Some of the 12z WRF research runs are pretty interesting. Note that they aren't much stronger than the operational model (in some cases, actually weaker)... but they're a bit more rambunctious with QPF. I'm wondering if that has something to do with the finer resolution and convective schemes?
As a general rule, higher resolution makes able the simulation of stronger vertical velocities, thus it would be expected that QPFs would be a little more rambunctious with the hi-res models. I see the same thing in my own WRF runs (I run at 3km).
