基于R／S分析的黑河出山年径流量灰色预测

摘要

Runoff process has both fractal and gray characteristics .So R/S gray prediction model for predic-ting runoff sequence is proposed by means of combining gray system theory with R /S analysis.Based on annual runoff data from 1949 to 2011 in Yingluoxia Station, the out-mountainous runoff sequences of Heihe River is ana -lyzed,and use R /S method, to determine the Hurst exponent and the average cycle firstly ; then in one cycle, out-mountainous annual runoff is predicted by gray .It is showed that the average cycle( T ) of out-mountainous annual runoff sequence of Heihe River is about 20 ~25 years, and if T was 20, the result of gray prediction would best . The prediction accuracy of R /S gray prediction model is 91.04%, which is higher than the accuracy of the gray prediction model.R/S gray prediction model enlarges the applied scope of fractal theory and grey model , and pro-vides a new scientific method for predicting annual runoff sequence .%径流过程具有分形和灰色特征。基于此，将 R/S 分析与灰色系统理论相结合，提出了 R/S 灰色预测模型以预报黑河出山径流量。针对1949-2011年莺落峡水文站年径流量资料，首先进行 R/S 分析，确定径流量序列的Hurst 指数和平均循环周期 T；然后在一个周期内进行径流量灰色预测。结果表明：黑河出山径流量循环周期在20～25年之间，在进行 R/S 灰色预测时，取 T ＝20为宜； R/S 灰色预测结果的精度高于直接进行灰色预测。该方法拓宽了分形和灰色理论在径流过程研究的应用范围，为径流量的科学预测提供了一种新方法。