Weather Nerd

Will Edouard turn northwest?

The Houston Chronicle‘s Eric Berger notes that “there have been some hints today about Edouard possibly taking a northwest turn before making a Texas landfall, which would bring the system closer to the Texas-Louisiana border than to Galveston.” Berger explains the possibilities to his Houston-area readers:

[T]here are now a couple of scenarios in play for the Houston area.

Edouard strikes Galveston Island: If this happens then conditions tomorrow should be pretty much be as I describe below. It will be very wet and very windy, a day to avoid being outside if possible. Expect rainfall totals of 3 to 5 inches across the Houston region, with isolated totals of 8 inches. This remains the official forecast.

Edouard turns toward the Texas-Louisiana border: Such a scenario is possible and Jeff Masters, of The Weather Underground, believes this will happen. He’s usually a reliable source. If this plays out the Houston metro area won’t get much wind, and we’ll probably get less rain than the 3 to 5 inches currently being forecast. And tomorrow people will be saying, “Where’s that storm the media promised us?”

Also, as Berger notes, a northwest turn “would be a bad thing for the state’s drought.”

I mentioned in my previous update that the current official forecast track is on the left, or south, side of the computer-model “consensus,” according to the National Hurricane Center discussion. This is because the NHC is waiting — as it often does, when there’s a shift in the computer models’ thinking — to see whether the next model run will confirm the shift. In essence, the NHC is hedging its bets, splitting the difference between the previous official forecast and the current model consensus.

This practice is designed to prevent ping-ponging, whiplash-inducing forecast tracks that rapidly shift to and fro when the computer models prove fickle, as sometimes happens. I understand that rationale, but I often wonder why the NHC doesn’t at least move the “probability cone” to more accurately reflect the true probabilities in such situations, instead of always having the track “line” be smack-dab in the middle of the “cone,” even when it’s concededly not in the middle of the consensus forecast range (and particularly when that “range” is itself trending in a particular direction). For example, instead of having the forecast map look like this…

…why not leave the “track line” in place, but move the “cone” to something like this?

That would — roughly speaking — more accurately reflect the forecast thinking, no? After all, nobody seriously thinks Edouard is headed toward Corpus Christi, yet according to the official forecast map, a landfall there is equally likely as a landfall near Lake Charles, Louisiana (which is, in reality, much more plausible). That’s just silly, and misleading.

On a related note, I’d also like to see the NHC adjust the width of the forecast “cone” to reflect the actual level of uncertainty in any given forecast, so “high confidence” forecasts would have narrower cones, while “low confidence” forecasts would have wider ones. Again, I think that would be more accurate that than the current one-size-fits-all approach.

Just my two cents.

P.S. In case this wasn’t clear, the “cone” in the second map above was crudely drawn by yours truly, and is totally unofficial.

And watch, now that I’ve said this, Edouard will head toward Corpus Christi, just to make me look like an idiot. :P

P.P.S. Meanwhile, an update on the previous get-off-the-roads kerfuffle. Ubu Roi, who started the whole controversy with his comment that the Houston Chronicle‘s Eric Berger was unjustifiably hyping Edouard by suggesting that people should stay off the roads and businesses should shut down, posted a lengthy discussion of the issue on his own blog. And then he posted an update revealing that the mayor has changed his mind, and has ordered all “non-essential” city employees to stay home tomorrow.