Impact of Lateral Flow on Surface Water and Energy Budgets over the Great Plains – A Modeling Study

Abstract:

The land surface water and energy budgets are closely linked to its hydrologic state, which is scale-dependent and shall be resolved differently at different resolutions. While it is appropriate for land surface models (LSMs) to consider only 1-dimensional (1D) vertical exchange of water and energy at coarse resolutions, new requirements emerge for better resolving land surface physical processes, including the 3-dimensional (3D) lateral flow, in numerical weather prediction and climate models at higher resolutions. As the horizontal grid spacing decreases, treatment of hydrologic processes, such as the lateral flow of surface and subsurface flow, need to be explicitly represented. In this study, the offline WRF-Hydro is employed to study the lateral flow impact on soil moisture and energy fluxes over the Great Plains (GPs). The vast amount of measurements over the GPs provide an unique opportunity to assess the model behavior. In addition, newly developed land surface characterization and input forcing are ingested in the model, in an attempt to reduce uncertainties associated with the initial and boundary forcing and help to identify model deficiency. Our results show that the more realistic inputs (parameters, soil types, forcing) lead to larger underestimation of latent heat flux and dry bias, indicating model’s inability to properly represent SM and energy fluxes and the existence of model structural uncertainty (embedded errors) in WRF-Hydro that need to be characterized to inform future model development efforts. Including lateral flow processes partly mitigates the model deficiencies in representing hydrologic processes and alleviates the dry bias. In particular, both surface and subsurface lateral flow increases SM mainly over the lower elevations, except that subsurface flow also affects SM over steeper terrains. Additional simulations have been performed to evaluate the effect of routing resolution on model results. When LSMs resolution is high, noticeable differences in SM are exhibited between different routing resolutions especially over steep terrains, whereas when LSM resolution is coarse, difference between routing resolutions become negligible especially over flat terrains.