phenology and environmental determinants of explosive breeding in gliding treefrogs: diel timing of rainfall matters

The influence of abiotic and biotic factors on the temporal pattern of calling and breeding for many temperate anurans is well understood. However, few studies have documented patterns of reproduction in explosive-breeding tropical frogs or incorporated multiple environmental factors in their analyses, especially across multiple breeding seasons. We combine long-term natural history observations and automated data collection methods with boosted regression tree (BRT) analysis to determine the phenology and determinants of explosive breeding in the gliding treefrog, Agalychnis spurrelli. We monitored breeding for a total of 418 days across three breeding seasons and determined the relative importance of several environmental factors on the probability of calling and breeding activity. Our study population of A. spurrelli on Costa Rica’s Osa Peninsula forms breeding aggregations up to 11 times per year during 1–2-day long explosive-breeding events, from late May to mid-September. Calling and breeding activity are strongly and positively related to accumulated rainfall during the previous 24 and 48–24 hours before, particularly rainfall during the afternoon and evening. Day-of-year, days since breeding occurred, and lunar phase also influence reproductive activity. This study provides the first description and analysis of the phenology and factors that predict explosive breeding in A. spurrelli and illustrates the value of using automated data collection paired with BRTs for analysis of complex ecological data.

诸多温带无尾两栖类的鸣唱与繁殖时间格局受非生物与生物因子的影响机制已得到充分阐释。然而，针对爆炸式繁殖的热带蛙类，相关繁殖格局的记载研究极少，且鲜有研究纳入多环境因子开展分析，尤其是跨多个繁殖季的研究。本研究结合长期野外自然历史观测与自动化数据采集方法，辅以提升回归树（boosted regression tree, BRT）分析，旨在明确斯氏叶泡蛙（*Agalychnis spurrelli*）爆炸式繁殖的物候特征及其驱动因子。本研究在三个繁殖季内累计开展了418天的繁殖监测，明确了多种环境因子对鸣唱与繁殖活动发生概率的相对重要性。栖息于哥斯达黎加奥萨半岛的该种群，其繁殖活动集中在5月下旬至9月中旬，每次为持续1~2天的爆炸式繁殖事件，每年最多可形成11次繁殖集群。鸣唱与繁殖活动与前24小时及前24~48小时的累积降雨量呈显著正相关，其中午后与晚间的降雨相关性尤为突出。当日日期、距上次繁殖事件的间隔天数以及月相，同样会对繁殖活动产生影响。本研究首次对斯氏叶泡蛙的爆炸式繁殖物候及其预测因子进行了系统描述与分析，并证实了将自动化数据采集与提升回归树分析相结合，可有效用于复杂生态学数据的解析。