基于海浪谱分解与重构的资料同化试验

Assimilation experiment based on wave spectrum decomposition and reconstruction

针对有效波高资料提出一种海浪谱分解与重构的资料同化方案:利用历史时段内的有效波高观测资料和模式计算波高场,采用最优插值方法得到分析波高场;在WAVEWATCH-Ⅲ模式的波浪能量密度谱和有效波高分析值之间引入一个变异系数矩阵,描述模式的误差,以此为状态向量构建卡尔曼滤波系统,对分解过的海浪谱进行修正和重构,得到同化后的海浪谱初始场。利用美国阿拉斯加湾北部海域的7个浮标站进行同化和72h预报试验,对连续1个月的预报结果进行统计表明:采用该同化方案后24h预报结果的有效波高均方根误差比未同化的结果降低了0.13m;同化方案对预报效果的影响可持续36h左右,随着预报时效延长,同化的效果减弱。

In the near-shore wave forecasting model,the ocean buoy data assimilation method based on wave spectral decomposition and optimization is proposed.The calculated wave energy spectrum before the initial time is studied with orthogonal decomposition,and the result,combined with synchronous buoy observations of significant wave height value is used to construct a Kalman filtering system,and the initial wave energy density spectrum of wave model is revisesd by the multi-time significant wave height value.This method has been applied to 72 hwave forecast experiments with assimilation 7buoys significant wave height data in the Gulf of Alaska.One month experiments show that method can improve the forecasts of significant wave height at different degree of different prediction time.The mean square error for 24 hforecast of significant wave height reduces 0.13 m.Meanwhile,the effect of initial field after assimilation to forecast will extend 36 hor so,but the assimilation effect is weakened by extend the prediction time.