A stream keeps flowing during a dry spell even though almost no rain has fallen for days. Where can the water be coming from? Some may be entering from the surrounding aquifer through the streambed. At another reach, the opposite may happen: the river feeds groundwater. These hidden exchanges influence water supply, contamination, flood recovery and river ecosystems.
The gold standard of groundwater–surface water interaction is a testable explanation that links river flows, aquifer heads, bed properties, chemistry and time. This guide develops eighteen controls, two original fictional calculations, twelve FAQs and connections across hydrology, hydrogeology, groundwater modelling and ecology.
This root belongs to The Gold Standard. “Gold standard” refers to our editorial framework for careful evidence, not a water-quality certificate, legal water-rights decision or approved groundwater extraction plan.
Quick information: six checks at the water interface
- Identify the river, wetland or lake reach and relevant aquifer layers.
- Compare surface-water stage and appropriately located groundwater heads.
- Account for tributaries, diversions, storage and measurement timing.
- Separate net reach exchange from local hyporheic circulation.
- Investigate sediment conductance and possible disconnection.
- Test consequences for water supply, chemistry, habitats and nearby pumping.
Chapter index
- 1. Begin by treating groundwater and rivers as one system
- 2. Distinguish gaining, losing and mixed stream reaches
- 3. Use water-surface and groundwater heads correctly
- 4. Recognise connected and disconnected losing streams
- 5. Define baseflow without equating it with all groundwater
- 6. Make the river reach a water-balance control volume
- 7. Use seepage runs as one line of evidence
- 8. Observe the streambed and near-stream sediments
- 9. Understand the hyporheic exchange zone
- 10. Explain bank storage during a rising river
- 11. Connect pumping to streamflow depletion carefully
- 12. Include wetlands, lakes and springs in the same framework
- 13. Track dissolved chemistry across the interface
- 14. Understand contaminants and receptor protection
- 15. Use models with an explicit stream–aquifer boundary law
- 16. Separate observations, model calibration and validation
- 17. Consider environmental and social consequences together
- 18. Transfer the method across seasons and catchments
1. Begin by treating groundwater and rivers as one system
A river is not always only surface runoff moving downstream. Water from the adjacent aquifer can enter its bed and banks, while river water can move into subsurface materials. The direction and amount vary with groundwater heads, stream stage, sediment properties, pumping and season. The USGS’s foundational Circular 1139 describes ground and surface water as a single resource in many environmental and water-supply questions.
Start by asking which water exchange is being investigated, over what reach and period, and why it matters. The gold standard is an explanation supported by both surface-water and groundwater observations, with a clear boundary and stated uncertainty.
2. Distinguish gaining, losing and mixed stream reaches
A gaining reach receives net groundwater discharge; a losing reach contributes net water to the subsurface. A single river can contain both types at different locations, and the same reach can change behaviour with hydrologic conditions. An apparently dry streambed may still exchange water with deeper ground.
Map local context rather than assigning one label to a whole river basin. A streamflow difference measured across a long reach is a net result and may conceal local gains and losses. Identify each measurement’s spatial resolution and what the reach classification can establish.
3. Use water-surface and groundwater heads correctly
Where connected conditions apply, groundwater head relative to the river water surface helps indicate the tendency for exchange. A higher adjacent groundwater head may support discharge toward the river; a higher river stage can promote water entering the bank or bed. But the detailed exchange depends on bed conductance, vertical gradients and subsurface connectivity.
Measure heads with consistent elevation references and relevant screen depths. A well far from the stream may not represent the gradient at the bed. Do not infer a complete three-dimensional flow system from one pair of head measurements.
4. Recognise connected and disconnected losing streams
A losing stream can transfer water directly into a saturated connected aquifer or infiltrate through an unsaturated zone above a deeper water table. These settings can have very different responses to changes in groundwater levels and pumping. USGS Circular 1139 explains that disconnection conditions and transient recharge affect the interpretation.
Investigate geology, streambed layers, unsaturated thickness and head observations. A lower nearby groundwater table does not automatically establish the rate of leakage. The field evidence must support the type of hydraulic connection assumed.
5. Define baseflow without equating it with all groundwater
Baseflow is a component of streamflow sustained between precipitation events, commonly associated with groundwater and other delayed storage pathways. Its estimation may rely on hydrograph methods or direct hydrologic studies. The interpretation varies with catchment conditions, and a separated hydrograph component is not always a direct sample of one aquifer.
State the estimation method, timescale and assumptions before comparing baseflow fractions. Streams can also receive water from bank storage or other transient sources. A larger baseflow fraction does not automatically mean more groundwater resource is safely available for extraction.
6. Make the river reach a water-balance control volume
A reach balance compares upstream discharge, downstream discharge, tributaries, diversions, rainfall, evaporation and changing storage over an appropriate period. The residual may indicate net exchange with groundwater, but only under the measured and assumed terms. A few discharge observations cannot resolve a complicated flood pulse if flow conditions change between station measurements.
Define the signs of every input and output. In a seepage run, close-in-time flow measurements during stable conditions can help estimate net gains or losses. Report measurement uncertainty rather than treating a small residual as a precise new aquifer discovery.
7. Use seepage runs as one line of evidence
A seepage run measures discharge at multiple locations along a stream and accounts for tributary inflows and relevant withdrawals. The calculated difference can estimate net water exchanged along a chosen reach. USGS field guidance notes that changing flow conditions and uncertain discharge measurements can limit the result.
Choose conditions and stations suited to the study question, and recheck them under more than one season if the exchange is expected to vary. A net zero result does not prove every metre of streambed is inactive: local gains and losses can cancel.
8. Observe the streambed and near-stream sediments
The streambed can contain sand, gravel, clay, organic material and biological deposits with different permeability. A low-conductivity layer can restrict exchange despite a clear head difference, while permeable gravels can support strong local pathways. Sediment properties may change through scour, deposition or clogging.
Document bed material, channel geometry, hydraulic conditions and disturbance history. A core taken from one pool cannot automatically represent every riffle or bank. Conductance used in a model should be tested against observations rather than chosen solely to obtain the expected flow.
9. Understand the hyporheic exchange zone
The hyporheic zone is a region beneath and alongside streams where surface and groundwater interact over local flow pathways. Water can enter through one patch of streambed and return nearby, producing chemical and ecological conditions different from the open river or deeper aquifer. Pool–riffle sequences, meanders and bed structures can influence these paths.
Treat hyporheic exchange as spatially and temporally variable. A chemical tracer measured at one depth may reflect a mixture, not one pure endmember. The role of the zone in nutrient cycling and habitat needs both hydrologic and biological evidence.
10. Explain bank storage during a rising river
When river stage rises above adjacent groundwater heads, some river water can move temporarily into the banks. As the river level falls, stored water may drain back. USGS Circular 1139 describes this transient bank-storage mechanism, which can complicate the apparent relation between streamflow and groundwater.
The time sequence matters: a stream can lose water during a high-water event and later gain part of that water. Monitoring only one moment may assign the reach the wrong long-term behaviour. Compare stages, groundwater heads and relevant chemistry across the rise and recession.
11. Connect pumping to streamflow depletion carefully
A well pumping from a hydraulically connected aquifer can reduce groundwater that would otherwise reach a stream, or under some conditions induce river water to enter the aquifer. The resulting streamflow depletion can lag pumping and depend strongly on aquifer properties, distance, boundaries and time.
USGS research published in 2024 highlights the challenge of separating pumping impacts from climate and other human influences. A well located near a river is not sufficient evidence of a particular amount of depletion. Use a defensible flow model and observations suited to the decision.
12. Include wetlands, lakes and springs in the same framework
Groundwater–surface water interaction is not limited to rivers. Wetlands, lakes and springs may receive or lose groundwater depending on local and regional gradients, substrate, seasons and changes in water levels. Wetland vegetation and ecological processes can depend on these exchanges even when the water movement is not visually obvious.
Measure both hydraulic and ecological conditions before inferring dependency. A pond’s changing level might reflect rainfall, evaporation, seepage, pumping or controlled releases. An appropriate water balance and local geology can help distinguish those alternatives.
13. Track dissolved chemistry across the interface
Surface water and groundwater can carry different temperatures, dissolved ions, oxygen and nutrients. Mixing and reactions at the interface can alter the chemical environment for aquatic organisms and downstream users. A concentration difference may arise from distinct sources, but can also reflect biological or mineral reactions.
Choose analytes and sampling positions matched to the hypothesis. A chemical signature alone seldom proves a unique direction or amount of exchange. Cross-check with temperature, hydraulic heads and independent tracers where feasible.
14. Understand contaminants and receptor protection
Polluted groundwater can discharge to a stream, while contaminated surface water may infiltrate toward wells. Exposure depends on contaminants, transport, reactions, dilution and actual water use. A low concentration at one surface-water station does not prove contaminated groundwater is absent or that downstream ecology is protected.
A risk assessment requires suitable contaminant-specific evidence and competent local authority. Monitor relevant source zones and receptors, and preserve the distinction between a hydrological exchange calculation and a regulatory water-quality determination. This educational guide does not certify any water source as safe.
15. Use models with an explicit stream–aquifer boundary law
Groundwater models can represent stream-aquifer exchange using conductance relationships or more elaborate coupled flow equations, depending on their purpose and software. A simple linear conductance calculation may help explain direction and magnitude, but it assumes a particular connection and effective bed property.
State the stream stage, groundwater head, bed elevation, geometry and conductance method. The same model can fit groundwater levels while misrepresenting river depletion. Independent streamflow, chemistry or temperature evidence can be important when the decision concerns surface-water outcomes.
16. Separate observations, model calibration and validation
A model tuned to one set of monitoring wells and seepage-run data may not predict conditions after new pumping or a severe drought. Numerical closure and a low calibration residual do not prove the conceptual representation or boundary conditions are correct. Alternative aquifer structures can sometimes explain the same measured heads.
Preserve which observations were used for calibration and which were reserved for independent tests. Compare predictions across time and space, and state where data are sparse. If the result drives a consequential decision, investigate whether plausible model alternatives change that decision.
17. Consider environmental and social consequences together
Water exchange can affect drinking-water availability, irrigation, wetland ecosystems, stream temperature and aquatic habitat. Reducing groundwater pumping or restoring riverbank vegetation might improve one outcome while shifting burdens elsewhere. A purely volumetric analysis can miss ecological and community priorities.
Define the affected users and ecosystems before choosing the modelling scale and indicators. Include appropriate professional, regulatory and community input. A decision about interconnected waters should not use the artificial boundary between ‘groundwater’ and ‘surface water’ to conceal impacts.
18. Transfer the method across seasons and catchments
A gaining stream in a wet mountain valley, a losing reach in a dry basin and a pumped urban aquifer need different models and measurements. The transferable method is to define the interface, gather consistent heads and flows, build a water and solute balance, and test physical explanations over meaningful time.
A gold-standard report can say where exchange was measured, where only inferred, how uncertain it is and which future observations would improve the answer. It treats groundwater and surface water as a connected resource rather than separate administrative boxes.
Worked fictional case A: a river gains water after the known tributary is accounted for
In a fictional stable-flow seepage run, upstream discharge is 2.5 cubic metres per second and downstream discharge is 3.1 cubic metres per second. A measured tributary contributes 0.2 cubic metre per second along the reach, while a documented diversion withdraws 0.1 cubic metre per second. Assume no significant storage change, evaporation or other unmeasured inflow during the selected interval.
Net unaccounted gain is downstream minus upstream minus tributary plus diversion: 3.1 − 2.5 − 0.2 + 0.1 = 0.5 cubic metre per second. Under the stated assumptions, that is a candidate net groundwater contribution. It does not prove every section of the bed is gaining, and the 0.5 figure has no assessed uncertainty until instrument and timing errors are considered.
Worked fictional case B: a conductance estimate is a model, not a measurement
Suppose a simplified saturated streambed area of 200 square metres has hydraulic conductivity 0.00001 metre per second and an effective flow-path thickness of 0.25 metre. With a stream-to-aquifer head difference of 0.5 metre under the chosen reference, Darcy’s idealised leakage estimate is Q = K × A × Δh ÷ L = 0.00001 × 200 × 0.5 ÷ 0.25 = 0.004 cubic metre per second, or 4 litres per second.
This invented calculation assumes a representative uniform bed, saturated connection and suitable one-dimensional flow. Sediment variability, partial disconnection, local head gradients, clogging or flow path geometry can change the true exchange. A river-stage reading and one piezometer are not enough to validate the computed flow.
Integrated stream–aquifer investigation checklist
- Define the study reach, aquifer and water-budget interval.
- Verify vertical datums, screen depths, field methods and timestamps.
- Measure upstream, downstream, tributary and diverted flows where relevant.
- Assess both gaining and losing local zones despite net reach results.
- Record streambed geometry and plausible hydraulic conductivity.
- Cross-check hydraulic interpretation with tracers or temperature where suitable.
- Model pumping impacts with attention to delayed response.
- Preserve data-quality exceptions, uncertainty and receptor-protection implications.
Frequently asked questions
What is groundwater–surface water interaction?
Exchange of water and dissolved substances between rivers, lakes, wetlands and the subsurface.
What is a gaining stream?
A reach receiving net groundwater discharge through its bed or banks under the conditions studied.
What is a losing stream?
A reach where surface water moves into subsurface materials, potentially recharging groundwater.
Can the same stream both gain and lose groundwater?
Yes, at different locations or periods, including temporary reversals during water-level changes.
What is a seepage run?
A series of streamflow measurements used to infer net reach gains or losses after accounting for relevant inflows and withdrawals.
Does a zero net gain mean no groundwater exchange?
No. Local gains and losses may cancel across the measured reach.
What is the hyporheic zone?
A subsurface region beneath and beside streams where water exchange, chemistry and ecology interact.
What is bank storage?
Temporary movement of stream water into adjacent banks during higher river stage, with potential later return.
Can groundwater pumping affect river flow?
Yes, where the systems are connected; the magnitude and timing depend on geology, pumping and boundaries.
Is baseflow always entirely from one aquifer?
No. Estimation and delayed water-source contributions depend on catchment and method.
Does a head difference prove a particular leakage rate?
No. Bed properties, connection, geometry and modelling assumptions determine exchange.
Why does interface chemistry matter?
Exchange can transport nutrients and contaminants and influence aquatic or groundwater conditions.
Connected reading: hydrology, aquifers and freshwater ecology
Hydrology follows the water balance, hydrogeology provides aquifer structure, hydrochemistry studies the exchanged solutes, groundwater modelling evaluates selected scenarios and ecology considers the consequences for living systems. These are separate but complementary parts of the evidence.
- The Gold Standard of Hydrology
- The Gold Standard of Hydrogeology
- The Gold Standard of Hydroinformatics
- The Gold Standard of Groundwater Modelling
- The Gold Standard of Groundwater Remediation
- The Gold Standard of Hydrochemistry
- The Gold Standard of Wetland Hydrology
- The Gold Standard of Hydrodynamics
- The Gold Standard of Geospatial Analysis
- The Gold Standard of Ecology
Official USGS sources and interpretation limits
- USGS Circular 1139 — Ground Water and Surface Water: A Single Resource
- USGS — Interaction of Groundwater and Streams
- USGS — Rivers Contain Groundwater
- USGS — Interconnected Surface Water and Streamflow Depletion
- USGS — Streamflow Depletion Research Priorities (2024)
- USGS — Groundwater and Surface-Water Interactions
USGS sources were reviewed on 29 October 2026. The 1998 Circular 1139 is a foundational account of exchange mechanisms; newer USGS resources discuss pumping-related streamflow depletion and updated public explanations. Both flow calculations use fictional data and do not certify an actual water body’s condition, contamination status or extraction capacity.
Return to chapter index · The Gold Standard hub · How X Works