1. Local Weather

Did National Hurricane Center flub Erika forecast? Expert responds

An image from the National Oceanic and Atmospheric Administration's GOES-East satellite shows Tropical Storm Erika as it loses strength while passing over over Haiti on Saturday. [Photo by NOAA/NASA GOES Project via Getty Images]
An image from the National Oceanic and Atmospheric Administration's GOES-East satellite shows Tropical Storm Erika as it loses strength while passing over over Haiti on Saturday. [Photo by NOAA/NASA GOES Project via Getty Images]
Published Sep. 1, 2015

FORT LAUDERDALE — Did the National Hurricane Center flub Tropical Storm Erika's forecast?

James Franklin, the center's top hurricane specialist, said Monday that the forecast errors were considerably larger than normal, particularly when the system was four and five days away from a possible hit on Florida. But, he said, officials noted all along the forecast held a high degree of uncertainty because of the storm's disheveled nature and the atmospheric obstacles it faced.

Erika, predicted to aim everywhere from South Florida to the Carolinas and Alabama, fell apart on Saturday morning near Cuba after encountering strong wind shear and the rugged mountains of Hispaniola.

"I wish the forecast had been better," Franklin said. "We always want to make a perfect forecast. But we know that's not going to be true."

Bryan Norcross, hurricane specialist for The Weather Channel, said the hurricane center did a better job than people think but fell down on communicating the forecast uncertainties. He said the "right message" would have been: "There is high confidence Erika will affect Florida, but very low confidence in how strong it will be."

Jeff Masters, director of meteorology for Weather Underground, the online weather site, said Erika's track forecasts were far off the mark and "will be a big drag on the National Hurricane Center's accuracy for 2015."

Here is Franklin's postmortem:

What made the forecast so difficult?

After emerging in the central Atlantic on Aug. 24, forecasters knew Erika would interact with land and face strong wind shear when it entered the Caribbean. But they didn't know how much land it would cross and couldn't be sure if it would hang together or get shredded apart. Further complicating the forecast was the system's mushy nature; it was never fully organized.

"Weak and disorganized systems are always more challenging than the Category 3 hurricane," Franklin said.

What was the biggest challenge?

Trying to predict Erika's strength, because its strength would dictate its path.

A stronger system would be pulled north by atmospheric steering mechanisms and a weaker one would have drifted more to the west, possibly toward Florida.

If the storm plowed over land, it would weaken; it would strengthen if it remained over water.

"The land complicated matters because you needed to hit the track forecast just right to get the proper amount of land interaction," Franklin said. "It had more land interaction than we initially thought it would."

Did the models fall down?

To some degree, yes, because they couldn't grasp how the storm would react to all the complex atmospheric factors. That, in turn, was mainly because the models couldn't get an accurate picture of the storm's structure.

As a result, they swung the projected path back and forth, Franklin said.

"The models have never handled this type of storm particularly well," he said. "What we saw in this case was a great deal of inconsistency from run to run."

What went right?

On Friday afternoon, when Erika was about 48 hours away from a possible hit on South Florida, the hurricane center decided not to issue a tropical storm watch or warning for this region.

"We elected not to do so because we thought there was significant chance those watches and warnings would never be necessary," Franklin said. "So we waited to see what the interaction with Hispaniola would be, and in fact, the next morning, there was nothing left of the storm."

What could the hurricane center have done better?

Even though forecasters emphasized the high degree of uncertainty, residents focused on forecast track instead of the cone of uncertainty.

"We still need to do a better job of conveying uncertainties of tropical cyclones in the forecast," Franklin said.

In the future, the uncertainty might be highlighted in the Public Advisory portion of a forecast, he said. Now it's mentioned in only the Forecast Discussion for a storm.

How can residents better understand the forecast?

They should pay closer attention to the hurricane center's graphics that show the probability of a storm's winds hitting certain areas of the coastline, Franklin said. Had they done so for Erika, they would have seen the chances of any particular spot in South Florida being hit by hurricane-force winds were no more than 4 percent and with tropical-storm force winds less than 50 percent — during the entire Erika forecast.

"Was the hype out there commensurate with a 4 percent risk?" he said.

On other hand, he added: "Anybody out on Wednesday and Thursday buying supplies and going to the store was doing exactly the right thing, getting ready in case a watch or warning needed to be issued for South Florida."

How likely are we to see more off-target forecasts?

Very likely, Franklin said. "We have scenarios like this every year," he said. "This one happened to occur when Florida was in the five-day cone, which calls a lot of attention to it. But this is not a new problem."


Here's how the hurricane center forecast Erika:

Aug. 24: The center aimed Erika toward South Florida from the moment the system emerged in the central Atlantic.

Aug. 25: Forecasters placed South Florida in Erika's cone of uncertainty.

Aug. 26: First the system was predicted to hit South Florida dead-on as a minimal hurricane. Then the track shifted toward Georgia or the Carolinas. By the end of the day, the system was back at South Florida - and projected to hit with top winds of 85 mph.

Aug. 27: Erika still was projected to scrape South Florida as a hurricane and possibly aim north.

Aug. 28: The track shifted significantly to the west and Erika was no longer projected to be a hurricane. But it still was predicted to hit the southern tip of the state as a tropical storm.

Aug. 29: Erika dissipated.

©2015 Sun Sentinel


This site no longer supports your current browser. Please use a modern and up-to-date browser version for the best experience.

Chrome Firefox Safari Edge