FWIW, this is the percentage of time spent above a certain rpm by Vettel on his pole laps on different tracks (selection is purely random amongst the ones I had available already analyzed...):
The dotted lines are relative to "on throttle" areas (= I removed parts of lap corresponding to decelerations), while the solid lines are for areas "on throttle" and with speed > 130km/h, arbitrarily picked as speed at which acceleration stops to be grip limited and is power limited.
Likely the actual value is different, possibly higher (and varies on various tracks, or even corner by corner, depending by tyres/tarmac aero setup, lateral acceleration etc) but as first approximation for a reference I decided to play safe, knowing that if it was higher it would mean to further move the solid lines to higher %.
As you can see the engine spends roughly 85% of power limited accelerations revving above 16k, and around 50% (slightly more in some tracks, slightly less in others) above 17k.
To complement that, this is rpm vs speed for Turkey (points from on throttle areas only), each line obviously represents a different gear, from 2nd to 7th (velocity plot is a bit filtered while rpm is not, that's why some data, especially corresponding to gear changes, don't exactly match the proportion):
It clearly shows again how the engine stays in basically the uppermost 2-2.5k rpm, except in low/medium speed areas in low gears when obviously grip is what limits the acceleration.
Worth analyzing in this graph the overlap of 2nd and 3rd gear, with the data points highlighted in red corresponding to red areas of track map, clearly areas where acceleration wasn't purely longitudinal but composed.
In these areas, even if speed was low enough to potentially use 2nd gear, Vettel preferred to stay on the longer ratio, keeping engine at lower rpm, as anyway he couldn't have used the additional power that it would have had in 2nd.
That way, even if probably the power available could still overcome the grip, the excess was smaller and the delivery smoother, more controllable.
These are the circumstances where a wide usable power band is helpful, allowing to keep a longer gear, letting the rpm drop to 12-13k, with still enough power to exploit the grip, but not much more, so that the throttle control is easy and the grip increment more accurately followed.
Ideally one would want the power available in that rpm range at WOT to be roughly as much as grip allows to use and also the increment with rpm to match nicely the grip increment that will arrive thanks to increment of speed (thus of downforce) and reduction of lateral acceleration.
In that ideal situation all the driver has to do is just floor the throttle and be done with it not having to worry anymore about power delivery, the closer you get to that ideal, the better. (as rule of thumb, the less a driver has to do, the less can go wrong...)
If the engine lacks power in that low rpm range and the driver is forced to drop to a shorter gear, all becomes more difficult with following acceleration being strongly dependent on control, via throttle, of a peaky power delivery thus way more difficult to get right and exploit all the available grip.
In theory the difference between these two scenarios is small (or even non-existent if we assume the perfect driver with perfect throttle control) but in practice in real life it can change a lot on performance, on the single lap and, even more, on repeated performance lap after lap, on tyre wear and so on.
In other parts of lap though, the power available under 15.5k-16k, is scarcely relevant and only the uppermost 2k rpm matter as that's where the engine is used.
Needless to point out that you don't see data points overlapping two different gears in the rest of the graph, at each speed Vettel is always selecting the shortest gear available, to keep engine at the highest possible rpm, as close as possible to peak power (which for current F1 engines is presumably very close to 18k).
