How the two differ in kind
Excavating below the water table means dealing with water that wants to flow into the hole. There are two ways: lower the water table so the hole is above it, or build a barrier so the water cannot get in. They are not interchangeable, and the ground usually chooses.
PERMEABILITY decides whether pumping is even an option. It spans roughly ten orders of magnitude between a clean gravel and a clay, which makes it the most variable parameter in geotechnics by a wide margin. In sand and gravel a wellpoint system lowers a water table quickly and cheaply. In clay there is nothing to pump: the water does not move on any timescale that matters, and the problem in clay is not inflow but the base heaving or a perched layer letting go.
The second constraint is what a cut-off can key into. A wall that stops short of an impermeable stratum does not stop the water — it lengthens the path, and the flow comes UNDER it instead. A partial cut-off is a legitimate design, but it is a design that still needs pumping inside, and it should be chosen knowingly rather than discovered when the excavation fills.
The factors that actually differ
Show
| Dewatering — wellpoints or deep wells | Cut-off — secant or contiguous piles | |
|---|---|---|
| What the ground has to be | Permeable enough for water to flow to a well. In silt it is marginal and needs vacuum assistance; in clay it does essentially nothing. | Indifferent to permeability for the wall itself — but it needs a competent impermeable layer to toe into, or the water underflows. |
| What happens outside the works | A cone of depression extending well beyond the site. Lowering the water table raises effective stress in everything under it, so compressible ground consolidates and what is standing on it settles. | Almost nothing. The water table outside is largely unchanged, which is the reason this option exists on a congested site. |
| Dependence on continuity | Absolute. The system runs for the whole duration, and an interruption refloods the excavation and can float a partly built structure. | None. The wall works while nobody is watching it. |
| Mobilisation | Light and fast — a header main and a row of wellpoints can be running in days. | A rig, a guide wall, a concrete supply and a programme. It is a construction operation in its own right before the excavation starts. |
| What it becomes afterwards | Nothing. The kit leaves and the water table returns, which is a consideration for anything buoyant that was built in the dry. | Frequently permanent — a secant wall is often the basement wall, so part of its cost is not a temporary works cost at all. |
| Consents | Abstraction and discharge are both usually regulated: taking the water out needs permission, and putting it somewhere needs more. Contaminated ground changes the discharge problem completely. | A construction operation rather than an abstraction one, so the regulatory load is lighter — though the wall's toe may still need agreement where it affects a groundwater body. |
| Certainty of the design input | Rests on permeability, which is the least reliable number in geotechnics unless it came from a pumping test on the actual ground. A table value can be out by an order of magnitude. | Rests on knowing where the impermeable stratum is and that it is continuous — a stratigraphy question the boreholes may or may not have answered. |
| Shape of the cost | Low to start, then a rate per week for as long as the excavation is open — which makes a delay expensive in a way a fixed cost is not. | High and largely fixed at the start, insensitive to programme, and partly recovered where the wall is permanent. |
Which one, and when
Choose dewatering — wellpoints or deep wells when…
- The ground is sand or gravel, where a modest system lowers the table quickly.
- The excavation is open ground with nothing within the drawdown area that settlement would damage.
- The duration is short, so a weekly rate does not accumulate into the cost of a wall.
- The excavation is shallow enough that a wellpoint system reaches, without staging.
Choose cut-off — secant or contiguous piles when…
- There are buildings, services or a highway inside the radius a drawdown would reach.
- There is a reliable impermeable stratum at a depth the wall can toe into.
- The wall is wanted permanently — as a basement wall or a retaining structure — so most of the cost buys something that stays.
- The ground is contaminated, and drawing the plume toward the works or having to treat the discharge is worse than walling it out.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- How far away does the drawdown reach?
- Further than most people expect, and the honest answer is that the figure used to describe it is an empirical fitting parameter rather than a measurement. The classical well equations need a radius at which drawdown is taken as zero, and the expressions for it — Sichardt's among them — are correlations, with different sources giving different ones. They survive because the radius enters the yield calculation logarithmically, so a factor-of-two error changes the pumped flow modestly. What it does not do is tell you where the settlement stops. A real aquifer has no boundary; it has a cone that thins with distance and interacts with rivers, boundaries and other wells. Where the answer matters, the route is a pumping test with monitoring wells — not a formula.
- Why would pumping water out damage a neighbour's building?
- Because water in the ground is carrying part of the load. Terzaghi's effective stress principle says the stress that matters at the grain contacts is the total less the pore water pressure — so removing the pore pressure INCREASES the stress the soil skeleton carries, and compressible layers consolidate under it. The ground settles, and it settles across the whole cone of depression rather than only inside the excavation. Buildings on shallow foundations within that area settle with it, timber piles above a lowered water table can rot, and a nearby well can go dry. This is why dewatering schemes near existing structures come with monitoring, with trigger levels agreed in advance, and sometimes with recharge wells whose only purpose is to hold the water table up outside the works.
- Can I dewater in clay?
- Not in any useful sense. Water does not move through clay on a construction timescale, so wellpoints pump air and achieve nothing. The problems clay presents below the water table are different ones: base heave, where the weight of the remaining clay is insufficient to resist the pressure in a permeable layer beneath it, and perched water in sand lenses within the clay that produces sudden local inflows. The remedies are also different — relief wells to depressurise the underlying stratum, or a deeper cut-off — and reaching for a wellpoint system because it worked on the last job is the commonest way to waste a fortnight before a site investigation is finally read properly.
- What is a partial cut-off, and is it a failure?
- A wall that does not reach an impermeable stratum, so groundwater flows underneath it into the excavation. It is a deliberate design rather than a mistake: the wall lengthens the seepage path and cuts the inflow substantially, and a modest pumping system deals with what gets under. The reason to do it is that the impermeable layer may be far too deep to reach economically. What matters is knowing which you have bought, because a partial cut-off still needs a pump, still needs power, and still produces some drawdown outside. Designing it as a full cut-off and finding out on site is the version that goes wrong, and it goes wrong at the moment the excavation reaches formation.
- What about ground freezing?
- It works in ground where nothing else does — silts, mixed fills, ground with obstructions — because it makes a barrier out of the water rather than around it, and it can form shapes a wall cannot. The costs are time and energy: a frozen wall takes weeks to establish and has to be maintained continuously until the permanent structure is in, so a power failure has a consequence measured in days rather than minutes. It also expands the ground as it freezes and settles it as it thaws, which is its own third-party issue. It is a specialist technique for cases the ordinary two cannot handle, and it belongs in the conversation when the site investigation shows ground that defeats both.
- Do I need permission to pump groundwater?
- In most jurisdictions, twice — once to take it and once to put it somewhere. Abstraction above a threshold volume needs a licence or a notification, and discharge needs a consent whose conditions depend on what is in the water. Both take time to obtain, and both can be refused or conditioned in ways that change the scheme: a discharge limit that the flow exceeds means settlement tanks, treatment, or a phased excavation. Contaminated ground is the case that most often makes dewatering impossible on paper rather than on site, because drawing a plume toward the works or discharging treated water at volume is a regulatory problem rather than an engineering one. Establishing that early is worth more than refining any figure on this page.