摘要
利用Kalman滤波方法,对T106数值预报产品进行了误差修正和预报优化。用1995-1996年夏季(5-8月)T106数值预报产品资料建立预报模型,并提供递推参数初始值,其后以1997年夏季(5-8月)T106资料为独立样本对副高面积指数和脊线指数进行动态订正预报。预报效果的对比分析表明,Kalman滤波方法较其它统计预报方法的自适应能力更强,能够对预报对象提供更为准确、有效的跟踪和描述,对数值预报产品的副高预报误差修正效果良好。
Kalman filtering was used to revise errors and optimize the forecast by T106 numerical forecast products. Based on the data during the period from May to August in 1995 and 1996, the forecasting model was developed and the initial value of recursive parameters was identified. The data of numerical forecast products in 1997 were used as an independent sample to revise the subtropical high index dynamically. Analysis and comparison to forecast effect showed that Kalman filtering is more adaptive compared to other statistics forecast method, which can provide a more, accurate and effective description and track to the predictands, and it also has good effect to error correction of numerical forecast products for the subtropical high.
出处
《热带气象学报》
CSCD
北大核心
2006年第6期661-666,共6页
Journal of Tropical Meteorology
基金
国家自然科学基金项目(40375019)
中国博士后科学基金项目(2004036012)
江苏省博士后科研资助计划项目(2004087)共同资助
