Which temporal resolution to consider when investigating the impact of climatic data on population dynamics? The case of the lesser horseshoe bat (Rhinolophus hipposideros)

Pierre-Loup Jan, Olivier Farcy, Josselin Boireau, Erwan Le Texier, Alice Baudoin, Pascaline Le Gouar, Sébastien J. Puechmaille, Eric J. Petit

Climatic variables are often considered when studying environmental impacts on population dynamics of terrestrial species. However, the temporal resolution considered varies depending on studies, even among studies of the same taxa. Most studies interested in climatic impacts on populations tend to average climatic data across timeframes covering life cycle periods of the organism in question or longer, even though most climatic databases provide at least a monthly resolution. We explored the impact of climatic variables on lesser horseshoe bat (Rhinolophus hipposideros) demography based on count data collected at 94 maternity colonies from 2000 to 2014 in Britanny, France. Meteorological data were considered using different time resolutions (month, life cycle period and year) to investigate their adequacy. Model averaging was used to detect significant predictors for each temporal resolution. Our results show that the finest temporal resolution, e.g. month, was more informative than coarser ones. Precipitation predictors were particularly decisive, with a negative impact on colony sizes when rainfall occurred in October, and a positive impact for June precipitations. Fecundity was influenced by April weather. This highlights the strong impact of climatic conditions during crucial but short time periods on the population dynamics of bats. We demonstrate the importance of choosing an appropriate time resolution and suggest that analogous studies should consider fine-scale temporal resolution (e.g. month) to better grasp the relationship between population dynamics and climatic conditions.