Main content area

Assessing the effects of reservoirs on extreme flows using nonstationary flood frequency models with the modified reservoir index as a covariate

Su, Chengjia, Chen, Xiaohong
Advances in water resources 2019 v.124 pp. 29-40
basins, hydrologic data, magnetic resonance imaging, models, rivers
The purpose of this study is to ascertain whether the changes in extreme hydrologic time series can be attributed to the reservoir operations by using nonstationary flood frequency analyses. On the basis of the reservoir index (RI), a modified reservoir index (MRI) was put forward and used as a covariate to represent the effects of reservoirs on extreme flows. Using the generalized additive model for location, scale, and shape (GAMLSS), nonstationary flood frequency models were developed with the MRI covariate. The models were applied to fit the annual maximum daily flow (AMDF) from 1956 to 2013 at Boluo, the control gauge station of the Dongjiang River (DJR) basin, which is strongly influenced by reservoirs. In addition, to avoid the potential impacts of climatic variability on extreme flows, a precipitation covariate that characterizes the key climatological properties is taken into account in the nonstationary flood frequency models. The results show that the MRI is superior to the RI to reflect the effects of reservoirs on the extreme flows which is more effective and practical. And it is a significant explanatory variable for the extreme flows in the nonstationary modeling. A significant effect of the reservoirs on the extreme flows is detected with the MRI and this effect is greater than the precipitation effect. In addition, with the increasing demand for water resources and the transformation of reservoir functions, the effects of the reservoirs on extreme flows in the DJR basin will further increase.