FlowSOM clustering - A novel pattern recognition approach for water research: Application to a hyper-arid coastal aquifer system.

Monitoring and understanding of water resources have become essential in designing effective and sustainable management strategies to overcome the growing water quality challenges. In this context, the utilization of unsupervised learning techniques for evaluating environmental tracers has facilitated the exploration of sources and dynamics of groundwater systems through pattern recognition. However, conventional techniques may overlook spatial and temporal non-linearities present in water research data. This paper introduces the adaptation of FlowSOM, a pioneering approach that combines self-organizing maps (SOM) and minimal spanning trees (MST), with the fast-greedy network clustering algorithm to unravel intricate relationships within multivariate water quality datasets. By capturing connections within the data, this ensemble tool enhances clustering and pattern recognition. Applied to the complex water quality context of the hyper-arid transboundary Caplina/Concordia coastal aquifer system (Peru/Chile), the FlowSOM network and clustering yielded compelling results in pattern recognition of the aquifer salinization. Analyzing 143 groundwater samples across eight variables, including major ions, the approach supports the identification of distinct clusters and connections between them. Three primary sources of salinization were identified: river percolation, slow lateral aquitard recharge, and seawater intrusion. The analysis demonstrated the superiority of FlowSOM clustering over traditional techniques in the case study, producing clusters that align more closely with the actual hydrogeochemical pattern. The outcomes broaden the utilization of multivariate analysis in water research, presenting a comprehensive approach to support the understanding of groundwater systems.

Nitrate prediction in groundwater of data scarce regions: The futuristic fresh-water management outlook.

Nitrate contamination in groundwater poses a significant threat to water quality and public health, especially in regions with limited data availability. This study addresses this challenge by employing machine learning (ML) techniques to predict nitrate (NO3--N) concentrations in Mexico's groundwater. Four ML algorithms-Extreme Gradient Boosting (XGB), Boosted Regression Trees (BRT), Random Forest (RF), and Support Vector Machines (SVM)-were executed to model NO3--N concentrations across the country. Despite data limitations, the ML models achieved robust predictive performances. XGB and BRT algorithms demonstrated superior accuracy (0.80 and 0.78, respectively). Notably, this was achieved using ∼10 times less information than previous large-scale assessments. The novelty lies in the first-ever implementation of the 'Support Points-based Split Approach' during data pre-processing. The models considered initially 68 covariates and identified 13-19 significant predictors of NO3--N concentration spanning from climate, geomorphology, soil, hydrogeology, and human factors. Rainfall, elevation, and slope emerged as key predictors. A validation incorporated nationwide waste disposal sites, yielding an encouraging correlation. Spatial risk mapping unveiled significant pollution hotspots across Mexico. Regions with elevated NO3--N concentrations (>10 mg/L) were identified, particularly in the north-central and northeast parts of the country, associated with agricultural and industrial activities. Approximately 21 million people, accounting for 10 % of Mexico's population, are potentially exposed to elevated NO3--N levels in groundwater. Moreover, the NO3--N hotspots align with reported NO3--N health implications such as gastric and colorectal cancer. This study not only demonstrates the potential of ML in data-scarce regions but also offers actionable insights for policy and management strategies. Our research underscores the urgency of implementing sustainable agricultural practices and comprehensive domestic waste management measures to mitigate NO3--N contamination. Moreover, it advocates for the establishment of effective policies based on real-time monitoring and collaboration among stakeholders.

Disentangling nitrate pollution sources and apportionment in a tropical agricultural ecosystem using a multi-stable isotope model.

Fertilizers increase agricultural productivity and farmers' income. However, intensive agriculture frequently overuses fertilizers, which in turn can contaminate surface and groundwater. In this study, hydrochemical and multi-isotope (δ15NNO3, δ18ONO3 and δ18OH2O) data have been combined to identify nitrate pollution sources in Ghana's Densu River Basin, trace the Nitrogen (N) biogeochemical processes in the basin and apportion the contribution of each pollution source. Surface water NO3- ranged from 0.3 to 10.6 mg/L (as N), while groundwater NO3- ranged from 0.9 to 34 mg/L. Hierarchical cluster analysis classified the water samples into three spatial categories: upstream, midstream, and downstream, reflecting river and land use patterns. The multi-isotope model considered five primary NO3- sources: atmospheric deposition, manure/sewage, NH4+ in fertilizers, other NO3- based fertilizers and soil N. Nitrification was identified as the major biogeochemical process upstream, whereas mixing of sources and denitrification dominate the midstream to downstream sections of the basin. Nitrate source apportioning using a MixSIAR model reveal that N fertilizers (40 %) and soil N (34 %) contribute the most to nitrate pollution upstream of the river. From the midstream to downstream sections, manure/sewage (43 %) become the dominant nitrate source, reflecting the transition from agriculture to peri-urban and urban land use. This study has shown that soil erosion and runoff contribute to nitrate pollution in the Densu River, at levels comparable to N fertilizers, and groundwater across the basin is impacted mainly by manure/sewage. The multi-isotope analyses allowed the partitioning of N sources in other ways not possible using only classical hydrochemical methods.

Seawater intrusion pattern recognition supported by unsupervised learning: A systematic review and application.

Seawater intrusion is among the world's leading causes of groundwater contamination, as salty water can affect potable water access, food production, and ecosystem functions. To explore such contamination sources, multivariate analysis supported by unsupervised learning tools has been used for decades to aid in water resource pattern recognition, clustering, and water quality data variability characterization. This study proposes a systematic review of these techniques applied for supporting seawater intrusion identification based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and subsequent bibliometric analysis of 102 coastal hydrogeological studies. The most relevant identified methods, including principal components analysis (PCA), hierarchical clustering analysis, K-means clustering, and self-organizing maps, are explained and applied to a case study. Although 74 % of the studies that applied dimensional reduction methods, such as PCA, associated most of the database variance with the salinization process, 77 % of the studies that applied clustering methods associated at least one water sample cluster with the influence of seawater intrusion. Based on the review and a practical demonstration using the open-source R software platform, recommendations are made regarding data preprocessing, research opportunities, and publishing information necessary to replicate and validate the studies.

Hydrochemical controls on arsenic contamination and its health risks in the Comarca Lagunera region (Mexico): Implications of the scientific evidence for public health policy.

Nearly half of the world's urban population depends on aquifers for drinking water. These are increasingly vulnerable to pollution and overexploitation. Besides anthropogenic sources, pollutants such as arsenic (As) are also geogenic and their concentrations have, in some cases, been increased by groundwater pumping. Almost 40 % of Mexico's population relies on groundwater for drinking water purposes; much the aquifers in semi-arid and arid central and northern Mexico is contaminated by As. These are agricultural regions where irrigation water is primarily provided from intenstive pumping of the aquifers leading to long-standing declines in the water table. The focus of this study is the main aquifer within the Comarca Lagunera region in Northern Mexico. Although the scientific evidence demonstrates that health effects are associated with long-term exposure to elevated As concentrations, this knowledge has not yielded effective groundwater development and public health policy. A multidisciplinary approach - including the evaluation of geochemistry, human health risk and development and public health policy - was used to provide a current account of these links. The dissolved As concentrations measured exceeded the corresponding World Health Organization guideline for drinking water in 90 % of the sampled wells; for the population drinking this water, the estimated probability of presenting non-carcinogenic health effects was >90 %, and the lifetime risk of developing cancer ranged from 0.5 to 61 cases in 10,000 children and 0.2 to 33 cases in 10,000 adults. The results suggest that insufficient policy responses are due to a complex and dysfunctional groundwater governance framework that compromises the economic, social and environmental sustainability of this region. These findings may valuable to other regions with similar settings that need to design and enact better informed, science-based policies that recognize the value of a more sustainable use of groundwater resources and a healthier population.

Predicting adverse scenarios for a transboundary coastal aquifer system in the Atacama Desert (Peru/Chile).

The Caplina/Concordia transboundary coastal aquifer system, located in the Atacama Desert, is the primary source of water supply for domestic use and irrigation for La Yarada-Los Palos (Peru) and Concordia (Chile) agriculture districts, and to a lesser extent, for Tacna province public supply use (Peru). Despite the scarce amount of rainfall (<20 mm/year) in the area and the limited recharge coming from the Andean highlands, this transboundary aquifer system has been overexploited mainly for agriculture since before the 2000s on the Peruvian side. Consequently, this has caused groundwater depletion and seawater intrusion. In this study, comprehensive hydrogeological information was integrated to understand the aquifer system's behavior and the effects to which it has been subjected to groundwater overexploitation. To that end, a 3D hydrogeological framework was developed using the LEAPFROG software and a constant-density groundwater flow model with equivalent heads was generated in FEFLOW software, which was adjusted with Monte Carlo analysis and conventional automated calibration. Finally, eight scenarios, considering various water resource management options proposed by the authority and potential climatic trends (CMIP6), were simulated from 2020 to 2040. The results showed that between 2002 and 2020, the increase in the seawater wedge and the average groundwater level decline were 216 hm3/year and 7 m, respectively. It is expected that the depletion will continue with a groundwater level decline between 5 and 8 m and an increase in the seawater wedge between 1120 hm3/year and 1175 hm3/year for the forecast period. The study concludes that the aquifer system will remain unsustainable for the next 20 years, regardless of the selected scenarios, and suggests that any mitigation measure requires the participation of stakeholders from Peru, Chile, and Bolivia.

Mapping salinization and trace element abundance (including As and other metalloids) in the groundwater of north-central Mexico using a double-clustering approach.

This study aimed to determine the reliability of the double-clustering method to understand the spatial association and distribution of major and minor constituents in the groundwater of an arid endorheic basin in central Mexico (Comarca Lagunera Region). The results of the double-clustering approach were compared with well-known spatial statistics such as spatial autocorrelations (Moran index) and the local indicator of spatial association (LISA). Fifty-five groundwater samples were collected from diverse wells within the basin, and the major ions, metalloids, and trace elements were determined. Overall, the double-clustering analysis was an effective tool for identifying lithogenic/anthropogenic processes occurring in the basin and for establishing zones with high or low abundance of major ions and trace elements, even where processes affecting the groundwater quality were spatially dispersed. Although 89% of the samples showed As higher than the threshold value of 10 µg/L proposed by the World Health Organization for drinking water, both the double-clustering and LISA analyses identified As hotspots in the alluvial aquifer, where the extraction of deeper and warmer groundwater might promote the concomitant release of the metalloids As, Sb, and Ge and the trace elements V and W. Similarly, both statistical analyses identified mountainous sectors where the weathering of silicates and carbonates plays a key role in the abundance of HCO3-, Ga, and Ba. However, the LISA analysis failed to identify hotspots of carbonate-derived elements such as Ca, Mg, Sr, and U and silicate-derived elements such as Ca, Mg, K, Sr, Rb, Cs, Pb, Ni, and Y. Otherwise, the double-clustering analysis clearly defined high- and low-concentration zones for all these elements in the study region. Unlike the LISA analysis, the double-clustering approach was also successful in determining alluvial areas with high concentrations of Si and Ti and areas where the concentrations of Na, Cl-, SO42-, NO3-, B, Li, Cu, Re, and Se in groundwater were elevated, increasing the groundwater salinity. Overall, this study demonstrated that the double-clustering is an easy-to-apply approach, capable of visualizing disperse zones where specific anthropogenic processes may threaten the groundwater quality.

Determining nitrate and sulfate pollution sources and transformations in a coastal aquifer impacted by seawater intrusion-A multi-isotopic approach combined with self-organizing maps and a Bayesian mixing model.

Over the past few decades, the La Paz aquifer system in Baja California Sur, Mexico, has been under severe pressure due to overexploitation for urban water supply and agriculture; this has caused seawater intrusion and deterioration in groundwater quality. Previous studies on the La Paz aquifer have focused mainly on seawater intrusion, resulting in limited information on nitrate and sulfate pollution. Therefore, pollution sources have not yet been identified sufficiently. In this study, an approach combining hydrochemical tools, multi-isotopes (δ2HH2O, δ18OH2O, δ15NNO3, δ18ONO3, δ34SSO4, δ18OSO4), and a Bayesian isotope mixing model was used to estimate the contribution of different nitrate and sulfate sources to groundwater. Results from the MixSIAR model revealed that seawater intrusion and soil-derived sulfates were the predominant sources of groundwater sulfate, with contributions of ~43.0% (UI90 = 0.29) and ~42.0% (UI90 = 0.38), respectively. Similarly, soil organic nitrogen (~81.5%, UI90 = 0.41) and urban sewage (~12.1%, UI90 = 0.25) were the primary contributors of nitrate pollution in groundwater. The dominant biogeochemical transformation for NO3- was nitrification. Denitrification and sulfate reduction were discarded due to the aerobic conditions in the study area. These results indicate that dual-isotope sulfate analysis combined with MixSIAR models is a powerful tool for estimating the contributions of sulfate sources (including seawater-derived sulfate) in the groundwater of coastal aquifer systems affected by seawater intrusion.

Estimation of nitrate pollution sources and transformations in groundwater of an intensive livestock-agricultural area (Comarca Lagunera), combining major ions, stable isotopes and MixSIAR model.

The identification of nitrate (NO3-) sources and biogeochemical transformations is critical for understanding the different nitrogen (N) pathways, and thus, for controlling diffuse pollution in groundwater affected by livestock and agricultural activities. This study combines chemical data, including environmental isotopes (δ2HH2O, δ18OH2O, δ15NNO3, and δ18ONO3), with land use/land cover data and a Bayesian isotope mixing model, with the aim of reducing the uncertainty when estimating the contributions of different pollution sources. Sampling was taken from 53 groundwater sites in Comarca Lagunera, northern Mexico, during 2018. The results revealed that the NO3- (as N) concentration ranged from 0.01 to 109 mg/L, with more than 32% of the sites exceeding the safe limit for drinking water quality established by the World Health Organization (10 mg/L). Moreover, according to the groundwater flow path, different biogeochemical transformations were observed throughout the study area: microbial nitrification was dominant in the groundwater recharge areas with elevated NO3- concentrations; in the transition zones a mixing of different transformations, such as nitrification, denitrification, and/or volatilization, were identified, associated to moderate NO3- concentrations; whereas in the discharge area the main process affecting NO3- concentrations was denitrification, resulting in low NO3- concentrations. The results of the MixSIAR isotope mixing model revealed that the application of manure from concentrated animal-feeding operations (∼48%) and urban sewage (∼43%) were the primary contributors of NO3- pollution, whereas synthetic fertilizers (∼5%), soil organic nitrogen (∼4%), and atmospheric deposition played a less important role. Finally, an estimation of an uncertainty index (UI90) of the isotope mixing results indicated that the uncertainties associated with atmospheric deposition and NO3--fertilizers were the lowest (0.05 and 0.07, respectively), while those associated with manure and sewage were the highest (0.24 and 0.20, respectively).

Tracking nitrate and sulfate sources in groundwater of an urbanized valley using a multi-tracer approach combined with a Bayesian isotope mixing model.

Over the past decades, groundwater quality has deteriorated worldwide by nitrate pollution due to the intensive use of fertilizers in agriculture, release of untreated urban sewage and industrial wastewater, and atmospheric deposition. Likewise, groundwater is increasingly polluted by sulfate due to the release of domestic, municipal and industrial wastewaters, as well as through geothermal processes, seawater intrusion, atmospheric deposition, mineral dissolution, and acid rain. The urbanized and industrialized Monterrey valley has a long record of elevated nitrate and sulfate concentrations in groundwater with multiple potential pollution sources. This study aimed to track different sources and transformation processes of nitrate and sulfate pollution in Monterrey using a suite of chemical and isotopic tracers (δ2H-H2O, δ18O-H2O, δ15N-NO3, δ18O-NO3 δ34S-SO4, δ18O-SO4) combined with a probability isotope mixing model. Soil nitrogen and sewage were found to be the most important nitrate sources, while atmospheric deposition, marine evaporites and sewage were the most prominent sulfate sources. However, the concentrations of nitrate and sulfate were controlled by denitrification and sulfate reduction processes in the transition and discharge zones. The approach followed in this study is useful for establishing effective pollution management strategies in contaminated aquifers.

Groundwater Flow Processes and Human Impact along the Arid US-Mexican Border, Evidenced by Environmental Tracers: The Case of Tecate, Baja California.

With the increasing population, urbanization and industry in the arid area of Tecate, there is a concomitant increase in contaminants being introduced into the Tecate River and its aquifer. This contamination is damaging the usable groundwater supply and making local residents and commercial enterprises increasingly dependent on imported water from the Colorado River basin. In this study we apply a suite of chemical and isotopic tracers in order to evaluate groundwater flow and assess contamination trends. Groundwater recharge occurs through mountain-block and mountain-front recharge at higher elevations of the ranges. Groundwater from the unconfined, alluvial aquifer indicates recent recharge and little evolution. The increase in salinity along the flow path is due to interaction with weathering rock-forming silicate minerals and anthropogenic sources such as urban wastewater, residual solids and agricultural runoff from fertilizers, livestock manure and/or septic tanks and latrines. A spatial analysis shows local differences and the impact of the infiltration of imported waters from the Colorado River basin. The general trend of impaired water quality has scarcely been documented in the last decades, but it is expected to continue. Since the groundwater system is highly vulnerable, it is necessary to protect groundwater sources.