Managed aquifer recharge for fluoride mitigation in crystalline hard rock aquifers: an integrated hydrological and hydrogeological approach

Researchers have developed and tested a new integrated framework for identifying where managed aquifer recharge (MAR) can safely and effectively mitigate fluoride contamination in fractured crystalline aquifers, and their results from a fluoride-endemic watershed in southern India carry a sobering warning: most places that look good for recharging groundwater are not actually safe bets for improving water quality. In a study published in Environmental Earth Sciences, Shahwaz Khan, P. D. Sreedevi, Tanvi Arora, and Shakeel Ahmed combined two decades of hydrochemical monitoring with geospatial modeling and geophysical validation in the Maheshwaram watershed of Telangana, and found that of 63 hydrogeologically promising recharge locations, only 14 demonstrated a consistent history of fluoride dilution when water levels rose.

The Maheshwaram watershed, covering roughly 53 square kilometers in Rangareddy District, is in many ways a microcosm of the groundwater crisis facing semi-arid hard-rock regions across peninsular India and beyond. The area sits on Archaean granites of the Eastern Dharwar Craton, where groundwater is stored and moves only through secondary porosity created by weathering and fracturing. An upper weathered saprolite zone acts as the principal reservoir, while the fractured bedrock beneath provides preferential flow paths. More than 700 irrigation borewells tap this system, driving water levels steadily downward, and prolonged interaction between groundwater and fluoride-bearing minerals such as biotite, fluorapatite, allanite, and epidote has produced some of the highest fluoride concentrations recorded in the region, with previous studies reporting values up to 17.3 milligrams per liter and local extremes in Telangana exceeding 20 milligrams per liter, far above the World Health Organization guideline of 1.5 milligrams per liter.

Fluoride at low concentrations is an essential trace element, but chronic consumption of water exceeding the WHO guideline can cause dental and skeletal fluorosis along with neurological complications. India is among the most severely affected countries, with more than 66 million people exposed to elevated fluoride in groundwater. Conventional defluoridation technologies exist but carry operational and maintenance costs that are often prohibitive in rural, water-stressed communities. Managed aquifer recharge, by contrast, offers a nature-based alternative: intentionally directing low-salinity, low-fluoride water into aquifers to raise water levels and, ideally, dilute contaminants. The catch, as the new study makes clear, is that recharge can cut both ways chemically. Depending on local mineralogy, fracture connectivity, and groundwater chemistry, infiltrating water may dilute fluoride, or it may leach additional fluoride from shallow weathered horizons and accelerate mineral dissolution.

To tackle this problem systematically, the team delineated groundwater potential recharge zones (GPRZs) by integrating twelve geo-environmental parameters in a GIS-based Analytical Hierarchy Process, a multi-criteria decision analysis technique that uses expert pairwise comparisons to weight each factor. The parameters spanned surface and subsurface controls: geology, geomorphology, soil, slope, land use and land cover, vegetation index derived from Sentinel-2 imagery, rainfall, drainage density, lineament density, aquifer transmissivity, infiltration rate, and fissured-zone thickness. Parameters directly controlling groundwater occurrence, such as lineament density, transmissivity, and infiltration rate, received higher weights than indirect influences like vegetation cover. The consistency ratio of the expert judgments fell below Saaty’s 0.10 threshold, indicating internally reliable weighting.

A key methodological refinement was the comparison of two recharge-zone maps: one built from surface parameters alone and another incorporating the subsurface hydrogeological layers. Adding transmissivity, infiltration rate, and fissured-zone thickness reduced the extent of the “Very Good” recharge class from 20.7 percent to 19.2 percent of the watershed, demonstrating that surface indicators alone can overestimate recharge suitability. The refined map, the authors argue, better captures the hydrogeological realities of crystalline aquifers, where weathering thickness and fracture distribution, not surface appearance, ultimately govern how much water can infiltrate and circulate.

Because model-based recharge maps can be circular if validated only against the same assumptions used to build them, the researchers independently checked their delineation using two lines of evidence. First, they reinterpreted 25 Vertical Electrical Soundings conducted with Schlumberger arrays at maximum current-electrode spacings of 300 meters, deriving Dar-Zarrouk parameters from the inverted layer resistivities and thicknesses. High transverse resistance, which ranged up to 10,074 ohm-square meters, indicated thicker and more transmissive aquifer zones, with values above 4,000 ohm-square meters covering 44 percent of the area and marking moderate to high groundwater potential. Longitudinal conductance values between 0.07 and 0.9 siemens pointed to moderate-to-good aquifer protective capacity, and electrical anisotropy values between 1.0 and 1.5 flagged favorable fracture connectivity at 21 sounding points. The high-recharge zones mapped by the AHP model coincided well with these geophysically favorable areas, particularly valley fills with high lineament density. Second, groundwater-level fluctuation maps showed that high and very high seasonal fluctuations fell within the good and very good recharge zones, corroborating the classification.

The heart of the study, however, lies in its use of long-term monitoring data to test whether recharge actually improves water quality. The team analyzed 798 groundwater samples collected from 19 representative borewells between 2003 and 2023, in both pre- and post-monsoon seasons, with fluoride determined by ion chromatography at the CSIR–National Geophysical Research Institute in Hyderabad. The record revealed distinct seasonal behavior. Pre-monsoon fluoride concentrations remained persistently high, between 1.24 and 1.88 milligrams per liter, largely insulated from rainfall variability and reflecting long residence times and mineral dissolution under alkaline conditions. Post-monsoon concentrations, ranging from 0.85 to 1.78 milligrams per liter, tracked rainfall more closely, with wet years bringing dilution and drought years, such as those during El Niño episodes in 2002, 2009, 2015, and 2018, bringing evaporative enrichment. Worryingly, fluoride has shown a gradual rising trend since 2015, suggesting that natural recharge is no longer sufficient to offset geogenic release and over-abstraction.

Groundwater levels told a parallel story. Pre-monsoon levels declined steadily over the 2001–2023 record, a signature of unsustainable pumping, while post-monsoon levels fluctuated with rainfall. Statistical analysis showed a significant positive correlation between annual rainfall and post-monsoon water-level recovery, but the persistent long-term decline despite several above-normal rainfall years confirmed that excessive pumping, not climate variability alone, is the dominant driver of aquifer depletion in the watershed.

The critical insight emerged when the researchers examined the relationship between water-level rise and fluoride response at each monitoring location. Only 14 of the sites exhibited a consistent inverse relationship, meaning that when groundwater levels rose after recharge, fluoride concentrations reliably fell. These sites, the study concludes, are where MAR structures such as percolation tanks, check dams, recharge shafts, or recharge wells are most likely to deliver simultaneous gains in groundwater quantity and quality. The remaining locations showed weak, inconsistent, or adverse responses. In some areas, recharge temporarily increased fluoride, likely because infiltrating water interacted with fluoride-rich minerals in shallow weathered granitic horizons before dilution could take hold, or because alkaline, bicarbonate-rich conditions promoted mineral dissolution and cation exchange that mobilized fluoride. Previous work in the watershed had documented two mechanisms for such enrichment: surface-derived or anthropogenic fluoride entering the aquifer during recharge events, and greater leaching potential of fluoride-bearing minerals at shallow depths compared with deeper levels.

A map-removal sensitivity analysis reinforced confidence in the underlying recharge-zone model. Removing one thematic layer at a time and re-normalizing the remainder showed that lineament density, drainage density, geomorphology, geology, infiltration rate, and fissured-zone thickness exerted the strongest control on recharge classification. Excluding lineament density, for instance, shrank the poor recharge zone by 21.7 percent while expanding the very good zone by 18.8 percent, underscoring how strongly fracture networks govern recharge in crystalline terrain. Drainage density had the largest single effect on the very good class, reducing its area by 21.6 percent when removed. In contrast, slope, vegetation index, rainfall, and transmissivity played secondary roles, suggesting the model rests on physically meaningful structural and hydrogeological controls rather than arbitrary weighting.

The authors are candid about the limitations of their approach. The AHP method inherently involves subjective expert judgment, even when consistency thresholds are met. Hydrogeological parameters were interpolated using inverse distance weighting, which smooths over the localized heterogeneity in weathering thickness and fracture distribution that characterizes hard-rock aquifers. Geophysical validation relied on only 25 sounding locations, and resistivity interpretation is inherently non-unique, with similar values potentially reflecting different subsurface conditions. Monitoring data for 2020–2022 were unavailable, creating a gap in the two-decade record, and the study did not explicitly model future climate-change impacts on recharge or fluoride behavior. Nevertheless, the convergence of multiple independent datasets, geospatial, geophysical, water-level, and hydrochemical, provides a more robust basis for site selection than any single method could offer.

The broader implications extend well beyond one watershed. Fluoride-affected crystalline aquifers are widespread across semi-arid regions of Asia, Africa, Australia, and South America, and conventional MAR planning typically prioritizes recharge enhancement without evaluating water-quality outcomes. This study demonstrates that such an omission can backfire: structures built at hydrogeologically suitable sites may still mobilize geogenic contaminants if aquifer geochemistry is ignored. By screening candidate sites against two decades of fluoride behavior before any construction begins, the framework offers a transferable, cost-effective pre-implementation tool for water managers. The authors suggest that broad weathered zones with gentle slopes and drainage convergence suit percolation tanks and check dams, while fractured zones intersected by major lineaments favor recharge shafts and wells, though detailed engineering design, storage-capacity assessment, and cost-benefit analysis remain tasks for future site-specific investigations. As climate variability intensifies water stress in hard-rock regions, the lesson from Maheshwaram is clear: where you recharge matters as much as how much you recharge, and the groundwater’s chemical memory of past recharge events may be the best guide to where new interventions will heal rather than harm the aquifer.