HomeTopicsWater Harvesting › Managed Aquifer Recharge
layers Water Harvesting

Managed Aquifer Recharge: Banking Water Underground

Storing seasonal surplus water in aquifers costs 50 to 400 USD per megalitre of capacity. Building an equivalent surface reservoir costs 800 to 3,000 USD per megalitre, plus ongoing evaporation losses of 1 to 3 metres per year. The physics of underground storage are straightforward: water below the surface cannot evaporate. The engineering question is how to get it there quickly enough to be useful.

schedule 11 min read article ~2,500 words update April 16, 2026
Dig Deeper

The Problem MAR Solves

Aquifer depletion is a practical economic constraint before it becomes a crisis. Irrigated agriculture globally withdraws approximately 70 percent of freshwater consumed by human activity, and in arid and semi-arid regions, a significant fraction of that withdrawal comes from aquifers recharging at rates far below extraction. The Ogallala Aquifer in the US Great Plains declines at 0.3 to 1.0 metres per year across most of its footprint while recharging naturally at 0.5 to 1.5 cm per year. The Central Valley aquifer in California has experienced cumulative subsidence of 3 to 9 metres in some areas due to groundwater overdraft over the past century (Farr et al., 2015, Water Resources Research 52(4):2131-2148; USGS California Water Science Center).

Managed aquifer recharge (MAR) is the deliberate redirection of surface water, floodwater, or treated reclaimed water to sites where it can infiltrate and enter aquifer storage. It is not a new concept: the practice has been used formally in parts of Europe and the US Southwest since the early 20th century. What has changed in the past two decades is the sophistication of hydrogeological site assessment tools (satellite-based land subsidence monitoring, geophysical surveys, digital groundwater models), the availability of treated reclaimed water as a source, and the regulatory frameworks in several jurisdictions that now explicitly authorise and encourage MAR as a water management tool.

The operator asking about MAR is typically in one of three situations: they are farming in a region where groundwater levels are declining and they want to use seasonal flood surplus or recycled water to offset their extraction; they are a water utility or rural water authority looking to replace or delay expensive surface reservoir expansion; or they are a landholder who wants to capture seasonal flood surplus from a stream crossing their property and use it to recharge the shallow unconfined aquifer that supplies their bores. All three scenarios are legitimate MAR applications with documented precedent. This page covers the mechanisms, economics, and implementation pathway for farm to district-scale applications, connecting back to the water harvesting pillar's core argument: water stored underground cannot evaporate.


Methods: How Water Gets Into the Aquifer

The four primary MAR methods differ in cost, geology requirements, and scale. Spreading basins are the simplest and cheapest: a shallow pond, typically 0.5 to 2.0 metres deep, constructed in an area with highly permeable soil or alluvial sediments, into which water is pumped or diverted during high-flow periods. The pond bottom allows water to percolate downward through the unsaturated zone (vadose zone) to the water table. Percolation rates in coarse alluvial gravel can exceed 1 metre per day; in fine sand they fall to 0.1 to 0.3 metres per day; in silt or clay they are insufficient for practical MAR without pre-treatment of the substrate. Basin size requirements for a given recharge rate are determined by the percolation rate: achieving 1,000 m3/day recharge in fine sand soil with 0.2 m/day percolation rate requires 5,000 m2 of basin area. In alluvial gravel at 1.0 m/day, 1,000 m2 suffices.

Infiltration galleries are buried perforated pipe systems installed in shallow trenches backfilled with clean gravel. They distribute water horizontally through the vadose zone over a larger footprint than a single concentrated basin, reducing clogging risk and maintaining more uniform percolation. Installation cost is higher than a simple spreading basin, typically 80 to 200 EUR per linear metre of gallery, but infiltration galleries work in soils with lower permeability than spreading basins require, and they can be installed below agricultural land without removing productive area from use.

Injection wells are drilled boreholes with casing, screen, and pumping infrastructure that inject water directly into a confined or semi-confined aquifer, bypassing the vadose zone entirely. Injection wells are necessary when the target aquifer is confined (i.e., overlain by an impermeable layer that prevents surface infiltration) and when very high recharge rates are needed in a small footprint. Capital cost for a drilled injection well is 15,000 to 80,000 EUR per borehole depending on depth and geology, with additional pump and treatment infrastructure. Water quality requirements for injection are the most stringent of any MAR method, as clogging of the aquifer formation around the well screen is the primary operational failure mode.

Bank filtration, the fourth method, uses pumping wells placed near rivers or lakes. When water is pumped from the wells, hydraulic gradients draw water from the river through the riverbank sediments and into the aquifer. The riverbank acts as a passive filter, removing suspended solids, pathogens, and some organic compounds. This method is widely used in Germany and the Netherlands for municipal water supply pre-treatment: approximately 50 percent of Germany's drinking water supply is produced by bank filtration (Hiscock and Grischek, 2002, Journal of Hydrology 266:139-144).

Soil and Aquifer Profile for Spreading Basin MAR
Basin floor (surface) Scraped to permeable substrate. Fine surface sediment removed to expose coarse alluvium. Flooded to 0.5-1.5 m depth during recharge cycles. Percolation rate determines basin area needed. Rate: 0.1-2.0 m/day
Vadose zone (unsaturated) Unsaturated zone between basin floor and water table. Water percolates downward through pore spaces. Soil aquifer treatment (SAT) occurs here: pathogen removal, organic carbon removal, nitrate attenuation. Depth: 2-30 m
Unconfined aquifer (water table) Water table rises as recharge water enters. Each 1,000 m3 added to a 10-hectare aquifer footprint with 25% porosity raises the water table by 0.04 m. Seasonal cycles of recharge (wet) and extraction (dry) make this a managed bank. Target storage zone
Confined aquifer (if present) Separated from unconfined zone by clay aquitard. Injection wells required to reach this layer. Higher pressure, often artesian. Separate regulatory approval typically required for confined aquifer injection. Injection wells only

The Economics of Underground Storage

The cost comparison between MAR and surface reservoir storage is compelling at most scales. The World Bank water sector infrastructure cost database and Australian Water Resources Council data document surface reservoir costs in the range of 800 to 3,000 USD per megalitre of active storage capacity for new construction, varying by terrain, dam height, and concrete versus earthfill construction. Spreading basin MAR systems in suitable alluvial geology cost 50 to 400 USD per megalitre of annual storage throughput (vault_atom_TBD: World Bank Water P&E Technical Papers; CSIRO AWP cost database). The difference of 3x to 20x in capital cost is compounded by the evaporation factor.

A surface reservoir in a semi-arid climate with 2,000 mm potential evaporation (typical for much of southern Europe, California, and southern Australia) loses 1.5 to 2.5 metres of water depth per year from the open water surface. For a reservoir 3 metres deep with 1,000 ML of active storage, that represents an annual evaporative loss of 500 to 833 ML, or 50 to 83 percent of storage volume. An aquifer storing the same 1,000 ML loses effectively zero to evaporation. Over a 30-year asset life, the aquifer delivers 30,000 ML of stored water; the surface reservoir delivers approximately 10,000 to 18,000 ML after evaporation losses, at 2 to 5 times the capital cost per megalitre delivered.

The Orange County Water District in California operates the Groundwater Replenishment System (GWRS), the world's largest advanced water purification facility for indirect potable reuse via aquifer recharge. The GWRS treats 130 million gallons (approximately 492 ML) per day to drinking water standards and recharges it to the Orange County Basin via spreading basins. Total capital cost for the expansion to current capacity: approximately 487 million USD. Cost per megalitre per year of sustainable yield from the recharged aquifer: approximately 180 to 240 USD, significantly below the 800 to 1,200 USD per ML that new surface reservoir construction would cost in the same region (OCWD Annual Report 2023; vault_atom_TBD).

Cost per Megalitre of Storage Capacity (Capital Only)
MAR: 3-20x Cheaper Than Surface Storage
Sources: World Bank Water Infrastructure Cost Database; CSIRO AWP; OCWD Annual Report 2023; vault_atom_TBD
Spreading basin MAR (suitable alluvial site)50-400 USD/ML
Advanced treatment + spreading basin (GWRS scale)180-240 USD/ML
Injection well MAR (confined aquifer, urban scale)400-800 USD/ML
Earthfill surface reservoir (new construction)800-1,500 USD/ML
Concrete dam surface reservoir (challenging terrain)1,500-3,000 USD/ML

The recovery efficiency of MAR systems is the ratio of water recharged to water recovered by pumping. In a well-designed spreading basin with a confined extraction well field, recovery efficiency is 60 to 80 percent; the remainder mixes with ambient groundwater or migrates beyond the recovery zone. Injection-recovery wells (ASR, aquifer storage and recovery) achieve 80 to 95 percent recovery efficiency because the same borehole used for injection is used for recovery, drawing back the water banked in the formation around the well (Dillon et al., 2019, Water 11(7)).


Site Selection, Design, and Operation

The first requirement for a spreading basin MAR site is permeable geology within 10 metres of the surface. Alluvial fans, braided river channels, river terraces, and glacial outwash deposits are the ideal geological settings: coarse sediment, high hydraulic conductivity (10-3 to 10-2 m/s or higher), and good connectivity to the target aquifer. Checking geological maps for alluvial deposits is the starting point. Follow that with a percolation test: dig or auger a hole to 1.5 metres depth, fill with water, and measure the rate of fall over 30 minutes. If the fall rate is less than 25 mm per hour, the site will not support viable spreading basin MAR without significant soil amendment.

Water quality of the source must match the aquifer's intended use. Floodwater sourced directly from rivers during high-flow periods typically carries sediment loads of 500 to 5,000 mg/L total suspended solids. This sediment will clog the basin floor within hours to days if not addressed. Pre-treatment using a settling basin upstream of the MAR basin, with 24 to 48 hours of residence time, reduces suspended solids to below 50 mg/L before the water reaches the recharge zone. Without pre-treatment, spreading basins require frequent scraping and regrading of the basin floor to restore percolation rates, adding operating cost and logistical complexity.

Basin size calculation: determine the daily recharge volume target (say, 500 m3/day during a 90-day flood season), divide by the measured percolation rate (say, 0.5 m/day in sandy loam), to get required basin area of 1,000 m2. Add 20 percent for margin, giving a design basin of 1,200 m2. Earthworks for a 1,200 m2 basin 1.0 metre deep, graded flat: approximately 1,500 to 3,000 m3 of excavation. At 4 to 8 EUR per cubic metre for bulk earthworks, the basin itself costs 6,000 to 24,000 EUR. Add inlet structure, outlet overflow, and sediment trap: total installed cost typically 15,000 to 40,000 EUR for this scale.

The Clogging Problem

All spreading basin MAR systems experience clogging of the basin floor over time as suspended solids, algae, and fine sediment accumulate in the surface pores. Recovery of percolation rate requires periodic drying and mechanical scarification of the top 5 to 10 cm of basin floor. Systems that alternate between wet (recharge) and dry (rest/recover) cycles maintain higher long-term percolation rates than continuously flooded basins. Design the operational regime with planned dry periods of 2 to 4 weeks for every 6 to 8 weeks of active flooding.

Monitoring consists of: groundwater level monitoring wells in the aquifer (installed at 3 to 5 locations around the basin before operation begins), flow meters on inlet and outlet structures, and turbidity monitoring of inlet water. Monthly groundwater level readings allow calculation of the actual recharge volume entering the aquifer each season, enabling comparison with design targets and identification of changes in percolation performance over time.

MAR Method Selection by Site Geology and Scale
Method Geology Required Capital Cost Scale Key Constraint
Spreading Basin Permeable alluvium, unconfined aquifer 50-200 USD/ML/yr Farm to regional Requires high permeability; clogging management
Infiltration Gallery Moderate permeability; shallow water table 100-300 USD/ML/yr Farm scale Higher installation cost; limited to shallow unconfined
Injection Well (ASR) Any; confined or unconfined 400-800 USD/ML/yr Urban; large scale Strict water quality; borehole construction cost
Bank Filtration River alluvium; high transmissivity 150-400 USD/ML/yr Municipal water supply Requires proximity to river; riverbank character

MAR in the Water Harvesting Stack

Managed aquifer recharge is the natural downstream destination for water captured by any of the earthworks-based water harvesting methods. On-contour swales and keyline cultivation redistribute surface water and accelerate infiltration at the farm scale, functioning as distributed, unmanaged MAR. Check dams slow stream flow and hold water in alluvial stream bed sediments, creating managed infiltration zones in the streambed. The connection to the water table recharge page in this pillar is direct: that page covers the passive, distributed earthworks approach; this page covers the engineered, concentrated approach. Both are moving water from the surface to the aquifer; the method choice depends on scale, geology, and how quickly recharge is needed.

In the context of constructed wetlands for water purification, MAR is the endpoint for treated effluent that meets groundwater quality standards. Constructed wetland output directed to a spreading basin represents full closure of a greywater loop: treated water infiltrates to aquifer, is recovered from extraction wells, and returns to the farm irrigation system. In jurisdictions that credit recovered aquifer water as a private property right attached to the recharging land parcel, this loop also creates a legal water entitlement from previously unavailable recycled water, which is an asset with real market value.

The climate adaptation case for MAR parallels the argument for swales and earthworks at the pillar level: water banked underground during wet years is available during drought years without the capital cost of new surface infrastructure. India's National Aquifer Mapping and Management Programme has identified 411 principal aquifer systems covering most of the country's agricultural zones, and MAR across these systems is increasingly treated as a national water security strategy rather than a niche technique (Central Ground Water Board of India, 2018). The economics are compelling: the cost of drilling 30 monitoring wells and constructing 10 spreading basins across a 500-hectare watershed is typically 40,000 to 80,000 EUR, less than one year of irrigation water cost for the same watershed from purchased or pumped surface water in many semi-arid regions.

For agroforestry operators concerned about the interaction between deep-rooted trees and groundwater, MAR provides a positive mechanism: trees in the vadose zone above an MAR-recharged aquifer draw on the recharged bank during dry periods, reducing dependence on surface irrigation while maintaining the productivity of the system. The regenerative agriculture connection runs through soil organic matter: high-SOM soils hold more water in the vadose zone between recharge events, extending the effective reach of each MAR pulse and reducing the frequency of recharge operations needed to maintain target water table depths.

MAR Position in the Water Infrastructure Chain
Flood surplus / seasonal flow diverted to Spreading Basin MAR recharges Unconfined aquifer
Constructed wetland effluent fed to Spreading Basin MAR closes loop to Irrigation extraction
Swale and earthworks infiltration distributed recharge into same Aquifer recovered by Bore pumping

FAQ

Common Questions About Managed Aquifer Recharge

What is managed aquifer recharge and how is it different from natural groundwater recharge?

Natural groundwater recharge occurs passively when rainfall infiltrates over years to decades. Managed aquifer recharge deliberately accelerates this by directing surface water to sites optimised for infiltration rate and aquifer connectivity. MAR methods include spreading basins (shallow ponds over permeable substrate), infiltration galleries (buried perforated pipe systems), injection wells (direct injection into aquifer), and bank filtration (wells near rivers that draw riverbank-filtered water underground). MAR systems can achieve recharge rates of 0.1 to 2.0 metres per day of water table rise per unit of input. Source: Dillon et al. (2019) Water 11(7):1-30; IGRAC Global MAR Portal.

How much does managed aquifer recharge cost compared to building a surface reservoir?

Spreading basin MAR systems in suitable alluvial geology typically cost 50 to 400 USD per megalitre of annual storage capacity. Large surface reservoirs cost 800 to 3,000 USD per megalitre of active storage capacity, plus ongoing evaporation losses of 1 to 3 metres per year. In semi-arid climates with 2,000 mm potential evaporation, a surface reservoir loses 15 to 25 percent of stored volume annually to evaporation alone. An equivalent aquifer storage loses effectively zero. Over a 30-year asset life, the total cost per megalitre delivered is typically 3 to 6 times lower for MAR than surface reservoir. Source: CSIRO AWP; World Bank Water P&E Technical Papers; vault_atom_TBD.

What water quality is needed for managed aquifer recharge and are there regulatory requirements?

For agricultural irrigation recovery, most jurisdictions require turbidity below 5 NTU, BOD below 10 mg/L, and no heavy metal contamination in the source water. For drinking water recovery, source water must meet primary drinking water standards or achieve those standards through soil aquifer treatment during infiltration. Many jurisdictions require pre-approval for MAR projects, with review periods of 6 to 24 months. In Australia, the National Water Quality Management Strategy covers MAR explicitly. In the US, the Underground Injection Control program regulates injection well MAR under the Safe Drinking Water Act. Source: WHO MAR Overview (2012); US EPA UIC Program.

Water Harvesting

Store Water Where It Cannot Evaporate

Underground storage costs a fraction of surface reservoirs and eliminates evaporation losses entirely. Explore the full water harvesting framework or see how water table recharge at the farm scale complements managed aquifer recharge.

Dig Deeper