Main content area

Performance evaluation of hydrological models using ensemble of General Circulation Models in the northeastern China

Faiz, Muhammad Abrar, Liu, Dong, Fu, Qiang, Li, Mo, Baig, Faisal, Tahir, Adnan Ahmad, Khan, Muhammad Imran, Li, Tianxiao, Cui, Song
Journal of hydrology 2018 v.565 pp. 599-613
General Circulation Models, hydrologic models, observational studies, planning, rivers, runoff, soil water, watersheds, China
The Songhua River Basin (SRB) plays a vital role in supplying water resources to northeast China. This river has a substantial effect on agricultural grain producing region. Therefore, the current study was carried out to evaluate the performances of multiple hydrological models driven by bias corrected precipitation estimates from a group of General Circulation Models (GCMs). The variations in discharge obtained from multiple hydrological models were also analyzed. For this purpose, bias correction methods for GCMs and five hydrological models, lumped rainfall runoff (Nedbor Afstromnings Model (NAM), GRJ4), semi-distributed (the Water Evaluation and Planning tool (WEAP), HBV), and the distributed (Soil Water Assessment Tool (SWAT)) were used in the SRB. Before applying the bias correction method, raw GCMs were also plotted against observational data. The results revealed that none of GCM performed well against observational data. The results also showed that quantile mapping method corrects the biases of GCMs better than daily and monthly scale factors. Distributed and semi-distributed hydrological models performed better than lumped rainfall runoff models in the selected study area of SRB. By forcing hydrological models with bias corrected precipitation data, the differences in the simulated discharge by SWAT, HBV, NAM, GR4J and WEAP were −3%, −2%, −4%, −7 and +3, respectively, when compared with observed discharge. The results also show that the overall performance of SWAT, HBV, NAM, and WEAP was better than GRJ4. The results also suggest that the application of multiple hydrological models is necessary to make improvements in the planning and management of agricultural and water resources in the SRB region.