Hi there!
Certainly, munt strives to implement a reverb model found in the emulated MT-32 hardware yet simply provides you for full and direct control over the reverb parameters available on real devices via software (i.e. SysEx messages). For instance, the four reverb modes are listed in the owner's manual along with the two other parameters. The manual also mentions that the user may control reverb via the control panel and (in another chapter) via software. But the manual does not provide a good hint to answer your question wrt. the mapping nor what's what, so that's basically all one can get from that source of information.
Apparently, others have already been wondering about that stuff, and we can find a lot clearer answer in Bob O'Donnell's book. He says the control panel allows you to set the reverb time parameter only leaving the other two at their default. I'm yet to figure what happens when you set a different mode / level via software, then change the time via the control panel - whether those are reset to defaults or not. That's also a bit strange and I'm curious how the 8 possible values of reverb time got stretched over 11 values settable via the control panel...
Speaking of the meaning of the parameters, I'm not sure I can describe that well (being not a musician) - the code does that in much more details. In short, the first three reverb modes are quite similar and only differ in the depth and sound colour (that is, the lengths and coefficients of the comb filters within). The Tap Delay mode is distinct, so that each reflection is very audible, they also come in stereo. The reverb time parameter determines how fast the reverb will decay, and the reverb level is exactly what you might think of it 😀
There is one function munt implements which the hardware lacks: the reverb settings can be overridden permanently (i.e. got fixed), so that any further SysEx messages that would normally change the reverb on real hardware will be ignored by munt. You can also disable reverb altogether or smoothly adjust the gain factors of the dry and wet outputs separately. Otherwise, the emulation is fair enough.