Cross-section takeoff vs. surface-to-surface volumes: what holds up on a pay app

Every super who has fought over an earthwork pay app knows the drill. The earthmoving sub submits a load count, the GC's estimator runs a cross-section takeoff off the design file, and the two numbers don't match. Someone eats the difference, usually the party with the weaker backup.

So which method should be the basis for the number you sign off on: the cross-section takeoff your survey crew has always run, or a surface-to-surface volume comparison pulled from drone capture? They're not the same calculation, and they don't fail the same way.

How a cross-section takeoff works

A traditional cross-section takeoff samples the site at stations, usually every 50 or 100 feet along the alignment, and assumes the ground between those stations behaves the way the interpolation says it should. That assumption is fine on a straight grade with consistent slope. It falls apart on anything irregular: a stockpile with a ragged toe, a cut that daylights unevenly, a borrow pit that's been worked from three different directions by three different operators.

The crew still has to walk or drive the site to shoot those stations, which means weather windows, access around active equipment, and a field day that competes with the schedule you're trying to protect. And once the shots are taken, you've got a snapshot of lines, not the ground itself. Anything that happened between stations is invisible to the model, even though it's very visible to the sub's own load count.

That's the gap that turns into a pay app argument. The cross-section model says one volume. The truck tickets say another. Neither side has a full picture of the ground to point to, just a set of sampled lines and a stack of receipts.

What surface-to-surface differencing changes

Surface-to-surface volume comparison skips the sampling step entirely. Instead of shooting stations and interpolating, you fly the whole site and build a continuous model from the imagery, then difference that model against a prior capture (or against the design file) to get cut and fill volumes for the full footprint, not just the stationed lines. Earthwork Volumes runs this off weekly drone (or crewed aircraft) capture at under 10 cm GSD, RGB, so the resulting model reflects the actual ground, ragged toes and all.

The practical difference shows up at the pay app meeting. Instead of a cross-section model that covers maybe 60% of the site's actual relief and interpolates the rest, you've got a volume number tied to a full capture taken that week. When the earthmoving sub's load count and your volume number disagree, you're not arguing lines against tickets anymore. You're pointing at a dated capture and asking them to explain the gap.

It also changes the haulage math. A cut/fill volume broken out by area of the site tells you where material still needs to move and roughly how much, and that area-by-area number is what haul planning runs on.

Where each method still has a place

None of this means cross-section takeoff disappears. Design verification against a stationed alignment, drainage grade checks, anywhere the geometry itself matters more than the bulk volume, that's still cross-section territory, and a project engineer isn't going to stop running those checks. The cases where surface-to-surface differencing earns its keep are the ones cross-section takeoff was never built for: irregular stockpiles, active borrow areas, and any pay app where "trust the sub's count" isn't good enough backup.

If you're deciding which number to cite on your next draw, Earthwork Volumes is built for exactly that argument, a weekly cut/fill volume from surface differencing you can put in front of the owner's rep without reaching for a cross-section model that only covers part of the site.

Next time the load count and the takeoff don't match, it's worth asking which one saw the whole site this week.