Climate drives the long-term ant male production in a tropical community.

Forecasting insect responses to environmental variables at local and global spatial scales remains a crucial task in Ecology. However, predicting future responses requires long-term datasets, which are rarely available for insects, especially in the tropics. From 2002 to 2017, we recorded male ant incidence of 155 ant species at ten malaise traps on the 50-ha ForestGEO plot in Barro Colorado Island. In this Panamanian tropical rainforest, traps were deployed for two weeks during the wet and dry seasons. Short-term changes in the timing of male flying activity were pronounced, and compositionally distinct assemblages flew during the wet and dry seasons. Notably, the composition of these distinct flying assemblages oscillated in consistent 4-year cycles but did not change during the 16-year study period. Across time, a Seasonal Auto-Regressive Integrated Moving Average model explained 75% of long-term variability in male ant production (i.e., the summed incidence of male species across traps), which responded negatively to monthly maximum temperature, and positively to sea surface temperature, a surrogate for El Niño Southern Oscillation (ENSO) events. Establishing these relationships allowed us to forecast ant production until 2022 when year-long local climate variables were available. Consistent with the data, the forecast indicated no significant changes in long-term temporal trends of male ant production. However, simulations of different scenarios of climate variables found that strong ENSO events and maximum temperature impacted male ant production positively and negatively, respectively. Our results highlight the dependence of ant male production on both short- and long-term temperature changes, which is critical under current global warming.