非一致水文序列频率计算方法应用效果比较分析

Comparison on application effects of frequency calculation methods in inconsistency hydrological series

受气候变化与人类活动等因素的影响,水文要素时间序列频率计算的一致性假设难以满足,需要寻求一套适合非一致性水文序列计算的方法。收集了寸滩水文站资料,利用回归分析、Mann-Kendall法进行趋势分析,采用滑动T与Lee-Heghinian法进行突变检验。通过对非一致序列修正后进行频率计算和直接基于非一致序列进行频率计算2种方法进行对比研究,探索了不同处理方法的应用效果。结果表明:2种方法均能提高频率计算成果的可靠性。对于样本容量比较大的序列,2种方法处理效果相当;当样本容量比较小时,修正方法更实用,但改善程度有限,资料代表性对计算成果影响也很显著。实际工作中应尽量扩大样本容量,提高资料代表性,对具有一定趋势变化的序列进行一致性修正。

Due to the influence of climate change and human activities,it is difficult to meet the consistency assumption in time series frequency calculation of hydrological elements.Therefore,it is need to figure out a calculation method for inconsistent hydrological series.In this paper,the hydrological data of Cuntan station,with a long station construction history,were collected for the hydrological frequency calculation and analysis of long series.Regression analysis and Mann-Kendall method were used for trend analysis,and sliding T test and Lee-Heghinian method were used for mutation test.In order to explore the application effects of different methods,two methods were adopted,one was frequency calculation after correction of inconsistent series and the other was frequency calculation based on inconsistent series.The results showed that both methods could improve the reliability of frequency calculation.For the sequence with large sample size,two methods have the same application effect.When samplesize was small,the correction method was more practical but the improvement degree was limited,and the representativeness of the data also had a significant impact on the calculation results.In practice,we should enlarge the sample capacity as much as possible,improve the representativeness of the data and modify the consistency of series with a trend.