← All summaries
Bedrock / aquitard stream hcrit threshold ground surface water table saturated zone vadose zone
Figure 1. Cross-section of a stream–aquifer system. The first slider sets the regional groundwater head relative to stream stage; the regional head is capped at ground surface. The second slider sets the depth below the streambed at which true disconnection occurs (the orange dashed hcrit line). Three regimes result: gaining (groundwater discharges into the stream), losing‑connected (a continuous saturated column persists between streambed and water table, even when the water table is below the streambed), and disconnected (the water table has fallen below hcrit and a vadose zone has formed under the bed). Adapted from Winter et al. (1998); the parametric disconnection threshold follows Brunner et al. (2009, 2011) and Morel‑Seytoux et al. (2018), who show the threshold depends on streambed thickness, the Kaquifer/Kclog contrast, the air‑entry pressure, and stream width — disconnection rarely coincides exactly with the streambed.
The three regimes
Gaining

Groundwater feeds the stream

The water table sits above the stream stage on both sides, so groundwater flows toward the channel and discharges as baseflow. In Mediterranean and snowmelt climates, this baseflow is what keeps streams flowing through the dry season.

Pumping risk: reducing the gradient toward the stream cuts baseflow directly — the dominant SGMA depletion pathway.

Losing — connected

Stream feeds the aquifer

The water table sits below the stream stage but still touches the streambed. Stream water leaks downward; the leakage rate depends on stream stage, streambed conductance, and the head gradient. Often seen on alluvial fans, recharge basins, and mid-elevation reaches.

Pumping risk: further drawdown increases the loss rate — still a depletion of surface water — until disconnection.

Disconnected

Unsaturated gap below the bed

A vadose zone develops beneath the streambed. Seepage flows through the channel sediments at a rate fixed by streambed properties (unit‑gradient flow) and is independent of how deep the water table has fallen below. Common in heavily‑pumped semi‑arid alluvial basins.

Pumping risk: further pumping doesn't increase per-unit leakage but can dry the stream entirely by exceeding inflow.

Why this is hard to measure

Exchange is spatially heterogeneous

A single reach can be gaining in one segment and losing 50 m away. Bedrock highs, paleochannels, beaver ponds, riffle-pool sequences, and clay lenses all create local reversals. Differential gaging picks up the net but misses the structure (Woessner 2020).

Practical implication: point measurements rarely scale up reliably. Multiple lines of evidence — gaging, temperature, head data, isotopes — are typically needed to characterize a reach.

The hyporheic zone matters

Between the open channel and the regional aquifer lies the hyporheic zone — sediments saturated with a mix of stream water and groundwater. Exchange here drives oxygen, nutrient, and temperature dynamics that determine whether a stream supports aquatic life and riparian vegetation.

For SGMA, this is where "depletion" becomes ecological harm: a stream that loses its hyporheic exchange and shallow groundwater connection can lose its salmonids, frogs, and willows long before it goes visibly dry.

Vocabulary the rest of the series uses

Head, gradient, and discharge

Hydraulic head is the energy elevation of groundwater, measured in a well. Gradient is the slope of head between two points. Discharge per unit area = hydraulic conductivity × gradient (Darcy's law). For a stream–aquifer interface, the exchange flux is Kbed · (hstream − haquifer) / bbed.

Connection vs. disconnection — and where the threshold really sits

A stream is connected if a continuous saturated column exists from stream bottom to the regional water table; it is disconnected if a vadose zone has formed underneath. The simple picture places this threshold at the streambed, but Brunner et al. (2009, 2011) showed that the true threshold hcrit typically sits well below the streambed. The depth grows with: (a) the Kaquifer/Kclog contrast, (b) clogging-layer thickness, (c) capillary fringe (air-entry pressure), and (d) stream width. Morel‑Seytoux et al. (2018) propose a refined exchange formulation that gives a smoother, dimensionally consistent transition. Disconnection is also hysteretic — recovery requires the water table to rise back above the streambed, not just back above hcrit.

Baseflow & bank storage

Baseflow = the portion of streamflow coming from groundwater discharge. Bank storage = water temporarily stored in alluvial banks during high flows and returned later. Both are first‑order budget terms; pumping affects baseflow directly and can convert temporary bank storage into permanent depletion.

Interconnected surface water (ISW)

SGMA term of art: "surface water that is hydrologically connected at any point by a continuous saturated zone to the underlying aquifer." A reach is ISW if and only if it is in the gaining or losing-connected regimes above. Once disconnected (per the regulatory test), it is no longer ISW — but it can become ISW again if heads recover.

Key references in the project library

  1. Winter, Harvey, Franke & Alley (1998). Ground Water and Surface Water — a Single Resource. USGS Circular 1139. (Foundational primer; figures here adapt its cross-sections.)
  2. Sophocleous, M. (2002). Interactions between groundwater and surface water: the state of the science. Hydrogeology Journal 10(1): 52–67.
  3. Brunner, Cook & Simmons (2009). Hydrogeologic controls on disconnection between surface water and groundwater. Water Resources Research 45, W01422.
  4. Brunner, Cook & Simmons (2011). Disconnected surface water and groundwater: from theory to practice. Groundwater 49(4): 460–467.
  5. Morel‑Seytoux, Miller, Mehl & Miracapillo (2018). Achilles heel of integrated hydrologic models: the stream–aquifer flow exchange, and proposed alternative. J. Hydrology 564, 900–908. (Refined exchange formulation; argues against locating disconnection at the streambed.)
  6. Woessner, W. (2020). Groundwater–surface water exchange. The Groundwater Project.
  7. Fuchs, Niswonger, Hopkins & others (2019). Quantifying disconnection of groundwater from managed-ephemeral surface water. WRR 55, 10377–10394.