Presented at the 2013 SMOS-Aquarius Science WorkshopA comparison is reported of Aquarius and SMOS brightness temperatures over land and ocean. This is done by reprocessing the SMOS data to match the footprints of the Aquarius radiometers. Although the focus of Aquarius has been calibration over the ocean to retrieve sea surface salinity, it is planned to provide a soil moisture product and consequently calibration of the full dynamic range from ocean to land is important. It is a challenge to validate TB over land (as opposed to oceans) using models because there are more factors that contribute to TB and the footprints are more heterogeneous. On the other hand, the inter-comparison of the two sensors provides a consistency check on the calibration.
Observations made at the same frequency, polarizations, and incidence angle by multiple instruments/platforms (at concurrent locations and times) should be consistent with each other. Both SMOS and Aquarius have an equatorial overpass time of 6 AM (SMOS-ascending; Aquarius-descending). Thus, the TB observations from the Aquarius and SMOS missions should be similar. This provides us with an opportunity to cross-calibrate and compare the TB observations. Consistent calibration across all satellite missions is critical to eventually developing a long term climate data record of L-band TB observations. A physically-based algorithm that spans multiple L-band missions requires consistent input observations for the development of a long term environmental data record.
The SMOS data were reprocessed to match the incidence angles and sizes of the three Aquarius radiometer footprints. Only the alias free portions of the SMOS orbit were used in the comparison. The alias free portions of the orbit provide brightness temperatures with the lowest NeÄT. The current version of the Aquarius TB (v1.3.7) shows a very strong correlation with the SMOS observations. The Aquarius TB compared well with SMOS observations over oceans, however, the Aquarius brightness temperatures are biased warmer than the SMOS observations over land. There is a bias of about 8K for h-pol and 6K for v-pol observations. These combined results provide strong evidence of inconsistencies in the relative calibrations of Aquarius and SMOS. The analysis will be re-done with the Aquarius data v2.0 (to be released in February 2013).