Oggers,
You're not entering into a game of semantics, you're missing the point. You get horizontal gravity variations of about 1 part in 100,000. They don't matter. Period. If anyone really sticks them into a model of falling debris, then they're casting doubt upon the quality of other bits in their model.
It is possible that the vertical changes in gravity, by about 3 parts in 1000 from 10km up to sea level might have a subtle effect on where heavy dense debris lands, since that slows less and spends less time lower down than lighter debris.