In general, the worst case is initial runup above idle to initiate taxi. Often a vortex forms (sometimes visible in high humidity) that starts scrubbing the tarmac and getting loose sand, pebbles, etc. bouncing. If they bounce high enough they get caught up in the inlet flow. A relative small patch of tarmac is the "feeder" zone - directly below the inlet.
Once the aircraft is rolling, the feeder zone only sees the vortex for a short time so it's less of an issue. Plus, the engine is soon back at idle, so the vortex is weaker.
Having spent much of my career "flying" engine test cells, I can tell you some cells are chronic producers of visible vortices - appearing from the floor, sidewalls, or sometimes the ceiling.
Addendum: Harold Klein of Douglas did some work (1957) referenced
here.