When 80% is not enough

Wednesday, August 17 at 2pm ET the National Hurricane Center issued the first outlook on what would become probably the most talked about area of interest (invest) in the history of mankind. That outlook began a process that kept scores of weather geeks, emergency managers, TV meteorologists, weather forecasting firms and the general public glued to their Internet-connected devices. From the get go it looked like this could be “the one”. It had potential to make it all the way across the Atlantic and possibly affect the United States after roaring through the Caribbean. And so here we are at the other side of that long journey and what became known as “99L” to millions of people will long be remembered for what it didn’t do: failed to develop.

As the process of tracking the tropical wave got underway, computer models were generally in agreement that the system would move westward at a fairly low latitude. However, it was clear early on that mid-level dry air, somewhat associated with the Saharan Air Layer (SAL), would be an inhibiting factor for development. Never the less, by early Saturday morning the 20th, it looked as though a tropical depression would form. Chances went from virtually nothing to 50% in just three days.

By Monday evening the 22nd, the probability increased to 60% and it looked as though a hurricane was possible for the southwest Atlantic Basin, maybe even close enough to Florida or the Carolinas to warrant concern. The social media hurricane machine was in full throttle mode with every expert (and non-expert) opinion you could imagine being thrown in to the mix. It was “invest 99L” overload and it would only get worse.

Leading the charge for development at first was the dynamic duo of the GFS and the ECMWF. Both seemed to latch on to 99L and make something of it. At first, it looked like a track towards Florida and then a turn north was likely. At times, the GFS had an enormous hurricane sitting not too far off the North Carolina coast, moving harmlessly out to sea in the longer term. It looked like this would be a close call but no guarantees yet for a landfall.

As the days went by, seemingly very slow considering we can watch the computer models come and go virtually 24 hours a day, things began to change. The GFS dropped the development almost completely while the ECMWF put memories of Katrina, Andrew and Betsy in our collective minds. It wasn’t just the Euro, the new and improved HWRF model, which is specifically designed to forecast tropical cyclones, showed similar forecasts of a hurricane headed towards south Florida and then turning west as it tracked south of a strong high pressure area over the mid-Atlantic. The hurricane world went in to full meltdown mode. It had been since before Twitter, Facebook as we know it, the iPhone and even the birth of One Direction that Florida had been hit by a hurricane. It was Tuesday, August 23 and 99L had a 70% chance of developing in to at least a tropical depression. The “H” word was bound for Florida.

To make matters worse, the ECMWF ramped up 99L to beast mode and sent it towards Louisiana, the LAST PLACE that needs a hurricane this season. Meanwhile, the GFS showed almost nothing at all. Just weak energy coming across and never really doing much. People were scoffing that if the Euro was right, congressional hearings needed to take place immediately to “fix the GFS!”. It was pure madness because, you know, the ECMWF nailed Sandy’s forecast from 8 days out. People were hung up on a hurricane from four years ago, forgetting the failures of even recent events like Erika last season and Fiona in 2010. Oh yeah, let me tell you about that one.

In 2010, we had Earl, headed for a possible landfall up the East Coast. Meanwhile, the ECMWF was advertising Fiona to be “the next Katrina”. A few in the weather business bit off on this and ran with it like the end times were coming. Earl was going to be a dud while Fiona would be a disaster of epic proportions. All the while, the GFS showed Earl doing exactly what it ended up doing: coming close to the NC Outer Banks then turning out to sea as a strong hurricane. As for Fiona, the GFS showed it as nothing more than a nuisance and that’s how it all turned out. GFS right, Euro wrong. Funny how few people remember that. I remember, trust me, I was on the Outer Banks where Earl brought 85 mph wind gusts to our anemometer that was set up next to Oregon Inlet.

By Wednesday morning, just 48 hours ago, the chance for 99L to develop made it to 80%. It looked like a sure thing now. I mean 80% is pretty good, right? In basketball an 80% free throw shooter is considered to be almost lethal from the line. You foul that player and you might as well put two points on the board. Eight times out of ten the shots go down.

But what happens when that player is in the title game and his team is down 1 with .90 seconds on the clock and he is fouled driving to the hoop? Everyone holds their breath as he lofts the first shot to tie the game. He misses. It’s stunning. The announcers balk about his percentage and how they can’t believe he missed. Must be the immense pressure. Time for try number two. He misses. No one gets the rebound and the clock expires. His team loses the National Championship for one simple reason: 80% is not enough. It’s not 100%. There was a 20% chance he missed either of those two shots. Now this is purely hypothetical but it makes my point. Sometimes high probability is mistaken for certainty.

In the case of 99L, it looks dead and gone now. The GFS, in its past few days of runs, turned out to be correct, for the most part. No hurricane coming for south Florida this weekend like the Euro showed. To be fair, the HWRF did as well and it busted big time. Instead, the Euro now has weak energy bringing possible heavy rain to portions of Florida. No hurricane in to Louisiana or elsewhere, just a strung out mess.

Odds of development over the next five days are down to 60%. Strong wind has all but beat the system in to oblivion. There is virtually no convection or thunderstorm activity with it and the USAF Hurricane Hunter crew has been grounded since there’s nothing there to investigate.

Sixty percent. Hmmmm. That’s not too high nor is it very low either. What is the reason behind this number? It’s because there is still a chance, apparently a 60% chance as of this writing, that 99L will develop some in the Gulf of Mexico. How could this be? The Euro “dropped it”. Well, the ever-excited HWRF sure didn’t and the GFS now shows limited development in a few days with some rather wacky tracks thrown in for good measure. To be honest, it’s giving me a headache to watch this day in and day out and all we have is an area of interest, even if fading to an area of blue skies.

I caution that while the gist of my post is aimed at pointing out the obvious, that anything short of 100% probability has a chance of falling short, it is also true that unless it’s zero, there’s still a chance. Even the 60% free throw shooter is sometimes the hero.

There is still some energy down in the region near the southeast Bahamas that might be able to survive long enough to warrant keeping an eye on. None of the computer models that show any development potential do so until later in the weekend – so let’s see what happens. Perhaps 80% was too high early on but 60% will be just enough for now.

We want to be able to trust the computer guidance to give us time to prepare if in fact something is going to develop. However, the public should understand probability and know the limitations of forecasts even in the relative short term. Most people don’t have time or interest to fully invest their energy in to making sense out of it all. I do my best and could not for the life of me understand why the GFS gave up on this system while the Euro did not. Sometimes you need to look out the window, so to speak, and view the actual weather and not just the predictions. The satellite presentation of 99L never really looked promising for development. It came close a couple of days ago but the reality was it didn’t have that “look”. So logic should have dictated that if it looks poor and one of the major global models insists on non-development, then maybe that’s why: because it won’t. It’s a simple concept and for now, it turned out to be right. Shear and lack of convection and overall organization has kept 99L from developing but it’s not necessarily over.

I think the next 48 hours will be all we need to know how this ends. If nothing happens by then and the wave of energy basically spreads out or dissipates completely, then we’re golden. Until that happens, it’s obviously prudent to keep an eye on things – just in case 60% is just enough to do what 80% could not.

I’ll have more in my daily video discussion posted here, to our app and on YouTube later this afternoon.

M. Sudduth 9AM ET Aug 26

Share

Model guidance duel continues as 99L tries to organize

Satellite image clearly showing that the low level center, weak as it may be, remains exposed due to strong upper level winds

Satellite image clearly showing that the low level center, weak as it may be, remains exposed due to strong upper level winds (click for full size)

The overnight runs of the various computer models did not really get us any closer to “knowing” the outcome of what becomes of 99L.

As it stands now, the tropical wave continues to slowly get better organized over the warm waters of the extreme southwest Atlantic – but I emphasize the word “slowly”. It seems that strong upper level winds are continuing over the system, preventing the thunderstorms from persisting and wrapping around the broad area of low pressure situated just north of eastern Hispaniola.

Despite the lack of organization, strong winds are being observed in the convection that itself is well removed from the low level center (poorly defined low level center). In addition, heavy rain is spreading over portions of Hispaniola and this is cause for great concern due to the risk of flash floods and mudslides. Remember, there are some fairly tall mountains in the region and tropical rain fall can lead to lethal flooding and tremendous damage.

The wave of low pressure is forecast to continue moving off to the west-northwest today and tomorrow, reaching the southeast Bahamas during that time frame. From there, it is likely to track through the Bahamas and toward Florida this weekend.

Early morning run of the HWRF model showing a strengthening hurricane in the eastern Gulf of Mexico in about 96 hours

Early morning run of the HWRF model showing a strengthening hurricane in the eastern Gulf of Mexico in about 96 hours (click for full size)

Now the tricky part: how strong does this system get – if at all? My answer: I honestly don’t know. The computer guidance is very confusing with the majority of the U.S. generated models, such as the GFS and the hurricane-specific HWRF model both showing little to no development, at least not in the short term. It is interesting to note that the HWRF, which stands for Hurricane Weather Research Forecast, is now indicating significant strengthening in the eastern Gulf of Mexico, around days 4 and 5. I am skeptical since that same model has had a very difficult time developing 99L much at all the past day or so.

On the other hand, consistency has been the rule for the much talked about ECMWF model or Euro. It has not strayed from its forecast of a possible hurricane impacting south Florida later this weekend. From there, it goes on to strengthen and make landfall up near the Big Bend area as a formidable hurricane. Needless to say, this is quite concerning but exactly how much stock do we put in to it? I just don’t know what to make of all this right now. On the one hand, the remarkable run to run track record of the Euro makes me think it could have the correct overall solution – that being a potential hurricane for some portion of Florida in the coming days. Conversely, the lack of development seen by the GFS and the waffling of the HWRF model for track and intensity makes me wonder: will anything happen at all?

ECMWF from the overnight run showing tropical storm conditions for the southern portion of Florida this weekend

ECMWF from the overnight run showing tropical storm conditions for the southern portion of Florida this weekend

All of this is not good for the public and the perception of how hurricane forecasting is “supposed” to be. Usually we don’t keep waiting and waiting just to see if a system will develop. The advantage, if you’re going to have to deal with a hurricane, is knowing it is coming in the first place. The “what if” scenario here is a little unsettling.

Let’s suppose that 99L does indeed wait until 24 hours before landfall in south Florida, assuming it does in fact take a path in that direction. If it were to quickly intensify over the very warm waters, how fast would it ramp up? Could it become a hurricane rapidly? Yes it could. We’ve seen it before but it’s been a while. Katrina in 2005 was on a fairly steady pace to strengthen as it approached SE Florida from the Bahamas but we at least knew it was a strong possibility well ahead of time. People will react better to a named storm or a hurricane headed their way than to a tropical wave. At least that’s my thought on the matter. This is a tough situation as the longer we wait, the less time there is to get ready if the need arises. While we would all like to wish that everyone along the coast was prepared anyway, we know the reality and it makes for quite the sitting duck in situations like this.

The afternoon runs of the various models (actually based on morning data) will be quite telling – or not. Questions about whether or not the GFS steps it up and develops 99L may be answered. What if it has the answer already and nothing much is going to happen to begin with? Maybe if the Euro suddenly shows little to no development we can at least say, ok, two major global models now show a low impact event. We will just have to wait and see. Either we know more or are stuck with dueling models once again.

I will post more here this afternoon including my daily video discussion.

M. Sudduth 8:50 AM ET Aug 25

Share

Joaquin poised to make historic landfall

Satellite photo of hurricane Joaquin

Satellite photo of hurricane Joaquin

It all began as a rather innocuous area of spin in the mid to upper levels of the atmosphere a little more than a week ago. What was once just an upper level low, producing some showers and thunderstorms over the warm Atlantic, is now hurricane Joaquin. Most hurricanes form from other sources such as tropical waves that emerge from Africa. Joaquin is unique – it is that rare hurricane whose origins can be traced back to a system that is cold in the middle, not warm like a hurricane. And so here it is and so here we go with the anguish of worrying about where it ends up. The potential for something historic is on the table and those who know my writing know that I rarely use terms like that.

First – the stats. As of 8am ET, Joaquin was a 75 mph hurricane moving towards the central Bahamas. This is the first region that will have to deal with the effects which may be quite intense as the hurricane continues to intensify over very warm ocean water. As such, hurricane warnings are up and people in the region are hopefully preparing. The slow movement is a problem too – it means a prolonged period of wind, rain and surge for the Bahamas.

Once Joaquin turns north, and it should according to the official forecast, things get very interesting. A lot was made about the fact that the ECMWF model, considered to be the world’s best by many, nailed the evolution of what eventually became devastating hurricane Sandy. Somehow, the model “saw” the track as far as seven days from the landfall in New Jersey. All the while, the American based GFS model handed Sandy’s energy off and sent it packing out to sea. We all know the end score – Euro > GFS.

Here we are three years later and another global model duel is at hand. This time, Joaquin is the name and the end result is still in question. Why? We are talking about less than five days, maybe six at the most. How can the global models not be locked on the solution that can give forecasters confidence that their track and intensity ideas have solid merit? Basically, it’s the pattern.

Sandy was very unique in that a Caribbean hurricane moved up in to the southwest Atlantic and was then pushed out to the northeast and away from the United States – only to be blocked by an enormous ridge of high pressure which sent it back towards the Mid-Atlantic where a deep, strong trough captured it. The set-up for Joaquin is similar yet different. This time, it’s going to be early October. The trough in question is going to cut off from the main flow and not be nearly as strong as the one that captured Sandy. Water temps are quite a bit warmer this time than what they were in late October 2012. Joaquin has a chance to make landfall purely tropical with a concentrated area of winds and the potential for a devastating storm surge. When and where that could happen remains to be seen.

GFS (left) vs ECMWF (right) and their positions of Joaquin at 102 hours

GFS (left) vs ECMWF (right) and their positions of Joaquin at 102 hours

This brings me to the GFS vs ECMWF duel.

Check out the graphic showing the GFS track from the overnight run (6z). Clearly it curves around the cut-off low and bends back towards the North Carolina coast. This would be a very bad scenario for obvious reasons.

Now look at the ECWMF position at the same time – 102 hours. The difference between the two positions of Joaquin is incredible and means a completely different outcome for each model.

One has to wonder why such a spread between the two global models? I wish I knew. Obviously, the GFS captures Joaquin with the cut-off low and swings it back towards the coast. The ECMWF, on the other hand, finds just enough of an escape route offshore to allow the hurricane to turn safely away from the United States. Which solution will turn out to be correct? Well considering that the ECMWF seemingly lies alone in its “thinking”, it looks more and more like the other models, the GFS included, have locked in on what will eventually be a nasty hurricane event for some location(s) along the East Coast.

If the ECMWF turns out to be correct, it will be an incredible turn of events and mean that the current track forecast that we see now will be turned almost sideways, pointed eastward instead of towards the coast. It’s possible but at this point, it’s hard to believe the GFS, which had more data from the NOAA G-IV mission last night, will be totally wrong and eventually flip to the “out to sea” track. I guess anything is possible with weather so we shall see.

Precipitation forecast over the next 5 to 7 days showing incredible rain fall for parts of the East

Precipitation forecast over the next 5 to 7 days showing incredible rain fall for parts of the East

All of that aside, what can you expect if you live along the East Coast? Well, for one thing, rain! The trough and upper level energy coming in to the Southeast and East will set off a significant heavy rain event even before the supposed arrival of Joaquin. Take a look at the precip forecast map from the Weather Prediction Center – notice how vast an area is covered by 6+ inches of rain over the next several days. Add the effects of a hurricane to the mix and we have the set-up for what I term a history making event. Flooding from freshwater is astonishingly lethal. The fact that the Appalachians could get excessive rain makes me very nervous. The Piedmont is also very vulnerable in this kind of set up. I urge people to make sure they are aware of the weather forecast for their local area. Use weather.gov as a source – read the warning info, the discussions and tune in to your LOCAL TV and radio sources. This much rain, combined with a potential landfalling hurricane, is simply too much to ignore and brush aside as hype. This situation could have lingering impacts for years to come and people better be paying attention.

As far as direct impacts from Joaquin – the Bahamas are up first, then we wait. If current forecast trends continue, it looks like a hurricane strike for North Carolina, Virginia or even points north. Swells will move out ahead of the hurricane which will make the already battered beaches even more battered. Surfers will love it but swimmers will need to simply stay out of the water. The rest is up to the hurricane and where it ultimately tracks. We can look at what impacts to expect when and if that time comes – there is still time for the ECMWF solution to be correct and save the day – wouldn’t that just be something?

I will produce and post a video discussing further outlining much of the content that I covered here. I expect to have that online by later this afternoon. As always, you can follow along in our app – Hurricane Impact (two words) in the App Store and on Google Play.

M. Sudduth 9:40 AM ET Sept 30

 

Share

GFS up to no good with its phantom hurricanes

GFS model output from a few days ago showing a hurricane looming in the Caribbean Sea

GFS model output from a few days ago showing a hurricane looming in the Caribbean Sea

It looks as though the GFS model, the main U.S. generated global model, is suffering from some kind of issue that creates hurricanes out of cumulus clouds.

As I mentioned in recent posts, there appeared to be a chance that we would see development take place down in the Caribbean Sea as the week progressed. While that could still happen, I don’t see it as being a big issue.

This year in particular, the GFS model has had an annoying tendency to develop hurricanes out of the slightest of disturbances when other models, especially the rival ECMWF, did not. Remember Cristobal back in late August? At one point the GFS had it becoming a powerful hurricane, parked right off of New Orleans. Social media went nuts about this since some decided to post images of the output without any explanation of how ludicrous the notion was, considering how far out in time we were talking about, etc. If every hurricane that the GFS developed in week two of its 16 day output actually happened, no one would want to live at the coast! It’s pretty bad and I hope it gets fixed sooner rather than later.

Fortunately, we do have other global models available when looking for what is called “tropical cyclogenesis”. This is just a fancy way of saying development. I do not claim to be a computer model expert, not by any stretch. I imagine there is quite a bit of physics and math that goes in to programming the powerful suite of models used around the world to predict the weather. Picking out the birth of a tropical cyclone is not always an easy task but some models seem to do better than others. This year, the GFS has not been one of those models and I think this is another case of crying hurricane instead of wolf.

Nevertheless, the Caribbean Sea is favored this time of year and it would not be smart to completely ignore the chance of development in that region. Remember too that it does not take a hurricane to cause significant problems. Even a tropical storm with torrential rain could impact land areas from the Caribbean northward to Florida and the Bahamas during this time of the season. While the GFS may be suspect, and perhaps even more so now, it’s worth keeping tabs on the Caribbean until the end of the season but as for now, any hurricanes that the GFS dreams up look to be just that, the stuff of fantasy land.

In the east Pacific, Simon continues to dwindle away and will soon be a remnant low pressure area. Periods of rain, sometimes heavy, will move in to the Southwest U.S. but this does not look to be as widespread an event as Norbert and Odile were. So far, there are no flood watches posted for the region. Check out this graphic from the NWS Tucson. It explains the upcoming event very well in one convenient graphic.

Meanwhile, a powerful typhoon is churning away in the west Pacific. It is currently one of the strongest tropical cyclones of 2014 and could threaten Japan, though much weaker than it is now, in several days. It seems as though the Pacific Basin as a whole is the place for all the action this year. If it were not for hurricane Arthur back on the 4th of July, the Atlantic season would seem almost non-existent which is pretty much in line with what was forecast for this season. I will discuss that in more detail in an upcoming blog post.

M. Sudduth 1:15 PM ET Oct 7

Share

Sandy nearly a hurricane and poised to become historic event one way or another

Sandy's wind field will ikely expand and could bring tropical storm conditions to a large part of the Southeast U.S.

Sandy's wind field will ikely expand and could bring tropical storm conditions to a large part of the Southeast U.S.

It is going to be a rough day in Jamaica as Sandy is nearing hurricane intensity this morning. Luckily, there is no eye readily apparent in satellite imagery which would indicate significant strengthening but I think it is only a matter of time before Sandy becomes the 10th hurricane of the season.

Jamaica will feel the effects today and tonight followed by a landfall in Cuba early tomorrow morning. The interaction with the higher terrain of eastern Cuba will disrupt the inner core of Sandy and should keep it from being too strong once in the Bahamas. However, sea surface temps are plenty warm there and it would not be surprising to see Sandy regain hurricane intensity while passing through the Bahamas.

Late tomorrow and in to Friday, Florida’s east coast will begin to feel the effects of Sandy wit an increase in wind and surf. Right now, the upper Keys and a good portion of SE Florida is under a tropical storm watch. Since the wind field of Sandy is forecast by the global models to expand significantly, I feel that it is almost a certainty that winds to at least tropical storm force, perhaps up to 50 mph, will be felt across portions of southeast Florida.

The other issue will be the huge wave set up that is going to happen as a result of Sandy’s massive wind field. Beach erosion is likely to be a major concern for east facing beaches along the Florida coast and working up the Southeast coast in to North Carolina. I cannot emphasize this enough and with the growing Moon phase towards full, we could be looking at a major coastal flood event for some areas of the Southeast U.S. coastline. A lot will depend on how far west Sandy tracks as some of the models are indicating a brief jog back to the northwest in a few days. Interests along the Florida east coast all the way up to the North Carolina Outer Banks should be paying close attention to this situation. The chance for substantial ocean overwash, especially in the Outer Banks, seems to be increasing with time.

Then we have the issue of the ECWMF’s idea of an unprecedented impact to the Northeast with Sandy or what ever it becomes once past about 35 N latitude. The model has not given up on its forecast of a general northward track, just passing the Outer Banks and then slamming the Northeast with what looks like hurricane conditions over a large area of coastline. While the GFS remains strong in its forecast of an out-to-sea track, it has been getting a little more west and north with each run. Even if the Euro forecast turns out to be dead wrong, Sandy will leave its mark down south along the Florida east coast and probably the North Carolina Outer Banks. If the Euro is right, then we will remember the ending of the 2012 hurricane season for many years to come.

I’ll have another update posted here by early this evening.

Keep in mind that we do have our iPhone app which is a great way to keep up to date with the latest on Sandy and other tropical news and info. I post video blogs to the app each day with several of them posted daily as needed during such events as this with Sandy. During field missions, our app is the ONLY one that offers live weather data from our own instrumented wind tower. Plus, we set up live web cams and post video updates from the field on a regular basis. You simply cannot find this level of dedication and information WHERE THE ACTION IS from any other hurricane tracking app. To get HurricaneTrack for iPhone, iPod Touch and even the iPad, click here.

Share