Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 71
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences 3 The Changing Context for Atmospheric Science A significant evolution and growth of the atmospheric sciences has occurred since the first National Academies’ review of the status of research and education in the field (NAS/NRC, 1958). The expansion of university, private-sector, and National Center for Atmospheric Research (NCAR) research and the development of new communications and computational infrastructure, coupled with greatly expanded research and operational efforts at other agencies and in other countries, has transformed understanding of the atmosphere, created new operational observational and modeling capabilities, and changed the way in which atmospheric research is conducted. The expansion of the field has also led to significant achievements and scientific discoveries with direct societal benefits, such as decreased economic losses in a number of private sectors due to improvements in weather predictions. New subdisciplines of atmospheric science have emerged, such as climate change and atmospheric chemistry, which grew out of an increased awareness of air pollution. The number and size of university atmospheric science programs has increased by nearly a factor of five, indicative of a more comprehensive and richer research endeavor. NCAR has grown to an institution that houses about 935 scientists and support personnel, builds and maintains observational and modeling facilities, and serves as a leader in organizing field campaigns, educational and outreach activities, and other community service efforts. Federal agencies other than the National Science Foundation (NSF), including the National Aeronautics and Space Administration (NASA), National Oceanic and Atmospheric Administration (NOAA), Department of Energy (DOE), Department of Defense
OCR for page 72
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences (DoD), and Environmental Protection Agency (EPA), have added to the support for atmospheric research, both internally and extramurally. These other agencies have focused efforts on their own missions and supporting research objectives (e.g., air quality) and have pioneered new approaches to research, most notably the introduction by NASA of space-based platforms for observing the atmosphere and near-space environment. International collaborations, including large multi-investigator and multinational field campaigns, now play a major role and require a significant fraction of the research budget. In this chapter, some aspects of the evolution of the atmospheric sciences from 1958, when the National Academy of Sciences (NAS) first considered the status of research and education activities in the field, to the present are analyzed. While illustrative rather than comprehensive, this consideration of a number of key factors that influence the field—including the broader intellectual and societal context, demographics, and technology developments—has helped inform the committee’s thinking about what factors are important in shaping future directions for the atmospheric sciences. INTELLECTUAL AND SOCIETAL CONTEXT During much of the 20th century, atmospheric scientists focused primarily on issues of weather, greatly expanding our understanding of the physical dynamics of the lower atmosphere required for weather forecasting. In the early years, the Navy, Department of Agriculture, the Army Medical Department, the Smithsonian Institution, the Signal Office, and other government programs supported research to develop accurate weather predictions for storm forecasting, aviation, and agriculture (Fleming, 1997). Indeed, the 1959 “blue book” report of the University Committee on Atmospheric Research (“UCAR,” 1959) that presented the scientific rationale for the establishment of a large national atmospheric sciences research center focuses on atmospheric physics topics relevant to meteorology, balanced by a recognition of cross-disciplinary research avenues such as aeronomy, atmospheric chemistry, and the possible impact of atomic weapons detonations on the atmosphere’s electrical structure. Basic and applied research in meteorology over the past several decades has contributed to remarkable advances in knowledge of the atmosphere, discoveries of relevance to scientific inquiry more broadly (e.g., the discovery of chaos theory, as described in Box 3-1), and greatly improved abilities to forecast atmospheric conditions. Atmospheric science has been deeply rooted in practical applications since its inception, so that the need for research to meet societal expectations and to lead to progress in operations has long been an organizing principle. Indeed, it is striking that many of the topics highlighted in the
OCR for page 73
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences 1959 NAS/National Research Council (NRC) report Proceedings of the Scientific Information Meeting on Atmospheric Sciences remain among the major focus areas for research and development, such as improvement of understanding and methods related to weather forecasting, pollution and its health effects, fire risk, droughts, agriculture, erosion, and water management, to name a few. Although a few topics identified in the 1959 report have so matured through technical advances that continued research is not as prominent a feature of the scientific landscape as it was in the past, these are the exceptions. Further, a number of topics have been added to the menu of societal concerns, particularly seasonal-to-interannual climate forecasting, global change, space weather, and atmospheric dispersion of chemical, nuclear, and biological contaminants. The range of products that are needed and expected by an ever more engaged and broader public continues to expand and deepen, building upon the successes and development since the 1950s. It seems apparent that the public’s interest in gaining access to the information relating to these topics has increased rapidly. Today’s citizen makes greater demands on research to deliver a far larger number of user-oriented products. Examples include urban air quality forecasts, agricultural forecasts tailored to specific farming areas or crop types, as well as lightning detection systems to assist in fire risk evaluation. New warning systems, such as online access to hurricane and tornado forecasts, are also among the products that now enjoy large constituencies due to the availability of the Internet as well as the greatly improved capacity of scientists to provide increasingly accurate and ever-faster response information, enhancing public safety. These are only a few examples of the many types of products that reflect the ever-increasing pace of application of research to operations and products (NRC, 1998b). Society also expects more finely tailored and wider ranges of information, provided in terms understandable to a broad audience. As scientific information and understanding have deepened on topics such as atmospheric pollution and climate change, there has been a far deeper appreciation of the policy relevance of atmospheric science for societal decision making (e.g., Box 3-2). Indeed, the findings of atmospheric science have provided the cornerstones for policy measures such as the Clean Air Act, the Montreal Protocol on Substances that Deplete the Ozone Layer and its subsequent amendments, and the Kyoto Protocol. Public interest in understanding how such policies work, the basis for their application, and the impact they will have has led to an increasing demand for research organizations to provide summaries aimed not just at the policymaker and other scientists, but to a far broader range of audiences, including the public, local and state governments, industry, and the education sector. Addressing the broader impacts of research beyond advancement of knowledge has been an important thrust of NSF in recent years. All NSF
OCR for page 74
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences BOX 3-1 Meteorology, Chaos Theory, and Edward Lorenz The Atmospheric Sciences have contributed many discoveries of great importance to the overall fabric of science, but perhaps none has had the impact on physics and mathematics as the introduction of chaos theory by Professor Edward Lorenz. The following discussion is taken largely from Chapter 3 of Lorenz’s (1993) excellent popular book, The Essence of Chaos.This discussion is an interesting example of the interplay among a large international field program, large-scale numerical modeling, and the brilliant work of a single investigator. In the early 1960s, several leading meteorologists (see Charney et al., 1966) were planning for one of the largest atmospheric field programs, the Global Atmospheric Research Program (GARP), that has been executed to date.The concept was to show how improved meteorological data would enable better weather forecasts to be made. Among the original aims of GARP was the goal of enabling two-week weather forecasts.Jule Charney was concerned, however, that the feasibility of making a two-week weather forecast might be proven impossible before the first such forecast was even attempted.During a special conference in Boulder, Colorado, Charney convinced all of the global atmospheric circulation modelers to perform experiments in which pairs of numerical forecasts were made starting from slightly different initial conditions.These experiments showed that the doubling time for small errors in the intial conditions was about 5 days. Professor Edward Lorenz investigated the mathematical basis of such behavior using both idealized mathematical models (Lorenz, 1963) and state-of-the-art numerical weather prediction models (Lorenz, 1982). Lorenz (1963) discussed how the time evolution of a physical system may be described by a trajectory in a multidimensional phase space. He introduced an idealized set of three seemingly simple nonlinear equations in three variables as an idealization of the equations for fluid convection and showed that for suitably chosen parameters, the solutions were unpredictable in the grant proposals are evaluated in terms of their broader impacts, which include educational objectives, broadening the participation of underrepresented groups, enhancing the infrastructure for research, wide dissemination of research results, and benefits to society. The NSF-wide small-center programs (i.e., Science and Technology Centers and Engineering Research Centers) have placed even more emphasis on education and outreach. As the 20th century closed, the content and context of the atmospheric sciences had expanded dramatically. In addition to discussing the major issues facing meteorology, the NRC’s report The Atmospheric Sciences— Entering the Twenty-First Century (NRC, 1998b) highlights the challenges of improving and maintaining air quality, protecting and improving ecosystems, sustaining the stratospheric ozone layer, understanding and man-
OCR for page 75
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences sense that two solutions with arbitrarily close initial conditions soon departed from one another so that after some time there was no way to identify that the two solutions were at an earlier time very close to one another in phase space. The implications of Lorenz’s work were enormous for meteorology. It became clear what the mathematical limits for deterministic weather forecasting were, and the goal for deterministic weather forecasting was from that time on to see how much improvements in weather prediction models and observational systems could produce forecast skill for times within the theoretical limit for predictability. It also set the stage for ensemble weather forecasting which is the present state of the art in which several forecasts are run that span the uncertainties in both initial conditions and physical parameterizations so that we are not only predicting the weather but also predicting the confidence in the spread of the model predictions. The implications of Lorenz’s work were not confined to meteorology. Chaos theory found its way into mathematics and physics and other fields, and books with titles such as From Clocks to Chaos (Glass and Mackey, 1988), Chaos: Making a New Science (Gleick, 1987), Order Out of Chaos (Prigogine and Stengers, 1984), and Chance and Chaos (Ruelle, 1991) appeared showing the generality of the concept. Clearly, Lorenz’s work had its beginnings in meteorology and had great impact on the field, but it also had far-ranging implications in areas beyond the atmospheric sciences. It was the result of a brilliant individual performing his own research, but its motivation was from a much larger community problem, whose importance was well recognized. It should be noted that the Lorenz (1963) paper has an acknowledgement of Air Force funding for this research, but would the Air Force fund such work now? It is worth asking whether any agency other than the NSF Division of Atmospheric Science would be funding this type of research today. aging climate variability and global change, characterizing the near space environment, and developing the ability to predict space weather. The central role of the atmospheric sciences in the addressing challenges of global environmental change was also addressed in a massive 1999 NRC report Global Environmental Change—Research Pathways for the Next Decade (NRC, 1999b). Both of these landmark reports emphasize the close coupling between atmospheric properties and processes occurring in the oceans, on land surfaces, in the near-space environment, and on the sun. They clearly demonstrate that atmospheric scientists need to collaborate closely with a wide range of colleagues from the physical, biological, Earth, and space sciences to meet the challenges facing atmospheric scientists during the 21st century.
OCR for page 76
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences BOX 3-2 Keeling’s CO2 record Charles D. Keeling, 1928–2005 Excerpts from Annual Review of Energy and Environment, 1998, 23:25–82 “When I began my professional career, the pursuit of science was in a transition from a pursuit by individuals motivated by personal curiosity to a worldwide enterprise with powerful strategic and materialistic purposes.The studies of the Earth’s environment that I have engaged in for over forty years, and describe in this essay, could not have been realized by the old kind of science.Associated with the new kind of science, however, was a loss of ease to pursue, unfettered, one’s personal approached to scientific discover. Human society, embracing science for its tangible benefits, inevitably has grown dependent on scientific discoveries. It now seeks direct deliverable results, often on a timetable, as compensation for public sponsorship.Perhaps my experience in studying the Earth, initially with few restrictions and later with increasingly sophisticated interaction with government sponsors and various planning committees, will provide a perspective on this great transition from science being primarily an intellectual pastime of private persons to it present status as a major contributor to the quality of human life and the prosperity of nations.” “In 1953, I complete a dissertation on polymers under Dr. Dole, taking what was then the extraordinarily long period of five full years. I had acquired a working knowledge of geology, weak in laboratory and field work….” “Although I hardly grasped it then, the opportunities for new Ph.D.s were at nearly an all-time peak. There had been a shortage of Ph.D. chemists ever since the recent world war…. I was offered employment by large chemical manufacturers, most of which were located in the industrialized cities of the eastern United States…. In more recent times it would have been risky to pass up such good job offers…. I accepted an invitation from Professor Harrison Brown of the California Institute of Technology RESEARCH SUPPORT The atmospheric sciences have enjoyed a slow but steady increase in funding by NSF since the late 1950s. NSF funding for atmospheric sciences was $16.3 million (in constant 1996 dollars) in 1958, increasing to $53.9 million in 1959. The ATM budget had increased to $122 million by 1972, reaching $196 million in 2004 (Figure 3-1). Much of the budget increase that ATM has experienced since the 1980s can be traced to new funds for facilities operated by entities other than NCAR ($27 million increase since 1982) and for NSF-wide priorities, such as “Biocomplexity in the Environment” and “Information Technology Research” ($25 million increase since 1989). The core grants program and NCAR have experienced modest increases in support over the past 30 years.
OCR for page 77
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences in Pasadena, California, where he had recently started a new department in Geochemistry. I became his first postdoctoral fellow.” “With Professor Brown’s consent, I postponed the study of uranium in granite and set about building a device to equilibrate water with a closed air supply. I acquired a hand-operated piston pump. Through a nozzle it could spray water from a natural source onto the wall of the glass chamber to bring about a thermodynamic equilibrium between the carbon dioxide dissolved in the stream of water and gaseous CO2 in the chamber…. I did not anticipate that the procedures established in this first experiment would be the basis for much of the research that I would pursue over the next forty-odd years.” “The highly variable literature values for CO2 in the free atmosphere were evidently not correct. Rather a concentration of 310 ppm of CO2 appeared to prevail over large regions of the northern hemisphere. I had detected this near-constancy under the implausible circumstances of studying air in old-growth forests where variability was to be expected.By 1956 my broader findings of surprising near-constancy seemed to me secure enough to communicate them to others…” “The consumption of fossil fuel has increased globally nearly three-fold since I began measuring CO2 and almost six-fold over my lifetime.” These early studies of atmospheric CO2 concentrations by Charles D. Keeling in California provided a foundation for the establishment of the now famous Mauna Loa CO2 measurements.The autobiographical text from Dr. Keeling is provided as an example of how atmospheric science has evolved, and underscores the importance of working across scientific disciplines and investing in transformative and sometimes risky work for individual investigators. This funding is currently directed to the modes of support including core grants, university facilities, NCAR facilities and science, and NSF priorities, as shown in Figure 1-2. These modes overlap in many ways, for example, because facilities are integral to the research process. Over these 30 years, core research has decreased from 50 to 38 percent of the overall ATM budget, and support for science at NCAR has decreased from 23 to 18 percent of the ATM budget. However, given the overall increase in the ATM budget, NSF core grant support has remained about constant in total dollars. At the same time, support for facilities at NCAR and at universities has increased from 23 to 33 percent of the ATM budget. Thus, facilities support has increased faster than core grant support, most likely due to the increasing sophistication of computing and observing capabilities. The com-
OCR for page 78
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences FIGURE 3-1 ATM funding for the atmospheric sciences since FY 1972 in constant 1996 dollars. The NCAR numbers include support for both science and facilities housed at the center. “Other facilities” refers to support for facilities operated by institutions other than NCAR. “Priorities” refers to NSF-wide initiatives, such as “Biocomplexity in the Environment” and “Information Technology Research.” mittee notes that the availability of facilities creates research opportunities for individual investigators. Many of the advances in the atmospheric sciences have been enabled by the availability of sophisticated, and expensive, facilities. These include supercomputers, research aircraft, and high-power radar systems, so it is not surprising that during the last 30 years, the fraction of ATM funding devoted to facilities has grown from 23 to 33 percent of its budget. Very valid arguments can, and will, be put forth for ATM purchasing bigger, and more expensive, computers and very valuable, and expensive, observing facilities in the coming years. At the same time, if ATM’s budget is not increasing faster than inflation, the funds to purchase and maintain these facilities will have to come from the research budget. Thus, ATM will be presented with a dilemma of how to make the trade-off between investments in “tools” at the expense of funding people when both will be necessary to generate and implement ideas. Ideally these difficult decisions
OCR for page 79
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences FIGURE 3-2 (Left) FY 2004 funding for weather and space weather research by the 10 agencies surveyed by the OFCM. The overall funding by these agencies for this year totals about $503 million. Note that the NSF funding only includes the foundation’s contributions to space weather research and the U.S. Weather Research Program, which together total about $14 million. The NASA proportion of the OFCM funding is composed of the estimated meteorology share of the supporting research and analysis programs as well as Earth Observing System (EOS) and Earth Probe instruments, EOS science, and the EOS Data Information System elements of the NASA Office of Earth Science budget (OFCM, 2004). (Right) Estimated FY 2003 budget for atmospheric-related climate change research (i.e., the atmospheric composition, climate variability and change, carbon cycle, and water cycle program areas) by the 13 agencies of the CCSP (CCSP and SGCR, 2004). Total funding for these program areas is approximately $1.4 billion. will be made in the framework of ATM’s strategic planning and with input from the broad atmospheric sciences community. Other agencies have experienced much larger fluctuations in their extramural funding for atmospheric science. It is not easy to track down exactly how much each agency spends on atmospheric research; Figure 3-2 shows efforts by the Office of the Federal Coordinator for Meteorology (OFCM) and the U.S. Climate Change Science Program (CCSP) to sum up the contributions of different agencies to research relevant to their individual mandates. Note that the agencies also support research on air quality and solar sciences, which neither of the charts in Figure 3-2 includes. These budgets include both intramural and extramural support for research. ATM is a relatively small player overall, but plays a significant role in supporting university and other extramural research. In the last 5–10 years, NASA and
OCR for page 80
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences FIGURE 3-3 Annual extramural funding for basic research in the atmospheric sciences at DoD and NASA (NSF, 2004). DoD have decreased their support for basic research in the atmospheric sciences (Figure 3-3). This reduced funding for basic atmospheric research by other federal agencies will likely cause more and more of the community to turn to ATM for basic research funding. The committee believes that this evolution of atmospheric science research since 1958 introduces not only new opportunities but also new challenges. For example, five years of steady growth in NSF budgets have given way to a new period of limited budget growth, while support of atmospheric science research by other federal agencies exhibits considerable volatility. The constrained budget environment combined with the expanded scope of scientific questions have increased the need for interagency and international coordination. DEMOGRAPHICS NAS (NAS/NRC, 1958) concluded that there was a strong need for more professionals in the atmospheric sciences. At the time, only about 10 to 15 doctorates were awarded each year. By the late 1970s, an average of 84 doctorates a year was awarded by a greatly expanded number of university atmospheric sciences departments in the United States, meeting
OCR for page 81
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences the needs for professionals in the field at that time (http://www.ametsoc.org/EXEC/TenYear/figs.html). Table 3-1 provides a number of indices for the growth in the atmospheric sciences research community. The table illustrates the significant expansion of educational efforts, professionals, and research funding over the past four decades, although it is difficult to pin down the exact size of the community because of its diversity. The size of the research workforce in atmospheric and related sciences seems to have been leveling off since the 1990s because of lower interest in the physical sciences, the growth of research programs overseas, and movement of some of the Ph.D. population to the private sector (Hoffer et al., 2001; Vali et al., 2002). On average, 133 atmospheric science doctorates were awarded annually in the late 1990s. The number of applicants to atmospheric science graduate programs declined between 1995–1996 and 1999–2000 (Vali et al., 2002), but increased slightly as of the 2002–2003 academic year, the most recent year with available data (Vali and Anthes, 2003). In the coming years, there is a projected shrinkage of the science and engineering research labor pool through retirements (NSB, 2002) coupled with a projected growth in science and engineering career opportunities. It is not clear exactly how these broad trends for the physical sciences and engineering might impact the atmospheric sciences. Note that not all atmospheric scientists are trained in atmospheric science, meteorology, astronomy, or Earth science departments. In particular, atmospheric chemists and cloud/aerosol microphysicists may be enrolled in chemistry, physics, applied science, chemical engineering, aerospace/ mechanical engineering, civil/environmental engineering, or public health programs. Solar physicists, aeronomers, and other near-space scientists may be trained in astronomy, physics, chemistry, or electrical engineering departments. Those who study marine meteorology or interactions between the atmosphere and the ocean may enroll in marine science departments. ATM supports research in all of these academic enclaves. Along with efforts to increase the size of the atmospheric sciences workforce, the meteorological community has worked to make the production and communication of weather information more professional (NRC, 2003b). Private-sector meteorology began in earnest in this country shortly after the end of World War II, when several thousand meteorologists trained to support the massive aviation activities of the U.S. armed forces left government service eager to apply their newly acquired skills (Mazuzan, 1988). The Weather Bureau made the decision to permit its weather data to be used by the emerging private sector, and the first group of private meteorological companies began operating in 1946. The emerging television industry was a natural outlet for weather information and forecasts, and the decision by the Weather Bureau that government employees would not provide television weathercasts prompted the development of an influ-
OCR for page 90
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences BOX 3-3 A History of Field Experiments Joachim Kuettner, Chair Atmospheric Sciences and International Research, UCAR/NCAR Ph.D., Physics (Meteorology), University of Hamburg, Germany My career in science has spanned much of the last century. My passion has been to combine both science and flying. Beginning with a study of the newly discovered lee wave in Germany’s Riesengebirge Mountains, I obtained a doctorate in physics from the University of Hamburg in 1939. I have worked with numerous aircraft for research, including low- and high-altitude, motorized and gliding aircraft. In pursuing this love of flight, I established several world and national gliding records, including an absolute altitude gliding record for high altitude (43,000 ft [13 km]; still the German gliding high-altitude record). Before becoming involved in many field projects associated with NSF, I joined Wernher Von Braun’s team at NASA’s Marshall Space Flight Center in Huntsville, Alabama (1958–1965), and became the Center’s Director of the “Mercury-Redstone” Project (the first space flights of the U.S. astronauts, Alan Sheppard and Virgil “Gus” Grissom). Subsequently, I was Director of the Apollo Systems Office, responsible for the integration of the Apollo spacecraft and the Saturn-V rocket for the lunar landing.After I left Huntsville I became the Chief Scientist at the NationalWeather Satellite Center inWashington;and in 1967, Director of Advanced Research Projects at NOAA in Boulder, Colorado. The many atmospheric researchers who have flown on research aircraft have learned that you don’t really know the Earth’s atmosphere until you have experienced it personally in flight. For example, during the Monsoon Experiment (MONEX,1979), the Electra penetrated the monsoon front over the Arabian Sea about two days prior to landfall in India. Our colleagues from India who had devoted a lifetime to monsoon studies were almost delirious at seeing the inner structure of the phenomenon of their life’s work.In addition to facility support by NSF, whenever I needed a break as scientific director, I would invite our NSF program managers, Jay Fein (summer MONEX in India) and Dick Greenfield (winter MONEX in Malaysia), to take over the project management and the planning meetings.They did a remarkable job in the field.
OCR for page 91
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences Prior to MONEX, in the early 1970s, all member countries of the World Meteorological Organization (WMO) agreed to implement GARP, the Global Atmospheric Research Program. Its first field project, the GARP Atlantic Tropical Experiment (GATE) in 1974, was a huge undertaking involving almost 4,000 participants from 70 nations.The USSR participants were particularly thrilled to engage in the active, open exchange of ideas that characterized GATE. The NSF’s participation in the GATE experiment was very important to its success. Among the many worries that I had as the WMO-appointed head of GATE was the daily deployment of the flexible observing systems, such as 13 aircraft and 39 ships, for a host of competing scientific objectives. Should the decisions be made in a more military fashion, by a “czar,” or more democratically, by a mission-planning team? Could such a team act in the short time available—usually about one to two hours? Through simulations of mission planning for various, sometimes surprising, scenarios, conducted at NCAR and attended by the lead scientists from several nations (United States, U.S.S.R., United Kingdom, France, and Germany), we created a congenial and efficient “Mission Selection Team” that set the standard for practically every international experiment since.Following GATE, I planned and directed for WMO the aforementioned Monsoon Experiment.That was followed by ALPEX, the Alpine Experiment, which explored the airflow over and around mountain ranges. Since 1985, my home base has been the National Center for Atmospheric Research (NCAR) and the University Corporation for Atmospheric Research (UCAR) in Boulder, Colorado, where I have been associated with many major field projects, such as GALE (Genesis of Atlantic Lows Experiment, 1986), TAMEX (Taiwan Area Mesoscale EXperiment, 1988), THERMEX (Thermal Wave Experiment, 1989), TOGA-COARE (Coupled Ocean–Atmosphere Response Experiment, Australia, 1992), CEPEX (Central Equatorial Pacific EXperiment, 1993), INDOEX (Indian Ocean [aerosol] Experiment, 1999), MAP (Mesoscale Alpine Program, 1999), and T-Rex (Terrain-induced Rotor Experiment, 2005–2006). In 1994, NSF created a Distinguished Chair for Atmospheric Sciences and International Research at UCAR/NCAR, which remains my current position. The Central Equatorial Pacific Experiment (CEPEX), mentioned above, was a good example of the close cooperation in the field between the NSF Program Director and the scientific community. CEPEX focused on surface temperature regulation in the western Pacific, and was led by Veerabhadran Ramanathan. Since Ramanathan had never directed a field project, Jay Fein suggested that I help him lead the experiment. Ramanathan and I were quickly able to communicate on the same wavelength and lead the project together.This collaboration among Ramanathan, Jay Fein, and myself has continued to this day, and has led to the development of unmanned aerial vehicles to study the role of aerosols in cloud formation and climate over the ocean. Looking back on a long professional life, it appears that I started my atmospheric research on mountain waves and rotors, and have just completed a research project on the same subject, the T-Rex project.In the 1930s, I was puzzled by observations of rotors and hope to have found some answers through the 2006T-Rex project.It should be mentioned that T-Rex was the first operational project for the new NCAR aircraft, HIAPER, recently acquired through NSF’s efforts.
OCR for page 92
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences structural mechanics, and thermal control; (2) the necessity to minimize size, weight, power consumption, consumable gases and chemicals, and cryogens; (3) the requirement of reliable field calibration. Further, it was noted that development time scales to move an innovative technique from laboratory proof-of-principle to a reliable field measurement tool was typically seven to ten years, a time scale difficult to manage with typical two-to three-year research grants, normal graduate student and junior faculty time frames, or private company time-to-market constraints. Finally, the 1990 report went on the recommend that agencies supporting atmospheric chemistry research to “encourage good, innovative instrument development proposals” and “that these projects can be viewed as a key R&D portion of an atmospheric research program and should be a significant (10 to 15 percent) of each agency’s overall budget.” It also recommended that federally funded laboratories with ongoing instrument development programs be encouraged to form partnerships with university and private-sector laboratories, noting that such arrangements might encourage students to take on instrument development projects because they would be collaborating with successful instrument-oriented professionals (NRC, 1990). The need for innovative observational tools was also highlighted in the NRC Board on Atmospheric Sciences and Climate’s report The Atmospheric Sciences—Entering the Twenty-First Century, which listed the development of new observational capabilities as an “Atmospheric Science Imperative.” That report stated, “the federal agencies involved in atmospheric science should commit to a strategy, priorities, and a program for developing new capabilities for observing critical variables, including water in all its phases, wind, aerosols and chemical composition, and variables related to the phenomena in near-Earth space, all on spatial and temporal scales relevant to forecasts and applications” (NRC, 1998b). In addition, a 1999 report prepared under the direction of the NRC’s Climate Research Committee focused on needed upgrades in the climate observing system (NRC, 1999a). This report called for agencies involved in the U.S. Global Climate Research Program (USGCRP) to “establish a funded activity for the development, implementation, and operation of climate specific observational programs” as a way of “providing essential additional capability to operational observing systems.” This activity would include the identification and measurement of “critical variables that are either inadequately or not measured at all.” NSF/ATM has played a leading role in funding global atmospheric chemistry, meteorology, aeronomy, and climate change research; including a key role in the USGCRP. While other agencies, such as NASA, NOAA, and various DoD agencies are assigned the lead role in developing satellite remote sensing systems, NSF/ATM has played a strong role in developing in situ as well as ground and airborne remote sensing observational
OCR for page 93
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences tools. However, the main NSF funding paradigm of grants to individual academic investigators is often not consistent with the wide skill sets and long time scales required for successful observational tool development and deployment. Given the ongoing need for innovation in observational instruments and systems, detailed above, additional effort may be required. The committee notes that traditional, non-ATM-specific NSF instrumentation activities can be useful. For instance the Major Research Instrumentation program now has an instrument development component, and, as noted in our interim report (NRC, 2005e), the NSF Small Business Innovation Research and Small Business Technology Transfer Research programs have produced valuable atmospheric science instruments. Instrument development partnerships among interested university groups, private sector organizations, and large government and federally funded research and development center laboratories could also be an effective way to access the full range of scientific and engineering skills and experience with field measurement requirements necessary for successful instrument and measurement systems development. Training in Observational Tool Development and Utilization A key to continued success in the earth sciences is the continued access to high-quality observations, use of multiple observation datasets, and the ongoing development of new tools. The next generation of weather radar technology, the development of new space-based instruments, and the growing sophistication associated with data processing, visualization, and analysis requires educating the next generation of scientists and engineers who are equipped with the knowledge to integrate earth science problems with engineering observational solutions. The nature of instrument development and prototyping is changing and that requires more partnership and more involvement with the private sector. It has been shown that universities increasingly are not investing in education programs in the observational aspects of the science for various reasons (Serafin et al., 1991; Takle, 2000). The Takle article provides a concrete set of recommendations for enhancing university instruction in observational techniques. A summary of potential actions includes: Provide opportunities for faculty and students to participate in NCAR and other instrumentation development programs and encourage active engagement in such programs Consolidate and leverage the COMET and Unidata resources as well as other instrumentation and observational educational materials for classroom use at a broad array of higher education institutions
OCR for page 94
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences Seek opportunities to collaborate with other universities and with other geoscience and environmental science programs and departments in the development and implementation of instrumentation courses Prepare videos and electronic media on specialized instrumentation and platforms for use in college and university education programs Develop Web-based materials or supplements that can be shared and used at other colleges and universities Provide field program fellowships and opportunities for students to obtain hands-on experience with instrumentation While there are opportunities for undergraduate and graduate students to participate in NSF-funded research projects that utilize or develop observational tools, it is rarer to have courses that provide the concepts of engineering design or data processing. The challenges of providing such education and training at a single university noted in the above articles may be overcome if a more community approach is used in the development of course materials that serve as modules and use of information technology for the delivery of courses. This approach provides for the opportunity to leverage existing materials not only in universities but also those developed at the national center (COMET and Unidata) and through the AMS. In addition several workshops or “schools” that present engineering and observational tool fundamentals have been developed and implemented (e.g., NCAR workshops, International Radar School) but have not been propagated into more formal coursework in our university curricula. The development of good online material that can be shared nationally as well as select fieldwork sites that encourage hands-on engineering internships for students should be the topic of an NSF-sponsored collation effort between atmospheric science and engineering departments, the national center, the AMS, and other federal laboratories that engage in observational tool development. INFORMATION TECHNOLOGY AND COMPUTATIONAL MODELING The extraordinary evolution in information technology over the last 50 years has had a huge impact on the atmospheric sciences. Roughly speaking, computational capability has advanced at nearly a hundred-fold per decade throughout the entire time period. Associated with this are increases nearly as great in internal memory, data storage, and data transfer. The advent of the Internet has served to connect the community in unprecedented ways, and presently allows practical exchange of vast amounts of information. These changes have allowed an entirely new dimension of research—that of simulation and prediction—to join theory, observation,
OCR for page 95
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences and analysis in underpinning the science. Numerical weather, climate, and air pollution “experiments” may now be conducted in an environment that can be controlled in ways not available naturally, and in great numbers compared with what can be observed in nature. Computational models allow a new means to learn as well as a new means to harness existing knowledge, toward the development of the best possible operational and research products. For example, efforts are now under way to improve seasonal climate predictions and make them relevant to numerous real-world applications (e.g., Box 3-4). The field of data assimilation has emerged and is just starting to fulfill its promise of improving prediction (see also Box 3-4). By the 1980s, the models were good enough to “accept” specially targeted observations (e.g., see Case Study 2 in Chapter 2) to improve hurricane forecasting. Modern data assimilation lies at the intersection of analysis and simulation, and is a critical part of both research and operational prediction. It is one of the most demanding and resource-intensive aspects of modern weather prediction. Beginning with the Fronts and Atlantic STorms EXperiment (FASTEX) in 1997, several field programs have investigated the impact on model forecasts of observations focused on locations found by adjoint models to produce the most forecast error. In the near future, enhancing use of satellite data in model-specified geographic areas to increase the certainty of forecasts will become even more promising because of easier availability and faster response time. The concept of climate “reanalyses,” that is, analyses of past observations using current models and assimilation methodologies optimized to represent climate parameters, is relatively recent but has provided extremely important products for research, despite known difficulties. The use of computer simulations as a tool to understand the Sun and the space environment has grown markedly in the last two decades. Models have been developed to study the aspects of the solar interior and to predict the intensity of the sunspot cycle (see Case Study 9 in Chapter 2). Simulations of the Earth’s magnetosphere and its interaction with the solar wind are now able to reproduce real events and, in the future, will be able to provide predictions of space environmental conditions. The NSF has funded a Science and Technology Center, the Center for Integrated Space-Weather Modeling that is developing a set of coupled codes extending from the surface of the Sun to the upper atmosphere of Earth. Techniques developed in tropospheric weather modeling, particularly data assimilation, are increasingly used in space physics. For example, the Space Environment Center assimilates the total electron content over the United States in near-real time to make predictions, and the solar-cycle model of Dikpati et al. (2006) uses sunspot-intensity data from the previous three solar cycles. ATM has also supported community access to space physics models by providing partial
OCR for page 96
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences BOX 3-4 Atmospheric Reanalyses and Dynamical Seasonal Climate Prediction Jagadish Shukla, Professor George Mason University Ph.D., Banaras Hindu University, Geophysics The NSF is unique among the federal agencies that fund academic research because it has the flexibility to entertain and support highly innovative basic research ideas that might be considered high-risk in the mission-driven agencies. I can think of no better example of that approach than the Atmospheric Sciences Division, with which I have worked for more than 20 years.In my own case, there have been several occasions when I have brought ideas to the NSF and always received a respectful hearing and an indication of interest. The following two instances are examples of many success stories that have been the result of the flexibility and vision of NSF and its program officers. Reanalysis In the mid 1980s, I became convinced that it was possible and essential to produce a reanalysis of data representing the four-dimensional nature of the global atmosphere for the past half-century. The technology available at that time was adequate to analyze observations of the global atmosphere using a state-of-the-art data assimilation system such as was in use for operational Numerical Weather Prediction (NWP) at the European Centre for Medium-RangeWeather Forecasts (ECMWF) or the U.S.National Meteorological Center (now the National Centers for Environmental Prediction, NCEP). I was also convinced that such assimilation would be far superior to the analyses made in real time over the period since NWP was initiated in the 1950s. The data archives were adequate to uniquely define the dynamic and thermodynamic state of the near-surface and upper atmosphere every day for the period since rawindsondes were in routine use, which began about the time of the end ofWorldWar II.I started a campaign to persuade the centers involved in NWP around the world to consider undertaking such a task. After several unsuccessful attempts to get the backing of the operational agencies, I approached the Atmospheric Sciences Division of NSF (Jay Fein) with the idea of conducting a pilot project as a proof-of-concept for reanalysis. The reviewers of the proposal were impressed and we received funding to produce a multiyear reanalysis of observations that proved to be demonstrably superior to the analyses available from the real-time NWP archive. The results eventually appeared in a 1994 article. Consequently, NCEP, in partnership with NCAR, and ECMWF had become convinced of the value of reanalysis for atmospheric research purposes and for the improvement of their own NWP skill.The adoption of reanalysis as a methodology enabled a huge
OCR for page 97
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences number of new capabilities for NWP. It ensured the continuing integrity of the observational data archive—whose longevity was in grave danger due to dwindling curatorial resources—and led the publication of hundreds of scholarly papers that pushed the boundaries of our understanding of atmospheric dynamics and physics.All this can be ascribed to the wisdom of the NSF in seeing the value of such an enterprise. Dynamical Seasonal Climate Prediction As early as the late 1970s, I began to suspect that, despite the chaotic nature of day-to-day atmospheric fluctuations we normally ascribe to “weather,” the Earth’s climate might be predictable beyond the so-called deterministic limit due to slowly varying conditions and processes in the climate system. We began to explore this possibility with admittedly crude global atmospheric models at that time, and found encouragement in the relationship between variations in the sea surface temperature and the large-scale atmospheric circulation. Later, we found that there was a potential predictive relationship between land surface variations and the atmospheric circulation and precipitation on seasonal time scales. A firm scientific basis was needed for exploring and quantifying this seasonal predictability.Hand-in-hand with that academic question, there was a real nuts-and-bolts issue of how to exploit that predictability. At that time, seasonal predictions were made entirely on an empirical basis. We approached NSF, along with NOAA and NASA, to consider establishing a national research center to explore, understand, and quantify seasonal and longer time scale predictability and to develop the capability for regular dynamical seasonal prediction. Again, NSF took the lead, and the other agencies responded with enthusiastic support.After a rigorous peer-review process, the project was launched. The result of that and subsequent proposals was the establishment of the Center for Ocean-Land-Atmosphere Studies, which has thrived for over 20 years. Ensuing research revealed great disparity among the results obtained by the various modeling groups that explored the question of seasonal prediction.As before, NSF recognized the importance of a national program to critically and quantitatively compare the various models and supported us in establishing the Dynamical Seasonal Prediction program. A sister program called PROVOST (PRediction Of climate Variations On Seasonal and inter-annual Timescales) was established in Europe. More recently, after the World Climate Research Program announced a new strategy for Coordinated Observation and Prediction of the Earth System (COPES), the Atmospheric Sciences Division of NSF has taken the leadership role in the United States to engage the scientific community and other federal agencies in serious discussions about this emerging strategy.
OCR for page 98
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences support to the Interagency Community Coordinated Modeling Center, where users can request specific model runs and visualize the results. With the enormous successes have come significant challenges. More and more, numerical weather forecast models are assimilating increasing amounts of satellite data, with assimilation of radar data promising to improve short-term convective storm forecasting in the near future. Furthermore, use of numerous runs of the same model with slightly different initial conditions (“ensembles”) or combining ensembles of runs with different models (“superensembles”) are not only improving forecasts but also allowing an evaluation of model uncertainty and development of probabilistic predictions. The climate challenge is even more significant, because simulating future climates needs to involve the interaction of the oceans, the Earth system, and the cryosphere, as well as the atmosphere and the Sun. Policymakers are demanding such runs be urgently produced at higher resolution than presently feasible (NRC, 1998a, 2001b), with effective horizontal grid spacing of 5 km or less. As in the case of weather forecasting, the use of ensembles and superensembles is essential to estimate uncertainty, and to generate probabilistic seasonal predictions and climate change predictions. Meeting the demand for improved climate modeling capability will likely require substantial increases in computational resources. For several decades, the continuing rapid development of computer capability has enabled the ATM community to more or less meet its computing needs, but this seems no longer to be the case. Advancements in simulation, data assimilation, and prediction capabilities have in recent years begun to place serious demands on existing computational resources— demands that, increasingly, are not being matched by new investments. As pointed out in several reports (NRC, 2001b, 2004), this problem is growing, and is in fact more acute within the United States than in many other countries. For example, Japan, Germany, and the EU continue to make new investments in computing resources for the Earth sciences that match advancements in scientific capabilities. Those considering the future needs of the community call for computers so large and with such significant cooling and power requirements that the decades-old solution of housing the NCAR supercomputer in the NCAR Mesa Laboratory is not going to be feasible in 10 years (Kellie, 2004). NSF, realizing the approaching challenge for the geosciences and other areas of science that rely on supercomputing, has issued a Request for Proposal (RFP) to support the development of a petascale computing environment (NSF, 2005b). UCAR coordinated a response (Ad Hoc Committee and Technical Working Group for a Petascale Collaboratory for the Geosciences, 2005). Also, while one can significantly increase computing power by combining multiple processors, the “latency” or time lag in
OCR for page 99
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences communicating among the processors is a significant barrier to increasing computational speed (NRC, 2001b). In the meantime, through a contract with IBM, NCAR has worked hard to improve its computer infrastructure to keep up with the demand, and climate scientists in both NCAR and the university community have gained access to other computing facilities, including the Japanese Earth Simulator and the DOE Leadership-Class Computing Facility (LCF) at the Oak Ridge National Laboratory. Computational scientists have joined in the development of community models, such as the Weather Research and Forecasting Model and the Community Climate System Model. As the scientific boundaries between atmospheric and geophysical and astrophysical sciences become more blurry, there will be benefits from pooling computing and intellectual resources. NRC (2001b) calls for national coordination in the form of “common modeling infrastructure” to facilitate model improvements and data formats to streamline the research process. NSF (2006) calls for a GEO computer and NRC (2005a) calls for multi-agency investment and minimizing barriers to international collaboration. An additional challenge is storage and analysis of the enormous amount of data produced by the runs. Unless adequate storage is provided, in conjunction with computational capability to allow detailed analysis of the data, much of the research value is unrealized. NSF is to be commended for putting out an RFP for analysis of the products of recent climate assessment runs. Looking to the future, it can be anticipated that the gap between research needs and available computational resources is going to become even wider unless action is taken to enhance such resources very substantially. NRC (2001b) makes the case that about 100–1000 times more computational capacity is required to meet existing needs than was available at the time. NRC (2004) emphasizes that a sustained program of support and development is needed to meet future needs. This issue is obviously larger than what can be addressed by ATM alone, and is arguably larger than what can be addressed by NSF alone. The committee concurs with the position taken by NRC (2004) that the government agencies that are the major users of supercomputing must take joint responsibility for the strength and continued evolution of supercomputing infrastructure in the United States, and that adequate and sustained funding must be allocated in the national budget. It has long been recognized that strong computing facilities are of primary importance for advancing the atmospheric sciences. That need remains today and may be greater. How best to direct future investments in computing resources for the atmospheric sciences is a complicated issue that requires more careful study than possible in this report. Nonetheless, the committee is convinced that good science and important social impact
OCR for page 100
Strategic Guidance for the National Science Foundation‘s Support of the Atmospheric Sciences would be enabled by better, faster models, which require more and more powerful computers. Supporting ever-larger and more capable computing infrastructure should be a high priority, but must be balanced by the other needs of the community, so as not to jeopardize maintaining observational facilities, and, especially, continued support in basic research. Meeting this demand will not likely be possible with the approaches used today and may require new organizational mechanisms, sources of funding, and partnering with other agencies, the private sector, or other nations.
Representative terms from entire chapter: