Dewatering a well means removing water as fast as the aquifer supplies it so that digging can go deeper, and the Thiem equation for flow to a well, Q = πK(H² − h²) ÷ ln(R ÷ r), shows how fast that is: in a sand with K = 1×10⁻⁴ m/s, a crew bailing 30 L/min holds the water only about 0.9 m below its natural level, because inflow grows with every centimetre of extra drawdown. Measure your own well’s inflow rather than trusting a textbook K, never pump the shaft so low that the bottom boils, and keep engine exhaust out of the shaft.
Part of Deep Wells
Dewatering is the removal of groundwater from an excavation so that work can continue below the water table. In a hand-dug well it is done by bailing or pumping while the diggers deepen the shaft into the saturated zone.
Inflow Rises With Drawdown, So Bailing Has A Ceiling
Once the shaft passes the water table it behaves like a pumped well in reverse: the lower you hold the water inside, the harder the aquifer pushes water in. Dewatering works only while you remove more than flows in. The balance point is where the inflow at a given drawdown equals the rate at which your crew or pump can lift water.
Outflow. A crew hauling a bucket of volume V every t seconds removes V ÷ t. A 15 L bucket every 30 s is 15 ÷ 30 = 0.5 L/s = 30 L/min = 5.0×10⁻⁴ m³/s.
Inflow, from the Thiem (Dupuit) equation for an unconfined aquifer. It is a standard hydrogeology result, recalled here from the textbook form rather than derived: for a well of radius r_w, water held at depth h_w above an impermeable base, in an aquifer of saturated thickness H and hydraulic conductivity K, with a radius of influence R beyond which the well has no effect,
Q = π × K × (H² − h_w²) ÷ ln(R ÷ r_w)
with Q in m³/s, K in m/s and all lengths in metres. It assumes a uniform aquifer on an impermeable base and a steady state; real aquifers are layered and take time to reach it, so treat the result as an order of magnitude.
Worked Example: A Sand Well
Inputs (illustrative values, not measured ones): K = 1×10⁻⁴ m/s (a sand; 8.6 m/day), H = 5 m, r_w = 0.6 m (a 1.2 m shaft), R = 100 m (assumed).
ln(R ÷ r_w) = ln(100 ÷ 0.6) = ln 166.7 = 5.116.
- Water held at h_w = 4.5 m (0.5 m of drawdown): Q = π × 10⁻⁴ × (25 − 20.25) ÷ 5.116 = 2.92×10⁻⁴ m³/s = 17.5 L/min. A 30 L/min crew copes.
- Water held at h_w = 1 m (4 m of drawdown): Q = π × 10⁻⁴ × (25 − 1) ÷ 5.116 = 1.47×10⁻³ m³/s = 88 L/min. The crew loses.
How far can 30 L/min hold it down? Set Q = 5.0×10⁻⁴ m³/s and solve for h_w: H² − h_w² = Q × ln(R ÷ r_w) ÷ (π × K) = 5.0×10⁻⁴ × 5.116 ÷ (π × 10⁻⁴) = 8.14. So h_w² = 25 − 8.14 = 16.86, h_w = 4.11 m, and the drawdown is 5 − 4.11 = 0.89 m, about 0.9 m.
Digging 2-3 m into the saturated zone, as the Deep Wells guide specifies, therefore cannot be done dry with bucket bailing in this sand. The work below the first metre or so of water is done standing in water, or by sinking rings from which soil is removed from under the water, or with a faster pump.
Sensitivity. The result depends on the log of R, so it is forgiving: R = 30 m gives 1.31 times the inflow of R = 100 m, and R = 300 m gives 0.82 times. K is not forgiving. Sands range over several orders of magnitude, so an assumed K can be wrong by 100 times. Measure the real inflow: bail the well down, time the recovery, and measure the actual L/min at a known drawdown (Well Digging gives the overnight recovery test). Use the measured inflow, not the equation, to size the crew.
Pumping The Bottom Too Low Makes The Sand Boil
If water outside the lining stands much higher than water inside, it pushes up through the bottom. When the upward gradient reaches the critical value the grains lose contact and the bottom turns to a quick, boiling slurry; the lining loses its support and the shaft can collapse.
- Critical gradient i_c = (γ_sat − γ_w) ÷ γ_w = (18 − 9.81) ÷ 9.81 = 0.83, for an assumed soil weight γ_sat = 18 kN/m³
- Rough estimate of the actual gradient: water travels down outside the lining and up inside, a path of about twice the depth d the lining toe is embedded below the bottom, so i ≈ s ÷ (2d), where s is the head difference. This is a simplification of my own, not a design formula
- Example: s = 0.89 m (the drawdown above) and d = 1 m gives i ≈ 0.45, below 0.83; with d = 0.5 m it gives i ≈ 0.89, above it
If sand bubbles up at the bottom or the floor turns to mud, stop pumping at once and let the shaft refill, or flood it, so that the pressures balance; then continue by sinking the lining deeper or digging under water.
Methods, From Simple To Fast
| Method | Rate | Notes |
|---|---|---|
| Bucket on windlass | V ÷ t, e.g. 30 L/min | The usual method; needs a hoist and a crew |
| Suction hand pump with hose | Pump limit; lift of 6-7 m at most | The pump must be near the water (Suction Pump) |
| Force pump with cylinder in the shaft | Pump limit; any lift | See Deep-Well Pumps |
| Dig under water; sink rings | No dewatering needed | Soil is removed with a grab or bucket from under the water (Concrete Rings) |
Common Mistakes
- Assuming a textbook K. Measure the inflow of your well.
- Digging dry in running sand. The bottom boils; dig under water instead.
- Forgetting that every metre of drawdown multiplies the load. Inflow rises with drawdown, so the last metre is the hardest.
- Stopping the bailing for a rest with the diggers still below. The water rises back and the sump fills.
Safety
Water, sand and exhaust all kill in a shaft
The shaft is a confined space; ventilate and use a harness, a line and two people above on every entry (Safety Protocol). An engine-driven pump gives off carbon monoxide, which has no smell: keep any engine and its exhaust well away from the mouth of the shaft and never lower it in. A boiling bottom and a collapsing lining give little warning; the first sign is sand moving or the floor going soft. Haul the digger out on the line and never go down to rescue.
What’s Next
- Learn next: Well Protection — finishing the well once the water is reached
- Combine with: Concrete Rings — sinking the lining while digging under water