UCAR > Communications > UCAR at 40 Search

Measuring and Modeling
  • Measuring and Modeling: timeline
  • In their own words: Joachim Kuettner
  • In their own words: T.N. Krishnamurti
  • 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


    Portable and solar-powered, NCAR's fleet of automated mesonet stations collect weather data around the clock with only occasional maintenance from technicians.

    In the 1920s, atmospheric science leaned on largely the same equipment it had in the 1720s. Wind, temperature, pressure, and humidity could be measured at a particular spot, but not from a distance. Balloons could sample the air above the ground, if only for a brief up-and-down flight. Scientists had managed to infer the physical laws that drive how weather systems grow, move, and die. To go further, though, they needed far more observations and a way to process them. In 1922, British physicist L.F. Richardson imagined a "forecast factory," a vast room filled with clerks performing calculations to predict the weather.

    By the time UCAR and NCAR were conceived, technology had begun to catch up with scientists' dreams, but society had not. Radar, satellites, rockets, balloon-borne radiosondes, computers, and instrumented aircraft had come onto the scene. Yet no single university had the funds to provide all these tools to its researchers. And many government studies of weather and climate were carried out with minimal involvement from academia.

    Gradually, UCAR and NCAR helped fill the void. Balloon, computing, and aircraft facilities were established within a decade, all for the express use of university and NCAR scientists. As a university-governed organization with centralized facilities available to academic institutions, UCAR fulfilled a critical national need to support both basic and applied research.

    NCAR's computational resources bolted forward in 1965 with the commissioning of a long- awaited Control Data 6600 machine.

    NCAR has retained its unique role as a home and testbed for some of the most advanced technology used by atmospheric scientists. In the past decade, that role has encompassed remote sensing based on sophisticated radars and Global Positioning System (GPS) satellites, as well as computing based on distributed processors. UCAR is now exploiting the World Wide Web to organize field projects and catalog data in new ways. With funding from NSF and other sponsors, all of these facilities support a long-held goal: No matter where they are, scientists facing the biggest problems in atmospheric research will have the best tools for the job.

    The crunch of numbers

    Few disciplines demand as much from a computer as does atmospheric science. The microchips of a computer are the discipline's laboratory—the only place where scientists can tinker with the atmosphere in freedom and double-check their results. Numerical models—the software that produces weather and climate in the laboratory of a computer—have become more complex each year. In turn, their depictions of the atmosphere have sharpened, like binoculars coming into focus.

    NCAR's computing center was founded in 1964 with a machine far less powerful than today's home computers. The facility has since grown to become one of the largest in the world serving atmospheric scientists. In the early days, most users were at NCAR's Mesa Laboratory. Now some 1,000 researchers around the globe access NCAR's supercomputers each year. A panel made up of NCAR and university specialists carefully allocates computer time.

    Starting in 1976, NCAR oriented its computing facility around machines produced by Cray Research. These vector supercomputers featured high- powered processors acting in sequence to run programs with increasing speed. As the landscape of computing technology diversified in the 1990s, NCAR experimented with a variety of other platforms while continuing to rely on Cray machines. In 1999, NCAR acquired an IBM RS/6000 SP as its flagship computer, heralding a new approach: symmetric multiprocessors. Instead of running a single sequence of operations at ever-faster speeds, the IBM is made up of 167 sets of four processors each. Every set has its own memory and shares data with other processors through a special protocol. By the next year, with this and other upgrades, NCAR had more than tripled the amount of computing power available to its university users.

    This IBM SP supercomputer, acquired in 1999, has helped NCAR triple the computing power it makes available to university users.

    NCAR computing has always provided clients with a place to keep the vast amounts of data generated by their work. The center's mass storage system includes some 250 terabytes (trillion bytes), the equivalent of nearly a million compact discs. The archives can be accessed in minutes with the help of NCAR staff and automated silos that store and retrieve tapes and disks. Along with a destination for data, the center also provides a starting point. Teaming with Eugenia Kalnay (now at the University of Maryland) and others at the National Centers for Environmental Prediction, NCAR's Roy Jenne led a group that conducted an exhaustive retooling of 50 years of global weather observations, the data that jump-start computer models. The NCEP/NCAR global reanalysis was released in 1998 after nearly a decade of work to remove systematic errors and fill data gaps. Dozens of university research projects already have made use of this improved picture of a half-century's climate.

    Model output is brought to life through NCAR's visualization laboratory. Its high-end software can portray a data set in five dimensions: for example, two elements (such as temperature and wind) can be followed through time and three dimensions of space. A sophisticated blend of visual supercomputers and virtual-reality technology, the "viz lab" allows researchers to explore massive data sets and turns equations into visible structures—some not discernible until visualized. The lab follows in the tradition of the NCAR Graphics software package and other tools that help university and NCAR scientists view their work clearly and vividly.

    Flying high

    If computers are an ever-shifting tool in the study of the atmosphere, aircraft are a constant. Since NCAR's aviation facility was founded a few miles east of Boulder in 1964, ten aircraft have flown over 4,000 missions for NCAR and the university community, observing the atmosphere over six continents.

    Twice each year, scientists apply for use of the two versatile aircraft currently owned by NSF and maintained by NCAR's skilled, attentive staff. The fleet has evolved in line with aviation developments and the needs of research. Chemistry experiments, for example, require a large cargo capacity to handle heavy and bulky instruments. NCAR's C-130 Hercules, acquired from the U.S. Navy in 1993, carries up to 12,000 pounds and can fly up to 2,900 nautical miles on a single trip.

    Sometimes NCAR's planes are modified to suit a particular project; other refinements are for the long haul. Among the biggest of the latter is the Electra Doppler Radar (ELDORA), which brings the power of Doppler wind-tracking technology to the skies. Built in partnership with the French government, ELDORA was installed in 1992 on the Electra, the NCAR airplane that complements the C-130. ELDORA's dual-scanning technology is among the most advanced in the handful of airplanes worldwide that support Doppler weather radar. Roger Wakimoto (University of California, Los Angeles) used ELDORA in 1995 to analyze a tornadic column of wind that extended nearly to the top of a thunderstorm—something scientists had never been able to observe before.

    Along with instruments developed in-house, the NCAR air fleet carries sensing tools created at universities. In 1999 the Electra flew over the North Pole bearing the Iron Boltzmann Temperature Lidar, a new instrument built at the University of Illinois by Chester Gardner and colleagues. The lidar sends light toward tiny particles of iron far above the plane and measures the returning signal. The lidar has yielded the first temperature measurements from the world's coldest air mass: the summertime polar mesosphere, more than 50 miles (80 km) above the Arctic and Antarctic, where readings can drop as low as -140°C (- 220°F).

    In the field

    When scientists need to blanket a region with instruments, the fabric is often woven from NCAR equipment. Since the 1970s the center has produced three generations of its Portable Automated Mesonet (PAM) station. The compact, solar-powered device collects standard weather data near the ground every five minutes. It uses satellite and modem signals to report back from some of the world's harshest environments: Arctic tundra, Mongolian desert, the swamps of Florida. In the early 1990s NCAR built a smaller group of Flux-PAMs to measure heat and moisture migrating from the air to the ground and vice versa. By 2000 all PAMs were upgraded to measure other types of fluxes, along with soil moisture and soil temperature.

    Radiosondes have been a mainstay of upper-air sensing for over 60 years, but NCAR's are among the few that can be launched from the road. Six NCAR-based vans and trailers can deploy radiosondes wherever needed. NCAR's instrument specialists work with others in the United States and Europe to incorporate the latest compact electronics in their sonde packages. In the mid-1990s, the mobile radiosonde units moved to GPS technology, as did NCAR's GPS dropsonde. (See "When Weather Matters Most" for more on the dropsonde and on the S-Pol radar, another major NCAR facility available to universities.) UCAR scientists have also discovered ways to infer the presence of water vapor from the bending of GPS signals as they approach ground-based receivers. (See "Knowledge for All" for more on this and related techniques.)

    To make sensing as trouble-free as possible, from ground level to miles high, NCAR bundles PAM stations and mobile sondes as part of four integrated systems. They also include two ground-based instruments that can scan upper levels of the atmosphere remotely: a profiler to measure winds, and a radiometer to sense temperatures.

    Paving the way for science

    After managing NCAR's field observing facilities in the 1970s and 1980s, Robert Serafin served as NCAR director from 1989 to 2000.

    UCAR provides far more than equipment to help get international, multi- institutional field programs up and running. A single experiment can take years of planning. Permissions have to be obtained from governments and landowners. Dozens—sometimes hundreds—of people must be housed. Someone needs to track each participant's contribution and how all of these roles will dovetail. Starting with the Line Islands Experiment in 1967, UCAR has taken on these and many other related tasks for major field projects each year. In the last decade alone, the Joint Office for Science Support (JOSS), part of the UCAR Office of Programs (UOP), has handled logistics involving roughly 50 nations and some 200 universities and laboratories.

    UOP remains involved well beyond the life of many field programs. JOSS's CODIAC software houses and catalogues data collected from the field using an innovative, Web-based interface. The amounts of data can be vast. For a single European campaign in 1999, the archive reached one trillion bytes, a volume that might have boggled the countless minds in the forecast factory dreamed up by L.F. Richardson 80 years ago.


    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