies within a night as temperatures decline. An important implication is that acoustic surveys conducted in winter should not assume equal likelihood of detecting all species in a given assemblage. These dynamics present an intriguing opportunity for future community ecology studies. For example, assemblage structure is the product of both interspecific interactions (Stevens and Willig 1999) and species-specific responses to the environment (Andersen et al. 2022; Kunberger and Long 2022) with relative contributions being context-dependent (Leibold and Chase 2018). One context is seasonally variable temperatures and the species-specific differences that drive responses. Moreover, as responses to low temperatures in winter described here appear to be strong, the relative contribution of temperature to community structure may be greater in winter than summer when conditions are generally favorable for all species. Another intriguing consequence is that although remaining physically present in the system, torpid individuals are inactive and do not interact with other species of bats or the arthropods that they consume. The temporal scale of the active species assemblage (i.e. at what timescale do individuals or different species respond to variation in weather conditions), and the implications for aspects of ecosystem processes, presents opportunities for future study.

Our study provides a baseline of winter bat activity in the region (see also Kunberger and Long 2022). White-nose syndrome (WNS) has caused declines of hibernating bat populations across eastern North America (Frick et al. 2015). Although, at the time of this study, WNS had not been documented in the South Central Plains ecoregion (Limon et al. 2019), the disease has begun to encroach from the north and west (whitenosesyndrome.org). WNS alters the activity of overwintering bats by causing more frequent arousals (Reeder et al. 2012; Warnecke et al. 2012; Bernard and McCracken 2017). However, it is unclear how differences in WNS responses may affect, or relate to, the lower temperature thresholds we observed. Winter bat activity is also likely to be affected by changing climate, although few studies currently consider short-term behavioral responses to weather, focusing instead on range expansions and contractions and large-scale demographic consequences (Festa et al. 2023). Notably, Smeraldo et al. (2021) highlight the challenge of predicting the consequences of changes at the assemblage level. Range shifts have already been observed or suggested for several of the species in our study—N. humeralis (Kaarakka 2018), L. seminolus (Perry 2018), and T. brasiliensis (McCracken et al. 2018), but we highlight the potential for altered winter activity patterns as warming conditions differentially affect species with varied active temperature thresholds.

Acknowledgments

Thanks to E.B. Smith and the U.S. Forest Service for providing lodging throughout the study. We appreciate the support of D. Greene (Weyerhaeuser Company) and T. Pagels (Manulife Investment Management) for coordinating land access. Thanks to B.J. Dennis, J.P. Wilson, W.R. Conway, C.J. Garcia, B.A. Ward, and B.N. Ward for providing assistance in the field. Thanks also to H. Amarilla-Stevens for assistance with preparing the manuscript.

Author contributions

BRA: conceptualization, data curation, formal analysis, investigation, methodology, writing—original draft; RDS: conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, writing—review & editing; JRG: formal analysis, writing—review & editing; LPM: conceptualization, formal analysis, funding acquisition, methodology, supervision, writing—review & editing.

Funding

The research was funded by the National Council for Air and Stream Improvement, Inc. (NCASI) and Texas Tech University. LPM was supported by the Natural Sciences and Engineering Research Council of Canada.

Conflict of interest

The authors declare that they have no conflict of interest.

References

Author notes