最优组合预测的短期气候预报建模研究  

Study on Short-term Climate Prediction Model of the Optimal Combination Forecasting

作　　者：农吉夫[1] 金龙[2] 谭福锦[1] 主毅 

机构地区：[1]广西民族大学数学与计算机科学学院,广西南宁530006 [2]广西气象减灾研究所,广西南宁530022 [3]贵州六盘水市气象局,贵州六盘水553001

出　　处：《数学的实践与认识》2008年第8期91-97,共7页Mathematics in Practice and Theory

基　　金：国家自然科学基金项目(40675023);广西民族大学青年科研基金项目(2007QN23)

摘　　要：以预报量序列建立均生函数短期气候预报模型及根据500hPa月平均高度场预报因子分别建立的BP网络模型、回归预报模型为基础,用"误差绝对值和最小"作为最优准则,建立月平均降水量的短期气候组合预测模型.采用线性规划方法计算得到组合预测模型的各权系数,对这种短期气候组合预测模型的预报能力进行了分析研究,结果表明,该组合预测模型的预报精度优于各子方法,具有很好的应用价值.The short-term climate prediction model of mean generating function established by using the prediction, BP network model and regression prediction model established by using the predictors of 500hPa height, the optimal combined forecasting model with the monthly mean rainfall is established by adopting ＂the least sum of the error absolute value＂ as the optimal rule. The weight of the combined forecasting model is determined by linear programming. The prediction ability of this short-term climate model of the combined forecasting is studied. The result shows that the Combined forecasting model is superior in the precision of prediction compared to the submodels and is of practical value.

关 键 词：短期气候预测 组合预测 线性规划 

分 类 号：P456[天文地球—大气科学及气象学]