一种短期气候预测集成技术的应用研究

以递归正权综合决策方案为基础，结合实际情况，首先对多种预报方法进行筛选，选取预报效果较好并且较为稳定的几种方法参与集成，并将集成结果写成Micaps数据格式，最终以图形方式显示出来。评分结果表明：该集成方法对温度预报和灾害指数预报较权重平均集成法效果好；而对月雨量预报和季度雨量、温度预报则与权重平均方法相当；但是对季度雨量、温度预报来说，集成后预测精度有明显提高。同时，将集成结果通过Micaps工作平台以图形方式显示，不仅客观、定量，而且更一目了然．

The Application of a Composite Technique for Short-Term Climate Forecasts

Based on the synthetic decision plan of recurrent positive weights, many predicting methods are filtered and the steady and more effective methods are composed, then the result is written in Micaps data format and showed in graphics mode. The score result shows that, the composite method is more effective than the composite method of mean weights to the temperature and disaster index prediction, but it is the same to the month rainfall and the season rainfall predictions, the season rainfall and temperature predictions, it enhances greatly the forecasting precision. Additionally, by Micaps working platform showing the composite result in graphics mode, it is not only objective and quantitative, but also visual.