Yes, ReaVerb is the stock plugin included with Reaper. I'm not married to the idea of using it if it's a poor choice, but I thought I would start with it since it's already there.
Offtopic, but since it sounds like you have both tools, here is my breakdown of preference:
Verb: Ozone > ReaVerb, any day of the week. Something about ReaVerb's settings just isn't conductive. I used it once in a final mix because I didn't have the processing power remaining to bring in another instance of Ozone. Anymore and I'll just render a stem with it...
Compressor: ReaComp > Ozone. ReaComp has oversampling, and better controls in terms of knee, lookahead, etc.
EQ: realistically a toss up. I like Ozone's better in terms of functionality, but ReaEQ is much easier to use in a multi-track mix. Often I'll get the settings I want in ozone and then do it in ReaEQ which sounds fine and has a lower processor requirement.
Limiter/Dither: Ozone. (period).
The vocals are the most noticeably dry. A dobro/guitar channel could use some as well. The other guitar and the bass sound OK. The drum overhead could probably use some.
From my reading it looks like I should go with a short reverb (although I don't really know how short is short) on a separate track which is being fed an aux send from the channels I want reverb on. I'm just unsure what (if anything) should be done to keep from interfering with the audience mics.
What's the purpose of the audience mics? Is it crowd reaction, spaciousness, tone? If you just need crowd reaction, consider dropping the volume of the non-interesting sections by about 4-5db, then mix to taste. That gives you people, but you don't have to worry about interfering with any reverberant information because it's low enough in the mix now that it's sort of moot. If you need reverberant information for the entire mix and that's available in the audience pair, then I'd have a tendency to look at that first before I soak an entire mix (or a large portion of it) in any amount of fake verb. If after I got appropriate amounts out of the audience pair, and still needed some on the vocals, then I'd go to work on just that track with the fake verb and use it sparingly. I'm pressed to think of an occasion where I'd take fake verb over real verb
when I had an honest option between the two, but I guess it could happen. Have you tried doing compression on just the audience pair? That removes the peaks, but accentuates the audience and reverb tails, then let the sbd spots provide the dynamics and clarity that way. Lots of ways to approach it, but I'd have to hear samples in order to figure out what method I'd prefer.
Just my two cents. Hope at least some of it is useful.