The Vellore district lies between 12° 15' to 13° 15' North latitudes and 78° 20' to 79° 50' East longitudes in Tamil Nadu State. The Palar River is the main river of the study area and passes through the whole Vellore district, where the Palar River covers 64 km from the north side of the Palar River and 49 km south of the Palar River in the Vellore district, India (Fig. 1). The Vellore district has five blocks: Katpadi, Guidyathum, K.V. Kuppam, Anicut and Vellore. The district experiences three separate seasons: summer, monsoon, and winter. The summer season extends from March through June, and the monsoon season extends from October to December. Post-monsoon samples were collected in November 2021, and pre-monsoon samples were collected in May 2022 (Climate data). The average rainfall in the district is 1,034.1 mm (40.71 in) each year.

Surface water was collected from thirty-four different locations to evaluate the underlying aquifer system of the research area, as shown in Fig. 1. The study region is characterized by low-lying plains interspersed with hills. Lowlands dominate the eastern and central parts of the district, while rocky terrains are prevalent in the western regions. The Palar River serves as a vital resource, supporting agriculture, industries, and domestic needs, with the area's soil primarily comprising sandy loam and red lateritic formations. Vellore district has a tropical climate, with very hot summers reaching 30°C–43°C and mild winters between 18°C–30°C. It has moderate monsoon rainfall of 1,034mm per year. Groundwater and the alluvial aquifer system of the Palar River support the district; but overexploitation has begun to compromise water quality. Samples were collected during the day for 15 days each season using a random sampling technique. Quality assurance and control procedures were meticulously followed to ensure accuracy and precision. All the samples were collected in reusable plastic bottles multiple times with sample water before collection and then transferred to the laboratory for analysis. A digital pH meter was used to determine the sample pH (MK VI), providing insights into the acidity or alkalinity. To assess the Total Dissolved Solids (TDS) and Electrical Conductivity (EC) of the water, an Instrument 308 Systronics TDS meter was used to determine the mineral content and conductivity. Turbidity, an indicator of water clarity, was measured using a Systronics Nephelo turbidity meter 132. The concentrations of nitrate and sulphate were determined using a Jasco V-730 UV–Vis spectrophotometer, which provided insight into the presence of these important ions. Sodium and potassium levels in surface water samples were tested using a flame photometer (Systronics 128), which provided data on the concentrations of these essential elements. Ammonia levels were measured through titration against standard N/10 HCL, providing information on the presence of this nitrogen compound. Standard methods such as IS-1500:2012 were used to assess the total hardness, calcium content, and total alkalinity, while the argentometric method was used to determine the chloride content. Finally, magnesium concentrations were computed using the APHA (2^nd Edition) mathematical method, and the results were compared to the national (BIS) and international (WHO) standards for comprehensive analysis and comparison. All equipment was pre-calibrated with standard solutions using standard techniques, and electrodes were cleaned with double-distilled water before use. For pH calibration, buffer solutions of pH 7 and pH 9 were used, while for nitrate analysis, blank samples were provided to verify accuracy. Before each measurement, probes were conditioned in the sample to achieve the optimal stabilization time. Blank samples have been used to perform chemical analysis for the entire study, and normal operating procedures and safety protocols have been conducted at all steps of the process.

Fifty milliliters of each water sample were placed in a beaker and digested with 2.5ml of concentrated HNO₃ on a hot plate. After digestion, the samples were filtered through Whatman No. 44 filter paper and transferred to 100ml volumetric flasks, and volume make-up was done with distilled water. For standard preparation, a stock solution containing 1000mg/L of Cd, Cu, Ni, Co, Fe, Mn, Zn, and Cr was prepared using analytical-grade chemicals. Working standards for each metal, with concentrations of 2.0, 4.0, 6.0, 8.0, and 10.0mg/L, were created from the stock solution using micropipettes. Deionized water was used as the reagent blank.

The Atomic Absorption Spectroscopy instrument was configured to optimize the flame conditions and absorbance. Subsequently, blanks (deionized water), standards, sample blanks, and samples were introduced into the flame. Analysis of heavy metals was conducted using a Perkin Elmer 400 atomic absorption spectrometer. Calibration curves correlating the concentration with the absorbance were generated. Statistical analysis involved fitting a straight line through the least squares method. Additionally, a blank reading was recorded, and appropriate adjustments were applied to determine the concentrations of different elements [20–21].

Spatial analysis of various physicochemical parameters was carried out using GIS (Geographical Information System) contouring methods along with QGIS software (3.34.1 Prizren) [8]. For every physicochemical characteristic, spatial distribution maps were produced using the interpolation method known as Inverse Distance Weighting (IDW). The spatial distribution of the drinking water standards followed as per the Bureau of Indian Standards [9].

The fuzzy logic model was introduced by Zadeh [10], which aimed to develop systems that more closely mimic human thinking processes. Some studies have demonstrated the efficiency of this method in addressing complex systems under uncertain conditions [22–23]. Matlab (version 2023b) was used to develop the Mamdani Fuzzy Inference System in this work. A fuzzy controller system consists of three primary functions: fuzzification, inference, and defuzzification, as illustrated in Fig. 2. Within the Fuzzy Inference System (FIS), human expertise is encoded into a series of fuzzy linguistic rules that drive approximate decision-making. This system can be realised not only through human knowledge but also through a combination of a Fuzzy Rule-Based System (FRBS) and a fuzzified component, reducing reliance solely on human expertise.

A comprehensive statistical analysis was conducted utilising various tools and techniques, including the hydrochemistry of surface water piper diagrams, Durov diagrams, Principal Component Analysis, Gibbs diagrams, and Pearson’s correlation matrices. The analysis aimed to elucidate the relationships between twenty-five parameters. Origin 2024 software was used to facilitate this investigation. The surface water hydrochemistry piper diagram offered insights into the compositional relationships, visually representing their interactions. The Durov diagram further illustrates the distribution patterns of elements and their concentrations in the water samples. Gibbs diagram analysis enabled the examination of the hydrogeochemical processes influencing surface water chemistry, highlighting the dominance of certain factors such as rock weathering or evaporation. Principal Component Analysis was utilised to extract significant variables and elucidate the underlying patterns in the dataset, aiding in the interpretation of complex relationships among variables. Pearson’s correlation matrix was employed to quantify the degree of association between pairs of parameters, providing a comprehensive overview of their interdependencies and facilitating the identification of meaningful correlations.

The Nemerow Pollution Index is a water quality metric used to assess pollution levels [24–25]. This calculation enabled the identification of parameters contributing to pollution for each heavy metal.

The NPI was calculated by using the following Eq. (3)

where Cn represents the observed concentration of the nth parameter, and Ln represents the acceptable limit of the nth parameter. The units for Cn and Ln in Eq. (3) should be the same. Each NPI value reflects the contribution of each single parameter to causing pollution. The NPI is a unitless measurement. An NPI greater than 1.0 indicates the presence of impurities in water. Using Nemerow's pollution index, the pollution parameters were computed for each station.

The Geoaccumulation Index (Igeo) serves as a valuable tool for mitigating the impact of geological factors when evaluating heavy metal pollution [13]. It is fabricated using the following Eq. (4)

where:

A scatter diagram was proposed by Gibbs to explain the relationship between chemical compounds dissolved in surface water and aquifer lithology [15]; in turn, this diagram describes the mechanism of major cations and anions present in surface water that participate in three natural mechanisms, namely, rock dominance, precipitation dominance and evaporation dominance. The Gibbs ratio was determined using the following Eqs. (5), (6)

Heavy metals and metalloids pose significant health concerns due to their toxicity, even at low concentrations, and their ability to accumulate in the tissues of living organisms over time. Metals such as Chromium (Cr), Nickel (Ni), Cadmium (Cd), and Cobalt (Co) have no beneficial effects on the human body. Prolonged exposure to these metals can disrupt normal organ functions and lead to health issues. While some trace metals, such as Copper (Cu), Zinc (Zn), Iron (Fe), and Manganese (Mn), are essential micronutrients required for metabolic processes, excessive amounts can have adverse health effects [16]. Anthropogenic sources of trace metals in surface water include leaching from natural materials in soil and rocks, residues from agrochemicals, releases from sewage treatment plants and industrial runoff, and accidental releases from landfills and chemical accidents.

Chromium (Cr³⁺ and Cr⁶⁺) is a naturally occurring trace metal typically found in very low concentrations in surface water and remains unaffected by point-source contamination [17]. The Vellore district is surrounded by significant sources of chromium discharge, including the tanning industry and landfills. In contrast, Cr⁺³, Cr⁺⁶ can easily penetrate cell membranes [18]. Prolonged exposure to chromium can result in its accumulation in vital organs such as the lymph nodes, kidneys, liver, lungs, and spleen, potentially causing organ damage [19]. The WHO, US-EPA, and BDWS recommend a permissible maximum value of Cr in drinking water of 0.05 mg/L. However, the chromium concentrations surpassed this threshold in the eastern part of the study area during both seasons. The post-monsoon concentrations ranged from 1.89 to 3.33 mg/L, while the pre-monsoon concentrations varied from 0 to 115.78 mg/L. The elevated levels of chromium in the surface water in this region are likely attributed to chromium plating industries situated in the eastern area of the Ranipet district due to industrialisation near the study site (Fig. 4a and b). Chromium, a prevalent contaminant in the region, is recognised for its carcinogenic potential, particularly hexavalent chromium, which can cause lung, liver, and kidney damage upon chronic exposure [20].

Cadmium (Cd), a highly toxic trace element known for its prolonged half-life, naturally coexists with zinc minerals and can be released into surface water from various sources, such as buried metal refinery by-products, electronic components, and coal combustion [16]. The consumption of vegetables cultivated in contaminated soil or fish and seafood from polluted water bodies can result in cadmium exposure. Acute exposure to cadmium can lead to symptoms such as nausea, cancer, diarrhoea, anaemia, bone marrow disorders, muscle cramps, liver injury, and kidney failure. In the study region, cadmium concentrations ranged from BDL to 0.15 mg/L in the post-monsoon season and from BDL to 1.69 mg/L in the pre-monsoon season (Fig. 4c and d). The spatial distribution revealed higher cobalt concentrations in the western part of the study area during both seasons, indicating persistent contamination in those regions. According to WHO guidelines, the permissible limit for cadmium is 0.01 mg/L. Cadmium can have both beneficial and harmful effects on human health, depending on its concentration and isotopic composition. As a significant environmental pollutant, excessive cadmium exposure is linked to serious health issues, including renal dysfunction, bone demineralization (osteomalacia and osteoporosis), and an elevated risk of cardiovascular diseases [21]. Although cadmium is an essential component of vitamin B12, prolonged intake beyond permissible limits can adversely affect cardiovascular health [26].

Copper (Cu) plays a vital role in the metabolism of animals and plants, yet excessive exposure to potable water can induce gastrointestinal discomfort. Prolonged exposure may result in copper toxicosis, leading to liver and kidney damage, anaemia, hepatic cirrhosis, and basal ganglia degeneration [27]. Elevated copper levels in aquatic environments pose significant threats to fish and other aquatic life forms [28]. During the pre-monsoon season, the copper concentration in surface water ranged from BDL to 3.16 mg/L, with an average of 0.68 mg/L (Fig. 4e and f). Interestingly, copper was not detected in surface water during the post-monsoon period. Elevated copper concentrations were notably observed on the western side of the study area during the pre-monsoon period, surpassing the permissible limit for drinking water in some samples. While the overall values remain within permissible limits, heightened concentrations in specific areas may result from agricultural activities. For instance, the leaching of copper-based fertilisers and fungicides into surface water due to rainwater and irrigation can increase copper levels [22]. Additionally, leaching from open dumping sites of solid wastes and the corrosion of household materials also contribute to copper concentrations [23]. Near Vellore, several industries, including and companies involved in metal processing, operate within the region. These industries have potential implications for environmental and water pollution similar to the issues associated with the Vedanta Sterlite Copper plant in Thoothukudi, Tamil Nadu [29]. These establishments may be a factor in the heavy metal pollution of nearby water sources, which might have negative effects on the environment and public health.

Zinc (Zn), an essential nutrient found in most natural foods, has permissible limits in drinking water ranging from 5 to 15 mg/L [9]. In the study area, zinc concentrations range from BDL to 0.49 mg/L post-monsoon and from BDL to 1.29 mg/L pre-monsoon, with the eastern side showing the highest contamination levels. The average zinc concentration in both seasons exceeded the WHO limits. Fig. 4g and h, shows that the eastern side of the study area exhibits the highest contamination levels of zinc. Excessive zinc intake can disrupt metabolism and immune function, increasing the risk of infections, delayed sexual maturation, anemia, and birth defects [30]. Contamination often stems from household corrosion and industrial activities. While zinc is essential for metabolism and development, elevated levels in drinking water can lead to gastrointestinal distress, neurological issues, and long-term health complications [25]. Moreover, excessive zinc can interfere with the absorption of other vital nutrients, further impacting overall health [31].

During the pre-monsoon period, Manganese (Mn) concentrations in samples range from BDL to 0.60 mg/L, exceeding permissible limits and potentially causing neurological effects upon continuous consumption. This value is greater than that of post-monsoon samples, where manganese levels are negligible (Fig. 4i and j). The contamination pathway involves industrial effluents and domestic waste initially spreading on the surface, with rainwater helping to dilute the contamination. Excessive consumption of manganese-rich water (1.8 mg/L) can lead to symptoms resembling Parkinson's disease, memory impairment, hallucinations, disorientation, and impulsive behaviour. The secondary extreme contaminant level for manganese is 0.5 mg/L to avoid offensive taste, odour, colour, staining, and corrosion [24].

Nickel typically exists in the Ni²⁺ form in natural water at a pH of 5–9 [25]. In the study area, nickel concentrations range from BDL to 3.50 mg/l post-monsoon, with an average of 0.209 mg/L. During the pre-monsoon season, nickel contamination ranged from BDL to 3.058 mg/L, with an average of 0.38 mg/L, indicating higher concentrations than those during the post-monsoon period. During the pre-monsoon period, all locations exceeded the permissible limit for nickel according to [9] standards, as indicated in (Fig. 4k and l). In contrast, during the post-monsoon period, nickel concentrations were lower overall. However, most samples still exceeded the permissible limit, except at five locations that reported that prolonged consumption of nickel above the maximum contamination level can result in damage to the heart and liver, body weight loss, and dermatitis [32]. In the study area, the primary sources of nickel contamination are sewage water and the corrosion of nickel alloy materials. Alarmingly high nickel concentrations observed in this study are associated with allergic dermatitis, respiratory disorders, and an increased risk of cancer [33].

Cobalt (Co), a naturally occurring element with properties akin to those of iron and nickel, is commonly present in rocks, soil, water, plants, and animals in trace amounts [34]. It is often found with other elements, such as oxygen, sulphur, and arsenic. The cobalt concentrations in the water samples analysed during the pre-monsoon season ranged from negligible to 1.69 mg/L, while the post-monsoon levels were minimal (Fig. 4m and n). Although no specific permissible limit has been set by [25], cobalt possesses significant radioactivity, posing considerable health risks at elevated concentrations. This study revealed notable concentrations primarily in the western region during both seasons. The potential health hazards associated with high cobalt levels include respiratory issues such as pneumonia, asthma and wheezing [30]. Lower exposures may manifest as nausea, skin rashes, and vomiting, whereas prolonged exposure could lead to more severe symptoms such as diarrhoea, bleeding, coma, and even fatality [35]. Radioactive cobalt is also carcinogenic, causing cellular damage resulting in skin blisters, burns, hair loss, and temporary sterility [36]. Cobalt, while essential in trace amounts, becomes toxic at elevated levels, potentially causing cardiomyopathy and thyroid dysfunction [37]. Elevated manganese levels are known to impair neurological function, leading to symptoms similar to Parkinson's disease, particularly in vulnerable populations such as children and the elderly [38].

Iron (Fe) is a significant concern in India’s rural drinking water. Although a low level of iron is essential for human and plant metabolism, it promotes the growth of objectionable bacteria, known as 'iron bacteria', which can coat piping systems with a slushy residue. High iron levels (over 0.3 mg/L) can lead to stomach issues, vomiting, diabetes, nausea, and hemochromatosis [23]. Iron is a crucial nutrient for human nutrition, playing a vital role in metabolic processes and preventing anaemia, as it is a key component of haemoglobin and myoglobin. It is also necessary to function with enzymes such as cytochromes, peroxides, catalase, and various haemoproteins and flavoproteins [39]. In the study area, the iron concentrations ranged from 0 to 0.62 mg/L in the post-monsoon season and from BDL to 1.72 mg/L in the pre-monsoon season, with averages of 0.038 mg/L and 0.26 mg/L, respectively, as shown in the spatial distribution map in Fig. 4o and p. The eastern side, characterised by industrial activity, experiences heavy metal contamination. The spatial distribution suggests that iron contamination in surface water is primarily due to natural weathering and erosion processes, which release iron from rocks and soils into water bodies. Additionally, agricultural runoff, potentially carrying iron-containing fertilisers and pesticides, contributes to the elevated iron levels observed in these waters [40]. The increase in iron levels during the pre-monsoon period is attributed to interactions with rock and water. The surface water iron concentration may also result from the corrosion of household materials and leaching from pipes, fittings, and industrial activities such as smelting and welding [27]. Iron, another dominant contaminant in Vellore's water sources, can lead to conditions such as hemochromatosis, characterized by excessive iron accumulation in vital organs, resulting in liver, heart, and pancreatic damage [41]. Additionally, the bioaccumulation of these metals in the food chain exacerbates their impact, as agricultural products irrigated with contaminated water become a secondary source of exposure.

Fig. 5 presents the minimum and maximum concentrations of 16 physicochemical parameters concerning distances from the Palar River during both the pre-monsoon and post-monsoon seasons. It highlights significant variations in water quality, including pH, EC, TDS, turbidity, alkalinity, hardness, major ions (Calcium, Magnesium, Sodium, Potassium, Sulphate, Bicarbonate, Carbonate), and key contaminants such as Ammonia and Nitrate. The spatial distribution indicates changes in contamination levels based on proximity to the river, with notable differences observed between seasons. These findings provide crucial insights into the impact of hydrogeochemical processes, anthropogenic activities, and seasonal variations on drinking and household water quality [44–45]. The contamination levels are significantly elevated near the sampling points in the proximity of the Palar River Basin, indicating a close correlation with nearby industrial activities, including those of the BHEL (Bharat Heavy Electricals Limited), leather, and tanning industries. The industrial zone, located adjacent to the banks of the Palar River, has emerged as a critical source of pollutants, adversely impacting the quality of river water [46–47]. These contaminants, including those from leather tanneries and small-scale dyeing industries, impede the natural flow of the river and contribute to the dispersion of contaminated surface water [48]. The spatial-temporal mapping of heavy metal concentrations has been categorised based on the WHO drinking water standard limits, providing a clear assessment of contamination levels and their potential impact on water quality [25]. It utilizes a color gradient from faint green to dark green, indicating varying levels of contamination. Darker shades signify contamination surpassing permissible limits for specific physicochemical parameters [49]. The Supplementary materials includes spatial variation maps for all 16 parameters (Fig. S1–S16). The concentration of contaminants is generally lowest at sample locations farther from the Palar River, while the highest concentrations are observed closer to the riverbed. This trend suggests that contamination levels are elevated near the Palar River, particularly in areas adjacent to the riverine habitat.

Seasonal variations, particularly between the pre-monsoon and post-monsoon periods, significantly influence water quality in the Palar River Basin. During the pre-monsoon season, high temperatures and reduced water flow often lead to increased concentrations of pollutants, as lower water volumes reduce the natural dilution effect. This is particularly evident for parameters such as TDS, EC, and heavy metal concentrations (e.g., chromium and nickel), which are typically higher in the pre-monsoon months due to evaporation and reduced groundwater recharge [50]. In contrast, the post-monsoon season is characterized by increased rainfall and surface runoff, which can dilute certain pollutants but also introduce new contaminants from agricultural fields and urban areas. The seasonal variability highlights the dynamic interactions between hydrological processes and anthropogenic activities in the study area. By incorporating references from previous studies that have documented similar seasonal influences on water quality, this section now provides a more comprehensive discussion of the seasonal dynamics affecting the Palar River Basin.

In Fig. 6, the Three-Dimensional fuzzy surface is displayed as a graphical user interface tool that illustrates how the input variables interact and influence the creation of the output variable. The figure's lowest values stand for "Igeo" and "NPI," while the highest values stand for the DWQ. Consequently, the quality of surface water used for drinking deteriorates when "Igeo" and "NPI" values are low. By combining many characteristics into a single fuzzy model, the MFISM membership function reduces the uncertainties associated with conventional classification methods and enables a more thorough assessment of water quality. Some studies have successfully assessed water quality using fuzzy logic-based models, particularly in urban and industrial regions where pollution sources are diverse and ever-changing. Zehra et al. [51] demonstrated how fuzzy models may effectively capture spatial and temporal variations in water quality by evaluating the heavy metal contamination of the Yamuna River using a fuzzy inference technique. Fuzzy logic was also used by Wagh et al. [52] to integrate hydrochemical parameters for evaluating groundwater quality, demonstrating the usefulness of this method for handling variability and uncertainty. Through MFISM, the study technique creates a comprehensive assessment framework that integrates multiple indices, such as Igeo and NPI. The integrated approach improves the accuracy of interpretations that examine trends in water quality over time and geography. PCA analysis and correlation evaluation indicate that geographical locations that possess a higher Igeo and NPI values have been attributed to the quality of drinking water being poor. When MFISM technology and GIS spatial analysis are combined, an integrated water quality assessment system is created that helps make better judgments regarding pollution control and water resource management. This process applies to a variety of hydrogeochemical settings since it aligns with recent advancements in water quality modelling.

Drinking Water Quality (DWQ) was the output parameter in this study, while the Nemerow Pollution Index (NPI) and the Geoaccumulation Index (Igeo) were integrated as input parameters using the Modified Fuzzy Inference System Model (MFISM). According to the classification results, the research area's water quality varies greatly, ranging from "Excellent" to "Unsuitable for Drinking." The MFISM model offers a strong and adaptable framework that enables the integration of several elements, such as temporal, geographical, and physicochemical aspects, into a single decision-making procedure. The findings show that pre-monsoon seasons, when concentrations of prioritised contaminants, heavy metals, and dissolved solids are higher, and areas with higher industrial activity have lower water quality. The categorisation results are consistent with other studies conducted in industrial river basins, where fuzzy logic-based methods have also been applied to assess the level of water contamination [46–47]. However, by adding expert-based fuzzy rules that take into consideration local hydrogeochemical circumstances, MFISM improves interpretability in contrast to traditional approaches.

With 25 fuzzy rules (Fig. S17), the model's rule-based design can effectively differentiate between various water quality classes. Interestingly, regions with high NPI and Igeo values correlate with poor water quality, confirming the impact of geogenic and industrial sources on pollution. To evaluate the effectiveness of fuzzy logic in evaluating intricate hydrochemical interactions and to ascertain the water quality of river basins like the Ganga and Yamuna, other researchers have used fuzzy inference algorithms [53–54]. The study's findings emphasise the significance of region-specific water management, especially in densely populated and industrialised areas. This study offers a thorough grasp of water quality changes by combining MFISM with spatial analysis, helping policymakers create efficient, focused mitigation strategies [55].

In the defuzzification step, the DWQ score ranges from 0 to 125. A lower fuzzy score, particularly below 50, indicates better drinking water quality (Table S2). An essential aspect of model development involves assessing the agreement between the results derived from MFISM and expert knowledge. This entails ensuring that the system yields appropriate responses under various conditions that may be introduced. Several examples illustrating the operational model of the MFISM are depicted in Fig. S18 and S19. Similarly, the results for all the locations are obtained and presented in Table S1. Khadka and Khanal [56] conducted a study on the Environmental Management Plan (EMP) for the Melamchi Water Supply Project in Nepal, emphasizing the impact of seasonal variations on water quality. Their findings demonstrated that water quality parameters fluctuate significantly due to climatic conditions and anthropogenic influences, necessitating continuous monitoring and adaptive management strategies. Muller [14] conducted a study on heavy metal pollution in sediments of the Rhine River, Germany. This study introduced the Geoaccumulation Index (Igeo) as a quantitative measure for assessing metal contamination in sediments. The classification system developed in this study categorizes pollution levels into different classes, ranging from unpolluted to extremely polluted, based on metal concentration relative to background levels. The Igeo index has since been widely applied in environmental studies worldwide to evaluate sediment contamination. Our study employs this index to assess heavy metal pollution in the Palar River Basin, demonstrating its applicability in understanding sediment quality and pollution severity [57].

Piper diagrams depicting anions and cations for the post-monsoon and pre-monsoon periods are shown in Fig. S20a and S21a During the post-monsoon period, most surface water samples clustered within the 1^st field, indicating the dominance of calcium and bicarbonate ions, which are typical of waters influenced by the break-down of carbonate minerals such as calcite and are common in limestone and dolomite rocks. These waters often have relatively high pH levels and are associated with carbonate aquifers. A minority of samples fell within the 3^rd field, indicating elevated concentrations of calcium, magnesium, and bicarbonate ions, suggesting the dissolution of both carbonate and silicate minerals in areas with diverse lithologies. Some samples extended toward the 4^th field, indicating increasing sodium and chloride ions, which are typical of saline water compositions, suggesting saline water intrusion and anthropogenic contamination. The Durov diagram is a valuable tool for interpreting hydrochemical processes in surface water systems, offering insights into ion exchange, mixing of different water types, and reverse ion exchange. Major ions are represented as percentages of milli-equivalents in two triangular plots, with total anions and cations summing to 100%. By projecting data points onto a square grid perpendicular to the third axis in each triangle, the diagram helps identify similar water compositions. During the post-monsoon period (Fig. S20b), surface waters showed transitional compositions, predominantly between calcium bicarbonate and magnesium bicarbonate waters, indicating interactions with magnesium-rich minerals such as dolomite. Some samples fell into the calcium and bicarbonate water category, which is typical of regions with calcium-rich minerals such as limestone. During the pre-monsoon season (Fig. S21b), most samples had balanced concentrations of major ions, reflecting interactions with various rock types in the Vellore district. Fewer samples showed high bicarbonate concentrations with calcium and magnesium, characteristic of bicarbonate waters from carbonate rocks such as limestone or dolomite. Samples with high sulfate, sodium, and magnesium concentrations suggested interactions with anthropogenic sources such as industrial activities. The current study provides a complete hydrochemical analysis of surface water in the Vellore district area, emphasizing seasonal fluctuations utilizing Piper and Durov diagrams. The study observed a larger concentration of calcium and bicarbonate ions during the post-monsoon period, which is more accurately seasonal differences and similar to the findings of Srinivasamoorthy et al. [65].

TDS concentrations in the study area normally range between 100 and 1000 mg/L, according to the Gibbs diagrams in Fig. 7 suggesting that hydrochemistry is primarily influenced by water-rock interactions. Cl⁻/Cl⁻ + HCO³⁻ and Na⁺/Na⁺ + Ca²⁺ ratios imply subsurface accumulation and Na⁺ dissolution. The majority of samples plot in the rock dominance zone of binary ion plots, which emphasise geochemical processes, including silicate weathering and cation exchange. From north to south, Cl⁻ in relation to HCO³⁻ increases. By differentiating between areas where rock weathering and evaporation predominate and offering vital information on hydrogeochemical processes as well as cation absorption and release, pre-monsoon Gibbs diagrams illustrate the influence of geological and climatic conditions on water chemistry [66]. Gibbs diagrams have been used in earlier research, albeit with differing degrees of detail, to identify hydrogeochemical processes in various Indian river basins. Without doing a thorough ionic ratio analysis, Kumar et al. [50], concluded that evaporation and rock-water interactions were the main controls in their study of the hydrochemistry of the Yamuna River Basin. The Cauvery Basin's groundwater was also examined by Ramya et al. [40], who identified weathering processes but did not evaluate seasonal fluctuations or patterns of regional distribution. By analysing TDS concentrations between 100 and 1000 mg/L and connecting them to subsurface accumulation and Na⁺ dissolution in minerals, the current study, in contrast, offers a more thorough analysis. To illustrate geological processes like cation exchange and silicate weathering, this study also uses binary ion plots and concentrates on particular ion ratios, such as Na⁺/Na⁺ + Ca²⁺ and Cl⁻/Cl⁻ + HCO³⁻. This study differentiates between rock weathering and evaporation dominance regions seasonally (pre- and post-monsoon) and geographically (north to south gradients), in contrast to other research that mostly concentrated on broad classifications. A thorough grasp of the geological, climatic, and human influences on water chemistry is also provided by the combination of fuzzy logic and GIS-based spatial distribution analysis, which improves the interpretation of Gibbs diagrams [43].

In this study, PCA was employed to reduce the dimensionality of the dataset and identify underlying patterns among the physicochemical parameters and heavy metals. PCA is particularly effective in simplifying complex datasets by highlighting the most significant factors influencing water quality. By analyzing variances, PCA allows for the identification of key parameters, such as pH and bicarbonates (HCO_3⁻), which contribute the most to water quality variations. In our study, PCA revealed that the top three principal components explained a substantial portion of the total variance—64.1% during the post-monsoon season and 53.8% during the pre-monsoon season. This reduction in complexity facilitates a more focused interpretation of the data, making it easier to implement targeted water management strategies in the Palar River Basin. The PCA results highlighted that pH and HCO_3⁻ were the most influential parameters among the second and third factors, respectively. These findings underscore the complex interplay between natural processes—such as evaporation, salt domes, weathering, and soil erosion—and human activities like agro-field fertilization and domestic/industrial waste disposal, all of which shape water quality. Fig. S22 illustrates the concentrations of water quality elements across various monitoring stations, reflecting the combined influence of geological attributes, anthropogenic factors, and climatic conditions on the surface water quality of the study area [51]. Our findings are compared with those of previous studies conducted in industrial river basins. Consistent with our results, Khan et al. [43] found that PC1 accounted for 67% of the variance in their analysis of the Ganga River Basin, with sulphate, TDS, and EC as the dominant contributing factors. Their study also reported that bicarbonate and pH were the main factors influencing PC2, indicating similar hydrogeochemical controls in semi-arid river systems. In contrast, Kumar et al. [50] observed a higher contribution of heavy metals, including Chromium (Cr), Nickel (Ni), and Zinc (Zn), in PC1 in a highly industrialized river basin, suggesting widespread contamination. However, our study shows that heavy metals exhibited a lower variance in PCA, indicating that their distribution is more localized rather than affecting the entire basin. These differences highlight the varying degrees of industrial and agricultural impacts across different study sites.

Pearson’s correlation matrix, depicted in Fig. S23, reveals varying degrees of association among surface water metal parameters. Correlation reflects the reciprocal relationship between two variables; when one parameter increases, the other increases [44]. During the post-monsoon period, strong correlations are observed between TH and Cl⁻ (r = 0.95), between sulphate and Cl⁻ (r = 0.78), and between sulphate and TDS (r = 0.79). Additionally, strong correlations are found between TDS and Cl⁻ (r = 0.87), between EC and Cl⁻ (r = 0.87), and between Ca²⁺ and TA (r = 0.75). Moderate correlations included Magnesium (Mg²⁺) with TDS (r = 0.50) and EC (r = 0.46), Na⁺ with SO₄²⁻ (r = 0.43), nitrate with TH (r = 0.39), and TA with turbidity (r = 0.39). Conversely, pH exhibited negative correlations with TDS, EC, Nitrate, Hardness, Alkalinity, and Calcium. Dissolved oxygen (DO) and Ammonia (NH₃) showed weak correlations with all the parameters.

In the Pre-Monsoon period, strong correlations included Cl⁻ with TH (r = 0.89), EC (r = 0.72), and TDS (r = 0.83), as well as between SO₄²⁻ and TDS (r = 0.98), EC (r = 0.80), and TH with EC (r = 0.84), TDS (r = 0.87). Moderate correlations are observed between TA and TDS (r = 0.53), between Mg²⁺ and TDS (r = 0.59), between TH (r = 0.46), between Na⁺ and TDS (r = 0.41), between EC (r = 0.42), between Mg²⁺ (r = 0.44), between K⁺ and TDS (r = 0.59), between TH (r = 0.50), and between TA (r = 0.56). Similar to the post-monsoon period, pH had negative correlations with several parameters and weak correlations with turbidity, Mg²⁺, Na⁺, K⁺, NH³, and DO. Mg²⁺ exhibited weak correlations with turbidity (r = 0.14) and TA (r = 0.22), while NH₃ showed weak correlations with all other parameters. Most heavy metals displayed weak correlations with each other across both periods. Weak or negative correlations, on the other hand, suggest independent behaviors of parameters, which is useful for isolating the impact of specific contaminants. The use of Pearson’s correlation matrix strengthens the findings by offering a comprehensive understanding of interdependencies, which is vital for making informed decisions about water quality management . The higher percentage of variance explained by the PCA (64.1%) indicates that the post-monsoon data has clearer and more distinct patterns in surface water quality parameters. This suggests that the data is less complex and more homogenous, where it explains clearer and more interpretable data patterns [45]. Our correlation trends align with previous studies on industrialized river basins. Ramya et al. [40], in their study of the Cauvery River Basin in India, reported high correlations between total hardness, EC, and chloride, similar to our findings. Likewise, Ramachandran et al. [46], in their study of the Adyar River, Chennai, observed strong sulfate-TDS relationships, suggesting similar contamination sources from industrial and urban effluents. However, our study contrasts with the findings of Kaur et al. [47], who examined the Yamuna River Basin and found strong correlations between heavy metals, including Cr, Ni, and Zn, indicative of widespread pollution. In contrast, we observed weaker correlations between heavy metals, suggesting that heavy metal contamination in the Palar River Basin is more localized rather than basin-wide [67]. The high variance explained by PCA, along with strong correlations among ionic parameters, underscores the significant role of anthropogenic activities—such as industrial discharge, agricultural runoff, and domestic wastewater—in influencing water quality in the Palar River Basin. Additionally, seasonal variations in correlation trends highlight the importance of year-round monitoring to account for hydrochemical shifts. The integration of PCA and correlation analysis provides a robust framework for identifying key water quality drivers, ultimately aiding in the development of targeted mitigation strategies.