Hurricane Irma model consensus on Carolina hit, but is it real?
which reads, in part:
11 AM 9/4/17: As I noted below, one of the problems with the Carolina forecast was that models were still trending south... and that won out. See my new blog entry for details.
10 AM 9/3/17: After the GFS insisted on a mid-Atlantic hit from Irma yesterday, while the Euro track missed the U.S. entirely, there was remarkable consensus on Hurricane #Irma on a U.S. landfall with the overnight models. The GFS, Euro and Canadian models now are all predicting a hit on southeast North Carolina. But...*
*It's a remarkable shift from the chaotic model predictions of the past few days, which showed landfall all over the place... but it could just be a fluke, as we're still 8 days out, too far in the future for most meteorologists to commit to this forecast.
If the models continue to show this track today, and tonight, then we may be on to something. Extremely strong, historic storms are sometimes predicted by the models days in advance with remarkable accuracy... but with thousands of model forecasts for hundreds of storms each year, the odds are against this being the one.
Additional problems I see with this forecast:
- The European model's average ensemble forecast still points the storm out to sea, even though the Operational version has the Carolina hit. This means the possibilities the Euro sees are not yet reflected in its own forecast, which could mean it's just averaging things out.
- Very few of the ATCF models, which show the consensus track as a hit in southeast North Carolina, are making that actual forecast. The wide swath of their forecasts is still Florida to Canada, with no significant grouping around the Carolina track.
- Models are still moving tracks south, and sometimes the trend is more important than the forecast itself.
What about intensity?
The next question for Irma is maximum intensity. Last night, the GFS had her at 882 mb, matching the lowest pressure ever recorded in Hurricane Wilma in 2005 (Tim Vasquez), but most of the ATCF models this morning are split Cat 3/4.
