Winter outlook ingredients 

Posted: 4:37 am Thursday, August 16th, 2018

By Kirk Mellish

Long time readers of my blog already know it is far too early to make a stab at a winter forecast, they also know there are hundreds of variables that impact the outcome of any given winter. Some of those are NOT even knowable more than a few weeks in advance, for others it’s too early for the data to be used because months of change lie ahead.

Of course I stick to science so I don’t use squirrels, other animals, wooly worms, bee and bird nests, thunder days or fog days or any other old wives tales which have been formally studied and found lacking.

Winter forecasts this early are made by those just wanting to get people excited and worked up and use it as clickbait, or to move energy markets so they can make money. But they are so far off so often they are not worth anything.

On the other hand, research shows long-range forecasts made just before the start of the month or season in question are more accurate than a guess over the long run but not by a lot.

Here are some of the early indicators being monitored that will serve as inputs for my winter outlook in the months ahead:

ENSO El Nino Southern Oscillation (el nino and la nina):

Screen Shot 2018-08-15 at 6.22.36 AM

Last year at this time…

Screen Shot 2018-08-15 at 6.29.33 AM


This year…

Screen Shot 2018-08-15 at 6.42.56 AM

You can see a reversal in the Nino 3.4 region from cooler to warmer than normal.

MODELS indicate for DJF a weak El Nino of a half to one degree Celsius. :


From the Australian BOM:

Screen Shot 2018-08-15 at 6.56.33 AM


El Nino comes in different flavors. No two are ever exactly alike. They can come on fast or slow be strong or weak, fade fast or fade slow.

An “east-based” El Nino winter differs from a “West-based” or central Modoki El Nino winter.

Screen Shot 2018-06-27 at 8.59.45 AM

The Nino 1.2 region has been cooler than normal most of this year while the Nino 4 region has been the warmest indicating the coming ENSO event is unlikely to be a canonical or conventional basin-wide El Nino.

When Nino3.4 is warmer than Nino1+2 it favors a cooler pattern from the Plains eastward with an active sub-tropical jet stream.

Also modeling shows the frequency of polar vortex disruptions (sudden stratospheric warmings SSW) increases by around 40% in El Nino winters.

As of now a weak central-based or Modoki El Nino is expected, that type favors a mean jet stream ridge western U.S. and mean jet stream trough in the eastern U.S. as do weak El Ninos in general:

Screen Shot 2018-08-15 at 7.07.46 AM

Screen Shot 2018-08-15 at 7.10.01 AM

Screen Shot 2018-08-15 at 7.34.20 AM

There is also a suggestion of a split-jet stream pattern with dominant North and South streams simultaneously. As of now I would not expect a lot of brutal cold as the source region for air masses in El Ninos tends not to be from Alaska or Siberia but rather Canada.

Some research has shown ENSO of recent decades similar to the estimated pattern of the late 1870s, but unfortunately climate data from then is obviously difficult to obtain.


Screen Shot 2018-08-16 at 7.09.25 AM


We are in a low sunspot cycle heading toward and near a solar minimum. There have been near record few sunspots in recent months and for the year 132 spotless days. That can slow the El Nino and also favors high latitude blocking in winter which favors colder and snowier east U.S. but there can be a lag so low solar does not guarantee any given summer or winter will be cool.

There is an approximately 0.3 W/m2 change in solar irradiation from solar max to solar min.

Remember, there has been a down trend in solar activity since around 1980 yet plenty of global warmth and both up and down summers and winters here so it’s not a straight line that quiet sun automatically equals cool.

Screen Shot 2018-08-15 at 8.00.40 AM


Screen Shot 2018-08-15 at 7.50.32 AM

Bottom expected in a year or two part of long predicted 11-year cycle.

Screen Shot 2018-08-15 at 9.27.02 AM

Screen Shot 2018-08-15 at 7.57.42 AM


Keep in mind we’ve had plenty of warm summers and warm winters during low solar periods.

North Pacific:

Other SST cold and warm pools matter, not just the Pacific ENSO regions. We monitor all oceans in both hemispheres. For the Southeast what is sometimes referred to by the press as “a warm blob” (coined by Washington state climatologist) South of Alaska is an important key to monitor:

Screen Shot 2018-08-15 at 8.25.38 AM

That North Pacific warm pool favors western Canada upper level ridge in winter and downstream trough. It figured into the winters of 2013-15.

Screen Shot 2018-08-16 at 6.28.06 AM

These are just some of the bigger known “drivers” of the winter pattern, you can google these and others to learn more. Others we look at are the MJO, QBO, PDO/AMO, PNA/TNA, WPO/EPO, NPO, MEI/ONI, GWO/GLAAM, AO/NAO, TNI, IOD etc. plus hurricane season behavior in both oceans, how the summer plays out, how the fall plays out, Atlantic dipole and Eurasian snow cover in autumn and weather in the Southern Hemisphere in the preceding season. Again, feel free to google them and learn as much as you want. I’ve covered them all many times in the past so don’t want to be redundant or turn the blog into a textbook.

Most of the data is incomplete or not even known yet.

I can tell you that the Canadian and European models are pointing toward below-normal temperatures East of the Missouri River and above-normal precipitation along the South and East U.S. on average this coming winter. (In general, not everywhere).

Also note, because TIMING is everything a below-normal temp/above-normal precip pattern does not have to translate into above normal snow or ice.

In fact, the latest explicit snow output from the ECMWF and UKMET show below normal snow in Georgia despite colder mean temperatures. The France model is below normal on snow until February when it goes above.

The CFSv2 depicts a blow-torch winter for much of the country as does the Jamstech model from JMA.

Again, I do not yet have a forecast as I’ve seen too many fall short when made RIGHT BEFORE the season starts and too many fail like a bad joke when made before October.

Long-range forecasting as a science is still in it’s infancy, no reason to pretend otherwise.

Averaged together past decade of winters have averaged mild (WxBell research):


The winters of 2011, 2015 and 2016 were the warmest for the nation as a whole, while these averaged colder (WxBell research):



Screen Shot 2018-08-15 at 7.28.26 PM

Follow me on Twitter @MellishMeterWSB.