TOGA accomplished much in its decade, especially in observing, understanding and predicting ENSO in the tropical Pacific. However, many questions about ENSO and other types of interannual variability around the globe, especially in the middle latitudes, remain unanswered. We recommend for the future: maintenance of observing systems; creation of an institute for developing applications of short-term climate forecasts; and a program for continued research on seasonal-to-interannual climate variations and their predictability.
TOGA went far towards fulfilling its goals—perhaps further than anyone could have foreseen at the beginning of the program. There were many successes. In particular, researchers associated with TOGA:
conducted process experiments, especially the TOGA Coupled Ocean-Atmosphere Response Experiment (COARE), and processed, distributed, and archived the resulting data sets;
built and maintained the TOGA Observing System and developed the new technology involved in TOGA Tropical Atmosphere/Ocean (TAO) array of moorings;
made the data from the TOGA Observing System widely available, in real time, through electronic and paper distribution;
developed theories for El Niño and the Southern Oscillation (ENSO), using coupled models;
developed methods of ENSO prediction and demonstrated predictive skill through the establishment and maintenance of the TOGA Program on Prediction (T-POP) and the establishment of the coupled-modeling project at the National Meteorological Center;
developed effective management and advisory structures to maintain and support the TOGA effort, both nationally and internationally; and
planned for an International Research Institute for Climate Prediction.
All these accomplishments contributed to a successful program, well balanced and integrated among theory, modeling and predicting, and observations.
TOGA was able to achieve these successes in part because it concentrated on the strongest interannual climate variation: ENSO in the Pacific. This concentration came about for both scientific and programmatic reasons. Scien
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 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program 7. THE FUTURE TOGA accomplished much in its decade, especially in observing, understanding and predicting ENSO in the tropical Pacific. However, many questions about ENSO and other types of interannual variability around the globe, especially in the middle latitudes, remain unanswered. We recommend for the future: maintenance of observing systems; creation of an institute for developing applications of short-term climate forecasts; and a program for continued research on seasonal-to-interannual climate variations and their predictability. TOGA went far towards fulfilling its goals—perhaps further than anyone could have foreseen at the beginning of the program. There were many successes. In particular, researchers associated with TOGA: conducted process experiments, especially the TOGA Coupled Ocean-Atmosphere Response Experiment (COARE), and processed, distributed, and archived the resulting data sets; built and maintained the TOGA Observing System and developed the new technology involved in TOGA Tropical Atmosphere/Ocean (TAO) array of moorings; made the data from the TOGA Observing System widely available, in real time, through electronic and paper distribution; developed theories for El Niño and the Southern Oscillation (ENSO), using coupled models; developed methods of ENSO prediction and demonstrated predictive skill through the establishment and maintenance of the TOGA Program on Prediction (T-POP) and the establishment of the coupled-modeling project at the National Meteorological Center; developed effective management and advisory structures to maintain and support the TOGA effort, both nationally and internationally; and planned for an International Research Institute for Climate Prediction. All these accomplishments contributed to a successful program, well balanced and integrated among theory, modeling and predicting, and observations. TOGA was able to achieve these successes in part because it concentrated on the strongest interannual climate variation: ENSO in the Pacific. This concentration came about for both scientific and programmatic reasons. Scien
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program tifically, the engagement of the challenges in the Pacific proved so fruitful that once these were addressed, the TOGA community maintained its enthusiasm as oceanographic, atmospheric, and coupled-system problems were attacked. Funding and support were always only barely adequate, so focusing the resources on a limited area of the Pacific became a necessity. WHAT TOGA DIDN'T DO TOGA's limited resources were focused on studying interannual variability in the tropical Pacific associated with ENSO. Consequently, other problems of importance, either ones known at the beginning of TOGA or issues that were identified during the program, remained unexplored. Although the focus on ENSO was consistent with the initial U.S. plan for TOGA (NRC 1983), the program did not cover the broader initial objectives for the international program (WCRP 1985), which were agreed on by U.S. scientists (NRC 1986). It is well known from empirical studies that the global impacts of ENSO are strongly controlled by the annual cycle, and, in fact, can be thought of as modulations of the mean annual cycle. The very concept of anomalies requires the annual cycle to be known. While the climatology for sea surface temperature is well characterized, the climatologies of other basic quantities in the tropical Pacific—thermocline depth, surface currents, subsurface currents, and subsurface temperature—are still largely unknown, and will be determined only through many years of measurements. Furthermore, the annual cycle itself has components that result from strong coupling among ocean, atmosphere, and land processes. The skill of predictions of sea surface temperature in the tropical Pacific has proven to be strongly dependent on which seasons lie between the time at which and the time for which a prediction is made. Hence, further study is needed on the nature of the annual cycle and its strong variations around the globe, the impact of the annual cycle on interannual variability, and the predictability of interactions between the annual cycle and interannual variations. Even the nature of ENSO, the basic interannual variability in the tropical Pacific, was not completely determined by TOGA. A ten-year program simply was not long enough to define the nature of ENSO's interannual variability. The nature of interannual variability outside the tropical Pacific remains also underexplored. For example, interannual variability of the strength and onset of the Asian monsoon and the midlatitude connections to ENSO present many unsolved problems. Much remains unknown about seasonal-to-interannual variability induced by midlatitude interactions on large scales between the atmosphere and oceans; induced by the interactions of the atmosphere with sea ice, snow, and land; and induced by the high-frequency forcing by the atmos
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program phere on each of these slow reservoirs (the so-called Hasselmann (1976) mechanism). In considering these issues, it will undoubtedly prove true that the radiative effects of natural and anthropogenic aerosols in producing seasonal-to-interannual variability, and in affecting such variability produced by other physical mechanisms, will have to be considered (see NRC 1996). Interactions between events in the tropics and extratropics, and the possibility of predicting midlatitude climatic variations, provided some of the justifications for the TOGA Program. However, some issues did not receive the attention they deserved from the TOGA Program. These issues include the role of tropical sea-surface-temperature anomalies in perturbing the extratropical atmosphere, the generation of midlatitude sea-surface-temperature anomalies through this interaction, and the regions of the globe where skillful forecasts of tropical sea surface temperature can be translated into useful regional forecasts. They have been addressed to some degree by other independent efforts, but still need significant attention. We have come to understand that the warm phase of ENSO can be identified with the eastward extension into the central Pacific of a significant portion of the precipitation normally found in the far western Pacific (the “maritime-continent heat source”). The global circulations arising from the thermal forcing from this gigantic heat source influence other tropical heat sources, in particular the South Pacific Convergence Zone, the Inter-Tropical Convergence Zones, and the heat source over South America. These in turn influence the maritime-continent heat source. Their interactions propagate ENSO influences throughout the entire tropics. A basic question that needs to be addressed is: What influences the position, strength, and interactions among all of the tropical heat sources on monthly to interannual time scales? The interactions can take many forms. The Asian monsoon interacts with ENSO. Interactions of ENSO with the South American heat source influence South American rainfall. ENSO affects the Inter-Tropical Convergence Zone in the Atlantic and, consequently, rainfall in northeast Brazil. ENSO also interacts with the Inter-Tropical Convergence Zone in the eastern Pacific, with attendant interactions between the atmosphere and surface conditions over South America, Mexico, and the United States. All these interactions, which occur relatively slowly, are influenced by the patterns of sea surface temperature and by land processes. The relative slowness of the evolution of the system offers the possibility of being able to predict sea surface temperature using coupled atmosphere-ocean models. In turn, sea surface temperature is a boundary condition for models in which the interactions between the land surface and atmosphere are more accurately simulated, e.g., using nested-grid high-resolution regional models. It has become clear that there are other modes of interannual variation besides ENSO. Land wetness induces interannual variations of weather, inde
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program pendent of connections to ENSO (see, e.g., Delworth and Manabe 1989). The middle-latitude ocean, atmosphere, and land system would have interannual variability in the absence of ENSO (Manabe and Hahn 1981, N.-C. Lau 1981). Whether these variations are predictable and whether they are enhanced by signals from the tropics remain crucial questions. In addition, the interactions of higher-frequency fluctuations in the tropics and middle latitudes with seasonal-to-interannual climate variations, and the predictability of the resulting climate variations, need to be assessed. Shortly before TOGA started, a clear picture of the “canonical” evolution of ENSO (Rasmusson and Carpenter 1982) was generally believed. However, after the 1982-83 El Niño, the TOGA decade began with a knowledge that the Rasmusson and Carpenter description could not be correct. As TOGA ends, the picture is still not clear. From 1990 to 1994, the climate system appeared to be locked in an unusually extended warm state. The NINO4 region (160°E to 150°W, near the equator) showed persistent positive (0.5–1.0°C) anomalies of sea surface temperature, and the Southern Oscillation Index was negative for that entire period. Such conditions are unprecedented in the instrumental record (Trenberth and Hoar 1996). Is this the result of interactions between interannual and interdecadal variability, or is it a manifestation of global warming? Clearly the nature of these interactions needs to be better understood and quantified. The key processes that are essential in the coupling of the ocean and atmosphere on seasonal-to-interannual time scales need to be better identified and parameterized in models. The COARE field program was designed to address some of these processes, but was conducted so late in the TOGA Program that it had little influence on prediction and modeling during TOGA. COARE analysis is scheduled to proceed for several years beyond the TOGA time frame, and it is likely that improvements in cloud, surface flux, and boundary layer parameterizations will be made. These studies no doubt will point to additional improvements needed in the convective and mixing parameterizations of coupled ocean-atmosphere-land models. As models become more comprehensive, and as practical forecasting experience is gained, the observing-system requirements for initializing the forecast system will require refinement and re-evaluation. The crucial upper-ocean and land-surface variables requiring initialization need to be identified, and routine observations of these need to be implemented. OBSTACLES TO PROGRESS The development of short-term climate prediction and its supporting research enterprise is fraught with difficulties. The cooperation and collaboration of diverse communities is required because ENSO is a coupled atmosphere-ocean phenomenon. Progress in climate prediction will depend on additional coupling
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program in models of the atmosphere and ocean to the land and ice. Development of this complex enterprise will put strain on the existing institutional structures. Compartmentalization Traditional departmental structures on college campuses serve to maintain excellence in basic scientific disciplines (e.g., physics, chemistry, biology, geology, and mathematics). However, these structures can discourage study and research on climate problems, which cross disciplinary boundaries. Only the exceptional student can overcome the lack of flexibility in course requirements to develop a path for multidisciplinary climate studies. Only an exceptional faculty member can overcome a lack of institutional rewards for collaborative work. Some parts of the federal funding structure are organized around traditional disciplinary lines. This organization makes it difficult to plan cross-disciplinary ventures. The investment required for a serious effort to predict seasonal-to-interannual climate variability by means of quantitative physical models is modest in view of the potential economic benefits of skillful forecasts. There has been a lack of communication between scientists and decision makers in both the public and the private sectors. Most scientists lack knowledge of the day-to-day requirements of decision makers in climate-sensitive portions of the economy. Most decision makers believe that scientists are incapable of understanding and responding in a significant way to needs in the public and private sector. Building and Maintaining the Infrastructure The systematic development of a climate-observing system requires long-range planning and commitment by both scientists and governments. The social and political environment evolves on time scales shorter than many climate variations. The geographic coverage of the present in situ climate observation network continues to decline at a time when new initiatives are needed (NRC 1994a). Under stringent budgetary conditions, the building of infrastructure always suffers. A new area needs facilities and resources that are dedicated to its own problems. For short-term climate prediction, these include educational programs, supercomputers, facilities for maintaining the observations needed to initialize and evaluate models, and funding structures that recognize the need to support and coordinate both development and research. TOGA implemented a prototypical observing and prediction system for limited aspects of ENSO. Such an extensive in situ observing system was not
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program foreseen at the beginning of TOGA. As a result, the TAO part of the observing system came into existence very late in the program, almost as TOGA was ending. Thus, while TOGA designed and put an observing system into place, there was no time within the life of the program for evaluation of that system. There is a need to see what effects the system's meteorological observations have on the global forecasts produced by the world's (atmospheric) operational forecast units or what effects its oceanographic observations have on the regular and systematic production of forecasts of sea surface temperature in the tropical Pacific. Furthermore, recent analyses of the decadal variability of tropical sea surface temperature indicate that the observing system may have been designed with too narrow a spatial extent, and may neglect regions important for the understanding of ENSO and its connections to other parts of the globe. Much remains to be done for implementing a comprehensive system for seasonal-to-interannual climate prediction. The upper-air observing system seems to be decaying. The status of operational satellites is always in question, primarily because of the large resources required. Research satellites do not seem to have the continuity required for obtaining the long, homogenous records needed for the study of climate variations. The effects of the decline of the global observing system on TOGA was discussed in NRC 1992. While outside the direct control of the TOGA community, these problems existed for the duration of the TOGA Program and affected its ability to improve climate understanding and prediction. For future progress in the study of climate variations, it is essential to maintain what we already have, including the upper-air observing network, satellite altimetry, and the upper-ocean and surface-meteorological measurements made routinely in and over the ocean. We re-emphasize the main conclusions of NRC 1994a: Present TOGA observations should be continued. The single most critical effort to be sustained, because of its late establishment and because of the central importance in TOGA predictions of the fields it measures (tropical wind stress, sea surface temperature, upper-ocean thermal structure), is the full TOGA TAO array of approximately 70 moorings.” Furthermore, “Observations now in place to support prediction … or added later for this reason … should thereafter be sustained until such time as a serious study of their impact on predictions reveals them to be of marginal value or until a new and more cost-effective technique is demonstrably ready to replace them without disrupting or biasing the geophysical time series.”
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program AN INTERNATIONAL RESEARCH INSTITUTE FOR CLIMATE PREDICTION (IRICP) In response to the applications being developed for the countries of the tropical Pacific, the Intergovernmental TOGA Board urged the United States to take the lead in designing an institute for experimental short-range climate prediction. This institute would employ the scientific knowledge being developed about ENSO and work with the most affected countries to learn how to use these forecasts. The initially proposed goals for the International Research Institute for Climate Prediction (IRICP) (Moura 1992) are: to continually develop dynamically and thermodynamically consistent coupled models of the global atmosphere, ocean, and land, to serve as a basis for improved climate prediction; to systematically explore the predictability of climate anomalies on time scales out to a few years; to receive, analyze, and archive global atmospheric and oceanic data to improve the scope and accuracy of the forecasts; to systematically produce useful climate forecasts on time scales of several months to several years on global space scales; and to shape and augment these forecasts by incorporating additional physical, agricultural, economic, and other appropriate data, to the explicit social and economic benefit of national societies. An international meeting to codify the institute's mission and architecture was held during November 1995 in Washington, D.C. This meeting was a major step toward the institute's formal establishment. The intention is that the IRICP provide an international structure that combines worldwide research capabilities, operations, and applications expertise. Regional application centers would certainly play a central role in the interpretation and dissemination of forecasts tailored to local needs. The mission-oriented focus of the IRICP, will help realize the goals of improved predictions and the development of applications of those predictions, with obvious benefit to societies worldwide. In whatever final form it takes, an IRICP will surely become one of the most important legacies of the TOGA program. TOGA has shown that a well conceived and well executed research program can play a central role in both basic research and policy applications. The application of ENSO predictions for the benefit of the societies and economies of the tropical countries affected by ENSO has taken TOGA out of the realm of pure research. As a corollary, seasonal-to-interannual climate research can justify additional demands on the public treasury. The IRICP will be a multinational facility with strong connections to the international research community and to application centers in member nations.
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program It is designed to make the best possible research forecasts of global seasonal-to-interannual climate variations, initially in and around the tropical Pacific, and to distribute those forecasts to applications centers, interested researchers, and operational weather-prediction centers. Training of people from member nations in the nature and use of prediction systems, and tailoring forecasts for specific region, will be central tasks. The recent Seasonal-to-Interannual Climate Prediction Program (SCPP) proposal (NOAA 1994), as forwarded by NOAA in partnership with other U.S. agencies, is an evolution of the concepts expressed in the IRICP proposal. It is a broad vision of how the United States, both nationally through the National Centers for Environmental Prediction (NCEP, McPherson 1994) and in a multinational framework through the IRICP, would implement these concepts. The SCPP provides mechanisms for the next steps in developing experimental forecasts, in research, in observations, in data management, and in applications of climate information on seasonal-to-interannual time scales. Laying the foundations for a program that will systematically produce and disseminate climate forecasts for applications in societal and economic planning is central to the SCPP. Support for an institute to develop applications of climate prediction can be found in NRC 1994b and 1995b. We re-emphasize a recommendation of NRC 1995b: Establish an international research prototype prediction capability, including a focused facility (the proposed International Research Institute) and a supporting research program in order to accelerate the application of demonstrated prediction capabilities; secure multinational support for global scale observing systems and international research programs; and focus research to extend predictive capabilities and applications. GOALS AND CLIVAR Comprehensive coupled ocean-atmosphere-land models, and prediction systems based on such models, are in their infancy. Development of these models and systems require process studies and the maintenance of appropriate observing systems. The questions raised during TOGA point to the need to develop more comprehensive prediction systems and expand the scope of climate prediction to the entire globe and to longer time scales. The new international program for A Study of Climate Variability and Predictability, CLIVAR (see
OCR for page 132
Learning to Predict Climate Variations Associated with El Niño and the Southern Oscillation: Accomplishments and Legacies of the TOGA Program WCRP 1995), especially its seasonal-to-interannual component, the Global Ocean-Atmosphere-Land System (GOALS, see WCRP 1995 and NRC 1994b), is designed to address these problems. We re-emphasize the words in NRC 1994b: The ultimate scientific objectives of the GOALS Program would be to: understand global climate variability on seasonal-to-interannual time scales; determine the spatial and temporal extent to which this variability is predictable; develop the observational, theoretical, and computational means to predict this variability; and make enhanced climate predictions on seasonal-to-interannual time scales. The focus of the GOALS program is an assessment of the global interannual climate variation that can be understood, simulated and predicted…. It is proposed that the GOALS program be an important component of the Climate Variability and Predictability (CLIVAR) program, which is a broader new initiative of the World Climate Research Program[me] (WCRP) addressing the variability and predictability of the coupled climate system. TOGA opened the way to the future of seasonal-to-interannual climate prediction. These follow-on programs will further develop the means of predicting the climate for the ultimate benefit of humankind.