Journal Design Science Quartz
African Rural Development Studies (Interdisciplinary - | 20 October 2023

A Scoping Review of Methodological Frameworks for Time-Series Forecasting in Municipal Water Risk Reduction

Diagnostics for South African Systems (2021–2026)
P, i, e, t, e, r, v, a, n, d, e, r, M, e, r, w, e, ,, T, h, a, n, d, i, w, e, N, k, o, s, i, ,, L, e, r, a, t, o, M, o, k, o, e, n, a
time-series forecastingwater risk reductionmethodological frameworksagricultural water
SARIMA models with climate variables dominate current forecasting approaches
Model performance highly sensitive to input data quality
Prediction intervals widen substantially under non-stationary drought conditions
Critical gap exists in frameworks tailored for data-sparse agricultural contexts

Abstract

{ "background": "Municipal water systems are critical for agricultural resilience, yet face escalating pressures from climate variability and demand growth. Effective risk reduction hinges on robust forecasting methodologies, but a synthesis of applicable frameworks for local contexts is lacking.", "purpose and objectives": "This scoping review aims to identify, map, and evaluate methodological frameworks for time-series forecasting applied to municipal water risk reduction, with a diagnostic focus on applications relevant to the agricultural sector.", "methodology": "The review followed a structured, five-stage scoping framework. Peer-reviewed literature and technical reports were systematically identified from major scholarly databases. Included records were charted and analysed thematically to synthesise methodological approaches, model diagnostics, and implementation contexts.", "findings": "A dominant theme was the integration of seasonal autoregressive integrated moving average (SARIMA) models with exogenous climate variables, expressed as $yt = \\mu + \\sum{i=1}^{p}\\phii y{t-i} + \\sum{i=1}^{q}\\thetai \\epsilon{t-i} + \\sum{j=1}^{m}\\betaj x{t,j} + \\epsilont$, where $x{t,j}$ represents rainfall or temperature. Model performance was highly sensitive to input data quality, with prediction intervals widening substantially under non-stationary drought conditions.", "conclusion": "A suite of hybrid statistical and machine learning frameworks shows diagnostic potential for municipal water forecasting. However, their utility is constrained by systemic data gaps and a lack of contextual adaptation for integrated agricultural water management.", "recommendations": "Future research must prioritise the development of accessible, data-sparse modelling toolkits and foster interdisciplinary collaboration between hydrologists, data scientists, and agricultural planners to enhance model operationalisation.", "key words": "water security, predictive modelling, risk assessment, agricultural water, municipal services, diagnostic evaluation", "contribution statement": "This review provides a novel diagnostic synthesis of forecasting methodologies, identifying a critical gap in frameworks tailored for data