» 2008 » September

Weather Nerd

Monthly Archives: September 2008

Kyle causes little damage; Laura forms

September 29th, 2008 - 7:09 am

Hurricane Kyle caused a rainy and blustery night, but apparently “no major problems,” in Nova Scotia and New Brunswick. The brunt of the storm, such as it was, hit southwestern Nova Scotia, knocking “lots of trees down, power lines down” but otherwise causing little damage. At the storm’s height, some 40,000 customers lost electricity; that number is already down to 14,000. Alan Sullivan writes: “Nova Scotia is a stormy place, and this was nothing out of the ordinary. It is quite common for this outstretched piece of land to catch a dying tropical cyclone, but the winter storms are worse.”

[UPDATE: In a hot-off-the-presses post, Dr. Jeff Masters suggests that Kyle probably wasn't a hurricane at landfall: "Kyle did generate one hurricane force wind gust -- 77 mph at Baccaro Point, on the extreme southernmost point of Nova Scotia -- but it is questionable whether it really was a hurricane over Nova Scotia. Kyle weakened dramatically right at landfall. ... Even though Yarmouth was on the strong (right) side of Kyle where the highest winds should have been, the airport measured top winds of only 30 mph, gusting to 50 mph. Not even minor damage was reported there, according to news reports."]

Meanwhile, the National Hurricane Center has initiated advisories on Subtropical Storm Laura, out in the middle of nowhere in the North Atlantic. This is a “fish” storm, a threat only to shipping interests, so there’s not much to say, aside from debating whether its designation constitutes “count-padding.” Unsurprisingly, Sullivan says yes: “This is the most dubious call of the year for NHC, perhaps trying to salvage the broken forecast of a much more active season.” Just as unsurprisingly, I’m skeptical of that conclusion.

If the NHC wanted to pad the storm count, surely they would have given a name to the subtropical-ish system that made landfall in North Carolina a few days ago, about which Sullivan himself wrote: “I won’t grumble if NHC designates this hybrid cyclone.” If the NHC had done that, “Laura” would be “Marco” today. But the NHC didn’t do that, which, it seems to me, deflates the conspiracy theory a bit. Laura may be “marginal,” but surely reasonable people can disagree on such subjective close calls, particularly when there is now monolithic pattern of always erring in one particular direction.

In any case, for what it’s worth — and I’m on record, repeatedly, as stating that seasonal hurricane forecasts are overhyped and overrated — the consensus was that we’d have between 11 and 16 named storms, between 6 and 9 hurricanes, and between 2 and 5 major hurricanes. Right now, with two months left in the season, we’re sitting at 12, 6 and 3. So, if the season ended today (and it does appear to be winding down, climatology notwithstanding), we’d be just at the low end of the predictions, and just above the long-term climatological averages (9.6, 5.9 and 2.3).

Also, Dr. William Gray — the dean of seasonal hurricane forecasting, and, incidentally, a global warming skeptic — predicted an ACE of 150; we’re sitting at 120 (the average is 102). All things considered, I’d say it’s at least premature, and arguably just downright erroneous, to talk about “broken forecast[s] of a much more active season.” Marginally erroneous forecasts of a slightly more active season, maybe.

P.S. As an aside, I don’t think the “another weak season!” meme is going to get much traction, after a year in which we’ve seen, according to Wikipedia, 28 landfalls, 852 deaths (and counting), more than $50 billion of damage, and a string of 7 consecutive named storms hitting the U.S. mainland (Cristobal, Dolly, Edouard, Fay, Gustav, Hanna and Ike). Of course, those aren’t really legitimate measures of seasonal activity, meteorologically speaking. But as a practical matter, nobody’s going to buy the “weak season” argument in 2008. I’d be better, I think, for the skeptics to stick with playing defense against the “it’s all caused by global warming” meme, and not try to advance the competing, and equally misleading, “it was a weak season, which disproves global warming” meme.

Hurricane Kyle made landfall this evening near Yarmouth, Nova Scotia, with 75 mph winds enhanced by a 30 mph forward speed, but then weakened and was declared extratropical at 11:00 PM EDT as it passed over the Digby area.

Kyle’s track, and its tropical-storm-force wind field as of 11pm EDT.

Radar and satellite reveal that most of Kyle’s rain is on the west side of the center, over Maine and New Brunswick (with some additional bands several hundred miles to the east, near and over Cape Breton Island). Here’s another radar view, and here’s a look at some Doppler-estimated storm totals.

The strongest winds, waves and storm surge, however, will have been just east of the center, in Kyle’s right-front quadrant, where the storm’s forward motion of 30 mph must be added to the estimated maximum sustained winds of 75 mph (now down to 70 mph as of 11pm). In addition, Kyle hit near high tide, increasing the likelihood of coastal flooding.

Here is what Dr. Jeff Masters wrote this afternoon about Kyle’s damage potential:

A landfall intensity of 60-65 mph is a good bet. Winds of this level should cause widespread power outages and tree damage over western Nova Scotia, but only light structural damage. Kyle’s expected storm surge of 1-2 feet should not cause major flooding, but its large 10-foot high battering waves could cause considerable coastal erosion. Kyle is moving fast enough that rains are only expected to be 2-4 inches, which should not cause major river flooding.

The 11pm advisory was the National Hurricane Center’s last on the now-extratropical storm, which is expected to move north as a strong non-tropical cyclone, crossing the Bay of Fundy into New Brunswick and eventually Quebec.

For coverage of Kyle’s impact, go to CTV, CBC, the Chronicle Herald of Halifax, the Toronto Star, and the Globe and Mail.

Meanwhile, in Texas, more than 400 people are still missing two weeks after Hurricane Ike hit, including about 60 who lived on the Bolivar Peninsula, which was so devastated by Ike’s storm surge that some places looked like they were hit by a nuclear bomb. I’m afraid the death toll — currently at 31 in Texas — will rise substantially when all is said and done.

Kyle may become a hurricane; bound for Fundy Bay?

September 26th, 2008 - 8:38 am

Fundy’s long and Fundy’s wide,
Fundy’s fog and rain and tide;
Never see the sun or sky,
Just the green wave going by.

Gordon Bok

*   *   *   *   *

Tropical Storm Kyle’s maximum sustained winds increased to 60 mph overnight, and Tropical Storm Warnings have been issued for Bermuda, as a precaution — although Kyle is expected to stay well to the island’s west, and tropical storm force winds are unlikely to impact folks there, barring a course change.

The forecast calls for Kyle to become a Category 1 hurricane in 36 hours, as it passes several hundred miles east of the Carolinas late Saturday and accelerates northward toward eastern Maine or the Canadian Maritimes. However, this projected intensification may not happen. Kyle is currently suffering from heavy shear, and as Dr. Jeff Masters notes, “unless the shear relaxes, allowing Kyle’s heavy thunderstorms to wrap all the way around the center, intensification into a hurricane will be difficult.”

The NHC discussion says the intensity forecast is considered “problematic.” But, for what it’s worth, slow intensification into a weak hurricane is the current, official expectation. After 48 hours,


Masters notes, however, that “the west side of Kyle will remain relatively thunderstorm-free at landfall, due to strong upper-level winds from the west creating high wind shear. According to the forecast wind radius from NHC, tropical storm force winds of 39 mph and higher will miss Massachusetts, but may affect eastern Maine.” Bottom line, this is not mainly a wind event, but a rain event:

Kyle’s main threat is heavy rain.  Kyle’s rains will primarily affect Nova Scotia and New Brunswick at landfall on Monday morning. However, Kyle should pull copious amounts of tropical moisture and the remains of the unnamed storm that hit South Carolina last night northwards into Canada and northern New England. This will create potential serious flooding problems early next week in the region. NOAA is forecasting up to eight inches of rain could fall in New England over the next five days.

The UKMET model is forecasting that Kyle will stall after landfall. If this forecast verifies, there is the possibility that extremely heavy rains in excess of twelve inches will fall over northern New England, New Brunswick, and Nova Scotia next week. Near-record flooding with heavy damage would likely result. However, the other models do not go along with this scenario, and rain amounts in the 6-8 inch range are more likely for northern New England, New Brunswick, and Nova Scotia.

Interestingly, the current official forecast track would bring Kyle’s right-front quadrant directly over the Bay of Fundy, between Nova Scotia and New Brunswick. I say this is “interesting” because Fundy has the highest tides in the world. I honestly don’t know how much damage a potential storm surge could do there — on the one hand, a Category 1 hurricane doesn’t cause a huge surge, but on the other hand, the right-front quadrant of an accelerating storm can be rough because you have to add the forward speed to the wind speed. In any case, because of Fundy’s extremely variable tides, it would make a big difference precisely when Kyle would hit.

Another question that I don’t know (or remember) the answer to is, how much low-lying development is there along Fundy’s shore that would be particularly damaged by a moderate storm surge? Perhaps not much, as they surely must be accustomed to tidal problems up there. On the other hand, I believe a New Brunswick or Nova Scotia landfall by a tropical cyclone moving almost due north would be unusual; it’s my impression that storms at that latitude are usually recurving to the northeast already, and thus just sort of clip the Maritime provinces. (Somebody correct me I’m wrong on this.)

(Full disclosure: when I was a kid, my parents and I traveled numerous times to Nova Scotia and New Brunswick for summer vacations. I’ve seen the tidal bore at Truro not once, but twice. And I’m a big fan of Maritime folk music; hence the Gordon Bok quote above.)

P.S. Dr. Masters also writes about the unnamed storm, 94L, that hit the Carolinas yesterday:

The unnamed storm that moved ashore over South Carolina/North Carolina last night continues to bring heavy rain and strong winds all along the eastern U.S. … The storm dumped 4.16″ of rain in Wilmington, NC, setting a new daily rainfall record for that city.  A storm surge of four feet was observed in Carteret County, NC, and the road to the North Carolina Outer Banks was flooded by the ocean at several points during the storm.

Evidence suggests the storm was probably subtropical or tropical at landfall, and could have received the name Laura.  However, one of the criteria for getting a name is that a storm must persist as a subtropical or tropical storm for a “reasonable period of time”.  This season, it seems that NHC has been waiting longer than in the recent past (the 1990s and 2000s) to give storms names.  However, in the 1970s and 1980s, a “reasonable period of time” was usually judged to be a day or longer.  I doubt that yesterday’s storm would have gotten a name during Nell Frank’s tenure as director of NHC from 1974-1987.  Thus, yesterday’s decision not to name this storm is probably consistent with how things would have been done back in that era.  There will always be a grey area in this regard, and NHC will inevitably get complaints about decisions to name or not name storms.  If they had named this system Laura, they would have gotten complaints that are too quick to give names to storms that do not deserve them, and thus are artificially inflating tropical storm statistics to make it appear that global warming is increasing the number of tropical storms.  Last night’s unnamed storm fell solidly in this grey area, and there is no clear-cut “right” answer as to whether the storm deserved a name or not.

Elsewhere in the tropics, there are two other areas of disturbed weather to watch, but neither appears likely to imminently develop.

Two proto-Kyles, no Kyle? [UPDATE: Kyle!]

September 25th, 2008 - 7:08 am

I’ve been too busy blogging about politics to pay much attention to the doings in tropics over the last 48 hours or so, but suffice it to say, there are now two systems competing for the name “Kyle” — the aforeblogged Invest 93L, now moving away from Puerto Rico and Hispaniola, and a coastal storm off the Carolinas, a.k.a. Invest 94L, which is presently extratropical but which could potentially become subtropical. Both storms are covered in the National Hurricane Center’s latest Tropical Weather Outlook.

Alan Sullivan was bullish yesterday about the coastal storm’s chances, but this morning he doubts that either storm will earn its stripes as a tropical cyclone. In fact, he speculates: “With polar influences expanding so swiftly, there may never be a Kyle at this rate.” Well, not until 2014, at least.

Regardless, the Carolina coastal storm (94L) will produce “STRONG WINDS…COASTAL FLOODING…HIGH SURF…AND DANGEROUS RIP CURRENTS . . . ALONG PORTIONS OF THE U.S. EAST COAST DURING THE NEXT COUPLE OF DAYS.” The Raleigh News & Observer, the Wilmington Star-News, and the Virginian-Pilot have coverage.

UPDATE: Dr. Jeff Masters has a new, lengthy post on both systems, focusing on 94L.

UPDATE, 5:04 PM EDT: Invest 93L, the one north of Hispaniola, has become Kyle. The official forecast track takes it toward Nova Scotia. More later.

P.S. In case anyone was wondering, this season’s “ACE” — Accumulated Cyclone Energy, a more accurate measurement of seasonal tropical activity than simple storm-counting — is at 115.6 and counting. That slightly-above-normal count easily exceeds last year’s slightly-below-normal 71.7, even though we’re presently five named storms short of last year’s total. It also beats 2006′s ACE of 81.7. (That season had the same number of storms, ten, that we’ve seen so far this year.) Of course, we’re nowhere near the ridiculous 248.1 ACE of the record-shattering 2005 season.

Proto-Kyle update

September 23rd, 2008 - 7:07 am

The National Hurricane Center is once again giving Invest 93L a greater than 50% chance of becoming a tropical depression in the next 48 hours, according to the latest Tropical Weather Outlook. “SURFACE PRESSURES ARE FALLING IN THE AREA…AND UPPER-LEVEL WINDS ARE EXPECTED TO BECOME A LITTLE MORE CONDUCIVE FOR DEVELOPMENT.”

Meanwhile, Alan Sullivan says the latest computer model runs look less threatening for Wall Street:

Proto-Kyle is still caught on the terrain of Hispaniola. … There’s no telling what it will do when it finally breaks loose, but models have converged somewhat, and many yield a landfall on Long Island or Cape Cod.

For a storm accelerating at this latitude, only the eastern eyewall constitutes a real danger. New York Harbor would not suffer significant surge unless a hurricane made the freakish New Jersey landfall, southwest of there. This could only happen if a storm were moving along a course west of north. The future Kyle is now modeled heading almost due north — a much more normal course, climatologically speaking. This means greatest risk of surge for southeastern New England … But it remains to be seen whether any future Kyle escapes the trap of Hispaniola with enough strength to develop at all.

Meanwhile, I’ve updated my sidebar, eliminating most of the Ike-specific links — though I’ve left up several links to sites with good aftermath coverage, and I’ve also added a link to the Red Cross. If you have any suggestions for particularly useful links, whether Ike-related or otherwise, please suggest them in comments.

The Wall Street hurricane?

September 22nd, 2008 - 1:22 pm

Alan Sullivan, who has shifted the focus of his Fresh Bilge blog in the past week from hurricanes to the financial crisis, manages to combine the two topics in a single post this afternoon:

Markets continue to be profoundly disturbed. The Dow and other stock averages have crashed again. Regional banks have come under particular pressure: they have not been clearly included in the mortgage bailout. Meanwhile oil has shot up a record $25. This destabilization may be too serious for any attempt at control now. It may simply have to play out, like some Greek tragedy. And some respectable models still run a hurricane Kyle into New York Harbor Friday night. I don’t think this will happen, but if it did, Wall Street is the lowest part of Manhattan. Any respectable surge would shut the whole subway system, probably for weeks. It might also fill Ground Zero with a nice lake.

“Kyle,” it should be emphasized, does not even exist yet, and indeed the latest Tropical Weather Outlook has downgraded its chances of formation in the next 48 hours from “high” to “medium.” But, for what it’s worth, here’s what the HWRF model shows on Friday night and early Saturday morning. And here are some more computer models.

It’s far too early for anything remotely approximating hype, however. Earlier today, Sullivan summarized the multifarious possibilities:

It is not yet clear how proto-Kyle will behave in the coming days. Some models make it a hurricane. Outlier models bring it north-northwest and ashore in New Jersey as a cat one storm, with New York Harbor in the right front sector. Disaster watchers are sure to give this long-shot scenario much play. But it remains to be seen whether Kyle will develop at all, or whether its interaction with an extratropical low in the southeastern US will keep the tropical system sheared and weak throughout its span. During my many years on the East Coast I watched many such storms form, take threatening courses, then stay offshore over their natural habitat — the warm water of the Gulf Stream.

I should say a little more about the models. They divide into three camps. I favor the group that posits very slow northwest movement for five days, with the storm weak and drifting off the Bahamas. Presumably Kyle would hook northeast and away after a week or so. Another camp recurves and accelerates Kyle much sooner and takes it harmlessly out to sea as the cold continental low reaches the East Coast and follows the tropical storm. The final camp posits a rare interaction between the polar low and a strengthening hurricane. In this scenario, the cold system retrogrades inland, and it shifts the steering flow to run Kyle toward a landfall somewhere between Cape Hatteras and New York City. I don’t think this will happen, but there are actually three respectable models in this camp.

The more immediate problem is the extremely severe flooding in Puerto Rico, which is, after all, part of the United States. We may have “disaster fatigue,” but the people on that island are going to need help, after this:

Storm total rainfall amounts have exceeded 20 to 30 inches in parts of southeast Puerto Rico where rivers are up to 14 feet above flood stage. Flash floods and mudslides have been reported across the east, southeast, and southeastern interior Puerto Rico. An additional 10-20 inches of rain is expected over western and southwestern Puerto Rico today, due to the very slow motion of 93L. The rains from 93L are the most that have fallen on the island since Hurricane Georges ten years ago.

Here’s the link to donate to the Red Cross.

P.S. I realize my sidebar is now out-of-date, with its focus on Ike-related resources. I’ll be making some changes tonight or tomorrow.

Ike and proto-Kyle

September 21st, 2008 - 6:55 pm

First of all, I apologize for the lack of updates over the past week. There is, of course, tons and tons of news to report on the aftermath of Hurricane Ike, which I won’t even be able to scratch the surface of here. I hope interested readers have been visiting the Houston Chronicle website, which continues to have lots of good information. Also, for some great photos, check out The Big Picture from the Boston Globe. And John Little, Melissa Clouthier (also here), Ubu Roi, and the Houston Press have been blogging form the affected areas.

For my part, aftermath-blogging has never been my forté, but I certainly would have liked to post more than I have. However, I’ve been very busy, among other things, planning my upcoming move. It’s been kind of a crazy week, and blogging has been necessarily placed on the back burner (after being on the front burner for quite a while there). In addition, I have to admit, I feel a bit like I’ve “overdosed” on hurricane season after the ridiculous string of six consecutive storms hitting the United States: Dolly, Edouard, Fay, Gustav, Hanna and Ike.

In any event… almost a million customers remain without electricity in the Houston/Galveston area — though significant progress is being made, and most are expected to have power restored by Thursday — and the death toll is at least 60, including 26 in Texas. (Many others died well inland as the storm’s soggy, blustery remnants swept north.) And in Galveston, after some false starts and some cutting in line, all residents will be allowed back onto the island on Wednesday, though city officials warn that conditions remain very difficult, and a curfew is still in effect.

Speaking of Galveston: anyone who doubts the additional devastation that Ike could have caused there, if the storm had come ashore only a few miles to the west, need only look at the incredible photos from the Bolivar Peninsula, where the devastation was utterly complete. The Bolivar Peninsula took the blow from Ike’s right-front quadrant that Galveston Island almost took. This was no “false alarm,” folks, and anybody who says otherwise is criminally stupid. “Hype,” my eye. As hard hit as Galveston was, that city was very, very lucky Ike made a last-minute northward wobble. And the same goes for vulnerable areas along Galveston Bay. Something very closely approximating a “worst-case scenario” was averted by a mere 10 to 20 miles. Such wobbles are literally impossible to predict more than a few hours before landfall. Thank God this one happened when it did, or the death toll could have  been perhaps two orders of magnitude higher.

The worst devastation of all occurred in Gilchrist, Texas, a town of about 1,000 permanent residents that was wiped off the face of the earth by Ike’s storm surge: “[O]f the approximately 1000 structures existing in the town before Hurricane Ike, only about five survived the hurricane. … Rescuers who have reached Gilchrist have not been able to find any victims in the debris because there is no debris. Ike’s storm surge knocked 99.5% of the 1,000 buildings in Gilchrist off their foundations and either demolished them or washed them miles inland into the swamplands behind Gilchrist.” Why did this happen? “Gilchrist was built in an unusually vulnerable place,” and Ike made landfall in just about the worst possible spot. Read the whole thing.

Anyway… I mentioned earlier that Ike was the sixth consecutive named storm to hit the U.S. (seven if you count Cristobal, which grazed North Carolina but never actually made landfall). Josephine broke the streak, and now we’ve had a week since Ike’s overland demise without any named systems in the tropics. I’m sure the National Hurricane Center has greatly appreciated the breather, and though I’m a mere weatherblogger, I must say that I have appreciated it, too.

But it looks like the break may be about to end. “Invest 93L,” currently causing torrential rain, flooding and mudslides in Puerto Rico, has inspired a pair of “Special Tropical Disturbance Statements” today, and it appears likely to develop into Tropical Depression Eleven, and then perhaps Tropical (or maybe Subtropical) Storm Kyle.

Thankfully, proto-Kyle appears unlikely to threaten the battered Gulf Coast, and indeed unlikely to be a major threat to anyone in the U.S. mainland, at least in terms of wind. Here’s what Dr. Jeff Masters has to say, as of this evening, about the proto-storm’s prospects:

The intensity forecast
Wind shear is forecast to remain 10-20 knots over the next five days, and most of the reliable forecast models predict that 93L will develop into a tropical storm by Tuesday. The GFDL and HWRF models predict 93L will strengthen into a Category 1 or 2 hurricane by Friday. However, there will be high wind shear very close to 93L for the next five days, and the storm may struggle at times with this high shear. Water temperatures are a warm 29.5°C and ocean heat content will be moderate to high over the next five days. The NHC is giving 93L a high (>50% chance) of developing into a tropical depression by Tuesday. A Hurricane Hunter aircraft will investigate 93L Monday afternoon.

The track forecast
The models agree on a slow north-northwesterly motion for 93L over the next 3-4 days, which would bring the storm to a point between Bermuda and South Carolina. A major complicating factor in the long-range track forecast is the expected development of an extratropical Nor’easter storm off the coast of South Carolina on Wednesday or Thursday. The Nor’easter could bring hostile wind shear over 93L, weakening it, and potentially converting it into a subtropical storm. The two storms may rotate cyclonically around a common center (the Fujiwhara effect), sending the Nor’easter west-southwestward into the Southeast U.S., and 93L northwestwards towards North Carolina. This is the solution of the 18Z (2 pm EDT) GFDL and HWRF model. The NOGAPS model predicts that the Nor’easter will not develop at all, and instead 93L will absorb the energy that would have gone into creating the Nor’easter. This might convert 93L into a hybrid subtropical storm that would affect the coast of North and South Carolina late this week with sustained winds in the 50-60 mph range. I don’t have a good feel for what will happen in this complicated situation, but it currently appears that coastal North Carolina may get tropical storm force winds from the extratropical storm beginning as early as Wednesday night. It is possible that 93L may impact the mid-Atlantic or New England regions early next week as a strong tropical storm.

Dr. Masters also notes an unusual competitor for the name “Kyle” — an extratropical system approaching the coast of Portugal: “This storm has developed some heavy thunderstorm activity near the center, and has winds of 40 mph, according to this evening’s QuikSCAT pass. However, in NHC’s subjective judgment, it does not yet have enough tropical characteristics to be named subtropical storm Kyle. This system should make landfall in southern Portugal Monday afternoon, bringing heavy rain and tropical storm force winds of 40 mph to the coast.”

Ike’s aftermath

September 13th, 2008 - 9:13 pm

Almost 24 hours after Hurricane Ike made landfall over Galveston, we still aren’t entirely sure what the full extent of its impact was. We know there’s a ton of damage. We know millions are without power, and probably will remain that way for quite some time. But many questions remain unanswered, particularly with regard to the human toll. We’ll undoubtedly learn a lot more in the coming days. The Houston Chronicle website is probably the best news source for information on the storm’s aftermath. See also the various links at right.

Meanwhile, here’s an aerial video of the Galveston Island flooding, taken from a Coast Guard helicopter. You can even download it, if you like. (It’s a six-minute clip, and a 154 MB file.)

By the way, Ike is now a tropical storm… and, in case you were wondering, there’s nothing brewing in the tropics that appears likely to develop. Thank goodness.

Almost far worse

September 13th, 2008 - 11:33 am

As I mentioned below, updates will be limited for the rest of the day. For full coverage of Ike’s aftermath — which, I hasten to say, is plenty destructive, notwithstanding all the “could have been worse” talk below — I recommend the various blogs and local media outlets listed in my sidebar at right. Local sites like the Houston Chronicle staff blog and the CBS 11 and ABC 13 blogs have more and better coverage of the storm’s impact than I would in any case, from my distant perch in East Tennessee.

Anyway… Dr. Jeff Masters says that, notwithstanding the miscalculations of the storm-surge computer models, Ike was almost far worse for Galveston:

Although Ike caused heavy damage by flooding Galveston with a 12-foot storm surge, the city escaped destruction thanks to its 15.6-foot sea wall (the wall was built 17 feet high, but has since subsided about 2 feet). The surge was able to flow into Galveston Bay and flood the city from behind, but the wall prevented a head-on battering by the surge from the ocean side. Galveston was fortunate that Ike hit the city head-on, rather than just to the south. Ike’s highest storm surge occurred about 50 miles to the northeast of Galveston, over a lightly-populated stretch of coast. Galveston was also lucky that Ike did not have another 12-24 hours over water. In the 12 hours prior to landfall, Ike’s central pressure dropped 6 mb, and the storm began to rapidly organize and form a new eyewall. If Ike had had another 12-24 hours to complete this process, it would have been a Category 4 hurricane with 135-145 mph winds that likely would have destroyed Galveston. The GFDL model was consistently advertising this possibility, and it wasn’t far off the mark. It was not clear to me until late last night that Ike would not destroy Galveston and kill thousands of people. Other hurricane scientists I conversed with yesterday were of the same opinion.

These are the sorts of last-minute variables that cannot be predicted, with any precision or certainty, more than a few hours in advance — which is further proof of why it’s so ridiculous and irresponsible to cry “hype” every single time a prediction just barely goes amiss. Hurricane forecasting can truly be a game of inches.

Those people who refused to evacuate Galveston Island, and other vulnerable low-lying areas, are still fools. They’re just fools who lucked out. Unfortunately, that probably means that some of them are now emboldened fools. But at least they’re not dead fools, and thank God for that.

Credit where it’s due

September 13th, 2008 - 10:56 am

I’m going to be traveling tomorrow, so I need to run some errands, pack for my trip, etc. In addition, my two alma maters, Notre Dame and USC, both have big games today (against Michigan and Ohio State, respectively), which I intend to watch. So I’m afraid I won’t be able to provide much Ike aftermath coverage this afternoon and evening.

I recommend the various blogs and local media outlets listed in my sidebar at right. Local sites like the Houston Chronicle staff blog and the CBS 11 and ABC 13 blogs have more and better coverage of the storm’s impact than I could possibly do anyway.

*   *   *   *   *

Meanwhile, on the issue of Ike’s weaker-than-expected storm surge — which was destructive but not cataclysmic — Alan Sullivan has a good initial wrap-up post, which he calls the “Lessons of Ike.” He writes:

Let’s talk about coastal flood modeling. That went badly wrong yesterday. I used a common sense analogy to help readers understand how the sea surface deforms during hurricanes with different characteristics. When a hurricane is tightly wound like Charley, striking near Fort Myers, Florida in 2004, its core pulls up the sea like a pimple. When a hurricane is broad and comparatively weak, like Ike yesterday, the surface of the ocean resembles a hive — swelling covers a much larger area, but it has no sharp peak. I don’t know how the forecast models are constructed. I am not a mathematician or a meteorologist. But if the output of a computer defies common sense, it’s usually wiser to ignore the computer.

He then proceeds to discuss the possibility of “built-in bias” in models “devised by and for the government that wants to keep everybody safe from everything.” I’m a bit skeptical of that aspect of his analysis. I would imagine the models are designed to be as accurate as possible, and any “bias” caused by the “nanny state mindset” would come in with the interpretation of the models and dissemination of the model data to the public (see, e.g., “certain death“). I’m certain that the flaw in the modeling, whatever it is, that led to this major error, will be studied and, if possible, corrected. Even taking the most cynical view, false alarms do not advance any “nanny state” agenda. They detract from it, in fact, since they tend to cause people to grow cynical and trust the government’s warnings less. Surely even the most “nannyish” bureaucrat understands that.

But Sullivan’s “common sense” approach definitely has merit. Keep in mind, Sullivan, like me, is not a meteorologist; he’s just an amateur weather buff who likes to blog about hurricanes. It’s difficult and risky to challenge a consensus as strong as yesterday’s “storm surge catastrophe” conventional wisdom, especially when you’re not an “expert,” and all the experts are convinced by the models. So Sullivan deserves a lot of credit for bucking the CW. I was looking at the very same “common sense” facts that he was, but I assumed the experts and models were right, and thus came to the opposite conclusion. I don’t think this was unreasonable on my part — but it was, as it turns out, wrong, whereas Sullivan was right. That counts for something. Indeed, if you’ll indulge me for a moment, allow me to put it in perspective:

Three years ago, I got a bunch of attention in the media for, supposedly, presciently “predicting” Hurricane Katrina’s impact on New Orleans. This made for a good storyline — the 23-year-old law student in Indiana saw it coming, and President Bush didn’t! — but it was mostly nonsense. As I tried repeatedly, and usually in vain, to explain to various reporters, I didn’t really “predict” anything, and certainly nothing unique or extraordinary. I said what everyone “in the know” was saying about the threat to New Orleans; I just said it loudly and repeatedly, got Instalanched, posted frequent and link-rich updates, and ultimately built quite an audience. Sure, I was ahead of the curve as compared to the MSM and some elements of the government, but that’s about as difficult as being the best football team in the ACC. And anyway, countless meteorologists, disaster planners, other bloggers, and assorted weather nuts were just as “ahead of the curve” as I was.

Sullivan’s coverage of Ike is a different story. He truly stuck his neck out, and saw what the pros failed to see — twice.

1. First, at a time when the experts were nearly unanimous that Ike still had an excellent chance of intensifying explosively into a high-end Category 3 or 4, Sullivan dissented. “This is not The Big One — except in size,” he wrote Wednesday evening, predicting a “category one or two” landfall. “If Ike can develop no more than it has over the Loop Current, it will not be spinning suddenly to category four, fifty miles off Galveston.”

On Thursday morning, while many others were still waiting for Ike’s circulation to tighten, and its winds to rise in reaction to its pressure drop the previous night, Sullivan predicted that the storm would retain the “weird” internal structure that was preventing such strengthening. “Hurricanes often develop distinctive individual characteristics and retain them through much of their cycle,” he wrote. He said he doubted the NHC’s “concern that this sprawling system might contract (and thus develop stronger winds) just as it approaches the coast” would come to fruition. “Ike’s damage may be considerable, but it will be scattered and not record-setting in any one locality, IMO.”

On Thursday afternoon, Sullivan called Ike “overhyped,” saying, “The inner core has dissipated. We now have an eyeless hurricane with greatest winds in a ring far from the center. This is not a configuration for strengthening. This is not The Big One.” Gradually, the consensus evolved, and came around to Sullivan’s position: Ike would not strengthen significantly before landfall. This proved true, of course. Ike came ashore as a Category 2 — not the Cat. 4 monster once envisioned by forecasters.

2. Sullivan made his second bold prediction on Friday morning, as experts began to propound a new fear: that Ike, although not a “major hurricane” according to the Saffir-Simpson scale, would nevertheless produce a catastrophic storm surge, more typical of a Category 4 storm. This fear, it should be noted, was not based on “hype” or “hysteria,” but rather, on raw data: an Integrated Kinetic Energy level rarely seen in recorded history, and SLOSH computer models that predicted a storm surge of 15-20 feet — even 25 or 30 feet in some spots — along the vulnerable Texas coast. But, again, Sullivan dissented from the direst of forecasts:

Extreme surge scenarios are being propounded for this storm. I suspect we will learn that they are excessive. Ike is too broad and weak to follow models based on the tight core structure of a normal hurricane. The danger will arise not so much from the height and speed of surge; but from its duration, and the immense wave heights built in the huge storm. I do not believe any location will be as severely affected as Mobile Bay in Katrina, where surge heights really did exceed twenty feet. But the lesser surge of Ike will be spread along a much greater length of coast. It will be very destructive.

Analogy: The elevation of ocean surface resembles a pimple in a typical hurricane. Ike’s deformation of the sea will resemble a hive — broader and flatter.

Sullivan urged evacuation of vulnerable coastal areas, but described the National Weather Service’s “certain death” pronouncement as “over the top.” He also rejected commenter Steve Sadlov’s comparisons of Ike to a Gulf of Alaska storm or, worse, the North Sea flood of 1953. “But… it’s not all piling up in one place,” Sullivan wrote. “Remember my analogy: this is a hive, not a pimple.” Even as the advance storm surge, hundreds of miles from the storm’s center, seemed to confirm everyone else’s worst fears, Sullivan stuck to his guns. “I still doubt the extreme projections for surge height,” he wrote at 2:06 PM yesterday. “People have never seen the ocean rise before a hurricane, but this is a storm in a class by itself.”

A couple of hours later, Sullivan did say that “we may be looking at the most expensive hurricane in history,” and his concern grew last night. “Galveston will probably be a total loss,” he wrote. But he added that he still wasn’t “buying” the direst scenarios. “The trouble,” he wrote, was not the height of the surge itself, but “the track: right front of eye coming straight at the mouth of Houston Ship Channel — at high tide.” As he later explained:

When it appeared for a few hours that landfall would be 20-30 miles SW of where it actually was — and I was fully persuaded of that — I was alarmed about Houston. This was the one place where I thought it was possible that model heights might be better justified. If the course of the eye had held NW for longer, onshore wind would have kept piling surge into the mouth of the bay, while flooding rains fell inland.

But as soon as it became apparent the eye was coming right over Galveston, that concern was alleviated.

Bottom line: Sullivan never bought into the computer model scenarios. He trusted his “common sense.” The rest of us trusted the models. We were wrong. He was right.

Importantly, Sullivan’s wrap-up post explicitly endorses the wisdom of evacuations, and chides those who failed to get out of harm’s way:

No matter what the controversies about models or codes, when a hurricane approaches, coastal residents should secure their property and leave, if they live in flood zones or flimsy buildings. What good is the security of a brick house, if it is only one story high, and water fills it to the ceiling? What protection is a mobile home on high ground, when the big pine blows on top of it?

This, too, is a matter of “common sense,” he says. Indeed.

Alas, those who would thoughtlessly deride prudent preparations and evacuations as an unnecessary byproduct of “hype” are sorely lacking in common sense — or any sense at all. Hurricane forecasting is an imprecise thing. Tiny, last-minute wobbles, like the one that spared Galveston Island and Galveston Bay the destruction that everyone (even Sullivan!) feared last night, cannot be predicted more than a few hours in advance. Perfection in forecasting is, and will always be, impossible. As a result, most alarms are false. This is true of necessity, and is unlikely to change anytime soon. It is a byproduct of our limited ability to precisely forecast the chaotic machinations of our atmosphere and our oceans.

Not every “false alarm” is purely “hype.” Very few are, in fact. Of course, even true threats can be overhyped — and, needless to say, media overhype is everpresent. But there is almost always a solid basis for actual evacuation orders, and Ike was no exception.

For what it’s worth, I stand by everything I have posted here regarding Ike. I believe the information I disseminated, and the warnings I relayed, were accurate and reasonable at the time I posted them. And that, as I know Sullivan would agree, is the only standard by which statements about hurricanes can be fairly evaluated: the standard of contemporaneous reasonableness, not 20/20 hindsight. I don’t believe my dire warnings about Ike’s potential were “hype,” so long as they are understood as just that: warnings about potential, not statements of certainties. (I admit, I became fairly convinced last night that Galveston’s utter destruction was a near-certainty. But so did Sullivan, and so did almost everyone else. That last-minute wobble was truly a godsend.)

It comes down to this: over the last several days, Alan and I looked at the same information and interpreted it differently. I think both interpretations were plausible, and I think I stated things reasonably. As it turns out, though, my interpretation (and nearly everyone else’s) was incorrect, and Sullivan’s was correct. So kudos to him. *tips cap*