改进型经验正交函数海洋声速剖面预报方法

An improved empirical orthogonal functions(EOF) on the forecast of ocean sound speed profile

鉴于深海温跃层以下往往声速值缺乏,声速剖面不完整的原因,提出一种声速剖面的预报方法:在传统经验正交函数预报法基础上,首先改进协方差矩阵的求解方法,将原始数据的空间信息和时间信息有效地融合到协方差矩阵中,通过由大量实测数据统计得出的时间函数的经验公式,得到合成剖面,将二者结合,把不完整剖面垂直向下延拓到海底,较为有效地解决了传统方法求解协方差矩阵和时间函数较粗糙的问题,给出了完整的海洋声速剖面的准确预报。实测数据检验结果表明,改进方法的预报精度比传统方法有了很大提高。

Based on empirical orthogonal functions（EOF）, the article gave a prediction method of sound speed profile, the depth of which is below the thermocline. First, it improved the method to get the solution of covariance matrix and mix the position information of the original data and the time information together into the covariance matrix; and then, gave several empirical time function formulas calculated with a lot of real statistics. It could predict the acoustic speed profile. Compared with the traditional prediction method with the real statistics, the precision of the improved method could make a better improvement than the prediction that used the traditional EOF.