HYCOM模式SST的预报误差订正

Study on the correction of SST prediction of HYCOM

尝试了一种新的通过自回归模型进行模式预报订正的方法:通过1992—2006年的卫星遥感和NERSC-HYCOM模式模拟的SST资料,计算模式15 a的历史误差序列,建立自回归模型AR(p),来估算2007年和2008年两年的误差,并用来对输出的模式资料进行修订,对修订后的模式预报结果进行检验。可以得出:经过数据订正以后,模式SST的均方根误差明显减小,相关系数增大,模式误差在很大程度上得到消减,订正效果明显;同时,订正效果也存在时间和空间上的差异,不同时刻,不同海域订正效果不尽相同。该方法计算过程简单,计算量小,便于实现。总体而言,利用本文方法对模式SST的预报订正具有很强的操作性和可应用性。

A new error-correction model was proposed based on the AR（p） in the paper. Based on the historical SST error sequence derived from the satellite remote sensing data and model data from 1992 to 2006,the AR（p）was built to estimate the model error of 2007 and 2008, and to correct the model data.The results showed after being corrected, the MSE（Mean Squared Error） of model SST reduced significantly and the correlation coefficient increased, reflecting good effect of this method. The effects of correction were changed in time and space. The method had its own advantage： simple process, small calculation, easy to implement. The method we used to correct model error was very operable and applicable.