In my local area there is a creek which dries up completely in the summer and is flowing strongly in the winter. Is it the case that, at the location where the creek dries up, it continues flowing underground?

If so does it take the same trajectory as the creek? At what typical depths does it start flowing "forward"?

  • 1
    $\begingroup$ Yes, at least sometimes, as you can perhaps verify by digging a hole in the creek bed. $\endgroup$
    – jamesqf
    Commented Jul 7, 2015 at 17:24

2 Answers 2


Water in liquid state always takes the fastest route to lower altitude. All rainwater would flow vertically down to the groundwater surface if not the permeability was limited by the material properties.

The permeability controls how fast water infiltrate and penetrate the material. For example, can surface runoff water flow on top of a hard road coating, while water from a watering-can goes into the soft soil in the flower-bed. Gravel, sand and fractured rocks have high permeability whiles shale and non-fractured igneous rock have low permeability. So low, that in many cases it appears to be totally impenetrable on a human time scale.

The same goes for a river. If the river flows on permeable rocks, typical sediments deposited by the river itself (see picture), ground water will also flow in the sediments under the river bed, permitting that water is still entering upstream.

A better way to understand a river or lake in this context, is that it is the surface manifestation of the groundwater table. If you dig a well next to a body of water you’ll reach ground water at the same altitude as the river water surface. The gradient of the groundwater surface and the permeability of the material controls the direction and velocity of the flow.

However, if more water enters a system than the surface can infiltrate, the surplus will flow as surface runoff, typical for wadis (See picture) in desert areas.

Wadi. Photograph taken by Mark A. Wilson, The College of Wooster

So, to answer your question. It depends on the local geology, but ground water will always flow from the precipitation area to the oceans (or a lower point where it can evaporate or being pumped up). Ground water exists everywhere and the gradient of the hydrostatic pressure controls the direction of the flow. As viable rivers shows the ground water table, it is usually rather safe to assume that the ground water under the dry river flows more or less the same direction.

However, ground water flow can be changed by e.g. pumping from irrigation wells. Imagine a flat lying river bed with a well upstream. When enough water is removed from the well and eventually evaporated, ground water will flow towards the well. The water don't 'remember', but follows the law of gravity.

  • 1
    $\begingroup$ Water always takes the fastest route to lower altitude — unless it evaporates. $\endgroup$
    – gerrit
    Commented Jul 8, 2015 at 13:45
  • $\begingroup$ Good point @gerrit! Water in liquid state always takes the fastest route to lower altitude would be correct. I'll update and expand the answer. $\endgroup$
    – user2821
    Commented Jul 8, 2015 at 21:41
  • $\begingroup$ "Ground water exists everywhere and the gradient of the surface controls the direction of the flow." except when it doesn't earthscience.stackexchange.com/a/270/117 $\endgroup$
    – mart
    Commented Jul 13, 2015 at 14:03
  • $\begingroup$ Answer updated to clarify that it's the surface of the hydrostatic groundwater table, not the topography. (See earlier in the answer). $\endgroup$
    – user2821
    Commented Jul 13, 2015 at 14:31

To answer this question, I'll back up a bit and first talk about what happens to rain as it flow to streams then about what happens when streams dry up.

When it rains, some of the rain hits the ground or trees/plants and is evaporated before it flows away from where it hits and before it sinks into the soil very far. Some enters the soil but is taken up by plant roots and transpires back into the atmosphere. We don't need to worry about those components for this question.

Some of the rain flows down the ground surface, more or less as sheet flow and eventually merges into little flow paths to the creek or river. That's called runoff.

Typically, some rain will enter the ground and flow down to the water table, which is the surface of the saturated groundwater. By that I mean permanent water in the subsurface where all the void spaces in the rock or sediment are completely filled with water (there are some different definitions of groundwater that include the unsaturated zone but I'll keep them separate). Groundwater flows downgradient under gravity towards a stream, river, lake, or the coast.

In between there is sort of a fuzzy situation where water flows in the unsaturated soils, moving downward and or sideways until it concentrates in a perched aquifer or a seepage face at the stream. That water will discharge to the surface water system. Importantly, it typically reaches the surface water more quickly than the groundwater flow but slower than the runoff. Its often called quickflow, interflow, or perched flow. Let's not worry about this too much - you can think of it as slow runoff or fast groundwater that kind of smears out the streamflow peaks.

Streamflow, then, is mostly made up of these 3 components. Why do streams flow between rain storms? It does take runoff some time to reach the stream and flow downstream so it doesn't stop instantly but it goes away the fastest of the three. Quickflow will keep going a bit longer and depends a lot on the geology, soils, and topography. Simplistically, it will keep streams flowing longer than pure runoff would. Saturated groundwater is a slower process in terms of its discharge to a stream and, particularly for big rivers can keep them flowing year round. The first take-away message is the groundwater and surface water are interconnected - at least in the big picture.

A couple of interesting things can happen to the groundwater-surface water connection. First, there might not be enough groundwater flow to keep it discharging to a stream, either seasonally like your creek or permanently in some places. If the stream still has water it can reverse the connection and lose water to the groundwater - a losing stream instead of a gaining stream. Without more rain or flow from upslope the stream will dry up. In some cases the groundwater level can drop faster than the stream drains into the sediments and the system can become disconnected. The groundwater flow direction shifts so it is not towards the stream but towards a more regional discharge location. Since streams also flow down to the regional surface water system, the groundwater will often take a similar path but overall I don't think of it as an underground stream.

Finally though, things aren't always that straight-forward. One thing that often happens is that valleys fill up with sediments that are more permeable than surrounding sediments or rock formations and allow more flow. In that case the groundwater flow is mostly along the stream flow direction in the valleys and stream bed. It then looks similar to an underground river. That can happen when the river or creek is flowing or when it stops. The fancy word for it when the river is flowing (particularly when you are talking about pretty local scale effects) is hyporheic flow - meaning flow below the river.

So that's probably more than you wanted to know, but for your creek it might go each year from pretty much all runoff, to gaining groundwater, to losing to the groundwater, to disconnected, to dry with some subsurface flow through the stream bed to more overall groundwater flow that doesn't see any influence from the stream location. But it depends on the weather, the sediments, how steep the land surface is, and perhaps some other complicating factors.

  • $\begingroup$ Another thing to consider is that in some places, like the Sierra Nevada (where I live), most of the precipitation doesn't come as rain, but as winter snow. So the water in a creek mostly comes from melting snow that moves directly to groundwater, then seeps downhill until it enters the creek from springs &c. $\endgroup$
    – jamesqf
    Commented Jul 10, 2015 at 17:53
  • $\begingroup$ Good and pedagogic explanation of quickflow and recharge. $\endgroup$
    – user2821
    Commented Jul 11, 2015 at 18:19
  • $\begingroup$ Yes snow melt is another factor. You can get much more runoff, especially when the ground is still frozen, or as you say induce more recharge and groundwater flow. Snow pack is an important water store in mountainous areas and melting snow maintains streamflow into summer. $\endgroup$
    – haresfur
    Commented Jul 12, 2015 at 22:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.