Tropical Storm Matthew formed this morning out of Invest 97L about 35 miles southeast of St Lucia in the Leeward Islands, or about 2250 miles west-southwest of Berwick Bay. Yes it is a long way off and only a tropical storm with winds of about 60 mph. It is really just starting to spin up and pull the winds and weather in around the center. That said most of the models put the system as a 90-knot hurricane a couple hundred miles south of Jamaica late Saturday. Barring crossing any major landmasses, conditions will be ripe for further intensification and many of the models are predicting a Cat 4 if not a Cat 5 storm in Matthew’s future. A long ways off for sure but the trend will definitely be towards a big one!
So where is this monster in the making headed? That is definitely the $64,000 question this morning and no one has a clue what will actually happen after the weekend. All of the models turn the system more northerly, but in degree of turn they diverge greatly. The major models have landfalls anywhere from the Mexican coast to no landfall at all out in the central Atlantic. So why even predict? In truth, it is a little clearer than that. We will look at that in a bit.
First, let’s look at what is causing all of the variance in the models. In essence it is the wonderful first taste of fall weather that began yesterday and will truly blossom over the next three days. Highs not making 80 and lows near 60 with wonderful dry air and clear skies. It has been a while! But it is an odd scenario generating this wonderful weather. Normally we would see a big deep low pressure center crossing through the Great Lakes, with a cold front trailing to the south followed by a big high pressure area building in its wake. But this time, a piece of the big low is going to split off and drift south over Georgia and stall as it is out of the main upper air flow. Eventually, it pulls out to the north ahead of the next front sometime over the weekend. It is that cut-off low which holds the key to where Matthew will head. If the low pulls out quickly and weakens with high pressure rushing in to replace it, the Gulf is definitely in the cross hairs. If the low is stronger and/or slower to pull out, it is Florida and the East Coast under the gun. Here are the two model ensembles of potential tracks from the two top models in the world. Courtesy of Mike’s Weather Page, by the way. Not exactly a match!
On top is the European (ECMWF) Model for the next 14 days. On the bottom is the US (GFS) Model for the same period. As I explain below, interpreting ensembles is an art. But basically, the two top models in the world have very different ideas. Ignore all those squiggles and look at the red outlined cones I have added. You will note I have also ignored the major outliers in each case. For instance, some of the ECMWF runs show the storm going west into Mexico near Tampico. That is highly unlikely at least for now, so I have it outside the cone. But you can see how difficult this is to judge.
As I said, ensemble models are not the easiest to understand. The best analogy I can think of is this. Someone has started cooking a huge gumbo to feed hundreds and just walked off leaving you to you to finish it. All you know is what ingredients went into the pot but not how much of each. Health regulations say you cannot sample the whole gumbo pot to check its flavor until it is completely finished. In fact you can’t taste it at all, even a small sample in another bowl, until you have made your changes to that small bowl. So you take about 20 small samples from the big pot and add a bit of this or that to each, covering all your bases, then taste them to see what needs to be added to the big pot to make it Louisiana good.
That is sort of what the ensembles do. They tweak the potential outcomes to cover all the minor changes which might take place. Unfortunately, we can’t sample them until the very end in this case, when we actually know where Matthew went. However, as we watch each successive model run and see which changes seem to be matching what is actually happening, we learn how to target our tweaks for a narrower cone. But it takes many runs and there are only a couple runs of each a day.
Now I know some of you 😉 You want to know what my gut says and will message me as soon as I say “we just have to wait.” So here it goes but this is a flat out guess and confidence is not high. At the moment, I am leaning towards the ECMWF solution for two reasons. First, the ECMWF, while far from perfect, is the best model in the world. It is the only 4-demensional model out there until the GFS fully integrates temporal data analysis in the near future. Second, the GFS has played this cut-off low as stronger and bigger than appears to be materializing. It would have had us in the mid-70s for highs and mid-50s for lows. That is quite a big miss from what appears to be coming. The ECMWF just appears to have a better handle on the cut-off low right now. If that trend holds, the ECMWF should also have a better handle on the future track of Matthew. But man that is a lot of “ifs.” Too many for any degree of confidence.
Even if the ECMWF scenario verifies, the cone runs from Corpus Christi to Savannah right now, over a thousand miles. Not exactly a pinpoint scenario. But given the extreme uncertainty not just in these models but many others, that wide cone also seems more appropriate to me right now than the narrow GFS cone. But only time will tell. Keep an eye on Matthew as you enjoy our first gorgeous Fall weekend after months of hot and muggy/downright wet weather. By early next week the models will have a better handle on things and we will know who needs to be taking action to protect life and property ahead of Matthew. If this were not complicated enough already, Miche and I are set to fly to Boston over the Columbus Day weekend to visit Evan. So we have a potential of dealing with Matthew either here and up there depending on which track verifies 🙁 Got to love those tropics. For now, just enjoy this wonderful weather!