Nearly a quarter of Jammu and Kashmir samples show some degradation; Budgam, Baramulla and Srinagar record the highest index values

SRINAGAR: Groundwater in Jammu and Kashmir is under more stress than in Himachal Pradesh or Uttarakhand, according to a regional assessment published in the journal Water (MDPI). Nearly one in four samples from the Union Territory were classed as poor or worse, a much larger share than in the two neighbouring Himalayan states. The authors say the deterioration is spread across several districts rather than confined to a few exceptional readings.
The paper, Regional Variability and Spatio-Temporal Dynamics of Groundwater Quality in the Western Himalayas: An Integrated WQI and Hydrochemical Assessment, is by Kusum Pandey of the G. B. Pant National Institute of Himalayan Environment (Garhwal Regional Centre, Srinagar, Uttarakhand). Fenil Gandhi and Vipul Anand are at the Department of Civil Engineering, Sardar Vallabhbhai National Institute of Technology, Surat, Gujarat. Saurav Kumar, the corresponding author, is at the Department of Geography, Pachhunga University College, Mizoram University, Aizawl. Chandan Roy is at the Department of Geography and Resource Management, Mizoram University. Nikola Milentijević is at the Faculty of Sciences and Mathematics, University of Priština in Kosovska Mitrovica, Serbia. Milana Pantelić and Dragan Dolinaj are at the Faculty of Sciences, University of Novi Sad, Serbia.
Study Design
The team analysed Central Ground Water Board (CGWB) monitoring data from 338 observation wells for 2019–2022. Of these, 150 are in Uttarakhand, 108 in Jammu and Kashmir and 80 in Himachal Pradesh. They scored samples with a weighted arithmetic Water Quality Index (WQI) built on Bureau of Indian Standards (IS 10500:2012) and World Health Organization (2021) drinking-water guidelines.
The index uses ten parameters: pH, electrical conductivity, total dissolved solids, total hardness, calcium, magnesium, chloride, sulfate, nitrate and fluoride. Only samples with at least five valid parameters received a score. The authors classed results using a scale first proposed by Brown and colleagues in 1972. Scores up to 25 are “excellent” and 25–50 “good”. Scores of 50–75 are “poor”, 75–100 “very poor”, and anything above 100 is “unsuitable” and unsafe for human consumption.
Regional Gap
In Jammu and Kashmir, 54.7 percent of samples were “good” and 22.2 percent “excellent”. But 14.9 percent were “poor”, 5.6 percent “very poor” and 2.6 percent “unsuitable”. In Himachal Pradesh, 60.5 percent were “excellent” and about 37.5 percent “good”, together about 98 percent, with 1 percent “poor” and 1 percent “unsuitable”. In Uttarakhand, 68.7 percent were “excellent” and 25.5 percent “good”, with 3.9 percent “poor”, 0.9 percent “very poor” and 1 percent “unsuitable”.
Most Jammu and Kashmir scores fell between 20 and 80, but a long tail passed 200 and reached nearly 700 at a few sites. The authors say these extremes point to site-specific contamination or intensified mineralisation. Possible causes include human pressure, natural geology or poor waste management. In Himachal Pradesh, most scores sat between 15 and 40, and the authors describe pockets of deterioration there as localised.
District Hotspots
District medians show sharp contrasts within the Union Territory. Badgam recorded the highest median WQI, at roughly 68, followed by Budgam (about 46), Baramulla, Srinagar and Anantnag (all in the low-to-mid 40s). Handwara, Kupwara and Reasi recorded the lowest values. Pulwama, Kathua, Jammu, Samba, Rajouri and Udhampur fell in the middle range. The authors link the higher readings in Budgam and Srinagar to stronger urban and land-use influence.
In Uttarakhand, Udham Singh Nagar and Haridwar scored worst, which the authors tie to intensive agriculture, industrial clusters and high population density in the Tarai–Bhabar and Ganga plains. Pauri Garhwal and Dehradun scored best. In Himachal Pradesh, Solan and Una recorded the highest medians, though both remain within the “good” category. Kullu recorded the lowest.
Mineral Signals
Summary statistics show Jammu and Kashmir with higher average mineral loads than the other two states. Mean total dissolved solids were 333 mg/L, against 275 mg/L in Uttarakhand and 262 mg/L in Himachal Pradesh. Mean total hardness was about 266 mg/L, against 201 mg/L and 153 mg/L. The Jammu and Kashmir sample ranges were also wider. Nitrate peaked at 421.67 mg/L, far above the 45 mg/L limit, and pH reached 9.99.
Correlation analysis points mainly to natural mineralisation. In all three states, electrical conductivity and dissolved solids move closely with bicarbonate, hardness, calcium and magnesium. The authors read this as rock–water interaction and mineral dissolution driving most variation. In Jammu and Kashmir, nitrate and chloride show more scattered high readings than elsewhere, which the authors read as localised human pressure. They point to wastewater loading, agriculture, urban expansion and the confined geography of valley aquifers.
Spatial Clusters
The team also used Local Moran’s I, a spatial statistic, to find clusters of high and low readings. In Uttarakhand, geology-linked parameters such as bicarbonate, calcium and magnesium formed many high–high clusters, mainly in the west and south. Bicarbonate had 63 such clusters in 2020. The number of significant carbonate clusters rose from 21 in 2019 to 108 in 2021 and fell to just 10 in 2022.
Himachal Pradesh showed fewer clusters, dominated by low–high outliers. In 2019, chloride had 29 of its 47 significant locations in that category. In Jammu and Kashmir, calcium showed 40 high–high and 44 low–high clusters in 2021. Dissolved solids showed 19 and 23. Bicarbonate had 100 significant locations in 2019, concentrated in the southern and central parts of the Union Territory. The authors say clustering shifts from year to year and that deterioration is concentrated in foothill and valley-transition districts, not spread evenly.
Mild Improvement
Annual median scores indicate moderate improvement in groundwater quality between 2019 and 2022, particularly in parts of Jammu and Kashmir. The authors also say the greater year-to-year variability in Jammu and Kashmir suggests valley aquifers may respond more to short-term environmental and land-use changes than upland recharge areas in Himachal Pradesh and Uttarakhand.
Land Use
The paper puts land use in context. Open forest covers 33.36 percent of the study area, followed by agriculture (22.15 percent), snow (15.36 percent), barren land (10.82 percent), dense forest (7.52 percent), water bodies (7.43 percent) and built-up land (3.35 percent). Agricultural and built-up areas cluster in foothills, river valleys and cities, while forests and snowfields dominate higher ground. The authors argue this pattern helps explain why high-altitude recharge zones keep lower WQI values while foothill and valley districts show higher ones.
Key Caveats
The authors acknowledge several limitations. The data are annual, so seasonal shifts are not captured. The dataset excludes trace metals, microbial contamination, isotopes and emerging pollutants. The CGWB records do not state whether wells are used for drinking, domestic or agricultural purposes. Missing values were filled using a two-stage machine-learning imputation, validated by district-blocked cross-validation. Himachal Pradesh had too few observations for several parameters in 2021 and 2022. A single index can also oversimplify complex water chemistry, and district-level analysis may miss finer variation within watersheds, aquifers and spring catchments.
Policy Pointers
The paper recommends district-level monitoring that combines water-quality indices, multivariate hydro-chemical analysis, spatial clustering and land-use evaluation. It also calls for stricter wastewater regulation, protection of mountain recharge zones and better land-use planning in foothill and peri-urban districts. It singles out agricultural and urbanising foothill districts for priority attention. Processed datasets are available from the corresponding author on request, and the CGWB source data are public.














