Contrasting water and energy fluxes during dry (2015 El Nino) and wet (2008 La Nina) extreme events at an Amazonian tropical forest.

Tropical forests are important biodiversity and biomass reservoirs - determining this ecosystems response to climate change is important, as the Amazon hydrological cycle is intensifying. Across a network of eddy covariance (EC) forest sites at the equatorial Amazon, seasonal ET has been shown to be driven by net radiation (Rn). This relationship indicates a strong environmental (e.g. energy availability as key driver), and although important, a less significant biological control (e.g. stomatal resistance) over water fluxes. However, little is known about changes in the ability of forests to drive reductions on ecosystem-scale ET via stomatal closure during drought conditions. As we expect the frequency and severity of extreme droughts to increase, three questions remain unanswered: (1) Is there a precipitation/plant available water threshold after which canopy resistance will assume a more significant role in determining seasonal ET fluxes? (2) Does the effect of sustained low rainfall periods change the observed seasonal relationship between Rn and ET? (3) Do different vegetation strategies (e.g. leaf abscission, and lower canopy conductance) contribute to reducing water loss? Here, we contrast water and fluxes during dry (2015 El Niño) and wet (2008 La Niña) extreme events at an Amazonian tropical forest. During the El Niño, ET decreased -- changes in seasonal water flux mainly driven by reduccions in transpiration (associated to vegetation controls) rather than by evaporation (associated to increasing atmospheric demand and available energy) -- the low Gs values reported during El Niño, and several years after, the 2015-2016 event. Therefore, during the drought, an unusually high fraction of available energy was allocated to H rather than to LE --as described by abnormally high Bowen ratio values. By contrast, during La Niña we observed both H and LE decreasing relative to the mean seasonality, these reductions driven by lower demand (Rn). Changing the turbulent flux partition (H>LE) with associated feedback effects (e.g. higher temperatures and atmospheric demand) will have a significant influence on local and global precipitation patterns and the energy cycle and a direct impact on the land-atmosphere exchange, biomass, and forest structure.