Streamflow Prediction in Highly Regulated, Transboundary Watersheds Using Multi-Basin Modeling and Remote Sensing...

Area of Science:

• Hydrology
• Remote Sensing
• Water Resource Management

Background:

• Reservoir operations significantly impact downstream hydrology.
• Accurate monitoring of reservoir storage and releases is crucial for water management and hydrological modeling.
• Existing global reservoir databases have limitations in representing local operational details.

Purpose of the Study:

• To assess the efficacy of remote sensing for inferring reservoir operations across various scales.
• To develop and validate a framework for improving reservoir representation in global databases and hydrological models.
• To enhance streamflow prediction in ungauged cascade reservoir systems.

Main Methods:

• Evaluated remote sensing imagery for estimating daily storage volumes of 21 reservoirs in the Greater Mekong region.
• Developed and applied an Integrated Reservoir Operation Scheme within a multi-basin hydrological model.
• Assessed prediction accuracy using correlation coefficients (CC), normalized root mean squared error (NRMSE), and Kling-Gupta efficiency (KGE).

Main Results:

• Remote sensing accurately estimated storage for within-year and over-year reservoirs (CC ≥ 0.9, NRMSE ≤ 31%) but not for run-of-river types.
• The integrated operational scheme significantly improved streamflow prediction in cascade systems.
• Competent daily regulated streamflow predictions were achieved across reservoir types (median CC=0.65, NRMSE=8%, KGE=0.55) and downstream stations (median CC=0.94, NRMSE=8%, KGE=0.81).

Conclusions:

• Remote sensing is a valuable tool for monitoring certain reservoir types but has limitations for others.
• The proposed framework enhances reservoir data for hydrological models, improving human impact assessments.
• The integrated operational scheme provides a robust solution for streamflow prediction in data-scarce river basins, aiding adaptation to extreme events.