Bioacoustic evidence for a continuous summer presence of the greater noctule bat, Nyctalus lasiopterus, in the Italian Alps

The greater noctule bat, Nyctalus lasiopterus (Schreber, 1780), is one of the rarest bat species in Europe. Data on its distribution in the central and western parts of the continent are largely lacking. For the first time, N. lasiopterus was observed in South Tyrol (Northern Italy) using bat detectors in the summer of 2016 and in the following years until 2021. The five sites where the species was detected are located in a limited area of the Adige Valley, which is characterised by low elevation and a mild climate. Multiple recordings indicated a continuous summer presence of N. lasiopterus on the southern side of the Alps. Additionally, we present insights on echolocation and species identification, including statistically significant differences in the call metrics of N. lasiopterus compared to Tadarida teniotis. These results are important at a European level given that data is lacking on the populations of N. lasiopterus in Europe outside the Mediterranean.


Introduction
Knowledge on the distribution of the greater noctule bat, Nyctalus lasiopterus (Schreber, 1780), in Europe remains scarce. Populations of this species are scattered and have low densities (Dietz and Kiefer 2020), and vagrant individuals have frequently been recorded outside the known distribution range (Hutterer et al. 2005). This has resulted in an increase in the lack of clarity regarding the distribution of the species. The distribution range of N. lasiopterus spans the Iberian Peninsula, France, Eastern and Central Europe, North Africa and Anatolia, and reaches its eastern limit in the Caucasus region and the Ural (Dietz and Kiefer 2020;Snit'ko and Snit'ko 2021).
In the Mediterranean Region, the species occurs mainly in the Iberian and Balkan Peninsulas, Morocco and Corsica. All the Italian regions are within the known distribution range of N. lasiopterus (Lanza 1959). The most up-to-date data for the Italian Alps include few records from the Friuli Venezia Giulia region and the province of Trento (Lapini et al. 2014;Scaravelli and Priori 2018). However, these records represent occasional sightings along the migratory routes and do not identify individuals from resident populations (Agnelli and Lapini 2023).
Regarding the Alpine region, there are no reports of N. lasiopterus in Austria (Spitzenberger and Bauer 2001), whereas the species has only rarely been reported in Switzerland (Zingg and Aellen 1995). Most of the recent and verified records originate from autumn catches at bird ringing stations on Alpine passes (Vallotton 2007;Mazenauer 2015). In Germany, only three records of the species exist, all of which are from the state of Bavaria (Dietz and Kiefer 2020).
Nyctalus lasiopterus has been reported from sea level to mid-mountain areas up to an altitude of 1,350 m, even though it can occasionally be found up to 2,000 m during migratory movements (Col de Bretolet pass in the French-Swiss Alps; Arlettaz 1985).
Nyctalus lasiopterus is protected by the EU Habitats Directive (Annex IV) and its conservation status is, thus far, assessed as Endangered in Italy (Rondinini et al. 2022), and it has been declared as Vulnerable on the IUCN Red List (Alcaldé et al. 2016). Data on its distribution and ecology are, therefore, substantial for the development of conservation policies in the EU member states.
The lack of data on the presence and distribution of N. lasiopterus may be due to several technical challenges associated with the monitoring of this species. Direct observations of N. lasiopterus are particularly difficult as these animals mostly shelter in tree holes (Dietz and Kiefer 2020). Moreover, they are fast and agile fliers and hunt and fly at high altitudes (Ibañez et al. 2011;Naďo et al. 2019). Therefore, capture attempts using mist nets are particularly challenging. During the past 10 years, the use of bat detectors and automatic sound recording systems has been widely established, allowing easier detection of the greater noctule bat. However, the echolocation calls of N. lasiopterus partly overlap with the social call characteristics of the sympatric species, Nyctalus leisleri (e.g., peak frequency). Moreover, the echolocation calls of N. lasiopterus can be confused with those of the European free-tailed bat, Tadarida teniotis (Barataud 2020). These issues may lead to the underestimation or misinterpretation of the presence of N. lasiopterus using bioacoustic methods, particularly in areas where T. teniotis is present simultaneously.
In the present study, we i) report the evidence for a continuous summer presence of the greater noctule bat on the southern side of the Alps, ii) accurately discriminate the calls of the greater noctule bat from those of the European free-tailed bat, as our recordings did not capture the full range of echolocation calls of N. lasiopterus, and iii) discuss the ecological aspects of the area where the species was detected.

Materials and methods
A large part of the data for this study was collected during a long-term project named Biodiversity Monitoring South Tyrol, (BMS 2022). From May to October in 2019, 2020 and 2021, Elekon Batloggers A + (www. batlo gger. ch) bat detectors were operated at 234 locations ( Fig. 1). The sites were chosen using a stratified selection approach, which included all the habitat types that occur in South Tyrol. The detectors were programmed to automatically record echolocation calls throughout the night for at least three consecutive nights per location (from sunset to sunrise), and they were positioned at a height of 1.5 m above the ground.
At one of the selected sites in 2019, possible calls of N. lasiopterus were detected for the first time. Hence, we proceeded with the installation of a Batlogger A + close to an irrigation pond, once per month for three consecutive nights (from May to August 2021), not far from the site where the presumed N. lasiopterus calls were detected ( Fig. 1-Terlano). Subsequently, greater noctule bat calls were detected at 3 of 20 other sites, which were monitored between 2016 and 2021 using Batlogger A + , in the same manner as mentioned above. In total, echolocation calls were recorded during 780 nights at 254 sites between 2016 and 2021 (altitudinal ranges from 207 to 2,788 m). To determine which Batlogger sequences may contain the calls of N. lasiopterus, all the sequences were first scanned manually using BatExplorer software (BatExplorer 2020). Subsequently, the sequences containing the suspected calls of N. lasiopterus were analysed again using the semiautomatic bat identification software package BatScope 4.1.1 (Obrist and Boesch 2018) and automatically assigned to a species. To investigate the differences in the species identifications, the presumed calls of N. lasiopterus and the calls of T. teniotis were measured manually using Raven Pro 1.6.1 (www. birds. corne ll. edu/ raven).
From the calls of the two species, which were used for all of the statistics mentioned below, we measured two time variables and six frequency variables, as shown in Fig. 2 Tables S1;  S2. The robust measurement methods implemented in Raven Pro 1.6.1 (Cortopassi 2006) were used, based on the energy stored in the selection of the call ( Table S1). The calls emitted in an assumed social context were excluded from the classifications. The differentiation of the calls of N. lasiopterus from Nyctalus noctula was also considered to be less critical, since Estók and Siemers (2009) did not highlight this issue in Hungary, and in South Tyrol, N. noctula occurs relatively rarely (11% of 254 sites) in contrast to the closely related N. leisleri (58% of 254 sites).
Based on the presumed N. lasiopterus and T. teniotis call measurements from South Tyrol, three methods were used to classify these two species using JASP 0.16 software (JASP Team 2021): i) random forest (RF), ii) K-nearest neighbours (k-NN) and iii) linear discriminant analysis (LDA).

Results and discussion
A total of 42 sequences of N. lasiopterus were identified from 6,096 h of recordings on 14 days in June (32/42), July (7/42) and August (3/42) between 2016 and 2021. At least three distinct call types were recognised in the call sequences of the greater noctule bats, in addition to transitional forms T. teniotis (n = 140) recorded in South Tyrol, Italy. The calls of the two species differed significantly for all the variables (Wilcoxon rank-sum test; Kassambara and Kassambara 2020). For the call measurement descriptions, please refer to Table S1 ( Fig. S1). The activities of the animals were mainly recorded during three nocturnal periods: between 21.20 and 22.30 (n = 11 echolocation sequences), between 0.00 to 02.30 (n = 21 echolocation sequences) and between 03.00 and 05.00 (n = 10 echolocation sequences) local time (air temperatures ranged from 15 to 28 °C; for details, please refer to Table S3). In Terlano, two N. lasiopterus individuals were recorded twice simultaneously (27th and 30th June 2019).
All the recordings of N. lasiopterus were registered at low altitudes (213-706 m) and in mild-climate areas of the Adige Valley ( Fig. 1; Table 1).
The validation and test accuracies for the classification algorithms (T. teniotis or N. lasiopterus; 20% of each of the data samples were used for testing and training) were 0.98 and 1.0 (RF ; Table S4), 0.97 (LDA Table S5), 0.98 and 1.0 (k-NN; Table S6), respectively.
Our results highlight that most of the studied echolocation calls of N. lasiopterus in South Tyrol were accurately distinguished from those of T. teniotis. Regarding the call variables of T. teniotis and N. lasiopterus, the range of values overlapped only slightly, except for the Delta Time. For N. lasiopterus, the frequency towards the end of the call was often in the range of 14 to 16 kHz, whereas for T. teniotis, it was between 10 and 12 kHz (Fig. 2, Table S2); N. lasiopterus displayed shorter calls in a higher frequency range than T. teniotis.
In this study, the N. lasiopterus calls were particularly variable regarding their shape in the frequency spectrogram when more than one individual and/or other bat species were calling simultaneously. The very-low-frequency modulated calls of N. lasiopterus (with frequencies of approximately 12 kHz in the terminal range, as described by Haquart and Disca [2007]) may be difficult to distinguish from the calls of T. teniotis. Moreover, the calls of T. teniotis (with frequencies in the terminal range greater than 12 kHz) may overlap with the calls of N. lasiopterus. However, in a study performed by Zbinden and Zingg (1986) on the southern side of the Alps, in 101 search pulses, the end frequencies of T. teniotis did not exceed 13 kHz, which coincides with the results of the present study.
The most striking deviation from the literature (e.g., Haquart and Disca 2007;Skiba 2009) was the predominantly short duration of the N. lasiopterus calls. Conversely, calls with a duration greater than 17 ms (Delta Time) were virtually absent. Presumably, in most cases, the greater noctule bats did not fly very far from the recording sites during the surveys, i.e., they did not fly at high altitudes (Close to reflecting objects, short calls prevent overlapping with the echo). The social calls of N. noctula and N. leisleri, with their lowest frequencies extending into the range of those of N. lasiopterus (below 15 kHz), are generally of much longer call duration (> 15 ms); however, if they are of short call duration, they are often followed by a sequence of several short-interval calls (< 15 ms;Zingg 1988a, b;Pfalzer 2002). We did not detect these types of signals in the 42 sequences analysed.
Previously, the greater noctule bat was sporadically captured in several Alpine areas (Arlettaz 1985;Vallotton 2007;Mazenauer 2015); however, the presence of resident populations of this species has not been determined. The results of the present study revealed a local and stable presence of N. lasiopterus on the southern side of the Alps. The repeated recordings of N. lasiopterus at five localities from late June to mid-August throughout several years suggest that a continuous presence of the species in South Tyrol, at least in summer. The presence of the species during summer could indicate a resident male population (mating and overwintering sites), as was previously reported in the Mediterranean during spring and summer (Helversen and Weid 1990;Ibáñez et al. 2009) or a breeding colony, with females present in spring and summer and then departing in mid-August to mating and overwintering sites (Ibáñez and Juste 2022). Hence, further comprehensive studies on the migration phenology and winter residency of the greater noctule bat in South Tyrol are still needed.
Notably, the Alps may not be optimal regarding the climatic requirements of the greater noctule bat. Nevertheless, the only evidence of a nursery of N. lasiopterus in the Alps originates from the Amsteg region (Reuss Valley, Switzerland; Fatio 1869), presumably at an elevation of  (Guariento et al. 2021;Ladurner et al. 2021). It is unknown whether N. lasiopterus dispersed only recently to South Tyrol or if its presence was overlooked in the past. Historical data is lacking due to methodological difficulty concerning the recording of this species. However, despite intensive surveys throughout the entire province, the species has only been encountered in the southernmost and sub-Mediterranean parts. This indicates a recent expansion of N. lasiopterus from the south or a lack of necessary environmental conditions in other areas.
N. lasiopterus is highly dependent on the presence of mature forests, where it often rests in the cavities of trunks and large branches (Estók et al. 2007;Naďo et al. 2019); however, individuals also occasionally use bat boxes, attics in human dwellings and crevices in large caves (Dietz and Kiefer 2020). The lower elevations of the Adige Valley are relatively poor in cavity-rich trees since the valley bottom is mainly covered by intensive apple orchards and settlements, whereas the slopes are covered with forests of low-growing deciduous trees. The large trees in the Adige Valley mostly consist of old-grown chestnut trees in several mixed forests and small areas with traditional cultivation; old deciduous and coniferous trees can also be found in parks and settlement alleys and their cavities could be selected as roosts, as already reported in the Mediterranean (Popa-Lisseanu et al. 2008). Furthermore, large deciduous trees exist in the small remnants of the riparian forests along the Adige River and several small protected areas of the valley bottom. For these reasons, the main conservation measures should focus on protecting large trees. This is particularly important at low elevations, since large trees are scarce, and their protection directly supports suitable roosts for this forest-dependent species.
In South Tyrol and in the Italian Peninsula, further research on N. lasiopterus is highly recommended to facilitate an accurate update of its conservation status in Italy on the IUCN Red List.
To the best of our knowledge, this study provides the first evidence of a continuous presence of N. lasiopterus on the southern side of the Alps during summer. Our results focus on bioacoustic identifications and demonstrate the effectiveness of non-invasive acoustic methods to collect occurrence and activity data on the greater noctule bat, supporting further large-scale applications.