Leveraging machine learning to optimise index-based aquifer water quality characterization
Résumé
Abstract The Cape Flats Aquifer (CFA) is a strategic water source for the City of Cape Town that is threatened by several contamination sources (including informal settlements and agricultural activity). In this study, a Water Quality Index (WQI) was developed using five years (2018–2023) of groundwater quality data. Subsequently, an artificial neural network (ANN) was constructed to optimise the monitoring of the aquifer, and a pollution source delineation was conducted. It was found that there were consistent spatiotemporal trends in the WQI distributions over the study area, with the Northern extent consistently producing the lowest quality water. Furthermore, the pollution source delineation indicate that expansion of agricultural activities (including irrigation and manure runoff) is a large contributor to contamination within the CFA. Finally, our constructed ANN model was able to predict WQI values for 2024 and 2025 using a fraction of the original input parameters (from 23 to 5) while maintaining high accuracy (training R 2 = 0.981, test R 2 = 0.976). This study therefore provides valuable insights into contamination sources within an important groundwater system while simultaneously presenting the value of machine learning approaches to optimise monitoring, with implications for data scarce regions in Sub-Saharan Africa.
Citer ce document
Exporter : BibTeX · RIS (Zotero, Mendeley, EndNote)
Accès au document
Texte intégral archivé sur SPHAERO, en accès ouvert : lisible gratuitement en ligne après connexion. Le document reste aussi disponible à sa source. Se connecter
Voir l'article sur le site de la revueLicence et provenance
Licence : CC BY
Notice moissonnée depuis OpenAlex le 16/11/2026. Le document reste hébergé par sa source.
Voir le document à la source →