Two-months of ASAR (ENVISAT) wave data (February and March 2003) provided by the European Space Agency, have been analysed and compared to wave parameters computed by the wave model WAM. To remove erroneous data, filtering criteria depending on the ratio of signal to noise and the retrieved wind speeds have been used. Thereafter, we performed a statistical analysis between the observed and modeled wave parameters. In most cases, the low-frequency wave heights obtained from the ASAR wave spectra, overestimate the wave model prediction. In other respects, strong correlation (greater than 0.85) is obtained between retrieved ASAR and modeled wind speeds, and this concern more than 75% of the data. A run of the wave model with assimilation of ASAR wave spectra has been performed for a period of five days in March 2003 and compared to runs without assimilation and with assimilation of altimeter wave heights from ERS2 only. A statistical analysis of the wave heights obtained with different runs at crossover locations of JASON-1 and ENVISAT is presented and shows the best estimate when using the ASAR spectral information.
The possibility of using directional wave spectra as a tool to improve ocean wave forecasting is an active area for basic and applied research. Since last year, detailed spectral information of the sea states have been globally and continuously provided by the ASAR on board ENVISAT satellite (European Space Agency, ESA). In a previous work (Aouf et al. 2003), which describes an assimilation of a limited 2-day of ASAR wave spectra, it has been shown that one obtains a better estimate of swell mean parameters (wave height, direction and period) when assimilating ASAR data. It was also indicated that in the forecast period the impact of the assimilation stays longer in comparison with the assimilation of altimeter data.