Understanding the influence of land surface heterogeneity on surface water, energy and carbon fluxes is crucial for modeling earth system variability and change. Using the E3SM land model (ELMv1) as the testbed, this presentation explores the factors that impact the modeling of land surface heterogeneity via two case studies.

In the first study, we investigated the effects of major heterogeneity sources on water and energy simulations over the CONUS. Four heterogeneity sources are considered that include atmospheric forcing (ATM), soil properties (SOIL), land use and land cover (LULC), and topography (TOPO). The Sobol' total and first-order sensitivity indices were utilized to quantify the relative importance of the four heterogeneity sources. ATM and LULC are the most dominant heterogeneity sources in determining spatial variability of water and energy partitioning, mainly contributed by their own heterogeneity and slightly contributed by their interactions with other heterogeneity sources. Their heterogeneity effects are complementary, both spatially and temporally. The overall impacts of SOIL and TOPO are negligible, except TOPO dominates the spatial variability of the runoff-to-precipitation ratio across the transitional climate zone between the arid western and humid eastern CONUS.

In the second study, we further conducted multiple ELM experiments over the Mid-Atlantic region, considering the factors related to mesh generation and input data sources and resolution that affect modeling of surface heterogeneity: (1) modeling using structured mesh (0.125 degree) vs. unstructured MPAS mesh (2–5km); (2) two land surface data with different resolutions, 0.125 degree vs. 1km; (3) three atmospheric forcings, GSWP3 (0.5 degree) vs. NLDAS (0.125 degree) vs. MSWX (0.1 degree). Multiple-source datasets were collected to evaluate our ELM simulations, including remote sensing datasets, gauge observations, and reanalysis datasets. Our preliminary results show that high-resolution land surface data improves the gross primary production and evapotranspiration simulations. Different atmospheric forcings also have large impacts on the simulation of snow water equivalent and river discharge.