Cliffs, what cliffs?

I was looking in detail at the (poorly georeferenced) sketch map overlaying the topo map.  underneath is a pseudo-coloured altitude wash:

Here's roughly the same area with red polygons for the topo's cliffs ... *way* out:

Here's a GPS track (in blue)

 Same GPS track with pseudo-coloured slope wash, derived from 1sDEM

Note that the particular slope calculation (gdaldem slope) is really chunky ... not unexpected, really, as the scan is 3m vertical res but only something like 30m horizontal res (whatever 1 arc-second is at the surface of the earth.)  Part of it, too, is the colour quantization - I'm guessing the display is only doing 8 bits of colour.

The contours are much more informative, even though they're derived from the same scan ... this is, I think, because contours can extract information from more than one data point ... they presuppose continuity which permits that.

Just for completeness' sake, here's the GPS track, the topo track and the topo cliffs overlayed on google's imagery.