Modeling the Impact of Climate Change on Low Flows in Xiangjiang River Basin with Bayesian Averaging Method

Abstract

This paper investigates the impact of climate change on low flows in Xiangjiang River Basin in central China. Projections from four global climate models (GCMs) under representative concentration pathway (RCP) 4.5 and RCP8.5 are used to drive the hydrological models. An ensemble prediction in the future period (2021–2050) from three competing hydrological models is generated using the Bayesian model averaging (BMA) method. Though hydrological models do well in simulating daily discharges, all underestimate the observed low flows to some extent. Such underestimation of low flows could be compensated by application of the BMA method. Uncertainty from GCMs is a predominating source for monthly mean discharges. An increase in intensity of 7Q10, 7Q20, and 7Q30 is found under both RCP4.5 and RCP8.5 for most cases. The increase of 7Q10, 7Q20, and 7Q30 could be more related to minimum monthly precipitation rather than the amount of monthly mean precipitation. The magnitude of the increase is smaller under RCP8.5 than RCP4.5, which could be explained by the higher temperature increase under RCP8.5.