浙江省温度和相对湿度释用技术及其效果检验分析  

Interpretation method and effect validation of temperature and relative humility forecast in Zhejiang province

作　　者：李超 裘薇 娄小芬 王霁吟 郝世峰 

机构地区：[1]浙江省气象台,杭州310017 [2]浙江省气象科学研究所,杭州310008 [3]南京大学大气科学学院,南京210023

出　　处：《气象科学》2016年第4期556-561,共6页Journal of the Meteorological Sciences

基　　金：国家自然科学基金资助项目(41405047);公益性行业(气象)科研专项(GYHY201306010);浙江省气象科技计划项目(2013ZD02)

摘　　要：基于多模式的要素预报和浙江省乡镇站点观测资料,结合全省天气统计特点,利用最优集成方案,改进了温度和相对湿度的1～7 d预报。统计检验发现,模式的温度预报在浙江省中南部主要表现为系统性偏差,在浙江省北部平原地区主要为随机误差。使用滑动平均误差订正后,浙中南等地形复杂地区温度预报的系统性偏差明显减小。在模式订正基础上,使用动态集成进一步减小了浙北平原地区温度预报的随机误差。基于温湿关系,使用改进后的温度预报对相对湿度预报进行订正。与传统的加权平均方法相比,改进后的温度预报均方根误差减小16.7%,相对湿度预报均方根误差减小13.8%,对改善浙江省精细化预报有一定参考意义。Based on the multi-model forecast products and observational data in Zhejiang rural sta- tions, combined with the statistical features of climate in the whole province, the prediction skills of 1-7 days＇ prediction of Surface Air Temperatures （SATs） and relative humidity at 2 m are improved by using the objective consensus forecasting system. The errors of SATs tend to be systematic bias in the central and south areas of Zhejiang, while the stochastic errors are dominated in the northern part of Zhejiang. The sys- tematic biases significantly decreases in the prediction of the complex terrain in the central and south areas of Zhejiang after correcting by moving average error . The stochastic errors are further deduced by using multi-model ensemble after the SAT revisions. On the basis of the improved SAT, the relative humidity at 2 m are corrected by the equation of temperature-humidity relation. Compared to the traditional products, the Root Mean Square Error （RMSE） reduces by 16. 7% with the improved SAT products, and 13. 8% for the relative humidity at 2 m. The improved predictions are helpful and important for the weather forecast.

关 键 词：最优集成方法 误差分析 释用技术 

分 类 号：P456[天文地球—大气科学及气象学]