Bulking, Swell and the Three Volumes of One Soil

One material, three volumes

Soil in the ground is packed as densely as its own history has made it. Dig it and the grains lose that packing, take in air, and the same material occupies more space. Place it in a fill and compact it, and it is squeezed into a third arrangement — usually denser than it was loose, and often denser than it was in the ground.

The three states have names and every earthworks quantity belongs to exactly one of them. BANK — also called in-situ or solid — is the volume the material occupied before it was disturbed, and it is what a survey of the excavation measures. LOOSE is the volume after digging, and it is what a truck body holds. COMPACTED is the volume it occupies after being placed in lifts and rolled.

Nothing about the material changed. Its mass is the same in all three states, and it is the mass that is conserved — which is the only reliable way to move between them.

Everyone on the job measures a different one

Excavation is normally measured and paid in the BANK, because that is what the drawing describes and what a before-and-after survey gives. Haulage is priced on LOOSE volume, because that is what fills a truck and what determines the number of journeys. Fill is measured COMPACTED, because that is what the specification requires in place.

So a single operation — dig here, cart there, place it — is quantified three times in three different states, by three parties who each believe they are talking about the same soil. Every earthworks dispute worth the name lives in that gap.

The practical discipline is to state the state with the number, every time. A quantity written as a bare volume is ambiguous in a way a bare length is not, and the ambiguity is worth twenty to thirty per cent.

The magnitudes, and why rock is a different problem

Ordinary soils swell modestly. A common earth might occupy a quarter more space loose than it did in the ground, so a hundred units of bank excavation becomes about a hundred and twenty-five to cart — and recompacted into a fill it may finish at around ninety. **One hole, three numbers, and the largest is nearly forty per cent above the smallest.**

Rock is a different order of problem. Blasted or ripped rock swells far more than soil, because the fragments cannot re-pack into the interlocking arrangement they had as an intact mass — and they never will, which is why rock fill has a permanent volume penalty. A cut through rock generates markedly more material to move than its bank volume suggests, and the excess has to go somewhere.

Materials also differ in whether they can be put back at all. Topsoil, peat and highly organic material compact poorly and are usually excluded from structural fill by specification, so their bank volume is a disposal quantity rather than a fill credit — and treating them as available fill is how a balanced earthworks scheme turns into an import.

Where it fails: a cut-and-fill balance struck in the wrong state

This is the error the page exists for. A scheme is said to balance when the cut equals the fill — but cut is naturally measured in the bank and fill in the compacted state, and those two numbers cannot be compared directly. Converting one to the other is the whole job.

A scheme showing equal bank cut and compacted fill volumes is not balanced: it is **short of fill**, because the cut shrinks on its way into the fill. The shortfall is roughly the shrinkage factor times the quantity, and on a large site it is an import that nobody priced.

The reliable route is through mass. Bank volume times bank density gives a mass; that mass divided by the required compacted density gives the compacted volume it will make. Both densities are measurable and the mass is conserved, which is more than can be said for any of the three volumes.

The second failure is applying a single factor to a whole site. Swell and shrinkage belong to a material, and a cutting that passes through topsoil, clay and weathered rock has three of them. A weighted figure taken from a borehole log beats a single number taken from a table.

Where it fails: the truck is not full

Haulage counts assume a truck carries its rated body volume of loose material, and it frequently does not. Two limits apply and either can govern: the body's volume, and the vehicle's legal payload MASS. A dense material fills the weight limit before it fills the body, so the truck leaves part-full and the journey count computed from volume alone is too low.

The opposite case is a light, bulky material — cleared vegetation, insulation, demolition arisings — which fills the body long before the axle limit and hauls air. This is why container and skip quantities are so often disappointing: the limiting factor is how well the material packs, not how much it weighs.

Loading practice matters too. A body heaped above its sides carries more than its struck capacity, and a body loaded by a grab rather than a bucket packs differently. The figures these pages produce assume a normally loaded body and say so; where the count matters, the honest measurement is a tally of actual journeys.

The alternatives: survey, weighbridge, or density test

Each of the three states has a direct measurement and they are all better than a factor. BANK volume comes from a survey before and after — traditionally cross-sections, now routinely a drone or laser scan differenced against the original surface. LOOSE volume comes from a journey tally against a known body size. COMPACTED volume comes from the placed layer's geometry, with the density verified in place.

Where money is involved, the weighbridge is the arbiter, because mass is the only quantity that does not change between states. Paying by the tonne removes the entire argument, which is why bulk materials increasingly are.

The factors on these pages are for planning: sizing a fleet, pricing a disposal, checking whether a scheme is roughly in balance before anyone commits to a level. They are not a substitute for a survey, and where the two disagree the survey is right.