Revealing year-round activity of cave-dwelling insectivorous bats with a sonotype classifier in data-deficient areas.

Context: Bats are the only mammals who have conquered the skies and the leading mammals in terms of total biomass with the exception of humans and domestic mammals. However, the ecology of many bat species is still unknown, especially in poorly sampled regions such as Central Africa. Aims: We present the first application of an acoustic method for studying the nocturnal activity patterns and annual phenology of bat activity of cave bat communities in regions where no bat acoustic classifier exists. Methods: In two caves (Mont Belo and Boundou) in the Republic of Congo, which are home to several species of insectivorous bats throughout the year, we set up a passive acoustic monitoring (PAM) over a period of 19 months. To identify the massive recordings collected, we used a sonotype classifier. To assign a species name or species group to the different sonotypes identified by the classifier, we collected local reference calls by capturing bats at the cave entrances. This enabled us to identify two bat groups and three species despite the absence of local call libraries. The capture sessions also allowed us to collect information about the biological status of the individuals, which we put in perspective with the annual phenology of activity. To quantify bat activity, we defined the activity of each acoustic group as the number of bat passes, where a bat pass corresponds to the detection of one or several echolocation calls within a five-second interval. This standardized measure allowed us to quantify nightly and seasonal variations in activity. Key results: The nightly activity displayed strong peaks in the evening and the morning, except during the long rainy season, when a tri- or quadri-modal activity was detected. This more evenly distributed activity throughout the night is related to the period of juvenile rearing (gestation, parturition and lactation). For all acoustic groups, we also observed a higher activity during the rainy seasons and the short dry season, compared to the long dry season. This high activity is linked to the lactation period, the independence of young bats and potentially the mating period. Finally, the reproductive activity of insectivorous bat species is synchronized with the period of high resource availability. Conclusion : This study represents the first long-term acoustic monitoring of cave bats in the Republic of Congo. This was allowed by the use of a classifier built without any pre-existing regional call data, highlighting a practical pathway for studying bat activity in other data-deficient regions. Implications : Information about the biological status of individuals captured at the monitored caves in association with the cave annual activity phenology produced with this new method are key to progress in the understanding of the ecology and the conservation issues of cave bat communities, which are particularly threatened globally.