Latest Tropical Storm Irene Track and Model Output

Latest track from National Hurricane Center. More tracking available from the link on your left.

Irene 5-Day Forecast

Don’t “Tie Yourself to One Model”, Irene Threatens US

To track TS Harvey, click here. To track Irene, click here. To track 98L, click here. For an overall look at the tropics, go to MYL’s Hurricane Tracking page.

Tropical Storm Harvey, the 8th storm of the 2011 Hurricane Season, is moving through central Belize tonight. It will likely degenerate over Guatemala and Mexico, but is producing quite a bit of rainfall in and around the center of circulation.

Infrared Satellite, TS Harvey, Image: NOAA

Harvey formed in the central Caribbean and didn’t pose a threat to the United States from its inception. At the same time Harvey or TD 8 was forming, a very unimpressive tropical wave was catching our attention in New Orleans, merely based on the long-range computer models.

The models, mostly notably the GFS, or Global Forecast System, was projecting a major hurricane to form as it approached the Caribbean and making landfall somewhere along the US coastline in the end of August.

In 2006, the GFS predicted a similar occurrence about 10-15 days out.

2006 Hurricane Ernesto, Image: Unisys

It turned into Hurricane Ernesto, a category 1 hurricane, that weakened substantially over western Haiti and eastern Cuba, before hitting Florida as a tropical storm. The only reason I bring up Hurricane Ernesto is to touch on the current unreliability of long range computer models. And really what I mean by long-range, is past 3 or 4 days. As I state this, the global computer models have improved. Last year, I noticed that the GFS was doing a fairly good job of forecasting the inception of a storm, meaning in 10-15 days, the global computer models could pick up on a closed circulation (or a tropical depression) forming. I emailed an expert the National Hurricane Center about it, and they concurred, that they had made some improvements to the GFS. What the GFS could not do and still has trouble telling us, is how strong this system will be when it arrives. However, there is another global computer model that is being heavily relied on by hurricane forecasters that is doing a fairly good job in telling us which storms will strengthen and where they will go.

National Hurricane Center

Last year, while attending the annual Broadcaster’s Conference of the American Meteorological Society in Miami, attendees were privileged to meet, not only the entire staff of hurricane forecasters from the National Hurricane Center, but we also got every insight they could offer to how they put together the forecast. One of the forecasters/researchers also told us the ECMWF, the forecast put other by the European Centre for Medium Range Weather Forecasts had been beating their track forecast for the past couple of years. The percentage was small, but somehow, their global forecasting model was able to beat forecasters with 20-30 years of experience. This may get some guffaws from readers, but keep in mind that the official track by the National Hurricane Center was statistically better than all of the other computer models out there. That means—don’t tie yourself to one model.

Ever since they told me that, I can’t help it.  I rely on the “European”.

But then, last year, I noticed the GFDL, the United States’ global hurricane model, nailing the forecast track as well. I was hopeful, proud. I’m not sure why, but I guess it bothered me a little bit that Europe, which is barely affected by hurricanes, could forecast them better than us! Then the GFDL started really performing well on the track and intensity.

That was last year. This year, I’m back to the European. I don’t know what happened to the GFS and GFDL, but this hurricane season, it’s been having some wild swings when it comes to storm tracks. It also ratchets some storms into a major hurricanes 10-15 days out like it did back in 2006.  The European only goes out 8 days, but in the last few years, I felt confident telling my viewers which way the storm was headed and relatively how strong the storm was going to be when it got there. This year, it’s been guiding me through Arlene, Bret, Cindy, Don and Emily. It forecast Emily forming briefly, dying somewhere north of Hispaniola and then possibly reforming as a weak system before getting carried northeast with a trough of low pressure off the East Coast. Guess what happened? Exactly that. Now, the GFS did eventually forecast a similar scenario. And, it’s much easier to be confident with a forecast when more of the computer models are in alignment.

And the European only guides me in track and possible strength. It doesn’t give particulars, what’s causing the weakening/strengthening, how weak/strong will the storm be, what areas could suffer torrential downpours or high winds. It’s merely guidance. We still have to do all of our own analysis of the wind shear, upper air pattern, the water vapor (these tropical systems really struggle with dry air), the Saharan Air Layer, everything that tells us what kind of storm we’ll be dealing with. But, it gives me confidence to let me know if I need to be gearing up for a 5-day stretch of spending the night in the weather center. (Technically I got to go home during Gustav—I lived 3 blocks from the station!)

Infrared Satellite, Invest 97L, Image: NOAA

In terms of Invest 97L, it’s been undergoing some interaction with dry air and wind shear that has kept it from developing into Tropical Depression # 9 or Tropical Storm Irene. It looks like it’s going to be a tropical storm in the next 24-hours. It also looks like despite dry air still to its north and west and its possible interaction with Hispaniola and Cuba, it will most likely hit Florida as a tropical storm or possibly even category 1 hurricane.

It looks like the persistent trough off the East Coast will keep it from affecting the northern Gulf Coast states.

I say this in confidence at this point because of my faith in the European. Also-because the other computer models are lining up with a similar forecast.

But… since I live along the Gulf Coast and not along the beaches of the Mediterranean, I still keep my hurricane plan in place until October 15th.

Dawn Brown, FOX 8 New Orleans

P.S. This is my anecdotal account of forecasting with models, Wunderblogger Jeff Masters has a much more studious approach to explaining the performance of models in his blog, click here.