Cortés-Torres et al. (2026) Scalability and Computational Performance of an Ecohydrological Model Using Machine Learning-Based Prediction

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2026
• Date: 2026-02-11
• Authors: Nicolás Cortés-Torres, Sergio Salazar-Galán, Félix Francés
• DOI: 10.3390/w18040466

Research Groups

Not explicitly stated in the provided text.

Short Summary

This study developed a general methodological framework to systematically quantify the computational scalability of distributed hydrological models, using TETIS v9.1 as a case study, and found that spatial resolution and output-gauge density are the strongest influences on runtime, while also creating a highly accurate predictive tool for performance.

Objective

• To develop a general methodological framework for systematically and reproducibly quantifying the computational performance and scalability of distributed hydrological models.
• To evaluate this framework using the TETIS v9.1 ecohydrological model.
• To train a predictive model for computational performance based on user-defined configuration variables.

Study Configuration

• Spatial Scale: Varied spatial resolutions and catchment sizes were explored to assess their impact on computational performance.
• Temporal Scale: Varied temporal resolutions were explored to assess their impact on computational performance.

Methodology and Data

• Models used: TETIS v9.1 ecohydrological model, Random Forest regression model.
• Data sources: Systematically recorded runtimes from TETIS v9.1 simulations under varying spatial and temporal resolutions, input/output gauge densities, and hardware configurations.

Main Results

• Spatial resolution and output-gauge density exert the strongest influence on runtime.
• Temporal resolution shows nonlinear effects on runtime, dependent on catchment size.
• The developed predictive tool (Random Forest model) achieved high accuracy for large hydrological simulations.
• Increased uncertainty in predictions was limited to extremely short runtimes on high-speed processors.

Contributions

• Introduction of a transferable methodological framework for assessing computational scalability in hydrological modeling.
• First reproducible characterization of TETIS v9.1 computational scalability.
• Development of a robust predictive tool for anticipating runtimes and supporting efficient experimental design and operational modeling.

Funding

Not provided in the paper text.

Citation

@article{CortésTorres2026Scalability,
  author = {Cortés-Torres, Nicolás and Salazar-Galán, Sergio and Francés, Félix},
  title = {Scalability and Computational Performance of an Ecohydrological Model Using Machine Learning-Based Prediction},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18040466},
  url = {https://doi.org/10.3390/w18040466}
}