Kirk's Winter Forecast for 2009-2010 season

A moderate El Nino event is expected to start the winter peaking early then weakening by the end of the season. But this El Nino Winter does not look like it will follow the El Nino norm of a lot of warmth for the North USA and just slightly cool for the South. The negative departure from normal in the SE looks significant. Nationwide it may be the coldest winter since at least 2004 or even 2000.

One of the most reliable features of an El Nino is a split-flow jet stream pattern with an active Sub-tropical Southern Jet stream flow. We will see that this winter as we already have in the fall. This pattern will be a challenge for the winter Olympics in Canada with possible warmth and snow drought. It will also mean an active Southern storm track sub-tropical jet stream. This presents problems for even short-term forecast accuracy as it means many systems will come toward Georgia that originate in the data-void areas of the South Pacific, Mexico, and the Gulf of Mexico. Thus we may not have a good handle on system behavior. Also these split flow patterns are handled poorly by the numerical models because they have trouble with the heat and energy exchanges between separate jet streams and whether or not they will phase/merge or remain separate.  So expect a lot of model flip flops, changeable weather and changeable forecasts.

The average of the 3-month period (not every day or every week) is for temps to end up below normal by -1.5F to -1.8F with total precipitation ending up plus or minus 10% of normal. The second part of winter looks to be the harshest, so expect a back-loaded winter with Jan./Feb./early March having the greatest deviation from normal temps on average but December will be no day at the spa.

If the analogs are right historic cold for the winter as a whole is not expected. But they do suggest some record threatening cold spells could occur (which is unusual for El Nino winters as a whole). The snowfall amount from the analog years came out to 1.8 inches, which is close to the long-term normal for Atlanta of 2 inches. The range of snowfall in the analog years was from zero to 4.8 inches, 15 of 17 years had snow with 2 years missing data.

The odds of snow look normal to greater than normal. The odds of light to moderate ice events looks normal to greater than normal. The odds of a widespread SEVERE ice storm look less than normal. The odds of at least one severe weather episode in the winter look a little above normal.

Should the ENSO event unfold in a manner different from my expectations then this forecast would be expected to bust. This is also true of other underlying assumptions that went into the long-range forecast such as the prevailing state of the PDO and AO/NAO etc. My forecast is at odds with many calling for an historic “throw-back” winter for much of the country including the Southeast, but I can certainly see how that could happen.

A very warm and very dry winter can be ruled out.
A wild winter possible but NOT boring.
Wetter and stormier earlier than late.
Real Cold will come and go but Real warmth will be hard to come bye.
Below-normal 3-month average temps but not a bitter historic cold winter.
Estimated 80-90% chance of below-normal winter temp average.
Several ice threats but probably no widespread damage type.
Snowfall prospects about par for the course.
Above-normal severe thunderstorm risk Dec./Jan but not Feb./March.
Mountains best chance to be drier than normal, South Georgia best chance of being wetter than average.
A normal or early start to spring.

It seems there is a better chance that the winter will be wintrier than I am forecasting than less. In other words, if this forecast turns out wrong it’s more likely that I forecast too mild and too dry a winter. UNLESS the El Nino spikes to strong then we would end up with a mild season.

Confidence in this forecast is low due to dueling atmospheric signals that are more difficult to decipher than in most winters.

Data and assumptions FACTORED INTO THE FORECAST:

Negative East QBO
Solar cycle minimum
Hurricane season behavior: quiet season, re-curving storms, and November Gulf storm
SOI/MJO oscillations
Negative AO/ neutral to negative NAO weak to moderate
Neutral to positive PNA
Negative EPO
Neutral to negative PDO
Moderate West-based El Nino
Neutral to negative Lower than normal GAAM 250mb vector wind earth torque
Cool summer, Cold October warm November, Wet Fall pattern
Northern Hemisphere snow cover
Autumn storm tracks Jet stream patterns and height anomalies
Stratospheric warming events
Atlantic and Indian Ocean Dipoles

The favored analog years are: 1903, 1925, 1951, 1957, 1963, 1965, 1967, 1968, 1969, 1972, 1974, 1976, 1977, 1986, 1987, 1994, 2002, 2004, 2006.

These aspects of recent and/or expected land-sea-air patterns allow “matches” to be found in the past. These “look-backs” are predictors of the future. The premise is what’s past is prologue. However, the matches are really near-matches as there is rarely a perfect fit. Nature repeats general patterns but NOT exact patterns, as the weathers options are infinite while our data is finite. In addition data for some analog years such as the 1800s and early 1900s is not available in full or in part and this could impact the outlook by giving an incomplete result. Snow data for Hartsfield for example is only available from 1929 but some of the analog years are from before that year so cannot be captured. Likewise the compositing tool limits me to 20 analog years even though I may have more than 20 in my total. See my blog slide show for national maps of the outlook.