CHARACTERISTICS OF WINTER BAT AGGREGATIONS IN POLAND: A REVIEW

The distribution of natural wintering roosts of bats in Poland is not homogenous, they are found almost exclusively in the south of Poland. Manmade underground shelters, on the other hand, are found across the country, including areas outside the range of natural caves. Based on literature data and my own observations, the size and species composition of winter bat aggregations occupying natural and artificial roosts were compared. The data in the records of both were published and collected between 1992 and 2019. The starting date of the range has been assumed due to the well-established formula of nationwide bat monitoring, which started in 1988, while the end date was based on the availability of publications. I’m arbitrarily adopted 50 individuals as the minimum number of bats in a certain wintering site, found on a minimum of one occasion during the study period. In total, I collected data on 128 wintering sites: 42 natural and 86 man-made. The overall size of the analysed aggregations a minimum of 88,800 individuals, with manmade roosting sites accounting for a minimum of 77,600 individuals, and natural sites almost 12,200 individuals. The most abundant bat aggregations were recorded in manmade underground sites - up to 38,594 individuals - while in natural sites there were only up to 2,900 individuals. However, the average aggregation size did not differ greatly between the two types of shelter (169 vs. 156, F = 3.368, df = 1,128; p = 0.069). In contrast, the number of species wintering in natural shelters is higher than in manmade ones (8 vs. 6 species, F


Introduction
The strategy for surviving the winter food shortage, especially the shortage of invertebrates, which are an important part of predators' diet, is to migrate or hibernate [Heldmaier & Klingenspor 2010]. In the temperate zone, insectivorous bats are subject to periodic fluctuations in food availability [Geiser 2011]. The start and end of the hibernation period is highly species specific, and it depends on the type of food preferred [Speakman & Rowland 1999]. Moreover, microclimatic parameters are important: some bat species require stable conditions during hibernation (the greater mouseeared bat Myotis myotis, the lesser horseshoe bat Rhinolophus hipposideros [Nagel & Nagel 1991]), while others winter in the near-entrance zone, where conditions are largely determined by external parameters (Plecotus sp., Barbastella barbastellus). Typical hibernation sites in nature are underground cavities -caves [Furey & Racey 2016]. The appearance of natural caves is strictly conditioned by the occurrence of karstic (karstic caves), sulphate or sandstone rocks (tectonic). In Poland, the presence of caves is limited to the south of the country: the Kraków-Wieluń Upland, the Sudetes, the Carpathian and the Świętokrzyskie Mountains [PGI-NRI 2022]. In addition to natural cavities, there are a number of other underground shelters of anthropogenic origin -bunkers, cellars, mines or even factories -which can provide wintering sites for bats [Wołoszyn 1996]. Contrary to most natural shelters, the vast of anthropogenic undergrounds have more than one entrance, frequently with additional ventilation openings. As a result, artificial underground sites often have a much larger dynamic zone range than caves [Lesiński 1996]. Differences in microclimates also have unequal attractiveness to bats: individual species differ in their thermopreferendum [Nagel & Nagel 1991]. In fact, a number of sites of anthropogenic origin gather large aggregations of wintering bats. Certain sites, such as the MRU, represent important wintering refuges of not only national, but also European importance [Mitchell-Jones 2016].
In 1988, the first census of wintering bats in Poland was conducted, called "the Decade" (DSN): monitoring was to take place during the first 10 days of February . A summary of the first 5 years of monitoring resulted in a variety of important information -it identified important wintering locations, and comparison with historical data allowed the identification of trends, both locally and nationally [Wołoszyn 1996]. The action was initially based on the amateur chiropterology network, and in the course of time, universities and other associations joined in, with a database being established at CIC ISEZ PAN in Krakow (CIC -Chiropterological Information Center). After 1992, most regional chiropterological groups were functioning independently of the CIC.
While discoveries of new natural sites have been rare in recent years are rare [Nowak & Grzywiński 2017, authors data], many man-made underground sites are constantly being explored and impressive hibernation aggregations are being found [Wojtaszyn et al. 2013a]. This article summarizes the available information on bat wintering roosts in Poland, and it provides a preliminary comparison of the fauna of natural and artificial roosts.

Materials and methods
This overview uses data from my own winter bat monitoring (Kraków-Wieluń Upland) and materials available in the literature (see Annex). The compilation includes sites with more than 50 wintering individuals recording during the period from 1992 to 2019. I arbitrarily adopted 50 individuals as the minimum number of bats in a certain wintering site, found on a minimum of one occasion during the study period. The start date of the range was assumed due to the well-established formula of nationwide bat monitoring, which started in 1988 (important factor: correctness of species identification), while the end date was based on the availability of publications. The underground sites were divided according to their genesis into: natural -caves (karst, flysh); manmade: wartime (bunkers, fortress), and non-wartime (cellars, mines, drainage systems). Some natural underground sites have been transformed during mining activities. Hence, the appropriate category was determined by the proportion of remaining natural cavities: if this was less than half it was considered to be an artificial site [Bochotnica: Kowalski & Dróżdż 2002], while if it was over half it was regarded as a natural one [Szachownica cave: Ignaczak 2017].
For every site, the bat fauna was described by: i) maximum abundance of each species in the surveyed year range (indeterminate excluded); ii) maximum abundance of all species (including indeterminate), iii) number of species found. The species of most individuals were determined by external diagnostic features (bats were not removed from the walls). For hard-to-determine species such as: M. alcathoe/M. brandtii and M. mystacinus, they were treated together as Mbra/Mmys, while all Pipistrellus were categorised as Pipistrellus sp. Moreover, each underground object was described based on: i) its origin: natural (caves) on manmade (shafts, mines, forts, bunkers, cellars); ii) its latitude and longitude; and iii) bibliographic data (author, journal/source).

Statistical analyses
Differences in maximum bat abundance and the number of wintering species between natural and artificial wintering sites were tested by one-way ANOVA after log-normal transformation. In addition, cluster analysis (Ward's method), was used to visualize the similarity of the proportion of the fauna of wintering bats. To determine the similarities of the diets between bat species, cluster analysis (Ward's method) with bootstrapping was performed using the R package. The distances of objects were calculated in Euclidean space, using the average linkage method.
Among the natural sites, the largest wintering one held slightly more than 2,000 bats (Szachownica cave), while only 2 caves held between 2,000 and 1,000 bats, and the remaining 39 caves housed less than 1,000. In total, artificial wintering roosts gather more than 6 times the number of wintering bats (76,603 individuals) than natural underground sites (12,197 individuals). Nearly half of the bats hibernated in wartime undergrounds sites dating back to WWII, while considerably lower numbers were found in fortresses from before WWI -only 11%. On the other hand, among the manmads non-war structures, drainage systems were the most important, followed by cellars, and finally, with a small contribution, mines. In contrast, the smallest proportion of bats overwintered in natural sites, the vast majority in karst caves (Fig. 2).
In the underground sites used in the analysis, 17 bat species were found to be hibernating. The most numerous species was M. myotis, while less abundant were M. nattereri, M. daubentonii and B. barbastellus. Small percentages were recorded for R. hipposideros, P. auritus, M. brandtii/mystacinus, M. emarginatus and P. pipistrellus. The species M. dasycneme, M. bechsteinii, E. serotinus, E. nilssonii, P. austriacus, and those known from single individuals -N. noctula and R. ferumequinum -were found incidentally. The number of bat species found wintering in the natural underground sites was 7 on average, and this was significantly higher than the number of species wintering in manmade sites, with an average of > 5 species (one-way ANOVA: F = 9.79, df = 1, 128 DF, p-value: 0.0022) (Fig. 3).
The one species found only in anthropogenic shelters was Pipistrellus sp. (most probably P. pipistrellus), but it was present in low proportions. In manmade shelters, the highest proportion of bats among species M. daubentonii, M. myotis, M. nattereri, P. austriacus, E. serotinus and B. barbastellus. In the caves, on the other hand, there was a clear predominance of M. brandtii/mystacinus, M. bechsteinii, P. auritus and M. dasycneme. Several bat species were recorded almost exclusively in caves: R. hipposideros, M. emarginatus, E. nilssonii and N. noctula, while only single individuals of R. hipposideros and M. emarginathus were found in manmade underground sites. The species composition of hibernating bats therefore differentiates between natural and manmade roosts (Fig. 4).
Abandoned military facilities with their extensive underground corridors are therefore very suitable roosts for the wintering of many bat species [Voigt et al. 2014]. Most of the bat wintering sites used in the analysis have been known since the first winter counts, but recent years have brought a very impressive discovery of new winter aggregations, ranging from several hundred to even several thousand: the drainage systems of Koszalin [2004 Wojtaszyn et al. 2013a], the drainage systems of Piła [2008; and the bunker in Stargard Szczeciński [2012]. The only natural site is the Niedźwiedzia cave in Kletno, where the discovery of new sections has also resulted in an increase in bat counts from 240 to more than 1,100 individuals .
Of the manmade underground sites, however, the urban drainage systems gather bat populations in their thousands. While similar sites in central Europe have already been mentioned [Godlevsky 2000], initially this information was rare and the abundances recorded at the time were not as high [Grzywiński & Kmiecik 2003]. Interestingly, only a small part of the drainage system of Poznań provided a wintering site for bats comparable to the largest above-ground sites located in the Poznań Fortresses [Grzywiński & Kmiecik 2003].
The monitoring of drainage systems is neither simple nor secure: there is the potential presence of carbon dioxide, methane or even hydrogen sulphide, and there are no plans or maps for most of them. However, new large wintering sites are found year after year: currently, record abundances have been recorded mainly in the cities in northern Poland: in Olsztyn, with 3412 ind. [Wojtaszyn et al. 2013a], in Piła, with 3403 ind. [Wojtaszyn et al. 2013a], and in Koszalin, with 955 ind. . It can therefore be hypothesised that these systems, considering their extent, represent some of the most important wintering habitats in areas where natural shelters are lacking.
On the other hand, the bat fauna found up to now in these wintering sites is quite sparse in species: M. nattereri dominate with up to 97% [Grzywiński & Kmiecik 2003;Wojtaszyn et al. 2008, Wojtaszyn et al. 2013a] and M. daubentonii account for up to 85% ], while other species, P. auritus and B. barbastellus, were was found incidentally.
Cellars, often the remains of factories, also gather large aggregations of wintering bats [Police (proper name of location, not "milicia"), Dzięgielewska et al. 2007], but they are less numerous than the other categories. The least numerous among the manmade wintering sites are mines, and they do not gather large groups of bats. The reason for their lower importance as wintering habitats may be that they are mostly located in the mountains and highlands of southern Poland, where there is also an abundance of natural shelters [Kowalski 1953, Piksa & Nowak, 2013, Postawa & Zygmunt, 2000Szkudlarek et al. 2008;Furmankiewicz et al. 2016].
On the other hand, among the sites of natural origin, only three of them provide shelter for more than one thousand individuals: Szachownica cave [Ignaczak 2017], Niedźwiedzia cave in Kletno  and Studnisko cave (author data), while the next in size do not exceed a few hundred individuals. Low abundance, on the other hand, may be the result of the much higher density of natural shelters than of manmade underground sites: only the upland and Carpathians have to date inventoried over 8,000 natural underground sites of various types [PGI-NRI 2022]. Despite lower winter aggregation, caves are one of the most important roosts -they are target wintering sites for bats from southern and central Poland [Wojtaszyn et al. 2019]. The abundance of wintering bats in PL is probably several times higher than the value obtained, because: (i) bats overwinter in many places such as cellars or wells (small numbers, but objects are numerous); and (ii) many wintering sites are unknown to us, as demonstrated by discoveries, for example, in Niedźwiedzia cave in Kletno, or other new caves with wintering aggregations of more than 100 individuals (Niedźwiedzia Górna cave -author data, Twarda cave [Nowak & Grzywiński 2017]). The results of winter censuses of bats confirm the trend observed for more than two decades of increasing numbers of these mammals in Poland (start 1999, citations). However, the dynamics of change are species specific [Piksa & Nowak 2013;Lesiński et al. 2011;Ignaczak 2017;Bernard et al. 2019], and numbers also vary at different times and even between different regions [Gottfried et al. 2019]. Despite a continuous increase in numbers, the population of wintering bats has not returned to the numbers it was before the collapse of the 1960s and 1970s, when abundances were much higher [Kowalski 1953].
The species composition of wintering bats clearly differentiates natural from artificial roosts, while only slightly between artificial shelters: war and no-war function. This is probably the result of a geographical trend: in the south, especially in areas rich in natural underground sites, there are many more species than in the north of Poland [Sachanowicz & Ciechanowski 2006]. On the other hand, most manmade underground sites are characterized by a dynamic microclimate (or its dominance): artificial shelters were built for human use and in most of them ventilation was important (hence the predominance of a dynamic microclimate). Natural objects, on the other hand, have an unpredictable pattern, and static microclimates predominate for the most part, which is preferable for most temperate bat species [Nagel & Nagel 1991].
Currently, some of these sites are protected as reserves, Natura 2000 areas, or even documentary sites. However, many are still at risk, especially manmade ones: several constructions have become utilized commercially, abandoned or demolished. Restoration works are starting in some of them in order to save their historical values, but this could present serious threats for wintering bats [see: Wojtaszyn et al. 2015]. Drainage systems appear to be especially difficult to protect -their function conflicts with bat conservation, and renovations can result in the loss of microhabitats and/or a change in microclimate.