Correlation begins by choosing what to hang the sections on. Hang them on the modern ground and you are drawing topography; hang them on a marker that was once flat — an ash, a flooding surface, a coal — and what you see is what the basin did.
A Wheeler diagram redraws the panel with time up the side instead of rock. Everywhere that deposited becomes a block; everywhere that did not becomes a gap. The white space is the point — most of geological time is missing from most sections, and this is the only diagram that shows it.
A metre of mudstone at three kilometres depth was several metres of mud when it was laid down. Porosity falls off exponentially with depth — φ = φ₀e−cz — and the grains themselves are conserved, so the original thickness follows by putting the water back.
A rate calculated over a long interval is always smaller than one calculated over a short interval in the same rocks, because longer intervals contain more of the gaps. The effect is real and enormous — an order of magnitude per order of magnitude of time — so a rate quoted without its interval says very little.
The logs themselves are drawn in Measured sections, an ash bed correlates two of them in Ash and hazard, and the age that turns thickness into time comes from Dating a rock.