Preliminary winter outlook 

Posted: 7:44 am Thursday, November 2nd, 2017

By Kirk Mellish

The forecast methodology is the analog method combined with some long-term model output. The idea is that what’s past is prologue, or the future is history. So we search for past years when factors influencing seasonal weather are a match to what has been seen this summer and so far this autumn and what is expected to be seen in the months ahead in terms of large scale drivers, like the global ocean sea-surface temperature patterns (not just in the Pacific).

Once an acceptable list of analog years is obtained they are composited together to get the mean outcome of those past years which give clues about the future.

However, due to the complexity of the earth-ocean-atmosphere system there are never perfect matches, so we have to settle for similar.

No two El Nino or La Nina seasons are ever the same, there may be similarities but exact matches are a rarity in mother nature. Also, some of the things we need to project are not yet predictable in any reliable sense beyond 1-3 weeks in advance such as the QBO, SOI/MJO, AO/NAO-Greenland blocking etc. So there are built-in limitations.

This is why long-range outlooks for months ahead are just an educated estimate.

As they say in those investment disclaimers: past performance does not guarantee future results. So just because a similar weather driver USUALLY resulted in pattern X does not mean it ALWAYS does or will.

We start this year expecting at least a weak La Nina or cool Pacific Ocean in key zones. But it could end up being weaker or stronger than expected. It looks like an east-based La Nina with the 1.2 ENSO zone cooler than the 3.4 ENSO zone in the NOAA charts:




The rest of the worlds oceans offer mixed signals for a cold or warm winter east of the Rockies, neither fully supporting nor fully contradicting the ENSO signature.

Typically for the Southeast a La Nina tends to mean a less active sub-tropical jet stream with fewer low pressure systems and lower precipitation (small odds for many Miller A or B type winter storms).

A stronger La Nina would decrease odds of a real winter, while a weaker or non-existent La Nina would at least “open the door” to a more notable winter in the Southeast U.S.

IF and when we get snow of any significance in the deep south it’s usually in the January-mid-March time frame, and I see no reason to think that would not apply this season.

I’ve previously covered the various ocean, atmosphere, solar and cryosphere indices considered so I won’t rehash those factors.

Other warm and cool pools across the oceans can modulate the La Nina and can change in the months ahead, so they need to be monitored going forward.

NOTE that a milder end of Fall and start of Winter also raises increased concern for severe thunderstorms, doesn’t mean it has to happen but generally speaking cool weather is more stable. IF in fact it turns out to be warmer than normal next 90 days it’ll be something to monitor.

Screen Shot 2017-11-05 at 10.56.46 AM

Screen Shot 2017-11-05 at 10.57.06 AM

Analog years: 1931, 1949, 1950, 1952, 1954, 1956, 1960, 1961, 1962, 1966, 1967, 1970, 1974, 1981, 1983, 1984, 1988, 1995, 1996, 1998, 1999, 2000, 2001, 2005, 2007, 2008, 2010, 2011,  2013, 2014, 2016.

From this I culled the best fit matches to this point for the following preliminary outlook:

Snowfall at the official airport station in the analog years ranged from ZERO to a high of  10.3 inches. The analogs came in with 1.4-1.9 inches of snow for Atlanta.

I have to admit I was surprised by that, but that’s what they show. I would lean toward less for now.

Using the analogs here are the results so far:

Screen Shot 2017-10-31 at 5.13.40 PM

The long-term average snowfall for Atlanta is 1.9 inches.

The above map shows the AREAL AVERAGE, it does NOT mean that every state and town will be above or below but that the region should be as indicated in the mean.

Remember, this is a 3-month mean it does not say every day or every week or every month will be like the maps below.


Screen Shot 2017-11-01 at 6.41.52 AM



Screen Shot 2017-11-01 at 6.57.57 AM

So while odds are reduced for significant winter weather here lasting months and certainly a harsh winter is not expected, the variability projected could still lead to at least one day with just the right timing to yield an inch or so snow in Atlanta.

So it does NOT look like it will be AS warm and dry as last winter,  but closer to an average winter than would normally be expected in a “typical La Nina” winter over the 3-month period here in the Southeast.

By the way, the LATEST measurable snow in Atlanta occurred on April 1, the AVERAGE LAST measurable snow in Atlanta is Feb 1 to March 1st. (“Official reporting station”).

The greatest number of days with snow in Atlanta was seven in 1968.

No measurable snow has ever been recorded in Atlanta in October or November.

It’s interesting to note that data shows the average date of the first measurable snow in much of the South and East is now 7-12 days later in the period 1978-2017 compared to 1938-1977.

Stay tuned for updates. Follow me on Twitter @MellishMeterWSB.