ALL long-range modeling has been severely cold biased for quite some time now and cannot be trusted. bluewave has made some very informative posts on why this is the case and what can be done to improve the accuracy of long-range forecasting going forward.
if you’re referring to early next week’s cold shot, that looks to be quite short-lived. Mild Pacific air will flood the mid-section of the country shortly afterwards.
you can’t really ignore the MJO imo. phase 6 is quite a strong warm signal for this time of year even when taking into account the ENSO. don’t really care what that awful new AI model is showing at hour 336 or whatever.
I've come to learn that the OP GFS is just a god-awful model. I wouldn't really take whatever it's showing in the medium-long range seriously. Although to be fair, the other deterministic models are quite putrid too.