Reference

Pipe Flotation and Minimum Cover at Water Crossings

An empty pipe below the water table displaces more water than it weighs, and it will float. The cover required to hold it down is a force balance, not a rule of thumb - and on large diameters it governs the crossing depth outright, overriding whatever minimum the feature class would have set.

The force balance

From ACPA Design Data 22, Flotation of Circular Concrete Pipe, per linear foot of pipe:

net downward force = W_pipe  -  W_water  +  (W_soil / FS)

W_pipe - the pipe’s own weight per foot
W_water - buoyant uplift, the weight of water displaced by the outside diameter (62.4 pcf, fresh)
W_soil - the weight of backfill above the pipe, divided by a factor of safety in the range 1.0 to 1.5

If the sum is negative, the pipe floats. Solved for cover rather than evaluated at a single depth:

cover = (uplift - pipe_weight) x FS / (w_I x D)

where w_I is the inundated soil unit weight and D is the outside diameter in feet.

Using inundated rather than dry unit weight is the part most often gotten wrong. Submerged backfill weighs substantially less than the same soil dry, and a calculation that uses the dry figure overstates the restraint by a wide margin - in the unsafe direction.

What it works out to

Cover required to resist flotation. Empty pipe, permanently submerged, sandy backfill at 110 pcf dry and specific gravity 2.65 - ACPA’s own worked Example 1 values.

Nominal diameter	Cover required
8 in	0.72 ft
12 in	1.07 ft
18 in	1.61 ft
24 in	2.15 ft
30 in	2.68 ft
36 in	3.22 ft
48 in	4.29 ft
60 in	5.37 ft

The crossover sits almost exactly at a 2 ft rule

Below about 22 inches, a 2 ft minimum-cover requirement already dominates and flotation never binds. At 24 inches flotation takes over, and from there it grows linearly with diameter - by 60 inches it is asking for more than twice the flat minimum.

Which means a design that applies one cover figure across all diameters is over-conservative at the small end and under-designed at the top of its own range.

The governing rule

Required depth at a water crossing is the larger of the two:

depth = max( minimum cover requirement,
             cover required to resist flotation )

Not one or the other. Two checks reported independently will eventually contradict each other at the same station - a crossing reported as clearing its 2.0 ft requirement while a buoyancy check simultaneously reports the pipe floating at that exact depth. Deriving the flotation depth from the same force balance the check uses makes that disagreement structurally impossible rather than merely unlikely.

Conservative assumptions worth naming

Anyone relying on a number like this deserves to know which way it errs.

Empty pipe. A pipe in service carries flow that helps hold it down. Empty is the worst case, and it is also the real case during construction and after a line is taken out of service.
Permanently submerged. At a water crossing this stops being conservatism and becomes simply accurate.
Soil prism above the pipe rather than the full trapezoidal backfill wedge. The prism is the smaller of the two volumes, so this understates restraint - it can ask for more cover than strictly needed, never less.
Pipe self-weight ignored where it is not known. Ignoring a real resisting force can only ask for more cover.

Every one of those errs toward more cover, which is the correct direction for a screening calculation. A study should never report a pipe as safe that is not.

What this does not cover

Project geotechnical data replaces the default soil values. The 110 pcf / SG 2.65 figures are ACPA’s own worked-example values - real and cited, but a stand-in for a soils report rather than a substitute for one.

Scour is a separate question with a separate method. Flotation sets a minimum; scour can set a deeper one.