Impacts of GCM Credibility on Hydropower Production Robustness under Climate Change: CMIP5 Vs CMIP6

Yuxue Guo,Yue-Ping Xu,Xinting Yu,Jingkai Xie,Hao Chen,Yuan Si
DOI: https://doi.org/10.1016/j.jhydrol.2023.129233
IF: 6.4
2023-01-01
Journal of Hydrology
Abstract:Climate change has emerged as a vital issue for hydropower management in the future. General circulation model (GCM)-informed climate is commonly applied to assess the impact of climate change on hydropower. However, the effect of GCM credibility on hydropower robustness has been rarely explored. One objective of this study is to seek valuable and robust policies, and the other is to evaluate the effects of GCM credibility on hy-dropower robustness based on the pre-optimized operating policies. A multi-objective robust optimization approach is first proposed to design operating policies by coupling stochastic simulation, heuristic parameterized policy search, and robust optimization. The GCM credibility is then evaluated based on multi-criteria ranking scores, and the hydropower robustness is illustrated by stepwise culling the lowest credible GCMs. The climate -induced impacts on hydropower robustness rely on an ensemble of 18 GCMs under three representative con-centration pathways in Coupled Model Intercomparison Project 5 (CMIP5) and 17 GCMs under three shared socioeconomic pathways in CMIP6. The Longyangxia and Liujiaxia reservoirs in Upper Yellow River Basin are selected as a case study. The results show that (1) the proposed stochastic approach outperforms the deter-ministic approach regarding hydropower generation and firm output where the mean annual hydropower generation of stochastic operation is improved by -0.50-19.53% compared to deterministic operation, indicating that the operating rules derived using historical stochastic inflows can adapt flexibly to future inflow changes than that using deterministic inflow. (2) CMIP6 maintains a more robust performance than CMIP5 with higher satisficing measures and lower regret measures. (3) Hydropower robustness tends to be higher when considering credibility and will be lower with time due to increasing uncertainty. Ignorance of GCM credibility may lead to biased climate change adaptation decisions.
What problem does this paper attempt to address?