{ "background": "Water treatment systems in Ethiopia face significant operational challenges, leading to variable performance and public health risks. Current risk assessment methods often lack the capacity to integrate sparse, multi-level data and quantify uncertainty for infrastructure management decisions.", "purpose and objectives": "This study aimed to develop and evaluate a novel Bayesian hierarchical modelling framework for the methodological assessment of risk reduction in water treatment facilities. The objective was to provide a robust tool for quantifying performance improvements and associated uncertainties.", "methodology": "A Bayesian hierarchical model was constructed, integrating facility-level operational data with regional environmental covariates. The core model structure is $y{ij} \\sim \\text{Normal}(\\alphaj + \\beta X{ij}, \\sigma^2)$, $\\alphaj \\sim \\text{Normal}(\\mu{\\alpha}, \\tau^2)$, where $y{ij}$ is the risk metric for facility $i$ in region $j$, $\\alphaj$ are region-specific intercepts, and $X{ij}$ are covariates. Model inference used Hamiltonian Monte Carlo sampling.", "findings": "The model demonstrated a high predictive capacity for system failure risk, with posterior credible intervals for key performance coefficients excluding zero. A principal finding was that improved coagulation control was associated with a median estimated 34% reduction in turbidity-related risk across the evaluated facilities. Uncertainty was successfully partitioned into facility and regional components.", "conclusion": "The proposed Bayesian hierarchical model provides a statistically rigorous methodology for evaluating risk reduction in complex water treatment systems. It effectively synthesises heterogeneous data and quantifies uncertainty, offering a superior alternative to conventional deterministic assessments.", "recommendations": "Adoption of this modelling framework is recommended for asset management planning by water authorities. Future work should focus on integrating real-time sensor data to enable dynamic risk forecasting.", "key words": "Bayesian inference, hierarchical modelling, risk assessment, water treatment, infrastructure reliability, uncertainty quantification", "contribution statement": "This paper presents a novel probabilistic framework for infrastructure risk assessment