ESA's Soil Moisture and Ocean Salinity (SMOS) satellite was launched on the 2nd of November 2009 from northern Russia. The SMOS single payload consists of a synthetic aperture radiometer operating at L‐band (1.4 GHz). It is a challenging mission since this is the first time that such an instrument is put into orbit, and that surface salinity and soil moisture are measured from space. SMOS aims at measuring sea surface salinity (SSS) over the ocean with an accuracy of 1 psu for each overpass at 30‐50 km spatial resolution or 0.1 psu after averaging areas of 200 x 200 km in 10‐30 days. The instrument provides global salinity and soil moisture maps every 3 days. In the SSS retrieval (level 2 or L2) operational processing chain the Geophysical Model Function (GMF), which relates the emissivity of the sea to the SSS (among other geophysical parameters), is defined as the sum of two contributions: the first one is the emissivity due to a flat sea, which is presently assumed to be well explained by the Klein and Swift 1977 model; the second one is the change in emissivity due to the sea surface roughness. In the second term, three different models have been considered in the operational processor, therefore producing three different retrieved salinities. Two of the models are theoretically based, while the third is a fully empirical linear approach, which has been derived from two pre‐launch field campaigns. However, the limited amount of data of sea surface emissivity collected in the different prelaunch campaigns were not representative of the global ocean, nor representative of all sea state conditions, thus limiting the derivation of the empirical GMF roughness induced model. To tune the empirical model and decide on which parameters significantly modulate the emissivity, a large amount of auxiliary data, co‐located in space and time with the SMOS measurements, is required. With the SMOS launch, global calibrated brightness temperature measurements have recently become available. These data, together with in‐situ data (e.g., buoys) and model outputs (e.g., atmospheric and ocean models), are being used to review and redefine (inclusion of non‐linear terms) the GMF, in particular the fully‐empirical roughness term. The tuning methodology together with an assessment of the fully‐empirical roughness model performance will be presented at the conference. This work will set the grounds for the future development of a fully‐empirical GMF (i.e., where the contributions from all parameters, including the flat sea ones, are empirically derived)

I Encuentro de la Oceanografía Física Española (EOF), 13-15 de octubre 2010, Barcelona

Peer Reviewed