I’m just going to say it — Hurricane Matthew is going to be a pain.
Much like Hurricane Joaquin last year, Matthew’s possible track is perilously close to the United States, but far from a sure-thing landfall. The National Hurricane Center’s cone of uncertainty suggests the storm will make landfall in eastern Cuba or Haiti, and then move north into the Bahamas. From there — who knows. Really, it’s too far out to tell. We just have to keep watching the forecasts.
But it turns out that cone has very little to do with what the models are predicting or how uncertain forecasters are. Instead, it simply represents the National Hurricane Center’s average track error over the past five years. As the forecasts get better each year, the cone shrinks. But still, it is one fixed size for every storm that year, with no regard for whether the forecast is high- or low-confidence.
To be fair, forecasters at the National Hurricane Center pick the middle path down the center of the cone. That’s their best guess using all the information they have — including the models. So, in a small way, the models are “built in” to the cone. But it’s still a very small range of possibilities when you compile all of the data.
There’s no better way to illustrate the issue than to compare the model forecasts with the cone. This is Wednesday night’s European model forecast (multicolored, mostly blue) along with the cone of uncertainty (white). The “model” I’m talking about here is actually a combination of 51 models — or an ensemble. Each model is run with slightly different conditions, which in the end gives you a range of possibilities. We’re fairly confident that the final track will fall within the colored areas, but we get worried when they fan out this much.
Even though the cone of uncertainty seems to be pointed at the Mid-Atlantic right now, it may not even come close. It could curve out to sea — harmless for the East Coast. But it could also make landfall anywhere from the Northeast to the Gulf Coast. That is the true range of possibilities this model is suggesting.
We keep running up against this cone problem. Last year, Hurricane Joaquin frazzled the nerves of meteorologists up and down the East Coast, in large part because the National Hurricane Center cone did not adequately convey the true uncertainty of the situation. Brian McNoldy, our tropical expert, explained:
The spread in possible tracks from west to east is extremely large. The outcomes in these tracks, which include high-resolution hurricane models as well as full global models, range from a sharp west turn and landfall in the Southeast United States, to a track far away from the East Coast with no impact at all.
But the range of these possible outcomes — even to only include the most-trusted and well-performing models — are not represented by the National Hurricane Center’s static cone of uncertainty.
Sounds a lot like Matthew, right?