0700 EST 23 AUGUST 2012
One of the really big rules in forecasting in the medium and extended range has to do with basic principles of understanding of how the weather Models are SUPPSOE to work. The problem is that a lot of TV meteorologists and some private sector meteorologists don’t really seem to grasp this. Let’s assume you have one particular solution from one weather model. Lets call it the ABC . Now all the other weather models are showing something else besides ABC. One weather model is showing XYZ.. another is showing MNO another QRS.. etc . Obviously this can cause quite a bit of confusion and variability or uncertainty in the forecast . Of course there are a number of ways and techniques good forecasters can use to get around or reduce this uncertainty or variability but for the most part … a A lot of forecasters are bamboozled when they come across a large array of possible solutions .
However if we keep in mind that the basic premise of forecasting with weather models is that as you get closer to the event … each day model accuracy should increase and you should have model consensus. This concept in taught in most of the undergrad meteorology schools and it is a very important concept. In other words weather models which are dealing with 48 hours now should be a lot more accurate and in a lot better agreement then 7 days from now.
Got it? .
That being so HERE is the BIG rule. When the other weather models begin to turn towards one particular weather model … say the ABC solution … it usually means that this particular weather model solution is going to be end up being correct . In other words as you get close to the Day 1 … if other the weather models turned towards the Canadian model solution ( for example) it usually means that the Canadian model solution will be correct. Or if the weather models turned towards the European model solution or the GFS model it usually means those model and that being the correct solution after all.
OK taking this idea and applying it to what is going on with Isaac…we see a very clear and distinct trend that all the models.: They are shifting towards the the European model solution of the past four days which is
1 Issac NOT going cut directly across Haiti and eastern Cuba
2 .. NOT going into south or eastern Florida
3 NOT coming up the southeast US coast…
4 Issac will move into the Eastern Gulf of Mexico.
The European model has remained very consistent –unlike all the other models –for the past four days. We can see again that the model takes the system south of Haiti and Eastern Cuba and finally turns across Western Cuba into the Eastern Gulf of Mexico. From there the 0z Thursday AUG 23 run the European model shows a land fall either in the far western portions of the Florida panhandle or in Southern Alabama… as a category 3 hurricane.
The Canadian model which has consistently taken Isaac into the Bahamas then up the Southeast U.S. Coast …suddenly now shows a Florida landfall which is something this Model has not shown before. This is a major shift towards the European solution.
The British model ( the Ukmet ) now shows a Eastern Gulf of Mexico track which again is a major change to cause this model has not shown that one time over the past four days.
and with the 0z GFS and GFS based hurricane models… We now see a track that is now not even on the West Florida Coast — which is what the GFS Models were showing yesterday . — but now the 0z GFS Model also has a track in the eastern Gulf of Mexico. So again we see a another shift to a west and a big step towards the European solution
Of course as I am writing this out.. the new and wretched 6z GFS Model comes out with a new ” issue”. Yes the 6z GFS is significantly west of yesterday’s GFS model runs …which was west of what we saw on Tuesday ….which was west of what we saw on Monday. But the 6z GFS model run has to come up with the idiotic solution. The 6z GFS takes Issac (after taking Isaac into Western Florida very close to Tampa ) then up towards Northern Florida… then the model phases or merges it with the jet stream over the Great Lakes and pulls the remains of Issac north into the Carolinas and Virginia with extremely heavy rains.
This is BULLSHIT.
We know it is BS because the solution makes no sense to meteorologically. We know it is BS because the GFS model is famous for over phasing or over developing coastal systems. And We know it is BS because no other model is showing anything like the 6z GFS. It is obvious and self evident that the model after 138 hours views the remains of tropical storm Isaac as a coastal storm located off the coast of Georgia… And that somehow another the calenda is now November as opposed to late August.
If we compare this solution to the very persistent and consistent European model which has shown any Eastern Gulf of Mexico track since Monday… without any serious variance whatsoever — unlike the GFS the Ukmet and the Canadian — the contrast is quite amazing. But more importantly there are tenn of Millions if not hundreds of millions of dollars riding on this forecast because the European model takes the remains of Isaac up into the drought ridden areas of the Upper Delta then into the heart of the Midwest with widespread significant rainfall. So if you are going to follow the 6z GFS and you bought a lot of soybean futures and you forecast more dry conditions in the 6-10 day you are about to end up in the poor house.
2240 19 AUGUST 2012
As a general rule I do not update the web site at all on Monday because it’s such a busy day for me with regard to the commodity trades that I simply don’t have time. Unfortunately with the system 94L developing in the tropical Atlantic combined with some realy exceedingly bad forecasts out there on the Internet… I guess I am going to have to find the time to break my rule and update.
First let’s talk about the date. Last year August 27, 28 29 was the landfalling dates of hurricane IRENE. Specifically over eastern North Carolina and into eastern Virginia it was August 28 2011. Not only was that a very important day for me and WXRISK.COM and the Facebook page but it also affected me along with many Virginians and eastern North Carolina residents in a major way. So naturally with this system possibly becoming ISSAC which should be yet a another ” I ” storm… along with IRENE last August 28-29 and ISABEL in SEPT 18 . 2003 … we are starting to see a lot of concern regarding 94L and if it becomes ISSAC. I mean if 94L does become ISSAC … and IF it hits NC VA …what are the odds of the three significant landfalling hurricanes in the lower Middle Atlantic states all being the 9th named tropical cyclone ( the “I’ storm)?
It’s one of those things that make you look and wonder if there is something else at work beside science but actually it’ i the human mind which is making these sorts of connection. To begin with 94L has taken a lot longer to develop then what anybody thought and this includes the folks at the National hurricane center. The morning satellite picture here on the 20th did not look very good. As you can see from this picture the immense Sahara dust cloud has in fact been entrain or pulled into the system. At the end of last week it look like 94 L would stay south of the Dust cloud and not really be affected by it. Clearly that is not the case and it is significantly impacting 94L’s ability to develop rapidly.
Because of this amount of dust entering 94L…. the European model over this past weekend backed off of the idea of a very powerful tropical cyclone / hurricane hitting the Southeast US coast (If you recall the European model last Friday was quite bullish on 94L becoming a big threat for the Southeast U.S. Coast). During this past weekend every single run of the European model showed no development at all and some of the European model runs actually kill the system completely over the next seven days!! Interestingly even though the GFS last week did not even see 94L and showed NO development…. during this past weekend the GFS models began to jump up and down about 94L even though it appeared to be struggling.
And because 94L looks weaker and is taking longer to develop… 94L is for more likely to head in a westerly direction which would take it into the Caribbean and very close to the big island’s of Hispanola and Cuba. This of course would interfere in its ability to develop rapidly were into a significant hurricane for an extended period of time.
This is vastly different from what we sought with IRENE or ISABEL . Those East Coast hurricane as did NOT come up through and over the big island’s in the northern Caribbean. Here is the European model surface maps from the Penn state E-wall site… and we can clearly see the model shows a very weak tropical cyclone move into the Northwest Caribbean into the eastern or southeastern portion of the Gulf of Mexico and hitting Florida.
On the other hand if we look at the 0Z Monday GFS at 180 hrs ( day 7.5) and 216 hrs ( day 9) we see a moderate intensity tropical cyclone probably not a hurricane moving across the far southeast portions of the Gulf of Mexico and hitting the western side of Florida. From there the system lingers over Florida — remember that after 192 hours the model resolution on the GFS becomes a pile of crap so it’s going to handle tropical systems a lot more poorly and with a lot less detail.
The Canadian model and the Canadian model ensembels .. was a bit more bullish with a tropical cyclone and showed a stronger system but one that was off the Southeast U.S. Coast. The Canadian model was not taking this system through the Caribbean or into Florida but paralleling the U.S. east coast and then heading out to sea: a very reasonable and viable solution even though the Canadian model is not a particular good model with regard to handling t ropical cyclones.
The 6z GFS showed more of the same thing … a TC of moderate intensity moving through or over Cuba and the Northwest Caribbean and making landfall on the west side of Florida
( ft Myers according to the 6z GFS) AUG 27 … then lingering over the west coast of Florida .
Here is the 12 the GFS…. And as you can see at day six the model has 94L as ISAAC but it is located over Central Cuba and by August 28 is located over west Central Florida
( HELLO GOP CONVENTION!!!.. didn’t this happen in 2008?)
In looking at the 12z hurricane models we see some important features. Notice that most of the models take this system through the big island’s over the Northern Caribbean. This is not good for those forecasting significant or rapid development. And not surprisingly the models hurricane intensity charts are also showing very weak and slow development. Yet despite all this there were several interesting forecasts and speculation from some meteorologists and people who stayed at a holiday inn last night that 94 L/ Isaac would become a major hurricane .
That of course is a distinct possibility but as you can see there are a lot a reasons for that not to be the case.
Next the 12Z Canadian model came out and it showed a very powerful hurricane headed right for the North Carolina Virginia Coast in a manner very similar to what we sought with IRENE and ISABEL .
Those 2 hurircanes came across from the tropical Atlantic passing east and northeast of the Bahamas -they did not come up to the Caribbean. So the Canadian at least it is showing a scenario very similar to what we saw with IRENE and ISABEL. That being said the Canadian solution is unrealistic because the model takes 94L/ ISSAC across the western side of the W.A.R. ( the Bermiuda HIGH) which does not make a lot of sense meteorologIically speaking.
If we look at the Canadian ensmbles to see how many members support or MATCH the 12z operational Canadian and its big hit to the East coast scenario we find a few members supporting it but not a lot. Again this is a warning sign that the operational Canadian scenario is probably not correct.
On the other hand many people saw the 12z GFS Hurricane ensembles .. and went ..”WOW east coast hurricane !!..” This is also a track very similar to what we saw with IRENE and ISABEL. But here we have to remember that the GFS ensemble is run at a lower resolution then the operational model and as result the model appears to be phasing or merging the hurricane and a trough in the jet stream coming in from the Great Lakes . This causes the hurricane to come up the East Coast according to the GFS ensembles.
The Monday afternoon evening satellite pictures are little more promising and we can clearly see a closed circulation even though there is a gap in the convection on the northwest side of the disturbance.
Indeed the 12z European model came out and actually showed a somewhat better developed 94L that was no longer disintegrating into an open wave . The European model was now detecting an actual tropical cyclone (ISSAC) in the Southeast Gulf of Mexico and threatening the west coast of Florida.
What we do have really good agreement on is that the European Ensembles AND the GFS ensembles show some sort of decent size tropical cyclone in and around Florida or the Eastern Gulf of Mexico or just off the southeast coast AUG 28-29-30. But it does not look particularly strong or well organized.
SUMMARY… 94L is NOT developing as fast as what I and many others thought it would. It has been pulling a lot of dust and dry air which has inhibited its development and is allowing the system to take a more westerly track so that it passes south the Puerto Rico in all probability. This means that the system is going to interact with the big island’s and the Northern Caribbean and will probably interfere will restrict its ability to rapidly develop into a significant hurricane. Those calling for a major hurricane for the Florida and or the Southeast U.S. Coast at this time are doing so without any basis in fact and are ignoring the significance of the model data and the overall track. 94L will likely become a tropical storm and it will be given the name of ISAAC. But this system’s track is significantly different from what we saw with IRENE and with ISABEL … and folks in Virginia North Carolina at this time are probably looking at the remains of tropical storm or weak hurricane coming up from Florida and Georgia — in other words no big deal.
There is of course the possibility that the system could rapidly intensified over the Eastern Gulf of Mexico before it came into the west side of Florida… And there is even a small possibility that the system might end up tracking north of the big island’s in the Northern Caribbean and approach the southeast U.S. Coast from the Bahamas. This obviously would also oppose a much bigger threat to the southeast coast. But these are very small possibilities at this time.