Land Surface Model Intercomparison of Ecohydrological Processes over Complex Terrain and Their Impact on Land Surface Temperature

Ecohydrological processes are essential components of land-atmosphere interactions in regions of complex terrain and have the potential to affect convective rainfall formation. Yet, mountain biophysical phenomena are not well understood, and the superimposed impacts of topography, soil, and vegetation on land-atmosphere interactions have not been comprehensively quantified. In the North American Monsoon (NAM) region, the monsoon onset accompanied by dramatic vegetation greening adds spatial and temporal variability to local and regional ecohydrological processes. One of the land surface states, land surface temperature (LST), can be used as a reliable indicator of ecohydrologic processes since it is involved in both the energy balance and in water partitioning through varying evapotranspiration rates. In this study, we present results from a detailed analysis of distributed hydrologic model (TIN-based Real-time Integrated Basin Simulator, tRIBS) simulations which capture the seasonal evolution of LST and identify the key controls, related to meteorology, complex terrain, soil and vegetation cover, on the temporal and spatial distribution of LST. To accurately represent land surface conditions, the model utilized high resolution (30 m) terrain, soil classification, and time-varying vegetation parameters derived from remote sensing vegetation indices. Prior to the assessment, the high-resolution simulations were validated against multiple hydrologic variables from both ground sensor networks and remote sensing products at the catchment scale. The simulated spatial distribution of LST was compared with LST imageries from MODIS and ASTER with resolution of 1 km and 90 m, respectively. We then compare this high-resolution (100 m) benchmark to an intermediate-resolution (1 km) land surface model (the offline WRF-Hydro) to determine the skill in reproducing LST fields over a regional watershed. Our comparisons include an analysis of the effect of dynamic vegetation greening on the high and intermediate resolution models by withholding the remotely-sensed vegetation parameter changes in a leaf-off scenario. Lastly, we setup fully-coupled atmospheric and hydrologic simulations using the online WRF-Hydro model and discuss the implications of dynamic vegetation cover on land-atmosphere interactions in mountain areas.