Editorial: Achieving SDG 6: Remote Sensing Applications in Sustainable Water Management

Mhd. Suhyb Salama^*

• University of Twente, Enschede, Netherlands

This collection in Frontiers in Remote Sensing highlights innovative applications of remote sensing and Earth Observation (EO) that support the realization of SDG 6. We present four articles that span a diverse spectrum of technologies, methodologies, and geographic contexts, collectively showcasing the growing maturity and impact of remote sensing in sustainable water management. While the broader discourse around SDG 6 has often emphasized the quantitative aspects of water (such as availability and access) these contributions notably centre on the quality dimension of SDG 6. This focus is both timely and essential, as water quality has historically received comparatively less attention despite its critical role in achieving sustainable water outcomes. The first contribution, by Saranathan et al., (2024) introduce probabilistic neural networks to retrieve water quality indicators with associated uncertainty estimates, enabling confidence-based decision-making in data-scarce environments. Whereas, Balasubramanian et al., (2025) leverage machine learning and multi-mission satellite archives to reconstruct long-term water quality trends, facilitating retrospective assessments, essential for monitoring policy effectiveness and environmental change. On the other hand, Atton Beckmann et al., (2025) demonstrate the power of highresolution satellite imagery to monitor algal blooms in small, under-monitored inland lakes, significantly expanding the scope of water bodies that can be routinely assessed from space. Finally, Wilson et al., (2025) address systemic barriers to EO adoption, offering inclusive, action-oriented pathways to empower data-poor regions, particularly in the Global South.Together, these contributions underscore the critical role that remote sensing technologies and EO play in advancing the objectives of SDG 6. These technologies have matured to the point where they are actively enhancing water quality monitoring, expanding spatial and temporal data coverage, informing evidence-based policy, and enabling the early detection of ecological threats. Yet, realizing the full potential of EO requires urgent attention to persistent challenges: the difficulty of integrating EO products into national regulatory and decision-making frameworks; the limited availability of robust, transferable algorithms for optically complex water bodies; the current inability of EO to detect emerging pollutants or non-optically active water quality constituents; and the lack of integration with predictive water quality models and complementary sensing technologies, particularly in the Global South. Addressing these challenges will require sustained investment in EO infrastructure, open-data platforms, algorithm innovation, capacity-building, and interdisciplinary collaboration.In the sections that follow, we first define the scope of the articles included in this collection and clarify the central focus of this editorial essay. We then outline the role of remote sensing and EO in advancing SDG 6, with particular attention to the specific targets where these technologies offer the greatest impact. Next, we summarize the key contributions of the featured papers, highlighting their methodological innovations and practical insights. This is followed by a discussion of emerging directions aimed at enhancing the utility, adoption, and influence of EO in driving progress toward SDG 6. We conclude with reflections on the broader implications of these findings and offer recommendations for future research, policy, and implementation. The articles in this collection primarily focus on the remote sensing of water quality, reflecting a key dimension of ensuring clean water for all. While water quality is a central concern, it is important to recognize SDG 6 encompasses a broader and more integrated vision. The goal aims to ensure availability and sustainable management of water and sanitation for all by 2030, covering a wide range of targets related to water governance and resource management.Beyond improving water quality, SDG 6 addresses critical objectives such as equitable access to safe and affordable drinking water, universal access to sanitation and hygiene, and increased water-use efficiency across sectors. It also calls for the sustainable management of freshwater resources and ecosystems, the safe treatment and reuse of wastewater, and the creation of enabling environments through policies and institutional frameworks. Moreover, SDG 6 emphasizes the importance of international cooperation and capacity-building, particularly for developing countries, to strengthen water and sanitation-related programs and activities.These broader issues will be briefly addressed in the following section, "Setting the Scene" to provide context on how remote sensing aligns with the full scope of SDG 6. However, the main body of this editorial will primarily focus on water quality, reflecting the central theme of the contributions in this collection. The final section on Future Directions will explore opportunities to expand and deepen the role of EO technologies in advancing the remote sensing of water quality as a key component of SDG 6 implementation.Throughout this paper, the terms "remote sensing technologies" and "Earth Observation (EO)" are used interchangeably to reflect the integrated nature of EO systems, which encompass satellite, aerial, UAV-based, and in-situ sensing platforms. As regulatory agencies, national governments, and international organizations work to implement SDG 6, remote sensing and EO are playing an increasingly important role in supporting both policy development and regulatory compliance. In this context, EO helps close critical data gaps that hinder effective water governance. Its capabilities directly support the assessment of progress toward environmental standards and several SDG 6 targets, including: Changes in vegetation cover, wetland dynamics, primary production, sedimentation, or algal blooms can be detected through EO-data analysis, enabling rapid intervention to protect and restore critical ecosystems under threat from anthropogenic pressures or natural hazards. Remote sensing technologies, from satellites to Unmanned Aerial Vehicles (UAVs), have emerged as powerful tools for tracking water quality indicators across large spatial and temporal scales. Optical sensors aboard Earth-observing satellites can detect nearsurface concentrations of chlorophyll-a, coloured dissolved organic matter, total suspended solids, and other optically active substances, enabling regional to global assessments. This capability is increasingly critical as climate change and human activities intensify pressures on inland lakes and urban water systems. The integration of multi-mission satellite data streams has significantly improved the temporal resolution of observations, supporting more timely and actionable water management decisions.Over the past decade, remote sensing applications for water quality management have advanced considerably. Satellite-based EO, when combined with in-situ measurements, proximal sensing, and machine learning techniques, now enables monitoring from global scales down to small, individual water bodies.Key contributions to this collection include:- Saranathan et al., (2024) demonstrate that Mixture Density Networks and Bayesian Neural Networks with Monte Carlo Dropout can generate pixel-level probability distributions for key indicators such as chlorophyll-a. By quantifying uncertainty alongside the estimates, these models enable more informed decision-making and risk assessment. The broader topic of uncertainty quantification for EO-derived water quality variables has been the subject of extensive research (Hammond et al., 2020;Mélin, 2021Mélin, , 2019Mélin, , 2010;;Zhang et al., 2022). However, the use of neural networks models to estimate heteroscedastic and epistemic uncertainties at a pixel-level is relatively new. This development represents an important advancement in generating EO products that provide not only quantitative estimates of water quality indicators but also associated measures of uncertainty, enhancing their reliability for operational monitoring and decision-making.While machine learning models are rapidly advancing, their opaque internal structures remain a challenge for transparency and interpretability. The proposed uncertainty-aware approach by Saranathan et al., (2024) represents a step forward toward developing more explainable and biophysically-informed AI algorithms for retrieving water quality variables (Jiang et al., 2020;Roy et al., 2023).- Balasubramanian et al., (2025) reconstruct long-term time series of water quality indicators using historical EO datasets. The study harmonizes data from multiple satellite missions, including the Moderate Resolution Imaging Spectroradiometer (MODIS), Medium Resolution Imaging Spectrometer (MERIS), and Visible Infrared Imaging Radiometer Suite (VIIRS). It also applies machine learning models to the 40+ year Landsat archive. This integration enables the generation of multi-decadal records of inland and coastal water quality indicators, such as water clarity, algal blooms, and nutrient status. Crucially, the researchers went beyond algorithm development to operationalize EO-derived indicators across archived datasets, demonstrating their readiness for large-scale monitoring and policy support (Chen et al., 2022;Wilkinson et al., 2024).Although such reconstruction approaches are essential for assessing the longterm effectiveness of environmental policies (such as nutrient reduction strategies) they still require further development to effectively integrate with current and forthcoming observations from a wide range of EO missions.-Atton Beckmann et al., (2025) demonstrate the feasibility of quantifying phytoplankton in small lakes using high-resolution imagery from Planet SuperDoves and Sentinel-2. This capability is particularly valuable for mapping water quality in poorly monitored environments, such as small ponds and streams. Early demonstrations (e.g., Kwon et al., 2023;Mishra et al., 2020) established the feasibility of deriving water quality variables from meter-level satellite imagery. Building on this foundation, the study by Atton Beckmann et al., ( 2025) demonstrate that such imagery can accurately detect localized phenomena, such as algal scums and sediment plumes, with accuracy comparable to in-situ observations. As these high-resolution data sources become more routinely available, we can anticipate a significant expansion in the number of water bodies monitored globally, along with more granular integration of water quality data into urban management systems, e.g., smart city dashboards for water (Chen et al., 2023;Okoli and Kabaso, 2024).While high-resolution data hold great potential for mapping previously unmonitored water bodies, their application is often challenged by the need to correct for variable scene illumination, surface reflectance, sun glint and geolocation. These issues are especially pronounced when using Unmanned Aerial Vehicles (De Keukelaere et al., 2023;Windle and Silsbe, 2021).- Wilson et al., (2025) Despite progress in technical skill development, the transition toward sustained institutional capacity remains underdeveloped, limiting long-term and systemwide adoption (Pritchard et al., 2022;Thapa et al., 2019). As we enter the Fifth Industrial Revolution, characterized by the explosion of information and the integration of generative AI, new questions emerge around how to build adaptive and enduring capacity (Marino and Monaca, 2025;Ziatdinov et al., 2024;Gunderson et al., 2020;Orion, 2019). How can pedagogical approaches be designed to retain institutional knowledge while keeping pace with rapidly evolving technologies? Emerging learning tools such as virtual reality (Makransky and Petersen, 2021), digital twins (Hazeleger et al., 2024), serious games (Sajjadi et al., 2022), gamification and generative AI tools (Patra et al., 2024) offer promising avenues. Yet, these questions remain open and are the subject of ongoing inquiry as the landscape of AI-driven capacity-building continues to evolve.Across these contributions, a consistent set of technologies and data infrastructures emphasize the advancement of remote sensing for SDG 6. The Sentinel constellation (Sentinel-2 and Sentinel-3) and the Landsat series remain the backbone of optical water quality monitoring, widely applied in both regional case studies and global assessments. As remote sensing technologies and EO continue to mature and diversify, the scope of remote sensing applications for sustainable water management is rapidly expanding. Looking ahead, six strategic priorities emerge that can further strengthen the role of remote sensing in improving water quality targets of SDG 6. With the rapid depletion of groundwater resources and increasing hydrological variability driven by climate change, small surface water bodies are becoming vital sources of drinking water, particularly for rural, peri-urban, and marginalized communities (Jasechko et al., 2024). These lakes, ponds, and reservoirs, often less than 1 km² in area, remain among the least monitored yet most vulnerable to contamination and seasonal fluctuations. Historically overlooked by traditional water management systems due to their small spatial footprint and remoteness, their importance in local water supply is steadily growing.The use of sub-meter high-resolution EO platforms (such as Pléiades, WorldView, and UAVs) has the potential to substantially improve monitoring of small water surfaces by providing data at scales relevant for local water management and policy implementation. Netherlands, where floating solar panels have been installed. This pond, currently being explored for its potential to complement the local drinking water supply while also serving as a site for green energy production, exemplifies the integration of renewable infrastructure with small water bodies. This approach reflects the growing importance of multi-functional water resource management, particularly in regions facing increasing demand and water stress. High-resolution EO data, such as the image presented in Figure ( 1), enable detailed assessments of how floating solar installations influence water quality like photosynthetically available radiation and light attenuation coefficients, which in turn affect lake stratification and lake-atmosphere heat exchange (Heiskanen et al., 2015). Meanwhile, high-resolution thermal EO capabilities are being actively developed under missions such as the ESA Land Surface Temperature Monitoring (LSTM, Koetz et al., 2019) and NASA's Surface Biology and Geology (SBG, Stavros et al., 2023). These missions aim to provide thermal data at spatial resolutions fine enough (30-100 m) to detect water surface temperature anomalies. In parallel, commercial satellite constellations are emerging with even finer Thermal InfraRed capabilities. For instance, HotSat1 by SatVu captures detailed heat variations across the Earth's surface with a resolution of up to 3.5 meters, both day and night. These systems offer potential for fine-scale assessment of surface temperature gradients, particularly in the context of floating infrastructure and climate-sensitive inland waters. For example Prandini et al., (2025) showed, using detailed in-situ measurements, that floating solar modules created a microclimate above them with temperatures approximately 12% higher than those recorded by weather stations in the surrounding area. When coupled with biophysical lake models, EO data can support the quantification of energy fluxes (e.g., latent heat) and help infer biogeochemical processes like CO₂ exchange linked to algal primary production (MacIntyre et al., 2010).To fully unlock the potential of high-resolution, future research should prioritize the development of robust EO algorithms and integrating them with biophysical and ecological models specifically designed for small, optically complex water bodies.Incorporating these water bodies into national SDG indicator frameworks could help guide targeted investments and strengthen resilience in water-scarce regions, ensuring that these often-overlooked resources are effectively monitored, managed, and protected. One of the most pressing yet understudied frontiers in remote sensing is the detection of emerging contaminants in aquatic systems. These include microbial pathogens (e.g., Escherichia coli, Vibrio spp.), heavy metals, dissolved nutrients, pharmaceutical residues, endocrine-disrupting compounds (EDCs), personal care products (PPCPs), per-and polyfluoroalkyl substances (PFAS), and microplastics. Detecting these pollutants using satellite or airborne platforms remains challenging due to their typically low concentrations and the absence of distinct absorption or reflectance features detectable by EO sensors. Consequently, they are largely invisible to traditional multispectral sensors. These contaminants are often linked to land-based pollution sources such as household wastewater, agricultural runoff, and urban stormwater, posing significant ecological and human health risks, especially in regions with inadequate sanitation infrastructure. However, recent advancements in remote sensing technologies, particularly hyperspectral imaging, offer new opportunities. With its capacity to capture continuous, high-resolution spectral signatures across narrow wavelength intervals, hyperspectral remote sensing holds promise for detecting subtle spectral patterns that could be associated with these pollutants.For microplastics, hyperspectral imaging has already demonstrated success in both laboratory and environmental settings, enabling pixel-level identification of polymer types based on their distinct spectral features (Balsi et al., 2025;de Fockert et al., 2024;Gebejes et al., 2024;Faltynkova et al., 2021). Detection may occur directly (when high concentrations or surface accumulations are present) or indirectly through proxies such as fluorescence, water colour, or algal community shifts. Similarly, while direct detection of PFAS remains nascent (Tshangana et al., 2025), optical sensing technologies such as Raman spectroscopy and microcavity sensors have been shown to discriminate PFAS compounds based on their vibrational spectral characteristics (Chen et al., 2024;Rodriguez et al., 2020), suggesting a foundation for future hyperspectral adaptation.Beyond plastic and chemical pollutants, microbial contamination can also be inferred through remote sensing. For instance, recent studies have used Sentinel-2 data in combination with environmental modelling to predict Salmonella and Vibrio presence in river systems based on variables such as turbidity, temperature, and chlorophyll-a (Palharini et al., 2025). While these approaches do not detect microbes directly, they represent viable early-warning systems based on proxies.Heavy metals and dissolved nutrients have shown more direct optical detectability in some cases. Red and Near Infrared reflectance and hyperspectral indices, when integrated with machine learning models, have been used to estimate concentrations of metals such as copper, lead, and cadmium in inland waters (Xu et al., 2024). Similarly, advances in fluorescence spectroscopy and high-resolution EO data are enabling assessments of nutrient concentrations, dissolved organic carbon, and related water quality indicators (Ngamile et al., 2025;Kumar et al., 2024). Despite this progress, detection of pharmaceutical residues, PPCPs, and EDCs remains in the early stages of remote sensing applicability. Current research relies largely on in-situ optical and biosensing platforms to characterize the presence and dynamics of these compounds (Li et al., 2024).To move toward operational monitoring, future efforts should prioritize the integration of airborne and satellite hyperspectral missions (e.g., EnMAP, PRISMA, PACE) with in-situ spectral libraries, chemometric modelling, and machine learning approaches. Such efforts could significantly extend EO's capacity to monitor water quality beyond conventional optical parameters, bridging into the chemical and biological domains of contamination. Achieving SDG 6 requires monitoring systems that are not only accurate and scalable but also inclusive and grounded in local realities. A promising pathway lies in the integration of triple sensing approaches: combining satellite-based EO, proximal sensing, and citizen science. This fusion offers multi-scale framework to assess and manage water quality in ways that are technically robust and socially embedded.The papers in this collection have demonstrated how EO provides synoptic, repeatable, and large-scale coverage of surface water bodies, allowing for the routine monitoring of key water quality indicators. However, EO-derived water quality information is primarily restricted to the surface layer and can be significantly constrained by cloud cover (Uday et al., 2025), as well as by its inability to directly detect optically inactive contaminants (Vakili and Amanollahi, 2020).Proximal sensing technologies, including UAV-mounted sensors, in-situ IoT devices, and mobile-based field platforms, help bridge critical observational gaps by providing highfrequency, high-resolution measurements at local scales (e.g., Kuusk et al., 2024).Proximal sensing is particularly effective in detecting fine-scale spatial heterogeneity and enabling rapid responses in small or dynamic water bodies. When deployed as part of a coordinated measurement network, they offer valuable coverage in both urban and rural contexts where EO data may be limited or unavailable (e.g., the WATERHYPERNET Ruddick et al., 2024).Citizen science complements both modalities by engaging communities in data collection and validation through mobile apps and low-cost kits. Mobile phone applications such as HydroColor (Leeuw and Boss, 2018) or EyeOnWater (Novoa et al., 2015) train volunteers in collecting water colour observations.While citizen-sourced data offer promising opportunities for expanding water quality monitoring, they face critical challenges related to sensor calibration, data consistency, and scientific reliability (Pattinson et al., 2023). Variability in device types, user practices, and sampling protocols can introduce significant uncertainty. For instance, Mahama and Salama, (2024) reported notable inconsistencies in smartphone-based turbidity measurements, particularly in optically complex waters where such tools are often deployed. To address these limitations, calibration and validation against high-quality reference data-such as measurements from UAV-mounted sensors or IoT-enabled probes-are essential for correcting biases and improving usability (Bakar et al., 2025).To realize this triple-sensing model at scale, it is essential to ensure data interoperability, standardization of quality-control protocols, consistent metadata reporting, and calibration across devices. Embedding citizen-generated observations within frameworks validated by proximal and EO data can enhance data reliability, while institutional support is crucial for integrating community-sourced data alongside conventional datasets. When implemented effectively, this integrative approach has to potential of delivering comprehensive, scalable, and trustworthy water-quality insights, supporting more resilient, inclusive, and evidence-based governance in line with SDG 6. While digital twins have made substantial progress in hydrology, particularly for modelling river flows, flood dynamics, and drought forecasting (Yang et al., 2024), their application to water quality remains relatively underdeveloped (Hamzah et al., 2024). A digital twin for water quality entails a dynamic, real-time numerical representation of aquatic systems, continuously updated with observational data to simulate and predict spatial and temporal changes in key parameters such as nutrient concentrations, algal blooms, turbidity, and dissolved oxygen. To realize this vision, the integration of triple sensing frameworks with hydrodynamic, biophysical, and biogeochemical models (hereafter called process-based models) is essential. EO provides synoptic and temporally rich surface data; proximal sensors offer high-frequency, localized in situ observations; and citizen science fills critical data gaps through community-based monitoring. When assimilated into process-based models, these inputs enable threedimensional, time-evolving simulations of water quality dynamics, capable of supporting scenario analysis, forecasting, and policy evaluation.Although EO data assimilation with process-based models is a well-established approach for improving aquatic ecosystem simulations (e.g., Ciavatta et al., 2018), the assimilation of EO-derived water quality products remains relatively underutilized and is still not widespread in operational systems (Groom et al., 2019).However, developing a unified global digital twin for water quality remains impractical due to the complexity, variability, and context-specific nature of aquatic systems. Each river basin, lake catchment, or coastal zone presents distinct hydrological, ecological, and socio-environmental conditions that require regionally tailored modelling approaches and calibration strategies. For example, Cho et al., (2020) argue in their review that a major challenge in assimilating EO data into water quality models lies in the uncertainty of EO observations and the spatiotemporal mismatches between measurement scales and the model grid and time domains. Addressing these limitations requires not only methodological advances (e.g., Salama et al., 2022) but also a shift toward more integrated modelling frameworks. As digital twin frameworks evolve, future efforts should prioritize the fusion of triple sensing data with process-based aquatic models at regional and global scales. Doing so will enable a new class of predictive tools that go beyond static or near-real-time monitoring, supporting early warning systems, pollution control strategies, and adaptive water quality governance. This convergence marks a critical step toward more intelligent, anticipatory, and evidence-driven water management systems aligned with the targets of SDG 6. To ensure broader and sustained impact, remote sensing-based water quality products must move beyond research prototypes toward full operational integration. For EO-based services, reaching Technology Readiness Level 9 (TRL 9) signifies that a system is not only technically mature but has been demonstrated under real-world conditions and embedded within established monitoring workflows. This transition involves more than algorithm refinement: it requires integration into regulatory or institutional frameworks, development of automated and scalable processing pipelines, rigorous validation with in situ observations, and the establishment of user support mechanisms and training. Operational readiness also depends on sustained collaboration with stakeholders and a shift from individual capacity-building to institutional strengthening. This includes longterm organizational commitment, the establishment of technical and operational protocols, and alignment with existing standards for data quality, accessibility, interoperability, and reporting. While platforms such as Google Earth Engine, Digital Earth Africa, and Euro Data Cube have advanced the technical scalability of EO applications, broader and lasting adoption requires co-producing water quality information that is not only technically robust but also fit for use within regulatory frameworks and day-to-day decision-making. Despite significant advances in EO, a persistent challenge lies in completing the value chain. Specifically, bridging the gap between data availability and its effective integration into decision-making processes (Virapongse et al., 2020). To increase uptake, remote sensing products must be integrated into institutional workflows, policy frameworks, and regulatory tools. This involves:• Co-designing EO solutions with policymakers, water managers, and civil society to ensure relevance, usability, and alignment with real-world needs.• Translating EO outputs into actionable indicators, thresholds, or risk scores that match local water guidelines and standards.• Embedding EO-derived products into early warning systems, environmental impact assessments (EIAs), water safety plans, and SDG reporting dashboards.• Demonstrating the socio-economic value of EO by quantifying how satellitebased monitoring reduces costs, increases responsiveness, or mitigates risks, particularly in resource-constrained contexts.• Building institutional trust through capacity development, long-term partnerships, and transparent validation procedures that clarify limitations and uncertainties of EO products. In summary, Earth Observation (EO) and remote sensing technologies are powerful enablers for operationalizing SDG 6. Their integration into water governance systems enhances data availability and transparency, while supporting more adaptive, evidencebased decision-making in the face of climate change and increasing demand for clean water. However, system-wide adoption requires expanding technical capabilities to include fine-scale water bodies and emerging pollutants, strengthening the integration of multi-modal sensors with process-based models, and advancing explainable AI techniques to improve the biophysical relevance of water quality retrievals. Equally important is the evolution of institutional capacity to keep pace with rapid technological change. While emerging tools like digital twins, virtual reality, and generative AI offer promise, durable knowledge systems remain essential.Decisive next steps include aligning EO products with national SDG 6 reporting frameworks, expanding validation networks, particularly in the Global South, and securing long-term funding. These efforts must be supported by cross-sectoral alliances, open data practices, and strategic capacity building. Only then can EO move beyond its current technical role to become a foundational element of inclusive and resilient water quality management system.