Menapace et al. (2025) Sensors prioritisation for hydrological forecasting based on interpretable machine learning

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-09-08
• Authors: Andrea Menapace, André Ferreira Rodrigues, Daniele Dalla Torre, Michele Larcher, Manuel Herrera, Bruno Brentan
• DOI: 10.1016/j.jhydrol.2025.134015

Research Groups

• Faculty of Agricultural, Environmental and Food Sciences, Free University of Bozen-Bolzano, Italy
• Institute for Renewable Energy, Eurac Research, Italy
• Universidade Federal de Minas Gerais - UFMG, School of Engineering, Brazil
• Newcastle University, School of Engineering, United Kingdom

Short Summary

This study proposes an interpretable machine learning framework to prioritise hydrological sensors, aiming to enhance short-term predictions. The research demonstrates that identifying and maintaining critical sensors significantly improves forecasting accuracy and reliability, offering a data-driven approach to optimise monitoring system maintenance.

Objective

• To enhance short-term hydrological predictions by prioritising sensors based on interpretable machine learning.

Study Configuration

• Spatial Scale: South Tyrol watershed (northern Italy)
• Temporal Scale: Short-term hydrological predictions, with horizons ranging from 1 to 24 hours.

Methodology and Data

• Models used: Machine learning-based hydrological models, interpretable machine learning techniques. The framework involves tuning models for different horizons and applying leave-one-out cross-validation to simulate sensor failures.
• Data sources: Data from streamflow gauges and weather stations.

Main Results

• Specific sensors significantly impact hydrological forecasting accuracy.
• Sensor prioritisation, derived from the proposed interpretable machine learning framework, improves the reliability of hydrological predictions.
• The findings underscore the importance of maintaining critical sensors and offer a data-driven methodology for optimising resource allocation in monitoring system maintenance.

Contributions

• Proposes a novel evaluation framework for sensor prioritisation in hydrological forecasting using interpretable machine learning and leave-one-out cross-validation.
• Provides a data-driven methodology to identify critical sensors, thereby optimising resource allocation for monitoring system maintenance.
• Enhances the robustness of hydrological forecasting and risk mitigation strategies by improving prediction reliability in the context of IoT networks and extensive data management.

Funding

• No funding information was provided in the article text.

Citation

@article{Menapace2025Sensors,
  author = {Menapace, Andrea and Rodrigues, André Ferreira and Torre, Daniele Dalla and Larcher, Michele and Herrera, Manuel and Brentan, Bruno},
  title = {Sensors prioritisation for hydrological forecasting based on interpretable machine learning},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134015},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134015}
}