鉴于仅使用单一的中长期径流预报方法很难有效地拟合和预测整个径流过程,将逐步回归法和自回归方法相结合,对尼洋河流域1956~1999年逐年年径流序列进行非平稳序列逐步回归趋势分析、周期分析及平稳序列自回归分析,并与AR(1)模型的计算结果进行比较研究。结果表明,组合回归方法的预测效果较好。 In view of the fact that it is difficult to fit and predict the entire runoff effectively using only a single long-term runoff forecast method, the stepwise regression method and the autoregressive method are combined to analyze the non-stationary series of annual runoff sequences from 1956 to 1999 in the Niyang River Basin Regression trend analysis, periodic analysis and stationary series autoregressive analysis were performed and compared with the results of AR (1) model. The results show that the combined regression method has better prediction effect.