I just came upon the thread.
I used to work on Ringwood's low band system years ago. Back then, they were using Mitreks in consolettes in the basement of the police station (which was usually loaded with confiscated beer kegs). They aren't capable of NXDN, and I haven't seen any more recent equipment on the market that will do it on 37 MHz. Not that it's a bad idea, or that it couldn't be done, but it's just not out there. When you see something that's technically possible, but doesn't exist in real life, they went to a licensing service that has zero field experience in radio to prepare their application and just told them to "give me one of them" in terms of emissions. You'll see 11K2 emissions on low band like that, too. All wrong (and all that's wrong with the land mobile radio industry right now, too, as agencies are trusting secretaries, clerical employees, and marketing types to give them competent technical advice).
As for why would NXDN be an advantage on high band, it's all got to do with occupied bandwidth. A narrowband analog signal occupies 11.25 kHz in a 7.5 kHz channelspace. It spills over from each side, so the adjacent channel cannot be reused anywhere close by. It also means that a receiver system that's up high and pulls in a lot of signals will also pull in splatter from those adjacent channels as those users also spill over. Narrow NXDN is 4 kHz in a 7.5 kHz channelspace. It will reject those adjacent channels - and - in NJ, licensing the transmit frequency would be easier, because fewer adjacent channel systems would need to be considered. It makes sense that way.
As for P25 vs. other formats, looking at it protocol wise, P25 phase 2 would have been ideal as a narrow FDMA solution just like NXDN rather than a 2 slot TDMA (which "looks like DMR" but isn't, and for P25 phase 2 "has to be " implemented within a trunked environment instead of stand-alone conventional systems), as it could have been applied to VHF conventional systems and substantially improved interference.