Sustainable water quality underpins ecological balance and water security. Assessing and managing lakes and reservoirs is difficult due to data sparsity, heterogeneity, and nonlinear relationships among parameters. This review examines how Self-Organizing Map (SOM), an unsupervised AI technique, is applied to water quality assessment. It synthesizes research on parameter selection, spatial and temporal sampling strategies, and clustering approaches. Emphasis is placed on how SOM handles multidimensional data and uncovers hidden patterns to support effective water management. The growing availability of environmental data from in-situ sensors, remote sensing imagery, IoT technologies, and historical records has significantly expanded analytical opportunities in environmental monitoring. SOM has proven effective in analysing complex datasets, particularly when labelled data are limited or unavailable. It enables high-dimensional data visualization, facilitates the detection of hidden ecological patterns, and identifies critical correlations among diverse water quality indicators. This review highlights SOMs versatility in ecological assessments, trophic state classification, algal bloom monitoring, and catchment area impact evaluations. The findings offer comprehensive insights into existing methodologies, supporting future research and practical applications aimed at improving the monitoring and sustainable management of lake and reservoir ecosystems.
Water quality is fundamental to environmental management and sustainability, directly impacting both human and ecosystem health. The United Nations (UN) Sustainable Development Goals (SDGs), targeted for achievement by 2030, recognize access to clean water as a basic human need and a key factor linked to several global challenges. The UN emphasizes the need for international assistance and cooperation to reduce water pollution, enhance water quality, improve wastewater treatment, and minimize water losses (United Nations, 2012). Despite efforts by national agencies and international organizations such as the World Health Organization (WHO) to regulate and improve water quality (Boyd, 2019;World Health Organization, 2017), maintaining safe and sustainable water resources remains a global challenge. In the European Union (EU), only 40% of surface water bodies achieve good ecological status (European Environment Agency (EEA), 2021). Similarly, in the United States, data from the National Lakes Assessment (NLA) in 2012 and 2022 reveal ongoing concerns regarding nutrient pollution in lakes. In 2012, 35% of lakes had excess nitrogen and 40% had excess phosphorus. By 2022, these proportions had risen to 47% and 50%, respectively. Biological conditions also increased, with the percentage of lakes in poor biological condition rising from 31% in 2012 to 49% in 2022. These findings underscore the persistent and growing threats to aquatic ecosystems (U.S. Environmental Protection Agency, 2016Agency, , 2024)).
Reservoirs and lakes are the world’s primary sources of readily accessible freshwater for municipal, industrial, agricultural, and environmental uses. Increasing stressors related to ageing infrastructure, land-use change, climate variability, intensified seasonality, and rising water demand exacerbate the difficulty of protecting surface-water quality for both human and ecosystem health (Perera et al., 2021;Juracek, 2015;Asres et al., 2025). Common water-quality problems in reservoirs include sediment accumulation, turbidity increases, eutrophication, and recurrent algal blooms, all of which impair storage reliability and ecosystem function (Kondolf et al., 2014;Winton et al., 2019;Juracek, 2015).
Managing water quality in reservoirs and lakes remains challenging due to the dynamic nature of freshwater systems, where interactions among biological, chemical, and physical factors are often nonlinear, highly heterogeneous, and difficult to predict. These complexities make consistent assessment and timely intervention particularly important. Monitoring efforts are further constrained by data heterogeneity, limited spatial and temporal coverage, and environmental variability, all of which hinder compliance with national and international water-quality standards. These obstacles create increasingly complex management trade-offs and uncertainty (Wu et al., 2023), underscoring the need for continuous, reliable monitoring of water-quality dynamics in lakes and reservoirs (Miranda and Faucheux, 2022).
Advances in AI and computational tools offer promising solutions to previous challenges. AI can handle data from various sources, including in-situ measurements (Lap et al., 2023;Fernández del Castillo et al., 2024), remote sensing technologies (Tang et al., 2023;Li et al., 2024), IoT (Singh et al., 2022;Bhardwaj et al., 2022), historical archives (Ha et al., 2015;Rimet et al., 2009;Zhang et al., 2021), as well as recent approaches that integrate heterogeneous data resources (Anand et al., 2024;Xia et al., 2024;Mamun et al., 2024) to analyse patterns, optimize processes, and support predictive modelling. Various AI methods have been explored for water quality assessment, including machine learning (ML) algorithms such as k-means clustering, support vector machines, and neural networks (Pérez-Beltrán et al., 2024;Lowe et al., 2022).
Beyond technical complexity, water quality management is also a high-risk domain in which decisions carry direct consequences for environmental and human health, with direct implications for water security. In such contexts, human supervision together with trustworthy and responsible AI systems, is increasingly emphasized internationally to ensure that automated analyses remain reliable, transparent, and under expert oversight (NIST, 2023;European Commission, 2019;Australian Government, 2022;OECD, 2022). This orientation towards human-centred governance highlights the need for intI interpretable AI (IAI) models that promote transparency, reproducibility, and trust in environmental decision making (Gilpin et al., 2018;Rudin, 2019).
Among AI approaches, SOM introduced by Kohonen (1982b,a) and later formalized in his subsequent publications (Kohonen, 2001(Kohonen, , 2013)), offers distinct advantages for exploring and visualizing complex environmental data by preserving topological relationships between data points. SOM is a white-box model that is transparent, interpretable, and naturally aligne
This content is AI-processed based on open access ArXiv data.
