An expected cold change in January for North America is indicated by ECMWF which is vividly much colder in the extended range outlook. Most of the precipitation across the Northern U.S. is snow helping to enhance the cold.
ECM has carried a wet bias with their 15-day operational outlooks throughout mid-to-late spring and early summer. Recently, GFS wet bias has (also) increased and is similar to ECM. Presented is an excellent opportunity for machine learning forecasts to compete with GFS and ECM as they may understand the wet bias and make adjustments to the operational forecast.
The next cold temperature issue arises in the week-3 forecast when all “weeklies” models project an evolving upper trough in the East U.S. ECM continues with the strongest trough forecast. For the moment, arctic air is not anticipated with this upper trough.
The mega-cluster ensemble, combining ECM and GFS, projects “most likely” temperature anomaly scenarios for early January cooling off the U.S. to near normal while Canada stays warm. The most likely candidate for colder than normal temperatures is the Interior West and possibly Western Canada.