The strong 2015-16 El Ni?o Southern Oscillation event provided a rare opportunity to study the impact of severe drought on Central American tropical forest dynamics and was the focus of this RAPID project. Given that the severity and frequency of drought events are predicted to increase due to climate change, there is an urgent need to understand how tropical forests respond to severe drought events. In particular, a better understanding of drought impacts on the regeneration of tree seedlings is critical for determining future tropical forest composition, diversity, and function.

We examined seedling dynamics before, during, and after the 2015 - 2016 ENSO event to test several hypotheses related to the response of tropical tree seedlings to severe seasonal drought.  We measured seedling recruitment, growth, and mortality as well as soil and leaf water potentials and light across a network of 8 1-hectare forest inventory plots in Central Panama that are arrayed across a strong rainfall and seasonal drought gradient.  Historical records from Barro Colorado Island, Panama (approximately ? way across the rainfall gradient) indicate that soil moisture was reduced in the late dry season of 2016 relative to historical levels (1971 to 2015). Soil moisture measurements show a strong linear relationship to mean annual precipitation in 2014 across our eight plots. Soil moisture was lower and more variable across these plots in 2016. To assess the impact of El Nino drought events on forest regeneration, we monitored > 13,000 seedlings of 656 species in established plots across the gradient.  We measured growth, survival, and recruitment of all seedlings in our focal plots. We found that overall mortality was around twice as high in the dry season than the wet season in 2016, suggesting significant stress caused by reduced rainfall during the dry season can increase seedling mortality. However, annual mortality rates for seedlings varied widely across the eight plots, and only two sites showed elevated mortality in the El Nino year compared to previous years.  These sites were located not at the extreme ends of the gradient, but at the sites that were the second most extreme wet and dry sites (Sherman and Parque Metropolitano, respectively).  While the Parque Metropolitano site showed much higher seedling mortality in the dry season relative the wet season during the El Nino year, seedling mortality was higher in the wet season than the dry season at Sherman that year suggesting that El Nino-related drought stress did not drive the elevated mortality at that site.

Overall, these results suggest that the El Nino event did not result in widespread seedling mortality due to severe dry season drought stress. Instead, we saw that the impact of the El Nino event on community-wide seedling mortality varied across sites and did not show a clear relationship with mean annual rainfall across the precipitation gradient. This may be due to the fact that there is substantial turnover in species composition across the gradient, and individual species may be responding differentially to the El Nino event. We are in the process of exploring this possibility using published estimates of tree species sensitivity to moisture based on species distribution data from this region (Condit et al. 2013 PNAS). We are also in the process of linking available data on species functional traits (e.g. wood density, maximum stature, SLA, leaf area; Wright et al. 2010) to our seedling data in order to assess whether individual species response?s to drought stress are related to specific functional traits (or suites of traits).  This project involved training for 2 postdoctoral research scientists and 6 research technicians.  The outcomes of the research will likely identify specific tree species that may be resilient or susceptible to severe seasonal drought and how this will impact forest diversity and function in the future.

Last Modified: 11/07/2018
Modified by: Frank Jones