Urban swimming pools in Sub-Saharan Africa face increasing microbial contamination risks due to high bather loads, inconsistent disinfection, and weak regulatory oversight. Yet most facilities rely on reactive monitoring, where microbiological hazards are only confirmed after laboratory results become available, prolonging swimmer exposure to unsafe conditions. This study developed an integrated monitoring framework and a Nowcast logistic regression model to provide real-time contamination likelihoods for proactive public health protection. A systematic–stratified sampling approach was applied across four high-use outdoor swimming pools in Cape Coast, Ghana, generating 216 water samples for six physicochemical parameters (pH, temperature, turbidity, free chlorine, salinity, and dissolved oxygen) and two microbial indicators (total viable count and total coliforms). Data were analyzed using descriptive statistics, principal component analysis (PCA), and Nowcast-style logistic regression modeling. Results showed frequent exceedances of guideline targets, including unstable pH (4. 44–6. 93), variable free chlorine residual (0. 03–3. 71 mg/L), and elevated microbial loads (8. 33–28. 33 CFU/mL), reflecting intermittent disinfection failure and physicochemical instability. PCA identified free chlorine, temperature, pH, and turbidity as dominant variables separating low-risk and high-risk pool regimes, consistent with mechanistic controls governing disinfectant efficacy, chlorine decay, and particulate shielding of microorganisms. A logistic regression Nowcast model using these predictors achieved 77. 3% accuracy, 56. 5% precision, 100% recall, 67. 7% specificity, and an F1-score of 72. 2%, demonstrating strong sensitivity for detecting unsafe conditions. The model’s probabilistic output (Pₙow) enables rapid, evidence-based interventions such as corrective dosing, intensified monitoring, or temporary closure before laboratory confirmation. This study demonstrates that routinely measurable pool parameters can be translated into a practical decision-support tool, shifting recreational water management in tropical urban settings from retrospective compliance checks toward proactive, predictive risk control. • Quantifies physicochemical instability and microbial contamination across four high-use outdoor pools in Cape Coast (N = 216). • Shows frequent non-compliance with WHO guideline targets for pH and free chlorine. • Identifies free chlorine, pH, temperature, and turbidity as dominant mechanistic controls (PCA). • Develops a Nowcast logistic regression model to estimate real-time contamination probability (Pₙow). • Provides actionable control triggers for corrective dosing, intensified monitoring, or temporary closure.
Nyarko et al. (Wed,) studied this question.