Detection of Trend in Hydrologic Variables Using Non-Parametric Test: A Study on Surma River in Northeastern Bangladesh

Abstract

An initiative has been taken to investigate the trends in discharge and water level (WL) of the Surma River in northeastern Bangladesh. The daily time series data of discharge and WL from two stations named Kanairghat and Sylhet with a period of 42 years (1973 -2014) and 35 years (1980 -2014) respectively have been analyzed. Non parametric Mann-Kendall Test has been applied to detect the trend and Sen's slope estimator is used to measure the slope of the trends. In Kanairghat station, annual mean WL has significant trend (P: 0.03); while, annual mean discharge, mean monsoon discharge, annual maximum discharge, mean monsoon WL, and annual maximum WL shows insignificant trend (P: 0.24, 0.46, 0.14, 0.05, and 0.12). In Sylhet station, annual mean discharge, annual maximum discharge, and annual mean WL have significant trend (P: 0.03, 0.004, and 0.02). On the other hand, mean monsoon discharge, mean monsoon WL, and annual maximum WL in Sylhet station has insignificant trend (P: 0.46, 0.13, and 0.21). According to Sen's slope statistics, all of the detected trends, except annual maximum WL at Sylhet station, are downward. This study recommends a comprehensive water management scheme should be taken to ensure sustainable use of the river water.

Impact of China’s Urbanization on Water Use and Energy Consumption: An Econometric Method and Spatiotemporal Analysis

“…To determine the range of the possible trends in the WUI and ECI indicators and their statistical significance during the 2005 to 2015 period, the nonparametric Sen's slope method and Mann-Kendall test were used. Sen's slope was used to verify the range of the possible trends in the statistical sequence [47,48], and the Mann-Kendall test was selected to substantiate the statistical significance [49,50].…”

Streamflow characteristics of Sangu-Matamuhuri watershed in the southeastern part of Bangladesh

“…Serial correlation can be measured at different lag period. In this study all of the rainfall and runoff time series data were tested for the presence of lag-1 serial correlation [ 60 ]. In order to estimate the lag-1 serial correlation, following Eq.…”

“…Non-parametric Mann-Kendall test was carried out to find out the trend in time series data. The test compares the relative magnitudes of sample data rather than the data values themselves [ 59 , 60 ]. One benefit of this test is that the data need not conform to any particular distribution.…”

“…This test evaluates whether outcome values tend to increase or decrease over time. It is often used to quantify the significance of trends in hydro meteorological time series [ 29 , [60] , [61] , [62] ]. The hypotheses of Mann and Kendall's trend test are as follows:…”

“…To estimate the slope of trend in a sample of n pairs of data the Sen's estimator was used [ 60 , 63 ]. In order to obtain slope estimate Qmed, the slope (Ti) of all data was first computed using the following Eq.…”