Sakib et al. (2026) Machine learning in urban stormwater management: A review

Identification

• Journal: Environmental Modelling & Software
• Year: 2026
• Date: 2026-08-11
• Authors: Md Sakib, Omar Hegazy, Abhiram Siva Prasad Pamula, Walter McDonald
• DOI: 10.1016/j.envsoft.2026.107138

Research Groups

• Department of Civil, Construction & Environmental Engineering, Marquette University, Milwaukee, WI, USA

Short Summary

This study presents a structured synthesis of machine learning (ML) approaches for urban stormwater modeling, focusing on their ability to address limitations in traditional models. The review highlights the potential of ML in predicting water quality, flow and discharge, infrastructure inspection, and real-time control.

Objective

• Investigate the role of machine learning in addressing key challenges in urban stormwater management

Study Configuration

• Spatial Scale: Urban catchments
• Temporal Scale: Real-time to long-term system management

Methodology and Data

• Models used: Supervised, unsupervised, reinforcement learning, random forest, artificial neural networks, deep belief networks, generalized linear models, partial least-squares regression, hierarchical clustering analysis, gradient boosting, extreme gradient boosting
• Data sources: In-situ sensors, remote sensing, distributed monitoring networks, supervisory control and data acquisition (SCADA) systems

Main Results

• Machine learning techniques can capture nonlinear relationships between land-use variables, pollutant dynamics, and hydrologic drivers that conventional linear regressions are unable to capture.
• ML models can produce predictions quickly and scale effectively across large, heterogeneous stormwater networks after training.
• Ensemble and deep-learning models continue to function well in scenarios with sensor noise and incomplete stormwater monitoring data.

Contributions

• This review provides a comprehensive synthesis of machine learning approaches for urban stormwater management, highlighting their potential to address limitations in traditional models.
• The study identifies the need for further research on hybrid physics–ML modeling, benchmark datasets, uncertainty-aware AI, digital twins, and transferable modeling frameworks.

Funding

• Not specified

Citation

@article{Sakib2026Machine,
  author = {Sakib, Md and Hegazy, Omar and Pamula, Abhiram Siva Prasad and McDonald, Walter},
  title = {Machine learning in urban stormwater management: A review},
  journal = {Environmental Modelling & Software},
  year = {2026},
  doi = {10.1016/j.envsoft.2026.107138},
  url = {https://doi.org/10.1016/j.envsoft.2026.107138}
}