Introduction

The western barbastelle Barbastella barbastellus (Schreber 1774) is a species known to hibernate in relatively cold temperatures ranging from − 3° to 6.5°1,2,3. In well-isolated deep underground systems such as Nietoperek Natura 2000 site PLH080003 barbastelles tend to occupy colder parts, often near entry points3,4,5. Their preference for cold is further apparent in that they are frequently found in less isolated hibernacula such as bunkers, cellars6,7 and even trees8. This adaptation allows the species to be among the first bats in central Europe to finish hibernation, usually leaving wintering sites mid-march5. Early spring poses a severe challenge for temperate zone bats. After months of dependence on energy reserves gathered before hibernation, they urgently need to find energy sources (prey). Simultaneously, insect abundance in early spring is still limited (life cycle, winter dormancy maintained by photoperiod and temperature)9,10,11. While most European species depend on prolonged hibernation until prey becomes more numerous barbastelles are already active. This behavioural adaptation is interesting because barbastelles are highly specialized moth predators rather than opportunistic foragers2,12,13. Evolutionary armed barbastelles with sophisticated echolocation, utilizing low amplitude calls14, allows them to bypass the acoustic defence of owlet moths Noctuidae—an adaptation based on ultrasound hearing combined with evasive manoeuvres triggered by high-intensity echolocation calls15,16. This trait makes barbastelles an excellent aerial hawking animal whose diet consists of over 90% moths12,13. The only other European species: that is known to effectively hunt owlet moths are long-eared bats—Plecotus sp.17,18, which we also included in our analysis. In this case, Plecotus sp. uses passive hearing17,19.

To consider why a highly selective bat leaves hibernation so early when its prey availability is restricted, we explored the available data on barbastelle prey12,13. We discovered that in the early spring season, most of its prey consists of the genus Orthosia—a highly specialized genus of noctuid moths, which includes several large species, all of which, as adults, feed on the pollen and nectar of Salix flowers at night20. Willow trees and shrubs are mainly formed by the goat willow Salix caprea (Linné, 1753). Since this species forms numerous hybrids in Europe with other species of the genus Salix, we will use the name Salix sp. Dependent on willows—a primary producer of nectar and pollen in early spring21,22, Orthosia moths are active earlier in the season than most other owlet moths23. A few other nocturnal moths feed on Salix flowers early in the season, such as Eupsilla, Conistra and Agrochola, as well as some species of Geometridae and Pterophoridae, but the most prominent genus is Orthosia20,24,25. Several other insect taxa visit early blossoming trees, such as bees (Hymenoptera, Apidae) and hoverflies (Diptera, Syrphidae). However, none of those groups represents a viable food source for the moth specialist barbastelle.

We designed a study based on pairwise deployment of acoustic detectors near hibernation sites to investigate this potential phenomenon. We hypothesize that in early spring, barbastelle activity will be significantly higher around flowering willows when compared to sites with tree species that flower later in the season. Additionally, we predict the activity will be large at the beginning of the vegetative season and procedurally decrease as other foraging options develop.

The western barbastelle is one of Europe's most threatened tree-dwelling bats26. it is included in Annexes II and IV of the EU Habitats Directive and is classified as Near Threatened in Europe27. We designed an experiment to better understand this species' ecology in a crucial period of the year and to improve conservation measures.

Materials and methods

Study areas

Nietoperek site (MRU)

Natura 2000 site PLH080003 “Nietoperek” lies in Lubuskie voivodeship in western Poland (central point: E 15.480600, N 52.394400). It spreads over 7377.37 ha above a 32-km underground system of tunnels (“Międzyrzecki Rejon Umocniony”, MRU) and on ground fortifications built in the 30 s by Germany, as a part of a larger defensive front “Ostwall” or “Festungsfront im Oder-Warthe Bogen”3,28. Currently abandoned by the military it is a critical hibernation site for bats in central Europe. With a record count exceeding 39,000 individuals of 12 species including over 600 western barbastelles, it is among the EU's top 10 largest hibernation sites3,28. The landscape above the tunnels mainly consists of young monocultures of Scots pine Pinus sylvestris (Linné, 1753) mixed with isolated patches of riparian woodlands with ash Fraxinus sp. and alder Alnus sp. Remaining mostly in depressions and along shorelines of lakes and river areas with restricted access for heavy timber harvesting machinery. This habitat covers approximately half of the protected area (46.19%, ca. 3400 ha). The remaining half is a rural landscape of fields, shrubs and woodlots with willows within cropland. The vegetation period in this part of Poland lasts for approximately 230 days29.

Schrony Brzeskiego Rejonu Umocnionego site (BRU)

Natura 2000 site PLH200014 “Schrony Brzeskiego Rejonu Umocnionego” (“Shelters of the Brest Fortified Line”, BRU) located in Podlaskie voivodeship, eastern Poland (central point: E 22.994400, N 52.392100) cover ca. 126 ha. In this area, there are 15 iron-concrete bunkers approximately 2 km west of the village of Anusin and 4 km to the east near Moszczona Królewska. The bunkers were built in the 1940s by the Soviet Union as part of a defensive line against the expected invasion of Nazi Germany (the so-called Molotov Line). They are a hibernation site for a winter population of Barbastella barbastellus (approximately 400 individuals) and several other bat species. It is an essential place for bat hibernation in this part of the country. The characteristic features of the landscape are the elevations of the postglacial moraines and river bends. Local forest habitats are sandy and consist mainly of Scots pine. Discontinuous pine forests surround the bunkers, and there are also a few fragments of riparian vegetation, including willows. There are small villages and agricultural areas in the vicinity. The vegetation period in this part of Poland lasts for approximately 215 days29.

Bioacoustics

We used a combination of four Song Meter SM4 Wildlife Audio Recorders (Wildlife Acoustics, USA) and six Audiomoths (Open Acoustic Devices, UK). The SM4s were deployed in two pairs in the “Nietoperek” site for 25 nights beginning on March 10th 2022. The Audiomoths were deployed in the same paired design in “Schrony Brzeskiego Rejonu Umocnionego” for 6–13 nights starting March 18th 2022. Detectors were placed after the first blossoming willows were observed in each study area. A pair consisted of one detector mounted on a willow tree with a microphone pointing upwards, and the other one installed similarly on a non-willow tree, i.e. non-early blossoming species, at least 200 m30 from and no more than 300 m away from the paired device to ensure the probability of recording independent data while maintaining similar geographic features. Pairs within a single study area were at least 500 m apart and no more than 3 km away (Table 1). The landscape in both study areas was best described as agricultural land with scrub bordering managed forest. The distance between the study areas was ca. 550 km. The geographic distance between both sites resulted in a phenological shift between them. All detectors were programmed to record from dusk till dawn, adjusted to their geographical location. The measure of activity was the number of bat passes. The term "bat pass" means a single flight of a bat understood as one well-defined series of echolocation signals. We take two or more single echolocation pulses as a series of signals. The series are separated by a minimum of 1 s between one and the next31. Recordings were manually identified to species, where possible and otherwise to genus, in BatSound 4.2 software (Pettersson. Elektronik AB, Sweden).

Table 1 Exact locations of detector deployment.

Statistical analysis

We used a linear mixed effects modelling approach to determine (i) whether bats selected willow tree species sites over non-willow tree species sites, and (ii) explore the importance of willow tree species sites for bats over survey nights. We first built a model that included western barbastelle calls as the response variable, and then built a separate model for ‘other’ recorded bat calls as the response variable, using the nlme R package32. The data was not transformed prior to model building and normality was not tested for due to the small sample size. The distribution of the residuals of each fitted model were visually checked for normality using a Q-Q plot; residuals did not deviate from normal. Homogeneity of variance was assessed using residual plots. The spread of residuals was even. Detector position (willow tree species site or non-willow tree species site) and survey night were included as fixed effects, and paired site included as a random effect. Models were fitted using Restricted Maximum Likelihood (REML). Model terms were tested for importance using ANOVA. All statistical analyses were performed in R 4.1.2 and RStudio 1.0.14333.

Research ethics

Our research relied on non-invasive passive recording devices which do not require specific institutional animal care approval.

Results

A total of 160 nights of bat recordings were made consisting of 80 nights at the combined five paired sites. Of these 80 night recording periods there were nine nights in which no bats were recorded in either the willow or control treatments. For the remaining 71 nights bat echolocation calls were recorded. When assessing the difference in the western barbastelle calls between willow and non-willow tree species sites our model identified a significant difference. The number of recorded calls at willow sites was significantly greater when compared to the number of recorded calls at non-willow sites (F(1,152) = 40.54, p < 0.001) (Fig. 1, Table 2). No significant difference was identified between willow and non-willow tree species sites when ‘other’ bat calls were analysed (F(1,152) = 2.72, p = 0.1) (Fig. 2). When assessing the change in the western barbastelle occurrence at willow tree species sites over survey nights we found a significant decrease in recorded western barbastelle calls from first survey night (F(1,152) = 13.55, p < 0.001). The number of western barbastelle calls were relatively high at the start of the survey period with incremental nightly decreases (Fig. 1). In contrast, when assessing the change in ‘other’ bat calls at willow sites over time we found borderline significance and an increase in recorded calls over time (F(1,152) = 3.90, p = 0.0501) (Fig. 2).

Figure 1
figure 1

Estimated mean averages of the western barbastelle (Barbastella barbastellus) calls in willow tree species sites (red) and non-willow tree species sites (black) over time, derived from linear mixed effects regression models. Dashed lines of corresponding colour represent trend lines for given categories. Error bars indicate 95% confidence intervals.

Table 2 Results of the linear mixed effects models used to test the effect of position (willow or non-willow tree), night and position over survey nights.
Figure 2
figure 2

Estimated mean averages of ‘other’ bat calls in willow tree species sites (red) and non-willow tree species sites (black) over time, derived from linear mixed effects regression models. Dashed lines of corresponding colour represent trend lines for given categories. Error bars indicate 95% confidence intervals.

Discussion

The relationship we found between an early-flowering tree species and a specialized bat species (the moth foraging specialist western barbastelle) is an example of a chain of dependencies formed between different groups of organisms. In early spring, Salix sp. is a key genus on which many species depend, primarily as a food resource and secondly as it attracts prey for insect-eating species such as birds and bats20,34. It can therefore be assumed that the increased activity of barbastelle around flowering willows is associated with the presence of potential food. The willow moths, like Orthosia (Noctuidae), are an established component of the barbastelles diet and have been found in their guano collected in early April13. Orthosia moths hibernate in adult form, ready for activity and end their dormant period early in spring. Thus, they appear early on the flowers of Salix sp., on which they feed. The western barbastelle is a species whose hibernation period ends earlier than that of other Palearctic bats5 and is a moth specialist adapted to capture Noctuidae14.

Salix caprea, like other Palearctic willow species—with which it often hybridizes—is pollinated by wind and diurnal and nocturnal insects20. The diurnal insects are mainly Hymenoptera and Diptera species, and the nocturnal insects are mostly Noctuidae moths. The latter, especially the genus Orthosia is specialised in feeding on Salix sp., and is therefore, also an important pollinator for the tree20,24,25.

The relationship between the willow and the bat is probably one-sided because bats benefit from feeding on moths attracted by willows. At the same time, this phenomenon is likely indifferent to the tree's survival. The loss of pollinating insects is negligible, considering the high prevalence of Salix trees and shrubs and the relatively low density of western barbastelles in the landscape. In this case, it would be a simple relationship where flowers provide food for a specialized insect species, which in turn is early spring prey for the western barbastelle (commensalism).

The western barbastelle leaves winter roosts early. Although the available literature is very sparse5, this species likely ends its hibernation in March, generally. Moreover, it is already known that during hibernation, bats break their torpor, drink water, and also forage, replenishing their energy reserves35. Therefore, it is unclear whether the animals we recorded were feeding over Salix sp. trees during the final stage of hibernation or immediately afterwards. However, food availability is considered one of the major factors influencing the spatial–temporal patterns of bat activity36. We found that the period of occurrence of barbastelles in the vicinity of flowering willows was relatively short. After 1 week, their activity decreased, and the barbastelles mostly disappeared (Fig. 1). Therefore, we conclude that barbastelles at that time used temporary roosts, which may be the neighbouring hibernation sites. The areas surrounding both study plots were covered by a forest composed mainly of young (x < 50 years) pines (https://www.bdl.lasy.gov.pl/portal/mapy). Such a habitat does not offer good roosting opportunity for the barbastelle, which in summer uses cracks in broken trunks or space behind protruding bark37,38,39 and in some cases also buildings, especially wooden ones40,41,42. Regardless, if we take the results of our research as evidence, spring foraging by willows will provide a temporary but critical supply of prey—allowing fast regeneration just after hibernation.

It is recognized that diversity in a rural landscape is an integral part of biological diversity, also in an area of Europe where agricultural activity has a long history and the vast majority has been permanently transformed. For this reason, efforts are made to protect and increase diversity for economic and biological reasons. The list of agricultural landscape elements important for maintaining diversity is still expanding43,44,45,46,47, and bats are one of the subjects of interest48,49,50,51,52,53, both as an element of the ecosystem and as a factor regulating the number of insects54,55,56,57,58,59,60. Willows are a frequent element of Europe's rural landscape (which includes agricultural areas often mixed with tree stands and fragmented woodlands) appearing both as large trees and as shrubs, especially in meadows and wetlands. In many regions with a temperate climate, they are part of the traditional landscape of pastures and meadows, which is an important factor in the conservation of bats61. Our research shows that willows can play an important role in the ecosystem for some species of bats. This is evidence for the protection of moth specialist bats such as western barbastelles and possibly long eared bats Plecotus sp., as willow tree stands growing in the area where hibernation sites of these species are located should be treated as an important element for local bat populations, and thus should be protected against removal or degradation. We also propose that complementary planting of willow in the vicinity of bat wintering grounds will increase the potential prey availability at a critical time of the year.