Every summer, the public faces the recurring dilemma of beach closures triggered by Escherichia coli (E. coli) contamination, events that disrupt recreation, threaten public health, and create economic setbacks for coastal communities. These closures traditionally come as reactive measures, often after individuals are potentially exposed to harmful bacteria. However, a groundbreaking advancement from the FAMU-FSU College of Engineering signals a shift toward anticipatory water quality management. This innovative framework harnesses artificial intelligence to predict E. coli contamination risk with unprecedented lead time, allowing intervention before health threats materialize.
At the helm of this research, Assistant Professor Nasrin Alamdari has pioneered an AI-driven predictive modeling framework that integrates vast streams of environmental and hydrometeorological data. This system analyzes upstream hydrology, rainfall intensity, turbidity levels, water temperature, and streamflow dynamics to forecast E. coli concentrations in recreational waters. The model’s ability to predict unsafe conditions with approximately 85% accuracy up to 24 hours in advance represents a significant leap beyond current standards, where laboratory water quality results are delivered only after a day or more.
The traditional reliance on manual sampling followed by laboratory assays introduces inherent delays in water safety notifications. Such delays expose swimmers unwittingly to contaminated water, increasing the risk of gastrointestinal illnesses and other infections. In contrast, Alamdari’s approach offers near real-time risk assessment, allowing water managers to issue early warnings, enact preventive closures, and protect both public health and economic interests more effectively. This shift from a reactive to predictive paradigm carries profound implications for how municipalities manage recreational waters.
Central to the success of this framework is the comprehensive integration of multiple environmental variables that influence E. coli fluctuations. Rainfall events, particularly in urbanized watersheds, can induce rapid spikes in bacterial concentrations by mobilizing contaminated runoff. Streamflow and turbidity serve as proxies for this increased risk, while water temperature affects bacterial survival and proliferation. By processing these complex, interrelated factors through machine learning algorithms, the model discerns patterns and signals indicative of imminent contamination events before traditional methods can detect them.
An illustrative case validating the model’s utility occurred during a 2023 sewage spill at the Big Creek Water Reclamation Facility. This failure led to a swift surge in pollutants downstream. The predictive model, equipped with current and historical environmental datasets, was capable of estimating the spike in contamination well before laboratory analyses confirmed it. This proactive detection capability not only aids rapid public notification but also facilitates targeted monitoring and resource allocation during sudden contamination crises.
The repercussions of delayed detection extend beyond health concerns. Unexpected closures reverberate through local economies heavily reliant on tourism and water-based recreation. Hotel bookings plummet, outdoor outfitters suffer losses, and municipalities face increased emergency response costs. Alamdari emphasizes that predictive monitoring equips communities to mitigate these economic damages by informing stakeholders in advance, allowing businesses and agencies to adapt and respond preemptively rather than reactively.
Importantly, the new model accounts for evolving landscape and climate factors exacerbating contamination risk. Urbanization between 2007 and 2023 in the study area raised impervious surface coverage by 4%, altering runoff characteristics and intensifying the transport of pollutants into waterways. These changes complicate traditional water quality prediction, but the model’s consideration of land use transformation and watershed wetness enhances the precision of risk assessment across varying environmental conditions, including moderate rainfall events often overlooked by conventional approaches.
Moreover, the model addresses the increasing unpredictability of precipitation patterns driven by climate change. Short-term heavy rains can lead to abrupt contamination spikes, challenging standard lab monitoring that operates on lagging schedules and often misses these transient yet critical surges. By synthesizing meteorological data with hydrologic indicators in near real time, the AI framework anticipates these volatile water quality changes, facilitating timely advisories that align with rapid environmental shifts.
The public health stakes tied to E. coli exposure in recreational waterways are significant. Infection can cause symptoms ranging from mild gastrointestinal distress to severe illness, with children, the elderly, and immunocompromised individuals at heightened susceptibility. Early warnings derived from predictive models reduce the risk of exposure, enabling healthier recreational experiences. This preventive edge could potentially lower the incidence of waterborne illnesses and associated healthcare burdens.
Alamdari’s team also highlights the broader implications for urban planning and infrastructure. Every decision impacting impervious surfaces and stormwater management influences downstream water quality and public safety. The research advocates for green infrastructure solutions that enhance natural filtration and mitigate contaminated runoff, thereby reducing the frequency and severity of contamination events. The predictive model serves as a tool not only for immediate health protection but as an informative guide for sustainable watershed development.
In essence, this AI-powered framework represents a paradigm shift in environmental health surveillance and management within urbanized watersheds. By enabling near real-time and next-day forecasts of E. coli concentrations, it empowers communities with actionable intelligence to protect both people and economy. The fusion of advanced data analytics with environmental science embodied in this research paves the way for smarter, more resilient water quality monitoring that adapts to the complexities of modern ecosystems.
This transformative research was recently published in the journal Water Research, showcasing the model’s empirical validation and underlining its applicability in diverse urbanized watershed contexts. Supported by grants from Florida State University, the work epitomizes the collaborative innovation necessary to confront emerging environmental health challenges in the 21st century.
The future of water quality management lies at the intersection of technology and nature, where timely data-driven insights inform proactive decisions. As recreational water users, public health officials, and local economies stand to benefit from this breakthrough, the integration of AI predictive tools could become the new standard, enabling safer and more sustainable interactions with our precious water resources.
Subject of Research:
Predictive modeling of Escherichia coli contamination in urbanized recreational waterways using artificial intelligence.
Article Title:
Near real-time and next-day prediction for Escherichia coli (E. coli) concentrations in highly urbanized watersheds
News Publication Date:
15-Feb-2026
Web References:
https://dx.doi.org/10.1016/j.watres.2025.125030
Image Credits:
Scott Holstein/FAMU-FSU College of Engineering
Keywords
Water quality, Artificial intelligence, E. coli prediction, Urban watersheds, Environmental modeling, Hydrometeorological data, Recreational water safety, Predictive monitoring, Public health protection, Climate impact on water, Green infrastructure, Waterborne illness prevention
