Future Droughts as Drivers of Ecosystem Change: Impacts on Vegetation Traits and Land-Atmosphere Fluxes

Climate change poses significant threats to society and ecosystems in general. The occurrence of more frequent and intense climate extremes is intensifying worldwide, driving economic losses, fatalities, and reduction of biodiversity. In this context, droughts are among the most impactful hazards, with the capability of spreading through large areas and persisting for extended periods. Their effects include jeopardizing food production and water supply for human use, as well as pushing natural ecosystems out of their adaptive limits. How ecosystems respond to this stress at the ecosystem scale remains an open question. Water stress drives plants to change their photosynthetic activity and carbon allocation strategies, with profound impacts on energy, carbon, and water fluxes. This dissertation investigates how droughts influence plant-trait dynamics, and the resulting land-atmosphere fluxes, under both current climate and future projections. First, drought events were identified under historic conditions and projected into the future based on outputs of the Coupled Model Intercomparison Project Phase 6 (CMIP6) following four socio-economic development pathways, in increasing order of greenhouse gases emissions, SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5. Drought events were identified through two established indices, the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI). Second, the drought indices were combined with other climate variables and remote sensed data sets of Leaf Mass Per Area (LMA) and Leaf Area Index (LAI) to derive Partial Least Square Regression (PLSR) models. LMA and LAI are important plant traits with direct effect on photosynthesis and carbon allocation strategies. By modelling these traits with PLSR models and climate forcing, it was possible to investigate the predominant ecosystem response to current climate conditions and project their future characteristics. Lastly, the ecohydrological model Tethys‑Chloris (T&C) was used to quantify how accounting for climate‑driven LMA changes alters water, energy, and carbon fluxes through their direct effect on canopy structure. Results indicate drought conditions worsen towards the end of the century for high emission future scenarios. Trends in drought severity are more pronounced for SPEI than SPI, driven by the accentuated increase of potential evapotranspiration with rising temperatures. Comparison of PLSR models based on spatial and temporal variability of LMA demonstrated that space-for-time substitution is restricted to moisture and energy axis at deciduous forests. Response of LMA at evergreen forests, and LAI at both forest types indicate slower processes such as compositional changes may be determinant in the temporal dynamics. The water stress caused by droughts drives vegetation toward more conservative leaf carbon allocation, indicating climate pressures would push LMA to higher values by the end of the century. Sensitivity analysis on LMA dynamics suggests that plant‑trait interannual variability have first order impacts on land–atmosphere fluxes. Results demonstrated that including LMA variability in ecohydrological models could change annual LAI in nearly -15% to + 21%, while Gross Primary Productivity (GPP) and transpiration in +/- 7%. These estimates isolate the structural pathway by which leaf construction cost (LMA) governs canopy area (LAI), holding photosynthetic capacity fixed, so that the simulated response is attributable to a single mechanism and constitutes a benchmark on the consequences of climate-driven trait adjustment. Quantifying the full feedback would require coupling trait acclimation to the atmosphere, and the results presented here provide the reference against which such an experiment can be evaluated. The findings presented in this dissertation can aid multiple engineering applications on different scales. Drought projections identify global hotspots of increasing drying trends, informing where resources should be applied to mitigate impacts on infrastructure, food production and hydropower generation. The effects of plant trait variability on land-surface fluxes yield insights into global and localized systems. As plants modulate carbon, water and energy fluxes, the accurate representation of these processes over land is key to improving outputs of Earth System Models. The same principle applies to finer scales, as the design of localized systems used in green infrastructure (e.g., rain garden, green roof) can be further improved by accounting for the role of dynamic vegetation response to climate.