The research, conducted by Tapas Ranjan Patra of the Department of Geography at Rajendra University in Balangir, Odisha, began with an unusually rigorous data-cleaning exercise. Groundwater-quality monitoring programs generate enormous multi-parameter datasets, but raw records are riddled with duplicates, inconsistent sampling locations and measurements that fail basic chemical plausibility checks. Starting from the Central Ground Water Board’s monitoring archive covering 2019 to 2023, the study applied geospatial deduplication to remove repeated observations at the same sites and charge-balance screening to discard samples whose major-ion concentrations did not balance electrically. The result was a curated dataset of 1,268 site-level samples, each carrying measurements of electrical conductivity, total dissolved solids, pH, hardness and the major cations and anions that define groundwater chemistry.

With the cleaned dataset in hand, the analysis turned to the central question of whether Odisha’s groundwater behaves as one continuous chemical spectrum or as a small number of distinct regimes. Using Gaussian mixture modeling, a statistical technique that identifies hidden subpopulations within multivariate data, alongside hierarchical cluster analysis and principal component analysis, the study identified two stable hydrochemical regimes separated along a dominant mineralization gradient. The stability of this two-regime structure was assessed with cluster-stability measures, including the adjusted Rand index, and the number of groups was selected using Bayesian and extended Bayesian information criteria rather than arbitrary thresholds. In other words, the split was not imposed by the analyst but emerged from the data itself and proved reproducible under repeated evaluation.

The two regimes differ in a chemically meaningful way. Regime 1 showed consistently higher electrical conductivity, total dissolved solids and major-ion concentrations than Regime 0, indicating a more heavily mineralized groundwater setting. Such mineralization typically reflects longer residence times of water in the subsurface, greater interaction with host rocks, and in some settings the influence of salinity-related processes such as those documented in coastal aquifers of the region. Regime 0, by contrast, represents a less mineralized, more dilute chemical environment. The study interpreted these contrasts through standard hydrochemical tools, including Gibbs diagrams and ion-ratio analysis, to characterize the processes governing each regime’s major-ion signature.

The most consequential finding concerns the two contaminants that matter most for drinking-water safety in the region: nitrate and fluoride. Both are regulated under Bureau of Indian Standards and World Health Organization guidelines because chronic exposure carries serious health consequences. Excessive nitrate, often linked to agricultural runoff, septic leakage and organic waste, is associated with methemoglobinemia in infants and other adverse outcomes, while elevated fluoride, typically geogenic in origin, causes dental and skeletal fluorosis. When the study cross-tabulated guideline exceedances against the two hydrochemical regimes, a clear pattern emerged. Nitrate exceedance rose from 4.45 percent of samples in Regime 0 to 13.81 percent in Regime 1, roughly a threefold increase, while fluoride exceedance climbed from 0.91 percent to 4.86 percent, more than a fivefold jump.

A critical concern in any observational study of this kind is that such contrasts might simply be artifacts of uneven sampling. If the more mineralized regime happened to be sampled more intensively in heavily contaminated districts, the apparent association could reflect geography rather than chemistry. To rule this out, the study employed district-adjusted statistical models, including generalized linear mixed models with district-level fixed effects, which compare samples within the same district and thereby absorb any district-wide confounding. These models confirmed that the elevated exceedance rates in Regime 1 persisted even after accounting for uneven district-level sampling, indicating that the hydrochemical regime itself carries genuine information about contamination risk.

The analysis went a step further by examining how the chemical parameters within each regime relate to one another. Using conditional-dependence network analysis, a method that maps the partial correlations among variables while controlling for all others, the study found that the higher-mineralization regime was more strongly organized by salinity-related hydrochemical structure. In Regime 1, the major ions moved together in a tightly connected network consistent with shared salinization processes, whereas the network structure in the dilute regime was weaker and differently arranged. This regime-specific organization suggests that the two settings are governed by distinct geochemical processes, not merely different points on a single continuum, and that contamination pathways may therefore differ between them.

The practical implications extend well beyond Odisha. Groundwater supplies the majority of rural drinking water in eastern India, and monitoring agencies face chronic constraints on laboratory capacity and field resources. The study’s regime-based framework offers a way to prioritize that effort: samples drawn from high-mineralization settings deserve more frequent nitrate and fluoride testing, while resources in low-mineralization zones can be allocated more sparingly. Because the underlying data are publicly available through the Central Ground Water Board portal, and the analytical approach relies on standard statistical tools, the framework could be replicated in other spatially heterogeneous groundwater systems across India and beyond, where monitoring networks similarly struggle to convert scattered measurements into actionable priorities.

The study also contributes to a broader scientific conversation about how large environmental datasets should be interpreted. Rather than treating each contaminant measurement in isolation, or collapsing entire aquifers into a single water-quality index, the regime-based approach preserves the natural structure of the data and uses that structure to sharpen surveillance. As the author notes, nitrate and fluoride screening burdens are concentrated in specific hydrochemical settings rather than being uniformly distributed across the monitoring network. Recognizing and mapping those settings transforms routine monitoring from a passive record-keeping exercise into a targeted risk-prioritization tool, one that could help public health authorities intervene before contaminated wells reach the families who depend on them.

Subject of Research: Hydrochemical regimes and nitrate–fluoride contamination surveillance in Odisha's groundwater

Article Title: Hydrochemical regimes and nitrate–fluoride surveillance in groundwater of Odisha, Eastern India

Article References: Patra, T. R. (2026). Hydrochemical regimes and nitrate–fluoride surveillance in groundwater of Odisha, Eastern India. Environmental Monitoring and Assessment, 198(10), Article 1064. https://doi.org/10.1007/s10661-026-15901-1

Image Credits: AI Generated

DOI: 10.1007/s10661-026-15901-1

Keywords: groundwater, hydrochemistry, nitrate, fluoride, Odisha, water quality monitoring, Gaussian mixture model, mineralization gradient, drinking-water safety, Central Ground Water Board, surveillance prioritization, eastern India