Conservation Value to Bats: Assessing Multiple Functional Habitats in a Nature Preserve at the Urban-Agricultural Interface via Temporal Ecology

Abstract

In grassland ecosystems, agriculture and urbanization are two main anthropogenic disturbances to native fauna. Nature preserves at the urban-agricultural interface may include diverse vegetation habitats, providing opportunities for native fauna. Limited research has examined the conservation value of such preserves to bats. We examined bat activity patterns at multiple temporal scales at Glacier Creek Preserve in Omaha, Nebraska, USA, between 2018 and 2020 via acoustic monitoring to identify what functional habitats it might provide to bats. We placed bat detectors along the forest edge and in the open, restored native prairies and open agricultural fields. A total of nine species were recorded at the preserve, including the endangered Myotis septentrionalis. Foraging activities were recorded for six species. The edge habitat had higher overall acoustic activities for three species (Eptesicus fuscus, Lasiurus borealis, and Lasiurus cinereus) and a higher proportion of foraging activities for two species (Lasiurus borealis and Lasiurus cinereus) than the open habitat. Lasiurus cinereus displayed activity peaks early at night, whereas Lasiurus borealis had activity peaks late. Results suggest that a medium-sized nature preserve at the urban-agricultural interface can provide roosts, commuting corridors, and foraging grounds for different bats. The conservation value of such nature preserves should not be overlooked.

1. Introduction

In grassland ecosystems, agriculture and urbanization are two main processes of anthropogenic disturbances [1]. Both processes remove native vegetation and disrupt native fauna. However, new vegetation is also introduced, such as ornamental trees in urban areas or hedgerows in farmlands, which might provide new resources for local fauna [2,3,4]. Nature preserves at the urban-agricultural interface tend to preserve or restore native vegetation while maintaining some urban or agricultural features on site or adjacent sites, forming a cluster of diverse habitats [5,6,7]. Even though there is a trend of decreasing size in more recently established nature preserves [8], only limited research is available to document the conservation value of these small or medium-sized preserves. Specifically, in the grassland ecosystem, empirical research has mostly focused on the conservation value of small or medium-sized preserves to plants, insects, or birds (e.g., [9,10,11]). Charismatic grassland megafauna such as the bison (Bison bison) or the gray wolf (Canis lupus) are less likely supported by these smaller preserves due to their home range needs [12].

Similar to charismatic grassland megafauna, bats have large home ranges. Most species of bats in the Midwestern United States have a home range much larger than the size of many preserves in the region (>300 ha, see [12,13,14,15]). Presumably, because of the scale mismatch between bat home ranges and preserve sizes [16,17], the conservation value of small or medium-sized nature preserves to North American bats has not been documented except for a few local baseline surveys [18,19]. Urbanization in grassland ecosystems increases structural complexity via trees and manmade structures [20]. Therefore, habitat diversity and its benefits to bats on nature preserves at the urban-agricultural interface should not be overlooked. The current consensus in ecology is that habitat is more than just a vegetation patch [21,22]. In other words, one vegetation habitat can provide multiple functions to fulfill organisms’ biological needs. For bats, a vegetation habitat can serve as roosts, commuting corridors, foraging grounds, and beyond [23]. The mobility of bats due to flight allows them to explore across vegetation habitats for different resources. Thus, nature preserves, even small ones, could provide one or more of these resources for bats within the night or across nights.

Temporal ecology of bats has a long history of indicating how bats use habitats and resources. Within a night, different species are active at different times to partition the niche, forming species-specific diel activity patterns [24]. Variations of a species’ diel activity patterns may occur when bats respond to changes or disturbances in the environment, such as seasonality, anthropogenic light, or noises [25,26,27]. When a species’ diel activity patterns vary across different habitats, it indicates differences in how bats use these habitats and what resources each habitat might provide [28,29,30]. Another aspect of temporal ecology is examining bat activity patterns across nights, months, or even years. At these scales, variation in bat activity may indicate life history events, such as arousals from hibernation [31], the arrival of migrators [32], preparation for lactation [33], or fall swarming and overwintering [34], which correspond to different resources provided by different habitats. In terms of foraging, bats generate unique feeding buzzes that can be recorded and identified to quantify foraging activity patterns [35,36]. A temporal approach can be applied to foraging activity patterns specifically to examine the quality of foraging habitats and food resource availability.

North American bats are experiencing drastic declines due to the spread of white-nose syndrome and the growth of the wind energy industry [37,38]. Those declines reduce the vitality of ecosystem services rendered by bats, such as nuisance insect control in cities [39] and crop pest suppression in agricultural lands [40]. Evidence is needed to demonstrate that nature preserves as a conservation initiative can mitigate the impact of urbanization and agriculture and contribute to the sustainability of bats and their ecosystem services [41,42]. In this case study, we present evidence generated via acoustic monitoring on how a medium-sized nature preserve at the urban-agricultural interface can provide multiple functional habitats to different bat species. The nature preserve investigated is Glacier Creek Preserve in Omaha, Nebraska, United States. The preserve is approximately 2.1 km² with multiple tracts. Prior to the establishment of the preserve, the land was cultivated for corn (Zea mays L.) and soybean [Glycine max (L.)] from 1870 to 1970 [43,44]. In 1970, the first tract, approximately 0.5 km² of land, was seeded to restore the tallgrass prairie. Since then, additional tracts have been added to the preserve. Currently, the preserve includes restored tallgrass prairies, agricultural lands, a few administrative buildings, a spring-fed perennial stream, and riparian zone woodlands [45]. The preserve is surrounded by agricultural land as well as single-family housing developments that have been constructed mostly in the last two decades.

In the summers of 2018–2020, we conducted acoustic surveys on the preserve to investigate its conservation value to bats. Firstly, we aimed to document what bat species were present on the preserve. We hypothesized that all species that had been previously documented in Omaha and its vicinity by previous research [46,47] and/or the local bat rehabilitation facility records would be recorded on the preserve. The specific species list (four-letter species abbreviations used in figures henceforth) is: big brown bat (Eptesicus fuscus, EPFU), eastern red bat (Lasiurus borealis, LABO), hoary bat (Lasiurus cinereus, LACI), silver-haired bat (Lasionycteris noctivagans, LANO), little brown bat (Myotis lucifugus, MYLU), northern long-eared myotis (Myotis septentrionalis. MYSE), evening bat (Nycticeius humeralis, NYHU), tricolored bat (Perimyotis subflavus, PESU), and Mexican free-tailed bat (Tadarida brasiliensis, TABR). Secondly, we compared whether bats used vegetation habitats (woodland edge vs. open [prairie or agriculture]) differently in terms of acoustic activity and foraging activity within and across years. We hypothesized that higher acoustic and foraging activity would be found in the woodland edge than in the open for edge-adapted species [48], such as the big brown bat, the eastern red bat, and the silver-haired bat, whereas the opposite patterns would be found in open-adapted species such as the hoary bat (Lasiurus cinereus). Lastly, we compared whether diel activity patterns (both acoustic activity and foraging activity) differed between vegetation habitats and/or across species. We hypothesized that diel activity patterns would vary among species. Certain species, such as the eastern red bat, were predicted to have an activity peak earlier than other species, such as the hoary bat or the silver-haired bat, based on research conducted in the similar region [24].

2. Materials and Methods

2.1. Study Area, Recording Sites, and Detector Setup

Glacier Creek Preserve is located approximately 20 km northwest of Omaha, Nebraska, United States (latitude 41.34° N, longitude 96.15° W, elevation 330 m, Supplementary Material Figure S1). The preserve has a mission to maintain an ecologically sustainable, landscape-level wildlife preserve focusing on tallgrass prairie and related ecosystems, such as the first-order watershed of Glacier Creek. Detailed information regarding its history, climate, vegetation, soil, and hydrology have been described by previous studies [43,44,45,49]. On the preserve, the most common vegetation habitats are riparian woodlands, restored tallgrass prairies, and agricultural fields that rotated between corn and soybean. Based on vegetation structures, henceforth, woodlands are referred to as edge and prairies and agricultural fields are open. In 2018, we selected four recording sites in the open (two in prairie and two in agricultural fields) and four sites along the edge of woodlands. The open sites were at least 100 m away from any woodland. All recording sites were also at least 100 m away from each other. For 2019, we kept all four edge sites and two prairie open sites. For 2020, we only kept two previous sites, one edge site, and one open prairie site. We selected four new recording sites, two at the edge and two in the open prairie. All sites within the same year were monitored simultaneously.

At each site, a Song Meter SM4BAT FS combined with aSMM-U1 ultrasonic microphone in 2018 and a SMM-U2 ultrasonic microphone in 2019 and 2020 (Wildlife Acoustics Inc., Maynard, MA, USA) was used to record bat acoustic activity. The change of the microphone model between years was one reason for incorporating year as a survey covariate in the statistical analysis. The bat detector microphone was positioned on top of a 3 m pole. The detector was set to record nightly from sunset to sunrise. The detector had a sampling rate of 256 kHz and a minimum signal duration of 1.5 ms. The trigger was set to a minimum frequency of 12 kHz, minimum amplitude of 12 dB, and a 3 s trigger window. The recording file was set to have a maximum file length of 15 s. In 2018 and 2019, acoustic recording was conducted for nine consecutive nights from 27 June to 5 July. In 2020, acoustic recording was conducted from 10 to 26 June, 1 to 11 July, and 21 July to 4 August. These efforts yield a total of 402 detector nights. For each recording night, we calculated its Julian date as a covariate [32,33]. We extracted nightly minimum temperatures from VisualCrossing (https://www.visualcrossing.com/ accessed on 31 January 2023) as a survey covariate as bat activity responds to temperature variations [27].

2.2. Acoustic Analysis

All recordings were saved as WAV files on the SD card and then transferred and stored on external hard drives. We used SonoBat Next Generation (version 30.0 [testing version directly received from the software developer], SonoBat, Arcata, CA, USA) for the automatic acoustic analysis to assign species identification and detect feeding buzzes. We first used the Batch Scrubber to exclude any recording file with only noises. The specific settings in Batch Scrubber were: thorough scan, 15 kHz cutoff, and medium call quality. Next, we used the Batch Classification to process the remaining recording files with a classifier that included only the species listed in our first hypothesis in the introduction as candidate species. We set the acceptable call quality parameter at 0.7, sequence decision parameters at 0.92, and max number of calls to consider per file at 32. We chose these parameters for medium to conservative species identifications and manually verified their accuracy, as described below. We defined each recording that was assigned with species identification as a bat pass for further analysis. Lastly, we used the Batch Buzz Detector to automatically detect potential feeding buzzes in bat passes.

After the automatic acoustic analysis, the first author conducted the manual acoustic analysis alone for consistency. We verified all bat passes that had been automatically identified as little brown bat, northern long-eared bat, and tricolored bat to confirm the identification. For bat passes that had been identified as other species, we randomly selected 100 passes per year per species to confirm the accuracy of the automatic classifier by comparing our recordings with previously collected voucher recordings (see [50] for details of the voucher recording call library). For all passes that had been identified to include feeding buzzes, we visually verified the presence of feeding buzzes based on the description in [35]. From acoustic analysis, we generated the two dependent variables: species-specific bat acoustic activity (number of passes per night per site) and species-specific foraging ratio—number of passes that included feeding buzzes divided by total passes in a night for a species calculated as a proportion [35,36]. We also extracted the timestamp of each bat pass that had been identified to a species (with or without foraging events). We further converted the recording timestamp as a percentage of the night (total minutes elapsed since sunset divided by the total length of the night). The percentage timestamp was used to generate nightly diel activity patterns for statistical analyses as described below. The night length was calculated via the R package “suncalc” [51].

2.3. Statistical Analysis

We conducted all statistical analysis in R version 4.2.1 [52] and used 0.05 as the statistical significance criterion and package ggplot2 for data visualization [53]. To compare bat acoustic activity and foraging ratio between open and edge sites, we first conducted preliminary analysis to explore the importance of survey covariates, which included year, Julian date, and nightly minimum temperature (referred to as temperature henceforth; no correlation between Julian date and temperature was found). For year and temperature, we constructed generalized linear models with year as a categorical independent variable and temperature as a numeric independent variable. We found that year had an effect on bat acoustic activity but not on foraging ratio. Thus, we included year as a random effect for bat acoustic activity. We found that temperature was positively correlated with both dependent variables and thus included it as a numeric covariate. As the effect of Julian date on bat acoustic activity was found to be non-linear in previous research [32,33], we plotted either dependent variable against Julian date and visually examined the relationship. We observed similar patterns between bat acoustic activity and Julian date described by previous research [32,33]. Thus, we chose to construct generalized additive mixed models (GAMM) with Julian date as a smooth term and year as the random effect [54] to further analyze if site (open vs. edge) had an effect on bat acoustic activity instead of a linear model. REML was used for the smoothing parameter estimation method, and GAMMs were constructed via the R package “mgcv” [55]. We did not observe patterns between Julian date and foraging ratio. Thus, we constructed generalized linear models (GLM) to investigate if site (open vs. edge) had an effect on foraging ratio. We modeled bat acoustic activity with the Poisson distribution and foraging ratio with the quasi-binomial distribution.

To examine how bat diel activity patterns might vary/overlap among species between open and edge sites, we followed the nonparametric kernel density estimate procedure via the R package “overlap” [56]. This procedure has been used in similar bat research [28,30]. We converted the timestamp of each recording into a radian scale (sunset/0% of the night as 0 and sunrise/100% of the night as 2π). The diel activity pattern was estimated based on a continuous circular distribution. The optimal bandwidth estimation parameter (kmax) was set to 3, and the bandwidth scalar adjustment parameter (adjust) was set to 1. We chose to calculate the overlap estimator Δ₄ based on previous studies as it was more suitable for large sample sizes [56]. This estimator has a value ranging from 0 (completely non-overlapping) to 1 (completely overlapping). It was used to infer the relative degrees of diel pattern overlapping among species between habitats instead of inferring statistical significance or directly comparing values across different estimations. To visualize the diel activity patterns, we plotted activity kernel density against the time of the night scaled to the percentage of the night. We further standardized the kernel densities using the peak kernel density as 100%. The rest of kernel densities would proportion to the peak. In this way we could better visualize when each species’ activity reached the peak and allow the comparison of peaks across species or habitat type (edge vs. open). For the kernel density estimation and visualization procedure, we first implemented it within a species between habitat types (4 Δ₄ estimated) then across species within a habitat (6 Δ₄ estimated per habitat), generating a graph with two panels (habitat comparison on the top, species comparison at the bottom). We implemented this procedure for bat acoustic activity and foraging activity separately.

3. Results

In 2018, we recorded 28,900 sound files that met identification criteria and identified 18,245 files to species, including 7909 at the edge sites and 10,336 at the open sites. In 2019, we recorded 12,305 sound files that met identification criteria and identified 7562 files to species, including 5046 at the edge sites and 2516 at the open sites. In 2020, we recorded 150,356 sound files that met identification criteria and identified 124,683 files to species, including 74,565 at the edge sites and 50,118 at the open sites. In all years, big brown bats were the most common species, followed by hoary bats, eastern red bats, and silver-haired bats. We recorded both commuting calls and feeding buzzes for all four species every year at both edge and open sites. Additionally, we recorded evening bats and Mexican free-tailed bats every year at both edge and open sites. However, feeding buzzes were only found for these two species occasionally. We recorded tricolored bats only at edge sites every year. In 2018, we recorded little brown bats on one night at one open site and northern long-eared bats on a different night at one edge site. In 2020, we recorded little brown bats twice on two different nights, once at an open site and once at an edge site. We never recorded feeding buzzes for tricolored bats, little brown bats, or northern long-eared bats.

When comparing acoustic activity between edge and open sites, we found significant differences for big brown bats, eastern red bats, and hoary bats (all p < 0.001, Figure 1A). Acoustic activity was 1.2, 4.8, and 2.7 times higher at edge sites than open sites for big brown bats, eastern red bats, and hoary bats, respectively. Acoustic activity did not differ between edge and open sites for silver-haired bats (p = 0.223, Figure 1A). When comparing the foraging ratio between edge and open sites, we found significant differences for eastern red bats and hoary bats (both p < 0.001, Figure 1B). For eastern red bats, the foraging ratio at edge sites was approximately 0.12 ± 0.10 (mean ± standard deviation) as compared to 0.05 ± 0.08 at open sites. For hoary bats, the foraging ratio at edge sites was approximately 0.08 ± 0.08 as compared to 0.04 ± 0.06 at open sites. We did not find differences in the foraging ratio between edge and open sites for big brown bats or silver-haired bats (p = 0.145 and p = 0.223, respectively, Figure 1B).

The big brown bat acoustic activity diel patterns between edge and open sites had the highest level of overlapping (Δ₄ = 0.965, Figure 2A). The eastern red bat acoustic activity diel pattern at the edge sites was more evenly distributed over the night than the diel pattern at the open sites, which peaked approximately 40% into the night (Figure 2A). Both hoary bats and silver-haired bats had more open activity concentrated early in the night and more edge activity spreading later into the night. Hoary bats showed the lowest level of overlapping of acoustic activity diel patterns between edge and open sites (Δ₄ = 0.874, Figure 2A). When comparing acoustic activity diel patterns across species (Figure 2(B-1)), big brown bats and eastern red bats had the lowest level of overlapping, regardless of open (Δ₄ = 0.804, Figure 2(B-2)) or edge sites (Δ₄ = 0.801, Figure 2(B-2)). At both open and edge sites, eastern red bats had the latest activity peak, whereas hoary bats had the earliest activity peak (Figure 2(B-2)).

When examining foraging activity only, big brown bats also had the highest level of overlapping between edge and open sites (Δ₄ = 0.936, Figure 3A). Besides big brown bats, all three other species shown in Figure 3A had more foraging activity later in the night at edge sites than open sites. It is worth noting that eastern red bats had foraging activity spreading throughout the night and lacked a clear peak pattern as compared to the three other species. Silver-haired bats showed the lowest level of overlapping of foraging activity diel patterns between edge and open sites (Δ₄ = 0.787, Figure 3A). When comparing foraging activity diel patterns across species (Figure 3(B-1)), eastern red bats and hoary bats had the lowest level of overlapping at open sites (Δ₄ = 0.728, Figure 3(B-1)), as hoary bat foraging activity was earlier than eastern red bats foraging activity. At the edge sites, big brown bats and eastern red bats had the lowest level of overlapping (Δ₄ = 0.739, Figure 3(B-1)).

4. Discussion

Over three years, we recorded all nine species of bats that occur in eastern Nebraska at Glacier Creek Preserve. Big brown, eastern red, hoary, and silver-haired bats were the most commonly recorded species, which is generally consistent with the results of acoustic research conducted recently in the same region, the northern Great Plains [57,58]. It is also not surprising that we recorded the evening bat regularly as this species is commonly captured in eastern Nebraska [59] and experiencing northward and westward range expansion [60,61]. However, the regular detection of Mexican free-tailed bats is somewhat unexpected as only a few capture or voucher records of the Mexican free-tailed bat were available in Nebraska [62]. The most recent record of a Mexican free-tailed bat in Omaha was an individual found in 2018 and taken to the local wildlife rehabilitation facility. As a long-distance migrant, Nebraska is considered part of the exploring zone for the Mexican free-tailed bat [62]. It seems plausible that individuals occasionally fly into the region undetected. Our acoustic recordings could represent such exploring individuals, or perhaps Mexican free-tailed bats have expanded their range northward in the central United States in recent years. This species has recently expanded its range northward in the southeastern United States [63] and has also recently been documented in western Canada [64]. As it is difficult to use mist-netting or other techniques to live-capture high-flying species [65,66], we encourage considering novel survey methods such as environmental DNA [67] to determine whether it is also expanding northward in the Great Plains. In contrast to those commonly recorded species, the little brown, northern long-eared, and tricolored bats have been experiencing drastic population decline due to white-nose syndrome [37,38]. The northern long-eared bat was listed as endangered by the U.S. Fish and Wildlife Service in 2022 [68], while the other two species are currently being reviewed for listing. The year 2018, which is the last year to record the northern long-eared bat in our area, is consistent with the local wildlife rehabilitation admission records and other studies in the region [69]. Nevertheless, the presence of all nine species at the preserve over the 3-year period of this study indicates that the preserve provides habitat for bats.

To better understand what habitats bats utilized, we compared whether bats were more active along the woodland edge or in the open over grasslands and agricultural fields. In the existing literature, it is shown that big brown, eastern red, and silver-haired bats are generally considered edge-adapted species, whereas the hoary bat is an open-adapted species [48]. In the grassland ecosystem, however, woodland edges become even more important due to the lack of physical and structural complexity in the environment [18,49]. Our results support that both the big brown bat and the eastern red bat are edge-adapted species, being more active at the edge sites than at the open sites. However, we also found that the hoary bat was more active at the edge sites than at the open sites and that no difference was found between sites for the silver-haired bat. We speculated that such contradictory results might be explained using prey availability differences. For example, when insect availability was lower in open habitats or when there was no difference among habitats [57,70], bats might respond to physical structures, which also provided roosting habitats. However, when insect availability varied, and more insects were available in open habitats, some species might use open areas more [35]. Thus, it was important to examine the actual foraging activity to gain a better understanding of how bats used the preserve.

Consistent with our prediction, the eastern red bat foraged more often at edge sites than at open sites, indicating it is an edge-adapted forager [48]. This result also suggests that woodlands on the preserve served as foraging habitats for the eastern red bat. Contradictory to our predictions, the big brown bat and the silver-haired bat did not forage more often at the edge sites. These bats might be more responsive to variations of prey availability and less responsive to physical habitat structures. Aerial insect abundance in agricultural landscapes shows non-linear spatial-temporal patterns and leads to fluctuations in food availability [71,72]. When insects at the preserve were less abundant, these bats might explore other areas. Additionally, both species have been found to switch roosts frequently [73,74]. It is probable that the preserve was one of several roosting areas being visited and used by these bats periodically. Our results on the hoary bat were also contradictory to the original prediction. We recorded hoary bat foraging activities more often at the edge sites than at the open sites. However, this result is generally consistent with the movement biology of the hoary bat. For a high-flying species [75,76], foraging events of the hoary bat are likely to be out of the detection range of our bat detectors. The open grassland is more likely to function as the commuting space for the hoary bat. On the other hand, woodlands can be used as either day roosts or night roosts for tree-dwelling hoary bats [77]. Thus, we likely recorded hoary bats as they departed from and arrived at roosting areas.

The diel activity pattern of each species may provide additional insight into how bats used different vegetation habitats on the preserve. Acoustic activities of big brown bats and eastern red bats reached the peak slightly earlier at the edge sites than at the open sites. In contrast, hoary and silver-haired bat acoustic activities reached the peak slightly earlier at open sites. Such differences might indicate that hoary and silver-haired bats have day roosts somewhere else and visit the preserve for foraging or night roosts. Night roosts are important habitats for bats to rest, feed/digest, and socialize [78]. A wide range of structures can serve as night roosts for bats with more flexibility than in day roost selection [79,80]. Diverse structures such as trees and administrative buildings on the preserve and in the urban-agricultural interface provide numerous night roosting opportunities. Night roosts are usually in close proximity to foraging habitats [81,82]. The diel patterns of foraging activity for the eastern red, hoary, and silver-haired bats all showed that relatively high levels of foraging events remained throughout the night after the activity peak. These results suggest that the species mentioned above might fly frequently along the woodland edges to forage between night roosting events. Furthermore, night roosts have the potential to become day roosts, providing additional resources for adverse environmental changes [83,84].

One interesting finding in the diel patterns for both acoustic and foraging activities is that the hoary bat activity tended to reach the peak first, followed by big brown and silver-haired bats. The eastern red bat activity reached its peak late, almost near the middle of the night. These patterns, even though indicating temporal niche partitioning among these species, are contradictory to previous knowledge generated from the Midwest grassland-dominated ecosystem [24]. In previous research, the eastern red bat was the first one to reach its activity peak, whereas the hoary bat was the last one over the night. We suspect that such contradictions might be due to the differences in local environments. The previous research was conducted at multiple sites near water sources [24]. On Glacier Creek Preserve, even though there is water available, the small stream is not large enough to serve as a waterbody for drinking or influencing bats’ distribution on the landscape [85]. Rather, our temporal niche partitioning results are similar to those of research conducted in the southeastern United States [30], where water might not be a limiting factor in concentrating bats on the landscape [50]. Additionally, the previous research was conducted in more natural environments, whereas ours is conducted at the urban-agricultural interface. Urbanization has been known to alter bat temporal niche partitioning [28,86]. Another plausible explanation is that the fieldwork time window in the previous research covered more months within a year than ours. Seasonal variations of diel activity patterns due to changing resources have been found in several studies [25,29]. The spatial scale difference between our study and previous research might also explain the difference. More research should be conducted on diel patterns of bats at broader spatial (e.g., state- or range-wide) and temporal scales (e.g., year-round recording) to better understand temporal niche partitioning [30].

Our research shows that Glacier Creek Preserve can provide multiple functional habitats for different species of bats. Even though the size of the preserve is generally smaller than most bats’ home ranges, it can provide a place for bats to roost and forage. Furthermore, woodlands on the preserve, along with other urban trees, might form a tree canopy network for certain species to commute. For example, we recorded the tricolored bat at some edge sites every year. Multiple studies have demonstrated the importance of forests with dense canopy to this species [87,88,89,90]. At the broader spatial scale, landscape connectivity in agriculture-dominated or urban landscapes has also been well-documented to be important for bat conservation [91,92,93]. Even though one preserve might not cover a bat’s entire home range, it may function as a stepping stone for movements in less preferable landscapes [8,94]. Lastly, like many locally owned and managed small or medium-sized nature preserves, Glacier Creek Preserve welcomes thousands of visitors from the local community. Bats’ presence on the preserve allows an opportunity for the public to observe and learn more about bats, adding social value to the preserve for the conservation of less charismatic species [8,95]. We argue that the value of small- or medium-sized nature preserves for bat conservation should not be overlooked. Our results suggest that maintaining habitat functional diversity on nature preserves is the key to support a diverse bat community. Specially, riparian woodlands in grassland ecosystems can provide essential roosting structures; whereas restored grasslands can be the source of diverse insects as prey. Recognizing the full conservation potential of small or medium-sized nature preserves can benefit the long-term sustainability of both the preserve and local biodiversity.