一个稳态Kalman滤波风暴潮数值预报模式

利用Kalman滤波资料同化技术将海洋站水位观测资料融入二维线性风暴潮模式中,研制具有资料同化能力的风暴潮预报模式,改进风暴潮模式计算结果.通过在风暴潮模式的动量方程中加入模式噪声项来修正模式本身和气象强迫力的不确定性.确定性模式的输出通过带有观测噪声的观测方程与可利用的海洋站的潮位观测资料联系起来.假定初始的模式噪声和观测噪声满足均值为0的高斯分布,用迭代法得到计算区域的状态向量的稳态Kalman滤波,进而得到风暴潮模式输出的最优线性校正结果.利用这种资料同化技术,对1956年发生在东海的一次强风暴潮过程进行了后报试验,结果表明,该同化方法对短期风暴潮水位后(预)报有一定的改进.

A numerical storm surge forecast model with Kalman filter

Kalman filter data assimilation technique is incorporated into a standard two dimensional linear storm surge model. Imperfect model equation and imperfect meteorological forcing are accounted for by adding noise terms to the momentum equations. The deterministic model output is corrected by using the available tidal gauge station data. The stationary Kalman filter algorithm for the model domain is calculated by an iterative procedure using specified information on the inaccuracies in the momentum equations and specified error information for the observations. An application to a real storm surge that occurred in the summer of 1956 in the East China Sea is performed by means of this data assimilation technique. The result shows that Kalman filter is useful to storm surge forecast and hindcast.