Trigila et al. (2026) Knowledge-guided methodology reducing computational effort in the heuristic calibration of a river model

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-25
• Authors: Mariano Trigila, Adriana Gaudiani, Alvaro Wong, Dolores Isabel Rexachs, Emilio Luque
• DOI: 10.1038/s41598-026-71630-y

Research Groups

• Faculty of Engineering and Agricultural Sciences, Pontifical Catholic University of Argentina, Buenos Aires, Argentina
• Science Institute, National University of General Sarmiento, Buenos Aires, Argentina
• Computer Architecture and Operating Systems Department, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain
• Computational Hydraulics Laboratory, National Institute of Water (INA), Buenos Aires, Argentina (provided the simulator)

Short Summary

This study introduces the Agile Adjustment in Successive Steps (A2S2) methodology, a knowledge-guided framework that significantly reduces the computational effort required for heuristic calibration of river hydrodynamic models by leveraging spatial and temporal continuity, while maintaining low calibration error.

Objective

• To develop and evaluate the Agile Adjustment in Successive Steps (A2S2) methodology, a hypothesis-driven calibration framework based on spatial and temporal continuity, aimed at reducing computational effort in the heuristic calibration of hydrodynamic river models.

Study Configuration

• Spatial Scale: River sections of the Paraná River.
• Temporal Scale: Multi-year simulations (e.g., 1995 for evaluation), with calibration performed over consecutive adjustment intervals.

Methodology and Data

• Models used: EZEIZA V hydrodynamic simulator (developed by Argentina’s National Water Institute (INA)).
• Data sources: Observed river flow data (implied for calibration and evaluation).

Main Results

• The A2S2 methodology achieved an approximately 86% reduction in estimated simulator executions compared to exhaustive evaluation.
• It maintained low calibration error levels, demonstrating a maximum reduction of approximately 57% in annual mean squared error (MSE) relative to the baseline configuration, observed in 1995.
• The methodology was compared with a two-stage Optimization via Simulation (OvS) approach, highlighting trade-offs between localized parameter reuse and broader parameter-space exploration.

Contributions

• Introduction of the Agile Adjustment in Successive Steps (A2S2) methodology for computationally efficient heuristic calibration of hydrodynamic river models.
• Demonstration of significant reduction in simulator executions by leveraging spatial and temporal continuity in parameter exploration.
• Provision of a scalable calibration framework for complex hydrodynamic river systems, addressing the high computational costs of exhaustive parameter exploration.

Funding

• Agencia Estatal de Investigación (AEI), Spain
• Fondo Europeo de Desarrollo Regional (FEDER), European Union (under grant PID2023-147995NB-I00 through the Universitat Autònoma de Barcelona (UAB))
• Fundación Escuelas Universitarias Gimbernat (EUG)

Citation

@article{Trigila2026Knowledgeguided,
  author = {Trigila, Mariano and Gaudiani, Adriana and Wong, Alvaro and Rexachs, Dolores Isabel and Luque, Emilio},
  title = {Knowledge-guided methodology reducing computational effort in the heuristic calibration of a river model},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-71630-y},
  url = {https://doi.org/10.1038/s41598-026-71630-y}
}