Land surface spatial heterogeneity influenced by topography can be grouped into two categories: (1) spatial pattern of land surface properties including soil and vegetation types; and (2) spatial variability of atmospheric forcing. Despite of the importance of these spatial heterogeneities in controlling land surface processes, their relative importance is not well understood. This study focusses on the evaluation of the relative importance of subgrid level land surface properties and climate forcing to improve simulations of land surface processes in Earth System Models. For this purpose, the Energy Exascale Earth System Model (E3SM) Land Model (ELM) configured with topography-based subgrid structure and capability to downscale atmospheric forcing from atmosphere grid to land subgrids is applied using subgrid-level land surface properties and grid-level land surface data where subgrids in the same grid share the same land surface properties. Using ELM simulations with grid-level land surface properties driven by grid cell mean atmospheric forcing as a benchmark, we compare ELM simulations with subgrid-level land surface properties driven by grid cell mean atmospheric forcing; and ELM simulations with subgrid-level land surface properties driven by atmospheric forcing downscaled to subgrids. Furthermore, relative importance of subgrid level representations of soil and vegetation heterogeneity will be evaluated with and without downscaling of atmospheric forcing. Finally, simulation results will be compared against observations in topographically complex regions.