Work

Moisture Recycling and Vegetation Strategies: Spatiotemporal Dynamics of the Amazon Ecohydrological System

Year: 2026
Type: dissertation
Abstract: The Amazon Basin, a critical component of the global climate system, is increasingly affected by accelerating land-cover change and the growing frequency and intensity of extreme weather events. These pressures have heightened concerns about the basin’s ecohydrological resilience and potentially approaching a regime shift, in which feedbacks between vegetation, moisture recycling, and climate could lead to large-scale and potentially irreversible transformations from a forest to a stable grassland area. Yet, the role of plant traits in mediating this feedback is poorly understood. Trait-based approaches have been proposed as they represent the functional links between the vegetation and the atmosphere, thus enabling to assess the role of biodiversity in the land atmosphere interactions. Several traits can be linked to water fluxes, from the root zone, passing through the stems and woody components of vegetation as well as those related to gas exchange and photosynthesis. The functioning of these traits can be altered by deforestation, as the loss or disturbance of trait composition affects water fluxes across multiple timescales, and can further compound the impacts of extreme events. In addition, the composition and spatial distribution of trait values within a region can influence these processes in distinct ways across both space and time. In this dissertation, I address three questions to test these hypotheses. First, together with my co-authors, we investigate whether spatial patterns in plant functional traits are linked to hydrological and atmospheric processes across the Amazon Basin, whether these relationships vary with hydroclimatic extremes, and among sub-basins. The hypotheses are that plant traits regulate hydroclimatic variables, as they exert control on water fluxes entering and exiting the plant. Using remote sensing-based plant traits (Specific Leaf Area, Leaf Dry Matter Content, Leaf Phosphorus Content, and Leaf Nitrogen Content), we link these to Evapotranspiration, Potential evapotranspiration, Soil moisture, Vapor Pressure Deficit, and Land Surface Temperature. We find that Specific Leaf Area and Leaf Dry Matter Content exhibit strong yet opposing correlations with all hydroclimatic variables except Soil Moisture. Using quantile regression, we find that these relationships intensify under extreme values of hydroclimatic variables and become more distinct at the sub-basin scale. In contrast, Leaf Phosphorus Content shows weak and more spatially inconsistent correlations, whereas Leaf Nitrogen Content appears largely decoupled from the examined hydroclimatic variables. Second, we ask how the spatio-temporal dynamics of the Amazon's moisture source and sink processes have changed over time, to what extent these changes are attributable to deforestation, and what are the timeframes at which these effects are taking place. We used the UTrack Lagrangian moisture tracking model combined with high-resolution (2000-2021) deforestation data from MapBiomas to quantify these hydrological effects. We found substantial heterogeneity in the basin’s moisture recycling regimes over the period analyzed. The strongest effects of deforestation were not immediate but emerged at mid term (13–48 months) and long term (49–72 months) lags. These shifts included a reduced regional evaporation contribution of the Amazon as a moisture source and an increased dependency on regionally sourced rainfall as a moisture sink. Together, these patterns demonstrate the persistent and delayed influence of deforestation on the basin’s hydrological functioning. Third, we examined how spatial patterns of vegetation-induced evaporation following the 2010 extreme drought relative to the pre-drought baseline, the extent to which these ecohydrological responses can be explained by plant functional traits, and whether forest loss modulates this relationship. Using trait maps, GLEAMv4 evaporation composition data and the MapBiomas land cover change dataset, we quantified vegetation’s association with the shifts in vegetation-induced evaporation. We find that one year after the drought there are widespread negative z-scores in vegetation-induced evaporation ratio, with 65\% of the basin exhibiting reduced hydrological function. Traits reflecting a ``conservative" resource-use strategy (e.g., high Leaf Dry Matter Content) were linked to higher resilience, while ``acquisitive" traits (e.g., high Specific Leaf Area) corresponded to greater vulnerability. These associations were stronger in areas that had experienced net forest loss, indicating that hydrological responses become more tightly linked to functional composition where the deforestation occurred. In conclusion, we conclude that the Amazon’s ecohydrological resilience arises from the diversity of its plant functional traits, which sustain the basin’s capacity to regulate water fluxes. This is particularly linked to two plant trait strategies, which show associations with hydroclimatic variables. We find that plant functional traits enable the ecosystem to maintain hydrological balance and buffers it against short-term climatic shocks. However, the resilience provided by plant functional traits is being progressively eroded by cumulative deforestation, which alters evapotranspiration dynamics, and heightens vulnerability to climatic disturbances such as droughts and heat extremes. Together, these findings offer a first order understanding of the mechanisms through which plant functional traits shape ecohydrological resilience in tropical forests and point to the accelerating loss of this resilience under continued land use change. These results offer important insights for modeling and projecting the Amazon’s future dynamics in the Anthropocene, showing the importance of conservation and climate strategies that consider both ecohydrological functioning and the spatial integrity of the forest.
Author: Nguyen, Kien
Language: English
Open Access status: green
OpenAlex ID: W7166186323
Cites:
Cited by:
Related to: 9