基于SWAN模式的卫星遥感海浪方向谱集合同化研究

The main method about the data assimilation of directional wave spectra so far is called the statistical interpolation based on the partition technique of wave spectra. The kernal of this method is that several key wave parameters which are actually assimilated, such as significant wave height, wave mean period and direction, are integrated from two-dimensional ocean wave spectrum. So this method discards the essential characteristic and meaning of data assimilation of directional wave spectra. The background error covariance matrix, which is commonly denoted as the exponential function in wave data assimilation, is one of the most important parameters affecting the assimilation result. The background error covariance is in fact the function of physical processes governing the meteorological situation. And the exponential function distribution cannot get the properties of inhomogeneity, anisotropy and the time variability of background error covariance. This project will develop the new data assimilation method of directional wave spectra which directly assimilates the spectral elements from the advanced synthetic aperture radar on the satellite Envisat. This project also constructs the sample ensemble of the background error of directional wave spectra and discusses the estimation and structural analysis about the background error covariance matrix. The accomplishment of this project will result in the new method and model of data assimilation of directional wave spectra. It is also considered as the preliminary research of the operational application of directional wave spectra obtained from the Bopuyi which is carried on the China-France ocean satellite planned to launch next year.

目前的海浪方向谱数据同化主要采用的是基于谱分割技术的统计插值方法，方法核心仍然是从二维海浪谱积分得到主要波参数：有效波高、波平均周期和方向等，然后对这些积分波参数进行最优插值同化，这失去了海浪方向谱数据同化的本质特征和意义。模式背景误差协方差矩阵是影响同化效果的重要因素之一，它是实际海浪物理过程的函数；但在海浪数据同化中常被取成随距离增大而减小的指数分布函数，无法体现背景误差协方差的非均质、各向相异和随时间的变化性。本项目拟基于第三代海浪模式SWAN，开展新的针对谱值元素的Envisat卫星高级合成孔径雷达海浪方向谱的数据同化研究，开展海浪方向谱的模式背景误差样本集合的构造和背景误差协方差矩阵的估计分析研究。本项目完成将建立新的海浪方向谱数据同化方法和模型，并为计划明年发射的中法海洋卫星搭载的波谱仪获取的海浪方向谱数据的业务化应用进行预研。

海洋卫星遥感技术的进步带来了日益丰富的海浪观测资料，它们与第三代海浪数值模式的结合促进了海浪数据同化研究的蓬勃发展。虽然海浪数据同化的研究主要是有效波高数据同化，海浪方向谱的数据同化研究相对较少，但其实海浪方向谱数据包含更多的海浪信息，其数据同化研究少主要是因为海浪方向谱观测资料相对较少并较难获取。本项目主要进行了如下相关研究工作：建立针对全谱值元素的卫星遥感海浪方向谱OI数据同化模型与方法；构造和分析海浪模式背景误差的样本集合；建立基于误差样本集合的卫星遥感海浪方向谱OI数据同化技术，相关数据同化实验结果表明该同化技术有效改善了同化后的海浪数值模拟效果；建立观测误差协方差矩阵模型，构造和分析观测误差样本集合；研究探讨如何根据风速误差的发散性来改进海浪波高资料同化，基于有限风区风浪成长公式推导了有效波高偏差和有效周期偏差与风速及风速偏差之间的关系；研究潮汐潮波方程的高精度差分格式和高阶杂交格式，并在此基础上完成了渤黄东海潮汐现象的高精度数值模拟；开展了非线性控制系统的解的稳定性问题研究。