Disaggregation as a top-down approach for evaluating 40 km resolution SMOS data using point-scale measurements: Application to AACES-1

The SMOS (Soil Moisture and Ocean Salinity) satellite provides soil moisture data at about 40 km resolution globally. Validation of SMOS data using in situ measurements is complicated due to the large integrated scale of remote sensing observations. Nevertheless, different approaches can be used to circumvent the direct comparison. One is to upscale ground measurements using aggregation rules. Another is to downscale (or disaggregate) remote sensing data at the representativeness scale of ground measurements. This study combines both approaches to make ground and remote sensing data match at an intermediate spatial scale. On one hand, the local-scale in situ soil moisture data collected during the first AACES (Australian Airborne Calibration/validation Experiments for SMOS) are aggregated to 4 km resolution. On the other hand, a disaggregation methodology of SMOS data based on 1 km resolution MODIS (MODerate resolution Imaging Spectroradiometer) data is implemented at 4 km resolution over the Murrumbidgee catchment, the site of the AACES campaign. Results indicate a correlation coefficient between disaggregated and ground observations of 0.92. The y-intercept of the linear regression between disaggregated and ground observations is very close to 0. However, the slope of that line is 0.44 only. This seems to highlight an issue with either the dielectric constant model or the roughness parameter value currently used in the SMOS retrieval algorithm.