几种不同方法在舟山海岛阵风预报试验中的对比分析  

Comparison of performance of test forecast of extreme wind based on different methods in Zhoushan archipelago

作　　者：胡波 杜惠良 俞燎霓 

机构地区：[1]浙江省气象台,浙江杭州310017

出　　处：《海洋预报》2015年第3期43-50,共8页Marine Forecasts

基　　金：浙江省重大科技专项重点社会发展项目(2011C13044)

摘　　要：设计两种常用的阵风预报方法,即固定权重插值法和天气分类回归法,并在此基础上利用统计降尺度技术,结合地面精细化模式预报,分别设计了直接经验映射法和间接经验映射法。然后利用2006—2012年的NCEP/NCAR分析数据、舟山群岛68个自动站的小时极大风速数据,结合WRF模式初始风场输出,对2012年7—12月一天2次(08时和20时)进行试报。结果表明:直接和间接经验映射预报的阵风数据在方差和极值评估表现均优于固定权重插值法和天气分类回归法,在单站阵风预报方面,经验预报也表现良好,说明经验映射法能较好的将经验数据映射到预报空间,使预报数据的空间分布与当地地形保持一致,弥补单纯动力模式的不足。Two common methods including fixed weighted interpolation and weather classification methods were designed, and by the use of statistic downscaling technique and ground high resolution model forecast, another two methods including indirect empirical projection and direct empirical projection were developed. Using the NCEP/NCAR reanalysis data from 2006 to 2012 and extreme wind data at 68 observatory stations in Zhoushan archipelago, combined with WRF initial wind field output, the forecast test from July to December in 2012 was carried out. The results of assessment show that the indirect and direct empirical projection methods perform obviously better than other two common methods on wind characteristics of variance and extremum; and on station forecast the empirical projection method also performs well. The Empirical Similar method could fully interpolate the data from the experience cases to forecast space, and makes the forecast data are corresponding to geographical features, which make up for the inadequacy of numerical model.

关 键 词：经验映射 阵风 预报 评估 

分 类 号：P732.4[天文地球—海洋科学]