Collection Of The Kalman Filter Data Assimilation Methods In The Numerical Prediction Of Sea Surface Temperature

Remotely sensed observations of sea surface temperature(SST) have been assimilated into MIT general circulation model(MITgcm) for Bohai and Yellow sea, using the Ensemble Kalman Filter(EnKF).An EnKF assimilation system has been set up to have a study on ocean monitor and forecast for Bohai and Yellow sea.In chapter two and three the characteristic and difference scheme of MITgcm model and basic theory and application of EnKF method are introduced respectively.In chapter four, An EnKF assimilation system is set up and the properties of the assimilation scheme are examined through a hindcast validation experiment. A sensitivity analysis of cut radius shows that 9km in meridional size and 18 in zonal size is appropriate choice. The relation between the number of ensemble members and root-mean-aquare(RMSE) has been discussed, which showed an ensemble size of about 80 is sufficient to obtain an EnKF that works satisfactorily, further, A resolvent about how observations enter the system and how to bring EnKF into effect have been discussed.In chapter five, the difference between the analysis and observed data is similar to forecast in distributing trend and to observe in numerical value. In addition, the difference between forecast and free-run model data can not forecast SST effectively because their discrepancy is visible. The variance in forecast typically decreased as a function of time. Forecast absorbs the information of observation well to update the background field. On the side, Some questions, such as the influence on the variance in forecast with different numbers of ensemble members, has also been considered.