The study species and sites were first selected, from more than a thousand amphibian anurans and all around Brazil and surrounding countries, by a detailed analysis based on six taxonomic and five biogeographic criteria. This analytic process led to identify 37 study species and seven study sites across Brazil, Bolivia and French Guiana. The study sites corresponded to four biomes (Atlantic Forest, Cerrado Forest, Pantanal, and Amazon Rainforest) and were located at the climatic extremes of the study species. Data collection for each site was conducted by using passive acoustic monitoring, a new automated and non-invasive recording technique, in order to register calling activity of the study species and communities as well as weather conditions during one entire breeding season. This automated method has been proposed because showing diverse advantages, such as reducing disturbances from the presence of observes, obtaining larger and simultaneous sampling data or enhancing species detection.
Thereby, data collection provided above 400 thousand recordings (>2 TB) and additional metadata corresponding to a year of automated monitoring for all study sites. For analysing this big data set, an ad-hoc semi-automated sound recognition procedure was developed. This procedure allowed determining the daily and seasonal patterns of calling activity of each study species and associated them with specific weather conditions.
Overall, the daily and seasonal patterns of calling activity were found to be species-specific, with larger differences in seasonal patterns than in daily ones between populations located at climatic extremes. This suggests that the study species are more prone to shifts in their breeding phenology than in their day-night cycle of activity. On the other hand, in comparison with temperate zone anurans, tropical anurans showed calling activity at more restricted ranges of temperature and relative humidity, which meets the established predictions. Hence, our findings indicated that tropical anurans are more vulnerable to climate changes during their breeding and calling activity than temperate zone species.
Finally, the information collected was expanded into a larger temporal and spatial scale, following a biogeographic approach, in order to predict the impacts of the current climate change on phenology and geographic distribution of the study species. This analytic procedure was first tested in temperate zone anurans using previous data sets. According to a severe climate change scenario for 2050, our findings pointed out that shifts in phenology due to climate change are expected to be more probable than shifts in geographic ranges in temperate zone anurans. The phenology of the breeding and calling activity of the study species was predicted to be altered by climate change, showing prominent advances in comparison with the current breeding season.
