{ "background": "Persistent inefficiencies in water treatment yield present a critical challenge for infrastructure in many developing nations, directly impacting water security and resource management. Existing evaluations often lack robust, forward-looking analytical frameworks tailored to local operational data. ", "purpose and objectives": "This study conducts a comparative methodological evaluation of yield assessment techniques and develops a bespoke time-series forecasting model to predict and improve treated water output. The objective is to identify the most reliable methodological approach for local conditions and provide a predictive tool for strategic planning. ", "methodology": "A comparative analysis of three methodological paradigms—deterministic mass-balance, empirical regression, and stochastic time-series—was performed using historical operational data from multiple facilities. The forecasting model employs an autoregressive integrated moving average (ARIMA) structure, formalised as \ (B) (1-B) ᵈ yt = \ (B) \, where model selection was based on minimising the Akaike Information Criterion with robust standard errors. ", "findings": "The stochastic time-series methodology provided a 23% more accurate yield representation than deterministic approaches when validated against withheld data. The ARIMA (1, 1, 1) model forecasts a significant upward trend in potential yield, with a 95% prediction interval indicating an improvement range of 12–18% over the forecast horizon, contingent on targeted infrastructure investments. ", "conclusion": "The comparative evaluation establishes the superiority of stochastic, data-adaptive methods for yield assessment in this context. The developed forecasting model offers a validated, practical tool for predicting treatment performance, enabling proactive management. ", "recommendations": "Water authorities should adopt stochastic time-series analysis for routine yield monitoring and integrate the proposed forecasting model into long-term asset management and capital planning cycles to prioritise interventions. ", "key words": "water treatment yield, time-series forecasting, ARIMA modelling, infrastructure efficiency, comparative methodology, resource management", "contribution statement": "This paper provides a novel comparative framework for evaluating water treatment methodologies and introduces a specifically calibrated forecasting
Osei et al. (Sun,) studied this question.