I got curious about the actual SE US tendencies associated with each MJO phase after seeing the charts showing temperature anomaly patterns for each phase. I knew the left phases tended to be cold and the right phases mild. But I wanted to know how cold and how mild. When I had some time and since I love analyzing wx stats, I decided that it would be interesting to see how Atlanta, which I figured could be used as a proxy for the SE US, was for each phase. So, I started off just looking outside the circle by phase. I did just January to be more efficient. I did this for 1975-2014 since this was done before January of 2015 was completed. 1975 was the first year that MJO day by data is available from this source:
As I was going along, I noticed a strong warm bias with these ~800 days. I noticed that phases 4 & 5 were quite warm as one would expect but what confused me was that the remaining phases were near normal with no phases averaging cold! That’s when I figured out that INSIDE the circle is where many of the cold days must be hiding.
So, that’s when I decided to analyze the ~430 inside the circle days. I discovered that inside is, indeed, where many of the coldest days were hiding! I then learned that inside the circle is also within the 8 phases (think 8 pie slices). So, when you see a map showing a cold SE phase 8, for example, the cold average is largely being carried by inside the circle phase 8 though just outside the circle in phase 8 was also cold.
By the way, every phase came in 3-6 F colder when inside vs outside and even phases 4-5 inside averaged near normal. The remaining 6 phases inside averaged colder than normal with inside 7 & 8 being the coldest.
The following diagram (ignore the 1978 reference) shows the Atlanta normalized average daily temperature anomaly for each phase, both inside and outside the circle (inside the circle phase 8 has an asterisk because it is the coldest at -6 F)(you may have to enlarge it to see it well):
Based on these findings, I drew an optimal path for the best chance for lengthy cold in at least January MJOwise (see diagram below). Optimally, start around the cool phase 6 within the circle and slowly circle around counterclockwise within the circle through the cold phases 7-8 and then through the still chilly phases 1-3 within the circle. This could easily take 2-3 weeks or so (the longer the better):
The following are many of the closest examples I could find to the optimal path….note that all of them resulted in anywhere from a 10 to 19 day period of solid cold domination: