Modeling Intermittent Water Supply in SWMM: A Critical Review With Reproducible Recommendations and a Python Package

• Published in: Water resources research Vol. 61; no. 8
• Main Authors: Abdelazeem, Omar, Meyer, David D. J.
• Format: Journal Article
• Language: English
• Published: Washington John Wiley & Sons, Inc 01.08.2025
• Subjects: hydraulic modeling; Hydraulic models; Hydraulics; intermittent supply; Modelling; Networks; Numerical stability; Reproducibility; Stability; Storage; User behavior; Water consumption; water distribution; Water shortages; Water supply; Water supply systems; Water users
• ISSN: 0043-1397; 1944-7973
• DOI: 10.1029/2024WR039551

Distinctly, Intermittent Water Supply networks cycle between filling, pressurized supply and draining, leading users to withdraw water and store it for later consumption. While intermittent networks serve one in five piped water users, their characteristic hydraulic features cannot be readily represented in available, open‐source hydraulic modeling software. Several hydraulic modeling methods have been proposed in the literature, but these methods disagree in their construction and assumptions, and most are not reproducible, hindering the exploration of techniques to improve the quality and equality of service in intermittent networks. To improve the reproducibility, consistency, and numerical stability of hydraulic models of intermittent supply, we synthesize the best modeling practices in the literature into a recommended, reproducible method: SWMM for Intermittent Networks (SWMMIN). We outline and demonstrate how SWMMIN models network pipes and user behavior: withdrawing water subject to available pressure in the network, storing water, and consuming from storage for their various activities. For experienced IWS modelers, we provide quantitative evidence of numerical stability and mass conservation within SWMMIN (from >1,000 simulations of 3 network models); we recommend spatial and temporal discretizations that result in solution speeds between 40 and 200 m/s. To facilitate the adoption of our recommended modeling procedures, we share a Python package (GOSWMMIN) that automates the implementation of SWMMIN. Lastly, we propose a model reporting template to bolster reproducibility and call on fellow modelers to use it; accessibly and reproducibly described models of intermittent supply have the potential to accelerate research and transform practice. Plain Language Summary Computer simulations of water pipe networks allow researchers and utilities to test thousands of ways of improving performance. But traditional simulation tools cannot model the behaviour of intermittent systems, which make up one fifth of the world’s water distribution systems. Due to water scarcity, these systems supply water for only a few hours a day or less, forcing households to store water and causing pipes to drain and fill frequently. The few published methods of simulating intermittent supply disagree about how to represent draining, filling and storage; most of these methods are not described in enough detail for others to implement. This paper synthesizes the best of these methods and describes them in enough detail for other modelers to implement. We also provide new evidence and advice about model resolution; we found an ideal ratio of spatial and temporal resolutions that minimizes model error. To make it easier to use our recommended best practices and resolutions, we made an open‐source Python package that automatically implements them. Our work makes it easier for researchers and utilities to simulate ways of improving intermittent water supply systems for the one billion people who rely on them. Key Points Most models of intermittent water supply are not consistent or reproducible, omitting essential features and thwarting research We synthesized best practices for modeling intermittent supply in SWMM: pipe filling and draining, user withdrawal, storage, and consumption Stability and mass conservation improve in models of intermittent supply when discretization achieves solution speeds of 40–200 m/s