I am the MIC at the Minneapolis CWSU. We were part of a convective forecasting project involving AWC, several airline meteorologists and us. We forecast for specific thunderstorm areas for specific times over the US. See web page
http://www.air-transport.org/ops/ccfm/ccfp_1800z.html for an example. We at Minneapolis CWSU used the RUC-II for our primary forecasting model for this project. Here are our general impressions of the RUC-II model.

Early in the summer in May and June when the upper dynamics were strong the RUC-II was far superior to the ETA for thunderstorm forecasting. We quickly found that the RUC-II was much better 850mb and above than below 850mb. We concentrated on indicaters such as the K-Index and the Showalter Index and they together would faithfully indicate the area to focus for thunderstorms. We made several detailed forecasts for thunderstorm timing and area in the June time frame which worked out very well.

Later in the summer when the dynamics weakened and thunderstorm development was tied to more subtle low level features (outflow boundaries, weak surface lows etc) then the RUC-II did not do nearly so well and some of its weaknesses below 850mb became more apparent. We noticed too much moisture in the Plains. In one instance with upslope flow around a high the RUC-II forecast dewpoints increase from the 50s to above 70 at places like western nebraska (BFF) and near Rapid City in 12 hours. The moisture increase was greatly exaggerated. (Sorry, I don't have the date this happened. We see the opposite near the Great Lakes where the cool influence of the lakes spreads well beyond the normal lake breeze front locations. Using RUC-II alone you would seldom forecast thunderstorms for Northeast Minnesota, Northern Wisconsin and most of Michigan. I had a few misses on forecasts in those areas until I figured out the overenhanced affect of the Great Lakes on the surrounding land areas.

Another thing which happened twice late in the summer. The RUC-II would wipe out surface fronts during the day and the flow would go into weak surface lows further to the south. The decay of surface fronts was incorrect and led to forecast busts as thunderstorms would develop along the fronts. The wiping out of the fronts was observed in the streamlines. I wish I had looked closer at the data because I bet the RUC-II developed thunderstorms first near the weak surface low where moisture and instability was plentiful but no trigger except surface heating. The cold front thunderstorms developed late in the day but they were the main area of thunderstorms.

We hope in the future we hope to give you more timely input to things we see. RUC-II will continue to be the primary model we use for our short term mesoscale forecasting that we do.