Hibernation is a widespread adaptation in animals to seasonally changing environmental conditions. In the face of global anthropogenic change, information about plastic adjustments to environmental conditions and associated mortality costs are urgently needed to assess population persistence of hibernating species. Here, we used a five-year data set of 1,047 RFID-tagged individuals from two bat species, Myotis nattereri and Myotis daubentonii that were automatically recorded each time they entered or left a hibernaculum. Because the two species differ in foraging strategy and activity pattern during winter, we expected species–specific responses in the timing of hibernation relative to environmental conditions, as well as different mortality costs of early departure from the hibernaculum in spring. Applying mixed-effects modelling, we were able to disentangle population-level and individual-level plasticity in the timing of departure. To estimate mortality costs of early departure, we used both a capture mark recapture analysis and a novel approach that takes into account individual exposure times to mortality outside the hibernaculum. We found that the timing of departure varied between species as well as among and within individuals, and was plastically adjusted to large-scale weather conditions as measured by the NAO (North Atlantic Oscillation) index. Individuals of M. nattereri, which can exploit milder temperatures for foraging during winter, tuned departure more closely to the NAO index than individuals of M. daubentoniid which do not hunt during winter. Both analytical approaches used to estimate mortality costs showed that early departing individuals were less likely to survive until the subsequent hibernation period than individuals that departed later. Overall, our study demonstrates that individuals of long-lived hibernating bat species have the potential to plastically adjust to changing climatic conditions, although the potential for adjustment differs between species.

Bat capture and data logging

We studied arrival (first recorded activity per hibernation period) and timing of departure from the hibernaculum (last recorded activity per hibernation period) of individually marked Natterer’s bats (M. nattereri n = 540) and Daubenton’s bats (M. daubentonii: n = 507) at a hibernaculum in North-Rhine Westphalia (Figure S1) (Stumpf et al. 2017, Meier et al. 2022). We caught bats using harp traps at the entrance of the hibernaculum, sexed them and distinguished two age classes, young of the year (1st hibernation) and adults (>1 hibernation; for details see Reusch et al. 2019). At this site, bats have been marked with individual RFID-tags (ID 100, Trovan) since 2002 for Natterer’s bats and since 2008 for Daubenton’s bats. Here, we focused on the years 2011 to 2015, for which we had continuous monitoring of bats entering and exiting the hibernaculum. All capture and marking were conducted under the permits (50.0835.2.1, 84-02.04.2015.A508, 70.2-2012/0254, 70.2-0228/10, 70.2-0197/08 and 70.2.2.27), issued by the relevant authorities.

The studied hibernaculum has two small entrances (8 cm x 5 cm), which were monitored by RFID-logger-antenna systems (LID-650, EURO I.D.). These systems continuously recorded the ID and a timestamp for all tagged individuals without interfering with the bats’ behaviour (Reusch et al. 2019, Meier et al. 2022). As only one antenna was employed per entrance, we could not directly assess whether a tagged bat entered or left the hibernaculum when detected by the RFID-logger. Therefore, we defined an arrival period (Aug 1st - Dec 31st) and a departure period (Jan 1st - Apr 30th) for each hibernation (Ht) based on activity patterns known from direct observations in the field, light barrier recordings (that count the number of bats entering and emerging) and RFID recordings (as in Meier et al. 2022).

For each individual and year, we determined the date of the last recording during the departure period (departure date of Ht) and the date of the first recording during the subsequent arrival period (arrival date of Ht+1). An individual was considered dead if it was not recorded in the subsequent hibernation period (Ht+1, Aug-Apr). We confirmed this assumption based on the available data: In our 5-year study period of continuous monitoring at the hibernaculum only one out of 1,047 tagged individuals re-occurred after not being recorded during a complete hibernation period. Moreover, in 13 years of bat surveillance only 24 out of 1,111 bats (2.2 %) assigned dead according to the abovementioned criteria re-appeared at another RFID-monitored hibernaculum in close proximity to the study site (< 500 m distance) for at least one more year [Grosche and Meier, unpublished data].

To make sure that recorded individuals have indeed used the hibernaculum for hibernation, we excluded those individuals that had not been recorded during both, the arrival and the following departure period of a given hibernation period (Aug-Apr.; M. nattereri: 18 out of 540 individuals (3.3 %); M. daubentonii: 32 out of 507 individuals (6.3).

Data file may be opened with excel, any text editor, and R. All Analyses were performed with R.