Predicting subgrid variability of soil water content from basic soil information

Knowledge of unresolved soil water content variability within model grid cells (i.e., subgrid variability) is important for accurate predictions of land-surface energy and hydrologic fluxes. Here we derived a closed-form expression to describe how soil water content variability depends on mean soil water content (sigma(theta)()) using stochastic analysis of 1-D unsaturated gravitational flow based on the van Genuchten-Mualem (VGM) model. A sensitivity analysis showed that the n parameter strongly influenced both the shape and magnitude of the maximum of sigma(theta)(). The closed-form expression was used to predict sigma(theta)() for eight data sets with varying soil texture using VGM parameters obtained from pedotransfer functions that rely on available soil information. Generally, there was good agreement between observed and predicted sigma(theta)() despite the obvious simplifications that were used to derive the closed-form expression. Furthermore, the novel closed-form expression was successfully used to inversely estimate the variability of hydraulic properties from observed sigma(theta)() data.