Initialization of seasonal forecasts assimilating sea level and temperature observations

In this paper, the combined assimilation of satellite observed sea level an omalies and in situ temperature data into a global ocean model, which is us ed to initialize a coupled ocean-atmosphere forecast system, is described. The altimeter data are first used to create synthetic temperature observati ons, which are then combined with the directly observed temperature profile s in an optimum interpolation scheme. In addition to temperature, salinity is corrected based on a preservation of the model's local temperature-salin ity relationship. Coupled forecasts with a lead time of up to 6 months are initialized from the ocean analyses and the impact of the data assimilation on both the ocean analysis and the coupled forecasts is investigated. It i s shown that forecasts of sea surface temperature anomalies in the Nino-3 a rea can be improved by initializing the coupled forecast model with the oce an analysis in which temperature and altimeter data are assimilated in comb ination. The results further imply that a good simulation of the salinity f ield is required to make optimum use of the altimeter data.