Microtopographic hydrologic variability change resulting from vegetation acclimation response to elevated atmospheric CO2

Abstract:

The elevated concentration of atmospheric CO₂ increases the ratio of carbon fixation to water loss from plants or water use efficiency, which reduces transpiration. However, the magnitude of the effects of this vegetation acclimation on hydrologic dynamics, such as soil moisture content and surface runoff controlled by microtopographic variability on the land surface, remains unclear. Here we integrate a multi-layer canopy-root-soil model (MLCan) with a coupled surface-subsurface flow model (GCSFlow) to capture the acclimation responses of vegetation to climate change and predict how these changes affect hydrologic dynamics on landscapes at fine scales. The model is implemented on a hybrid CPU-GPU parallel computing environment to overcome challenges associated with the high density of computational grid and nonlinear solvers. The model is capable of simulating large-scale heterogeneities due to both microtopography and soils and lateral water fluxes at emerging lidar-scale resolutions (~1m). We demonstrate that hybrid computing is feasible for detailed, large-scale ecohydrologic modeling, which has been previously assumed to be an intractable computational problem. Simulations are performed for corn crop in the Goose Creek watershed in central Illinois, USA at present and projected higher concentrations of atmospheric CO₂, 400 ppm and 550 ppm, respectively. The results show a net decrease of 11% for the average annual evapotranspiration of corn, which increases water content in the soil and at the land surface. These results highlight the critical role of a warming climate on atmospheric-soil-vegetation interactions and the need to understand other dynamics near the soil surface associated with water and vegetation.