A two-stage hierarchical runoff bias correction framework for observation-constrained river flow reconstruction

Area of Science:

• Hydrology and Water Resources
• Climate Modeling
• Geospatial Analysis

Background:

• Accurate long-term river flow reconstruction is vital for water resource management.
• Existing modeling simulations often contain biases, and traditional correction methods neglect upstream-downstream connectivity and mass balance.

Purpose of the Study:

• To introduce and validate the Two-Stage Hierarchical Bias Correction (TSH-BC) framework for correcting ERA5 runoff.
• To improve the accuracy of river discharge modeling across all stream orders, particularly in large river basins.

Main Methods:

• Developed the TSH-BC framework, which applies a cascading water-balance logic to correct ERA5 runoff at the source.
• Established headwater baselines and corrected incremental biases of lateral inflows between gauging stations.
• Applied the framework to 40,000 reaches in the Yangtze River Basin for daily discharge reconstruction (1980-2024).

Main Results:

• The TSH-BC framework significantly improved model performance, elevating the median Kling-Gupta Efficiency from 0.45 to 0.74.
• A monthly correction strategy specifically enhanced low-flow representation, achieving a median logarithmic Nash-Sutcliffe Efficiency of 0.72.
• Reconstructed daily discharge data from 1980 to 2024, demonstrating full-period in-sample performance.

Conclusions:

• The TSH-BC framework provides a topology-aware, observation-constrained baseline for historical water resources assessment.
• The adjusted runoff effectively absorbs various uncertainties, leading to more reliable hydrological modeling.
• This approach offers a significant advancement in correcting biases in global reanalysis runoff data for improved water management.