The midday radar shows moderate to heavy showers and thunderstorms covering70% of northern Iowa and the southern third of Wisconsin. There is a second cluster of activity covering 70% of Arkansasandextends into far southeast of Oklahoma and in southeastern Kansas and northwest Mississippi. There is also some scattered activity across Eastern Tennesseenorthwest Georgia central Kentucky and southwest Ohio.

this first map is the 12z Thursday operational or regular GFS model RAINFALL forecast for the next 5 days. As you can see the rainfall amounts continue to decrease as the model goes further out in time. The GFS model does very well when it comes to producing rainfall from actual thunderstorms in a specific location. But what the GFS does not do well with is the rainfall outside of the thunderstorm. The model seems to think that if a particular location gets 2 inches / 50mm  of rain ….that the rainfall gradually decreases as you go further away from the thunderstorm. That is not how thunderstorms work. You are either in the rain in the thunderstorm while you are not. This results in the GFSoften shows a lot of areas getting 0.25”/6mm.....   0.50”/ 12mm......0.75”/20mm.  of rain with 50 or 60% coverage. So GFS operational model ends up producing a fairly wet looking forecast. That is exactly what the last few runs of theoperationalGFS models have been showing but todays GFSseems to have seems to have " seen the light "

This next image shows the jet stream map valid for Tuesday morning July 14th. There is a trough on the East Coast which is beginning to weaken and will move up into southeastern Canada. There was a Ridge across the southern States from Arizona to Alabama and there was a large trough that is moved into western Canada and the Pacific Northwest. Within this trough there was an upper Low in central Saskatchewan.

This is the problem that the GFS models are having- difficulty in handling this trough. The GFS model often over develops or intensify troughs. It is one of the flaws or biases within the model. In the weather business, this is known as the “GFSCOLD BIAS” .This image shows the jet stream maps from the operational GFS on the left side and the GFS ensemble on the right side.I have drawn in the surface LOW pressure area along the IS. Canada border and the trailing cold front. The operational GFS on the left side has a much stronger trough in Western and central Canada that pushes deeper into the upper Plains in the Great Lakes. This in turn forces the heat Dome to set up further to the South initially. It also allows cold front-- which is the thick black line I have drawn in --to push into the central Plains and the WCB … and hence more rain. The GFS Ensemble( right)has a much weaker trough with the cold front further to the north because the Dome has developed over the lower Midwest and the southern States. That means less rain in the 6-10 day in the WCB and the upper Plains

This image shows the operational GFS temperatures valid for July 16th on the left side and the rainfall for the 6 to 10Day from the operational GFS on the right side. The thick black line that I  have drawn in on the temperature map is the 92 degree line. Noticed that because the cold fronts as of  July 1`6    has pushed all the 92f/ 32C PLUS  temperatures south of the Midwest.   The central and lower Plains and Delta and the Gulf coast states are beastly hot with temperatures between 95 and 105 degrees but north of the black line the operational GFS has temperatures 92 degrees or cooler and even shows some 70s and portions of Northern Iowa Wisconsin and Minnesota. Because the cold front comes further south on the 12z Operational the GFS Model produces significant rains from eastern Nebraska to central Iowa into southern Wisconsin with amounts of 1-5”/ 25-125mm

This next image we compared the operational GFS rainfall for the 6 to 10Dwith the GFS Ensemble on the right side also for the 6 to 10D.Notice that the rain fall from the GFS Ensemble is dramatically less than it is on the operational model.

Because the 12z operational GPS solution is not supported by the GFS Ensemble or the European model I rejected it completely. It is clear in my opinion that the GFS model continues to have issues in figuring out how strong and deep the trough and upper Lowin south central Canada is going to be. This is going to cause the GFS model for flip flop back and forth for the next few days. The GFS Ensemble solution handles this aspect of the forecast much better…and I am following it with respect to temperatures and the rainfall.

This next image compares Upper air pattern valid day 9  July 18  again with the regular GFS on the left and the GFS Ensemble on the right. Notice that the shape of the dome in the Midwest is bigger larger and a little further to the north.

This has an impact on the temperatures. These two images show the forecast temperatures with the GFS model on the left side valid for July 19th. Notice there was a tremendous amount of heat from Interstate 70 South word but in Iowa East Nebraska the northern half of Illinois Indiana temperatures stay in the 80s.However the GFS Ensemble on the right hand side is significantly hotter with upper 90s getting into much of Missouri southwest Iowa all of Nebraska Central and Southern portions of Illinois. The GFS Ensemble continues to show this sort of impressive heat right through day 15 which takes us to July 24-25.