The utility of satellite soil moisture products as input for land surface models of water and energy fluxes can be greatly improved by robust downscaling algorithms, calibrated over a wide range of environmental settings. In this study, we present the application of a multifractal-based statistical downscaling scheme over contrasting climatic regimes, using 800-m aircraft-derived soil moisture data collected during three field campaigns conducted in: (i) a temperate region in Oklahoma (SGP97); (ii) an agricultural area with sub-humid climate in Iowa (SMEX02); and (iii) two semiarid areas in Arizona and Sonora (Mexico) (SMEX04). We first demonstrate the presence of multifractality in  fields over the scale range from 0.8 km (aircraft footprint) to 25.6 km (satellite footprint) over most wetness conditions. Next, we estimate in each site an empirical regional calibration relation linking model parameters with the spatial mean soil moisture and coarse-scale predictors that account for topography, soil texture and land cover. We show how the use of these calibration relations allows reproducing the sub-footprint soil moisture distribution in a broad range of conditions, except for cases where specific physiographic features introduce spatial inhomogeneity in the soil moisture field. Once calibrated, the multifractal-based downscaling model is then applied to study the relation between spatial variability and mean soil moisture. For SMEX04, we present an application of the downscaling method with real (AMRS-E) and synthetic (SMAP) coarse-scale observations trained on aircraft data at C-band (PSR) and L-band (2D-Star). This comparison yields insight into the added value of the downscaling algorithm for upcoming soil moisture missions.