Plant Hydraulic Traits Variation Across Temporal and Spatial Scales Affects Responses to Drought in the Tropical Dry Forest Biome

Abstract:

Drought events are increasing in their frequency and intensity within the tropical dry forest (TDF) biome, leading to a decrease in plant productivity, increase in mortality rates, changing forest structure and ecosystem processes. The magnitude of these consequences is predicted to be mostly driven by the distribution of hydraulic traits in the plant community, which is linked to the species’ abundances and how these change through time or space given environmental variation. Our main objective is to show the magnitude of hydraulic traits’ variation in TDF sites across spatial and temporal scales, and how this affects responses to extreme drought events. We also propose through a conceptual framework that the hydraulic traits’ space (i.e. mean, range and distribution) affect forest responses to drought and determines plant community drought sensitivity. To achieve this, we sampled hydraulic traits on 100 species across four countries in the Neotropics: Colombia, Costa Rica, Mexico and Puerto Rico. We also used 11 years of annual measurements in forest plots in Costa Rica to study the effects of El Niño Southern Oscillation (ENSO) on forest productivity measured as tree relative growth rates (RGR) and above ground biomass (AGB). First, our geographically extensive data showed that hydraulic traits varied greatly among species and sites, depending mostly on the site-specific proportion of deciduous species. Second, our long-term data show that during an ENSO-drought trees’ diameter growth was depressed relative to long-term means. Moreover, AGB declined due to an increase in tree mortality during the ENSO event. AGB losses were explained by the community weighted means of water potential at turgor loss point and hydraulic safety margin (HSM). While RGR reductions were explained by species’ HSM. Our results suggest that, within TDF sites, plant communities might respond differently to climate change due to their physiological traits’ distribution. A better representation of this variability, through community level sampling, will improve our ability to model and forecast the consequences of future drought events on TDF ecosystem processes and forest dynamics.