Skip to main navigation Skip to search Skip to main content

Optimal reservoir flood-control operation under successive rainfall events during the Meiyu season

  • Yuchang Mo*
  • , Yunfei Li
  • , Shuaisen Ma
  • , Qingan Qiu
  • , Rui Peng
  • *Corresponding author for this work
  • Zhejiang University of Water Resources and Electric Power
  • Zhejiang-Kyrgyzstan Joint Laboratory on Artificial Intelligence and Clean Energy
  • Beijing Institute of Technology
  • Beijing University of Technology

Research output: Contribution to journal › Article › peer-review

Abstract

Reservoir flood-control operation during the Meiyu season is challenged by successive rainfall events, compound flood-wave superposition, and the conflicting objectives of flood safety and operational economy. Existing studies often represent rainfall uncertainty through independent events or aggregated inflow scenarios, which may overlook the temporal clustering and self-exciting characteristics of successive Meiyu rainfall events. This work develops a dynamic staged spillway-operation framework for reservoir flood-control scheduling under clustered rainfall conditions. A marked Hawkes process is employed to characterize the stochastic occurrence and temporal dependence of successive rainfall events, while a duration-corrected Gamma response kernel is introduced to describe the runoff generation and superposition process. Based on these components, a threshold-driven scheduling model is established, in which spillway operation levels are adaptively adjusted according to the real-time conditional non-failure probability. The optimal threshold structure is obtained by minimizing the expected operating cost through a nested Monte Carlo simulation framework coupled with particle swarm optimization. A case study of the Xin’anjiang Reservoir demonstrates that the proposed framework provides a better balance between flood-control reliability and operational cost compared with empirical, conservative, and aggressive scheduling strategies. Further analyses reveal that rainfall clustering characteristics, runoff-response parameters, and initial reservoir storage conditions significantly influence the optimal operation strategy. The proposed framework offers a risk-responsive and cost-aware approach for reservoir flood-control operation under successive rainfall events during the Meiyu season.

Original languageEnglish
Article number136328
JournalJournal of Hydrology
Volume679
DOIs
Publication statusPublished - Oct 2026
Externally publishedYes

Keywords

  • Marked Hawkes process
  • Meiyu season
  • Reservoir flood-control operation
  • Staged spillway operation
  • Successive rainfallevents

Fingerprint

Dive into the research topics of 'Optimal reservoir flood-control operation under successive rainfall events during the Meiyu season'. Together they form a unique fingerprint.

Cite this