Dorian Saturday Evening (31 August) Update

Not much has changed other than Dorian continues on a westerly heading.  It’s an impressive storm, here is a late afternoon GOES false color image …

NHC didn’t really shift their track much.  It is presently in about the middle of the guidance envelope, which takes the storm just through the northern Bahamas, then about 70 miles offshore from Florida, Georgia, and South Carolina.  Given the asymmetric nature of storms, and frictional effects of land, that just puts tropical storm force winds right on the shoreline from around Fort Pierce Florida up through Wilmington NC.

In addition to the warnings in the Bahamas, there is now a tropical storm watch up for part of the Florida shoreline.  There are evacuation orders for some residents in Brevard County (the Melbourne/Cocoa Beach/Kennedy Space Center area).  But mostly we are in the “breath holding” phase, as the guidance seems to indicate a turn close to the shore rather than a direct hit.  So for GA/SC, given the slow speed, it’s still a waiting game – but so far, still no need to take precipitous action.

In past years, a storm on this track would have already started massive evacuations.  It is a testament to the great work by tropical cyclone researchers and forecasters that we can watch a storm this powerful get this close and not freak out.  Well, you know what I mean 🙂

Dorian Update, Saturday Morning (31 August) – Why you should use NHC forecasts

As described yesterday, the main feature guiding Dorian is a ridge of high pressure just to the north of the storm.  The forecast for how strong the western side of that feature will be in 3 days is the key to the track, and if the storm will hit Florida, Georgia, SC, even stay out to sea.  Note that for the northern Bahamas, looks like you are going to get at least part of this powerful storm, so continue to follow local guidance based on the warnings in place.

Here’s the latest “spaghetti map”; the various global models, as well as the specialized models like HMON and HWRF that kept the storm striking South Florida long after the global models had shifted, have now shifted well offshore.  In fact, the NHC track (the red lines) are now well to the left of the “envelope” …

So what’s going on, and, as some have criticized, why is NHC so “slow” to change their forecast? And what about the chorus of those singing “but my <insert favorite local TV weather person or blogger> got it right!”  It’s important to realize that where the storm is most likely to go, and where you should plan on the storm going, are often not the same thing.  And as always it is vital to separate actual empirical data from emotional impressions and anecdotal evidence.

The goal of the NHC forecast is the issuance of watches and warnings so that you have enough time to prepare safely for a storm, without disrupting your normal life.  It’s a tough balance.  Once plans are initiated, they are difficult to change.  In addition, too frequent warnings can cause complacency (“crying wolf”).  Also, as any follower of this blog knows, the guidance can shift back and forth from run to run in these models.  That can cause the “windshield wiper effect” where the track jumps back and forth.  NHC wants to avoid that, so they look for trends.  Sometimes that can mean they are not the first to jump on what later does turn out to be a real trend (such as may be happening here).  But it also means that over time they are more right, and more consistent.  Just how consistent?  Here is last year’s NHC forecast verification report.  They do one every year.  What about my modeling and research?  Try this.  Or this.  Does your favorite TV meteorologist or blogger have an r-square, threat score, or mean error for their five day forecast?  Maybe they got it right this time.  But how many times have they said “the storm is going to turn” and it didn’t?  Or said it wouldn’t and it did?  Got data?  That is not to be snarky, and not to say some broadcast meteorologists are not quite good.  They do care about being right: in fact, they often care too much about not only being right, but being the first to be right.  That leads to more errors in the long run. So for something highly specialized like hurricane forecasts, you’re literally betting your life and livelihood on this stuff.  You should probably make sure it has rigorous analysis to back it up, so for planning purposes, you won’t go wrong, and have a lot less stress, if you go with the NHC forecast.

I often hear the phrase “better safe than sorry.”  Well, to be blunt, that’s foolish.  Overreacting carries a cost – not just in economic terms, but in actual human lives.  Stress aside, evacuating too early kills people in traffic and other accidents, and is especially hard on the elderly and those with medical conditions.  Surgeries are delayed, and those recovering are placed at risk from being moved prematurely.  And in our society, economic losses often translate in to lives shortened or lost.

So, where do we stand, based on the NHC forecast track and potential damage?  Here’s the obligatory map …

As for actions, the basics are right here.  As noted the ridge and steering is forecast to weaken in about three days, causing the storm to stall, then turn north.  That still brings it very close to the Florida shoreline.  People in that area should be ready to act and wary of wobbles that could greatly increase the potential damage.  Even a brush with the coast, such as the NHC track is showing, is a $6 or $7 Billion dollar event.

What about Florida, Georgia and SC?  Again, NHC probably said it best:

Life-threatening storm surge and devastating hurricane-force
winds are still possible along portions of the Florida east coast
by the early to middle part of next week, but since Dorian is
forecast to slow down and turn northward near the coast, it is too
soon to determine when or where the highest surge and winds will
occur. Residents should have their hurricane plan in place, know
if they are in a hurricane evacuation zone, and listen to advice
given by local emergency officials.


The risk of strong winds and life-threatening storm surge is
increasing along the coasts of Georgia and South Carolina during
the middle of next week. Residents in those areas should continue
to monitor the progress of Dorian.

If you are in GA/SC and have a hurricane plan, there is plenty of time to wait another day or two to watch what happens.  Impacts would be Wednesday/Thursday for Savannah on this track, which is still rather uncertain.  It’s scary to see the line, and the “Severe/Catastrophic Damage” categories near your home.  But there is really no need to freak out just yet.  If the track is still on or near the GA Coast tomorrow, then it will be time to start taking some actions, and I’m sure the local emergency management officials will start to roll out their plans.

Also as noted earlier, the GA/SC coast is having higher than normal tides.  It doesn’t look like last night peaked as high as forecast, but it was right at flood stage.  Tonight’s high tide at 10pm will be in that same ball park, as will Sunday night’s high tide.  Then it should decrease a bit Monday, but after that, it depends on where the storm goes.

Dorian Update, Friday Evening, 30 August 2019 (Ridges and Steering)

The key feature everyone is talking about is the ridge of high pressure that is guiding the storm.  Here is what it looks like, using data from the GFS model, using the 500mb height and winds.  500mb is about 18,000 feet high, and is a key level in the steering of most storms.  First, here is the current situation (as usual, click to embiggen): 

Dorian is the round bright object in the lower right corner.  Above it is a high pressure ridge, and you can see the wind barbs flowing clockwise around it.  Note the “ridge” extending from east to west.  That is what is guiding the storm right now.  What about Saturday morning?

If you look carefully you will see the barbs have more feathers, and the “ridge” extends further towards the US.  This is what is forecast to steer the storm towards The Bahamas and Florida over the weekend.  That seems pretty certain at this point.  But what happens next is in flux.  Here is the forecast situation by Tuesday.  Note the storm didn’t move all that far in three days, and note the “ridge” isn’t quite so prominent:

That’s the solution from just the main GFS model run.  How uncertain is it?  Lets look at the whole GFS ensemble family of tracks.  The blue line is the main run (that is shown above), the brown line the average of all 20 runs:

Several other models (including the much worshiped ECMWF family) are showing a similar “weakness” developing in the main runs, but the ensembles tend to be more inland.  So what does that mean?  It means what it always means: follow the NHC official track forecast since it’s their job to sort all that out!  Hopefully this gives you a better idea of the challenges of this kind of work, and why you look for trends over time, not individual model runs.  I haven’t seen the news reports on this advisory yet, so I don’t know how much of a bit deal the usual suspects are making about this, but here is a comparison of the tracks between the 11am and 5pm forecast advisories:

In a big picture sense, while the models track shift seems significant, the NHC track shift is not really that much, so nothing much has changed since this morning from an action standpoint.  People in The Bahamas and Florida need to be getting ready for a major hurricane.  Following the advice of your local Emergency Managers is probably the best bet.  This could be a very severe, prolonged event.  With this shift the modeled damage is still an eyewatering 46 Billion dollars.  Here’s the swath of doom:

For the people of Georgia and South Carolina, this may not  seem like a great evolution of the models, but certainly not freaking-out worthy yet.  If you have a hurricane plan in place, that’s all you need to do at this point.  Impacts for the Coastal Georgia area are still about 5 days away (next Wednesday or Thursday), and the potential track and intensity errors are large enough that it is a time to be vigilant, but not scared.  The long-range runs I’ve looked at have tropical storm force winds along the coast (and the coastal flooding in usual places), but otherwise not severe damage. Given the slower speed of the storm, we’re still in the “watch and be wary” phase.

Dorian Update, Friday 30 August (And a rant on models. Again.)

Bottom Line: Hurricane Dorian has started the expected turn to the west, and has the potential to strengthen before hitting The Bahamas.  As noted in the NHC Key Messages product, there is a hurricane watch up for the northern Islands, and people there should be executing their hurricane plans. People on the Central and South Florida should be getting ready for watches, warnings, and evacuation directions.  Georgia may see some effects of the decaying or remnant storm in 6-7 days, but it is way too early to say much on that.  More details at the end of this post.

The big picture hasn’t really changed all that much since yesterday, but the details have, and those can have a big impact on potential damage.  NHC has shifted the track a bit further south, and again the forecast speed has slowed down.   The forecast environment isn’t quite so favorable as was thought a couple of days ago, but Dorian will still likely be a major hurricane, it may even reach Cat 4 (130mph) before landfall. It be a dangerous, powerful storm.  Here’s the scary model track map …

(begin rant)
Once again, for planning purposes, all you need to concentrate on is the red line of the official forecast track.  All the talk about models is just stressful and confusing.  I was watching CNN last night, and like many outlets they made a big deal about  “The American Model” vs  “The European Model.”  Guess what: there is no such thing in either case. There are multiple models, and operating modes. used by the US National Centers for Environmental Prediction, NHC, US Navy,  and so forth.  Likewise, the European Center for Medium range Weather Forecasting (ECMWF) has multiple operating modes for their Integrated Forecast System, including hurricane specific modes.  And the UK Met Office has a modeling suite (and they are I suppose European until Halloween at least!).  Not to mention the Canadian Meteorological Center.  What CNN and other members of the professional and amateur chattering classes are showing you are the graphics from the main forecasting runs of the GFS (the “American” model) and the ECMWF IFS (the “European Model”).  It especially irks me that IFS is touted as being so much better than any of the US models.  Yes, IFS is good, especially in some circumstances.  But last year the “consensus” models that blend multiple models (including GFS and IFS), such as the TCVA, or the Florida State Super Ensemble (FSSE) were better.

So which “model” was the best overall? Here’s the report from last year.  THE NATIONAL HURRICANE CENTER OFFICIAL FORECAST TRACK.  So just stop it already!  If you want to use a model to illustrate some point, fine, but showing them because the graphics are prettier and animated, or to set up some artificial horse race competition, is just scare mongering.
(end rant)

As for impacts, The Bahamas and Florida look to be hit hard.  Here is the forecast impact map based on the 5am NHC Forecast …

As noted above, the northern islands of The Bahamas needs to be getting ready.  Dorian will be Andrew-like in damage to those areas.  Abaco and Grand Bahama (Freeport) are likely to be hit hard.  Currently the forecast landfall location is just north of West Palm Beach, putting the worst of the storm North of the WPB-Fort Lauderdale-Miami corridor.  Even on this track it’s a $38 Billion dollar storm. But shift the track a mere 50 miles south and it becomes well over $100 Billion dollars in direct damages.  Woah.  When you throw in economic impacts of the Labor Day weekend, if Dorian makes landfall at the expected intensity it will easily be in the top 10, and is on track to perhaps be the most expensive hurricane in US History …

In Florida the Lake Okeechobee area is often overlooked, but is also at great risk for flooding and ruptured flood control works, especially if as forecast the storm stalls.  People in that area should be especially aware of their flood zones.

Florida evacuations are going to be a goat rope. Follow the advice of your local EMA’s where possible.  Keep in mind the mantra: evacuate from water, shelter from wind.  Don’t get caught out on the road if things jam up, evacuating from wind if you live in a sturdy structure above the flood zone may actually be more dangerous than staying put.  It may not be pleasant, but better than in your car! This is an especially problematic storm in that it will put much of south and central Florida in storm conditions, with those spreading up towards the north, so to get completely out of the track would mean a journey to Alabama or North Georgia.  If in a flood zone or mobile home just get out.

What happens after day 5?  The NHC track ends at 120 hours for a reason: the uncertainties are just too great.  The global models  take the storm (or remnants thereof) across Georgia then either out to sea or up the east coast.  A warning here: there is at least one popular web site that has been showing the 850mb maps from GFS and/or IFS.  These look scary for Georgia, but the SURFACE winds were only 20 or 25 mph.  I suppose I’ll have to rant about that later.  So the bottom line in Georgia is what it was the last two or three days: we’re still more than 5 days from seeing anything from the storm, so best to just watch and keep out of the way of folks who need to move.

Hurricane Dorian, Thursday Evening 29 August 2019 update

As the sun sets over the storm, Dorian is now well away from land, over the Atlantic east of The Bahamas (note you can click any of these images to to “embiggen”) …

The 5pm NHC forecast slowed down the approach a bit later in the forecast, and shifted slightly south, as has much of the dynamic guidance.  But that hasn’t changed the big picture; those are just minor tweaks:  we are looking at a potentially severe hurricane making landfall on the Central Florida coast in about 4 days.  Here’s the impact forecast from my Haetta/TC model, based on the Official NHC track and intensity estimate …

Nothing has really changed since this morning as to what all this means.  Some models are now showing a “hook” to the north after landfall, and at least one source was showing a very misleading upper-air (850mb) graphic from the GFS model showing the storm over coastal Georgia.  The surface winds in that case were a blistering 20 mph.

Will check in tomorrow morning …

A few thoughts about waiting for storms

First, nothing substantially changed about Dorian since this morning’s review.

I see there are a lot of new people around (Facebook is now over 15 thousand followers, which is scary/surprising, you people need a life!), so a few thoughts about real time storm events and this blog for those new to this.

Originally I had a blog for business purposes,and as part of research outreach.  It actually got started as part of a NOAA grant I was working on with the University of Central Florida in 2004-5 or so, trying to figure out better ways of conveying storm risk to the unwashed masses.  Those maps that you see here labeled with “plain english” categories like “branches breaking” or “severe damage” are one of the results of that project.  After that grant ended I let the blog lapse, but the real time discussions got started because friends and acquaintances were always asking what I thought, and I got tired of repeating myself so I started posting again. In truth, I’m disappointed so many people say this is the most calm and reliable info they have found, so I guess I’m stuck with it.

I don’t make any money off of this; you will see no blipverts here, any you see on FB or Twitter aren’t mine and have nothing to do with Enki.  It annoys me when people think I do, or am doing this for some self-aggrandizing reason.  You note Enki doesn’t even have my name on it.  Because that’s not important.  So, about storms.

My biggest piece of advise is don’t get frantic, and don’t feel you have to chase every little wobble or shift in some model or another to be informed.  As discussed in just about every post, don’t worry about individual track models.  Unless you work at the National Hurricane Center or one of the other major centers, or are doing research in the area of tropical cyclone meteorology, to be blunt, you don’t have the background to interpret them (and chances are neither does the TV meteorologist showing you that fancy animation).  If you notice, when I do discuss an individual track model, it’s virtually always in the context of the official forecast, or as a whole to illustrate a point.  Unless you have some specialized need for more warning (in which case the TV/Media speculation isn’t going to help you), for planning purposes all you really need are the NHC official forecasts.  As those who have followed for a while know I like their Key Messages product, You’re not going to do better.

So how often should you try to get updates?  Unfortunately, the TV/”news”/social media beast wants to be fed, and fed often.  But that’s counterproductive for hurricanes.  Things just don’t change that fast and when they do, if you’ve been following all along, you’ll have plenty of time to react.  Blown forecasts or genuine surprises are really rare – but people tend not to think that because “the beast” has to play up minor differences and “the latest tracks” to keep you engaged.  Unfortunately, that is stressful.  Don’t play that game.

Generally speaking, checking twice a day, in the morning and after school or work in the evening, are all you need in most cases.  Check the NHC forecast, check to see if watches and warnings are up for your area, check your local Emergency Management Agency or local media to see what actions are being recommended (try to set aside all the noise over what if’s).  Then decide what you are going to do, based on the PLAN YOU DEVELOPED BEFORE HURRICANE SEASON (or right now this minute drop everything, if you live in a hurricane prone area and haven’t yet)!   The basic rule for hurricanes is “evacuate from water, shelter from wind” so know if you are in a flood zone (don’t forget rainfall induced flooding of creeks, rivers, and stuff that becomes a creek or river if you get enough rain!)  A lot of communities also evacuate for wind.  I have issues with that for other than very high winds because it vastly increases the disruption (and risk) for the majority of people, but I certainly understand not wanting to be out of power for two or three weeks (months if you live in Puerto Rico, but that’s a different rant).  Of course, the calculations are a bit different for people with health problems.  But for 95% of people and businesses, that’s it.  The five days notice you get, assuming you have a plan, should be enough.  If they aren’t, the problem is the forecasts uncertainty explodes beyond that (which is why NHC only goes to five days) and the “downside” of disruption outweighs any benefits of early action.

I hope that helps.  Note that I generally only post twice a day, early (6-7am), and late afternoon (4-6 pm) unless something really  changes.  Also, it has reached the point where I get so many messages and requests for specific guidance that I just can’t answer them all.  Sorry about that!  Sometimes people ask if they can contribute to keep the site going.   I really appreciate that! Even though it’s not for this, I’ve got good funding at this point (even if it, like most research funding, isn’t always as stable as I’d like it!) so I’m fine.  If you want to help please contribute to a Caribbean relief effort.  I spent a lot of time working with Caribbean governments and agencies in the 90’s and 2000’s, and while I can’t stand the sun in the tropics (I burst in to flames) I love the region.  If not that, find an effort to help the less fortunate here in the US, or an animal rescue group.  I have two feral/rescue cats in the office you may see in pics from time to time … so their colleagues always appreciate a helping hand.

Dorian Update for Thursday Morning, 29 August 2019

Dorian is now well past Puerto Rico and the Virgin Islands. Fortunately damage seems fairly light, mostly power outages, scattered roof damage, and some trees down.  Here’s a farewell radar scan from the San Juan International Airport …

Dorian didn’t intensify much overnight, but is entering an area where conditions are more favorable.  The primary track and intensity models are in agreement for landfall somewhere on the Florida coastline between Miami and Jacksonville.  I’ve heard some snarky comments about that, but it is an amazing improvement in our ability to forecast hurricanes that the band is that narrow four or five days out.  In this dynamic situation, even 10 or 15 years ago the uncertainty band would have been from Miami to New Jersey.  Here’s the scary confusing “spaghetti model map” …

… honestly, these kinds of maps are pretty useless like this so let’s peel back the layers and see what is really going on. If you are in a hurry you can skip down to “The Bottom Line” below.  Many of these lines are what are known as ensemble members, secondary models, or model runs based on older data.  Here are just the current/recent “primary” track models:

Not so bad, is it?  The brown line taking off to the north is the “CLIPER” model.  It is a purely statistical model that shows us where historical storms that were near where Dorian is now, and moving about the same direction, ended up going.  It doesn’t know anything about the weather now, but is a good indicator that in about a day or so Dorian is expected to do something “unusual.”  The faint gray line, XTRP, is what would happen if the storm just kept going in a straight line.  So what about the GFS (the dodger blue line), which has been causing some FUD (Fear, Uncertainty, and Doubt) in Georgia.  Let’s look at just the GFS family of model runs.  Again, the blue line is the primary run, the cloud of thin gray lines are twenty (20) “ensemble members” – simulations that started in slightly different positions, and with slightly different initial conditions.  The brown line is the average of these scenarios.   Note that is right back in the central/south Florida area where the other models are showing landfall …

The Bottom Line

So what should you do?  Well, if you are not a tropical cyclone expert there is only one line you need to care about, the National Hurricane Center official forecast track.  They go through the above exercise every six hours, using their years of experience to blend all the model data along with real time observations and aircraft data to create a forecast. Here is the current (5am ET Thursday) forecast track, along with the estimated impacts based on my Haetta/TC model using their track and intensity …

The storm track forecast has slowed down a little.  People in the northern Bahamas need to start preparing for a major hurricane.  It’s maybe three days out from the initial impacts, and it looks to be bad.  For the Florida coastline from West Palm to Jacksonville, and inland areas in the post-landfall path, it’s time to start getting serious.  Given the holiday weekend, expect evacuation guidance and other plans to kick in to high gear a bit early, potentially later today or tomorrow.  Remember the key life saving mantra: evacuate from water, shelter from wind.

If you had plans to go to Florida for labor day, I hate to say this but better luck next year.  Airlines are already realigning their flights.  Best to stay away.

For Georgia, it’s still watch and wait.  For now it looks like the bad impacts will be south of us, although the far south Georgia (Brunswick and St. Marys) need to be more ready for the fringe impacts and, if things go squirrely, take action.  I’m hearing that SCAD is recommending “voluntary evacuation” Saturday.  That’s not helpful.  At this time it looks like impacts to the Savannah Georgia area will be minimal.

On an unrelated but related note, there are some concerns about coastal flooding along the Georgia/SC coast because of high astronomical tides.  The forecast at Fort Pulaski was already close to major flood levels (which means US 80 closed), and we will likely see that tonight at the 8pm high tide and the overnight tides on Saturday, Sunday, and Monday.  You can see the tide forecast here.  People who live on the Georgia Coast and it floods with spring tides can expect that.  I wouldn’t expect Matthew or Irma kinds of flooding, as Dorian is a small storm, but we could well have a couple days of onshore winds so that with the high astronomical tides means at least a foot or two above the normal higher high tide levels.

Hurricane Dorian update, Wednesday Evening, 28 August 2019

Although there are still some strong cells moving by, the center of Dorian has moved past the Virgin Islands, and is now a hurricane.  It even looks like one; here is a radar scan from the San Juan Airport Terminal radar from about 4:30pm, and a a recent GOES satellite image(click to see full size):

Not that much has changed from a dynamical track model standpoint since earlier today.  They jitter around a bit from run to run, but most still change the “steering” for the storm to take it due west in to Florida this weekend.  What has changed is that because Dorian avoided both Hispaniola and Puerto Rico, it enters the very warm waters and favorable atmosphere around The Bahamas poised to intensify into a major hurricane. Here is the latest (5pm Wednesday 28 August) official forecast track.  It has shifted a bit south in anticipation of the stronger ridge.  The biggest change from earlier today is they are now fully expecting landfall to be as a major hurricane:

Are there other scenarios?  Yes, it could just keep on the more northward turn, and end up out towards Bermuda if the expected steering doesn’t develop, or if it isn’t as strong as expected, end up striking the Georgia, South Carolina, or North Carolina coast.  But most of the guidance says the NHC scenario is most likely.  That said, it is a bit of a dicey situation because we won’t know for at least another 24-36 hours if the turn is going to develop, or how sharp, or if it won’t turn north again later.  Sorry about the uncertainty, but that’s just how it is with this storm.  However, there is still plenty of time to react no matter what happens.

I’m often asked when it is time to take a storm seriously.  That will probably be tomorrow if you are in Florida or The Bahamas.  In the southern Bahamas and Turks and Caicos, the storm is expected to stay to your east.  In the Northern Bahamas and Florida, pay attention to NHC for watch and warnings, and your local emergency managers for evacuation planning, as they will likely be coming soon.  Make sure you know your shelters, evacuation routes and zones.  If you didn’t have a plan already (seriously, you’re in FLORIDA and don’t have a hurricane plan???) then visit the FEMA site for some ideas and get it together.  It looks like a bad Labor Day might be in store.

Further up the coast in GA/SC/NC, no need to freak out yet and break the refresh key on your computer (or these days, smudge up your smartphone and exhaust your data plan), but check in Friday with your favorite source of reliable information (which should be the NHC “Key Messages” product) and see what’s going on.  There is still plenty of time to watch and see where this thing is going, but reviewing your plans and making sure you know what to do if the storm doesn’t turn enough (either way!) is smart.  If it makes landfall as expected, there is some concern for coastal flooding as we are approaching a new moon and higher than normal tides anyway.  The onshore winds will exacerbate that – but still a bit early to tell on that aspect, and the further shift to the south helps..



Dorian Doomwatch, Wednesday Morning 28 August 2019

Dorian continues to defy the odds to an extent, and is not only holding together but tracking to the right of the forecast tracks.  In part because of the small size, Dorian is not “entraining” dry air as a normal-sized storm would, so is holding together and even intensifying more than perhaps it should.  As always the two questions are where and how bad.

Here’s this morning’s “primary” track models.  GFS main has trended north towards the SEUS coast over the last 12 hours, indicating a ridge of high pressure will have a “weakness” the will allow the storm to turn.  The ever sainted ECMWF and the Navy’s model keep it strong, and Dorian headed to Central Florida …

There are lots of local (Savannah GA area) amateurs, semi-pros, and attention … um, seekers … talking a lot about GFS (probably because it’s scary for people on the Georgia coast and gets them more attention.  Let’s take a closer look at that.  Here’s the GFS family of tracks available as of this morning …

The blue line is the primary model that is often used to generate scary graphics.At the moment it shows the storm brushing Hatteras.  Last night it showed Dorian right over Savannah.  6 hours before that it was over central Florida.  Get the point?  It is too unstable right now to get excited about.  Note the brown line, and the cloud of gray lines.  That’s the “consensus” of the GFS ensemble model, based on multiple initial conditions.  and variations.  Again, very wide spread.  Also note that NONE of these got the last few hours right – Dorian is well to the right of almost all of these already.

That’s the background.  What should you do?   Simple: ignore all the noise and the amateurs and even “pros” throwing out this or that scenario based on one model track that happens to be hitting your house.  For planning purposes there is only one place you need to go: the National Hurricane Center’s “Key Messages” page. And be careful about using that “cone of uncertainty” – NHC has a video on it talking about what it does and doesn’t mean.

As for the estimated impacts, here’s the forecasted impacts in plain English, from my Haetta/TC based model using the NHC forecast track as of 6am …

Bottom lines: In Puerto Rico you should have already planned for a Cat 1 Hurricane.  Sadly, on this track and intensity, you’re going to lose power again, and the system is still pretty fragile so it will be out for a while.  The biggest physical damage risk to PR is probably flash flooding; the worst of the winds will likely be over the ignored and  neglected US Virgin Islands.  Estimates are if the intensity follows the NHC trend, damages will be on the order of $400 to $500 Million, but as noted yesterday, the humanitarian impacts are likely to be disproportionate given the still fragile infrastructure.

For Florida, the next 24 hours or so will start to tell the tale, and if folks should begin to freak out or not.  (You, of course, have a plan and don’t have to panic until maybe Friday when the track will be much clearer).

For Georgia, South Carolina, and North Carolina, “Calm down, eat some fruit or something.”  As discussed above, if the ridge builds, Florida gets whacked.  If it decays a little, you might be in play.  If it decays a lot, it will stay offshore (think 1999’s Floyd).  We just don’t know which scenario will happen, and you still have time (assuming you have that plan) to plan your end-of-summer barbecue before you plan to flee.

Dorian jumps sideways …

As happens sometimes with weaker storms in hostile environments, the center of circulation of Dorian reformed north of its previous track/position after crossing St. Lucia.  Take a look at this comparison; the brown line is the 11am forecast, the back line the actual track, the red line the new forecast:

This means two things: first, a direct hit on Puerto Rico is now much more likely.  Second, it might be a bit weaker, but because of the direct hit, given the small size of the storm, a bit more impacts are likely.  Third, (ok, three things) it will be weaker entering The Bahamas.  Power outages and flash flooding are likely in store in Puerto Rico.  Not good.  Economic impacts are “only” forecast in the mid hundreds of millions, but that is partly because economic activity is still depressed.  Real impacts in the form of suffering, even for a weaker storm, are likely to be higher depending on how well preparations are going.  On the optimistic side, I’m hoping NHC is being too generous with the storm, and the dry air will keep it more suppressed that they are showing.

Track guidance is fairly tight for the dynamical models, but the intensity guidance is pretty split.  NHC is forecasting a recovery of the storm to near hurricane strength before landfall in mid-Florida this Labor Day weekend.  That’s not good for all the tourist businesses, so economic impacts for even a weak storm could be pretty dramatic.  Will have a better picture of that Thursday.  Meanwhile, folks in Florida who have a hurricane plan can still just watch and wait – and hope things go well in Puerto Rico; you’ll have time to figure out what to do later in the week.  Here’s the latest impact map based on the 5pm NHC forecast.  

For narcissists in GA/SC/NC who just have to know what this means for them, if you have a hurricane plan nothing to do or worry about at this time. The reformation to the north doesn’t mean it is more likely to hit you.   More than likely it will stay south and be guided into Florida and the Gulf by the mid and upper level winds.  Again, if it does start to do something wonky and head towards the SE coast, we’ll have plenty of warning.