If anything the 12z Tuesday operational GFS Model is colder and more impressive with respect to the outbreak that occurs in the 6 to 10 day then it was this morning. The actual forecasts of the sequence of events touched does not change at all model -- it is just that it has more cold air which plunges further south and is more impressive.

By September 24 25 the model actually has a very impressive 1037 mb HIGH centered over eastern North Dakota and Minnesota. Winds around this large area of cold high pressure or completely calm so this allows for maximum radiation cooling which is why the model has temperatures getting this low.

You need to keep in mind that the GFS has a pretty strong cold bias so some of these temperatures may be overdone. But as I said I believe this event to be a real threat will be some portion of the area and lead model here is the European NOT the less reliable and trustworthy GFS.

Also keep in mind that the second outbreak which occurs around the end of the month is still showing up on the weather models. The 12z Tuesday GFS still has that event as well for September 30.

The following temperatures are simply model information based
They are not my forecast.!!

morning of 9/23 the GFS Model shows:
M- U 20s   over central MB
30 to 32°... eastern MN northeast IA eastern WI
M- U 30s... eastern ND eastern SD rest of IA northeast MO western ILL western WI

morning of 9/24 the GFS Model shows:
M-U 20s... all of south-central Canada... eastern ND northern half of MN
30 to 32°... the rest of ND all of SD northern NEB the southern MN all of IA all of WI
northwest 50% of LL... M and U 30s... southern NEB KS MO IND eastern ILL OH

morning of 9/25... the GFS Model shows:
M-U 20s ... all of sw Ontario... all of MN all of IA western WI northern MO .
30 to 32°... eastern ND eastern SD eastern NEB eastern KS the rest of MO all of ILL eastern WI all of IND ...all of KY most of TN and as far S as MO/ ARK state line.