Yes, it's probably one of those cases where an oversimplified statement just causes more confusion, rather than allaying it.
I am fully prepared to accept the likelyhood that the colour tone variations do not reflect commensurate variations in the density or assured appearence of filtered scatter objects,
but I would not call them reflecting "topology", as the CM did; Topology would absolutely, to my mind, be unchanged for same common areas, with different filters.
...so a poor choice of word, IMHO, of the: "lies to children" kind... Wildly misleading in stretching its semantics waaaay too far, but sufficiently close-ish to save you having to provide the
long answer, as long as nobody thinks about it (...deferring that, along with having to stamp out the inital planted misconception, to future teachers

).
I would assume each filter collates a
subset of the bitmaps that go into terrain generation - never all of them, like when you see a conspicuously defined squiggly shape, without any fringes, encircling a wet area, cutting a border straight through a crater and across a hill, with 0% ice on one side of it, and 100% on the other, and this same shape coming out starkly in the filter overlay.