Kalman滤波风暴潮数值预报四维同化模式研究进展

ADVANCE IN NUMERICAL STORM SURGE FORECAST MODEL WITH KALMAN FILTER

本文首先介绍了Ｋａｌｍａｎ滤波在风暴潮数值预报中的应用，特别介绍了近年来国际上发展的一些在实际中可行的次优化Ｋａｌｍａｎ滤波算法。并通过一个稳态Ｋａｌｍａｎ滤波风暴潮数值预报模式的实例表明，使用资料同化可以明显改进风暴潮后报结果；资料同化能够提供更为合理的预报初始场，对风暴潮的短期预报有较明显的改进。一旦没有资料同化到模式中去，预报结果很快接近确定性模式。

In this paper, the application of Kalman filter data assimilation in storm surge numerical model is introduced, and especially some suboptimal schemes (SOS) of Kalman filter developed in recent years are also introduced in the world. A case of storm surge numerical mode by using stationary Kalman filter algorithm indicates that the result of storm surge hindcast has a great improvement by using the data assimilation. Data assimilation can give a better initial field, therefore the short-range storm surge forecast has an obvious amelioration. Once no real-time data is assimilated in model, the fOrecast will approach to the result of deterministic model quickly.