| On purpose to fix a more accurate initial state of the ocean numerical Simulation with making the best of observations and provide a more robust reanalysis ocean data in order to make a correct climate prediction, National Climate Center develop the second generation oceanic data assimilation system (BCC_GODAS2.0), which can coordinately assimilate manifold observations using the Three- Dimension Variational scheme.In this thesis, the model and scheme which the system used are introduced. what's more, to evaluate the effect and stability of the system, we design two experimental: in the first experiment, we assimilate Several observation data such as the satellite altimeter data,sea surface temperature, the GTSPP data and so on from 1990 to 2009,the another one is a experiment with none observation data in the same mean time period. TAO, OISST and SODA data sets are used to check the experiment result.The assessment indicated that, the assimilation result can effectively reduce the surface temperature simulation error, allowing the salinity simulation is more accurate too. The simulation of the entire layer of the oceans have improved, especially obvious in the mixed layer, RMSE of temperature can reduced 1.5℃, and RMSE of salinity to 0.6psu. The assimilation can effectively reduce the SST warm biases in tropical Pacific. By contrasting the time series of TAO data, we can see the seasonal temperature variation of tropical Pacific are seized much better.In addition, diagnostic analysis of the experiment results are discussed. The results indicated that, assimilation can reflect NINO index and Indian Ocean Dipole more accurate, and spatial structure function of the EOF of observed monthly averaged tropical Pacific SST variability confirms that the interannual variation pattern of the sea surface temperature (SST) of the assimilation result is more close to the observations. The robust effect and stability of the system lay a foundation for the Climate Prediction in National Climate Center, what's more, the 20 years' assimilation result of the experimental could be used to the climate research as a good reanalysis ocean data. |
