Accuracy of SMOS Level 3 SSS Products Related to Observational Errors

Abstract

The Soil Moisture and Ocean Salinity (SMOS) mission will provide, for the first time, satellite observations of sea surface salinity (SSS). The satellite will acquire a large amount of data but with a high observational noise. At level 3 (L3) of the SMOS processing chain, the information will be summarized in gridded products with the aim of synthesizing the information and reducing the error of individual SSS observations. The technique chosen to generate the maps is optimal statistical interpolation (OI). The goal of this paper is to quantify the impact of the observational error on the accuracy of L3 gridded products. The accuracy of the products is estimated using the OI error formulation, which has been extended to include the convolution with a normal error filter. For a given observational error, we estimate the minimum scales that can be resolved with a prescribed accuracy (at the expense of losing the variance associated with shorter scales). Conversely, for a prescribed accuracy and spatio-temporal resolution of L3 products, we estimate the maximum observational error that will allow the fulfillment of the requirements. Results indicate that a maximum SSS error of about 0.8 (1.1) psu would be enough to obtain an accuracy of 0.1 psu for L3 products with a resolution of 100 km/30 day (200 km/10 day). The statistical errors produced by the OI formulation are compared with the errors obtained for several case studies in order to assess their robustness.