Soil moisture simulation: Achievements of the RICE and PILPS intercomparison workshop and future directions

Soil moisture simulation in current landsurface schemes is diverse: in terms of the realism (and complexity) of the parameterizations and in terms of the resulting predictions. Even after careful adjustment, the range in soil moisture among the results from the 14 schemes compared for the HAPEX-MOBILHY (S. France) siteʹs 1.6 m soil layer remains around 100 mm (the annual precipitation for HAPEX-MOBILHY being 856 mm). Results show that the partitioning of surface available energy into sensible and latent heat fluxes is closely coupled to the partition of precipitation into evaporation and runoff plus drainage. Runoff seems to be grossly (30–40%) underestimated by 13 of the 14 participating schemes but incomplete observations do not allow a full evaluation. Runoff from the nearly saturated soil outside the main growing season differs significantly among landsurface schemes although these differences do not greatly affect soil moisture values themselves. Transpiration operates at different rates depending upon the initial soil water content as well as the scheme formulation. The implications are far-reaching: (l) landsurface schemes forced by the same atmospheric conditions reached different equilibrium states; and (2) response times differ for schemes with different structures and different parameterizations. In combination, these two conclusions mean that observational data limited to a few weeks or months cannot be used for validation unless the initial conditions are known very precisely. Soil moisture remains an ill-defined, and inconsistently and, generally, inaccurately predicted quantity.