Answers are grounded in this paper and cite their sources.
Achieving consistent solutions in hydrologic model calibration requires incorporating diverse residual and signature‐based performance metrics. However, optimizing many objectives simultaneously reduces the efficiency of multi‐objective optimization. This study introduces Non‐negative Orthonormal Correlation‐based Principal Component Analysis (NOC‐PCA), a novel method for transforming performance metrics into optimizable principal components while prioritizing them based on their informational content, measured by variance. Building on NOC‐PCA, a PCA‐based calibration framework was developed to optimize three objective functions, enabling efficient dimensionality reduction while preserving critical information from the original metrics. Tested across four hydrologically distinct case studies, the PCA‐based calibration framework consistently outperformed conventional bi‐objective and binary objective approaches. It maintained more metrics within the acceptable bias range (≤20%) during calibration and validation, particularly for high‐flow, low‐flow, and water balance processes. The number of metrics that meet this threshold during calibration strongly predicted validation performance, with a strong correlation coefficient of +0.8, underscoring the framework's robustness. These findings demonstrate the value of integrating a diverse set of metrics representing dominant hydrological processes and clustering and prioritizing them using NOC‐PCA. The proposed calibration approach is scalable and adaptable to diverse watershed characteristics and hydrological conditions.
No takes yet. Share an insight, caveat, or question.
Asadzadeh et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: