基于Mann-Kendall法的水文序列趋势成分比重研究

Study on the Percentage of Trend Component in a Hydrological Time Series Based on Mann-Kendall Method

目前水文序列趋势研究主要集中于判断一个水文序列是否存在着显著趋势特征和该序列趋势是上升还是下降,但由于水文序列可分解为趋势、周期、随机等成分,论文提出了趋势成分在水文时间序列中的比重问题,并基于Mann-Kendall法,得到了趋势成分比重的计算公式。结果表明,趋势的显著程度与对偶数p和p0的差值有直接关系,差值越大,显著程度越明显;趋势成分比重取决于该差值及序列长度;显著趋势条件下趋势成分比重分析的结果,初步揭示了序列长度对趋势成分比重的影响,证明了用Mann-Kendall方法分析时间序列趋势特征时至少需要10个及以上的样本,且随着序列长度的增加,比重相应减小,因此当序列长度超过一定的范围时,Mann-Kendall方法是否适用于分析水文时间序列的显著趋势特征有待进一步探讨。应用于闽江流域竹岐站近72 a来径流系列的分析表明,序列呈下降趋势,但趋势不显著;趋势成分占整个时间序列近1/4。

The researches of trend in a hydrological series usually focus on detecting whether the trend characteristic is significant and judging whether the trend characteristic is increasing or decreasing.The paper proposes a concept about the weight of trend component in a hydrological time series which is usually decomposed as trend,period,random component and so on.The calculation formula of the weight of trend component is derived based on Mann-Kendall method.The results show that the difference of dual number p and p_0 affects directly the significant level of the trend,and the weight of trend component depends on the value of（p-p_0） and the sequence length of series.It also reveals the effects of the sequence length on the weight of trend component for a series with significant trending level.Firstly,it verifies the conclusion that at least 10 samples are necessary to conduct the trend analysis using Mann-Kendall method.Secondly,the weight decreases accordingly with the increase of the sequence length,which also indicates that whether the Mann-Kendall method is applicable for the analysis of trend characteristic should be further studied when the sequence length exceeds a certain range in a hydrological time series.The annual runoff series of Zhuqi Gauging Station in the Minjiang River Watershed from 1936 to 2007,about 72 years,has been taken as an example.The results show that the trend of the series is decreasing,but not significantly.And the weight of the trend component for the runoff series is 24.73%,which is nearly 1/4.