The outstanding weather problem in the United States is the unanticipated, sometimes destructive, always hazardous occurrence of small-scale weather disturbances.Just before 6:00 p.m. on Friday, 2 August 1985, Delta Air Lines Flight 191a wide-body Lockheed L-1011was on its final approach into DallasFort Worth International Airport. As the airliner broke out of heavy rain near the end of the runway, the flight controller in the tower realized it was coming in too low. He transmitted the frantic command, "Delta, go around," but it was too late. As it crossed a road close to the runway, the L-1011 dropped abruptly. The plane sheared off the top of a passing car, killing the driver, then plowed into the ground, bursting into flames as it bounced and broke apart. Of 164 passengers and crew members aboard Flight 191, only 31 survived the crash.
The National STORM Program: A Call to Action,
University Corporation for Atmospheric Research, 1983
Delta Air Lines Flight 191 was struck down by an unanticipated, destructive, small-scale weather disturbance known as a microbursta sudden, violent, highly localized downward rush of cold air beneath a convective cloud. These powerful downdrafts are virtually impossible to detect with the radar and wind instruments available to weather forecasters and air traffic controllers at most major U.S. airports. Microbursts have been blamed for more than a dozen airline accidents in the last 20 years, including a 1975 crash at New York's Kennedy Airport that killed 113 people and one at New Orleans in 1982 that took 154 lives.
The microburst is the deadliest form of wind shearan abrupt change in wind speed and directionfor aircraft that encounter it during a few critical moments of takeoff or landing. Wind shear is what meteorologists call a small-scale or "mesoscale" phenomenonit occurs on a physical scale that falls between the microscale processes of raindrop and hailstone formation and the macroscale motions of regional and global weather systems. The mesoscale is sometimes called the stormscale, because it encompasses severe storm phenomena such as microbursts, tornadoes, thunderstorms, and flash floods. Mesoscale weather also includes nonviolent events such as heavy local snowfalls, freezing rain, dense ground fog, low clouds that disrupt air traffic, and temperature inversions that produce severe air pollution. Much of the weather that affects our lives most directly occurs on the mesoscale, and it is clear that better forecasts and warnings of mesoscale weather events could save many lives and provide substantial economic benefits. Better forecasts will come from two sources: improved scientific understanding of mesoscale weather processes and phenomena and more effective application of the knowledge and technology that we now possess.
|If an aircraft on its final landing approach enters a microburst, it encounters a sequence of strong winds blowing outward in opposite directions that can force the plane into the ground.|
Edward Zipser, director of the NCAR Cloud Systems Division, concurs. "Mesoscale research is important to the nation as well as a tremendously challenging scientific frontier," he says. "But to make progress in understanding the mesoscale, we're going to have to define the key problems and come up with field research programs carefully designed to get the data we need to solve them."
UCAR played a leading role in planning the National STORM (Stormscale Operational and Research Meteorology) Program, designed to help weather services predict mesoscale weather phenomena with sufficient accuracy and reliability to protect the public and provide economic benefits. A second goal is to help government, industry, and agriculture use weather and climate information more effectively.
Working with support from two government agencies and a private foundation, UCAR formed a scientific steering committee, chaired by George Benton of The John Hopkins University, that produced The National STORM Program: A Call to Action, a report released in March 1983. This report proposed a two-pronged effort: the STORM Operations Program, which would apply new technology to improve observation, analysis, prediction, and dissemination of warnings about stormscale weather, and the STORM Research Program, which would develop a scientific foundation to ensure that this new-generation, high- technology operational system is used to best advantage in the long run.
Benton says that UCAR's role in the National STORM Program involved three steps. "First, we had to put together a good plan. We assembled a steering committee of very competent people who were working at NCAR as well as several universities and government laboratories. This committee developed a plan that made sense from a scientific point of view, but that was only the beginning of the fight." The next step was to get the atmospheric research community behind the plan. "This meant holding workshops and involving university scientists in refining the plan," Benton says. "This phase took nearly two years.
"We realized that having a good plan with the university community behind it would get us nowhere unless the right people in the federal government understood the need for the program," Benton says. "So UCAR put on a different hat and went to the executive and legislative branches of the government." Benton and then-UCAR President Robert White briefed agency heads and senior staff people at the National Science Foundation (NSF), the National Oceanic and Atmospheric Administration (NOAA), the National Aeronautics and Space Administration (NASA), the Department of Defense (DOD) and other federal agencies. They met with key legislative staff people as well as members of the Senate and House of Representatives.
"Without that kind of educational effort, the program wouldn't have had a chance," Benton says. "That was a new direction for UCARtaking the initiative to educate government decision makers about the need for an important research program conceived by the university community."
The current budgetary climate in Washington and other public policy factors have combined to keep the National STORM Program from coming to fruition as quickly and vigorously as its planners would have liked. Preliminary fieldwork began in Oklahoma and Kansas in the summer of 1985, and Zipser, Anthes, and other NCAR scientists had leading roles in an interagency group that prepared a plan for STORM Central, the first of a series of major field experiment to be managed by a project office established by NOAA.
Of all mesoscale phenomena, the nearly ubiquitous convective storm, or thunderstorm, probably affects human affairs most frequently and dramatically. Many thunderstorms are potential killers, spawning lightning, hail, flash floods, microbursts, and tornadoes that take hundreds of lives and destroy billions of dollars' worth of crops and property every year. But thunderstorms bring benefits as well as dangers. The North American "breadbasket"the great food-producing agricultural region of the central United States and Canadagets most of its growing-season rainfall from thunderstorms. If spring and summer thunderstorms fail to bring enough rain to the region, drought can wipe out many millions of acres of wheat, corn, and other food crops.
|Pendulous cloud structures, known to meteorologists as mammatus, often bulge downward from the bases of large thunderstorm clouds.|
|If enough heat and moisture are available, and the atmosphere is sufficiently unstable, a thunderstorm can grow explosively as enormous amounts of energy are released by physical processes within the cloud.|
Many attempts have been made to modify the behavior of thunderstorms by "seeding" them with crystals of silver iodide or other substances to try to increase rainfall or decrease destructive hail. But these efforts have not produced any conclusive scientific evidence of success, although some farmers and commercial cloud seeders are convinced that their weather-modification projects have been effective.
Trying to predict when thunderstorms will bring life-giving rain and when they will spread death and destruction is a major challenge for weather forecasters. Both forecasters and cloud seeders can testify that, in spite of its familiarity, the thunderstorm is a marvelously complex and perversely unpredictable system of interacting physical and dynamic processes. The mechanism that starts this giant machine going is convectionthe upward movement of a parcel of warm, moist air. Once this simple process starts, it can set off a virtual explosion of energy as water vapor in the cooling, expanding air changes state, forming liquid droplets and ice crystals and releasing latent heat that accelerates the upward motion of the air. If the atmosphere is sufficiently moist and unstable, a little cumulus cloud can grow in a matter of minutes into a gigantic cumulonimbus, or thunderhead, ten miles high and ten across.
Some thunderstorms occur in solitary splendor, mushrooming up against a brilliant summer sky and dwarfing the puffy little fair-weather cumulus clouds around them. Known to atmospheric scientists as an isolated or air-mass thunderstorm, this kind of storm can unleash as much energy during its brief life cycle as was released by the atomic bomb that destroyed Hiroshima. Although isolated thunderstorms are not rare, ranks of thunderstorms often form along advancing cold fronts, triggered by a wedge of cold air that forces warm, moist air upward ahead of the front, starting the convective process. According to Zipser, "Triggering by cold fronts is actually a special case of the great propensity of convective storms to grow into organized mesoscale systems such as squall lines. Understanding when and how these systems become organized is one of the greatest challenges facing us in mesoscale meteorology."
Like research on many other problems of the atmosphere, convective storm research has been characterized by a constant interplay of observation and theory. Field observations provide data that are used to develop numerical models on every scale, from the formation of tiny ice crystals inside the cloud to the dynamic behavior of the total convective system. The models are run on computers to simulate and analyze the processes that are being studied, and the results define the data that are needed from future field programs.
There have been many convective-storm modeling efforts in the UCAR Community. For example, since the mid-1970s, Joseph Klemp of the NCAR Atmospheric Analysis and Prediction Division and Robert Wilhelmson of the University of Illinois have been developing, improving, and experimenting with a three-dimensional cloud model that simulates the dynamics of convective storms. One of their early experiments, first run in 1977 on NCAR's CRAY-1 computer, simulated the splitting of a single isolated thunderstorm into two independent, self-sustaining storms. More recently, NCAR scientists have used the Klemp-Wilhelmson model to study how individual storms organize themselves into squall lines and mesoscale convective systemslarger structures that often spawn tornadoes, flash floods, and other severe-weather phenomena that take lives and destroy property. Many other scientists at UCAR Universities and NCAR have developed other convective cloud models.
Most of the data that the modelers use to construct, test, and run mesoscale models come from large, carefully designed field experiments. Zipser points out that, by contact, modelers who are studying large- scale features of the atmosphere usually use data collected primarily for operational purposes as input for large-scale weather forecasts. "Operational data just don't get the job done in mesoscale research," Zipser says. "Weather balloon soundings and surface measurements made twice a day at stations 500 miles apart don't provide the time and space resolution we need to study things like convective storms. They happen too fast over too small an area."
Much convective storm research in the 1970s, including NCAR's National Hail Research Experiment (NHRE), was undertaken to assess the feasibility of seeding clouds to make them produce less hail or more rain. NHRE was undertaken in response to a request from the federal government, stimulated by claims of dramatic success in hail-suppression projects in the Soviet Union. Planning began at a hail-suppression symposium organized by NCAR in 1965, and field work was done in northeastern Colorado in the early and middle 1970s under the successive leadership of William Swinbank, David Atlas, and Donald Veal. The results of NHRE and other scientific investigations into the feasibility of modifying precipitation from convective storms were inconclusive as far as the cloud-seeding experiments were concerned, but direct observation clearly showed that the complex structure and behavior of convective storms required a more detailed scientific understanding before any valid conclusions could be reached about modifying their behavior.
Whether the ultimate goal is prediction or modification, the most pressing need in convective storm research has always been for more accurate, comprehensive measurements in and around thunderstorm clouds during all their stages of growth and decay. But taking measurements inside a thunderstorm is not an easy task. Scientists have been penetrating thunderstorms with instrumented aircraft since 1947, when Horace Byers of the University of Chicago used a fleet of ten Army Air Corps P-61 Black Widows in the Thunderstorm Project, the first major field program to probe convective storms. Today's research aircraft have vastly more sophisticated instruments than those that were available four decades ago, but aircraft data are still far less comprehensive than the researchers need.
Patrick Squires, who served as the last director of NHRE and preceded Zipser as director of the NCAR Cloud Systems Division, once summed it up this way: "Aircraft provide reasonably clear-cut measurements of temperature, liquid-water content, and updraft. But they are at best line measurements along the flight path through the cloud. It isn't good enough just to fly through the cloud and say that's the way it is, because the situation will be quite different three minutes later or half a mile to one side or the other."
Another observing tool, Doppler radar, can provide an overall view of motions within and around a thunderstorm that complements the aircraft measurements. Doppler radar is valuable because it can measure the velocity of targets moving toward or away from its antenna. Instrumented aircraft and Doppler radar were the two most important data-gathering tools used in the 1981 Cooperative Convective Precipitation Experiment (CCOPE), the largest and most ambitious thunderstorm field research program ever mounted. The project was jointly planned and conducted by NCAR and the Bureau of Reclamation of the U.S. Department of the Interior. The field work was done in the summer of 1981 over a 27,000- square-mile area in eastern Montana. More than 100 scientists and technicians from 29 institutions, including universities, federal agencies, and foreign research groups, participated in CCOPE.
According to Patrick Squires, who was co-director of the project, "We were trying to understand the cloud as a wholehow ice particles form in various parts of the cloud, how they are transported to other parts, and how those parts get to be the way they are when the particles arrive there. To do this, we needed a four-dimensional view of the cloud, as it were."
Besides aircraft and radar data, most of the information on the atmospheric environment came from rawinsondesballoons equipped with sensors to measure temperature and humidity. A network of ground- based instruments measured what was happening beneath the storm. The component of the experiment that brought all these observing techniques together, focusing them on the same place at the same time, was the operations center. Developed from experience in previous field programs such as NHRE, it enabled one scientistthe operations director for the dayto orchestrate the whole ensemble of aircraft, radar, and other observing tools.
As Squires described it, "That control center was a masterpiece. It gave one capable scientist the ability to be aware of all the information that was coming in and to plan the next 20 minutes or so of operations. Everybody in the field had someone in the center to talk to. If you were operating a Doppler radar, there was a controller in the center to give you your azimuth and elevation and scanning procedures so you were illuminating the same volume of the storm that was being observed by the other radars and penetrated by the aircraft."
|Severe hail, such as these giant hailstones that fell from a Missouri thunderstorm, inflicts millions of dollars in damage to crops and property every year.|
When a research operation was in progress, controllers seated at consoles might be talking over as many as six radio channels at oncefour frequencies were available for communicating with aircraft, one for radar, and one for mobile ground units. Three stations around the network tracked each of up to ten aircraft, and an altimeter on each aircraft transmitted its altitude. A color display on the operations director's console showed the position and altitude of each aircraft, along with radar information on the storm itself.
Zipser describes the CCOPE fleet as "the biggest and most sophisticated research aircraft collection that's ever been put together, as far as I know." Brant Foote, one of the CCOPE operations directors, observes: "I'm not aware of any other experiment that has used this many aircraft operating this close together. On some days, we had three or four different missions going at the same time in different parts of the cloud. We designed some standard missions, but we had to be opportunistic. If nature produced something a little different from what we had written in the book, we had to adapt to it."
Looking back, Zipser notes that any field experiment in atmospheric research requires a certain amount of luck. "it's not like a chemistry experiment in a laboratory where you mix two compounds together in a test tube and measure the reaction rate," he says. "It's not a controlled experiment." The CCOPE scientists were depending on nature to provide the particular phenomena they wanted to study. But skill and experience also played an important part in their success.
"When you design a field experiment and pick the location, you have to study the climatology and make sure you have a high probability of getting the kinds of storms you want. For the first two months of CCOPE, we had only a handful of good storms. But during the last three weeks, we had three superb events and several others that were pretty good. The two best storms occurred during our last week in the field," Zipser recalls. "But that's part of the game. We had a well-designed network, and we got some excellent data sets."
|Edward Ziper of NCAR (right), Irving Watson of NOAA, and NCAR visitor George Chen of the National Taiwan University, on a research flight during PRE-STORM field work.|
Organizing, analyzing, and drawing conclusions from the CCOPE data have continued to occupy a major part of the time and effort of many scientists at NCAR and the universities that participated in the experiment since the field work ended in August 1981. "The publications that appear during the first two or three years after a major field experiment usually don't present important results." Zipser says. "They describe what was done, how a particular measurement was made. Around the third and fourth years, people start putting together solid case studiesthat's where we are now in 1985 with the CCOPE data. Around the sixth year, the volume of papers peaks outwe expect a lot of CCOPE publications in 1987. The synthesis and review papers come late, when the volume of publications is declining, but they are the ones you'd better pay attention tothey come up with the broad, important conclusions."
Without big computers like NCAR's CRAYs, Zipser says, it would be impossible to handle the masses of complex data sets from diverse sources that come out of an experiment like CCOPE. "But scientists still have to spend a lot of time looking at the data and thinking about themit usually takes two or three years before you are able to really start working with the data in a way that produces useful conclusions."
Although the CCOPE results ultimately will lead to practical applications in forecasting (and, perhaps, in modifying) convective storms, the project was a basic research project designed to advance fundamental knowledge of a mesoscale phenomenon. Another mesoscale research effort is focused much more directly on a specific practical needavoiding the kind of violent and unanticipated downdraft that brought Delta Air Lines Flight 191 down on its approach to Dallas-Fort Worth Airport. The Joint Airport Weather Studies (JAWS) Project began as a cooperative effort between NCAR and the University of Chicago. Its goal was to improve techniques for detecting and observing microbursts. After a microburst killed 154 people in a jet airliner crash in New Orleans in 1982, the Federal Aviation Administration (FAA) and NASA provided additional funding for JAWS. Other universities, as well as NASA, NOAA, and two British government agencies, augmented the project with additional instrumentation and research support. Working at Denver's Stapleton International Airport, which has heavy commercial air traffic and frequent summer thunderstorms, the JAWS researchers, led by John McCarthy of NCAR's Atmospheric Technology Division and Theodore Fujita of the University of Chicago, studied more than 70 microbursts and other wind-shear events.
|The structure of a microburst is clearly revealed by the upward-curling "foot" of rain reaching out from the dark shaft precipitation below this massive thunderstorm near Denver's Stapleton International Airport.|
In May 1984, a United Airlines 727 ran into a microburst as it took off from Stapleton. It struck a radar antenna beyond the end of the runway, but it was able to circle and land with no injuries to the passengers or crew. Spurred by this near disaster, the FAA funded a second project, CLAWS (Classify, Locate, and Avoid Wind Shear), also led by NCAR's McCarthy. The goal was to develop a prototype system for forecasting microbursts and other wind-shear events and for providing timely advisory information to pilots. For 45 days in July and August of 1984, two NCAR meteorologists were stationed in the Stapleton control tower and two at an NCAR Doppler radar site about 20 miles northwest of the airport. The CLAWS meteorologists provided the FAA air traffic controllers in the tower with daily forecasts of the probability of microbursts within 20 miles of the airport and issued an advisory warning whenever a microburst was observed within five miles of Stapleton.
The FAA recently established a task force, headed by McCarthy, that is assessing the FAA's needs for information on weather conditions in the vicinity of airports and how adequately those needs are being met by existing systems. The group will look at how aviation weather forecasts can be improved and how available technology can be transferred more quickly and effectively into practical application in operational aviation weather services.
Robert Serafin, director of the NCAR Atmospheric Technology Division, sees JAWS and CLAWS as a model for technology transfer. "A small pilot program like this, with researchers working closely and directly with operational people, can provide the most effective transfer of this kind of technology," he maintains. "It doesn't have to be big. Instead of conceiving a $500 million program to solve a problem like wind shear, and having to sell it to Congress and all the other people in Washington, why not set up small, effective programs at a few airports, and demonstrate how well the technology works?"
|A dark, towering thunderhead like this one can spawn severe hail, powerful winds, torrential rain, and sometimes killer tornadoes.|
Serafin believes that the important thing is for researchers to be actively involved. "There's a feeling that researchers don't know how to interact with operational people, and that the operational people don't have time to listen to the researchers," he says. "But we established a good working relationship, even in the very busy environment of the Stapleton control tower, and we maintained good, close cooperation. The result was a substantial improvement in operational efficiency and safety at the same time that we were demonstrating the effectiveness of the technology."
STORM, CCOPE, and JAWS/CLAWS illustrate three important ways in which UCAR and NCAR bring the resources of the nation's atmospheric science community to bear on national needs and concerns. In the STORM planning process, UCAR marshalled the scientific knowledge and experience of the community to plan a major national research effort designed to advance an important frontier of atmospheric research. CCOPE was a precisely orchestrated field experiment, conceived and carried out by a partnership of NCAR, government, and university scientists, that provided invaluable data to advance fundamental knowledge of a major atmospheric phenomenon. And JAWS/CLAWS tested an innovative approach to moving research technology into the operational sector where it can be applied to save lives and improve efficiency in the world of commercial aviation.