UCAR > Communications > Staff Notes > October 1998 Search

October 1998


Landlocked against the Rockies, NCAR and UOP haven't traditionally been hotbeds of research on tropical cyclones. However, as it turns out, some NCAR scientists are carving out time for basic hurricane research. Others at UCAR are collaborating on technology to better observe hurricanes or on educational materials to help people better understand them. We recently took a counterclockwise spin through the institution to find out more.

In its first U.S. encounter, Hurricane Georges pounded the Florida Keys on 25 September. (Photo courtesy of NOAA.)

Putting the drop on hurricanes

The atmosphere over land is typically monitored from the ground up--through radiosondes, radars, and profilers. The air over the ocean is more often scrutinized in top-down fashion, from satellites and aircraft. For over half a century, "hurricane hunter" flights conducted by the Air Force and NOAA have taken measurements from the heart of tropical cyclones. One key tool on these flights is the dropwindsonde, an instrument package that descends by parachute and radios back findings to the aircraft on the way down.

Until the past couple of years, scientists were happy to get dropsonde reports every 1000 feet (300 meters) as a unit fell to the ocean surface. Now, thanks to NCAR's Global Positioning System (GPS) dropsonde, that vertical resolution has increased nearly a hundredfold for wind-speed measurements. That means unprecedented detail on hurricanes where it counts the most: near the eyewall and close to sea level.

Why did this tree cross the road? Because of Hurricane Bonnie. After striking North Carolina, Bonnie weakened to tropical-storm status, then reintensified to a hurricane. This was one of hundreds of trees downed in the Norfolk, Virginia, area. See page 3 for more on Bonnie and the experiences of UCAR-connected people in the storm's path. (Photo by Bob Henson.)

"It's almost like having an instrument that never existed before," says Terry Hock, an engineer in the Atmospheric Technology Division (ATD) and leader of the GPS dropsonde team. "We're now measuring quantities down to the very surface of the ocean. The resolution is so good that people are asking us whether [an observation is from] two meters or eight meters. That's unheard of in any previous generation."

"Now every dropsonde being dropped into hurricanes uses GPS."

--Terry Hock, ATD

Some 800 GPS dropsondes were used during the Fronts and Atlantic Storm Track Experiment (FASTEX) in the winter of 1996-97. As Hurricane Guillermo raged across the Pacific in August 1997, the sondes helped provide the first-ever high-resolution data on hurricane eyewall structure. These sondes were dropped from NOAA's P-3 research aircraft, but NOAA research flights are only a small fraction of all flights into hurricanes. Routine reconnaissance flights for tracking tropical cyclones are performed from Keesler Air Force Base (near Tampa). This spring, ATD helped equip nine Air Force WC-130s with GPS dropsonde capability. And ATD's field deployment pool includes new sondes that NASA is using aboard its DC-8 this fall as part of a large synoptic-scale hurricane experiment. "Now," notes Terry, "every dropsonde being dropped into hurricanes uses GPS."

Terry Hock with one of the GPS dropsondes that are sailing through hurricanes this fall. (Photos by Carlye Calvin unless otherwise noted.)

The GPS sonde is the latest in a noteworthy dynasty. NCAR developed its first dropsonde using the Omega navigational system in the early 1970s; it entered use in hurricane reconnaissance in 1982. In the late 1980s came the lightweight LORAN digital dropsonde (L2D2), a microprocessor-based unit that was one-quarter the weight of its predecessor.

Both of these systems have fallen victim to the far-superior navigational accuracy of GPS. It's primarily GPS precision that enables dense, high-frequency reports from the new sondes. Even at its near-ground fall speed of about 25 miles per hour (11 meters per second), the sonde can provide wind data every half-second (see diagram). The high fall speed makes the new sonde well suited for deployment from NOAA's new Gulfstream IV and other high-altitude jets. ATD also improved the accuracy and range of the pressure, temperature, and humidity sensors; upgraded the electronics; and built a completely new in-flight processing system. Up to four sondes can now be launched and tracked as closely as 20 seconds apart. Only the square-cone parachute designed for the L2D2 remains constant.

Specifications for the GPS sonde were developed in collaboration with NOAA and the German Aerospace Research Establishment, both of which funded, with NCAR, the sonde's development. UCAR has licensed the technology to Vaisala for sonde manufacture since 1996.

The GPS dropsonde (solid line) battles it out with its Omega-based predecessor (dashed line) in this test from 14 August 1996 near 24.8 N, 77.3 W. The sondes were dropped from NOAA's Gulfstream IV and P-3 aircraft, respectively. Depicted is the south-to-north (v) wind component in meters per second. Along with their much-improved resolution, the new sondes can retrieve data within a few meters of the ocean surface, while the Omega-based sondes cannot provide wind data beneath about 500 meters (1650 feet). (Illustration courtesy of Terry Hock and James Franklin.)

NCAR saluted Terry and collaborators Ned Chamberlain, Hal Cole, Errol Korn, Dean Lauritsen, Ken Norris, and Chip Owens with a nomination for the 1997 Technology Advancement Award. Rave reviews have arrived from the user community as well. "We are entering a new era in hurricane reconnaissance and research," wrote James Franklin (NOAA Hurricane Research Division, or HRD) in a letter to ATD. "It is no stretch to say that the new dropsonde is going to have a profound effect on the way we design research and reconnaissance missions, on the kind of research that we do, and on the accuracy of [National Hurricane Center, or NHC] advisories and warnings."

Helping radar to analyze hurricane winds

Like snowflakes, hurricanes assume an enormous variety of forms derived from a few basic elements. One especially odd-looking storm was Hurricane Earl, which moved into the Florida Panhandle on 3 September. Instead of a classic, symmetric hurricane, Earl looked more like a strong extratropical cyclone, the kind more often seen parading through the U.S. in winter than the Gulf of Mexico in summer.

Wen-Chau Lee.

"It had a huge comma cloud. The southwest side had very few [radar] echoes," notes ATD scientist Wen-Chau Lee. As strange as Earl was, the hurricane served as a useful test case for a new analysis technique created by Wen-Chau.

The goal of Wen-Chau's ground-based velocity-track-display (GBVTD) technique is to better analyze the near-surface winds of an approaching hurricane. The National Weather Service's WSR-88D Doppler radars, recently installed along most of the Gulf and Atlantic coastline, allow incoming tropical systems to be monitored for winds as well as rainfall. However, the radars can only detect winds blowing directly toward or away from the receiver along the radar beam. Tangential winds are not measured directly; they must be inferred, a task that becomes complex when a hurricane has pronounced asymmetries. An intense hurricane that looks perfectly symmetric on satellite can have pockets of high wind scattered in unforeseen places around the storm center. At landfall, these pockets can produce some of a hurricane's worst wind damage, as happened with Hurricane Andrew in 1992.

The idea behind GBVTD is to locate the asymmetric wind maxima before a hurricane nears land. Over the past seven years, Wen-Chau has worked on the GBVTD algorithm in between his duties as scientist for the Electra Doppler radar (ELDORA). Last summer, he spent part of a sabbatical in Miami refining GBVTD, and by this year Wen-Chau was ready to test the technique in real time. His partner in this task was Shirley Murillo, a protégé since 1997 in UCAR's SOARS program (Significant Opportunities in Atmospheric Research and Science; see the June and July 1998 issues of Staff Notes Monthly). Shirley spent this summer in Boulder working with Wen-Chau and familiarizing herself with GBVTD. Their collaborators were Ben Jou (National Taiwan University), Peter Dodge and Frank Marks (HRD) and Colin McAdie (NHC). In August, Shirley began a stint at HRD, where she gave GBVTD a real-world test during Earl's landfall.

It wasn't the easiest forecast in hurricane history. "NHC had a hard time locating the center," says Wen-Chau. "Then all the numerical prediction models went bananas." At one point, the various models were projecting Earl striking anywhere from the vicinity of Nova Scotia to Texas. After dawdling in the Gulf for a day or so, Earl began its final move northeast toward Florida and Shirley tested GBVTD using output from the WSR-88D radar at Eglin Air Force Base.

"What we found was a very weak eyewall circulation," Wen-Chau says. "Most of the strong winds were more than 80 kilometers [50 miles] from the center." Reconnaissance planes reported a peak wind of 83 knots (roughly 95 mph or 42 m/s) at 2 km altitude, whereas GBVTD analyzed a peak of 95 knots at 1 km. "At this point I wouldn't bet on that 95-knot wind as accurate," says Wen-Chau. "However, the GBVTD technique inferred the wind maximum at the right place, and the magnitude is comparable [with the reconnaissance reports at a higher altitude]."

A server glitch at NHC prevented a real-time test of GBVTD for Hurricane Bonnie, a Category 3 storm on the Saffir-Simpson intensity scale that struck North Carolina in August (see " The Bonnie Chronicles"). Still, if they can obtain the needed radar data, Wen-Chau and Shirley may do a retrospective test on Bonnie. "It's pretty circular, and the precipitation is fairly extensive," Wen-Chau says.

Clarity out of complexity: from a sea of Doppler wind data, the GBVTD technique identifies the bands of strongest wind wrapping around the center of tropical cyclones. This depiction of Hurricane Earl was created from WSR-88D Doppler data taken at Eglin Air Force Base on the evening of 2 September (0141 UTC on the 3rd), as Earl approached the Florida panhandle (the coastline is visible at top). The strongest winds, exceeding 90 knots (104 mph or 46 m/s), extend in a band from northeast to southeast about 50 mi (80 km) east of the eye--further away than is typical for a hurricane's peak winds.

The biggest problem in verifying GBVTD's merit is also its raison d'etre: no other technique can provide a full two-dimensional snapshot of hurricane winds. GPS dropsondes can now collect high-precision wind data along their fall paths, but they are deployed by reconnaissance aircraft over the course of a lengthy flight--not simultaneously. "We're trying to compare [dropsonde] data gathered over a couple of hours to a nearly instantaneous [radar] analysis. A hurricane can do all kinds of things in even a half hour."

The VTD technique, a cousin of GBVTD and another joint development between NCAR (Wen-Chau and MMM's Rit Carbone) and HRD (Marks), has been running in real time for four years. VTD uses aircraft-mounted Doppler radar, such as those aboard the NOAA P-3s, to examine tropical cyclones far from landfall. The goal is to produce real-time VTD results every half hour that allow monitoring of the intensity of the circulation into which the recon plane is headed. (Researchers can later expand the two-dimensional data from VTD and GBVTD to three dimensions by combining the imagery taken from different altitudes or radar-beam orientations.) The circulations derived from these complementary techniques will likely be woven into other tools at NHC that provide guidance on the near-surface winds of an approaching hurricane.

ATD has already given the P-3 hurricane-hunter aircraft a boost by Dopplerizing their nose-cone radars. The upgrade was requested by NOAA after Craig Walther and Mitch Randall successfully did the same for the nose radars of the NSF/NCAR Electra and C-130. Such radars weren't designed to serve as research tools, but they can be adapted through the PC-based integrated radar acquisition card (PIRAQ) developed by Mitch and Eric Loew in the early 1990s. PIRAQ provides a compact means for processing radar signals within a personal computer, a task that used to require bookcase-sized units.

PIRAQ was installed on the two NOAA P-3s in August 1997. That fall, they successfully gathered information from the Pacific's intense Hurricane Guillermo. Since then ATD has improved the system's sensitivity and made other software tweaks. With this year proving to be more active in the Atlantic, the system had already seen action in two hurricanes, Bonnie and Danielle, by mid-September.

Inside the eyewall: fingers and beads

Just before Labor Day, Peter Hildebrand and Bob Gall flew through the eye of Hurricane Danielle aboard a P-3. The flight gave them a chance to see PIRAQ and VTD in action. It also allowed them to hunt in person for features they've been scrutinizing in radar displays for years--and to look for new ones.

Peter Hildebrand

Peter, an ATD senior scientist, and Bob, director of the Mesoscale and Microscale Meteorology Division, are carving out time from their schedules to collaborate on some unfunded but promising work on how small-scale features affect a hurricane's dynamics. In July a paper they cowrote (along with MMM's John Tuttle) appeared in Monthly Weather Review. The study uses radar imagery from Hurricanes Hugo (1989), Andrew (1992), and Erin (1995) to investigate long, narrow rainbands that spiral into each storm's center. Large rainbands are a classic feature of intense hurricanes, but these bands are smaller in scale--on the order of 6 miles (10 km) across, 6 miles tall, and 60 miles long. They move with the hurricane's rotating winds. Apparently they can cause variations in near-surface wind of up to 18 mph (8 m/s), with the winds suppressed inside the bands and more intense outside of them.

Bob Gall

"We have been able to describe many features of these small-scale spiral bands," write the authors, "but the question remains, what are they? At this point we can only offer suggestions." The closest analog is boundary-layer rolls, cloud "streets" that line up in the direction of wind shear (beneath a low-level jet stream, for instance). While a doctoral student at the Massachusetts Institute of Technology, Inez Fung (University of British Columbia) produced spiral structures within a hurricane model that strongly resemble the rainbands Bob and Peter have examined on radar. The bands are likely bringing heat and moisture upward, but it's still an open question whether they are affecting momentum transfer into or out of the storm.

Now the two scientists are looking at the hurricane core--the eyewall and the eye itself--to find other small-scale features that could be important. For years, it's been debated how much of the stratospheric circulation atop a hurricane is being entrained into the eye. "We think there are reasons to believe there's an oscillation [or swinging] between the eye rotating in almost a solid manner to a situation where stratospheric air is going into the eye." In the case of near-solid-body rotation, the winds decrease smoothly from hurricane-force in the eyewall to nearly calm at the eye's center. If and when the dropoff in wind speed just inside the eyewall is especially steep, that could enhance instabilities and eddies. Downward eddies, in turn, might pull stratospheric air into the eye in the form of intruding "fingers," thus affecting the hurricane's momentum balance and perhaps its intensity. Downward eddies could also be a factor in producing the asymmetric wind pockets being traced by Wen-Chau's GBVTD technique.

Data from NCAR's GPS dropsondes lends credence to the idea of small-scale eddies, says Bob. "They have dropped sondes in the eyewall that found a huge amount of variability."

Kevin Petty, a postdoctoral researcher in the Advanced Study Program, has been working with the P-3 data from Danielle and other hurricanes. "He's playing a really major role in data analysis," says Peter. It will be some time before Bob and Peter's hypothesis can be proven, but in the preliminary data, Peter says, "We're seeing the kind of oscillation between rotation and nonrotation that we expected, and it's very encouraging."

Meanwhile, for another upcoming paper, Bob is working on a line of research involving pockets of high radar reflectivity that appear to circulate around the storm. "If you subtract the mean reflectivity, you can see these little beads we call pearls that rotate around the eyewall. Some of them maintain their identity all the way around."

Spinning up computer-based education

If your goal is to grab the attention of college-level students taking perhaps the only science course they'll ever take, you'd better have something dramatic for your subject matter. Hurricanes are fitting the bill for a new NSF-funded Web project based at UOP's Cooperative Program for Operational Meteorology, Education and Training (COMET).

COMET's Sue Kemner-Richardson and colleagues have assembled a set of Web pages that explain the science behind hurricanes as part of an overall theme of satellite-based remote sensing. The site is designed for use in a classroom or lab setting as part of science-survey courses. The project is supported by the NSF Division of Undergraduate Education's Course and Curriculum Development Program. COMET has been collaborating with UOP's Program for the Advancement of Geoscience Education (PAGE) to put the Web site together.

"The idea was to come up with an example of how multimedia could be used in undergraduate science education," says Sue. The site is designed to be flexible and to allow easy classroom access to dynamic data. "That's been one of the big shortfalls in meteorology education," says Sue. Extensive experience with multimedia for NWS training made COMET a natural for this challenge.

"What we could pull from other COMET modules, we did," says Sue, "but we had to simplify things. We aren't teaching meteorologists here, we're teaching art majors."

Sue, an instructional designer, teamed with COMET colleagues and with PAGE meteorologist Katy Ginger on the project. PAGE is overseeing an evaluation of the Web site this academic year, to be conducted by the University of Georgia in earth- or atmospheric-science survey classes at its own campus and at 10 to 12 other institutions.

The "Explore Hurricanes" portion of the site tracks three recent events--Opal and Fran (Atlantic) and Nora (Pacific)--using satellite imagery. "Hurricane Structure" delves into spiral bands, eyewalls, and low-level inflow/upper-level outflow. As many as 20% of the students in science-survey courses may be future schoolteachers, so the site is also allowing master teachers to post tips on how the hurricane material might be applied in a K-12 context.

COMET's Vickie Johnson and Sue Kemner-Richardson are overseeing two multimedia projects that explain hurricane basics to diverse audiences.

Meanwhile, COMET's Vickie Johnson has been working with NWS and the Federal Emergency Management Agency (FEMA) on another multimedia-based, hurricane-related project. The goal: design an instructional tool that can be used by emergency managers, mayors, and other key people in hurricane-prone communities--people who don't need graduate-level meteorological detail but who do need substantive information about how hurricanes work and how they're forecast.

"It's more geared toward users of the forecast than forecasters," says Vickie. "It's easy for us to put in stuff about weather, but it's not easy to know what this particular audience knows about weather." Because emergency managers are a mobile group, each year some move from inland areas to the Atlantic or Gulf Coasts, often arriving with little or no experience with hurricanes. Also, because the targeted areas range from rural counties to major metropolitan areas, "we have no clue what kinds of equipment, software, video or sound cards they all have. So we're trying to design this to work for a wide range of platforms."

Aside from emergency managers, the site also aims to give other local decision makers "an understanding of what their emergency managers need to know and need to do, so they can support [evacuation] decisions that may be unpopular and that three-fourths of the time are going to be wrong [i.e., false alarms]."

The project is COMET's first major collaboration with FEMA, which has been working more closely with the NWS in recent years. Vickie is the project manager; other key players include FEMA hurricane expert Bill Massey (nicknamed "Hurricane Bill"), Chris Adams (Cooperative Institute for Research in the Environmental Sciences), and Rainer Dombrowski (NWS).

A Web-based prototype is now being reviewed, and a CD-ROM version to be distributed by FEMA should be completed by early next year. This fall COMET will be meeting with several emergency managers and with FEMA staff for in-person tests. However, the frequency of U.S. tropical-cyclone threats over the past two months has made the logistics of assembling people a challenge. "It's unfortunate we're at this stage of the process at this time of year," observes Vickie, "but it's consistent with Murphy's Law." •BH

Putting the focus on society

Roger Pielke, Jr.

Roger Pielke, Jr. (Environmental and Societal Impacts Group) has had hurricanes on his mind since the early 1990s, when he produced an in-depth study of the warnings, impacts, and responses associated with Hurricane Andrew. The interest goes back even farther for his father, an atmospheric scientist at Colorado State University.

Last fall, the junior and senior Pielkes joined forces to produce the most comprehensive overview to date on the relationship between people and tropical cyclones, especially in the United States. Hurricanes: Their Nature and Impacts on Society (Wiley, 1997) uses the concept of vulnerability to integrate the societal and physical aspects of hurricane impacts. It also illustrates both the benefits and limits of hurricane research for society.

As Roger has pointed out in numerous talks, a mismatch exists between reality and the views of some opinion leaders on the nature of the hurricane threat. U.S. losses in the 1990s have exceeded the previous 20 years' combined, even after adjusting for inflation. "This has led many to mistakenly conclude that severe hurricanes are becoming more frequent," says Roger. In fact, he notes, "1991 to 1994 was the quietest [period] in 50 years."

The bad news is that the quietude may be coming to an end. As documented by Bill Gray (Colorado State University), a multidecadal cycle in Atlantic Ocean temperatures and resulting hurricane activity appears to be on the upswing.

Even without such an increase in activity, the current swarming of people and wealth toward the Gulf and Atlantic coasts is guaranteed to exacerbate the impact of any hurricanes that do arrive. Roger brought home this point in a study conducted with Chris Landsea (NOAA/HRD). By correcting for increases in population and wealth, as well as for inflation, their study found that Andrew would have been eclipsed by a 1926 hurricane as the top damage-producer. That storm struck Miami and moved on to a second landfall in Alabama. "Our estimates suggest that such a storm [if it struck today] would cause more than $60 billion in losses in a Miami landfall and an additional $10 billion in losses in an Alabama landfall," says Roger. The total would more than double Andrew's impact.

"Because we had large losses in the past, we can expect they'll happen again in the future," Roger notes. "We can wait for those to happen to motivate us, or we can better understand our risks today by fortifying our buildings and improving our insurance practices." •BH

In this issue...
Other issues of Staff Notes Monthly


Edited by Bob Henson, bhenson@ucar.edu

Prepared for the Web by Jacque Marshall