Development of a near realtime SST product for the Mediterranean forecasting system

摘要

Operational oceanography requires availability of remotely sensed data, for example sea surface temperature (SST), in near realtime (NRT). A system is presented that makes use of a combination of state of the art NASA Pathfinder SST (PFSST) algorithm and cloud detection procedures both adapted to operate in NRT. A novel cloud detection algorithm that makes use of a reference image based upon recent SST fields recovers data over coastal areas affected by sharp SST fronts that are discarded by the standard PFSST quality flag. The resulting increased SST coverage is visually checked, to remove residual cloud contamination, by a trained analyst prior to input to the objective analysis package in turn adapted for satellite-derived SST observations. The output daily gridded data in which the gaps due to clouds have been filled by the optimal interpolation module are assimilated into the Mediterranean Forecasting System Toward Environmental prediction (MFSTEP) ocean forecasting system on a weekly basis.