FE233—Submarine Groundwater Discharge

Freshwater Lost, Not Lacking

Submarine Groundwater Discharge (SGD) is the movement of groundwater through coastal and subsea sediments into the ocean. Some SGD consists of recirculated seawater and some carries dissolved salts or contaminants, but a portion represents terrestrial groundwater that has moved through continental aquifer systems before ultimately reaching the sea. Once fresh groundwater mixes with seawater, recovering it again requires energy-intensive desalination. This raises a simple but largely neglected question: how much of the freshwater naturally destined for the ocean could instead be intercepted and put to beneficial human use before that mixing occurs?

FE233 proposes that this question deserves systematic investigation, particularly on the Indian subcontinent. Beneath the Indo-Gangetic Plain lies one of the world’s largest groundwater systems—sometimes described as the “Inverted Himalayas” because water-bearing sediments can extend to extraordinary depths. MacDonald et al. estimated approximately 30,000 ± 14,000 cubic kilometers of groundwater within just the upper 200 meters of the Indo-Gangetic Basin, more than twenty times the combined annual flow of the Indus, Ganges, and Brahmaputra rivers. Yet quantity alone does not determine availability. Much of the shallow resource is affected by salinity or arsenic, while groundwater depletion is severe in some regions and groundwater levels are stable or rising in others. The challenge is therefore not simply to extract more groundwater, but to understand where it comes from, where it goes, and which portions of its natural flow can be captured sustainably. [MacDonald et al., 2016]

The distinction between groundwater storage and groundwater flow is essential. A vast aquifer may contain thousands of cubic kilometers of water while receiving only a comparatively small renewable annual recharge. Pumping stored groundwater faster than it is replenished is groundwater mining regardless of the size of the original reservoir. A more promising opportunity is to identify groundwater already moving through the system toward natural discharge points—including groundwater ultimately destined for the ocean—and determine what fraction of that flow can be intercepted without unacceptable consequences.

The Bengal Basin demonstrates that this is more than a theoretical possibility. Chakrabarti et al. (2018) identified a distinctive strontium-isotope signature from Bengal Basin groundwater in Bay of Bengal water, particularly at depths of approximately 100–120 meters. Their measurements provide direct evidence that groundwater from the Bengal Basin reaches the Bay through submarine groundwater discharge. The magnitude remains uncertain and spatially heterogeneous, but the existence of the pathway is established. Earlier estimates of Bengal SGD varied enormously; subsequent regional groundwater modeling produced much smaller estimates, illustrating precisely why better measurement and modeling are required before the resource can be evaluated responsibly. [Chakrabarti et al., 2018]

The hydrogeology is also more complex than a simple underground river flowing from the Himalayas to the ocean. Mountain precipitation, snowmelt, monsoon rainfall, rivers, irrigation, canal leakage, local recharge, ancient groundwater, and repeated groundwater-surface-water exchanges all contribute to the regional system. The Himalayas ultimately supply enormous quantities of water and sediment to the Indo-Gangetic and Bengal basins, but the hydraulic head of the mountains should not be assumed to transmit unchanged through a single confined aquifer to the Bay of Bengal. Instead, groundwater moves through a heterogeneous sequence of sands, gravels, silts, and clays containing local aquifers and aquitards.

That complexity does not eliminate long-distance groundwater movement. Michael and Voss (2009) modeled the Bengal Basin as a strongly anisotropic regional aquifer system. Although individual clay layers can locally separate aquifers, available evidence did not support a single basin-wide confining layer. At regional scale, horizontal hydraulic conductivity was estimated to exceed vertical conductivity by roughly four orders of magnitude. Consequently, groundwater can follow deep flow paths extending tens to hundreds of kilometers despite the extremely small topographic gradients of the Bengal Plain. Pumping can substantially alter those natural flow paths. [Michael & Voss, 2009]

The quality of groundwater also varies strongly with depth and location. Shallow groundwater in large portions of Bangladesh and the Bengal Basin contains dangerous concentrations of naturally occurring arsenic, while deeper groundwater is generally much lower in arsenic. This deeper resource is therefore potentially valuable, but “deep” should not be confused with universally pure or inexhaustible. Salinity and other dissolved constituents vary geographically, and excessive pumping from deeper aquifers can change hydraulic gradients and potentially induce downward migration of contaminated shallow groundwater or movement of saline groundwater toward freshwater wells. Any strategy for expanded groundwater use must therefore manage water quality and hydraulic gradients together. [Michael & Voss, 2009; MacDonald et al., 2016]

Recent geophysical evidence adds another important dimension. Deep-sensing magnetotelluric surveys in coastal Bangladesh have identified large freshwater bodies at depth, including a seaward-dipping freshwater wedge extending to approximately 600 meters. Available groundwater ages indicate that substantial portions of these deeper coastal waters may be remnants of Pleistocene recharge rather than modern renewable groundwater. This is another reason that groundwater age, hydraulic head, salinity, and flow direction must be measured rather than inferred merely from the presence of freshwater. Ancient stored water and modern renewable flow may occupy portions of the same regional groundwater system but have very different implications for sustainable use.

Existing numerical models already provide a foundation for this work. Michael and Voss developed a three-dimensional regional groundwater model of the Bengal Basin capable of examining natural and pumping-altered flow paths and estimating groundwater discharge toward the Bay of Bengal. Later modeling of the South Bengal Basin has extended computational boundaries tens to hundreds of kilometers offshore and to depths approaching three kilometers while accounting for seawater density and bathymetry. These models demonstrate that the necessary numerical methods already exist. What remains incomplete is the observational characterization needed to constrain the offshore portion of the system and connect it confidently to the continental groundwater budget. [Michael & Voss, 2009; Sikdar, Banerjee & Chakraborty, 2022]

The next step should therefore be a new generation of integrated three-dimensional groundwater modeling rather than immediate large-scale extraction. Such a model should extend from the major recharge and river-interaction zones of the Indo-Gangetic and Bengal systems through the delta and well beneath the Bay of Bengal. It should assimilate groundwater levels, pumping records, borehole lithology, groundwater ages, arsenic and salinity measurements, river-groundwater exchange, monsoon recharge, offshore bathymetry, submarine groundwater discharge measurements, and modern electromagnetic or magnetotelluric imaging. Offshore geophysical surveys analogous to those used to identify large freshwater systems beneath other continental shelves could determine whether substantial fresh or brackish aquifers continue beneath the Bay of Bengal and identify preferential pathways through which groundwater reaches the ocean.

The purpose of such a model would not merely be to estimate how much water exists. It would determine where additional extracted water would actually come from. For any proposed pumping scenario, the model should partition the resulting supply among at least five sources: reduced natural SGD, increased recharge from rivers and surface water, depletion of stored groundwater, migration of contaminated shallow groundwater, and movement of seawater or saline groundwater. It should simultaneously calculate changes in hydraulic head, groundwater age, arsenic transport, salinity, land subsidence, river baseflow, wetlands, and coastal ecosystems. Sustainable recovery can then be defined by measurable limits rather than assumptions.

This approach changes the engineering objective. The goal should not be to “stop SGD,” because groundwater discharge can perform important ecological and geochemical functions and because much SGD is not usable freshwater. The objective should instead be to identify capturable groundwater discharge: the portion of terrestrial freshwater flow that can be redirected to human use without unacceptable depletion, contamination, saltwater intrusion, land subsidence, or ecological damage. Some of the most promising interception points may prove to be inland or immediately upgradient of the coast rather than beneath the ocean itself. Coastal interception wellfields could potentially capture groundwater before it enters saline transition zones, while continuous measurements of hydraulic head and salinity could regulate pumping to preserve the freshwater-saltwater balance.

Managed aquifer recharge could expand this opportunity further. During periods of abundant monsoon flow, suitable surface water might be intentionally stored underground in portions of the enormous Indo-Gangetic aquifer system and recovered during dry periods. Underground storage can reduce evaporation losses while using geological formations that already function as natural reservoirs. Coupling managed recharge with strategically located extraction could transform groundwater management from the mining of stored water into the deliberate management of renewable flow. Whether this is technically and economically practical at large scale is a question for modeling, field experiments, and pilot projects rather than assumption.

Only after the regional flow system is understood should more ambitious engineering options be evaluated. Directionally drilled coastal or offshore collector wells, for example, might intercept freshwater moving beneath the continental shelf before it mixes with seawater. Local subsurface barriers might prove useful where geological structures concentrate groundwater flow. Focused submarine springs, if significant ones are discovered, might even permit direct collection after groundwater leaves the seabed but before extensive mixing occurs. Each possibility is technically conceivable, but none should be presumed appropriate until the location, volume, quality, renewability, and ecological role of the groundwater have been established.

The larger lesson extends beyond the Indian subcontinent. Water scarcity should not be evaluated solely by measuring rivers, reservoirs, rainfall, or the volume stored underground. Water is continually moving through landscapes and aquifers, and some fraction of that freshwater ultimately crosses the coastline and becomes part of the ocean. Where that discharge is renewable, sufficiently fresh, and environmentally capturable, allowing all of it to mix with seawater before considering human use may represent a missed opportunity. The relevant question is therefore not whether every drop of SGD should be captured, but whether science and engineering can identify the portion that should be.

The Indian subcontinent offers an extraordinary place to find out. It combines the world’s highest mountain system, enormous monsoon and river flows, one of Earth’s largest sedimentary groundwater systems, severe groundwater-quality and depletion problems, deep low-arsenic freshwater resources, direct evidence of submarine groundwater discharge into the Bay of Bengal, and rapidly improving tools for three-dimensional groundwater imaging and simulation. Rather than assuming that this hidden freshwater is either inexhaustible or inaccessible, FE233 proposes something more disciplined: map it, measure it, model it, and determine how much of the water now flowing beyond human reach can safely be put to beneficial use first.

Selected Research

MacDonald, A. M., et al. (2016). “Groundwater quality and depletion in the Indo-Gangetic Basin mapped from in situ observations.” Nature Geoscience, 9, 762–766. The study estimates approximately 30,000 ± 14,000 km³ of groundwater within the upper 200 m of the Indo-Gangetic Basin while documenting extensive arsenic and salinity constraints and geographically uneven groundwater depletion.

Michael, H. A., & Voss, C. I. (2009). “Controls on groundwater flow in the Bengal Basin of India and Bangladesh: regional modeling analysis.” Hydrogeology Journal, 17, 1561–1577. The study develops a regional numerical groundwater model showing long flow paths at depth, strong horizontal-to-vertical hydraulic anisotropy, and substantial alteration of natural groundwater flow by pumping.

Michael, H. A., & Voss, C. I. (2009). “Estimation of regional-scale groundwater flow properties in the Bengal Basin of India and Bangladesh.” Hydrogeology Journal, 17, 1329–1346. Hydraulic heads, carbon-14 groundwater ages, and driller logs support an estimated regional horizontal-to-vertical hydraulic-conductivity ratio on the order of 10,000:1 and do not support a single basin-wide confining unit.

Chakrabarti, R., et al. (2018). “Submarine groundwater discharge derived strontium from the Bengal Basin traced in Bay of Bengal water samples.” Scientific Reports, 8, 4383. Strontium isotope measurements provide direct evidence of groundwater-derived material from the Bengal Basin entering the Bay of Bengal, with the strongest groundwater contribution detected primarily at approximately 100–120 m water depth.

Sikdar, P. K., Banerjee, S., & Chakraborty, S. (2022). “Understanding the Past-Present-Future Hydrogeologic System Through Numerical Groundwater Modeling of South Bengal Basin, India.” Frontiers in Water. The three-dimensional model incorporates regional groundwater flow, pumping, contaminant transport, seawater-density effects and an offshore model domain extending tens to hundreds of kilometers into the Bay of Bengal.

Le, H., et al. (2024/2025). “Deep fossil freshwater in coastal Bangladesh.” Magnetotelluric imaging along a 120-km transect identifies deep freshwater bodies, including a seaward-dipping freshwater wedge reaching approximately 600 m depth; groundwater ages indicate that substantial portions may represent preserved Pleistocene groundwater.