Hurricane forecast: unravelling the storms

April 25th, 2011  |  by  |  Published in Cover Story  |  1 Comment

three men in front of a large globe

Manoa hurricane experts Businger, Barnes and Wang

On Aug. 18, 1992, halfway around the globe from Hawaiʻi, a disturbance in the atmosphere appears to leave the coast of Africa. Tracked for several days across South and Central America, it arrives in the Pacific on Aug. 28. Continuing westward, it morphs into a tropical depression on Sept. 5 and, three days later, into a bona fide tropical storm dubbed ʻIniki, Hawaiian for “strong wind.”

Living up to its name, ʻIniki intensifies into a hurricane packing Category 2 winds and continues westward south of the Big Island. Steadily powering up, it turns into a Category 4 hurricane on Sept. 10 and takes a sharply northward turn.

That evening, hurricane warnings are issued for Oʻahu and Kauaʻi. Early on Sept. 11, ʻIniki turns northeast and blasts Kauaʻi with gusts up to 175 mph, causing $2–$3 billion in damage.

It had been just 10 years since Hurricane ʻIwa wreaked severe damage on Kauaʻi. But in the 18 years since ʻIniki, no hurricane has struck Hawaiʻi, despite formation of seven to eight hurricanes a year in the eastern and central Pacific.

Are the islands lucky? Or overdue?

Meteorology professors at the University of Hawaiʻi at Mānoa for once agree: Hawaiʻi’s climate doesn’t make it a frequent hurricane target…for now.

“Hurricanes form around the Baja Peninsula in the eastern Pacific when it has the right ingredients—warm water at least 100 feet deep (so that strong winds don’t bring up cool water from below), a low pressure system with surface flows that make the winds converge and spin, wind shear that’s too weak to tear the storm apart and a moist atmosphere where surface air rises high and forms deep convective clouds,” explains Gary Barnes.

From the eastern Pacific, a hurricane would take about 10 days to reach Hawaiʻi; it tends to collapse before it gets within a few hundred miles of the Big Island. “Once tropical cyclones move into the trade winds, they get smaller and smaller because the shallow layer of moisture over Hawaiʻi is unfavorable for hurricanes,” adds Yuqing Wang.

Why did ʻIniki and ʻIwa make it? Unusual disturbances to the west steered the storms north toward Hawaiʻi, recalls Steven Businger. Barnes adds: “ʻIwa and ʻIniki formed during an El Niño, when westerlies often replace the trades,” providing more background spin, or rotation, to the winds.

The Holy Grail of hurricane research

These three scientists work to provide information that for improving hurricane forecast models, the computer programs that use mathematical equations to represent the “engine” driving storms.

The models are still far from perfect. Research is hampered by a scarcity of open ocean observations needed to simulate a hurricane. Since hurricanes originate far from weather stations, forecasters must rely on satellites and reconnaissance aircraft.

But satellite data lack details on weather conditions. Costly, dangerous flights into storms yield only brief sampling. And ships, an important source of weather data, do everything they can to avoid hurricanes.

“To help see what’s going on, forecasters plug a bogus cyclone into their model,” explains Barnes. “Then they guess at a lot—the wind field of the storm, the upper level winds, wind shear, sea surface and air temperatures and especially humidity.”

“When hurricanes get close to Hawaiʻi, NOAA sends out reconnaissance aircraft, like they did for Hurricane Felicia in 2009,” he adds.

“We have good skill in predicting cyclone tracks up to a week,” notes Wang. “But there’s been no significant improvement in the past 30 years in predicting intensity.”

Cutting-edge research deals with how a cluster of disorganized thunderstorms quickly grows into an organized tropical cyclone with spiral rainbands, an eyewall and hugely destructive winds and rainfall.

“A big unsolved problem is whether we can simulate and predict intensity changes. That’s the Holy Grail of hurricane research now,” says Businger.

Observers and modelers

diagram of a mature hurricane

Diagram of a mature hurricane. Illustration by Nancy Hulbirt

Barnes, the observationalist, has been flying into the eyewall of hurricanes with NOAA hurricane hunters for more than 20 years.

“Hurricanes must acquire energy from the sea to create a strong updraft of warm moist air,” he says. This energy is collected in the 500-meter-thick layer of atmosphere just above the sea. He wants to identify when the warm core first forms and what energy changes occur to turn a storm into a hurricane with high and dangerous winds.

After flying a “figure 4” through the storm to find the center, aircraft jettison sensors to record temperature, relative humidity, wind speed, wind direction, ocean-wave size, even images of rain drops. GPS dropsondes take accurate measurements twice a second until plunging in the ocean. Although superior to older, less sturdy windsondes, they provide only limited sampling.

“It’s like making a few slices with a knife through a huge storm,” Barnes says.

His best observations came from Hurricane Humberto in 2001. NOAA and NASA had four to five planes in the storm over three days. More than 200 sondes were deployed.

“We got to sample the pressure when the beast was just forming. We have fields (observations) that no one has yet seen, and they show something unexpected and disturbing: a hurricane forms when a plug of warm air reduces air pressure in the eye and the eyewall. Maximum warming appears to occur much closer to the ocean surface than we expected.”

Businger combines observational research and modeling to understand how hurricanes form and gain energy. He examines small-scale structures called vortex rolls that influence energy exchange between the sea surface and the hurricane, ultimately impacting how destructive a storm is.

He has spotted vortex rolls in Doppler radar data but needs more solid observations. Since dropsondes plummet straight into the ocean, Businger is working with NOAA to build balloons that can gather more information. Made of the same sturdy material as bulletproof vests, these smart balloons will carry meteorological instruments, he says.

“We can communicate remotely with the balloon and can keep it in the layer we want to study.”

Wang has a joint appointment with the International Pacific Research Center, which specializes in climate research using computer models. The three-dimensional tropical cyclone model he first developed more than 15 years ago represents a laboratory under ideal conditions.

“Nature is so complicated; you see the phenomena, but you don’t know which parts are important,” he explains.

In the model, researchers can isolate and control different physical processes. For example, using the environmental conditions during Hurricane Katrina, his model generated a second eyewall. Formation of a new eyewall outside the first signals a very intense storm.

As in real-life Katrina, Wang’s second eyewall moved inward and replaced the inner eyewall. By running the model under different conditions, he determined that increasing humidity helped trigger the second eyewall.

“In the western tropical Pacific there’s a lot of moisture in the air. That explains why typhoons in the western Pacific usually start out as large storms,” he concludes. The largest and most intense ever, Typhoon Tip in 1979, had rainbands more than 1,300 miles across.

“Atlantic storms tend to be smaller because they start out, as ʻIniki did, as easterly waves off the African coast near the Sahara, where the atmosphere is so dry.”

Improving hurricane forecasting

Wang’s findings, like Businger’s, help improve operational forecasts. For example, in the Weather Research and Forecasting Model prediction system currently in use, storms quickly grow too large, generating hurricanes too often.

“Our research suggests this is partly because the initial disturbance, the bogus cyclone, put into the model is too large,” Wang says.

While on Wang’s research team as a postdoctoral fellow, Yokohama National University Associate Professor Hironori Fudeyasu analyzed results of a new computer model that for the first time showed tropical storm systems over the whole globe.

Called NICAM, the new model began with weather conditions on Dec. 15, 2006. Two weeks later, NICAM was generating a disturbance that grew into a tropical storm on nearly the same day and took the same track as an actual tropical storm that developed over the eastern Indian Ocean and drenched northwestern Australia.

A first in long-term hurricane “hindcasting” simulation, it yielded a lot of information about the conditions under which tropical cyclones form, intensify and weaken. A major atmospheric disturbance, the Madden-Julian Oscillation, supplied the right ingredients for a group of thunderstorms, called hot towers, to merge and form the tropical cyclone’s eye. The idea that hot tower mergers signal the origin of a powerful storm had been proposed 50 years prior, but has been difficult to test.

A new project of Businger’s may provide additional evidence. “When there is lightning it usually rains, and when it rains, heat is left in the atmosphere and becomes a source of energy for hurricanes,” he explains. To test whether lightning rate could tell us whether a hurricane might form and when it might intensify, he recently set up the Pacific Long-range Lightning Detection Network in collaboration with a private high-tech company and with funding from the Office of Naval Research to collect lightning data over the ocean.

Global warming increases Hawaiʻi’s risk

Tropical cyclones need surface ocean temperatures around 80°F or higher to fuel their whirling engine. So what happens if climate change causes sea temperatures to rise?

Running Wang’s model under a global warming scenario, IPRC postdoctoral fellow Markus Stowasser, Wang and IPRC Director Kevin Hamilton found that tropical cyclones in the western North Pacific did not become more frequent, but did become more intense, with higher wind speeds and heavier rainfall. The findings jibe with the Intergovernmental Panel on Climate Change’s 2007 report.

IPRC scientist and UH Mānoa meteorologist Bin Wang is a well-known monsoon and climate modeling expert. His work suggests that, while the frequency of the most intense tropical cyclones in the western Pacific, the super typhoons, hasn’t changed over the past 40 years, changes in ocean temperatures and winds due to global warming will cause a shift by the end of the century. Fewer typhoons will originate in the western Pacific, more in the eastern North Pacific.

Fellow IPRC expert and UH Mānoa Professor Tim Li finds a similar eastward shift. His modeling study suggests the frequency of tropical cyclones will decrease over the western North Pacific but increase over the central North Pacific, in part due to weakening tradewinds. Should the models be correct, more hurricanes will blast Hawaiʻi if Earth continues to warm.

Prospects for prediction

Will we be able to predict hurricane structure and intensity one day?

Businger isn’t convinced, but Barnes anticipates making measurable progress within 20 years.

“Everything in science is incremental—knowledge and skill, model and forecasting.” maintains Yuqing Wang. “I say we can do it one day. Not suddenly, but gradually, one day.”

Tags: , , ,


  1. George Curtis says:

    April 29th, 2011at 5:00 pm(#)

    A caution about public use of “categories”. These have not normally been used in Hawai`i since they were developed for the Atlantic and Gulf coasts. In Hawai`i, a Cat 2 just off shore can (and has) become a 3 or 4 due to orographic amplification.

    Many NWS records are drawn from the usual stations located at airports — mostly nearshore and flat.

    So, caution has to be used in discussing or warning about windspeeds in Hawai`i and storm categories. You may realize that, but not all forcasters do.

    George Curtis (ex-JIMAR)