多因子逐步回归周期分析在中长期水文预报中的应用

Application of multi-factor stepwise regression cycle analysis in medium and long-term hydrological forecast

利用多因子逐步回归周期分析法建立长江宜昌站年平均流量的中长期水文预报模型．结果表明：所选优势因子主要是一组反映冷暖空气活动的指数，冷暖空气的活动和强度对宜昌的年平均流量影响很大；各优势因子的优势隐含周期4～9个不等，且普遍都具有25a和22a的变化周期；多因子逐步回归周期分析法对长江宜昌站年平均流量的拟合和试预报合格率都达到了100％，具有较高的可信度．

A medium and long-term hydrological forecast model of the mean annual discharge at Yichang station on the Yangtze River was established using the multi-factor stepwise regression cycle analysis. The result shows that the dominant factors are mainly a group of indices that can reflect the activity of the cold and warm air, and the activity and the intensity of the cold and warm air have a great impact on the mean annual discharge at Yichang station. The various dominant factors have 4-9 dominant hidden cycles and generally have cycles of 25 years and 22 years. The simulation and forecast accuracy of the mean annual discharge at Yichang station on the Yangtze River with the multi-factor stepwise regression cycle analysis reaches 100% .The present method is highly reliable.