Your Eyes Can Help Improve the Historical Record of Tropical Cyclones

Do you love looking at satellite imagery of hurricanes and tropical storms?

If so, an ambitious, innovative crowdsourcing project called Cyclone Center is asking for your help to improve the historical record of tropical cyclones.

You don't need even need a degree in meteorology.

Why is improving the historical record so important?

The study of tropical cyclones is hampered by their often remote locations over the open ocean.

Frequently, the only data available on a given tropical cyclone's intensity and location comes from satellite imagery, or a few instrumented buoys or ships that happen to be in the storm's proximity.

Aircraft reconnaissance data, such as from the U.S. Air Force Reserve Hurricane Hunters, is the most reliable data for a storm's intensity and location.

Unfortunately, only about 30 percent of the Atlantic Basin's historical record contains aircraft data, according to a 2009 study by Ed Rappaport from the National Hurricane Center. Missions are not flown in the eastern or central Pacific unless there is a land threat, and routine missions over the western Pacific Ocean haven't been flown since 1987.

A robust, accurate historical record can help scientists search for any long-term trends in tropical cyclone intensity and frequency. It also helps place current hurricanes and tropical storms in proper context.

Due to differing satellite analysis methods among different agencies over time and changes in satellite technology, there is inherent uncertainty in the record of tropical cyclones.

The Cyclone Center project seeks to reanalyze all tropical cyclone infrared satellite imagery from 1978 through 2009 in the north Atlantic Ocean. 

If 32 years of data doesn't sound overwhelming, consider this corresponds to nearly 300,000 satellite images. According to an April 2015 paper about Cyclone Center, it would take a single person about 12 years to analyze all those images, without holidays or vacation.

Thus, the crowdsourcing idea.

Scientists from the University of North Carolina at Asheville, NOAA's National Centers for Environmental Information (NCEI) and the Cooperative Institute for Climate and Satellites teamed up with the Citizen's Science Alliance to launch Cyclone Center, hosted on Zooniverse.

Once on the site, you're presented a series of questions about various infrared satellite images.

You don't need to know about the Dvorak technique, the standard method used to estimate tropical cyclone intensity based on patterns in satellite imagery.

You're simply lead through it by each question, attempting to match the appearance and shape of the cloud patterns with cyclone types. Clicking on a cyclone type brings up a suite of sample images. Just pick the closest match to the image in question.

Other questions involve picking which image looks like a stronger storm, locating a center, or picking which colors wrap around the center.

Advertisement

If you're a little stumped, you can click on the "Guide" or "Tutorial" buttons below the choices for a quick primer.

After you answer a few questions about one storm, you can scroll down to find out more about that storm, and there are links to discuss both the image and storm with other volunteers and scientists.

Just over 19,600 images had been analyzed by a group of just under 9,800 citizen scientists as of early July 2015. 

"Initially we wanted to complete all those years, classifying every single storm," said project principal investigator Dr. Christopher Hennon from the University of North Carolina-Asheville. 

"However, we have recently narrowed our focus to completing individual years first, with 2005 the first year in the lineup. When it is completed, we will go to another year."

But can untrained citizen scientists get the reanalysis right?

"Just matching cloud pictures and applying some rules produces results that are competitive with a computer algorithm -- the ADT-HURSAT," said Hennon. "I don't think we were expecting something that good early in the project, especially with the untrained public."

"I'm just blown away by the quality of the classifications," said Dr. Peter Thorne, project collaborator from Maynooth University in Ireland. "Volunteers initially feel like they must be doing it wrong and therefore are, wrongly, put off. The results show that isn't the case at all."

Better yet, analysis from Dr. Ken Knapp from NOAA-NCEI meteorologist found that it only takes about 10 individual analyses of one satellite image to form a useful consensus.

Now, the Cyclone Center team just needs more enthusiastic volunteers to tackle just a few storms each.

"If everyone watching The Weather Channel or visiting weather.com over a 24-hour period came over and did five storms -- 100 images -- we'd have a global reanalysis since the start of regular satellite measurements done for all basins and all storms in less than 24 hours," said Thorne. 

Knapp says the team is trying to tap weather enthusiasts, those with a latent fascination with weather. 

(MORE: Hurricane Central | Tropical Update)

"We have one volunteer that performed 20,000-plus classifications and is still contributing."

"Our top classifiers are from places one would not expect, including France and Germany," said Hennon. "The Frenchman is from Montpellier and works as a nurse. Our most prolific classifier hails from Brisbane, Australia. All were co-authors on our paper, in order to recognize their important contributions to the project." 

The bottom line is you can make a difference, and learn some meteorology.

"We have emails from users about how interesting and useful the site is," said Knapp. "Many cite the feeling of contribution to science, or learning about the weather."

Sound like fun? Check out the Cyclone Center site and rate some storms yourself. The initiative also has a blog and Facebook page, with the latest news and results.

You can read the full paper about the Cyclone Center project here.

MORE ON WEATHER.COM: Hurricanes From Space (NASA Satellite Imagery)