中文摘要：

气候变化及人类活动的影响很大程度上改变了洪水或径流的孕育环境,使得现行水文频率分析的一致性假设难以满足。为此,发展变异水文序列的一致性修正技术尤为重要。本文基于假定＂发生趋势性变异的实测序列存在着某种理想化的平稳性（一致性）状态,且这种平稳状态所具有的振动中心（即均值）是序列某分割点前后两实测样本系列均值的线性组合＂,通过综合分割点前后两段系列,开展趋势性变异系列的一致性修正,并应用该方法进行了实例研究。

英文摘要：

Effect of climate change and human activities has caused a change in the breeding environments of flood and runoff, resulting in a loss of consistency in flood volume series, and because of this loss of flow series consistency the conventional frequency analysis methods are not valid. So, it is a must to develop a modified technique for non-stationary observations. In this paper, the authors propose an approach to modification of the non-stationary observations, based on the assumption that an ideal state for the trend variation of an observation series and its fluctuation center （mean） of steady state is a linear combination of two mean values of the two sub-series which are the parts of this observation series before and after its split point respectively. An example is presented to illustrate this approach.