用聚类预报模型预报三峡水库来水量的探讨

为了开展三峡水库入库来水量长期预报，从北半球500 hPa高度场、太平洋海温场、北半球100 hPa高度场、北半球地面气压场与74项环流特征指数作为预报因子，采用回归方程求解和回归效果检验，建立回归聚类预报物理模型，对三峡水库6～8月各月平均来水量进行了长期预报分析。结果表明，三峡水库6~8月份各月来水量回归聚类预报的技术路线是可行的，正、负距平符号一致的概率能达到80%以上，效果较好，基本上可以满足生产需要。回归聚类预报物理模型具有一定的应用推广价值。

Discussion on inflow of Three Gorges Reservoir forecasted by clustering model

In order to carry out mid-long period forecast of inflow of Three Gorges Reservoir,we established the regression clustering physical model taking 500 hPa height field of north hemisphere,sea surface temperature field in Pacific Ocean,pressure field on the ground of north hemisphere,100 hPa height field of north hemisphere,and 74 circulation feature indexes as the forecasting factors.The established model is solved by regression equation and checked by regression effects.We made a mid-long term period foreca...