利用T_(106)L_(19)进行台风路径预报试验

TYPHOON TRACK PREDICTION EXPERIMENTS USING T_106L_19

为能提供较长时效的台风路径数值预报,利用国家气象中心中期数值预报业务谱模式T_(106)L_(19),在其第一猜测场或分析场中加入人造台风模型,进行了96小时的台风路径预报试验。结果表明:由于人造台风模型的加入,使T_(106)L_(19)对台风路径的预报能力有较大的提高,尤其是当T_(106)L_(19)的第一猜测场或分析场中台风涡旋相当弱的情况下效果更明显。另外我们采取三种不同的嵌入方式在T_(106)L_(19)的资料场中加入人造台风。三种方式分别是:分析前加入人造台风模型(将人造台风模型加在第一猜测场上-T_(106)L_(19)的6小时预报场);分析后加入人造台风模型(将人造台风模型加在分析场上);分析前及分析后两次加入人造台风模型。个例试验结果表明:从平均情况来看分析前加入人造台风模型在台风路径预报前期有较好的表现,其24小时预报距离误差为184.5km,48小时为278.1km,72小时为515.6km。结果的分析还发现:T_(106)L_(19)的初值化使所加入的人造台风涡旋有较大的平滑,初值化后的中心气压值一般在995 hPa左右.这对于台风路径的预报会有较大的影响。

The 96 h typhoon track prediction experiments were carried out using medium range forecasting system of NMC by adding BOGUS typhoon into the first guess field or the analysis field in order to provide longer time typhoon track forecast. The results show that T106L19 could provide a better forecast to typhoon tracks when the BOGUS typhoon was added, especially when the typhoon vortex is even weaker. The sensitive experiments on where to add the BOGUS typhoon show that the results from adding the BOGUS typhoon into the first guess field are better. The results also show that the initialization smooth the BOGUS typhoon a lot and this will affect the typhoon track prediction.