Comprehensive evaluation of multiple methods for assessing water resources variability of a lake–river system under the changing environment

Abstract

Abstract The inherent regularities of hydrological processes will not be stationary under the changing environment, making the evaluation of hydrological characteristics and trend forecasting face great uncertainty. Based on the observed streamflow series in the Poyang Lake catchment during the period 1960–2013, this study investigated the evolution characteristics and future trend of water resources. The extracted trend component by wavelet decomposition and the reliability of trend forecasting by rescaled range analysis (R/S analysis) were verified and assessed by a non-parametric statistical test. Results indicate a change point at 1995 when the trend of annual streamflows across the catchment changed. Periodic components indicated that the lake basin seems to enter into an increasing process under inter-decadal scale, while decreasing under inter-annual scale in the near future. R/S analysis revealed that annual streamflow series of most stations in Poyang Lake show ‘weaker’ or ‘very weak’ persistence. For the whole study period, Hurst exponents of most stations are less than 0.5 (0.224–0.339), therefore, opposite from the increasing trend of historical series, decreasing trend is forecasted for the basin. However, due to the effect of intensified human activities in the Fuhe sub-basin, the reliability of R/S analysis on trend forecasting is disturbed at Lijiadu station.