WHY YOUR SMART PHONE WEATHER APPS or the Temp forecast on your favorite Weather web site suck.
Yesterday when I issued the NEXT THREE-WEEK Newsletter, I asserted that the threat of extreme heat (temps above 90) for the eastern US — outside of the Gulf Coast states — is not over. Sure enough, the operational GFS came out this Tuesday morning and has temperatures in the low of mid-90s on September 21 -22 in the lower Middle Atlantic. Indeed, If you look at your favorite weather smartphone app or if you go to your favorite weather website, you will probably see temperatures for various locations in the Middle Atlantic that reflect this.
This is because all of those temperature forecasts on those various different platforms are ALL based upon the operational GFS model and nothing else. Now, when you are dealing with just the next couple of days that is usually not a problem. But during the past year the changes that were made to the operational GFS model has caused the model to go absolutely batshit crazy with its Max temperature forecasts. This is exactly what we see happening with the operational GFS model from this Tuesday morning valid for September 21. The GFS temperatures are completely Goofy, and the following images show how and why they are completely goofy.
Please take a look at the four Images here. They are four different projections of the max temperatures in the lower Middle Atlantic region valid September 21. All 4 projections are based upon the data from early on Tuesday morning. ideally all 4 projections should have temperatures that are close to each other especially since none of the 4 models are projecting anything like rain or cold front or cloud cover which might interfere with temperature.
Image #1 is the operational or REGULAR GFS model that you see getting talked about on the Weather Channel or Weather Nation or on your local TV weather guy. as you can see it is extremely hot and has temperatures in the low to mid 90s for September 21.
IMAGE #2 Is the GFS ensemble. The idea here is that forecasters run the GFS model with slightly different versions 21 times and compare it to the operational GFS. The mean or average of 21 is far more likely to be correct than just one model which might have a bad model run. And as you can see the GFS Ensemble is vastly different with its temperature forecast for September 21. This Discrepancy is a warning sign that the operational GFS and its extreme heat for September 21 is likely to be wrong.
IMAGE #3 is the operational European model. Here we can compare the op-European model to the GFS and GFS ensemble. When we do that, we find that the European model temperature forecast for September 21 is very close to the GFS ensemble. The op-EURO does not support the operational GFS and its extreme heat for September 21.
IMAGE #4 Finally here is the European ensemble which arguably is the best weather model in the world. Its temperature forecast is identical to its operational European temperature forecast. This is a good thing because it means that the models are in agreement. But note the European Ensemble does NOT match the extreme heat of the Operational gfs.
SUMMARY — -only the operational GFS is forecasting temperatures in the low to middle 90s in the southern portions of the Middle Atlantic on September 21. No other model data of any kind supports the ridiculous heat showing up on the GFS model for September 21. Believe me I can show a lot more models out than these and all of them say that the operational GFS is goofy.
FINALLY, it needs to be pointed out that the operational GFS has been having problems all summer long with its temperature forecast. In the Midwest for example and during this past summer the operational GFS forecasted Max temperatures in Des Moines Iowa above 100° more than 20 times on the 12z or midday operational runs of the GFS model. Yet max temperatures never got that hot in Des Moines — not once!! The operational GFS forecasted temperatures above 105°, eleven times in Des Moines Iowa during this past summer and that did not happen.
Clearly something is wrong with the operational GFS model. NOW — as we move deeper into the autumn and the winter season, unless that problem with the GFS model is fixed, its extended temperature forecast in this coming winter beyond day 4 is going to be really off. This is important because the GFS model might be forecasting a max temperature of 45° with rain for your location where all of the other data is showing the max temperature of 32° with snow. Not only is this going to cause a lot of frustration for meteorologists like me, but it means that your smartphone apps could have huge wild swings in them this winter as the temperature suddenly change 15° or more when the GFS model finally gets a freaking clue.