The role of spatiotemporal plant trait variability in model predictions of ecohydrological responses to climate change in a desert shrubland

Although spatial heterogeneity of soil properties, topography, and climate is commonly incorporated into ecohydrological models, the spatial and temporal variability in plant functional traits is typically overlooked. The objective of our study is to evaluate the impact of trait parameter variability on modeled ecohydrological processes. We implemented a model-data fusion approach to constrain spatiotemporally dynamic parameters in plant functional traits at two desert shrubland sites located along a topographic and climate gradient in the Mojave Desert. Our results showed that the estimates for specific leaf area and rooting depth for the broadleaf-evergreen-shrub plant-functional-type showed spatial variability, with lower specific leaf area and deeper rooting depth found at the low elevation site. We also found that the specific leaf area estimates changed over time at both sites in response to water stress, but with different sensitivities, possibly depending on species and/or climate. The spatial variability in trait parameter estimates was greater than temporal variability and played a more important role in accurately simulating ecohydrological processes, but including the temporal variability in specific leaf area further improved seasonal predictions. In simulations forced by future climate projections under the Representative Concentration Pathway 4.5 (RCP 4.5) and 8.5 (RCP 8.5) greenhouse gas emissions scenarios, spatial variability in trait parameters impacted predictions of both carbon and water fluxes, while temporal variability in trait parameters resulted in predictions of higher ecological function and water use efficiency. The higher water use efficiency led to improved ecohydrological function in simulations under RCP 4.5, but it showed little capacity for buffering intensive water stresses under the more pessimistic RCP 8.5 scenario, indicating that with spatiotemporally variable trait parameters, the impact on predicted ecohydrological processes depends on the climate projections. Overall, our modeling results prompt further field-based examination of temporal and belowground trait variability in desert shrublands, and they raise the question of how combined spatiotemporal variabilities of multiple traits may support ecohydrological function under water stress.