It’s been a whirlwind 24 to 36 hours for Tropical Storm Nate’s track forecast. It’s gone from likely to hit Mexico … to a sudden shift toward threatening to the U.S. Gulf coast … back to likely, and now very likely, to stay away from the U.S. and hit Mexico (if anywhere). From the 10pm NHC discussion:
THERE HAS BEEN A LARGE MODEL CHANGE FROM SIX HOURS AGO…WITH THE GFS/GFDL/HWRF MODELS NOW HAVING A GENERAL WESTWARD TRACK TOWARD MAINLAND MEXICO DUE TO THE CYCLONE BEING TRAPPED BENEATH A NARROW RIDGE OVER THE NORTHERN GULF OF MEXICO. NATE IS FARTHER SOUTH THAN EXPECTED YESTERDAY…SO THE LIKELIHOOD OF A MORE POLEWARD TRACK AHEAD OF THE NEXT TROUGH OVER THE CENTRAL UNITED STATES HAS DIMINISHED. THUS THE NHC FORECAST IS SHIFTED WELL TO THE LEFT AT THIS TIME…BUT BASICALLY ALL RELIABLE MODELS ARE STILL FARTHER TO THE SOUTH. IF CURRENT TRENDS CONTINUE…THE FORECAST WOULD HAVE TO BE ADJUSTED SOUTHWESTWARD AT A LATER TIME.
So basically, the NHC is forecasting Nate to come nowhere near the U.S., and it’s likely the forecast will shift even further away from the U.S. tomorrow, if current trends hold. And so far, they’re holding: the all-important 00Z GFS run, which came out after the 11pm NHC advisory, shows basically the same thing as the 18Z run, with Nate staying to the south and, actually, pretty much falling apart just as it’s about to come ashore in Mexico. (It’s frankly not clear to me why.)
Since this blog’s charge is to track major tropical-cyclone threats to the U.S. coast, and it no longer appears that any of the active storms — Katia, Maria or Nate — pose such a threat, I’m going to call off my updates for now.
If the situation changes and Nate poses a threat again, I’ll be back. In the mean time, check my Twitter feed for the latest — on hurricanes, and college football, and conference realignment, and politics, and whatever else I feel like tweeting about.
No time for a lengthy post, but very briefly: I wrote this morning that “we may be seeing” a turning point whereby the computer models decisively change course and move Nate north toward the Gulf coast. Well…never mind! Most of the model tracks have now shifted back south again — a healthy reminder that the Hurricane Center’s tendency to wait for confirmation before adopting a new computer-model consensus is well grounded in long experience and common sense. That “ping-pong effect” I mentioned can and does really happen!
AFTER 36-48 HOURS…THERE REMAINS A VERY LARGE SPREAD IN THE GUIDANCE…HOWEVER THE MODELS HAVE GENERALLY SHIFTED WESTWARD. THE GFS IS ALONG THE EASTERN EDGE OF THE ENVELOPE AND SHOWS A SLOW NORTHWARD TO NORTH-NORTHWESTWARD MOTION OVER THE CENTRAL GULF OF MEXICO. MOST OF THE REMAINDER OF THE TRACK MODELS MOVE NATE MORE WESTWARD TOWARD MEXICO. THE OFFICIAL FORECAST HAS BEEN SHIFTED TO THE LEFT…BUT NOT NEARLY AS FAR WEST AS THE MULTI-MODEL CONSENSUS…OUT OF RESPECT FOR THE GFS AND THE EARLIER ECMWF ENSEMBLE MEMBERS.
Bottom line, it’s once again more likely that Nate will hit Mexico than the United States. But the U.S. Gulf coast isn’t yet out of the woods. That pesky outlier, the GFS, is one of the most reliable models on earth, and for now, it’s still projecting a more northerly motion, toward Texas/Louisiana. Its 18z run is due out around 6:45 PM Eastern Time, so look for that; you can view it here. And all the 00z model runs will be important, for reasons noted earlier. (Links and schedule here.) Stay tuned, as they say. Hopefully, forecasters will have a better handle on Nate by tomorrow morning.
Oh, and by the way, Nate is strengthening. 70 mph. He’ll be a hurricane soon.
Sometimes, in the life of a tropical cyclone, there is a distinct turning point in the evolution of the track forecast, where the computer models suddenly shift toward a new and markedly different solution. When this happens, the National Hurricane Center — wanting to avoid a “ping-pong effect” of models shifting back and forth, as can happen during periods of forecast uncertainty — generally waits to see a few more model “runs” confirming the new thinking before it fully adopts the new consensus in its official forecast. This memorably happened in 2005 with Hurricane Katrina, when the models shifted early Friday morning toward a solution targeting the New Orleans area, inspiring my “New Orleans in peril” post at midday Friday (which I would later read aloud in Spike Lee’s movie), published eight hours before the NHC actually adopted the track that the models had begun shifting toward almost 24 hours before. More recently, something similar (albeit less drastic) happened with Hurricane Irene, when the models’ eastward trend suddenly stopped, and the forecast began to settle on a track targeting New Jersey/NYC/Long Island.
We may be seeing a similar turning point with Tropical Storm Nate this morning. Yesterday, this storm looked like it was bound for Mexico. But things have changed overnight. This morning, shortly before the National Hurricane Center’s 11am EDT advisory, CharlesFenwicktweeted: “In the Gulf of Mexico, there was a good consensus on Nate heading to Tampico, Mexico. This morning, though, GFDL & HWRF shifted northeast with solutions south of Louisiana and in the middle of the western Gulf of Mexico respectively.”
Up until last night’s 8 pm EDT runs of the computer models, the models were in general agreement that Nate would meander in the Bay of Campeche for several days, until a ridge of high pressure built in to the north of the storm, forcing it westwards to a landfall in Mexico. However, the latest 2 am EDT run by the GFS model predicts that Nate may gain enough latitude to escape being forced westwards by the ridge, and instead move northwards to make a landfall along the U.S. Gulf Coast. The GFDL, which uses the GFS for its initial conditions, is also on board with this idea, as is the HWRF model, to a lesser degree. The 2 am EDT run of the NOGAPS model did not go along with this idea, though. We will have to wait until the NOAA jet makes its first mission to sample the steering currents in the Gulf of Mexico to get a better idea on how probable this northern path might be; their first flight will be tonight, and the data will make it into the 8 pm models runs that will be available first thing Friday morning.
Just as it looked like the overnight models were converging on one solution for the track of Tropical Storm Nate, this morning’s models made an almost wholesale change.
Instead of an eventual westerly motion into Mexico, a number of forecast models have switched back to a more northerly movement by late this weekend, bringing Nate toward the northern Gulf of Mexico coast by late Sunday or Monday.
It will be most interesting to watch the newest global model runs, which should come out between noon and 2 p.m. today. I’ll especially be looking to see if the European model switches back to the more northerly solution. If you’ll recall, until yesterday this had been the Euro model’s preferred track. …
Critically, if Nate does move northward it will cross over very warm Gulf waters, and would have several days to intensify. This raises the possibility of a major hurricane striking the New Orleans region.
It still appears Texas will not be directly affected by strong winds and rain from Nate, but this is indeed proving to be a difficult storm to nail down.
Fenwick added: “NHC will probably wait for 12Z models before making any drastic changes to the forecasts, so 11AM forecast package shouldn’t have big track changes.”
AFTER 36-48 HOURS…THE MODEL SPREAD BECOMES INCREASINGLY LARGE…AS SUBTLE DIFFERENCES IN THE STRENGTH AND ORIENTATION OF A MID/UPPER-LEVEL LOW AND TROUGH EXTENDING FROM THE OHIO VALLEY SOUTHWESTWARD HAVE SIGNIFICANT IMPLICATIONS ON THE FUTURE TRACK OF THE CYCLONE. THE LATEST GFS AND GFDL MODELS…WHICH SHOW A SLIGHTLY DEEPER TROUGH…MOVE NATE NORTHWARD TOWARD THE NORTHERN GULF COAST. THE ECMWF AND NOGAPS MODELS DEPICT A WEAKER TROUGH THAT DOES NOT CAPTURE NATE AND ALLOWS IT TURN TURN WESTWARD TOWARD MEXICO TO THE SOUTH OF A BUILDING RIDGE…ALTHOUGH IT SHOULD BE NOTED THAT THERE IS CONSIDERABLE SPREAD AMONG THE ECMWF ENSEMBLE MEMBERS. THE OFFICIAL FORECAST HAS BEEN SHIFTED TO THE RIGHT AND LIES BETWEEN THESE TWO VIABLE SCENARIOS BUT HAS NOT BEEN SHIFTED AS FAR AS THE TVCA CONSENSUS. THE NOAA G-IV AIRCRAFT IS SCHEDULED TO PERFORM A SYNOPTIC SURVEILLANCE MISSION LATER TODAY…WHICH WILL HOPEFULLY HELP REDUCE THE SPREAD IN THE TRACK GUIDANCE TONIGHT.
Today would be a good day to follow my schedule and watch as the new computer models come out. Already, the latest GFS model run has seen the track details shift a bit, though the basic concept — a more northerly track, threatening the Gulf coast — remains. As Fenwickexplains: “the 12Z GFS…has split the difference between its last 2 solutions & has Nate hanging out in the northwest Gulf of Mexico through day 7.” Here’s what the 7-day model forecast position looks like, with Nate on the left and Maria on the right:
Maria, incidentally, appears likely to recurve harmlessly out to sea, much like Katia (though that could change). It’s Nate we need to be watching.
To be clear, there is no reason to panic about possible threats to specific locales. The forecast can and will change in the coming hours and days. But with the Gulf waters being so warm, and so many vulnerable spots along the coast, this is definitely a situation to watch closely.
Given the favorable environment, in addition to a very warm pocket of sea surface waters in the central Gulf of Mexico, we expect that Nate will intensify modestly over the next few days. The National Hurricane Center forecasts that Nate will become the third hurricane of the season by Friday. The HWRF and ECMWF agree with this forecast–both of these models bring Nate to a category 2 hurricane by Saturday. The IVCN/ICON consensus models that the Hurricane Center relies on are more conservative, peaking at category 1 intensity. Nate’s maximum potential intensity is heavily dependent on its track, which, according to the weather models, has been up in the air for the past few of days. Until this afternoon’s run, the ECMWF has held true to its forecast that Nate will track north and make landfall anywhere from Louisiana to Florida. This afternoon, it backed off of that solution and is now forecasting a northern Mexico landfall. The GFS has consistently forecasted a track that lingers in the Bay of Campeche for a few days before ultimately making dive to the west into Mexico. Over the past few days the Canadian CMC model has been reluctant to develop Nate at all, but today is forecasting the system to track north into the Southeast U.S. states. Now that there is Hurricane Hunter data to ingest (as well as confirmed 45 mph surface wind speeds), we expect the models will come into better agreement on both track and intensity for Tropical Storm Nate.
The National Hurricane Center’s 11pm EDT discussion emphasizes the gradualness of the intensification:
NATE IS FORECAST TO REMAIN IN AN AREA OF LIGHT VERTICAL WIND SHEAR FOR THE NEXT 36-48 HOURS…AND THE INTENSITY GUIDANCE RESPONDS TO THIS LOW SHEAR AND THE WARM UNDERLYING WATERS BY FORECASTING STRENGTHENING. HOWEVER…THERE IS ABUNDANT DRY AIR TO THE NORTHWEST OF NATE WHICH COULD SLOW THE RATE OF STRENGTHENING IF THE STORM INGESTS IT…AND THE CURRENT STORM STRUCTURE IS NOT FAVORABLE FOR RAPID STRENGTHENING. BASED ON THIS…THE INTENSITY FORECAST SHOWS GRADUAL STRENGTHENING INTO A HURRICANE DURING THAT TIME.
You can see the “abundant dry air” in the water vapor image at the top of this post. It’s the orange part.
After 48 hours, there is “uncertainty” and divergence among the models about how strong Nate will get. One reason the intensity depends in part on the track is that the waters have more tropical cyclone heat potential a bit further north. Also, the longer Nate stays in one place, the more it will weaken the water directly underneath it through “upwelling” — its winds and surf bringing up colder water from just beneath the surface.
Anyway, again, with regard to Nate’s track, the general forecast thinking is a path toward the Mexican coast, but there is considerable uncertainty, and the Canadian model taking Nate toward the U.S. Gulf Coast remains, albeit as an outlier. Here’s a look, courtesy of Ryan Maue’s site, at what two models are forecasting in 4 1/2 days’ time: first the HWRF, then the Canadian model.
That’s Nate at left, Maria at right in the bottom (Canadian) image.
Updates tomorrow (Thursday) will likely be limited, but you can get the latest from my Twitter feed, and also track down the relevant data yourself, if you so desire, using the links from my “day in the life” post.
The Atlantic tropics have been busy since Irene’s demise 10 days ago, what with Hurricane Katia, Tropical Storm Lee, Tropical Storm Maria, and now, as of a few hours ago, Tropical Storm Nate.
Lee dumped copious rain in the Southeast, and now its remnants are causing further flooding in the already saturated Northeast. Katia, which at one point reached Category 4 intensity, threatened for a while to be Irene redux, but now the computer models are extremely confident it will “recurve” safely off to sea. Maria looks likely to remain weak for a while, but could eventually strengthen and become a threat to the East Coast — but it’s way too early to tell, and any possible impact would occur no earlier than mid to late next week.
Nate, however, could become a big deal, and fast. The Hurricane Center named the storm, presently located in the Bay of Campeche / southern Gulf of Mexico, at 5pm EDT today, estimating its winds “conservatively” at 45 mph — and there are signs its intensity could ramp up quickly. Florida State meteorologist Ryan Maue, whose Twitter feed and website are excellent sources for computer model information, tweeted this afternoon, “With the curved cloud band of Nate–I’d expect explosive intensification. Wouldn’t be surprised to see hurricane in < 24 hours. Major < 72?" The waters of the Gulf of Mexico, of course, are very warm with a lot of cyclone heat potential, and there’s virtually no wind shear nearby to weaken Nate. So — with the caveat that, as Irene reminded us, intensity forecasting is a very inexact science — conditions certainly look ripe for significant and perhaps rapid strengthening.
Nate’s location is the other reason for concern. It could potentially pose a threat to anywhere from the Mexican east coast to Florida. The computer models just aren’t sure yet, and the steering currents are weak. A slow track toward Mexico presently seems the most likely, as the official forecast indicates:
In light of the generally sparse population along much of the Mexican coastline, that’s preferable to some previous computer model scenarios like a direct, quick strike on New Orleans, a Florida panhandle landfall, or a Mobile hurricane. But, again, it’s just too soon to tell for sure what will happen. Hopefully the models will get a better handle on Nate within the next 24 hours, now that it’s gone from being a hypothetical storm to an actual existing one.
Another big question is what impact Nate might have on the disastrous Texas wildfires. The recent blazes were actually fanned by the winds of Tropical Storm Lee and the cold front that swept in behind it — a cold front whose southern portion then stalled over the Gulf, developed tropical characteristics, and turned into…Tropical Storm Nate. So could Nate, partially a product of the weather system that helped fuel the fires, now help put them out, by bringing tropical rains to Texas?
Unfortunately, that isn’t likely, says Dr. Jeff Masters: “None of the models is hinting at a track towards Texas, and the intense dome of high pressure associated with their record drought and heat wave will tend to discourage any tropical cyclones from making a Texas landfall over the coming seven days.” The Houston Chronicle‘s “SciGuy,” Eric Berger, agrees: “It does appear likely Texas, due to building high pressure over the state, will not see significant rainfall from Nate, especially areas outside of lower Texas.”
Speaking only for myself here, it seems like what Texas might actually need to “root” for is Nate becoming a fairly powerful hurricane — powerful enough to “make its own atmosphere,” and maybe, just maybe, push back a bit against that intense high-pressure dome. But of course, that’s a catch-22. What Texas needs is a tropical storm with soaking rains, not a disastrous major hurricane.
Wherever Nate goes, its formation is historic, or nearly so. Nate is the second-earliest forming “N” storm ever, trailing only the “N” storm of the record-shattering 2005 season — coincidentally also named “Nate,” thanks to the six-year rotating storm name list — which formed just one day earlier. The 2011 season hasn’t seen nearly as many intense hurricanes as we saw in 2005, nor as many devastating landfalling hurricanes. But in terms of sheer numbers, it’s keeping pace thus far, which is amazing. The 2005 season, you may recall, was the first ever to exhaust all 21 listed names (Q, U, X, Y and Z are skipped) and send us into the Greek alphabet, with the season’s final storm, “Zeta” (the sixth Greek letter, and 27th named storm), memorably forming on my wedding day, December 30, and not dying until January 6 of the following year (during my honeymoon). And it turns out there was actually another storm, making the total 28 — Zeta should have been Eta.
Anyway, the point is, we’re presently on 2005′s pace, or within a day of it, which is incredible; I never thought I’d see another season approaching 2005. Who knows if it will continue — probably not. But 2011 is certainly turning out to me one heck of a year. And “Ophelia” may be coming soon to a hurricane tracking map near you.
I’ll try to post an update on Nate (and the other storms) later tonight and/or tomorrow morning. In the mean time, you can get the latest — mixed in with a lot of nonsense about college football, conference realignment, tonight’s GOP presidential debate, and so forth — from my Twitter feed.
I’m not back in hurricane-tracking mode here on Weather Nerd — Katia, proto-Lee, and far off hypothetical proto-Maria aren’t really threats just yet — but for those who’d like to do some do-it-yourself storm-watching, I thought I’d make a schedule of a typical day of hurricane forecasting developments, with links to the data sources that update around the listed times.
First, a bit of a glossary, and some background. “NHC” is the National Hurricane Center, and their advisories usually come out a few minutes before the official top-of-the-hour timestamp. (So, for instance, the 5pm advisory comes out around 4:45 or 4:50.) “Full” advisories include the all-important meteorological “discussions” and updated 5-day forecasts. “Intermediate” advisories update the current conditions, but don’t generally add much else, and they’re only issued when watches or warnings are up for some coastline somewhere.
“NAM,” “GFS,” “NOGAPS,” “CMC (Canadian),” “HWRF,” “GFDL” and “ECMWF (European)” are seven of the computer models that NHC looks at, with “runs” either two or four times a day. Loosely speaking, the runs are identified by the time that they start (in “Zulu” or Greenwich Mean Time, e.g., “the 00Z run,” “the 06Z run,” etc.), but they take differing amounts of time to complete. So I’ve listed the approximate times that each model usually spits out new forecast maps for us to look at.
The GFS and ECMWF global models are the most accurate models; NAM, NOGAPS and CMC are the least accurate of these seven. HWRF and GFDL are somewhere in the middle. You can read more here about the various computer models, from Dr. Jeff Masters.
In all cases except HWRF and GFDL, the links below go to the individual, model-specific graphic for the latest “run,” from Ryan Maue’s site. (To check if the model has updated, look at the top middle of the linked image where it says something like “MSLP (hPa) 06Z01SEP2011.” Those first three characters of the timestamp, “06Z” in this example, tell you which model “run” you’re seeing.) For the HWRF and GFDL, whose URLs are storm-specific, my links simply go to Maue’s page; those models at the top of his left-hand column.
Keep in mind that computer models can shift wildly from run to run, and are subject to huge errors, especially in their longer-term forecasts. Do not make any life-or-death decisions based on a computer model forecast! If you’re confused by what the computer models show, you’re probably better off ignoring them, and trusting the NHC. Over an entire season, the official forecast generally outperforms any specific model.
That said, for those of us “weather nerds” who can’t help ourselves, the model maps are like, well, #stormporn. So, without further ado, here’s the schedule. All times Eastern Daylight Time. I’ve tried to use font size and weight to give a general idea “how important” each event is.
Of course, reconaissance data is also important, but the recon flights’ schedule varies from day to day, and the data can be a bit difficult to decode. But you’re certainly welcome to try your hand at it here. Unless you’re pretty technically minded, though, it’s probably better to wait for the NHC advisories to explain what the planes found.
Via the Catskills region blog The Watershed Post, here’s another incredible video of the epic river flooding caused by Hurricane Irene’s torrential rains:
Dr. Jeff Masters has more on the catastrophic flooding, including some nifty charts showing various rivers’ flood-stage situation. You can find more river flood gauges here. Great stuff from the National Weather Service. Also from the NWS, the final advisory on Irene from the Hydrometeorological Prediction Center includes a ton of eye-catching numbers, with more than 30 locations reporting double-digit rainfall totals. The highest figure is an incredible 20.40 inches in Virginia Beach, Virginia.
With that, I’ll be signing off this blog until the next hurricane threatens, or events otherwise warrant. My role here as Pajamas Media’s “Weather Nerd” is to provide updates on active, significant tropical cyclones threatening the United States, so Irene’s official demise last night is the blog’s cue to go dormant again for a while. Perhaps I’ll be back in a week or so to talk about Hurricane Katia, presently known as Tropical Depression 12, which formed this morning off the African coast and is expected to steadily strengthen as it churns west over open waters. But then again, perhaps not: there’s an excellent chance “proto-Katia” won’t ever threaten the U.S., per Dr. Masters: “Forecast tracks from the long-range GFS and ECMWF models suggest that Bermuda and Canada might be the only land area threatened by TD 12, but it is too early to be confident of this.”
Anyway, until the next storm threatens, you can follow me on Twitter (though, be warned, my timeline will soon have more talk of college football than hurricanes) or on my personal blog, The Living Room Times.
P.S. The Red Cross is conducting relief efforts in the areas hit hard by Irene. You can donate here. For all the talk about “overhype,” this is a tremendous catastrophe in Vermont, the Catskills, and other regions swamped by flooding rains, so I’d encourage you to help out if you can.
So… who had Vermont in the “State Hardest Hit By Irene” pool?
Hurricane Irene made landfall in three places Saturday and Sunday: first at Cape Lookout, North Carolina, then near Beach Haven, New Jersey (ten miles north of Atlantic City), and finally at Coney Island, New York. But this storm won’t be remembered for its impact on the coast. Oh, there was wind annd surge, and there was damage: the Outer Banks Highway was destroyed, Hoboken was flooded, and Lower Manhattan’s seawall was overtopped, among other effects. But when the best an AP reporter can do for a lead quote about a disaster is, “You could see newspaper stands floating down the street,” that’s a pretty good sign it wasn’t exactly the storm of the century — at least not along the coastlines where it hit. Of particular note, flooding in New York was far more shallow and limited than had been feared.
Instead, Irene will be remembered first and foremost for her inland impacts, specifically the catastrophic river flooding now underway in, among other places, the Catskills of New York and, most especially, the last state in the Northeast to make a pre-storm disaster declaration: Vermont.
(WARNING: Second clip contains profanity.)
The situation in Vermont is pretty terrible, and getting worse as rivers rise. (That appears likely to be the major Irene-related storyline going forward.) Even in New Jersey, where fears of a heavy impact along the shoreline caused Governor Chris Christie to famously advise residents to “Get the Hell off the beach,” the biggest problem was not storm-surge flooding, but river flooding. (“Get the Hell off the riverbank”?)
The inland flooding was no surprise; indeed, I’d been saying for days that it might be “biggest threat of all.” Still, it’s impressive and awful to see the rainfall totals, which are just freaking huge, and the results are devastating in many cases. Irene’s unusually slow movement for this latitude, plus its interaction with a stationary front, have combined to produce a truly historic flooding event that will be the hurricane’s biggest meteorological legacy.
Just as inland flooding was no surprise, the lack of devastating winds was likewise hardly a shocker. We’ve suspected since Wednesday afternoon, and known since Friday morning, that Irene’s winds would not be a huge problem. You can safely ignore the deranged, Drudge-linked conspiracy theory that this 950 millibar storm was somehow a trumped-up tropical depression: plenty of wind observations demonstrate that it was certainly a Category 1 hurricane at landfall in North Carolina, and at least a strong tropical storm in New York and New Jersey. Nevertheless, Irene’s winds were obviously nowhere near what was legitimately feared when it appeared, at midweek last week, that the hurricane might hit North Carolina as a Category 3 or 4, and weaken only to a Category 2 or 3 by its more northern landfalls. But, again, we knew by Friday that the winds would likely underwhelm, and anyone who expected otherwise by Saturday was either not paying attention, being misled by media hype, or both.
What was a surprise was the storm surge — or relative lack thereof. You’ll remember that, even after it became apparent (and was acknowledged in the forecasts) that Irene would make landfall in a weakened state, forecasters were still expecting a storm surge significantly higher than a “typical” hurricane of Irene’s anticipated wind speed. As Dr. Jeff Masters put it Friday morning, “we can…expect a storm surge one full Saffir-Simpson Category higher than Irene’s winds.” Instead, Irene produced a storm surge topping out at roughly 4 feet all up and down the coast — exactly what you’d expect from a storm on the borderline between Category 1 hurricane (4-5 feet) and a tropical storm (0-3 feet). The only reason New York flooded at all was because Irene came in at almost exactly high tide. The tide alone (slightly elevated due to the New Moon), 5 feet above the low water mark, did more than half the work of overtopping those 8-foot seawalls. And they were only just barely overtopped, which is why you saw reporters standing in ankle-deep waters rather than seeing Lower Manhattan transformed into a latter-day Atlantis.
Why did the storm surge (like so many pieces of economic data recently) underperform expectations? I’m frankly not certain. Remember, I’m not a meteorologist, and I’m certainly not a hydrologist. But I’ve seen this story before, as I pointed out Friday night, when surge fears were high:
But will the surge predictions be borne out? I remember well the predictions of a catastrophic, higher-than-the-winds-would-indicate storm surge with Hurricane Ike in 2008, and those predictions weren’t borne out by the reality (although the surge was plenty bad). Will the same thing happen with Irene, or will the surge meet (or exceed) forecasters’ expectations? I just don’t know.
Ike and Irene were both unusually large storms with unusually deep pressures, whose winds didn’t strengthen as much as expected before landfall, and both resulted in unmet expectations of terrible storm surge. I’m not qualified to opine on what exactly went wrong with these predictions — though if the late Alan Sullivan were still with us, perhaps he’d revive his pimple/hive analogy — nor to suggest changes to the models or anything like that. But it’s certainly worth noting that the last two hurricanes to make landfall in the United States were both huge storms that were expected to cause storm surges well in excess of their wind strength, and both failed to meet the expectations of the computer models. Hopefully the NHC’s post-season analysis of Irene will have some answers on this. Or maybe there’s a Ph.D. thesis in this question for some enterprising meteorology student.
Of course, the big question most people are asking about Irene isn’t why its storm surge was 4 feet instead of 6 or 8, but why it was only a minimal hurricane / strong tropical storm at landfall in the first place. It’s tough for me to get too exercised about that question, since that particular die was cast long ago, between Wednesday afternoon and Friday morning. But anyway, the New York Times does a pretty good job explaining it:
“We were expecting a stronger storm to come into North Carolina,” said James Franklin, chief of the hurricane specialist unit at the National Hurricane Center in Miami. “We had every reason to believe it would strengthen after the Bahamas.”
He added, “What we got wrong was the structure of the storm.”
Forecasters had expected that a spinning band of clouds near its center, called the inner eyewall, would collapse and be replaced by an outer band that would then slowly contract. Such “eyewall replacement cycles” have been known to cause hurricanes to strengthen.
While its eyewall did collapse, Irene never completed the cycle, Mr. Franklin said. “There were a lot of rain bands competing for the same energy,” he said. “So when the eyewall collapsed, there were winds over a large area.”
That led the storm to be much larger, but with the winds spread over a larger area, they were less intense.
All of which brings us to the inevitable questions about “hype.” Was Irene overhyped? Well, yes and no. The fearful possibility of a monster hurricane hitting North Carolina, then taking an exceedingly dangerous and potentially mega-destructive track up the densely populated East Coast, was very real and fully justified as of midweek. And indeed, Irene took precisely the near-worst-case track that was being discussed, with near-worst-case tidal timing for both Chesapeake Bay and New York harbor to boot. That’s 2 out of the 3 necessary conditions to produce a mega-disaster. But all 3 must happen for the “worst case” to occur, and Irene only managed to bat .667. Worst-case scenarios generally require a bunch of bad things to happen in just the right (or rather wrong) combination; otherwise, the feared disaster, no matter how real the threat was, doesn’t come to pass. Here, Irene simply failed to ever became a true “monster” in terms of intensity, and thus her effects were vastly less severe than what they could have been. There was nothing preordained or inevitable about that failure — indeed, the meta-conditions were broadly favorable for rapid strengthening — yet it didn’t happen, demonstrating once again forecasters’ admitted lack of skill at predicting hurricane intensity (in stark contrast to the ever increasing skill, sharply on display here, at predicting hurricanes’ tracks up to 72 hours or so).
As I wrote in my post about “misconceptions,” the mere fact that a worst-case scenario doesn’t occur is hardly proof that it should never have been considered a possibility, or that precautions taken against such a scenario were therefore unwarranted. That’s totally illogical. I’m sure NOAA officials and others would love to have access to the 20/20 Hindsight Computer Model that some commentators seem to possess, but absent that, I believe it was completely justified and necessary to evacuate the folks who were evacuated, given the uncertainties in the forecast at the time decisions had to be made (specifically with regard to the storm surge). It’s the nature of the beast, given the current limits of our forecasting ability, that most “alarms” will be “false alarms.” It’s simply impossible to know with certainty what a storm will do at the time when evacuation decisions must be made, so we have no choice but to “prepare for the worst,” knowing full well that, in most cases and in most places, the worst will not happen. Thus, the fact of a “false alarm,” without more, is not evidence of improper “hype.”
Yet overhype certainly exists, not so much in the forecasts or the precautions, but in the media coverage. “Preparation for the worst-case scenario makes sense,” writes the Telegraph‘s Toby Harnden, “and could have saved hundreds during Katrina. But the worst-case scenario was largely portrayed as inevitable.” That’s a big problem in the early stages of hurricane coverage: the tendency to filter out the uncertainties, and treat the worst-case possibilities as probabilities or near-certainties. This, in turn, feeds into a cycle of self-perpetuating hype, which at some point seems to pass a “point of no return,” after which any walk-back of the doomsday talk is seen as irresponsibly advising people to “let their guard down” — not to mention hurting ratings. That helps cause what I view as the primary problem, which I’ve observed many times over the years: the MSM’s failure to adjust the tone and substance of the coverage once it has become apparent that the worst-case scenario(s), despite having previously been realistic possibilities, have now become unrealistic. In other words, they fail to dial down the hype a notch when the hype, once reasonable, is clearly no longer justified. I tweetedFridaymorning about this, stating: “Media must be careful today. Fine line b/w preventing complacency & overhyping a weakened Irene (which breeds cynicism and…complacency). Ideally, you communicate that Irene is a big deal that people should take seriously, but no longer likely to be an apocalyptic hellstorm. But that’s hard to do in practice, especially when MSM weather coverage generally has two settings: 1. #Ignore. 2. #OMGApocalypticHellstorm!”
As I wrote Saturday morning, “we can now be quite confident this won’t be a world-historical disaster… even while being equally confident that it is a force to be reckoned with, and one residents should not blow off. Surely there must be some way to communicate both of these concepts simultaneously.” Needless to say, this properly calibrated, nuanced message didn’t win the day. Instead, many folks apparently continued to believe they were indeed dealing with an apocalyptic hellstorm, a world-historic disaster of the sort that had been, in reality, pretty much off the table since Thursday or Friday. Why did they believe this? Because many folks in the media didn’t convey otherwise, even though the National Hurricane Center had made the diminished threat clear. (It’s not like I was going out on a limb. I was just relaying the publicly available forecast information.)
This pattern is dangerous, because it can breed both complacency and arrogance — the latter exemplified by Anne Thompson’s comment on the NBC Nightly News that New Yorkers had gained their “swagger” back because “New York took the best that Irene could give, and made it through.” That statement might make sense, if Irene had given New York anything close to “the best [it] could give.” But Irene didn’t do that. It’s absolutely critical to understand that this was nowhere near the worst-case scenario for NYC & environs, thanks to Irene’s limited strength. That scenario will occur come to pass someday; it just wasn’t today, thank goodness. But my fear now is that, when the eventual day of reckoning comes, folks won’t take it seriously because “they said that about Irene too.” Complacency caused by media overhype can kill, just as surely as complacency caused by people “letting their guard down” due to underhype. Finding the proper balance is very tricky, and impossible to do perfectly — but the media certainly needs to do better.
Heck, maybe the “day of reckoning” will come sooner rather than later. Let’s hope not. But computer models are calling for “Invest 92L,” a tropical wave off Africa, to eventually become Hurricane Katia (the replacement on the six-year rotating name list for the retired name “Katrina”), and some models think it will eventually menace the East Coast in about two weeks’ time, potentially around the weekend of 9/11. (FSU meteorologist Ryan Maue, he of the beautiful computer model graphics, referred to a Saturday-night model run as the “oh s**t, not again” forecast.) Other models think it will slide harmlessly out to sea. Here’s what the European model predicts for September 7:
It’s way, way, way, WAY too early to tell what “proto-Katia” will do, or to worry about this hypothetical storm. But it just goes to show how active the tropics are right now.
Hurricane Irene came ashore this morning over Coney Island, New York, weakening to a tropical storm as it did so. I’ll save any damage assessment until later, as early impressions are often wrong. But it’s safe to say that New York is very, very lucky Irene didn’t intensify into a Category 4 monster south of the Carolinas, as had been reasonably expected by forecasters before the storm’s inner core structure fell apart Wednesday/Thursday and unexpectedly never recovered. Irene’s track (and timing, in terms of high tide) ended up being pretty darn worst-case for highly vulnerable NYC, but thankfully, its intensity wasn’t. Hence, the storm surge peaked around 4 feet — bad enough to cause significant flooding, but seemingly not catastrophic — and the sustained winds probably topped out at high-end tropical storm force. Had this hurricane become a monster down south, it would have had a much bigger storm surge all throughout its life, and its winds would have had a long way to go in order to weaken to tropical storm force. It’s very plausible New York could have been hit by a Category 2 hurricane in that scenario, maybe even a low-end Category 3, and there’d be no discussion of “overhype.”
As I’ve said repeatedly, the big storyline with regard to Hurricane Irene’s coastal impact isn’t the winds, but the storm surge. Even if Irene weakens to a tropical storm overnight, as she may, what we really need to watch is how high the surge gets. Of particular significance, New York’s subway system will reportedly flood if the storm surge is more than 3 to 4 feet above the already astronomically high tide tomorrow morning.
With that said, check out this NOAA chart, updating in real time (or close to it), of the water level at The Battery in Lower Manhattan:
Click on the chart for a larger version. The blue line is the expected, “normal” water level, as the tides come in and out. Actually, “normal” isn’t quite right, since we’re dealing with astronomically high tides, thanks to the New Moon. But anyway, the blue line is what the tides would be without Irene.
The red line is the actual water level. When you see the red line not declining, or only declining slightly, while the blue line is going way down, that’s bad news. It means the storm surge is preventing the tide from going out, so the next high tide will likely be much higher (assuming the surge is still present when the tide comes in).
But the one to really watch is the green line. This is the “residual” level, the difference between the red and blue lines. Basically, the green line is the storm surge. So, for instance, if the red line is declining as the tide goes out, but not as much as it “should” be declining (as in the scenario just discussed), that will cause the green line to go up.
Low tide is at 2:14 AM. High tide is at 8:07 AM. Irene’s expected closest pass is expected at around 8:00 10:00 AM — so it’ll be very nearby at high tide. Awful timing. (Once the storm passes, the winds shift to blowing offshore, and the storm surge begins to fall away. See, for instance, Yorktown, VA. But in NYC, that likely will not happen until after high tide.)
Oh, and wave heights, I should note, are on top of the surge.
Regarding the subway, if you want to get a little more into the weeds… based on what Dr. Jeff Masters has written about this, it appears the seawall that protects the subway is approximately 8 feet above the “MLLW,” or “Mean Lower Low Water” level. An average high tide is roughly 4.5 feet above MLLW. The New Moon adds another 0.5 feet to tomorrow’s high tide. That leaves a 3-foot margin for error, give or take a few inches. So a surge of greater than 3 feet at high tide would likely flood the subway, unless Dr. Masters is wrong or I’m missing something here.
Moreover, if the surge reaches 4.5 feet at high tide, here are some of the areas that would potentially be underwater at high tide, according to Climate Central:
Note: I’m not sure if that map takes into account seawalls and such. So take it with a grain of salt, giving you a general idea of the amount of territory at risk, but not necessarily a precise block-by-block map of who’d be flooded.
The Weather Channel’s Bryan Norcross just broached the possibility of a 7- or 8-foot surge. I have no idea what that would do, beyond a vague assumption that it would be very bad. It seems high to me — but then again, New York harbor is acting like a giant funnel, so maybe it could happen. Hopefully not.
* * * * *
Anyway, that’s the situation with New York City. But of course, Gotham isn’t the only place at risk. And I thought it might be helpful to try and collect all the relevant tidal information in one place. NOAA maintains an excellent site where you can view tidal gauges for a variety of locales all around the country, including up and down the East Coast, but they’re all on individual pages. It also has a great “QuickLook” page where you can view this data, and more, about Hurricane Irene. But for viewing the critical storm surge data, and only that data, all in one place, I’m taking the liberty of combining a bunch of these charts in this post, on the next page.
One big caveat to all of this: I’m not a meteorologist, but I at least know a fair bit about meteorology from a layman’s perspective. Hydrology is a different story. I am definitely not a hydrologist, and I can’t even pretend to play one on the Internet. So while I think I’m getting the basic facts right here, it’s possible I’m making some mistakes in interpreting these charts. If someone more knowledgeable than I sees an error, please let me know!
