Irma is a historic storm with the potential for historic damage to Florida and perhaps other southeastern states, depending on its track.

Over the past week, we’ve all seen a succession of maps with a lot of possible scenarios for Hurricane Irma’s eventual landfall here in the United States. And we’re seeing what this storm is capable of already in the Caribbean.

But for the general public, it can be difficult to understand what you’re looking at with all the different lines drawn on some maps. Many of those lines — and those maps — are not even that useful. They’re just shown for illustrating an idea, rather than a specific forecast.


Forecasters aren’t going to know until sometime Saturday where Irma will actually land in Florida, and even then, there will still be a level of uncertainty.

Different models and different versions of the same model are showing meteorologists a variety of possible answers for the direction this storm could take.

Ultimately, the path of this storm will determine whether major — perhaps even catastrophic — damage occurs in the mainland United States.

Though the cloud shield of a hurricane can cover hundreds of miles, the inner core with the strongest winds doesn’t branch very far from the center, at least comparatively speaking. Travel 50 miles in any direction from the center of the storm, and winds rapidly drop below hurricane force.

With Hurricane Irma, the wind field will actually expand over Florida, so a large portion of the state could feel wind gusts up to hurricane strength, but not Category 4 winds.

This uncertainty means more people in the United States need to be warned about the storm than will actually be impacted by the most ferocious part of any hurricane, including this one.

Worst case scenario

For southern Florida and the big cities of Miami and Fort Lauderdale, a track where Irma comes onshore around Key Largo would be devastating. This would bring the strongest winds and the biggest storm surge in that area.


Although building codes have improved over the past 25 years, newer construction remains untested in the force of such a storm. We’ve never seen a storm this powerful move over such a populated area.

Best case scenario

The storm moves east of Florida, keeping the strongest winds and storm surge over the ocean.

This would be similar to Matthew’s track a year ago. The storm would probably then move into the Georgia or Carolina area, sparing the densely populated region of Florida but obviously being horrible for states to the north.

Another scenario

The storm could also move further west and head up the western side of Florida. This would, depending on the exact path, cause big issues for the western part of Florida, but the storm would weaken inland and pose less of a catastrophic hit further north.

A lot of models

You often hear about the two major models: the one from Europe, Euro, and the one here in the United States, GFS.

These models are “run” every 12 hours for the Euro and every six hours for the GFS. With each new model, we receive a new prediction for the ways in which the atmosphere might behave.

Some models try to forecast the hurricane but aren’t made for this type of forecasting. In fact, they aren’t really worth looking at for forecasts at all, yet they are often shown on TV or on the Web. It’s important to remember this when viewing these so called “spaghetti plots.”


Meteorologists look for trends in these models to determine the storm’s track and strength. We also know certain models tend to have their own biases, such as moving a storm too fast, too far in one direction, or as portraying the storm too wet or dry. The Euro has been the most consistent model, and though it’s not perfect, it does outperform the GFS overall.

Philippe Papin‏/University of Albany

The European model has also been the most consistent so far with Irma’s track.

Different versions of the same model

Another tool we use is called Ensemble forecasting. This takes the same model, such as the GFS or the Euro, and runs it with slightly different conditions.

Think about using the same recipe, but changing the oven temperature by a degree or two, just to see if there is any difference with the outcome. When the ensembles agree (they have the same outcome), we gain more confidence in the forecast.

With Irma, the ensembles are still not converging on a singular solution for the impact of the storm on Florida.

In the map below, each circle is a different spot where the ensemble members of the GFS believe Irma will be this weekend. The storm can’t be in all those spots at once, but the forecast is made to the best of our ability.



The GFS ensembles still have a wide range of possible tracks for Irma.

As of Friday morning, the Euro ensembles were forecasting that Irma would have a direct hit on south Florida.

Consensus track

The Hurricane Center uses all this information to build its best guess of Irma’s future track, which can shift east or west every three to six hours as new forecasts are made. As the varying models change, several meteorologists conference to determine which will have the most influence on the consensus track.

The cone of uncertainty takes into account all these variables to tell residents about the possibility of a predicted storm shifting 100 miles or more in any given direction. I would love nothing more than for this storm to stay off the coast and weaken before making landfall, but authorities must tell residents about worst case scenarios. As we saw in Houston, those worst cases can actually happen.

Follow Dave Epstein on Twitter @growingwisdom.