Water Security: A Systematic Review of Definitions, Indicators, and Artificial Intelligence Applications

Abstract

Water security is crucial for human well-being and environmental sustainability. The rapid increase in urbanization, climate change, pollution, etc., leads to water scarcity in many parts of the world. Therefore, it is important to understand the concept and growing challenges of water security. Using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework, 146 articles were identified for this study. The results highlight a novel aspect of the definition of water security, presenting it in a simpler, broader way with key components. The indicators for assessing water security are categorized into quantitative, qualitative, and combined types, and are further arranged across different dimensions, domains, and spatial scales. The study also examines Urban Water Security assessment methods and categorizes them into distinct methodological groups. Additionally, the studies show that only 25 articles explore artificial intelligence in the context of water security indicators. This reveals the need to address the gap between artificial intelligence and the assessment of water security. From these limited articles, artificial intelligence types and models were identified, and their applications were grouped into thematic categories. In general, this study supports improved assessment, decision-making, and sustainable water security management.

1. Introduction

Securing water security is a major challenge in the 21st century [1,2,3]. With increasing population growth, urbanization, and climate change, the availability of freshwater resources has decreased [4,5,6,7]. These issues not only affect human and ecosystem health but also agriculture, food production, energy generation, and other aspects of global stability [8]. The concept of water security was initially developed by international organizations such as the Global Water Partnership (GWP) [9], UN-Water [10], WaterAid [11], and researchers [12,13,14]. Studies have used either a simpler definition [4,15] and broader definition [16,17]. However, there are no review articles that define water security in a simple, broader way with various components.

Water security nexus was previously limited to indicators such as availability and access, but today it encompasses social, political, economic, and environmental factors [18,19,20,21]. The drastic increase in water demand is directly influenced by anthropogenic activities, thereby enhancing ecosystem services related to water security [22]. Moreover, energy extraction activities, including reservoir and shale resource development, along with associated developmental protection measures, are important considerations for water security due to their potential impacts on water quality, water availability, flowback water management, and overall water resource sustainability [23,24]. Water security is always embedded in multiple dimensions, domains, indicators, and time [25,26]. Ahopelto et al. emphasize that water security indicators are not only tools for assessment but also serve as instruments for clarifying governance priorities [27]. However, existing review studies have not categorized water security dimensions into qualitative, quantitative, and combined types along with their corresponding dimensions, indicators, domains, and scales.

Ensuring a reliable and safe supply of drinking water has become a major challenge amid increasing urbanization, population growth, and climate change [6,28]. This issue is particularly critical in developing countries, where rapid population expansion and limited technological capacity make it difficult to sustain Urban Water Security [29,30,31]. The assessment approaches evaluate how well cities can provide safe, reliable, and sufficient water under these pressures. These approaches help to improve water management. Nevertheless, no review studies have provided methods used in Urban Water Security assessment.

In order to improve urban water management, Hidayat and Kurniawan [32] explore smart water technologies for water management, focusing on improving water use efficiency and water balance. Optimal water utilization in communities, countries, or river basins primarily depends on water supply systems that employ a wide range of intelligent techniques [33]. Nowadays, Smart water systems enhance water security by leveraging digital technologies to monitor, maintain, and optimize water resources in real time, reducing water loss. In addition, water resource management could be optimized using machine learning and deep learning methods for water security assessment, and through many other applications, such as automation of water usage monitoring, anomaly detection in water flow, malfunctioning of water meters, leak detection in water supply systems, real-time data water quality monitoring, drought prediction, and preventive maintenance of water pipes [34]. There have been many literature reviews and case studies on water security in recent decades, yet there remains a lack of information on emerging technologies such as Artificial Intelligence and machine learning, which address the fundamental goals of securing water. Nevertheless, there is a lack of studies detailing the artificial intelligence techniques and their applications used in water security.

In this study, 146 articles were selected from Google Scholar, Web of Science, and Scopus, and 25 were identified as reviews. As illustrated in Figure 1, these review studies are categorized into three main categories: General Water Security (GWS), Agricultural Water Security (AWS), and Urban Water Security (UWS). The focus areas and spatial scales addressed within each category are examined to provide the scope and distribution of existing review articles. Among the 25 articles, only five specified the type of review conducted. A narrative review by Akiyode and Akiyode [35] provides an assessment of the impacts of climate variability and change on Urban Water Security, with a focus on the urban scale. Cook and Bakker [25] presented a comprehensive review focused on the definition of water security on multiple scales. Likewise, Xenario et al. [36] carried out a bibliometric analysis focusing on the assessment and frequency of water security at the regional level. A systematic review by Zainuddin et al. [37] focused on monitoring and evaluation of water security at the urban scale, and Gerlak et al. [38] on the definitions, risk, and governance across city, regional, national, community, and transboundary scales. The remaining 15 articles did not explicitly state their review type and were therefore classified based on the review types proposed by Grant and Booth [39]. Following this, Aligholi and Hayati [1] conducted a critical review of theoretical perspectives on Agricultural Water Security. Governance-related aspects across multiple scales were discussed by Bakker and Morinville [13], while McNeill et al. [40] examined definitions across global, regional, basin, and national contexts. Octavianti and Staddon [41] focused on water security assessments across a wide range of spatial scales, from household to global levels. Water security Policy complexity was analyzed by Zeitoun et al. [42]. In addition, Allan et al. [43], Hoekstra et al. [44] and Mishra et al. [45] focused on definitions, governance, and assessment at urban and global scales, and these were categorized as narrative reviews. Further, a scoping review by Dickson et al. [46] examined water security assessment tools. Systematic reviews by Garrick and Hall [47] and Jepson et al. [48] focused on water security risks and water insecurity across multiple scales, including community, basin, and household levels. Overview reviews by Gelfan [49] and Marcal et al. [50] provide definitions of climate change and water security across multiple scales. A state-of-the-art review by Ludwig et al. [51] provides insights into the impacts of climate change in small and medium-sized catchments, while an integrative review by Su et al. [52] focuses on definitions of urban water security. In addition, five studies from the themed issue on Environmental Change Assessments were included in the synthesis focused on adaptive capacity: Scott et al. [14] and Lankao and Gnatz et al. [53] addressed assessment approaches, Kirchhoff et al. [54] examined adaptive capacity and management, Sun et al. [55] discussed definitions and metrics, and Varady et al. [56] analyzed adaptive capacity. More details on the review articles are provided in Supplementary Table S1.

Although existing studies provide valuable information, there is a lack of a clear, simple, and broader definition of water security. Lack of categorizing various indicators as quantitative, qualitative, and combined with respect to their dimensions, domains, and scales, lack of information regarding the methods to evaluate Urban Water Security, and limited availability of information about artificial intelligence in water security. This review addresses this gap by aligning with the study’s main objectives.

• Assess the spatial and temporal distribution of relevant scientific publications to understand geographic research emphasis and the evolution of research trends.
• To synthesize water security definitions into simple and broader concepts based on key components.
• To identify and categorize water security indicators based on quantitative, qualitative, and combined approaches across different dimensions, domains, and scales.
• To identify different analysis methods used in Urban Water Security assessment by grouping them into eight categories.
• To identify and examine the applications and types of artificial intelligence with their corresponding models used in water security, and synthesize them into six thematic categories.

2. Materials and Methods

2.1. Literature Search and Study Selection

This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. It has four stages: Identification, Screening, Eligibility, and Inclusion, as shown in the flow chart in Figure 2.

Stage 1: Identification: In this study, Google Scholar, Web of Science, and Scopus were used as the databases. The keywords used for this search were “Water security indicators,” “Artificial Intelligence,” “Water security indicators,” and “Machine learning”.

Using Google Scholar, an initial search was conducted on 7 February 2025, and 75 records were identified and downloaded. On 28 February 2025, a second search was conducted using the same keywords, yielding 25 articles. Furthermore, on 17 April 2025, an additional 84 records were identified and retrieved to enhance the quality of the analysis, yielding all together 184 articles from Google Scholar. Subsequently, on 4 May 2026, an expanded literature search was conducted across Scopus, which identified 6 articles, and across Web of Science, which identified 99 articles, yielding a total of 105 articles. From the total of 289 articles identified, 106 articles were excluded due to duplication, preprints, dissertations, and reports.

Stages 2 and 3: Screening and Eligibility: A total of 183 records were screened for eligibility for the study analysis. During the screening process, 21 records were excluded as irrelevant based on their titles and abstracts. During the eligibility stage, 162 records were assessed, and 16 were excluded after full-text assessment for being irrelevant to the study’s scope. Study selection was conducted by a single reviewer, who screened the articles based on predefined inclusion and exclusion criteria.

Stage 4: Inclusion: The final stage of the PRISMA framework is the inclusion. The 146 records were selected through eligibility criteria and included in this study’s analysis.

2.2. Information Retrieval and Interpretation Approaches

Data from eligible articles were extracted according to the objectives of this study and organized in Excel spreadsheets and Word tables to facilitate compilation and analysis. Data extraction was conducted by a single reviewer, who systematically collected relevant information from the included studies. To examine the spatial and temporal distribution of the studies, the publication year and study location were recorded in Excel. The results were then visualized using bar graphs and a pie chart to illustrate publication trends. In addition, geographical maps were generated using Map Chart to show the global distribution of the reviewed studies.

To develop a simple, broader definition of water security, descriptive and definitional information was extracted from eligible articles. Key components related to water security, such as access, quantity, quality, governance, sustainability, risk, protection, capacity, beneficiaries, outcomes, links, types, scales, impacts, needs, and key aspects, were identified from the information for a broader definition. These components were compiled and compared in a table. Based on this synthesis, the broader definition was visually organized into a conceptual figure to represent water security. Furthermore, the dimensions and indicators were identified from the selected articles and collected. Indicators are then synthesized into different dimensions and further categorized into quantitative, qualitative, and combined types, along with their respective domains and spatial scales, in a table format. Key information, such as assessment methods and spatial scales, was extracted from articles on Urban Water Security. Methods were then grouped in a table and organized into two levels; a linear dendrogram was used to show their relationships. The artificial intelligence techniques and their applications were extracted from the reviewed articles and compiled into a table. The applications were then grouped into themes, and a sunburst chart was created to show their distribution.

3. Result

3.1. Temporal and Spatial Distribution of the Studies

The temporal distribution of publications shows an increase in research on water security, as illustrated in the bar chart in Figure 3. Between 2008 and 2015, the number of studies was very low, averaging less than five per year. Since 2015, the number of publications has increased, indicating growing scientific interest in water security research. The number of studies peaked in 2024 with 23 publications, followed by 17 in 2023 and 18 and 16 in 2022 and 2021, respectively. This trend has led to a rapid increase in research on water security indicators over the past few years. The geographical distribution of studies in Figure 4 shows that research on water security is concentrated in several key countries. China presents the most studies (25), followed by the United States (11), India (9), and the United Kingdom (7). Ghana (4), Brazil (3), Germany (3), and Iran (3) also contribute research studies. Several other countries, including Malaysia, Vietnam, Indonesia, Australia, Canada, Nepal, Turkey, Mexico, and South Korea, each have two studies, while many other countries contribute a single study. Moreover, Figure 5 presents a world map of research activity distribution, highlighting the global significance of water security indicators. This increase in research on water security is mainly due to growing concerns about population growth, rapid urbanization, climate change, increasing water demand, water pollution, resource depletion, and challenges in water governance and management. These factors have significantly influenced water security.

3.2. Definition of Water Security in a Simple and Broader Context

Among the reviewed articles, the definition of water security appears in 54 articles. The definitions are then synthesized and categorized in a simple, broader context. In a simple context, water security is defined as a multi-dimensional concept provided by Doeffinger and Hall [15] and a dynamic concept given by Dou et al. [4]. While exploring the broader concept of water security, several key components have emerged to define it, as illustrated in Figure 6. The components are as follows:

• Access is one of the most emphasized components in definitions. The foundational definition from UN-Water [10] gives the importance of access as “safeguard sustainable access,” which has been widely adopted in many studies [27,57,58]. Similarly, the definition from the GWP [9] describes access to “sufficient safe water” mentioned in the articles [17,59,60,61], where Alavijeh et al. described it as “sufficient water” [29]. In addition, it has been further refined to include aspects such as “adequate access,” as defined by Berner et al. [62], “stable and affordable access” by Penn et al. and Wehbe et al. [63,64] and “equitable access” by Srisuwan [16].
• Water quality is widely recognized as a fundamental component of water security, where water is safe for the environment. The definitions of UN-Water [10] and GWP [65] emphasize acceptable quality, which has been widely adopted across multiple studies, such as [12,45,61,66,67,68]. In addition, UN-Water [69], WaterAid [11], and some studies describe quality using the terms “good and sufficient quality” [59,63]. Other definitions describe water quality as “suitable quality” for different uses and emphasize the importance of providing “provision of quality drinking and domestic” water [29,70]. Li et al. emphasize the importance of meeting quality requirements, and Scott et al. define “adequate quality” of water in terms of availability [14,71].
• The component quantity in the water security definition is commonly described in terms of “adequate” by UN-Water [10], GWP [65] which is mentioned in many articles, and this is also defined by Evengard et al. [67], Jesnen and Wu [68], Scott et al. [14] and Zakeri et al. [17]. It is also described as an “acceptable quantity” [12] and “sufficient water quantity” by Water Aid [11] and several studies [59,63,71,72].
• Water security definitions often specify the broad range of beneficiaries, including humans, ecosystems, the environment, people, economies, industry, and resilient societies and communities, as provided by the numerous studies such as [17,67,68,70,71,72,73,74,75,76] and in the definition of the UN-Water [10], Water Aid [11], and GWP [65].
• Governance component in the definition encompasses aspects related to managing and planning by Biswal et al. [77], water demand by Dawuni et al. and Biswal et al. [59,77,78]. Dou et al. give the stakeholder interest aspect in the definition [4]. While Kreuger et al. provide the aspect of services that citizens receive [79]. Jepson et al. mention the ascendance in policy circles, academic scholarships, and hydro-social processes [18]. UN-Water also mentions peace and political stability [80,81]. And finally, the importance of a stable political environment is given in the definition by Srisuwan [16].
• The component risk is associated with an acceptable and tolerable level of water-related risks [12,17,29,66,82,83]. It also includes poor water quality risk [12,84], reasonable risk [77], flood risk [4], and water-related hazards [85]. Together, these aspects highlight the role of risk management in definitions.
• Capacity for water security is primarily associated with population size, and in a city or urban context. The definitions proposed by UN-Water [10] and Xi et al. [86] highlight a population’s capacity, while the definition of several other studies gives importance to the capacity of an urban or city for water security [68,87].
• The component needs will represent the purposes that make water security essential. Existing definitions indicate that water security is needed to support human health, the needs of life, the ecosystem, and products [74], and finally societal needs [75].
• Definitions specify the impacts on water security, which include climate change, as given by Wehbe et al. [64], uncertain global changes [17], human activities and natural causes [76]. In addition, emerging risks such as food-borne threats are also identified from the definition [62].
• The component scale from the definition highlights that water security is addressed across multiple spatial scales, which extends from the individuals [18], from household to global [60,88], and from the city and its catchments [68].
• The component types reflect the different contexts of water security. Some existing definitions, particularly focusing on the type of water security, including watershed-based [45,65], water resource-based [85], food and drinking water security [62], Urban Water Security [68,79], human water security [70], urban water supply security [79] and water security management [76].
• The component sustaining and sustainability component in the definition emphasizes sustaining livelihood and human well-being [10,16,68,81]. In addition, the definition also specifies the sustainable use and protection [85], sustainable access [17,45,67], sustainable availability [14], and sustainable development of water resources [85], along with some metrics such as the ratio exceeding unity, which often signifies water surplus [59].
• Protection in definition represents safeguarding water resources from waterborne pollution and water-related disasters (floods and droughts), as mentioned in many studies, such as [10,16,27,68,81,89]. It also emphasizes ensuring against the natural environment [60,61,88] and protection against some other pollutions [16].
• Component links reflect the interconnected nature of water security with different sectors. Taka et al. give the definition, which specifies the linkages between water and food, energy, climate, and human security [75].
• The definition also contributes to a wide range of outcomes, including socio-economic development [81], industrial and agricultural development [71], and energy generation. Additionally, water security aims to ensure a clean and productive life [60,61,88] and maintenance of the ecosystem and biodiversity [70]. Moreover, the definition specifies adequate, reliable, and affordable water for a healthy life [73], to maintain health, and to enact livelihoods [63], safety, reliability, continuity, and affordability [79].
• The definition highlights several key aspects, including affordable costs [60,61,88]; addressing the lack of water [84]; availability, adequate supply, and adequate information [62]; fresh water supply [64]; water management and water scarcity [4]; human development [73]; performance of the system function [79]; availability of water resources [71]; water sources [74]; and water sufficiency and equity [75]. These aspects underscore the multidimensional nature of water security and highlight the importance of integrated management.

The studies that provide the definitions are listed in Supplementary Table S2.

Figure 6. Dimensions of water security for a broader definition.

Figure 6. Dimensions of water security for a broader definition.

3.3. Water Security Indicator Classification: Quantitative, Qualitative, and Combined

Among the 146 studies, 36 articles provided the dimensions and indicators used to analyze water security. From these studies, indicators are synthesized and categorized into quantitative, qualitative, and combined (quantitative and qualitative) indicators, along with the synthesized dimensions, domains, and spatial scales, detailed in

Table 1. Classification of water security indicators based on quantitative, qualitative, and combined, along with dimensions, domains, and scales.

EM	Dimensions	Indicators	Citations	Domain						Spatial Scales
				WS	WR	S&H	E&E	WMS	SE
Quantitative	Availability	Freshwater resources per capita, water consumption per capita, urban population covered by urban water (people), volume of surface water, groundwater volume, annual water yield, monthly water supply, demand ratio, reliability of supply	[4,29,57,100]	✔	✘	✘	✘	✘	✘	National
		Internal renewal of water resources volume, external renewal of water resource volume, outflow to the outside of the basin, total natural renewable water resources, water resources per capita, dependency to external water resources ratio, dependency to groundwater ratio, relative water stress index, consumption index, electrical conductivity, total water withdrawals, total actual renewable water resources, population, dam storage capacity, per capita water availability, meteorological variation, ratio of ecological water to total water demand, exploitation ratio of water resources, surface water pollution index, reclaimed water utilization, rainwater resource utilization, the degree of water satisfaction, total renewable water resources per capita, ecosystem vitality index, per capita renewable water, water consumption, types of water sources, per capita water needs for human consumption	[4,22,85,91,98,101,102,103,104,105]	✘	✔	✘	✘	✘	✘	Regional, global basin, sub-regional
	Access	Percentage of households with access to the piped water supply, supply duration by LCs per hour in a day, percentage of unsafe water supply systems, percentage of total population with access to water supply, the proportion of the water network that is old and worn out, number of household water subscribers, coverage of human water uses, water supply capacity, water supply coverage, piped water coverage, accessibility of running water, periodic water shortage, operational status of water source, length of the water distribution network, water use per capita percentage, population density, piped water coverage, sewage coverage affordability, cost recovery of water utilities, accessibility to clean water, accessibility to clean water from centralized water supply system, efficiency of the centralized water supply systems, access to piped drinking water, access to wastewater collection, one way distance to water source, waiting time, seasonal resource variability	[16,22,29,30,57,68,90,100,102,103,106]	✔	✘	✘	✘	✘	✘	City, basin, national, sub-regional blocks, regional
		The percentage of the city’s population having access to LC’s sewage system (percentage of network coverage), the percentage of old and worn-out sewerage network (percentage of network capacity), percentage of the total population with access to sanitation services, the percentage of efficiency of the wastewater treatment plant, wastewater discharged per capita, nominal capacity of water treatment plants, number of water laboratories, percentage of population covered by municipal sewage facilities, water treatment level, sewage water coverage, state of sewer, waste water coverage, domestic sewage discharge, industrial wastewater discharge, emission of industry, population in sewage network, population in drinking water sewage network, access to safe water, access to improved sanitation, access to clean water, accessibility to clean water from a centralized water supply system, water sanitation, access to waste water collection, water treatment capacity, percentage access to improved drinking water	[16,30,57,60,90,100,103,105]	✘	✘	✔	✘	✘	✘	City, provinces, regional
	Water economy	Non-revenue water (NRW): water loss (leakage and theft, etc.)/water produced, energy consumption for water production, energy consumption for wastewater treatment, contribution of alternative energy to the operation of the water supply, selling price of one cubic meter of water, selling price of one cubic meter of sewage services, WASH budget: percentage of national budget directed to water and sanitation services, operating revenue: operation and maintenance cost recovery, water productivity, value added by industries, industrial water withdrawals, water budget, drainage investments, household water costs, investing in water infrastructure, affordability (water tariff)	[17,29,30,57,60,68,90,98,107]	✘	✘	✘	✘	✘	✔	City, global, national, provinces, watersheds
	Social	Population, per capita renewable water resources, ratio of employment in the agriculture sector to total employment, ratio of employment in the industrial and mining sector to total employment, ratio of employment in the urban services sector to total employment, labor productivity in the agriculture sector, labor productivity in the industrial and mining sector, labor productivity in the urban service sector, labor productivity in the region, employment productivity in the agriculture sector, employment in the agriculture sector to water consumption in the agriculture sector, employment productivity in the industrial and mining sector, employment productivity in urban services sector, employment productivity in the region, per capita income, marginal change in labor productivity in the region, marginal change in employment productivity in the region, literacy rate, population density, income, inequality, informal dwellings, gender equality, corruption perception index, environmental performance index, citizen support (consumer awareness and interest: questionnaire five-point likert scale), total water withdrawal, total actual renewable water resources, per capita GDP, population density, water productivity, water supply per capita, industrial water scarcity, agricultural water scarcity, population pressure	[22,30,57,85,90,91,98,99,108,109]	✘	✘	✘	✘	✘	✔	City, regional, global, basin, country
	Economic	Ratio of water consumption in agricultural sector to total water consumption, ratio of water consumption in industrial and mining sector to total water consumption, ratio of water consumption in urban services sector to total water consumption, water productivity in the region, water productivity in agricultural sector, water productivity in industrial and mining sectors, water productivity in urban services sector, relative importance of agriculture in the economy, total water withdrawals, relative importance of agriculture withdrawal in local water balance, marginal change in water productivity in the region, marginal change in water productivity in agriculture sector, water intensity, water consumption for economic activities, water use for agriculture, water use for inland waterway transport, water use for industry	[91,99,106]	✘	✘	✘	✘	✘	✔	Regional, deltas of the transboundary river basins, city
	Water Demand	Water demand per 10,000 rmb of industrial added value, industrial added value growth rate, per capita domestic water demand, water demand per mu for irrigation, tertiary industry added value growth rate, water demand per 10,000 rmb of tertiary industry added value, effective irrigated area growth rate, population growth rate	[72]	✘	✘	✘	✘	✘	✔	City
	Water self-sufficiency and supply diversification	Local water abstraction ratio (%), supply internationalization, water import rate, sourcing, distribution diversity index, water quantity sufficiency, diversity of sources	[110]	✔	✘	✘	✘	✘	✘	City
	Climate change	Annual average temperature, annual average precipitation, greenhouse gas emission, temperature increase, extreme rainfall events, climate risk index, daily mean average minimum temperature by month, average precipitation by month	[22,30,90,103,111,112]	✘	✘	✘	✔	✘	✘	Basin, city, regional
	Efficiency	Water intensity, water loss rate, residential water use, energy usage efficiency, wastewater reuse (recycling), efficiency of centralized water supply systems, residential water use, water intensity, water use intensity, water loss rate, wastewater treatment efficiency, wastewater reuse rate	[30,90,93,110]	✔	✘	✘	✘	✘	✘	City
	Driving force	GDP annual growth rate, per capita GDP, natural population growth rate, urbanization rate, population density, pollutant emissions from planting per hectare of arable land, pollutant emissions from livestock and poultry breeding per square kilometer of land area, fertilizer consumption per hectare of arable land, pesticide consumption per hectare of arable land	[94,109]	✘	✘	✘	✘	✘	✔	City, basin
Combined	Sustainability	Water uses per capita, water stress, sustainability (utilities), utility performance indicator, percentage of study area under natural vegetation, wastewater treatment plant discharge per capita, renewable freshwater resources per inhabitant, water exploitation index	[22,60,63,95,108]	✘	✔	✘	✘	✘	✘	Basin, country, city, provinces, rural
	Risk and disasters	Flooding/runoff risk: the flood-prone area as a percentage of the total surface area, sewer system blockages: stormwater and wastewater network effectiveness, water replenishment, population geospatial distribution, flood historical data, vulnerability level of the population, risk to temperature variations, risk to precipitation variation, flood deaths, disaster mitigation, disaster preparedness, public health risk, flood frequency, drought frequency, flood prone areas, people living under hazardous zones, risk management, major drought blackspots, articles on flooding in local media, capacity to cope with disasters, saltwater intrusion factor, flood damage, drought damages, water contamination incidents	[15,16,66,68,87,93,98,106]	✘	✘	✘	✔	✘	✘	Basin, global, city, country, deltas of the transboundary river basins
	Water safety	Leakage rate of the pipe network, drinking water source, water quality compliance rate, network-based per capita, safety level of drinking points, safety level of sanitation facilities, drinking water related contaminants, waterborne diseases, authorized water quality assurance, safety plans, proportion of population using safely managed drinking water service, percentage of safely treated wastewater flows	[57,97,98,107,108]	✔	✘	✘	✘	✘	✘	City, global, country, global,
	Governance	Institutional factor: questionnaire five-point likert scale, adaptability factor: questionnaire, five-point likert scale, staff productivity: no. of staff in the water supply and sanitation directorate/1000 houses connection, residential water use, water intensity, water services (water coverage, water losses, continuity of water supply), number of local admin units, number of operational policies with the community preparation, overall management of the water sector, potential to adapt to future changes, citizen support, role and responsibility, access to data and information, stakeholder engagement, communication and access to information, public participation opportunities, equality and non-discrimination, WASH investment, organizational structure, strategic planning, disaster management, regulation, cooperation on water management, people service (lacking complete plumbing, iwrm (degree of implementation), strategic planning, disaster management, management system	[16,30,57,68,87,90,92,98,100,102,106]	✘	✘	✘	✘	✔	✘	City, global, sub-region, basin
	Water environment	Urban population growth rate, air temperature, environmental flows: measured using the annual surface water low, water deficit, geospatial data on river morphology, geospatial data on fragmenting entities, state of natural water bodies, effect of polluting factors, aquatic species of conservation concern, annual quick flow (+), total nitrogen export, total phosphorus export, green areas, environmental safety, water pollution level, sediment transport, runoff, estimated soil loss by water erosion by land cover, average proportion of freshwater key biodiversity areas covered by protected areas, ecosystem vitality index, qualitative assessment of water quality, protection of water source, number of pollution sources, number of environmental impacts, conflict over water resources (human-wild life and human-life stock), water quality factor, upstream development activity, river flow for the environment, minimum ecological water demand distance, average wetland area, non point source pollution control rate, groundwater over exertion rate, green surface area: the green surface area as a percentage of the total surface area, urban landscaping, river quality	[4,15,29,30,66,87,96,97,98,99,100,106,108]	✘	✘	✘	✔	✘	✘	Basin, global, city, national
	Capacity	Urban landscaping, government expenditure on environmental protection, government expenditure on resource affairs, investment in water conservancy, efficient agricultural practices (%), household income, annual water investments, income from agricultural activities, years of education, management system, ownership over water source, water association registered, records kept, financial control, funds	[60,93,100]	✘	✘	✘	✘	✘	✔	City, provinces
	Infrastructure	accessibility, sewage infrastructure, drainage infrastructure, wastewater, investment need, CWS not in compliance, service discontinuity, service reliability, metering level, water loss, investing in water infrastructure, length of water distribution network	[15,17,29,30,93]	✔	✘	✘	✘	✘	✘	City, national, watersheds
Qualitative	Hygiene	Hand washing facilities, waterborne diseases, recreational opportunities, water sanitation, waterways city, incidence of recreational diseases,	[50,57,90,98,99,106]	✘	✘	✔	✘	✘	✘	Global, city, basin
	Quality	GW quality: groundwater samples that meet applicable quality standards (WHO and locally), surface water quality, wastewater treatment, drinking water quality, residual chlorine: percent samples with residual chlorine within the permissible limits, impaired stream/river length, water consumption chemical and biological contamination of water reservoirs, quality of water supplied, COD, ammonia and nitrogen, meeting WHO standards, waterborne diseases, water diseases	[15,16,57,71,87,90,113]	✘	✘	✔	✘	✘	✘	City, national, regional, basin

Notes: EM: Evaluation Method, WS: Water Supply, WR: Water Resources, S&H: Sanitation and Health, E&E: Environment and Ecosystem, WMS: Water Management System, SE: Socio-Economic.

. The quantitative indicators are the most widely used and cover key dimensions such as availability, access, water economy, social and economic factors, water demand, water self-sufficiency and supply diversification, climate change, efficiency, and driving forces. These indicators are applied across domains, including Water Supply, Water Resources, Sanitation and Health, Environmental and Ecosystem, and Socio-Economic systems, and the Water Management System. For example, they mainly identified measurable aspects of dimensions such as availability, which has an indicator (e.g., freshwater resources per capita) given by Alwathaf et al. [57], and access (e.g., piped water coverage, supply reliability) provided by Marcal et al. [90], while indicators of water economy (e.g., water loss: leakage and theft; water produced) are shown in [57]. Social and economic factors (e.g., ratio of employment in the agriculture sector to total employment; ratio of water consumption in the agricultural sector to total water consumption) are outlined by Bagheri et al. [91]. Water demand indicators (e.g., water demand per 10,000 rmb of industrial added value) are given in [72], while for water self-sufficiency and supply diversification (e.g., local water abstraction ratio (%); sourcing), indicators are presented by Jabari et al. [92]. For climate change, the indicators (e.g., annual average precipitation and annual average temperature) are provided by Arias et al. [22]. In addition, indicators of efficiency (e.g., residential water use; water intensity) are highlighted in [93], and indicators of driving forces (e.g., urbanization rate; natural population growth rate) are mentioned in [94]. These indicators are applied across multiple spatial scales, including city, provincial, basin, regional, national, and global levels. The qualitative indicators focus on non-quantifiable indicators of water security. These include dimensions such as hygiene and quality, having indicators related to contamination, compliance with standards, and waterborne diseases, mainly under the domains of Sanitation and Health. The indicators under quality (e.g., residual chlorine: percent samples with residual chlorine within the permissible limits) and hygiene (e.g., waterborne diseases) were provided by Alwathaf et al. [43], and were typically applied at city, basin, and national scales, where local health and environmental conditions are more explicitly observed. The combined indicators integrate quantitative and qualitative measures to assess water security and represent broader dimensions such as sustainability, risk and disasters, governance, water environment, capacity, and infrastructure under domains like Water Resources, Environment and Ecosystems, Water Supply, Water Management System, and Socio-Economics, providing both measurable and qualitative aspects of assessing water security. For example, the sustainability dimension includes indicators (e.g., utility performance indicator) given by [95], while the risk and disaster dimension includes indicators like (e.g., flood frequency) provided by [30,68,96], and water safety indicators include measures such as (e.g., leakage rate of the pipe network), as highlighted by Xu et al. [97]. Governance has many indicators, where Jensen et al. give the indicators such as (e.g., Strategic planning, disaster management) [68]. Furthermore, the indicators for the water environment (e.g., geospatial data on river morphology) are introduced by Butte et al. [98]. Capacity (e.g., government expenditure on environmental protection) [93] and, finally, infrastructure (e.g., drainage infrastructure) [99] are also reported through specific indicators. They are applied across a wide range of spatial scales, from city to global levels, reflecting the interconnected nature of water systems. More detailed information is provided in Supplementary Table S3.

3.4. Overview of Assessment Methods in Urban Water Security Studies

The synthesis of the reviewed studies reveals that 25 articles focused on Urban Water Security assessment, which were widely applied across different locations. This assessment has been conducted using a wide range of methods and techniques, categorized into methodological groups: index-based [29,68], model-based [22,72], framework-based [94,110], spatial and geospatial [4,114], data-driven [58,109], governance and qualitative [115,116], climate and scenario-based [72,116], and risk-based [92] as provided in

Table 2. Methods used for Urban Water Security assessment.

Methodological Groups	Citation
1. Index-based
Assessment using indicators	[29]
Urban Water Security index	[16]
Water poverty index	[100]
Indicator development using process analysis methods	[68]
2. Model-based
System dynamics	[72]
WEAP model	[22]
Capital portfolio approach	[79]
Integrated system-based approach	[117]
Water neutrality index	[117]
Particle swarm projection pursuit model	[109]
3. Framework-based
DPSIR	[94]
PSR	[4]
DAPSR	[109]
OECD	[115]
UWM	[110]
City blueprint framework	[115]
Indicator-based multi-dimensional framework	[68]
4. Spatial and geospatial
GIS-based	[114]
Theil index	[30]
InVEST	[4]
5. Data driven
Regression	[58]
Clustering	[58]
Factor analysis	[84]
6. Governance and qualitative
Delphi method	[58]
Q-method	[84]
Thematic analysis	[77]
Grounded theory	[77]
Stakeholder-based assessment	[115]
Mixed methods (surveys, FGDs, and interviews)	[116]
7. Climate and scenario-based
Climate modeling (RCP Scenarios)	[116]
Scenario analysis	[72]
Dynamic adaptive policy pathways	[72]
Gray prediction model	[94]
8. Risk-based
Risk matrix analysis	[92]
SWARA	[92]

Notes: WEAP: Water Evaluation and Planning model, DPSIR: Driving Force Pressure State Impact Response, PSR: Pressure State Response, DAPSR: Driving Force Agriculture Non-Point Pollution Pressure State Response, OECD: Organization for Co-operation and Development, UWM: Urban Water Metabolism, InVEST: Integrated Valuation of Ecosystem Services and Tradeoffs, SWARA: Stepwise Weight Assessment Ratio Analysis.

.

• Index-based: Explicit aggregation into a composite score (weighted indicators to a single metric of water security.
• Model-based: Uses process-based or simulation models (e.g., hydrological or system models.
• Framework-based: Primarily conceptual or structural (e.g., DPSIR-type approaches organizing variables without necessarily aggregating them.
• Data-driven: Relies on statistical or machine learning from the datasets without explicit system representation.
• Spatial and geospatial: This category includes methods where spatial representation and geographic variability are main to the analysis, regardless of statistical or modeling technique.
• Governance and qualitative: These methods prioritize stakeholder knowledge, institutional dynamics, and qualitative interpretation over quantitative modeling.
• Climate and scenario-based: This category represents the prediction aspect of climate variables, using scenarios to evaluate the future states of water security.
• Risk-based: This category signifies the evaluation and prioritization of hazards, vulnerabilities, and the impacts to support decision-making under uncertainty.

We acknowledge that several studies span multiple methodological categories (e.g., index-based approaches embedded within frameworks based, or model-data-driven methods) [22,90]. In such cases, studies were classified based on their dominant methodological emphasis.

Figure 7 shows a three-level hierarchical structure of the Urban Water Security assessment. The first hierarchy represents the major categories of approaches, including index-based, model-based, governance and qualitative, framework-based, data-driven, climate and scenario-based, spatial and geospatial, and risk-based approaches. The second hierarchy includes the specific methods under each category. For index-based groups, include the assessment using indicators, the water poverty index, Urban Water Security index, and indicator development using process analysis methods. For model-based approaches, encompass System Dynamics (SD), the Water Evaluation and Planning (WEAP) model, the Capital Portfolio Approach (CPA), integrated system-based approaches, and the water neutrality index. Similarly, framework-based techniques comprise Driving Force-Pressure State Impact Response (DPSIR), Pressure State Response (PSR), Driving Force Agricultural Non-Point Source Pollution Pressure State Response (DAPSR), Organization for Co-operation and Development (OECD), Urban Water Metabolism (UWM), the City blueprint framework, and an indicator-based multi-dimensional framework. Spatial and geospatial covers Geographical Information System (GIS)-based, the Theil index, index-based, and Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) models. Furthermore, the data-driven group has regression, clustering, factor analysis, and particle swarm projection pursuit model. Governance and qualitative approaches include Delphi, Q-method, thematic analysis, grounded theory, and stakeholder-based assessments. The climate and scenario-based group consists of climate modeling (RCP scenarios), scenario analysis, dynamic adaptive policy pathways, and the gray prediction model. Risk-based approaches contain risk matrix analysis and Stepwise Weight Assessment Ratio Analysis (SWARA). The third level represents specific applications and implementations of methods in individual studies, while spatial scales (e.g., cities and regional) are considered as a main aspect rather than a hierarchical level. More details are provided in Supplementary Table S4.

3.5. Artificial Intelligence Techniques and Their Applications in Water Security

Although only 25 studies addressed artificial intelligence applications, the available evidence highlights the growing importance of artificial intelligence in water security. The details of specific artificial intelligence techniques and their applications in achieving water security are provided in

Table 3. Types of artificial intelligence (ML/DL), corresponding tools/models, and applications in water security.

Sl No.	Type of AI (ML/DL)	Tools/Models	Applications	Citations
1.	ML	FL, KBS, GA, AA, and ES	Applied in measurement platform, proactively water monitor, water access, water supply, water delivery, water complaints, risk prediction, preventive maintenance, decision making, time series prediction, real time monitoring, natural calamities, water management prediction plannings, scheduling, design analysis, dynamic modeling, quality prediction, quality inspection, signal identification, system diagnosis, fault detection, failure analysis, maintenance analysis, pattern recognition, image identification, anomaly detection, management, target tracking, guidance system, routing system, value prediction, telecommunication	[33]
	DL	ANN
2.	ML	ELM	Applied to enhance the hydrological response of SWAT	[118]
3.	DL	ANN	Applied to evaluate the adsorption process of polluted water bodies, removal of heavy metal pollutants from mining-affected water bodies, applied in chemical precipitation to aid in the removal of arsenic pollutants, to monitor the growth of plants in different contaminated soils and groundwater, estimation of membrane performance characteristics, used to build and train prediction models for water treatment systems, and used to discover materials for ion-selective membranes.	[133]
	ML	ANFIS, SVM, and MLR
4.	ML	TPFML, HC, and SOM	Applied to find polluted areas and to select the best water security strategies	[119]
5.	DL	ANN	Applied to analyze water distribution networks, identifies the hidden patterns in large water datasets, and applies to enhance the efficiency of water distribution, supports network analysis and management	[32]
	ML	SVM, CT, and ANFIS
6.	DL	ANN	Applied for precipitation from remotely sensed information	[101]
7.	ML	K-NN	Applied to fill in the missing data through spatial interpolation for Urban Water Security maps	[30]
8.	DL	LSTM, GRU, and TCN	Prediction of lake water levels, forecasting lake water levels for 1-day, 3-day, and 7-day periods, time series forecasting of lake water levels	[120]
9.	ML	RF, SVM, and GBDT	Predicting aquifer vulnerability, predicting groundwater, and nitrate contamination prediction	[121]
	DL	ANN
10.	ML	XG Boost	Predicting and modeling water quality indicators such as ammonia nitrogen, total phosphorus, and Chemical Oxygen Demand at the grid scale	[122]
11.	ML	RF and SVM	Assessment and prediction of the water–energy–food nexus security	[123]
12.	DL	LSTM, GRU, and RNN	Forecasting future water consumption, food production, and electricity production, time series modeling of water, energy, and food resources production and demand	[124]
13.	ML	DTR, RF, and GBR	Used to predict nutrient concentrations and nutrient flux relationships based on catchment characteristics and climate variables, modeling complex non-linear relationships between land use, climate, and nutrient concentrations, to improve nutrient prediction accuracy	[125]
	DL	MLP
14.	ML	RF and SVM	Used to estimate riverine nutrient concentrations such as total phosphorus, total nitrogen, and ammonia nitrogen, and also used for uncertainty analysis	[126]
	DL	BPNN
15.	ML	XG Boost and GBDT	Predicting time-to-next pipe break in urban water distribution systems, modeling relationships between pipe characteristics and failure timing	[127]
16.	ML	RFR	Predicting optimal pipe diameters and reducing high-velocity zones in water distribution networks	[128]
Studies with general discussions on ML/DL applications with no specific tools/models mentioned
17.	ML		Applied for spotting anomalies in water flow, identifying malfunctioning of water meters, wastage, or pilferage during water transmission, analyzing water data, predicting water flows, checking water leaks, and estimating current pipe corrosion and deterioration	[34]
18.	DL/ML		Applied in groundwater management by enabling predictive modeling, more accurate assessment of aquifer conditions, predicting how groundwater levels will change with rainfall, water extraction rates, and land use, identifying areas at high risk of groundwater depletion, optimizing water extraction schedules, reducing over-pumping, determining the best times and amount of watering crops, minimizing the waste while maintaining the productivity, applied to simulate the impacts of different water management strategies, continuously adjust the irrigation schedule.	[130]
19.	ML		Applied to control for aquifer properties to get more accurate measures of groundwater changes	[129]
20.	ML		Used in efficient cloud forecasting and targeting efficiency in rainfall enhancement	[64]
21.	ML/DL		Applied to monitoring and survey, applied in research, predicting and forecasting, early warning, applied in consulting, planning, and decision making	[86]
22.	ML/DL		Applied to enhance water system monitoring at multiple spatiotemporal scales	[81]
23.	ML/DL		Applied to analyze diverse data sources encompassing hydrological, meteorological, environmental, and socio-economic variables, to uncover the hidden patterns, to enhance predictability and reliability	[94]
24.	DL/ML		Applied to simulate different aspects of the real-world rainfall runoff process	[131]
25.	ML		Applied to remove the buildings and forests from the Copernicus DEM to improve basin delineation accuracy	[132]

Notes: ML: Machine Learning, DL: Deep Learning, ANN: Artificial Neural Network, FL: Fuzzy Logic, KBS: Knowledge-based Systems, GA: Genetic Algorithms, AA: Adaptive Agents, ELM: Extreme Learning Machines, ANFIS: Adaptive Neuro-Fuzzy Inference System, MLR: Multiple Linear Regression, SVM: Support Vector Machine, TPFML: Two-Phase Fuzzy Machine Learning-based framework, HC: Hierarchical Clustering, SOM: Self-Organizing Map, CT: Classification Trees, K-NN: K-nearest Neighbors, LSTM: Long Short-Term Memory, GRU: Grated Recurrent Unit, TCN: Temporal Convolution Network, GBDT: Gradient Boosting Tree, XG Boost: Extreme Gradient Boosting, RNN: Recurrent Neural Network, DTR: Decision Tree Regressor, MLP: Multi-layer perception, GBR: Gradient Boosting Regressor, BPNN: Back Propagation Neural Network, RFR: Random Forest Regressor.

. It was identified that machine learning and deep learning have been applied across different water sectors to enhance water security. However, in some studies, information on specific models or tools is not provided. Batisha et al. highlight the importance of artificial intelligence in smart water systems through deep learning tools like Artificial Neural Networks (ANN), and machine learning tools like Fuzzy Logic (FL), Knowledge-based Systems (KBS), Genetic Algorithms (GA), and Adaptive Agents (AG) for applications such as water monitoring, risk predictions, and scheduling [33]. Budamala and Mahindrakar present a hybrid framework that integrates the machine learning model Extreme Learning Machines (ELM) to enhance the hydrological response of the SWAT model, thereby improving prediction accuracy and supporting effective assessment of freshwater security [118]. The study by Nti et al. highlights that deep learning models, such as ANN and machine learning models like the Adaptive Neuro-Fuzzy Inference System (ANFIS), Multiple Linear Regression (MLR) and Support Vector Machine (SVM), are widely used to remove heavy metal pollutants from water bodies, monitor the growth of plants in different contaminated groundwater, and optimize water treatment technologies, enhancing water security. Furthermore, Zamri et al. used machine learning techniques, including a Two-Phase Fuzzy Machine Learning-based framework (TPFML), Hierarchical Clustering (HC), a Self-Organizing Map (SOM), and Autoencoder models, to identify polluted areas and select the best water security strategies [119]. Regardless, deep learning models such as ANN and machine learning models such as SVM, Classification Trees (CT), and the Adaptive Neuro Fuzzy Inference System (ANFIS) are applied in water distribution networks to identify hidden patterns in large water datasets, and support network analysis and management as described in [32]. Besides this, Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR), an ANN-based precipitation dataset applied to remotely sensed precipitation, is mentioned by Dey et al. as supporting water security assessment [101]. Marcal et al. present a machine learning tool, K-nearest Neighbors (K-NN), for filling missing data through spatial interpolation for Urban Water Security [30]. Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Temporal Convolution Network (TCN) are applied for short-term lake water level prediction and time series forecasting [120]. Similarly, Random Forest (RF), SVM, Gradient Boosting Tree (GBDT), and ANN are used to predict aquifer vulnerability, groundwater conditions, and nitrate contamination [121]. Furthermore, XG Boost is used to predict and model water quality indicators, such as ammonia nitrogen, total phosphorus, and Chemical Oxygen Demand, at the grid scale [122]. RF and SVM are used to assess and predict water–energy–food nexus security [123]. Moreover, LSTM, GRU, and Recurrent Neural Network (RNN) models are utilized to forecast water consumption, food production, and electricity generation, and to model the temporal dynamics of water–energy–food resource systems [124]. Decision Tree Regressor (DTR), RF, and Gradient Boosting Regressor (GBR) are utilized to model and predict nutrient concentrations and fluxes by capturing complex non-linear relationships among catchment characteristics, land use, and climate variables, thereby improving prediction accuracy [125]. While RF, SVM, and Back Propagation Neural Network (BPNN) are utilized to estimate riverine nutrient concentrations, such as total phosphorus, total nitrogen, and ammonia nitrogen, they also support uncertainty analysis [126]. XG Boost and GBDT are used to predict the time to the next pipe break in urban water distribution systems by modeling the relationships between pipe characteristics and failure timing [127]. Random Forest Regressors (RFR) are applied to optimize pipe diameters and minimize high-velocity zones in water distribution networks, thereby improving system efficiency and performance [128]. Iyer and Venkatagiri indicate that machine learning tools are used to detect anomalies in water flow, identify malfunctioning water meters, and predict water flows [34]. Mehta et al. [129] present a machine-learning-based calibration model to control aquifer properties to obtain more accurate measures of groundwater changes. Meanwhile, Sigh et al. have been focused on groundwater management by applying artificial intelligence techniques to predict groundwater levels in response to changes in rainfall, water extraction rates, and land use, identifying areas of high risk of groundwater depletion, continuously adjusting the irrigation schedules, and assisting farmers to determine the best times and amount of watering the crops [130]. To develop a more efficient cloud forecasting and targeting system for rainfall enhancement, Wehbe et al. apply machine learning techniques to support water security strategies under climate change [64]. The study by Xia et al. demonstrates that an AI-integrated modeling system (e.g., the Yangtze simulator enables monitoring and survey, prediction and forecasting, and early warning, decision making for eco water security at the basin scale [86]. Later, another study by Xia et al. demonstrated the use of AI to enhance water monitoring across multiple spatial scales [81]. Kacikoc et al. mentioned simulating different aspects of real-world rainfall runoff processes using an AI-based black box [131]. Machine learning techniques are used to remove buildings and forests from the Copernicus Digital Elevation Model (DEM) to improve basin delineation accuracy, thereby aiding data processing for water security assessment [132]. Machine learning models are applied to analyze diverse data sources from hydrological, meteorological, environmental, and socio-economic variables, identifying hidden patterns and enhancing predictability and reliability for water security assessment, as given by Li et al. [94].

Figure 8 illustrates the application of artificial intelligence in water security, synthesizing evidence from 25 selected studies. The identified applications are grouped into six thematic categories, with the number of articles and the corresponding articles for each application in the studies. The first theme, representing general applications, provides operational functions. These include telecommunications, waste minimization, estimation of pipe corrosion and infrastructure deterioration, preventive maintenance, and support for research activities. In addition, artificial intelligence is applied to early warning systems, dynamic modeling, and system diagnosis, enabling proactive data-driven decision-making. Further applications, such as target tracking, design analysis, and overall water management, highlight the role of artificial intelligence in optimizing water infrastructure and services. This theme also shows AI-driven improvements in service delivery, including water access, water delivery, water complaints, and measurement platforms, demonstrating its contribution to enhancing efficiency and reliability in water security systems. The second theme focuses on monitoring and detection. This theme includes monitoring of water systems across multiple spatiotemporal scales, enabling a better understanding of system dynamics. Artificial intelligence techniques are also applied to monitor plant growth in contaminated groundwater environments and to support image-based identification processes. Furthermore, these approaches facilitate the detection of water wastage and pilferage during transmission, as well as the identification of regions at high risk of groundwater depletion. In addition, they can detect malfunctioning of water meters, identify anomalies, and locate polluted areas. Functions such as maintenance analysis, failure analysis, and fault detection further enhance a system’s efficiency. Techniques like signal identification, continuous monitoring, and surveys contribute to improved data interpretation. Moreover, proactive and real-time monitoring capabilities enable timely interventions, thereby supporting effective water resource management and enhancing overall water security. The third theme focuses on water system management and decision-making. This theme includes applications such as the analysis of water distribution networks, as well as consulting and planning activities that support stakeholders’ decision-making. Artificial intelligence is also used to determine crop water requirements and irrigation scheduling strategies to improve agricultural water efficiency. In addition, routing and guidance systems enable efficient distribution of water resources. Applications such as optimizing water extraction schedules and controlling aquifer properties further show the importance of artificial intelligence in groundwater management. Moreover, these approaches support strategy selection and the assessment of the potential consequences. The fourth theme represents water quality and treatment. This theme includes quality prediction and inspection, enabling continuous evaluation of water quality and safety. It also applies to model and optimize adsorption processes in polluted water bodies and to enhance the removal of heavy metal contaminants. Moreover, artificial intelligence is employed in membrane modeling and the estimation of membrane performance characteristics, facilitating more efficient filtration and treatment processes. It also applies in chemical precipitation methods and the discovery of advanced materials for ion-selective membranes. The fifth theme focuses on prediction and forecasting, where artificial intelligence is used to estimate future water conditions. This includes time-series and risk prediction, which helps us to understand potential changes in water systems. It is also applied in planning for natural disasters such as floods and droughts. In addition, artificial intelligence techniques are used to estimate precipitation from remote sensing data and to simulate rainfall–runoff processes. It is also used to predict groundwater levels, water flows, and values, which are important for water resource planning. Other applications, such as leak detection, cloud forecasting, future water consumption, prediction of pipe break and optimal pipe diameter prediction of nitrate contamination, prediction of ammonia-nitrogen, total phosphorous, and COD, further support accurate and timely predictions. The sixth theme focuses on data analysis, where artificial intelligence helps analyze large amounts of water-related data and perform pattern recognition. These methods are useful for uncovering hidden patterns that are not readily apparent in conventional approaches. Artificial intelligence is also applied to spatial analysis tasks such as interpolation and basin delineation, thereby improving understanding of how water resources vary across regions. In addition, these techniques assist in analyzing water data to develop optimal water management.

4. Discussion

4.1. Challenges in Defining Water Security

Defining water security remains very challenging due to its multidimensional nature, encompassing many interrelated components such as access, quality, quantity, governance, and sustainability. For example, the availability of widely used definitions, such as those from UN-Water [10] and GWP [9], emphasizes safe, sufficient, and sustainable access; other studies provide component access in various aspects, using terms such as equitable, affordable access [16,29,63], showing that even the components are defined in multiple ways. In addition to this, some studies focus on water availability and quality [12,71], whereas so other emphasizing on governance [13]. The definitions also vary across scales, from households to global levels. Moreover, recent studies incorporate climate change water-related risks, making the concept of water security more complex [28,47,89,134,135].

The findings show a lack of evidence for describing water security in simple, broader terms. The simple definition developed in this study is a dynamic and multidimensional concept of water security [4,15]. While broader definitions incorporate multiple components, they are often described using key terms that are crucial to understanding water security. Synthesizing these key terms for each component from multiple definitions provides a better understanding of water security in a broader context (Figure 6). By bringing these aspects together, this study gives a clearer picture of water security.

4.2. Analysis of Indicators, Methods, and Artificial Intelligence in Water Security

The classification of water security indicators in this study highlights a comprehensive approach compared to many existing studies. While previous studies often provide only a limited set of indicators or dimensions, typically providing dimensions such as water availability, quality, and demand [2,71]. The present study expands this perspective by classifying it into quantitative, qualitative, and combined categories across multiple dimensions, domains, and scales (Table 1). The previous studies have also incorporated social, economic, and governance indicators [18,46], but these are often provided separately rather than integrated. Moreover, green water scarcity and land-use changes strongly affect water sustainability, which is evaluated by using water security indicators such as blue water scarcity and blue water vulnerability [136,137,138,139]. Whereas this study brings together a wider range of dimensions, including water economy, social, environmental, governance, and infrastructure, having multiple indicators under it provides a deeper understanding of water security. Furthermore, while many studies apply the indicators at specific spatial scales, such as urban or basin scales [68]. This study identifies and organizes indicators across multiple scales, from city to global.

The methodological grouping of Urban Water Security assessment methods in this study provides clarity for urban-scale water security assessment compared to the existing literature. In most of the studies, methods such as index-based approaches and model-based approaches are usually provided in separate ways [22,100]. This study brings these approaches together and groups them into eight methodological groups, including index-based, model-based, framework-based, spatial and geospatial, data-driven, governance and qualitative, climate and scenario-based, and risk-based (Table 2). Previous studies have highlighted individual approaches, such as governance methods for the assessment [13] or using spatial tools such as a Geographical Information System [90] and an integrated mapping and analysis system [140]. Moreover, evaluating how the City Blueprint Approach incorporates hydro-social and politico-ecological aspects in assessing Urban Water Security [141]. This study differs by combining all these approaches used in Urban Water Security assessment.

The review of artificial intelligence applications in water security shows that, although 25 studies focus on applications and techniques, rather than providing a broader approach. For example, studies such as [33] emphasize prediction and modeling, while [34] focus on operational applications, such as anomaly detection and groundwater monitoring. Similarly, refs. [130,133] provide applications in water quality treatment and agricultural management, while studies such as [64,81,86] focus on climate-related prediction, and basin-scale analysis. Although these studies demonstrate the usefulness of artificial intelligence, they are mostly limited to processes and specified techniques used to evaluate water security (Table 3). However, these studies are mostly presented separately, with each focusing on a single function such as prediction, monitoring, or management. This study brings these applications together and organizes them into six main themes: general, monitoring and detection, water system management and decision-making, water quality and treatment, prediction and forecasting, and data analysis (Figure 8). This makes it easier to understand how different artificial Intelligence applications are connected within water security. The integration of artificial intelligence with other approaches can help with water security assessment and management. For example, combining AI with hydrological models helps predict the availability and accessibility of water resources and risk management [120,142]. In addition, integrating AI with governance frameworks enables the inclusion of stakeholder and socio-economic perspectives in decision-making [143,144,145]. Climate-based approaches further help in understanding future risks and uncertainties [64,81,86,146]. Together, these integrated approaches enhance decision-making, improve resource management, and support sustainable and resilient water systems. In addition, while previous studies mention specific techniques such as ANN, SVM, and clustering methods, many studies focused mainly on technical model performance without explicitly linking these techniques to water security objectives such as sustainability, accessibility, risk reduction, governance, and decision-making. This study differs by linking these techniques to their practical applications and grouping them by role.

4.3. Limitations and Challenges

This study followed a PRISMA framework to screen the collected articles, include suitable studies, and remove those that were not relevant to the study objectives. A keyword-based search was used to collect relevant studies; however, some studies may have been missing if they did not use the exact keywords. The literature screening, article selection, data extraction, and eligibility assessment were conducted by a single reviewer, which may introduce bias into the selection process. Another limitation is that there are still only 25 studies on artificial intelligence in water security using the keywords used in this study, and it is difficult to fully understand its potential. Future studies are encouraged to use broader, multiple search terms to identify more relevant literature in this area. Artificial intelligence applications span a wide range, so organizing them into clear categories requires careful interpretation of the available information. Although there are some limitations, the results provide valuable insights that can support future research aimed at strengthening water security.

5. Conclusions

Water security is becoming increasingly crucial due to challenges such as population growth, climate change, urbanization, and overexploitation of water resources. This study systematically reviewed 146 articles on water security, including definitions, indicators, assessment methods, and the use of artificial intelligence. The results show that there is no single common definition, as different studies describe water security in different ways. To address this, the study brings these ideas together and offers a simple, broader definition of water security. The study also shows that water security indicators vary widely across dimensions and scales. By categorizing them into quantitative, qualitative, and combined types, this work makes it easier to understand how water security is measured. In addition, classifying assessment methods into eight groups helps organize the diverse approaches used in 25 Urban Water Security studies, which are often scattered across the literature. Another important finding is the application of artificial intelligence in water security in 25 articles, which provides valuable insights into its growing role in water security through improved monitoring, prediction, and management. Grouping these applications into six themes provides a clearer picture of how artificial intelligence supports water security. Techniques such as deep learning and machine learning identified in the studies will help improve decision-making, make water systems more efficient, and better manage water resources. Collectively, this study brings together different aspects of water security, making it easier to understand and apply. It can help researchers, policymakers, and stakeholders to make better decisions and improve water management. Strengthening the use of artificial intelligence along with integrated approaches will play an important role in ensuring sustainable water use and addressing future water-related challenges.