Assessment of water quality and microbial contamination in institutional water resources: a necessity to understand health risks Authors: Rekha Kumari, Chirashree Ghosh, Ritesh Kumar, Rashmi Shakya, Shailender Kumar, Adesh K. Saini Published: 2026 ABSTRACT Lack of regular monitoring of water sources may lead to undetected contamination, posing serious health risks and necessitating regular water quality assessments. Sampling for physicochemical, microbial analyses, and online surveys across three higher education institutions was done to evaluate water quality. Spatiotemporal variations among physicochemical parameters showed that the pH, EC, and TDS decreased during the wet season, reflecting the dilution effect of rain. However, DO increased from 0.67 to 4.83 ppm, indicating better aeration. PCA showed seasonal variability, whereas the correlation matrix highlighted both positive and negative interrelationships between temperature-pH (− 0.25), DO-ORP (0.11), and TDS-EC (1.00). Potentially toxic metals were either negligible or not detected. Metagenomics revealed the presence of 29 bacterial phyla, 61 classes, 124 orders, 241 families, and 457 genera. Canonical correspondence analysis showed the influence of Mo, EC, salinity, and TDS on Bacteroidota, Chloroflexota, Cyanobacteriota, and Planctomycetota, whereas Verrucomicrobiota, Acidobacteriota, Chlamydiota, Candidatus Melainabacteria, Bdellovibrionota, and Deinococcota were affected by Ni, pH, and COD. Pathogen mapping revealed the presence of Vibrio, Pseudomonas, Enterobacter spp., etc., responsible for diseases such as cholera, diarrhea, and typhoid. Also, occupants’ perception about the water quality emphasizes the need for better management of drinking water in HEIs.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-53672-4. Introduction Freshwater is essential for sustaining life, but only 1% of Earth’s surface water is potable and has a profound impact on health. However, escalating global populations, rapid urbanization, and anthropogenic activities are continuously challenging clean water availability1. Both point (industrial effluents and sewage discharge) and nonpoint sources (agricultural runoff) play a significant role in degrading water quality2. Also, geochemical processes can lead to natural contamination of groundwater (GW) with potentially toxic elements (PTEs) like fluoride (F), arsenic (As), etc.3. World Health Organization (WHO) estimates that ~ 80% of illnesses stem from exposure to polluted water, a key medium for the spread of pathogens4,5. Insufficient potable water increases the risk of transmission for waterborne diseases, respiratory infections, and Neglected Tropical Diseases (NTDs)6. The UN General Assembly declared access to clean water as a fundamental right, but inadequate availability and pollutants pose significant health challenges in developing countries7. For instance, National Institution for Transforming India (NITI) Aayog’s Composite Water Management Index (2019) highlights that ~ 600 million people in India face high to extreme water stress due to contaminated water resources (~ 70%)8. According to United Nations Children’s Fund (UNICEF) and Water, Sanitation, and Hygiene (WASH), ~ 35% of Nigerians are deprived of basic drinking water, with rural communities facing greater challenges (~ 43%) compared to urban settings (24%)9. The ‘academic revolution’ over the past two decades has led to more than 40% of the world’s population being engaged in higher education institutions10. Availability of clean drinking water is a must, and therefore, regular quality checks become a priority. There are many recent reports where outbreaks of water borne diseases due to biological contaminants have been reported from various educational institutions or other public settings worldwide. These studies showed presence of microbial contaminations and toxic metal like arsenic present above the permissible limits11–14. Inability to monitor the water quality regularly can have deleterious impacts on the occupants’ health. For instance, in Flint, Michigan, Pb contamination caused severe public health issues, particularly affecting children’s cognitive development15. Higher levels of As and Pb in the Ganges, India, were associated with cancer and neurological disorders. GW contamination with excessive F in Rajasthan, India, led to widespread fluorosis16,17. Microbial contamination, especially in developing countries, causes various waterborne diseases, e.g., E. coli can lead to gastrointestinal infections and neurological disorders, further exacerbated by pH, temperature, and TDS18–20. Monitoring of the physical characteristics (color, taste, odor, etc.), chemical parameters (pH, DO, PTE, etc.), and microbial diversity (CFU count, metagenomic analysis) is critical in the assessment and can provide a comprehensive understanding of water quality21,22. Investigating these parameters helps in identifying pollution sources, tracking spatiotemporal variations, and implementing remediation strategies. The interrelationships between these parameters may determine the overall health of water systems, e.g., a decline in DO often correlates with increased BOD and COD, indicating organic pollution and microbial activity23. Similarly, pH variations can influence the solubility of heavy metals, affecting bioavailability and toxicity. ORP can indicate the presence of oxidizing agents influencing metal speciation and microbial activity24,25. Therefore, understanding the relationships among physicochemical parameters and microbial contamination is essential for effective water quality management. In large institutional settings, diverse water sources, such as ground, tap, pond, and harvested rainwater, are utilized for various purposes, affecting health directly or indirectly26. Therefore, it is imperative to maintain the water quality and adopt appropriate management strategies. Although mapping of pathogenic microorganisms is of utmost importance, unlike physicochemical screening, detection of microbial contamination poses challenges due to cost, expertise, and time27,28. Despite the immense importance of water resources in institutions, there are limited studies assessing the microbial diversity and pathogen mapping. Detection of microbes can be done qualitatively and semi-quantitatively by the CFU method, but metagenomics is critical for the identification and quantification of microbes present in the environmental DNA (eDNA). Amplicon sequencing (16S rRNA) is one of the widely used techniques for microbial identification, diversity, and abundance29,30. Hence, the present cross-sectional study was taken up keeping in mind two major problems i.e. Outbreak of water-borne diseases in the institutional areas and requirement of the UN Sustainable Development Goal 6 (SDG 6) goals to achieve accessibility of safe water to all31–33. This study involved a thorough analysis of the physicochemical characteristics and microbial diversity of water sources from three higher education institutions (HEIs) in Delhi, India. Additionally, online surveys with meticulously designed questionnaires were employed to understand the level of awareness about the water quality among the occupants. The present study is one of the first which has focused on physicochemical and biological contamination (metagenomics) and their co-relation and inter-relationships in impacting the water quality of any source simultaneously. This study may help in filling a few gaps in the previous studies to understand the water pollution and people perception about the clean water accessibility by 2030 for all which is an important aspect mentioned under the SDG 634. The findings may provide important insights about the water quality and related health hazards and might be instrumental in developing effective strategies for improving quality to safeguard public health in densely populated urban centers. Methodology The methodology involved a multifaceted approach combining on-site physicochemical analysis, PTE profiling, and detection of microbial contaminants to assess water quality. Online surveys were designed to evaluate people’s perception of water quality and safe water availability, especially in terms of waterborne diseases. All methods were carried out in accordance with relevant guidelines and regulations. The present study carried out sampling in two seasons based on humidity and precipitation, i.e., dry (January-March; humidity < 60%) and wet (August-October; humidity > 75%). The data was analyzed to evaluate the water quality, its relationship to microbiological pollutants, and health hazards to occupants. Water sampling Sampling locations were strategically selected across Delhi (Site-I, Site-II, and Site-III) to understand water quality status. Criteria for site selection included accessibility and availability of tap, ground, and pond water. The study site map was created using QuantumGIS (QGIS) version 3.40.14 “Bratislva” with plugin Open Street Map (OSM) and geographical coordinates of each site are detailed in Supplementary Fig. 1 (https://qgis.org/downloads/qgis-3.40.14.tar.bz2). Three samples from each source and site were collected every month for both the seasons. A total of nine samples from each site monthly in both seasons were collected. Four liters of water for both physicochemical and metagenomic studies were collected. For the metagenomic DNA isolation, water samples from all sites and the same source were pooled. Institutional permissions were obtained for access to the sites and sample collection. Water samples were collected from various institutional sources between 8–9 AM across Delhi (28.676314° N, 77.224977° E), following standard protocols for water sampling and preservation35,36. Samples were transported to the laboratory under controlled conditions for immediate determination of COD, metal fixation, CFU counting, and metagenomic analysis, while some part of the samples was also stored at 4 °C for subsequent procedures. Physicochemical characterization In situ physicochemical parameters like pH, temperature, TDS, EC, ORP, and DO were determined using a multi-parameter water quality analyzer (Hanna HI98194, Italy). COD analysis for both seasons was performed in the laboratory using the titrimetric method37. The single-source samples from dry and wet seasons were pooled separately to detect the presence of PTEs using an absorption-based method, i.e., ICP-MS (AGT/ICP-MS/005, US)38,39. All samples were taken in triplicate for analysis. Water quality index The Water Quality Index (WQI) was calculated in multiple steps, beginning with the selection of key physicochemical parameters based on their environmental significance. The WQI was calculated using a modified NSF-WQI approach based on the physicochemical parameters that were consistently measured across all sites and seasons: pH, electrical conductivity (EC), total dissolved solids (TDS), dissolved oxygen (DO), and COD. The quality rating scale (qi) for each parameter was determined using the formula qi = (Vi/Si) × 100, where Vi is the measured value and Si is the standard permissible value as per WHO/BIS guidelines40. Each parameter was assigned a weight (wi) reflecting its relative importance, with the sum of all weights normalized to 1.0. The weighted sub-index (Wi) for each parameter was then computed using the equation, Wi = qi × wi and the overall WQI was obtained by summing the weighted sub-indices41–43. Parameter weights followed the NSF weighting logic (weights inversely proportional to standard values), and no additional variables were included. The index was deliberately restricted to measured variables as they were used throughout the study for other analysis to maintain uniformity in other statistical analysis also. Water quality of various sources was inferred using the National Sanitation Foundation (NSF) index as the standard. As per NSF standards, WQI range 0–25, 26–50, 51–75, 76–90 and 91–100 are considered to be poor, fair, good, very good and excellent, respectively. Microbial contamination analysis Quantitative analysis using CFU method Colony forming units per milliliter (CFUs/ml) were counted using the spread plate method. Spread plating was performed by inoculating the sample on the Nutrient Agar (NA) medium. Serial dilutions of 100 to 10−4 were plated on NA medium to enumerate CFUs. Cells were incubated at 37 °C overnight, and bacterial colonies were counted after 24 h44. Metagenomic DNA isolation, sequencing and analysis Metagenomic DNA was isolated from samples using a modified cetyltrimethylammonium bromide (CTAB) method to improve DNA yield45. The integrity of DNA samples was checked on 0.8% agarose gel. The concentration (260 nm) and purity (260/280 nm) of the isolated DNA were assessed using a NanoDrop® spectrophotometer (Thermo Scientific One, USA). DNA amplicon sequencing The metagenomic DNA samples for a particular source were pooled and named as GW, TW, and PW. All pooled DNA samples were subjected to amplicon sequencing targeting the V3-V4 variable region of the bacterial 16S rRNA gene. Sequencing was performed on Illumina HiSeq using a 2 × 300PE V3-V4 sequencing kit. Prior to this, the amplicons of each sample were purified with AMPure XP beads to remove unused primers. An additional 8 cycles of PCR were performed using Illumina barcoded adapters to prepare the sequencing libraries. Libraries were purified using AMPure XP beads and quantitated using the Qubit dsDNA High Sensitivity assay kit protocol as per the manufacturer’s instructions. The metagenomic datasets generated are available in the SRA repository at NCBI servers with accession number PRJNA1461976. DNA sequence data analysis The raw sequence data produced by Illumina sequencing was analyzed using the QIIME 2 version 2024.10 bioinformatics pipeline in a Conda environment46. The demultiplexed raw sequence data was filtered by removing the primers and adapters. Sequence reads with an average call quality of 0.6 for each base between 20 and 40 bp were retained for downstream analysis. DADA2 (via q2-dada2) was used for quality filtering. Denoising of the raw sequence data was executed by estimation of the error profile of each possible transition and identification of amplicon sequencing variants (ASVs). Taxonomy was assigned to each ASV using the ‘classify-sklearn’ command in q2-feature-classifier based on the V3-V4 regions. This algorithm is a pre-trained Naive Bayes machine-learning classifier, trained on the 99% Greengenes 13_8 reference database; feature tables representing ASV counts per sample were generated in QIIME 2. Differentially abundant bacterial taxa at the genus or species level were identified using ANCOM-BC (Analysis of compositions of microbiomes with bias correction) via q2-composition47. Data were visualized directly as QIIME 2 artifacts in QIIME 2 View and through ggplot2 (v3.3.1), using data extracted from QIIME 2 artifacts via qiime2R (v0.99.6; https://github.com/jbisanz/qiime2R/). Taxonomic assignments, together with the variant abundance table, were used to calculate diversity indices and expressed as relative abundance (RA%). The community profile was also analyzed through core microbiome analysis to identify the main taxa present in different samples based on their prevalence and RA. Pathogen mapping To identify pathogenic bacteria at the genus level, a BLASTn search was conducted for all ASVs against the NCBI non-redundant nucleotide (nr) database. Taxonomic assignment was based on best-hit sequences (100% query identity and 99% coverage) and then compared with the bacterial pathogen list reported by KARIUS® (https://kariusdx.com/karius-test/pathogens) and the NCBI bacterial pathogen list to determine the genus of the pathogen. The RA of pathogens was calculated according to the ASV table. Population-based survey to assess water quality An online survey was conducted to assess the perception of water quality, including use of water and health implications, among institutional populations. The online survey was a volunteering exercise. However, an informed consent was taken from the participants at the beginning of the survey via google form. The study objectives and the purpose were clearly stated in the consent. All the experimental protocols, especially the online survey questionnaire, were approved by the institutional review committee (IRC) under the R&D cell, Miranda House, University of Delhi, and an ethical waiver was given via reference no. MH-32/2024/IRC-3/10–02-2024. The survey constituted four sections as follows:Primary source of drinking water: Institutional RO (filtered water), own bottles, and others.Water quality concerns and observations: Chemical and microbial contamination, non-functional ROs, change in taste, periodic cleaning issues, odor, etc.Awareness: SDGs, WQI, sanitation, and hygiene.Health Impacts: waterborne diseases. Primary source of drinking water: Institutional RO (filtered water), own bottles, and others. Primary source of drinking water: Institutional RO (filtered water), own bottles, and others. Water quality concerns and observations: Chemical and microbial contamination, non-functional ROs, change in taste, periodic cleaning issues, odor, etc. Water quality concerns and observations: Chemical and microbial contamination, non-functional ROs, change in taste, periodic cleaning issues, odor, etc. Awareness: SDGs, WQI, sanitation, and hygiene. Awareness: SDGs, WQI, sanitation, and hygiene. Health Impacts: waterborne diseases. Health Impacts: waterborne diseases. Statistical analyses of data The statistical analyses in the study involved a range of methods to evaluate WQIs, PCA, and correlation matrices among various physicochemical parameters. To assess the distribution pattern of the microbial community, PCA and NMDS were carried out. The interdependencies among water quality parameters and microbial communities were studied using CCA and Bivariate Pearson’s correlation. For statistical analyses PAST 4.17, R (v4.4.1) with R Studio (v2024.12.0 + 467), and Excel365 were used whereas data visualization of correlation maps and graphs was performed using the vegan and ggplot2 packages. Results Spatiotemporal variations of physicochemical characteristics Spatiotemporal variations and the impacts of anthropogenic activities on the water quality were evident in the physicochemical parameters, and their patterns across dry and wet seasons are depicted in Table 1 and Fig. 1, respectively. The pH levels across various sources of water were higher during the dry season (7.30–8.075) as compared to the wet season (6.14–6.77). EC values were lower in GW (448.28 µS/cm at site III) in the dry season as compared to the wet season (12,435 µS/cm at site I and 6435 µS/cm at site III), but TW showed higher EC values during the dry season (2430 µS/cm at site I and 4221.70 µS/cm at site III). TDS values in TW decreased from 2119.99 (dry season) to 647 mg/L (wet season). DO levels in the dry season across all sites (lowest 0.60 ppm in TW at site III) increased during the wet season, reaching within the acceptable range (6.7 ppm in GW at site III). Temperature showed a distinct seasonal pattern, with higher temperatures recorded during the wet season, up to 29.0 °C, and dropping to 19.2 °C (PW at site I) during the dry season. A significant decline was also recorded in COD levels across water sources and sites during the wet season. For instance, in GW, COD levels were highest at site I (30.66 mg/L) during the dry season but dropped drastically (3.6 mg/L) during the wet season. TW also followed this pattern, with site I recording 27.6 mg/L in the dry season, reducing to 1.6 mg/L in the wet season. In PW, COD levels at sites I and II were 18.72 mg/L and 25.6 mg/L in the dry season, dropping to 9.6 mg/L and 14.4 mg/L, respectively, in the wet season.Table 1Physicochemical parameters across different water sample sites and sources.SitesSalinity (psu)DO (ppm)COD (mg/L)pHEC (µS/cm)Temperature (°C)TDS (mg/L)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Site—ITW1.285 ± 0.3781.35 ± 0.051.297 ± 0.4706.21 ± 0.11827.6 ± 13.21.6 ± 0.007.46 ± 0.1756.4 ± 0.4132430.458 ± 843.080735.625 ± 735.62524 ± 3.33429 ± 0.8351577 ± 349.446647 ± 317.063GW2.408 ± 0.6343.8 ± 3.101.035 ± 0.4895.61 ± 0.53930.66 ± 12.4823.6 ± 2.807.3 ± 0.2176.32 ± 0.1434467.846 ± 1099.47712,435 ± 155.0025 ± 1.35429.26 ± 1.2252233.996 ± 549.7294685.6 ± 1760.377PW1.28 ± 0.0881.5 ± 0.101.17 ± 0.4165.04 ± 0.37418.72 ± 7.9549.6 ± 8.007.46 ± 0.2386.77 ± 0.4332411.324 ± 151.2863090 ± 60.0019.17 ± 3.29729.13 ± 1.2391205.569 ± 75.6481104.333 ± 336.770Site—IITW0.218 ± 0.0161 ± 0.801.067 ± 0.5256.7 ± 0.09726.4 ± 10.21411.2 ± 0.007.54 ± 0.1066.26 ± 0.174435.017 ± 68.9811532 ± 980.17221.1 ± 3.39129.23 ± 1.235360.872 ± 135.6451011.1 ± 629.775GW1.652 ± 0.1711.7 ± 0.601.204 ± 0.4856.22 ± 0.40723.2 ± 4.1057.2 ± 0.807.56 ± 0.3516.6 ± 0.3513445.483 ± 312.074425 ± 425.0026 ± 1.47929.1 ± 1.2061862 ± 138.3191082.933 ± 509.862PW0.61 ± 0.0670.45 ± 0.051.37 ± 0.4465.66 ± 0.34125.6 ± 9.81114.4 ± 1.607.84 ± 0.3856.75 ± 0.2621469.619 ± 470.081907.333 ± 158.56620.2 ± 3.54329.26 ± 1.320728.110 ± 179.038404.667 ± 61.699Site—IIITW2.268 ± 0.5830.2 ± 0.000.598 ± 0.1056.5 ± 0.18115.46 ± 8.7194.8 ± 1.607.39 ± 0.4746.53 ± 0.2724221.706 ± 1037.872382.66 ± 43.10626.3 ± 3.10629.26 ± 1.1202119.996 ± 528.4291156.56 ± 1001.753GW0.928 ± 0.7083.25 ± 0.051.286 ± 0.7826.7 ± 0.04616.8 ± 10.7186.4 ± 0.008.075 ± 0.1496.14 ± 0.149448.284 ± 116.5706435 ± 305.0030 ± 3.89829.3 ± 1.026883.569 ± 656.9251110.133 ± 950.148The data denote mean ± standard error of the mean.Fig. 1Multivariable graphs showing seasonal (dry vs. wet) and spatial variation in physicochemical parameters across Sites I–III for TW and GW and PW at Site I and II. Table 1Physicochemical parameters across different water sample sites and sources.SitesSalinity (psu)DO (ppm)COD (mg/L)pHEC (µS/cm)Temperature (°C)TDS (mg/L)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Dry (SE)Wet (SE)Site—ITW1.285 ± 0.3781.35 ± 0.051.297 ± 0.4706.21 ± 0.11827.6 ± 13.21.6 ± 0.007.46 ± 0.1756.4 ± 0.4132430.458 ± 843.080735.625 ± 735.62524 ± 3.33429 ± 0.8351577 ± 349.446647 ± 317.063GW2.408 ± 0.6343.8 ± 3.101.035 ± 0.4895.61 ± 0.53930.66 ± 12.4823.6 ± 2.807.3 ± 0.2176.32 ± 0.1434467.846 ± 1099.47712,435 ± 155.0025 ± 1.35429.26 ± 1.2252233.996 ± 549.7294685.6 ± 1760.377PW1.28 ± 0.0881.5 ± 0.101.17 ± 0.4165.04 ± 0.37418.72 ± 7.9549.6 ± 8.007.46 ± 0.2386.77 ± 0.4332411.324 ± 151.2863090 ± 60.0019.17 ± 3.29729.13 ± 1.2391205.569 ± 75.6481104.333 ± 336.770Site—IITW0.218 ± 0.0161 ± 0.801.067 ± 0.5256.7 ± 0.09726.4 ± 10.21411.2 ± 0.007.54 ± 0.1066.26 ± 0.174435.017 ± 68.9811532 ± 980.17221.1 ± 3.39129.23 ± 1.235360.872 ± 135.6451011.1 ± 629.775GW1.652 ± 0.1711.7 ± 0.601.204 ± 0.4856.22 ± 0.40723.2 ± 4.1057.2 ± 0.807.56 ± 0.3516.6 ± 0.3513445.483 ± 312.074425 ± 425.0026 ± 1.47929.1 ± 1.2061862 ± 138.3191082.933 ± 509.862PW0.61 ± 0.0670.45 ± 0.051.37 ± 0.4465.66 ± 0.34125.6 ± 9.81114.4 ± 1.607.84 ± 0.3856.75 ± 0.2621469.619 ± 470.081907.333 ± 158.56620.2 ± 3.54329.26 ± 1.320728.110 ± 179.038404.667 ± 61.699Site—IIITW2.268 ± 0.5830.2 ± 0.000.598 ± 0.1056.5 ± 0.18115.46 ± 8.7194.8 ± 1.607.39 ± 0.4746.53 ± 0.2724221.706 ± 1037.872382.66 ± 43.10626.3 ± 3.10629.26 ± 1.1202119.996 ± 528.4291156.56 ± 1001.753GW0.928 ± 0.7083.25 ± 0.051.286 ± 0.7826.7 ± 0.04616.8 ± 10.7186.4 ± 0.008.075 ± 0.1496.14 ± 0.149448.284 ± 116.5706435 ± 305.0030 ± 3.89829.3 ± 1.026883.569 ± 656.9251110.133 ± 950.148The data denote mean ± standard error of the mean. Physicochemical parameters across different water sample sites and sources. The data denote mean ± standard error of the mean. Multivariable graphs showing seasonal (dry vs. wet) and spatial variation in physicochemical parameters across Sites I–III for TW and GW and PW at Site I and II. PTEs analysis PTE analysis was performed to detect the presence of heavy metals (Cr, Ni, As, Mo, Ag, Cd, Hg, Pb) in different water sources. The concentration of Ni, As, Mo, Cd and Pb was insignificant (< 0.0001 ppm) whereas Cr, Hg and Ag were even not detectable (ND) (Supplementary Table 1). Interrelationships of physicochemical parameters The correlation analysis showed interdependencies among physicochemical parameters, critical for understanding the interactions among water quality parameters and their collective influence on water chemistry. Temperature was negatively correlated with pH (− 0.25), whereas DO showed a slightly positive correlation with ORP (0.11). Both TDS and EC showed a perfect relationship with each other (1:00) and were positively correlated with DO (0.16) (Supplementary Fig. 1). Principal Component Analysis (PCA) was applied to the water quality data to reduce dimensionality and identify patterns while preserving key variability. The PCA distinguished seasonal and temporal water quality patterns across various water sources (Fig. 2). PC1 captured 38.9% variation in water quality, whereas PC2 captured 23.6% variation. EC, TDS, salinity, DO, and temperature showed increased values during the wet season. ORP, pH, and COD values were associated with the dry season, showing higher values in dry conditions. TW and GW samples showed greater sensitivity to seasonal changes (Fig. 3).Fig. 2Biplots of Principal component analysis showing the influence of various physicochemical parameters on the water quality. Also, a change during the seasons is depicted for all sampling sites.Fig. 3(a) Venn diagram showing shared total composition of bacterial community (core ASVs) at phylum, family, class, order and genus level in all the samples. (b) Violin plot showing the distribution of the community at phylum, family, class, order and genus level in all the samples. Biplots of Principal component analysis showing the influence of various physicochemical parameters on the water quality. Also, a change during the seasons is depicted for all sampling sites. (a) Venn diagram showing shared total composition of bacterial community (core ASVs) at phylum, family, class, order and genus level in all the samples. (b) Violin plot showing the distribution of the community at phylum, family, class, order and genus level in all the samples. WQI analysis The WQI values of different water samples across the sites revealed distinct seasonal trends (Supplementary Table 2). The WQI of GW decreased from 35.65 (Fair) in the dry season to 28.54 (Fair) in the wet season. TW showed a marked improvement, with an average WQI rising from 42.00 (Fair) in the dry season to 68.23 (Good) in the wet season, and for PW, it decreased slightly from 43.80 (Fair) to 39.48 (Fair). Assessment of microbial contamination and pathogen mapping Microbial contamination was checked quantitatively by the CFU method. CFU exhibited variability across the water sources; however, the differences were insignificant. The CFU counts (log CFU/ml) in the water samples ranged from 3.7 to 4.8 in GW, 3.7 to 5.5 in TW, and 3.8 to 6.7 in PW. Composition of microbial communities at various taxonomic levels The RA of bacteria in water samples was analyzed up to the genus level, identifying 29 bacterial phyla, 61 classes, 124 orders, 241 families, and 457 genera. The detailed analysis of taxonomic classification at the phylum, class, order, family, and genus level of each ASV is provided in Supplementary Table 3. Phylum level taxonomy The phylum-level taxonomic composition based on RA ≥ 1% in each sample is detailed in Fig. 4. A total of 29 bacterial phyla were detected cumulatively and showed high similarity (~ 75%) among GW, TW, and PW, representing 25, 26, and 24 phyla, respectively. The top 10 bacterial phyla present in all samples account for > 95% of total taxonomically assigned reads (Fig. 4a). Actinomycetota (70.4%) was the abundant and prevalent phylum in GW, Pseudomonadota (37.2%) in PW, whereas both Pseudomonadota and Bacteroidota (37.1% each) were dominant in TW. Four phyla, namely Deinococcota, Bacteroidota, Gemmatimonadota, and Nitrospirota, exhibited RA of 1–5%, whereas 7 phyla showed < 1%. Phyla Acidobacteriota, Chlamydiota, and Candidatus Melainabacteria were found to be significant only in PW.Fig. 4Relative abundance of Bacterial community compositions at the phylum, class, order, family and genus level. ASV for each level made up < 1% of total composition in at least 1 sample and remaining were classified as “others”. Relative abundance of Bacterial community compositions at the phylum, class, order, family and genus level. ASV for each level made up < 1% of total composition in at least 1 sample and remaining were classified as “others”. Class level taxonomy The 10 most frequently detected classes were Actinomycetes (44.8%), Flavobacteriia (29.0%), Coriobacteriia (25.44%), Betaproteobacteria (25.15%), Gammaproteobacteria (16.19%), Phycisphaerae (12.11%), Alphaproteobacteria (10.35%), Verrucomicrobiae (9.55%), Clostridia (9.5%), and Chlamydiia (8.68%) (Fig. 4b). GW showed dominance of Actinomycetes (44.8%), Coriobacteria (25.4%), Clostridia (9.5%), Flavobacteriia (4.3%), Betaprotobacteria (3.6%) and the exclusive presence of Actinomycetes, Coriobacteria, and Clostridia. TW harbored Flavobacteriia (29.0%), Betaproteobacteria (25.1%), Phycisphaerae (12.1%), Alphaproteobacteria (6.8%), Cytophagia (6.4%), and Gammaproteobacteria (5.1%) as the most abundant classes, with Cytophagia and Chorobiia exclusive to the sample. In PW, Gammaproteobacteria (16.2%), Betaproteobacteria (10.6%), Alphaproteobacteria (10.3%), Verrucomicrobiae (9.5%), Chlamydiia (8.7%), Terriglobia (8.4%), and Unclassified Candidatus Melainabacteria (7.6%) were dominant. Also, Unclassified Candidatus Melainabacteria, Bacteriovoracia, Deinococci, Chlamydia, and Terriglobia were found to be present only in PW. Order level taxonomy The 10 most frequently detected orders were Bifidobacteriales (44.55%), Flavobacteriales (29.00%), Coriobacteriales (24.78%), Burkholderiales (22.56%), Phycisphaerales (12.06%), Enterobacterales (11.91%), Eubacteriales (9.51%), Limisphaerales (9.53%), Parachlamydiales (8.68%), and Bryobacterales (8.42%) (Fig. 4c). In GW, Bifidobacteriales (44.55%), Coriobacteriales (24.79%), and Eubacteriales (9.5%) were the most dominant, prevalent, and exclusive. TW harbored Flavobacteriales (29.0%), Burkholderiales (22.57%), Phycisphaerales (12.0%), Cytophagales (6.39%), and Chlorobiales (4.16%) as the most abundant orders, with Cytophagales and Chlorobiales specific to the sample. In PW, Enterobacterales (11.9%), Burkholderiales (10.13%), Limisphaerales (9.5%), Parachlamydiales (8.68%), Bryobacterales (8.4%), Vampirovibrionales (7.63%), and Bacteriovoracales (6.44%) were dominant. Also, Limisphaerales, Bryobacterales, Vampirovibrionales, Bacteriovoracales, and Thermales were exclusive to PW. Family level taxonomy The 10 most frequently detected families were Bifidobacteriaceae (44.55%), Coriobacteriaceae (24.67%), Flavobacteriaceae (20.29%), Comamonadaceae (18.00%), Phycisphaeraceae (12.06%), Erwiniaceae (11.87%), Limisphaeraceae (9.25%), Lachnospiraceae (8.85%), Simkaniaceae (8.68%), and Crocinitomicaceae (8.67%). In GW, Bifidobacteriaceae (44.55%), Coriobacteriaceae (24.67%), and Lachnospiraceae (8.85%) were the most dominant and exclusively present. TW harbored Flavobacteriaceae (20.29%), Comamonadaceae (18.00%), Phycisphaeraceae (12.0%), Crocinitomicaceae (8.67%), and Cytophagaceae (5.0%) as the most abundant families, with Cytophagaceae and Chlorobiaceae specific to the sample. In PW, Erwiniaceae (11.87%), Limisphaeraceae (9.25%), Simkaniaceae (8.68%), Bryobacteraceae (8.42%), Unclassified Vampirovibrionales (7.6%), and Bacteriovoracaceae (6.44%) were dominant and exclusive (Fig. 4d). Genus level taxonomy The 10 most frequently detected genera were Bifidobacterium (44.55%), Collinsella (24.66%), Flavobacterium (20.29%), Limnohabitans (12.88%), Mucisphaera (12.05%), Pantoea (11.87%), Limisphaera (9.25%), Simkania (8.68%), Fluviicola (8.67%), and Mediterraneibacter (8.40%). In GW, Bifidobacterium (44.55%), Collinsella (24.66%), and Mediterraneibacter (8.4%) were found to be the most dominant and exclusive. TW showed Flavobacterium (20.29%), Limnohabitans (12.88%), Mucisphaera (12.0%), and Fluviicola (8.67%) as the most abundant genera, with Aquirufa, Polynucleobacter, and Chlorobaculum being specific to the sample. The PW was dominated exclusively by Pantoea (11.87%), Limisphaera (9.25%), Simkania (8.68%), Paludibaculum (8.24%), Vampirovibrio (7.6%), and Peredibacter (6.44%) (Fig. 4e). Spatial variations in microbial composition The PCA analysis indicated differences in the microbial composition among the three sources. PC1 accounted for 84.34% variation among samples across three sources (Fig. 5a). Microbes from GW separated from the other two sources (TW and PW), with the PC2 explaining 15.66% of the total variance. The source-wise grouping of samples as displayed by the non-metric multidimensional scaling (NMDS) corroborated these results with a stress value of 0.1318 and an R value of 0.995 (Fig. 5b).Fig. 5Plots showing the microbial communities of water samples (a) PCA based on Euclidean and (b) NMDS dissimilarity (Bray–Curtis). Plots showing the microbial communities of water samples (a) PCA based on Euclidean and (b) NMDS dissimilarity (Bray–Curtis). Microbial pathogen mapping The distribution of potential pathogens within the water sample revealed a total of 10 pathogenic bacteria with RA 0.01%-0.62% (Supplementary Fig. 2). GW showed low diversity of waterborne pathogens but with high RA (1.117%), followed by TW (0.714%) and PW (0.567%). The predominant human pathogens like Acinetobacter spp., Aeromonas, Enterobacter spp., Pseudomonas spp., Staphylococcus spp., Legionella, and Vibrio were omnipresent, while Mycobacterium spp. and Burkholderia spp. were found only in TW and PW. Enterobacter spp. was found to be the most dominating pathogen in GW with RA ~ 0.62%. The Pseudomonas spp. was found to be prevalent in TW and PW. Correlation of physicochemical parameters with microbial diversity and pathogens In Multivariate Canonical Correspondence Analysis (CCA), component 1 explained 67.64%, while component 2 supported 32.36% of the variance observed due to the influence of physicochemical parameters on microbial diversity. The CCA triplot revealed strong relationships and distinguished the samples significantly in terms of bacterial diversity (Fig. 6a). In TW, the distribution of Bacteroidota, Chlorobiota, Planctomycetota, and Cyanobacteriota was influenced by a combination of Mo, EC, salinity, and TDS. High pH, COD, and Ni in PW affected the presence of Verrucomicrobiota, Acidobacteriota, Chlamydiota, Candidatus Melainabacteria, Bdellovibrionota, Deinococcota, Gemmatimonadota, and Nitrospirota. Neither temperature nor DO affected bacterial diversity in any sample; however, physicochemical parameters had no influence on GW microbial diversity.Fig. 6(a) Canonical correspondence analysis of microbial communities and physicochemical variables. (b) Pearson’s correlation map showing the influence of the physicochemical variables on the abundance of the microbial communities. (a) Canonical correspondence analysis of microbial communities and physicochemical variables. (b) Pearson’s correlation map showing the influence of the physicochemical variables on the abundance of the microbial communities. Pearson’s correlation analysis showed a positive correlation of temperature with Bacillota and Actinomycetota and a negative correlation with pH. Verrucomicrobiota, Acidobacteriota, Chlamydiota, Candidatus Melainabacteria, Bdellovibrionota, Deinococcota, Gemmatimonadota, and Nitrospirota showed negative correlation with temperature, DO, TDS and positive correlation with COD, pH, and Ni. Bacteroidota was positively correlated with TDS, salinity, and Mo whereas COD showed negative correlation with Bacteroidota, Chlorobiota, Bacillota, and Actinomycetota. Phyla Verrucomicrobiota, Acidobacteriota, Chlamydiota, Candidatus Melainabacteria, Bdellovibrionota, Deinoccota, Gemmatimonadota, and Nitrospirota were negatively correlated with Mo and positively with Ni. Conversely, Chlorobiota and Bacteroidota showed a positive correlation with Mo and negative with Ni (Fig. 6b). Pearson’s correlation analysis of physicochemical parameters with pathogenic bacteria showed a highly positive correlation (> 0.75) of Mycobacterium, S. aureus, Campylobacter, and Aeromonas with Mo, EC, TDS, and salinity. Also, E. sakazakii was shown to have a positive correlation with DO; Acinetobacter with Ni and COD; Legionella with pH, Ni, and COD; Vibrio with DO and temperature; and B. pseudomallei with Ni, COD, and pH. A highly negative correlation (< − 0.9) of Mo, EC, salinity, and TDS was noted with P. aeruginosa; DO and temperature with B. pseudomallei; TDS and DO with Acinetobacter; pH with E. sakazakii; DO and temperature with Legionella; and pH with Vibrio (Supplementary Fig. 3). Population-based survey to assess water quality An online survey to assess institutional water wellness was conducted among the occupants of the sampling institutions. A total of 496 respondents (73% male and 27% female) participated in the survey. About 35% of respondents reported using their own water bottle brought from home as the primary source of drinking water, whereas 16.4% were using institutional RO water, and the others relied on buying commercially available bottled water. About 59% of the participants were unaware and more than 27% were ambiguous about the water sources. Instances of leakage, open tanks, and poor hygienic conditions were noted by ~ 40% of respondents. A change in odour (42%), taste (~ 67%), and discoloration (~ 15%) was reported due to lack of periodical cleaning of water sources (18%) and non-functional RO systems (~ 40%). Despite these factors, more than 70% of respondents were satisfied with the quality of drinking water. As per 66.5% of respondents, the change in colour, odour and taste could be the underlying cause for water-related health problems. The low water quality was attributed to microbial contamination and chemicals like pesticides, weedicides, etc. by ~ 50% and ~ 25% of occupants, respectively. Most of the participants (~ 63%) were aware of the water-related health concerns and instances of waterborne diseases, such as typhoid (50%), diarrhea (22.6%), etc. The general awareness and understanding of respondents regarding the water quality parameters and policies in place for clean water accessibility and sanitation were also assessed. About 42% of respondents were aware of the term ‘WQI,’ whereas ~ 72% were familiar with the UN’s SDG-6 related to clean water and sanitation. Approximately 75% of respondents were skeptical about the accessibility of clean water and the lack of adequate and equitable sanitation and hygiene facilities for the general population (Fig. 7).Fig. 7Staggered graph showing the People’s perception of water quality, hygiene and sanitation. Staggered graph showing the People’s perception of water quality, hygiene and sanitation. Discussion Evaluation of water quality and its impact on occupants of built-in environments, like academic institutions, corporate offices, etc., is critical to prevent waterborne diseases and to safeguard their health48,49. Designing strategies to manage/maintain water quality requires an understanding of water physiography and spatiotemporal fluctuations50. The water quality can be assessed using a set of physicochemical and biological parameters, which are influenced by the environmental factors, geological characteristics (runoff), and anthropogenic activities51. The present study reports spatiotemporal variations affecting water quality across study sites and sources. The overall quality improved during the wet season due to the dilution effect of rain and low demand. The improvements in the WQI shown for both TW and PW may be due to higher DO and reduced salinity during the wet season. However, the decreased WQI of GW might be a result of increased TDS, EC, and salinity caused by runoff. Among the physicochemical parameters, both temperature and ORP showed a negative correlation with pH. ORP is further influenced by chlorination of drinking water, which is evident by the higher values recorded for the TW as compared to PW52. Higher pH values in all water samples during the dry season could be attributed to buffering by alkaline aquifer material (enhanced carbonate), low-recharge conditions, and reduced biological CO₂ uptake due to insufficient photosynthetic activity. This trend is observed particularly in groundwater and processed water systems widely reported in urban and alluvial aquifer systems globally53–55. Seasonal dynamics can also affect acid–base equilibria, influencing pH trends as observed in PW (low pH) during the wet season (external disturbances, rainwater, atmospheric depositions, high aquatic vegetation, and increased microbial metabolism)56. Furthermore, increased levels of nitric acid from decaying vegetation and organic matter (enhanced biochemical oxidation) may have decreased the pH during the wet season57. The perfect correlation observed between EC and TDS is consistent with their well-documented association in reflecting ionic strength and salinity58. GW showed the higher EC (at site I and II) and TDS values, indicating elevated ionic concentrations due to prolonged rock-water interactions, anthropogenic leaching, limited natural dilution during dry periods, and changes in land use patterns59. EC and TDS showed high values during the dry season; that reflects possible infiltration of sewage and dust deposition. Also, construction activities at sites I and II may have contributed to the elevated levels of TDS and EC across all sources. TW demonstrated the lowest EC and TDS, indicating pre-treatment and centralized management (sewage treatment plants), with further notable improvement during the wet season likely due to increased source water mixing. PW showed moderate levels of EC and TDS but exhibited more seasonal variability than GW, likely due to anthropogenic activities such as construction, road dust resuspension, and atmospheric inputs. The significant decline in TDS during the wet season supports the rainfall-driven dilution hypothesis60. DO and COD values are crucial for water quality assessment and vary notably across seasons and sources. Higher DO values during the wet season may be due to increased photosynthetic activity, wind movement, and respiratory processes61. A general trend showing improved DO across all sources during the wet season could be attributed to the rainfall dilution effect and reduced demand-driven stress in the case of TW. Low DO levels of GW reflect microbial degradation and absence of aeration. PW remained problematic due to organic pollution often worsened by atmospheric deposition carrying ionic dust and organic debris62. Higher COD values for both GW and TW during the dry season suggest organic load and inefficiencies in removing contaminants during water treatment processes, respectively. These values were well above the permissible WHO limits, raising concerns about the safety of occupants18. Significant reductions in COD (organic pollution) were observed in both GW and PW samples during the wet season due to dilution and scavenging effects of rain. PW, however, retained higher COD in both seasons, which may be contributed by surface-borne pollutants63. The WQI analysis supported these findings; for instance, GW showed the lowest WQI because it continuously receives contaminants from nearby markets, mixed cluster habitation, and dissolved construction material leaching64. WQI for TW improved significantly, entering the “Good” category, whereas for PW it remained “Fair,” possibly due to anthropogenic activities. The ponds used in the present study were artificial and under constant disturbances (variable vegetation growth and volumetric changes). These disparities mirror global urban water systems, where infrastructure and hydrology shape source-dependent patterns65. Water microbiomes, especially bacteriomes, play an important role in the aquatic ecosystem for the transformation and cycling of nutrients66. Low bacterial diversity across all samples may be attributed to less heterogeneity of water sources, similar environments, and close locations51. Freshwater organisms such as Actinomycetota, Pseudomonadota, Bacteroidota, Bacillota, and Planctomycetota, etc., were found dominating all water samples. As reported earlier, GW samples showed the presence of both Actinomycetota and Pseudomonadota67. Phyla, namely Verrucomicrobiota, Acidobacteriota, Chlamydiota, Candidatus Melainabacteria, Bdellovibrionota, Deinococcota, Bacteroidota, Gemmatimonadota, and Nitrospira from PW, showed a positive correlation with pH and COD, which may be attributed to an increase in organic and inorganic substrates. Both TW and PW promoted growth of Pseudomonadota members due to a negative correlation with DO and temperature and a positive correlation with pH68,69. Similarly, Actinomycetota was found to be highly abundant in GW with higher DO and temperature and low pH values at site III, especially during the wet season68. Salinity appears to be a major factor driving the changes in microbial communities in TW, as it is linked to high TDS and EC in aquatic environments58. Conversely, the exclusive presence of Actinomycetota and Bacillota in GW suggested their potential resilience against any physicochemical parameters. These findings underscore the significant role of pollutants in shaping microbial ecology in open water sources70. Heavy metals have substantial influence on bacterial community structure; even trace amounts of heavy metals can influence the growth of bacteria in sediments71. Notably, in TW, three bacterial phyla, Bacteroidota, Chloroflexota, and Planctomycetota, have shown strong influence due to the presence of heavy metal Mo, suggesting metal contamination and ionic strength variations favor specific microbial groups72. Three classes of Pseudomonadota, i.e., alpha, beta, and gamma, were found to be abundant across all the sources, which agrees with the previous observations73. The health implications of contaminated water are severe and can range from gastrointestinal diseases to more serious infections. The presence of microbial colonies implied biological contamination of water samples, and subsequent mapping of pathogens further pinpointed the specific health risks to the occupants. Identification of pathogens in different water samples can be instrumental in implementing appropriate control measures18,74,75. Pathogenic bacteria were ubiquitous and present at relatively high abundance (~ 1.1%), especially in GW. Among them, Pseudomonas spp., responsible for bacteremia, pneumonia, etc.76, and Enterobacter, for UTIs, meningitis, etc.77, were the most prevalent in all samples, followed by Acinetobacter, responsible for lung infections and bacteremia78. Several other potential pathogens, such as Aeromonas (gastroenteritis), Mycobacterium (nontuberculous mycobacteria, i.e., NTM infection), Burkholderia pseudomallei (melioidosis), Campylobacter (campylobacteriosis), Enterobacter sakazakii (necrotizing enterocolitis, sepsis, and meningitis), Staphylococcus (pneumonia and myocarditis), and Vibrio (cholera), were also detected in samples79,80. Aeromonas spp. are common aquatic microorganisms, and a number of them have been associated with human diseases81,82. The correlation analysis between physicochemical parameters and pathogenic bacteria revealed significant associations, suggesting that water quality directly influences pathogen presence. A strong positive correlation of Mycobacterium with Mo indicates conditions favorable for its survival83. In contrast, P. aeruginosa and B. pseudomallei exhibited negative correlations with EC, indicating intolerance and inhibition of their growth. These findings underscore the impact of physicochemical parameters on the distribution of pathogenic bacteria. Promoting hygienic conditions and behavioral interventions is imperative to prevent the adverse impact of water pollution on health in HEIs. The survey-based assessment of clean water accessibility at the sampling sites revealed various dimensions about occupants’ awareness. The skepticism regarding the use of institutional drinking water showed poorly maintained water purification systems and particulate pollution resulting from ongoing construction in the vicinity84. The waterborne diseases and their symptoms appeared to be linked with the low water quality, like bad odor, coloration, and taste85. Although utmost care was taken to explain the motive behind this survey to the volunteers, but still a few biases could have been introduced due to recall bias, perception bias, and social desirability bias etc. and other limitations commonly associated with questionnaire-based assessments of environmental health risks4,86,87. While these biases cannot be entirely eliminated, they were mitigated through anonymous participation, informed consent, and structured questionnaires. Importantly, we emphasize that the survey findings were triangulated with physicochemical measurements and metagenomic evidence, thereby strengthening the reliability of interpretations. Thus, addressing challenges in water hygiene and ensuring clean water accessibility to the occupants are vital steps to reduce its associated health risks. Conclusions Assessing water quality in HEIs is essential for understanding the overall water, sanitation, and hygiene (WASH) situation. Water source dynamics play a pivotal role in understanding the spatiotemporal variations in water quality. TW is typically treated and is expected to adhere to certain standards; however, variations in treatment processes, pipe age, and distribution network conditions can impact its quality. GW is influenced by aquifer characteristics, recharge areas, and anthropogenic activities in the vicinity, leading to its contamination. PW, being an open water body, is susceptible to fluctuations in its quality due to environmental factors like rainfall, evaporation, and surrounding land use patterns. Maintaining water quality standards is indispensable to keep the occupants healthy. Evaluating waterborne pathogens is instrumental in guiding the selection of appropriate treatment protocols and making informed decisions for the maintenance of water purification systems. Thus, periodical monitoring of physicochemical parameters and microbial contamination can provide a comprehensive understanding of water quality and help in the development of effective strategies to mitigate health risks. This study provides an understanding of how spatiotemporal variations and microbial contaminants can adversely affect the occupants if not managed properly. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3. Supplementary Material 4. Supplementary Material 4. Supplementary Material 5. Supplementary Material 5. Supplementary Material 6. Supplementary Material 6.