Related Experiment Video
Updated: Sep 11, 2025

Watershed Planning within a Quantitative Scenario Analysis Framework
Published on: July 24, 2016
Exploring the influence of hydrological indicators on flow regimes through a data-driven modeling approach in the
Qing Wei1, Peipei Chen1, Zichen Jia1
1State Key Laboratory of Water Pollution Control and Green Resource Recycling, College of Environmental Science and Engineering, Tongji University, Shanghai, 200092, China; Ministry of Education Key Laboratory of Yangtze River Water Environment, Tongji University, Shanghai, 200092, China; Shanghai Institute of Pollution Control and Ecological Security, Shanghai, 200092, China.
Abstract:
Understanding the impact of hydrological indicators on flow regimes is essential for sustainable water resource management. This study presents a data-driven framework integrating eXtreme Gradient Boosting (XGBoost) with SHapley Additive exPlanations (SHAP) to quantify the influence and interactions of key hydrological indicators on annual runoff in the Huai River Basin during 1970-2020. Based on 5-fold cross-validation results, XGBoost achieved superior predictive performance with R2 values of 0.984 in the midstream and 0.967 in the downstream. SHAP analysis identified the 90-day and 30-day maximum flows, and monthly flows in July and September, as dominant predictors. Interaction analysis revealed clear thresholds and synergistic patterns. In the midstream, runoff exceeded 501.22 × 108 m3 when both 30-day and 90-day maximum flows surpassed 3531 m3/s and 4891 m3/s, respectively, but dropped below 173.45 m3/s when both were low. In the downstream, 90-day maximum flow thresholds at 2986 m3/s and 5137 m3/s corresponded to runoff outputs of 357.37 × 108 m3 and 391.30 × 108 m3. Strong interactions between July and August monthly flows revealed joint thresholds associated with high runoff clustering. These results highlight the nonlinear and region-specific responses of flow regimes to hydrological drivers. Based on these findings, targeted ecohydrological management strategies are recommended, emphasizing real-time monitoring and regulation during key high-flow periods. This study offers a transparent data-driven approach for decoding hydrological influences and provides practical insights for adaptive basin-scale flow management.
Related Concept Videos
Typical Model Studies
Design Example: Creating a Hydraulic Model of a Dam Spillway
Rapidly Varying Flow
Gradually Varying Flow
Modeling and Similitude
Design Example: Analyzing Capacity Contours for Flood Risk Assessment

