基于均生函数模型的高温预测

Heat prediction based on the Models of mean-generating function

目前,常用的高温预测方法主要有灰度、指数等模型。文章采用均生函数模型预测各气象观测站未来一个月的最高气温。气候数据按照时间序列有不同时间尺度震荡的特征,传统时间序列预测模型的预测值往往趋于平均值,而均生函数模型有很好的预测结果。最后,采用ArcGIS系统将未来可能超过35℃的区域生成图像产品,供用户用Web方式查看实时的预报信息。

At present, the commonly used heat prediction methods are mainly gray model, index modeL This paper predicts the next month highest temperature of weather stations by mean-generating function model. Climate series data have shock characteristic according to different time scales, predicted values of traditional time series forecasting model tended to mean, but mean-generating function model has good predictive results. Finally, the system will generate image products which marked areas that may exceed 35 ℃ by Arc GIS system, then users could view real-time forecast infor- mation by the web.