The Experimental Winter DJF 2014-15 NAO Forecast using Zonal Wind indicators from May & September

This is a NAO forecast based on research done in 2011, whereby certain zonal wind anomalies in the Northern Hemisphere during May and September are used as predictors. It has worked decent during 2 of 3 forecasts from the winters of 2011-12 & 2012-2013, but busted rather harsh during the Winter of 2013-2014.

Figure 1. Tropical Northern Hemisphere 500mb Loading Pattern in +Mode (NOAA/CPC)

In my attempt to try to figure out why, I remembered the extreme North Pacific regime and Tropical Northern Hemisphere (TNH) (Figure 1) blocking in the Winter 2013-2014 was dominant to the point where I believe it may have affected some way the expected Atlantic DJF blocking outcome (NAO). I noticed the 2 times the model failed quite badly was during periods of extremely positive TNH values, as a matter of fact the winters of 2008-2009 and 2013-2014 had the highest DJF TNH values since 1948. This meant a very strong Hudson Bay Low anomaly over Canada and above normal 500-hPa heights from the Gulf of Alaska to just of the coast of California. Out of all the winters with a TNH of +1 S.D., 4 out of 6 had inaccuracies, which was the largest cluster of inaccuracy found in the data. The model's R^2 has dropped from 0.79 to 0.69 due to the 2013-14 forecast failure, but still describes 69% of the NAO.

NAO Forecast Graphic

The Winter of 2014-2015 could very much be driven by North Pacific dominate variables, although that is not a guarantee. If that is the case, this forecast could be offset by a +TNH dominant winter, so take it as an "Experimental" tool not a definite. The NAO is essentially being forecasted as neutral with a value of -0.13 but with error -0.13 +/- 0.2 or -0.33 to 0.07.