Thanksgiving weather uncertain 

Posted: 7:25 am Thursday, November 16th, 2017

By Kirk Mellish

The computer models have been swinging wildly back and forth over the past week as they try to determine the week ahead and beyond. This seems likely due to their struggling to handle changing energy and ocean patterns in the Pacific.

There looks to be at least SOME chance of rain Thanksgiving give or take a day, but as of now the odds look SMALL. However, confidence in the forecast one way or the other is low due to the high chaos factor currently in the data.

This is also true for temperature specifics, but as of now the next 15-days look to be cool with just a few mild days in the mix.

Screen Shot 2015-12-09 at 1.06.52 PM

If you’re unfamiliar with Lorenz Chaos theory “butterfly effect” and weather forecasting with ensembles you can google the terms.

Here are the European and American model temperature guidance output from WeatherBell, when the lines are close together there is high confidence, when they spread apart and gray bars grow larger confidence is lower and uncertainty high due to increased chaos:


Therefore, confidence in output is LOW after next Tuesday:


The same is true as to if it will be wet or dry on or near the Thanksgiving holiday.

But at least for now the operational or deterministic (non-ensemble) models from Pivotal and WxBell both on the GFS and ECMWF show THANKSGIVING DAY dry around Atlanta:

Screen Shot 2017-11-16 at 7.10.37 AMecmwf_slp_qpf6_noram_32ecmwf_ir_noram_33

And they show mostly dry the day before and day after as well. But remember, confidence is lower than normal due to the model chaos.

Here is the 16-day GFS Ensemble temperature trends continuing the cool average with brief warm spells in-between:


FOLLOW me on Twitter @MellishMeterWSB.