论文部分内容阅读
根据预报值具有最小方差这一要求,详细推导了融合法在观测数据为一维、多维和维数不同的情况下的具体同化表达形式,同时还给出了不同情况下与同化表达式相对应的预报误差公式.利用这些公式,可以用融合法处理常见的海洋观测数据的同化问题.在陆架海模式HAMSOM基础上,以4月份的渤海海表温度为例,我们验证了同化公式的正确性,并给出了同化后较好的同化结果。最后将融合法的同化结果与卡尔曼滤波同化结果进行了对比.比较表明,融合法使用起来更简单,且能有效地处理常见的海洋观测数据.
According to the requirement that the forecast value has the minimum variance, the concrete assimilation form of the fusion method under the one-dimensional, multi-dimensional and the different dimensions of the observed data is deduced in detail. At the same time, the corresponding assimilation expressions are also given under different conditions We can use the fusion method to deal with the common assimilation problem of ocean observation data.On the basis of HAMSOM in the shelf sea model and the April sea surface temperature in the Bohai Sea as an example, we verify the correctness of the assimilation formula , And gives the better assimilation result after assimilation. Finally, the assimilation results of the fusion method are compared with the results of the Kalman filtering assimilation.The comparison shows that the fusion method is simpler to use and can effectively deal with the common ocean observation data.