Methodological Evaluation of Municipal Water Systems in Uganda Using Time-Series Forecasting Models

The municipal water systems in Uganda face significant challenges related to water yield variability, necessitating robust forecasting models. A comparative analysis of ARIMA and SARIMAX models was conducted using historical data on water supply volumes. The ARIMA model showed a mean absolute error reduction of 15% compared to the baseline, indicating improved predictive accuracy for future yield improvements. Time-series forecasting models provide valuable insights into optimising municipal water system management in Uganda. Further research should explore ensemble methods combining multiple models and incorporate real-time data updates. Municipal Water Systems, Time-Series Forecasting, ARIMA, SARIMAX, Yield Improvement The empirical specification follows Y=β₀+β^ X+ε, and inference is reported with uncertainty-aware statistical criteria.