Proto-Dean update

Dr. Jeff Masters believes that “Invest 90L” will become Tropical Depression 4 tomorrow or Tuesday, and he thinks it will ultimately become Hurricane Dean, and a threat to North America. In a post this morning, he writes:

Watching the computer model runs for 90L is not for the faint of heart. All the major models except the NOGAPS develop the system into a tropical storm or hurricane that tracks westward over the Atlantic, reaching the lesser Antilles Islands as early as Thursday night, August 16. There are four possible scenarios to consider:

1) A strong trough of low pressure is forecast to move off the East Coast of the U.S. at that time, and this trough may deflect 90L northwards so that it misses the Lesser Antilles Islands, and then recurves harmlessly out to sea.

2) In keeping with the steering pattern we’ve observed since late July, the trough is expected to rapidly move onward, allowing a ridge of high pressure to build in. If the trough is not strong enough to recurve 90L out to sea, the storm will be forced to the west once more and eventually hit the East Coast of the U.S. This is the solution of last night’s ECMWF model.

3) 90L will be far enough south and next weekend’s trough will be weak enough that 90L will plow through the Caribbean, and not be deflected north of the Lesser Antilles Islands. The storm would eventually track into the Gulf of Mexico. This is the solution preferred by this morning’s GFS model.

4) 90L will never develop, or will never become more than a weak tropical storm, due to unfavorable wind shear, dry air, or other factors. This is the solution of the NOGAPS model.

Of the four scenarios, I believe #2 or #3 are most likely to occur–90L will develop into a tropical storm or hurricane that will affect the Caribbean and/or U.S. East Coast. Residents throughout the Caribbean and U.S. should anticipate the possibility that 90L may become a hurricane–and possibly a major hurricane–that will not recurve.

With regard to scenario #3, recall that the Gulf of Mexico is really, really warm right now, so it would be seriously bad news if any well-organized storm were to reach it.

Meanwhile, out in the Pacific, monster Category 4 Hurricane Flossie is expected to pass 50-100 miles south of the Big Island of Hawaii late Tuesday or Wednesday. She will have weakened by then, though, probably to a Cat. 1.

7 Responses to “Proto-Dean update”

  1. Harry Eagar says:

    That’s not the actual prediction for Flossie.

    I have the Saturday night statement from the National Hurricane Center in front of me.

    It says the error in the track AVERAGES 200 nm on day 3 and 225 nm on day 4, so the ‘prediction’ is really a consensus of best runs for several competing models for it to pass somewhere between Maui and Johnston Atoll, with a most probable path within a couple of degrees of South Point, Hawaii.

    As predictions go, that’s like looking at a plot that gives a range from Key West to Wilmington and reporting to the public that the storm will ‘hit Fernandina Beach.’


    And the center statement says that AVERAGE errors in wind strength are 20 kts/day.

    In fact, what we have is an ordinary midseason eastern tropical Pacific storm that is wandering around in a vast empty space. Before satellite observation, this ‘monster’ storm probably would never even have been detected.

  2. ScottF says:

    Harry, We’re not being picky tonight, are we? I didn’t see the word ‘prediction’ in the blog post yet you put it in ‘quotes’ like you are accusing Brendan of misusing the term. Your own words are ‘most probable path’. Is that much different from Brendan’s ‘is expected to pass’? And I’m no expert, but is ’50-100 miles south of the Big Island of Hawaii’ anywhere near ‘within a couple of degrees of South Point, Hawaii.”?


    What happened? Did Brendan’s dog attack your cat or something? :-)

  3. Harry Eagar says:

    The prediction is not the estimated most probable path. The EMPP is part of a prediction that covers a huge area.

    People who paid too much attention to the EMPP lost their boats when, as happened with Iniki in ’92, the storm stays within predictions but wanders from the EMPP.

    With this post and the earlier one attacking Drudge, I’m beginning to suspect the Loy had never encountered the concept of error bars.

  4. ScottF says:

    It sounds like you may have a point to make but I still don’t understand your objection to the original post. As I mentioned above, Brendan didn’t use the word, prediction, in the post. It seems like Estimated Most Probable Path is a key term in hurricane forecasting. Perhaps you should have capitalized it? Did Brendan use a similar term that should have been capitalized?

    BTW, “Emergency workers mobilized yesterday to prepare, Big Island Mayor Harry Kim said.” It’s good to see Ensign Kim land in a command position. I was wondering what happened to him after ST:Voyager wrapped its last season. :-)

  5. David K. says:

    With this post and the earlier one attacking Drudge, I’m beginning to suspect the Loy had never encountered the concept of error bars.

    And it sounds like you are a little to “eagar” to find the smallest fault anywhere with anything Brendan says, even attributing to him things he’s not actually saying. I think ScottF has it write, apparently Brendans dog attacked your cat.

  6. Brendan Loy says:

    I based what I wrote on what Dr. Jeff Masters wrote in the linked post. I trust Dr. Masters on these matters; he knows vastly more than I do, and I’ve never known him to blog imprudently about a hurricane. That said, if Dr. Masters’s statement was in error, then so was mine. However, it sounds like his statement wasn’t in error, and what we have here is a simple difference of opinion over the proper way to express the nature of the NHC prediction. Unfortunately, in the course of illuminating that difference of opinion, you mischaracterize what I said rather badly:

    As predictions go, that’s like looking at a plot that gives a range from Key West to Wilmington and reporting to the public that the storm will ‘hit Fernandina Beach.’

    Two problems with this characterization of my statement (or rather, my paraphrase of Dr. Masters’s statement). One, “will” and “is expected to” are two very different things; I said “is expected to,” not “will.” Second, unless Fernandina Beach is 50 miles long, that would be a far more specific prediction than saying that a storm is expected to pass within a 50-mile range.

    You seem to want me to totally ignore the center line of the forecast track, and focus exclusively on the track “cone.” I am well aware of the importance of the cone, and I certainly would not advise anyone on the Big Island to let down the guard. But I still believe it is accurate to say that Flossie “is expected” to go where the forecast track says it will go — not a specific point, but a 50-mile “mini-cone,” if you will, in the center of the forecast track. Such a statement is especially accurate when the models are in good agreement, as it seems they are here (again according to Dr. Masters).

    The average forecast error rate is just that, an average. Just because there’s a large average error rate (which includes all forecasts, including those where the models are not in good agreement) doesn’t mean that seasoned forecasters like Dr. Masters can’t look at the current data and conclude that, with respect to a particular storm, it’s reasonable to focus on narrower portion of the “cone” as the “expected” target. Of course, there are some storms where this is impossible, because the models are in disagreement and the forecast highly uncertain, and in those cases neither I nor Dr. Masters would ever say that the storm is “expected” to go to a 50-mile-wide area 48-60 hours out. But it varies from storm to storm, and on those storms where the forecast confidence is high, such statements are entirely reasonable and proper.

    In other words, the average forecast error is not some sort of meteorological gospel truth. The track cone is not the beginning and end of forecast analysis. These things have to be considered on a storm-by-storm, and forecast-by-forecast, basis. And I trust Dr. Masters to consider them properly and characterize them fairly.

    In any event, your huffy tone is clearly uncalled for, even if you think I’m wrong about this.

  7. DuaneH says:

    I assume the center track in the forecast cone to be the most statistically likely path for the storm to take and the tracks adjacent to this path slightly less likely and so forth until you reach the tracks on the edges of the forecast cone which are the least likely path for the storm. These tracks still have a statistical likelihood of being the ultimate storm path but that likelihood is simply lower than that of the center track.

    Otoh, maybe Harry just needs a little laxative.