Context. According to the current trend of biodiversity loss, information on population trends at large temporal and spatial scales is necessary. However, well documented animal population dynamics are generally based on intensive protocols requiring animal manipulation, which can be impossible to conduct in species for which conservation is a concern.
Aims. For many bat species, an alternative approach entails performing an appropriate analysis of counts in roost cavities. Because of managers’ perception of chaotic variations through time, relatively few count monitoring surveys are regularly analysed. Here, we present the analysis of a twenty-two-year survey of a large hibernaculum of pipistrelle bats (Pipistrellus pipistrellus) located in a railway tunnel in Paris, France.
Methods. We propose that using combinations of population-dynamics modelling using demographic parameters from the literature and statistical analyses helps with identifying the biological and methodological effects underlying the dynamics observed in census analyses.
Key results. We determined that some of the observed year-to-year variations of population size cannot be explained only by the intrinsic dynamics of the population. In particular, in 1993–94, the population size increased by >40%, which should have implied a massive immigration. This change coincided with the end of the operation of the railway line. After consideration of a drastic trend of population decline (7% year–1), we were able to detect this event and several environmental effects. Specifically, the winter conditions and the temperature in July affected the colony size, presumably because of aggregative behaviour and reproduction success, respectively.
Conclusions. Emigration–immigration processes might have preponderant effects on population dynamics. In addition, our analysis demonstrated that (1) the study population suffered a large decline, (2) a combination of human disturbance and meteorological variation explains these dynamics and (3) emigration–immigration processes have preponderant effects on the population dynamics.
Implications. To conduct a meaningful analysis of non-standard time series and provide a source of data for implementing biodiversity indicators, it is necessary to include (1) the local knowledge of the people involved in the field surveys in these analyses (the existence of disturbances and site protections) and (2) meteorological information for the appropriate seasons of the year.
