UCAR > Communications > UCAR at 40 Search

When Weather Matters Most
  • When Weather Matters Most: timeline
  • In their own words: Richard Reed
  • UCAR at 40
    Who We Are
    Introduction
    One Planet, One Atmosphere
    Between Sun and Earth
    Measuring and Modeling
    •When Weather Matters Most
    Spreading the Word
    Knowledge for All
    Looking toward the Future
    UCAR at a Glance
    List of acronyms


    Johnstown, Pennsylvania, the site of a flood in 1889 that killed over 2,000 people, was again hit by deadly floods in 1977. (Photo by Chuck Manula, The Tribune-Democrat Johnstown, Pennsylvania)

    As the 20th century came to a close, two disasters confirmed that the United States is not quite weatherproof. The first billion-dollar tornado on record ravaged the Oklahoma City area on May 3, 1999, killing 38 people. Four months later, Hurricane Floyd triggered the nation's largest peacetime evacuation and brought catastrophic floods to North Carolina. Floyd's death toll of 56 was the largest related to a U.S. hurricane since Agnes in 1972.

    Both of these storms could have been far more deadly were it not for accurate forecasts made possible by years of research and development. One of the first goals of UCAR and NCAR was to better understand the weather features now referred to as mesoscale (roughly a county to a state in size). In the center's early years, this meant wringing as much as possible out of limited data and relatively crude computer models. By the 1990s, the challenge was how best to incorporate a torrent of new data types into ever more refined models and what to make of the weather features that finer-scale observations and models were uncovering. All the while, NCAR and its collaborators have sought insight into the atmospheric building blocks—radiation, turbulence, cloud physics—that affect mesoscale weather as well as global climate.

    Untwisting tornadoes

    NCAR itself has never issued weather predictions for the public, but ideas and technologies nurtured at the center have played an important role in the forecasts and warnings people rely on each day. Much of this benefit reached citizens in the 1990s as the National Weather Service (NWS) carried out a long-planned modernization. Doppler radar was installed nationwide, allowing forecasters to track winds as well as rainfall. NCAR helped pave the way with key Doppler radar research that began in the 1970s, led by David Atlas and Robert Serafin Doppler radars have helped the NWS to double its average lead time for tornado warnings and to quantify rainfall with far more precision. With its S-Pol radar, NCAR is now refining dual-polarization techniques that allow hail to be discerned from rain. The new hardware and software is expected to join the NWS Doppler network in the next several years.

    Along with a 45-minute warning of the Oklahoma City twister's approach, residents were alerted several hours ahead that violent tornadoes were possible. The advance outlook stemmed from an understanding of the ingredients that drive long-lived supercells and other destructive thunderstorms. A team of scientists including NCAR's Joseph Klemp, Morris Weisman, and Richard Rotunno, as well as Robert Wilhelmson (University of Illinois at Urbana-Champaign), forged a theory in the 1980s of why supercells rotate, based largely on vertical variations in wind (or wind shear). The leap from supercell to tornado remains mysterious: less than a third of rotating supercells spawn twisters. In the mid-1990s, NCAR joined other institutions for a large experiment in the southern Great Plains that trained ground-based and airborne Doppler radars and vehicle-mounted weather stations on severe storms. The results, still being analyzed, indicate that subtle variations in temperature and wind can tip the atmospheric balance toward or away from tornado formation.

    The secrets within storms

    With help from NCAR, the prototype for the nation's fleet of Doppler weather radars was installed and tested near Denver in 1989.

    Other storm threats—less dramatic than tornadoes but dangerous in their own right—are being clarified through research. Mesoscale convective systems (MCSs) can sprawl across one or more states, producing torrential rains. Dozens of MCSs over the summer of 1993 led to catastrophic Mississippi River floods. Christopher Davis and Stanley Trier later showed that a vortex lying at the heart of an overnight MCS can trigger another MCS the following night.

    Hurricane diagnosis improved dramatically in the mid-1990s with the advent of the Global Positioning System dropsonde. Like its predecessors, the sonde was developed at NCAR by Terry Hock and colleagues and is used by the U.S. Air Force and the National Oceanic and Atmospheric Administration (NOAA) as a tool to go where humans cannot. The compact instrument package collects and radios back weather data after being dropped from aircraft into the heart of a tropical cyclone. Thanks to its GPS precision and upgraded electronics, the new sonde provides data every 5 to 10 meters (16–33 feet) as it drops, compared to every 300 meters (1,000 feet) before. When dropped en masse around a hurricane, the sondes can help computer models to analyze and predict upper-level winds that steer the cyclone.

    In its first 20 years, NCAR's computing prowess was aimed largely at global-scale problems. The center's role expanded with the arrival of the PSU/NCAR Mesoscale Model, originally developed at Pennsylvania State University in the 1970s by Richard Anthes and Tom Warner (both now at UCAR). With development now led by Ying-Hwa (Bill) Kuo, the model is in its fifth generation (MM5). It predicts fronts, cyclones, and other mesoscale features at data points separated by as little as a kilometer (0.6 mile). This resolution is over 20 times sharper than that of the national-scale models typically used by forecasters. The MM5 is freely distributed to over 600 scientists at more than 200 institutions worldwide. Support is provided on line and through NCAR workshops. Soon the MM5 will be supplanted by the Weather Research and Forecast (WRF) model. This new tool—a collaboration among NCAR, NOAA, and the University of Oklahoma—will allow researchers and forecasters alike to analyze thunderstorms and other mesoscale features with greater precision.

    Parachute-borne weather stations developed at NCAR probe the inner workings of hurricanes with far greater resolution than ever before. (Photo courtesy NASA)

    Much of the recent storm-related modeling and analysis work at NCAR, including the WRF, is tied to the U.S. Weather Research Program. The USWRP aims to better understand several critical pieces of the forecasting puzzle: heavy rainfall and flooding, the behavior of hurricanes at landfall, and the origins of North American winter storms in the poorly monitored Pacific Ocean. Scientists at NCAR and elsewhere concluded in the 1980s that only a concerted national push could adequately address these key problems. Their persistence led to the creation of the USWRP and its sponsorship, through 2004, of a series of field projects and small, focused studies.

    Forecasts are being improved not only through better physical understanding but also through analyses of how people receive and interpret predictions. NCAR political scientist Roger Pielke, Jr., found that warnings issued weeks ahead of spring flooding across North Dakota and Minnesota in 1997 were taken by users to be more precise than forecasters intended them to be. When the Red River crested some five feet higher than forecast, the surprise element only worsened the resulting disaster.

    Behind the scenes of weather

    Leading up to each flood or twister is an elaborate chain of physical causes and effects. Many NCAR scientists focus on quiet, often-unseen elements—pulses of wind, pockets of moisture—that combine to give us the weather we all notice.

    The atmosphere's main stage, the place where storms and sunshine affect society, is its lowest kilometer, the planetary boundary layer (PBL). As the atmosphere meets the land and the sea, heat, moisture, and momentum are transferred in complex ways, moment by moment. A forest absorbs and reflects sunlight far differently than does a parking lot. Cornfields trap, and later release, more rainfall than does grassland. A blossoming of new instruments in the past 20 years—including sonic anemometers and thermometers, lidars, and radiometers—has allowed scientists to peer into the PBL's workings as never before. One long-term network launched in 1989 by the U.S. Department of Energy spans the pastures and prairies of Oklahoma and Kansas. From this data, specialists from over 50 universities and laboratories, including Margaret LeMone, Donald Lenschow, and Jielun Sun, have learned much about PBL cycles and how to portray them in weather and climate models.

    Residents of rural Missouri collected hailstones in the spring of 1975 for Project Dustorm, a collaborative project to examine thunderstorms and hail formation.

    Over the ocean, where long-term instrument networks are hard to sustain, several massive experiments have studied air-sea exchange using ships and aircraft. Over 20 nations, 100 institutions, and 1,000 people, including scientific leaders Peter Webster (University of Colorado, or CU) and Roger Lukas (University of Hawaii) were part of a landmark 1992–93 study (see timeline) in the western Pacific, the breeding ground of El Niño.

    Before sunlight reaches the PBL, it passes through a swarm of clouds and chemicals. NCAR's microphysicists study clouds at their roots by flying through them, creating them in the laboratory, or simulating them in small-scale computer models. A particular focus has been cirrus clouds, sheets of ice thrown off by the tops of thunderstorms. The ice crystals in cirrus help warm the Earth's climate by trapping outgoing radiation. As the crystals fall, they can help "seed" rainfall or snowfall below. Recent field studies have helped to nail down the crystals' descent rates and their impact on weather. Ice crystals are also key to lightning production, the focus of studies in 1996 and 2000 across northeast Colorado. The first study, led by chemist Brian Ridley and cloud physicist Jim Dye, showed that lightning within a thunderstorm top can produce significant amounts of nitrogen oxide, a precursor to ozone, well above ground level.

    The stealthy swirls of turbulence

    To learn how local weather and wildfires interact, NCAR researchers are combining fire-scale models with observations from infrared cameras. (Photo courtesy U.S. Forest Service Research)

    Turbulence is an inherent element of fluids in motion, including air, but its results—such as downed airplanes—can be devastating. Ever since NCAR was founded, turbulence has tested the ingenuity of its researchers and collaborators. The late Philip Thompson, Jackson Herring, John Wyngaard (now at Pennsylvania State University), Cecil (Chuck) Leith (now at Lawrence Livermore National Laboratory), Douglas Lilly (now at the University of Oklahoma) and other leaders carried out pioneering work using theory, observations, and computer models. Sometimes the goal has been to clarify what can't be known. Chaos theory holds that tiny, turbulent shifts in air flow can lead to huge changes in weather, making the prediction of individual storms beyond a few days impossible. This problem was identified in the 1960s by Edward Lorenz (Massachusetts Institute of Technology), a frequent NCAR visitor for much of the center's history. Other scientists simulated turbulence in computer models, enabling such progress as the discovery of vortex tubes by Eric Siggia and Stuart Patterson.

    A breakthrough in turbulence modeling came in the 1970s at NCAR, when Lilly and James Deardorff (later at Oregon State University) created the large-eddy simulation (LES) technique, widely used in and beyond the atmospheric research community. Instead of modeling each atmospheric swirl, LES statistically summarizes the effects of smaller-scale eddies on the larger air flow. In one of the most extensive applications of LES to date, Chin-Hoh Moeng and colleagues at NCAR and several universities mapped the behavior of marine stratocumulus. These giant sheets of cloud cover much of the ocean. Because they block sunlight and trap outgoing heat from Earth, marine stratocumulus strongly affect the global radiation budget, a key variable in climate change. The studies have shown, among other things, how turbulent eddies of drier air can engulf and erode the tops of stratocumulus, altering the cloud layer's structure and its important radiative effects.

    In addition to turbulence, mountains help spawn many other atmospheric phenomena, from devastating rains to dramatic wave clouds. NCAR used its mountainous backyard as a testbed in the early 1970s for research flights, many organized by Edward Zipser, to uncover the mechanisms behind downslope windstorms. The work helped trigger a series of observational, theoretical, and modeling studies, including an international field campaign held in and near the Alps in 1999.

    The steady march of computing speed and power has not only brought long- term goals within reach, it has also made once-unimagined research practical. In the last few years, NCAR's Terry Clark, Janice Coen, and Lawrence Radke have been studying the wildly turbulent growth of forest fires, working closely with the U.S. Forest Service, Monash University, and CU. Clark and Coen are using an atmospheric model with a resolution of less than 100 meters (330 feet). By pairing this model with others of vegetation and burn patterns, and adding the atmospheric preconditions, the researchers have teamed with NCAR visualization expert Don Middleton to show how fingers of flame emerge from a fire line and how fire- generated winds help fuel a blaze. Someday, they hope, firefighters might download such visualizations on portable computers—charting the progress of blazes even as they battle them.


    UCAR at 40
    Who We Are
    Introduction
    One Planet, One Atmosphere
    Between Sun and Earth
    Measuring and Modeling
    •When Weather Matters Most
    Spreading the Word
    Knowledge for All
    Looking toward the Future
    UCAR at a Glance
    List of acronyms


    Other UCAR PublicationsUCAR | NCAR | UOP

    Executive editor Lucy Warner, lwarner@ucar.edu
    Prepared for the Web by Jacque Marshall
    Last revised: Fri Jan 26 17:18:32 MST 2001