Trend analysis of monthly and annual runoff based on a segmented M-K method in Fuhe River basin, China

Abstract

Abstract It is very important for water resource management to analyse the change rule of runoff. In this paper, a segmented Mann-Kendall trend analysis method, which is called segmented M-K method in the following, is proposed to analyse the periods of significant trend within the time series. This method fills in the gap that the changes of internal trend for long sequence are ignored in M-K trend test, by finding out the significant change period within the long time series. It may be a new method for the analyse of runoff changes. Fuhe River, which is one of the main rivers of Poyang Lake in Yangtze River basin, is taken for example. Based on the runoff data of the Lijiadu section of the Fu River Basin from 1957 to 2015, the interannual variation of runoff over Lijiadu section in Fuhe River basin was quantitatively analysed by using a combined trend analysis method containing linear regression, sliding average, cumulative filter, Mann-Kendall trend test, Mann-Kendall mutation test and segmented M-K method. The results show a light interannual variety and a trend of insignificant increase in annual runoff for whole time series in Fuhe River Basin. However, an insignificant decrease trend, which might course a shortage of water for development and utilization, is found in the period from 1997 to 2015 while a significant increase trend is found between 1963 and 1985. The annual runoff alternates frequently between abundance and drought. However, its interannual variation of monthly runoff is large, and 3 months include January, December and August show a significant increase. From the result of segmented M-K method, both significant increase trend period and significant decrease trend period could be found in flood season months, while only significant increase trend period is found in other months. In addition, the increasing period is close to the middle part of the total time series in dry season months, to the front part in flood season months, and to the rear part in months after flood season.