Does it need a “ground”, or does it need a “ground-plane”?
Mounting the antenna with a thru-hole through the centre of the roof both grounds the antenna as well as provide as equi-spread a ground-plane as is possible to obtain on a car (short of then adding radials that stick out both sides of the car as far as the front and back stick out!).
In reality the performance difference you will experience (primarily in a receive-only scenario, but to a lessor extent also in a Tx scenario) with the antenna mounted in the centre of the roof (thru-hole) versus mounted on one or other side of the car to a roof rack (or bumper), will probably be measurable with lab instruments, but is going to have negligable effect in day to day use.
The variation in enviromental conditions your car/antenna set-up is going to experience, day to day as you drive around in it (i.e. variation in multi-path refelections off buildings, position of hills, changes in ground moisture & conductivity conditions, changes in relative position to contour lines, position of large volume steel mass objects - large steel framed buildings, car parks full of cars, blah blah, blah blah …...) is going to reduce to zero any advantage that could have been gained by way of mounting the antenna with a thru-hole in the centre of the roof.
That deals with ground-plane effect on antenna performance in the circumstances you are citing - and I emphasize in the circumstances you are describing.
But, from a pur grounding perspective i.e. bringing all parts in your radio/scanner set-up to a common and equal voltage potential - yes, as good a ground as is possible to establish is always the way to go – …. and then the subject starts getting complicated.
But your question really is: how is it going to impact the antenna performance?
Well, the antenna is going to be “grounded” in any event - thru the shroud/outer shield of the coax cable to the the receiver/scanner, which in turn is then “grounded” to the vehicle chassis by way of the negative (-) wire of the power supply that goes to the radio/receiver/scanner. So, in theory it is grounded.
Now - how technical do we want to get, and how effective is the above ground?
I have no experience with the antenna you are describing, but from what I can see on the net (i.e. pictures of it), the connection options offered with the example I looked at, would ground the antenna to the car at the fixture point. How effective that "ground" would be, if it was thru-hole, would be all about bare metal to bare metal contact (i.e. paint layers removed from between the antenna fixture and the panel it was been mounted to?), or how well any roof-rack was fixed to the car body (i.e. paint players between the roof and the rack fixture points) (?)
Will making an effort to deliberately isolate the antenna from “ground” at these points result in any descernable difference in performance?
mmmm …..... I see a big debate coming, so I'll answer it like this. Why make any effort to deliberately isolate it from ground? I tell you where it could result in a BIG difference (potentialy): in a lightening strike. Instead of the strike flowing around the outside of the car, on the body panels (Faraday Cage effect), it will now take the path of least resistance - down the coax to your radio!!!!!
In short: yes - a good earth is always a good idea - for performance reasons (if only theoretical, as often is the case) as well as safety and general common sense reasons.
As for it striking obstacles – well, most antenna manufacturers take that into account in ensuring that the spring located at the base of the antenna is sufficiently flexible to account for the torque of an antenna strike, and if I recall correctly the rule we used (years ago when I worked on mobile antenna design)was: so long as the contact/strike point on the antenna was no lower than around 50% or 66% of the length of the antenna (from the top), then the spring would bend before your bodywork! But you are going to be best to check that out when you get your antenna – don't take my word for it please.
