Sea Surface Salinity Retrieval Error Budget within the ESA Soil Moisture and Ocean Salinity Mision

Memoria de tesis doctoral presentada por Roberto Sabia para obtener el título de Doctor por la Universitat Politècnica de Catalunya (UPC), realizada bajo la dirección del Dr. Adriano Camps Carmona y de la Dra. Mercè Vall-llossera Ferran.-- 184 pagesSatellite oceanography has become a consolidated integration of conventional in situ monitoring of the oceans. Accurate knowledge of the oceanographic processes and their interaction is crucial for the understanding of the climate system. In this framework, routinely-measured salinity fields will directly aid in characterizing the variations of the global ocean circulation. Salinity is used in predictive oceanographic models, but no capability exists to date to measure it directly and globally. The European Space Agency’s Soil Moisture and Ocean Salinity (SMOS) mission aims at filling this gap through the implementation of a satellite that has the potential to provide synoptically and routinely this information. A novel instrument, the Microwave Imaging Radiometer by Aperture Synthesis, has been developed to observe the sea surface salinity (SSS) over the oceans by capturing images of the emitted microwave radiation around the frequency of 1.4 GHz (L-band). SMOS will carry the first-ever, polar-orbiting, space-borne, 2-D interferometric radiometer and will be launched in early 2009. Like whatsoever remotely-sensed geophysical parameter estimation, the retrieval of salinity is an inverse problem that involves the minimization of a cost function. In order to ensure a reliable estimation of this variable, all the other parameters affecting the measured brightness temperature will have to be taken into account, filtered or quantified. The overall retrieved product will thus be salinity maps in a single satellite overpass over the Earth. The proposed accuracy requirement for the mission is specified as 0.1 ‰ after averaging in a 10-day and 2ºx2º spatio-temporal boxes. In this Ph.D. Thesis several studies have been performed towards the determination of an ocean salinity error budget within the SMOS mission. The motivations of the mission, the rationale of the measurements and the basic concepts of microwave radiometry have been described along with the salinity retrieval main features. The salinity retrieval issues whose influence is critical in the inversion procedure are: • Scene-dependent bias in the simulated measurements, • Radiometric sensitivity (thermal noise) and radiometric accuracy, • L-band forward modeling definition, • Auxiliary data, sea surface temperature (SST) and wind speed, uncertainties, • Constraints in the cost function, especially on salinity term, and • Adequate spatio-temporal averaging. A straightforward concept stems from the statement of the salinity retrieval problem: different tuning and setting of the minimization algorithm lead to different results, and complete awareness of that should be assumed. Based on this consideration, the error budget determination has been progressively approached by evaluating the extent of the impact of different variables and parameterizations in terms of salinity error. The impact of several multi-sources auxiliary data on the final SSS error has been addressed. This gives a first feeling of the quantitative error that should be expected in real upcoming measurements, whilst, in another study, the potential use of reflectometry-derived signals to correct for sea state uncertainty in the SMOS context has been investigated. The core of the work concerned the overall SSS Error Budget. The error sources are consistently binned and the corresponding effects in terms of the averaged SSS error have been addressed in different algorithm configurations. Furthermore, the results of a salinity horizontal variability study, performed by using input data at increasingly variable spatial resolution, are shown. This should assess the capability of retrieved SSS to reproduce mesoscale oceanographic features. Main results and insights deriving from these studies will contribute to the definition of the salinity retrieval algorithm baselinePeer Reviewe