A New Scale Based 'Heterogeneity Matrix' for Improving Soil Moisture Spatial Scaling

Soil moisture is a dynamic state variable of interest to agronomists, hydrologists and climate modelers alike. But the spatial scales at which they require soil moisture data are very different from each other. Past studies have established that the spatial structure of soil moisture is dependent upon the heterogeneity in physical controls, namely, precipitation, soil, vegetation and topography of the region. However, the spatial structure of soil moisture has also been found to be severely affected by scale. In order to estimate the spatial structure of soil moisture at different scales, it is important to first identify an effective scale based representation of heterogeneity with respect to its effect on soil moisture spatial distribution. In this study, we are attempting to devise a 3x3 heterogeneity matrix to accurately represent the existing heterogeneity in a region at the field, watershed and regional scale. Heterogeneity in terms of soil, vegetation and topography is being investigated for 2 hydro-climates (humid and sub-humid). An exhaustive sensitivity analysis using the Community Land -Surface Model (CLM) is being conducted to determine the most appropriate scale based parameters (like NDVI, % sand, slope etc.) that can represent the different physical controls. The statistical entropy of each chosen parameter will be used to represent the heterogeneity magnitude of a particular physical control which will form the diagonal of the heterogeneity matrix. The remainder elements of the matrix will be estimated as the joint entropy and will represent the correlation between the different physical controls.