Not looking at models alone for cool off 

Posted: 1:32 pm Tuesday, October 2nd, 2018

By Kirk Mellish

Many look just at NOAA/NWS products or APPS or models. I try to look for and at everything. It’s meteorology not model-ology.

Sometimes the computer wins sometimes the human beats the computer models.

The “before the digital age” methods are useful for back-testing what the models project. If they are inconsistent or contradictory then what the model is showing will usually turn out wrong. They also tip forecasters off to what the computers should show down the line before they show it. Few forecast sources know this or use this old time approach.

Computer models have not always existed for weather forecasting, so other methods were used, such as rules of thumb, teleconnections and analogs.

Jet stream patterns tend to lock into recycled patterns for a period of time then switch to another pattern for an extended period of time. These patterns work there way around the hemisphere known as Rossby Waves. This recurring cycle or reoccurring Rossby Waves can help track repeating patterns. They typically last 30-45 days at a time.

Others that can clue a forecaster to coming changes before they show up in the model guidance include the MJO, SOI, IOD, the bearing straits rule, AAM and the typhoon curvature rule, timing of El Nino onset, South American weather patterns, Rubber Band Theory and analogues. They all can help project large-scale long-term weather pattern changes from weeks to 90 days in advance.

For example, in terms of analogs this summer and start of Fall temperatures in the nation have closely tracked a variety of past years so they are current analogs. A late starting El Nino provides more analog years. Tropical storm moisture into the Southwest U.S. in Sept-October provides more analog years.

Another example would be the abnormally strong October East Coast ridge which looks like it will trail only 2012, 1990 and 1985 providing more analog years to check.

Some of these reveal this for October and November:

Screen Shot 2018-10-02 at 9.40.19 AM

It tells us that warm first halves of October have reversed at the end and into November.

Screen Shot 2018-10-02 at 9.42.56 AM

Late starting El Ninos like the current one in November:

Screen Shot 2018-10-03 at 8.51.13 AM

That doesn’t mean it HAS to happen this year but it shows it has happened following a warm September and first part of October in the past and similar things are going on so far this year.

On the other hand we are fighting multi decade warming trends and the AAM argues for mild weather holding on in the Southeast until the end of October while areas just North and West get the cold fronts.

Either way weather history tells us that a warm October has been followed by rough winters in the U.S. plenty of times and mild winters have followed after chilly Octobers. It is one of the months with low persistence to ensuing months.

Last October was warmer than normal, then a few spots in the NW suburbs had about a foot of snow in December.

We look instead to the build up of snow and cold over Canada and Siberia/Eurasia in October, it has to happen there first if it is to have any chance of ever coming here in winter.

In the Western Pacific “Kong Rey” is forecast to head toward China before curving North toward Japan. This is a known teleconnection to a cooling upper level trough in the Eastern U.S. 6-25 days latter. When such storms track straight West inland it signals a ridge will build in the East U.S. 6-25 days latter. Whether it is on the shorter or longer end depends on the wavelengths at the time of the pattern.

Screen Shot 2018-10-02 at 9.47.06 AM

Another known impetus for a change in the jet stream pattern to a trough in the East in the past has often been an October or November tropical storm system up the East Coast.

This is something we are currently monitoring 6-10 days from now:

Screen Shot 2018-10-02 at 9.55.00 AM

Past weather history shows us the areas they’ve tracked this time of the year in the past:

Screen Shot 2018-10-02 at 9.56.20 AM

Anywhere in the green is on the table for now.

It also fits the well-known “ridge over troubled waters pattern” when upper-level heights build over the Northeast states a weakness of low pressure develops underneath:


So the bottom line is there are multiple non-computer indicators of a pattern change that suggest the models should eventually show some below normal temperatures late October. We will temper that “excitement” with the knowledge that the models showed this happen once already and it failed to materialize on our side of the Appalachians. We got less humid and “cooler” than the hot but remained above-average.

Just FYI here is what the European model ensemble shows for the jet stream pattern early to mid November:

Screen Shot 2018-10-02 at 12.40.07 PM

Yes this crazy lingering pool weather has legs but it’s days are numbered.  I’ll be watching to see if the cross-signals get more compelling on the late month pattern change and let ya know.

Oh and as I’ve covered in recent blog posts about the El Nino “Modoki” being a more wintry version compared to a “canonical” version, I am also noticing a warm ocean anomaly South of Alaska, which has been nicknamed the “warm blob”. (Bering Sea, Gulf of Alaska, sometimes off  West Coast of U.S.). In the past that has been a signal for a “fun” winter in the East and South. We’ll have to monitor to see if it persists into winter.


The warm blob.

Follow me on Twitter @MellishMeterWSB.