摘要
在普查影响中国降水量和气温的大气环流系统及其指数的基础上,选取高相关、独立性强的大气环流系统指数作为预测因子,采用过滤式的筛选因子方法,动态建立线性回归方法,构建区域月气候预测模型(Regional Monthly Climate Forecast Model,RMCFM),自主开发操作系统,设置人机交互界面,实现人机交流操作方式,并通过解释应用方法做出中国160个台站的月降水量和月气温预测。实践证明,RMCFM具备天气气候专家系统概念,具有运算速度快,结构清晰易操作等特点,预测准确率较高,应用RMCFM可提高区域气候预测能力。
Various atmospheric systems and the indexes affecting precipitation and temperature in China are examined. The parameters of high independence and high correlation coefficients were chosen as prediction factors to predict the monthly precipitation and temperature of 160 observation stations through downscaling method. By using the filtration method to determine primary prediction factors and establishing the linear regression equations dynamically, the Regional Monthly Climate Forecast Model （RMCFM） is built up, in which the operational interface for communication between people and computer is set up. RMCFM can make monthly climate prediction of monthly precipitation and temperature over 160 stations in China. Practice proved that RMCFM is of high calculating speed and clear structure, and easy to operate. The real-time daily grid reanalysis data in NCEP/NCAR were used as pretreatment data for RMCFM. It is proved that RMCFM improved the capability of regional climate prediction.
出处
《气象科技》
2014年第6期1028-1038,共11页
Meteorological Science and Technology
基金
国家重大科学研究计划(2012CB957803
2012CB957804)
中国气象局成都高原气象研究所高原气象开放基金课题(LPM2012004)
内蒙古自治区气候与气候变化创新团队项目共同资助
关键词
大气环流系统
释用方法
线性回归方程
区域月气候预测模型
atmospheric systems, downscaling method, linear regression equation, Regional Monthly Climate Forecast Model （RMCFM）
