I'm surprised at the suggestion that pilots routinely inhibit the EGPWS terrain floor function if they think there might be ''obstacles''. I would only do this if I knew that the landing runway was not in the database; after all, the whole point of it is to warn that you're following an unsafe path relative to terrain/obstacles. If you're close enough to trigger a caution or warning, then you need to do something positive quickly!
As they were on an NPA, my reading of FAR 91.175 (being an EU-OPS guy, although it's very similar) says the visual references required are fairly broad and don't mandate slope guidance. Just the approach lighting would be fine until <100' above the TDZ.
Given the falling ground under the approach path and the shorter runway, the expected visual illusion is that of being too low. This might have been countered by uncertainty from the 'black hole effect'. Puzzling. It would be very interesting to synthesise the view out of the window at various stages during the approach when accurate flight data becomes available...