I think I see where these questions are leading.
In theory, the best thing to do would be to go out and get an RTA, 'map' out your vehicle's transfer function via frequency response, and then model your sub enclosure's response to counter this exactly, leaving you with an in-car response that is naturally smooth.
Now to the real world. You cannot build an enclosure that counter's your vehicle's transfer function map perfectly. If you are serious about it, build an enclosure that gets you in the ballpark, then let your processor do the rest. Most people just build for the most predictable response form the enclosure (generally flat) and then adjust for their vehicle's transfer function via their sound processor exclusively.
Hope this helps.