Mar 28, 2014

Seeking an Auto Discovery Cycle Duration - Part 4

The patterns are cycling but the season they recur in is likely the missing feedback in the Trend. I am seeking to eliminate this seasonal hiccup. An accurate dominant harmonic correlation must be integrated into the formula. The short-term ISO may allow the trend to latch onto seasonal change and produce a more accurate projection. Testing out a new formula to calculate a dominant harmonic correlation, a recent analysis produced a short-term ISO of 18 days. This number led me to potential analogs at specific harmonics.

~18, 36, 54, 72

Correlations




The primary and second harmonic were the most dominant as there wasn't regional consistency in positive correlation past 40 days (see above images). I chose the second harmonic. Give or take a couple days, 36 days away from mid March was early/mid February. I knew that back on Feb 20 (see tweet below) the atmospheric setup produced a severe weather event that had recurred on two separate occasions in the long-term cycle.


----

The SPC had nothing in their D4-8 forecasts on Mar 20 or Mar 22. They did in the D3 on Mar 25 though. Very good.






----

The above observation is over a week old and the dominant ISO components have evolved. Below is the Mar 26 run. More to come.

Correlations


If there are any questions, comments, or suggestions on the material presented please let me know. Thanks for reading! More to come.