Assessment of the FOAM global data assimilation system for real-time operational ocean forecasting

An operational system to forecast the state of the global ocean a few days ahead has been implemented at the UK Met. Office (UKMO). The system, known as the Forecasting Ocean Assimilation Model (FOAM), consists of a 1°×1° resolution numerical ocean model driven by surface fluxes from the UKMO numerical weather prediction (NWP) suite and a modified successive correction data assimilation scheme for thermal observations. The assimilation scheme is assessed here in a series of 1-year integrations by comparison with ‘independent’ thermal profile observations and climatology. Assimilating temperature observations significantly reduces model errors in the upper ocean and results in temperature analyses that on average are closer to independent observations than climatology. The specific results depend on location and depth. The extent to which the data assimilation scheme is able to compensate for uncertainties in the surface forcing fluxes is also assessed by comparing integrations forced with climatological and NWP fluxes. Assimilating data is able to compensate for uncertainties in the surface heat forcing fluxes and significantly reduces the impact from uncertainties in the surface wind stress.