Since we are talking about whip antennas as usually installed on a vehicle they are also vertically polarized. If you were able to stack the antennas (on a different horizontal plane, separated by around a wavelength) you would get pretty good isolation from antenna to antenna.
Putting verticals on a flat surface; bonnet or roof of a vehicle, building top, etc... Places all of the other antennas directly in the main radiation lobe of the other vertical antennas. On building rooftops I have seen signal levels of +40 dBm (10 watts) on a receiver antenna from a nearby 500 watt vertical antenna that was operating several MHz away (928 MHz pager transmitter hitting a 932 MHz MAS point-multipoint repeater input). The desense was quite significant as the repeater would go into major AGC action, making it deaf to those sites out there that were trying to come in at -90 to -110 dBm. Weak sites would work intermittently (between gaps in the pager transmitter operation) but stronger sites could still capture the receiver.
There was a triple cavity filter on the repeater input but it did not have a high enough Q to attenuate the passband signal level low enough to prevent the desense. I ended up adding an 8 pole programmable pass filter;
I was able to knock down signals that were as close away as 25 KHz to the point where I did not have desense on the repeater. It also eliminated an intermod problem as the repeater was going non-linear from the high signal level.
If it is possible, I never place vertically polarized antennas that are anywhere near each other (frequency-wise) on the same horizontal plane. When spacing vertically polarized antennas in a stacked configuration on a tower or mast I always go at least a wavelength apart (minimum 3', irrespective of frequency, even up into the microwave bands).
For horizontally polarized antennas the problem will be exactly the opposite.
Remember too that whenever you end up in a complex antenna environment you can have the other antenna elements acting in a passive manner, creating "directors or reflectors" and your omni pattern can be significantly altered. If you do NEC modeling (wire models of vehicles and the antennas) then you need to represent all of the other antennas. You will really
be surprised by what that does to your expectations of how an antenna should behave.