I think it was just more to do with timing,
The old analog trunking systems of yore would have had a harder time getting all the additional frequencies those use in VHF rather than UHF or the relatively uncharted territory of 800 MHz. And 900 MHz for that matter.
When digital came along and narrow banding, and coupled with the fact that VHF was beginning to look like a ghost town, it once again became more viable once DTRSs came around. Just a guess. In absence of any replies that mention any,
That depends on where you are in the U.S. I've had to find frequencies for large VHF trunked systems after the equipment was sold and delivered to save the salesman's... behind. Folks, NEVER buy equipment before you know there are resources available to support it! To make matters worse, there were combiner limitations, the sites already had other equipment at them, so transmitters couldn't be within a certain kHz value of the existing receivers or transmitters, AND it was north of Line A (Canada). That meant that all of the omnidirectional sites the engineers specified had to be re-engineered to very low ERP and directive antenna patterns that pointed back into the jurisdiction to minimize signal incursion into Canada. There were still numerous HIA letters thrown across the border, with many leading down the path of three HIAs and an on-air test. It took two years between the FCC and Industry Canada to get this done - and it felt like building the pyramids every bit along the way.
The only VHF "pick of the litter" places I've seen have been where manufacturers lobbied to put in large 800 MHz systems - Like Florida. There, VHF makes the jurisdiction the "odd man out" and there is only patched interoperability potential.
Here's an interesting item - recently the FCC issued a Report and Order dealing with legacy trunked systems in VHF and UHF. If the trunked system has not been narrowbanded (on the license and in actual practice), it lost its FB8/MO8 protection and can be ignored. So, slowly, these things will all need to be narrowbanded if they have not already done so. For certain manufacturers, that means going to P25 or other digital technology. That's essentially a forklift upgrade of everything, including subscriber units most of the time.
Narrowbanding hasn't really created that glut of channels the pundits and industry bright lights had anticipated. The reason for that is that the channels are 7.5 kHz apart and have to nominally fit 11.25 kHz worth of signaling into them. The ONLY technology that fits entirely within the 7.5 kHz channelspace is NXDN. Everything else requires a certain amount of geographic separation from its neighbors - or the written consent to share frequencies and accept interference - in order to use. In areas where those extra frequencies were really needed, there still is nothing because there is usually a dominant user co-channel or immediately adjacent to the channel being considered. In my opinion, the only thing narrowbanding did was create an artificial crisis that killed off a number of working systems, mandating the licensees buy new equipment to replace them if the systems could not be reprogrammed. Manufacturers lobbied the FCC to disallow reducing the deviation and installing tighter IF filtering in receivers. We're seeing the backlash now, in reduced deployment of smaller systems. What should have happened was a channelization to 6.25 kHz to line up with the NTIA's plan, just like UHF. Then there should have been standard pairing with mobile frequencies that would only allow FX1 operations, but NEVER allow an FB or FB2/FB8 on them, with frequencies set aside for simplex use. Then "right size" the ERPs for the jurisdictions. No countywide coverage with countywide interference protection and 300 W ERP for a 2 square mile town. But none of that ever happened. Fail.
Public safety isn't assigned anything on 900. They can, however, be subscribers on privately owned SMRs.