Anticipating storm development based on what's happening vs. what the models predicted

So far, the NAM and other mid- to long-range models haven't handled the forecast very well. Even the short range models are having trouble with this system and predicting where storms are going to form. Take a look at this image:

This is a loop of simulated radar reflectivity starting at around 8PM EDT yesterday evening and progressing until right around 11AM. Notice that quite a few things are different at the end of the loop from what's actually happening right now:

First of all, there's no cluster of thunderstorms over Wisconsin. The blob of rain on the border between Indiana and Ohio is a bit further north than originally predicted, and has held together a bit better. Finally, there's a cluster of showers over western lower Michigan. Now, all of these features are quickly dissipating, however it tells you that the forecast wasn't very accurate in terms of placement of atmospheric activity.

The NAM model seems to favor a more east/west track of activity, while the HRRR is favoring a more north/south track. I think the sweet spot is somewhere in the middle, in which case I'm of the opinion that tonight's storms may affect the north portion of the state more than the south. In terms of new development, the HRRR definitely missed what erupted in Michigan, and if that happens tomorrow, it'll be more than just showers popping up. As such, I maintain the "severe thunderstorms likely" portion of the outlook for most of the eastern portion of the lower peninsula for tomorrow evening, and include the SPC's slight risk area. I think that otherwise the threat for severe weather is marginal at this point.