Tag: tropical

We’re at the point in storm tracking where I begin to get frustrated. I want to give a specific forecast. I cannot.

Well, I could, but it would be bulls**t. At this moment no one knows exactly where Hurricane Sandy will go!

Over the years researchers have developed dozens of computer models to try and predict storms through the laws of physics. Most only see a few days into the future. With Hurricane Sandy so far away, both in time and distance, the most used models are the ECMWF produced by the European Centre for Medium-Range Weather Forecasts and the GFS from the Weather Service.

Since each model uses its own shortcuts to allow computers to crunch the data each often has its own solution. That’s the case with Hurricane Sandy. As we get closer we hope the models come into alignment.

As of Wednesday evening the ECMWF was calling for landfall in South Jersey late Monday while the GFS brings Sandy to the Canadian Maritimes a few days later! That’s leading forecasters to believe this storm will affect the East Coast.

We’re still days away from knowing exactly where.

I just took a look at the European (ECMWF) and GFS models. They’re our best predictors for medium range weather. Both are now predicting an East Coast strike for Hurricane Sandy. Obviously that is not good news.

The GFS brings landfall to Maine while the Euro hits New Jersey just south of New York City.

Let’s get real for a moment. Both of these models have varied greatly over the last few days. They have changed trajectories by hundreds of miles. There’s a big difference between striking Cape Cod and Asbury Park as the European model has suggested over the last day.

I am more concerned than I was yesterday. Both of today’s scenarios would affect Connecticut and, of course, we’re the compromise position between the two.

It’s still too early to go nuts, but it is the time to start thinking about what you wish you’d had on hand during Irene.

Do you use rechargeable batteries? Are they charged?

Do you have food in your freezer you’d like to start using now?

Are your prescriptions up-to-date?

New England is nothing more than a potential target for Hurricane Sandy. It is still too early to think its post-weekend location can be accurately forecast.

As we get closer we’ll know more.

Here’s my problem with Sandy. I can look at all the computer runs and sense something’s wrong. Too many things are happening I’ve never seen before. You’d think in nearly 30 years here I’d seen it all.

I spoke to Bob Hart on my way home. He’s a professor at FSU. No one knows more about tropical weather. His mind is very attuned to math and physics which eliminates some of the interim steps folks like me have to take.

He was born to be a teacher. We’re lucky.

Bob said there were some storms in the 1800s that produced what is Sandy’s worst case scenario.

There’s a reason stuff like this happens infrequently. A huge number of conditions must come into alignment. If one or two aren’t as forecast it all goes to hell! With Sandy this far away in time and distance there’s no parameter that’s not suspect.

Neither Bob nor I want to believe Sandy could strike Connecticut.

We talked about the weaknesses of models.

Well, he talked. I listened.

Tropical systems are really small compared to the weather we most often see. The computer models are often too coarse to understand the complexity of these tightly wound storms that mathematically can fall between the cracks.

Sometimes tropical storms induce feedback loops. That’s a condition that overwhelms the model’s algorithms and allows the condition to grow out of control. It makes the model output suspect at best and poisoned so much it has to be discarded at worst.

To those who think meteorologists love big storms. No.

I’ll be busy for the next week.

It’s much too early to get worried about Tropical Storm Sandy. Unfortunately, at the same time, we can’t find a reason to dismiss it as a potential threat to Connecticut!

Early Tuesday evening Sandy is around 225 miles SSW of Kingston, Jamaica. Yesterday the storm was stationary. This evening it’s moving north at 8 mph.

We know Sandy is better organized! What was just a cluster of thunderstorms now has the curved bands you’d expect with a tropical system.

Sandy is also over very warm water. Think of that oceanic heat content as gasoline thrown on a smoldering fire.

Sandy is forecast to hit Jamaica Wednesday with hurricane force winds. From there it will continue toward Cuba. Both Jamaica and Cuba have mountainous interiors. That will likely slow the storm’s development

From there the official track stays east of Florida cutting through the mainly flat Bahamas.

What you’ve just read is what we think is likely. Now we come to the part of the forecast that’s much less certain… carved in cottage cheese, if you will.

In order to pin down the forecast we look at a variety of computer models. These take the storm’s characteristics, turn them into mathematical variables, then apply the laws of physics.

Each computer model uses a slightly different method to ‘solve’ the forecasting problem. We have many models because none of them is consistently right! Each has strengths and weaknesses.

One model, produced by the European Centre for Medium-Range Weather Forecasts, showed great skill last hurricane season, especially with Hurricane Irene. Its solution for Sandy is scary for Connecticut.

The ECMWF curves Sandy on a counterclockwise path through the Atlantic, hitting Long Island and cutting through Connecticut from southeast to northwest. It is an exceptionally rare path for a storm. It would have a devastating impact on our state.

A storm like this could easily rival the damage from Irene… but wait.

The ECMWF prediction is for early next week. That’s still a long way away, both in distance and time. Small errors in a computer model early on multiply with time.

At this time there’s just no way to dismiss the ECMWF’s forecast. That’s not to say it will come true, it’s just one of many possible scenarios. There’s little individuals can do and not much you should do today.

Hopefully, as we go forward the computer models will begin to agree and we’ll be able to tell you to relax or get ready.

I was working on my story, speaking with a video editor, when Rachael called across the newsroom.

“The 12Z’s in.”

She was referring to the European model. In these early stages of Tropical Storm Sandy the European has been the most consistent in bringing the storm to New England. The 12Z run is reasonably similar to early output.

The European is still out of line with nearly everything else we use, but it can’t be discounted.

There’s still no certainty or even likelihood Sandy’s coming here. There’s just no way to dismiss the possibility.

We’ll try and treat the storm with the respect it deserves without unduly alarming people. We’re still a long way from it’s possible arrival time early next week.

Stay tuned.

A lot of folks want to know about Sandy. There’s no Sandy yet. There is a cluster of clouds in the warm water south of Jamaica.

Both the European and GFS models build this storm and bring it up the East Coast. First effects for us would be after dark, Saturday night.

I am not putting water in my tub. We are not stocking up on canned goods.

The global models aren’t built to handle tropical systems like this. Even small errors over this much time can produce a totally different outcome. There will be forecast errors.

The GFS solution is strange. It looked like the classic setup for the Fujiwhara effect where “Sandy” would orbit an offshore low. That would push the storm into the coast! Bad.

Instead the GFS combines three lows to produce a vigorous system, but not a hurricane and a good chance the worst wind would be away from us.

There are four GFS runs every day. Each one will be somewhat different than the last.

Right now there’s nothing to do but watch.