Potential for downscaling soil moisture maps derived from spaceborne imaging radar data

The presence of nonlinear relationships between surface soil moisture and various hydrologic processes suggests that grid-scale water and energy fluxes cannot be accurately modeled without subgrid-scale soil moisture information. For land surface and energy balance models run over continental- to global-scale domains, accurate fine-scale soil moisture observations are nearly impossible to obtain on a consistent basis and will likely remain so through the next generation of soil moisture remote sensors. In the absence of such data sets, an alternative approach is to generalize the statistical behavior of soil moisture fields across the relevant range of spatial scales. Downscaling procedures offer the possibility that the fine-scale statistical properties of soil moisture fields can be inferred from coarse-scale data. Such an approach was used for a 29 x 200 km transect of 25 m active radar data acquired over Oklahoma by NASA's spaceborne imaging radar imaging (SIR-C) mission on April 12, 1994. Using a soil dielectric inversion model, the radar data were processed to provide estimates of surface soil dielectric values, which can be equated to volumetric soil moisture content. The soil moisture field along each strip was analyzed for evidence of spatial scaling for scales ranging from 100 to 6400m. Results suggest that a spatial scaling assumption may not always be an appropriate basis for a downscaling approach. Prospects for the development of a more robust downscaling procedure for soil moisture are discussed.