Utilization of data and modeling at multiple scales to compare varying formulations of the soil resistance term affecting evaporative flux from the soil surface.

Land Surface Models (LSMs) are used to predict heat, energy, and momentum fluxesoccurring at the land surface and the resulting effects in the soil and atmosphere at various scales.Evaporation from bare soil is an integral component of the water balance that is very difficult toaccurately predict since it is complexly affected by the coupled effects of atmospheric conditions andsoil properties. Inaccurate or simplifying assumptions can have drastic effects on regional and globalLSM predictions and cause available LSMs to predict conflicting values for the soil moistureconditions and surface fluxes (e.g. evapotranspiration, infiltration, run off). The goal of this work isto see how heterogeneities in soil properties can be properly represented with a soil resistance termthat accounts for physically based parameters of the soil system at the land-atmosphere interface.Utilizing a comprehensive, experimental dataset generated from a soil with known, heterogeneousproperties under highly controlled atmospheric conditions, we are able to compare the effectivenessof various parameterizations in two different models. The first being a multiphase, non-equilibrium,and non-isothermal model that minimizes the dependence on fitting parameters. The effects ofcertain mechanisms are better understood at this fine scale and incorporated into the land surfacecomponent of the Accelerated Climate Modeling for Energy project (ALM), which is focused oncapturing the interactions between the surface and the atmosphere at larger scales. The formulationsof the resistance parameter, soil water retention curve (SWRC), and diffusivity through partiallysaturated porous media are of particular interest. The fine scale model was used in conjunction withthe experimental data to test formulations before implementing them into the ACME Land Model(ALM). Effects of these alterations were compared to the existing mechanisms in ALM and thentested against lab and field scale data sets. Initial findings suggest the Tang and Riley (2013a) soilresistance more accurately reproduces results lab and field results on multiple scales whereheterogeneity is present. Further understanding of soil resistance will lead to more robust landsurface models which decrease the reliance on such empirical relationships.