摘要: 本文基于正规化周期回归模型，对新丰江流域月径流量序列的演变规律进行分析，得到序列演变趋势和稳定周期波序列。在此基础上加入人工干预筛选周期，一定程度上避免了提取到的部分伪周期；同时，考虑到不同月份径流量丰枯程度不同，对汛期和非汛期选取不同的前期时刻流量和雨量作为校正因子。结果表明改进后的方法可以提高新丰江月径流量预测精度，可用于中长期径流预报。
Abstract:
In this study we analyzed monthly runoff evolution in XinfengRiver Basinbased on normalized periodic regression forecasting model, and obtained the variable tendency and the steady period of the river channel discharge. Artificially selected period were added to avoid fakes to certain time scale and in consideration of the difference of runoff in different months, another earlier runoff and precipitation are added in both flood season and non-flood season. It was concluded that the improved method can raise the prediction accuracy as in theXinfengRiver basin, thus being used to forecast the mid-long-term runoff.