4
Elements and Growth of the Program

GOALS is envisioned as a 15-year program, running from 1995 to 2010. This period coincides with the lifetime of CLIVAR, the international program on climate variability and predictability being developed under the auspices of the WCRP.

ELEMENTS OF THE PROGRAM

Climate prediction on seasonal-to-interannual time scales is the focused objective of the GOALS program. In working toward this objective, GOALS would follow the successful example of TOGA by adopting these four major program elements:

  • Modeling,

  • Observations,

  • Empirical studies, and

  • Process studies.

Modeling that involves development and application of improved coupled ocean–atmosphere models for data assimilation and prediction is the unifying theme of the proposed GOALS program. Requirements of model initialization would help define the observational systems needed. Empirical studies would largely define the requirements of model improvement and would help evaluate which process studies are needed. These elements would provide the research framework for the development of operational climate monitoring and predictive systems.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis 4 Elements and Growth of the Program GOALS is envisioned as a 15-year program, running from 1995 to 2010. This period coincides with the lifetime of CLIVAR, the international program on climate variability and predictability being developed under the auspices of the WCRP. ELEMENTS OF THE PROGRAM Climate prediction on seasonal-to-interannual time scales is the focused objective of the GOALS program. In working toward this objective, GOALS would follow the successful example of TOGA by adopting these four major program elements: Modeling, Observations, Empirical studies, and Process studies. Modeling that involves development and application of improved coupled ocean–atmosphere models for data assimilation and prediction is the unifying theme of the proposed GOALS program. Requirements of model initialization would help define the observational systems needed. Empirical studies would largely define the requirements of model improvement and would help evaluate which process studies are needed. These elements would provide the research framework for the development of operational climate monitoring and predictive systems.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis PROGRAM GROWTH The range of GOALS activities would expand as knowledge expanded. The following sequential ordering of developments in program growth represents a best guess as to the way knowledge will improve. This sequence is determined by the scientific considerations summarized in Chapters 2 and 3 of this report. The GOALS program would proceed as a natural sequel to TOGA. It would concentrate initially on the ENSO phenomenon in the tropics. To improve skill in prediction, the interactions between the Indian Ocean (including the surrounding land masses) and the maritime-continent heat source would be examined. To exploit fully the predictability of the tropics, interactions with the Atlantic and its bordering South American and African land masses would next be considered. As the area of investigation expanded eastward over Africa, the interactions with the African land mass would have to be examined, and the observed decadal variations of rain in the Sahel would become of interest. The effect of the global tropics on the higher-latitude atmospheric circulation would become the next challenge. Finally, GOALS would examine the effects of higher-latitude ocean, land, and ice on seasonal-to-interannual climate predictability. As sound as the preceding outline of expected program growth may be, it is possible that flexible program management and continuous, routine assessment of program directions and achievements could result in adjustment to the ordering. For example, models are generally run in a global mode, and modeling breakthroughs can suggest new avenues of research. It may be possible to identify those qualities of the mean state that enhance or degrade the ability of anomalies to propagate out of the tropics into higher latitudes (see, for example, Trenberth and Branstator, 1992). If so, this would point to features of the mean state that must be measured and modeled correctly, and a major advance in midlatitude prediction could ensue. The rest of this chapter discusses the four program elements listed above. A section on each describes the expected contribution of that component to the overall program, together with the challenges and objectives in that particular area. MODELING The overall strategy for the GOALS modeling component is to work within the framework of the existing WCRP programs on climate modeling and to build on the success of the TOGA program. The

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis observational and process study elements of GOALS would, of necessity, be regional and would expand initially from the TOGA region of focus—the tropical Pacific—to the global tropics. The modeling component of GOALS would include a hierarchy of models but would emphasize large spatial scale. Atmospheric GCMs are global by nature, and the coupling of the atmosphere to the ocean and land would be with GCMs that are also global in extent. Therefore, many of the extensions, problems, and areas of uncertainty that GOALS would eventually face might first be encountered in a modeling context. From this point of view, interannual predictions would be global in scope before the initialization data were available for the global ocean and global land. The modeling component of GOALS therefore emphasizes the extension of coupled models of the tropical ocean and atmosphere to include the tropical land and extratropical atmosphere and upper ocean, and simulation of soil moisture, snow, sea ice, and vegetation. The deep ocean may be important to understanding climate variability on time scales of a decade or longer, but a working hypothesis for GOALS is that this complexity is not crucial to extending the range of short-term climate prediction. More important is the specification of the initial condition of upper-ocean, sea-ice, and land-surface properties that potentially have memories of a season or longer. Also necessary is the development of coupled models that are comprehensive, but balanced in terms of the complexity with which key atmospheric, oceanic, and land processes are represented. It is important to develop coupled ocean–atmosphere–land models that exhibit the full range of variability, not only on the annual scale but also interannually. To verify the spectrum of variability, coupled models must be executed for about 30 model years so that the annual cycle is well-defined. The interannual variations with respect to this annual cycle can then be isolated. Evaluation of the modeled annual cycle thus requires a good annual climatology of important atmospheric, land, and oceanic quantities. This climatology must include global SST, land wetness, and surface wind stress as primary quantities; as well as global boundary-layer depths in the atmosphere and ocean; sea-level pressure; upper-level horizontal fluxes of heat, momentum, and moisture in the atmosphere; and fluxes of heat and fresh water in the ocean. The preparation of a global land wetness climatology, which is needed for running and verifying climate models, would require close cooperation with the ongoing Global Energy and Water Cycle Experiment (GEWEX) program.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Atmosphere–Ocean Coupling The success of coupling the atmosphere to the ocean depends on correctly predicting the surface quantities: SST and surface fluxes of heat, fresh water, and momentum. The determination of these quantities requires much more than a correct parameterization of boundary-layer processes, as these quantities are sensitive to many atmospheric processes above the boundary layer, such as convection, radiation, and cloud formation. Clouds influence the surface radiation budget and enhance sensible heat flux, evaporation, and the surface wind stress in the vicinity of deep convection (Johnson and Nicholls, 1983). Clearly the development and evaluation of atmosphere and ocean models for use in coupled models will require long, accurate records of global SST and wind stress. The coupling of atmospheric GCMs to land-process models and to oceanic GCMs is a large and challenging undertaking, but one that is ultimately foreseen as offering the best hope for producing skillful forecasts of seasonal-to-interannual climate variations. The development of coupled GCMs has been slowed by the difficulty of simulating the correct spectrum of variability. This problem is gradually being overcome; several coupled GCMs now show reasonable annual and interannual variability. Essential to all these efforts is the correct modeling of the climatology—particularly the annual cycle. The annual cycle, forced by annually varying solar radiation, is not adequately simulated in the present generation of fully-coupled atmosphere–ocean models. Successful simulation of the annual cycle is therefore not only a validation of a coupled model, but a prerequisite for correctly simulating interannual variations. Nevertheless, because of its large amplitude and regularity, the annual cycle is relatively well observed and well documented. Some of its more robust features, such as the northern hemisphere winter planetary-wave configuration, are well-simulated in the present generation of atmospheric general circulation models and coupled models, although others are not. For example, the models have difficulty reproducing both the sharpness and subtle seasonal changes in the convergence zones over the tropical South Pacific and South Atlantic as well as the precise timing of the onset and cessation of the climatological mean rainy seasons in some areas. Some of the systematic biases in the models are larger than the anomalies associated with short-term climate variability. More accurate representation of the physics in the models will be needed to correct these deficiencies in the simulated annual cycle. In tropical latitudes, annual SST variations occur as a result of

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis the annual variation of both wind stresses and heat fluxes accompanied by both mixing and advection in the near-surface ocean. At higher latitudes, heat and momentum fluxes through the ocean surface are accompanied mainly by mixed-layer processes, while advection in the ocean is relatively unimportant (Gill and Niiler, 1973), at least away from coastal boundary current regions. Since cloud and humidity variations in the atmosphere affect the heat fluxes into the ocean, the problem of annual and longer-term SST variability in the ocean is fundamentally coupled to atmospheric variability. Since the cloud and atmospheric variability is partly forced by SST variability, the modeling problem is truly coupled. Existing coupled GCMs have proven capable of modeling the ENSO phenomenon with varying degrees of success (for example, N.-C. Lau et al., 1992; Philander et al., 1992). Recent models have even been able to simulate both the annual cycle and the ENSO cycle (Nagai et al., 1992; Latif et al., 1993a); but, to date, no coupled GCM has been capable of correctly simulating the entire annual and interannual spectrum of variability. The status of coupled atmosphere–ocean models and their success in simulating the observed climatology, both annual and interannual, is reviewed by Neelin et al. (1992) and is discussed in the context of prediction in Appendix A of this report. One of the objectives of GOALS would be to continue improving coupled ocean–atmosphere–land general circulation models with a view toward correctly simulating the spectrum of climate variability. Future modeling of the coupled atmosphere–ocean models must come to terms with properly defining criteria for successful coupling. These criteria will need to be defined both for the surface atmospheric and oceanic fields so that an atmospheric model that satisfies its own criteria, when coupled to an ocean model that satisfies its own criteria, will produce a coupled model with correct behavior. Until this is possible, the essence of coupled models will not truly be understood. Atmosphere–Land Coupling Changes in the land-surface properties of albedo, surface roughness, and soil moisture produce changes in ground temperature, evaporation, and sensible heat flux. Changes in horizontal gradients of ground temperature produce changes in the convergence of moisture. Changes in vertical gradients of temperature in the presence of moisture convergence produce changes in convection and rainfall, which subsequently further alter the soil moisture. The nature and degree of these interactions depend on the character of the dynamical circulation regime in which the land-surface changes are taking place. The

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis occurrence of prolonged droughts in subtropical regions (where subsidence occurs) and even the tendency of heat waves to persist in the extratropical regions can be explained, at least in part, by atmosphere—land interactions (Atlas et al., 1993). An accurate treatment of those atmosphere–land interactions provides potential for prediction of seasonal mean surface temperature and possibly rainfall over the continental United States. The development of realistic models of coupled ocean–land–atmosphere systems for prediction of short-term climate variations will require comprehensive land-surface process models that predict soil moisture, land-surface albedo, snow pack, and surface roughness, rather than prescribe them. Since the task of developing better land-surface models and preparation of data sets to validate land-surface models rests primarily with GEWEX, GOALS intends to take advantage of a successful GEWEX program to help incorporate the role of the land surface into global ocean–atmosphere–land models suitable for examining climate variability. Experimental Prediction The ultimate test of a program focusing on seasonal-to-interannual variability is the ability to predict variations. The development and evaluation of coupled upper ocean–atmosphere–land and cryosphere models, the development of a data collection and assimilation system, and a set of hindcasts for predictive validations are all required for the development of improved predictive capability. Sufficient data for proper model initialization and evaluation are crucial to the enterprise. Advances in prediction with coupled upper ocean–atmosphere–land models is heavily dependent on computational infrastructure. Access to the needed computer resources and the concomitant data storage, high-speed data transmission, and visualization facilities needs to be improved, and simulation and evaluation standards need to be developed. Such standards should include model documentation, simulation and intercomparison tests, standard formats for the exchange of model data, and visualization protocols. The best way of accomplishing the prediction objectives of GOALS is by establishing the needed computational infrastructure, by maintaining a strong and expanded research program focused on developing and making predictions (for example, by continuing the TOGA Program on Prediction [T-POP] into the GOALS era), and by a strong proposal-based research program investigating the model questions involving simulation and predictability on short climatic time scales.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis It is anticipated that GOALS would interact strongly with regular and systematic short-term climate prediction activities such as the proposed International Research Institute for Climate Prediction (IRICP) (Moura, 1992), the Coupled Modeling Project at the National Meteorological Center, and other similar activities. OBSERVATIONS GOALS requires observations (1) to increase understanding of processes not well parameterized in large-scale models; (2) to use for assessment and prediction, including initialization of forecast-analysis systems; (3) to provide the research foundation for operational monitoring; and (4) to provide ground-truth for satellite data products. The current TOGA observing system forms a nucleus for GOALS, which requires observations of the seasonal-to-interannual variability of the global upper-ocean circulation. The following sections are provided to help guide the future development of a global observing system; they do not fully specify such a system. A detailed implementation strategy must still be developed. Observations Over and In the Ocean The present TOGA ocean observing system forms a nucleus for development of a GOALS observing system to monitor variability on and in the tropical Pacific (NRC, 1994). Key parts of the TOGA observing system have only recently been fully established; their predictive benefits would be realized during GOALS. The emerging Global Ocean Observing System (GOOS) program will be crucial to making the long-term ocean observations needed by GOALS and will be essential for providing observations to operational government agencies. The TOGA TAO array, which measures surface winds and upper-ocean thermal structure in real-time in the equatorial zone, will not be fully established until the very end of TOGA in 1994. In addition to the TAO array, TOGA measurements of upper-ocean thermal structure, including SST, derive from a variety of sources: VOS XBTs (volunteer observing ships expendable bathythermographs), VOS surface data, drifters, and satellite retrievals. TOPEX/Poseidon provides data on sea-surface height variations and in tropical areas these data can provide the basis for an estimate of thermocline depth variations. The various techniques have different strengths, and some have dual applications. Over the next several years, a heterogeneous array of satellite sensors for SST will be flown. These include AVHRR

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis (Advanced Very High Resolution Radiometer), ATSR (Along Track Scanning Radiometer), MODIS-N (Moderate Imaging Spectrum-Nadir), and OCTS (Ocean Color and Temperature Scanner). Considerable new work will be required to learn how to produce and calibrate SST products to a known, climatically useful accuracy. TOGA has supported the development and deployment of radar-based atmospheric sounding systems at several sites across the tropical Pacific. Preliminary analyses of these data indicate they are likely to be important for GOALS. The wind-profiling radar at Christmas Island has identified serious problems in the way boundary-layer wind shear is analyzed in an operational forecasting model. Automated, integrated sounding systems formed the core of the soundings array in TOGA-COARE. Systems like these offer promise for filling key gaps in the operational sounding network across the global tropics. Current knowledge of seasonal-to-interannual variations of global upper-ocean circulation is poor. One of the measures of success of the GOALS program would be obtaining new results with respect to the seasonal-to-interannual variability of global upper-ocean circulation and its contribution to atmospheric variability. Knowledge of the seasonally varying upper-ocean circulation is important to evaluate results of the GOALS models. The optimum mixture of in situ and satellite observations to describe upper-ocean circulation variability must be determined. Altimeter and scatterometer data may need particular attention because of their potential utility in providing information on the upper ocean. In addition to thermal and flow measurements, accurate surface and subsurface salinity measurements should be obtained. Salinity measurements may provide the best opportunity to evaluate the hydrologic cycle in coupled models. Precipitation over the ocean is an important, but poorly sampled, quantity. Various techniques have been developed to measure precipitation and should be tested at sea. Autonomous ocean profilers of temperature, salinity, and current offer potential for extended time series in remote regions at a considerably lower cost than present systems. The penetration of visible irradiance to depths below the upper layer of the ocean has a potentially large influence on the evolution of SST. Observations of the ocean color field from satellites and from in situ measurements, in conjunction with high-quality surface irradiance fluxes, would be useful to evaluate this potentially important energy source in the mixed-layer heat budget.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Observations Over Land The existing system of climate-related measurements, especially upper-air measurements, is in a state of decline. From 1989 through 1992, reports to the National Meteorological Center (NMC) from upper-air stations in tropical Africa decreased by 50 percent. In the United States and at overseas military stations, serious consideration has been given to reducing the frequency of upper-air observations. Automated Aircraft Enroute Reports (AIREPs) have increased along air routes near the requisite ground equipment, but there has been a decrease in the frequency of manual reporting from remote routes. Automated sounding equipment should be considered as a way of improving the situation, particularly in the tropics, where strong land–atmosphere interactions give rise to a substantial part of the long-term predictability of the coupled system. In addition to the technological challenges for increasing meteorological observations in data-sparse regions, free and open scientific data exchange is being questioned in several countries. In recent years pressures have increased on many national weather services to recover the cost of their investments in observing systems. As a result, some countries have abandoned the standard practice of exchanging data for scientific research at the minimal cost of reproduction. If this trend continues, it would threaten the success of the GOALS project as data became too costly for investigators to purchase. The decline of the World Weather Watch (WWW) radiosonde network across the global tropics is of particular concern, considering the emphasis in GOALS on predictions of regional atmospheric variations. Boundary-layer processes in particular must be understood and accurately modeled to achieve GOALS objectives pertaining to climate fluctuations of moisture convergence and precipitation. Observations of winds, temperature, and humidity above the land surface would be much more important for the success of GOALS than they were for the more limited objectives of TOGA. Land-surface models usually recognize several land-surface types that differ in vegetation structure and physiology, and several soil types that differ in hydrologic properties and albedos. GEWEX will be concerned with the development and analysis of these data sets and of additional data to evaluate land-surface models. However, GOALS would require the accurate initialization of some components of land-surface models (for example, soil moisture).

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Satellite Observations Over Land and Ocean Remote sensing plays a critical role in providing the global data sets required to initialize, force, and evaluate models for GOALS. The GOALS time period, 1995–2010, coincides with a coordinated increase in earth-observing satellites scheduled by the National Aeronautics and Space Administration (NASA), European Space Agency (ESA), Centre National d'Etudes Spatiales (CNES), and National Space Development Agency of Japan (NASDA). Multisensor spaceborne platforms (such as NASA's Earth Observing System [EOS], Tropical Rainfall Measurement Mission [TRMM], ESA Envisats, and NASDA ADEOS) will provide unprecedented coverage of atmosphere, ocean, cryosphere, and land variability. For example, there will be radar scatterometer measurements to derive ocean-surface wind velocity; radar altimeter observations of sea-surface topography; radar measurements of tropical rainfall; passive microwave observations of sea-surface wind speed, sea ice, snow cover, and water vapor; and high-resolution spectrometer/radiometer estimates of SST, albedo, cloud fraction, surface irradiance, vegetation, and ocean color (see Table 4-1). There would be a need to keep abreast of developing technologies that have potential for the remote sensing of soil moisture and sea-surface salinity. GOALS requirements would assist in defining the spatial and temporal requirements for such quantities. GOALS must not become dependent on proposed remote sensing missions because, historically, delays or changes in missions have often resulted in missed opportunities for scheduled research pro- TABLE 4-1 Satellite Data Products for GOALS DATA PRODUCT SATELLITE PLATFORM AND/OR SENSOR Atmospheric Temperature Profile MSU, AIRS Clouds AVHRR, VISSR Land–Atmosphere Flux SAR Phytoplankton Abundance SeaWiFS, OCTS Precipitation TRMM Sea-Ice Extent SSMI Sea-Surface Height ERS-1, TOPEX/Poseidon, ERS-2, GFO, TPFO Sea-Surface Temperature AVHRR, ATSR, OCTS Snow Cover SSMI, MODIS-N Surface-Wind Speed SSMI, ERS-1, TOPEX/Poseidon, ERS-2, GFO, TPFO Surface-Wind Vector ERS-1, ERS-2, NSCAT, Seawinds Vegetation AVHRR Water Vapor SSMI

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis grams. However, the successful implementation of the currently planned earth-observing satellites (TRMM, EOS, Envisats, and ADEOS) would greatly assist the accomplishment of the objectives of GOALS. GOALS would take full advantage of these remote sensing missions when sustainable and verifiable data were obtained and were suitable for process and modeling studies on seasonal-to-interannual time scales. The ultimate suitability of the remote sensing data would depend critically on rigorous analyses, reanalyses, and continuous verification with corresponding in situ data. Continuity and calibration of satellite measurements throughout the GOALS program would be very important and should be achieved by planning for brief intervals of overlap between aging and replacement spacecraft and by planning for a well-maintained, in situ observational network. Expansions and Extensions Expansion of the observing network would be an important part of the GOALS program. The critical role of surface-wind forcing in the tropical ocean and the substantially different model outputs that arise from different tropical surface-wind products have driven the development of the TAO array in the Pacific. TOGA TAO is the only feasible near-term means for obtaining data for accurate, spatially well-mapped winds in this region. Many of the same dynamical considerations and sampling arguments that were used to advocate TAO in the Pacific also apply in the Indian and the Atlantic oceans. It is important to extend TAO measurements into the equatorial zones of these two oceans. If indeed an accurate (to within 0.2K to 0.3K) global mapping of SST is required for significant advances in extratropical climate predictions, then there is much observational work to do. Satellite retrievals alone cannot yield the desired accuracy; a substantial increase in ''sea truth'' measurements with which to calibrate satellite data will be needed. Development of an optimum mixture of techniques and optimum sampling schemes will be a complex task. Technical developments should be pursued to enhance and extend global observations. It is clear that the capability represented by the VOS fleet is not being used to full advantage. Efforts to design and deploy robust packages capable of accurate measurements and of accurate reporting via satellite should be encouraged. The primary beneficial impact would be on standard surface meteorological observations (such as sea-level pressure, SST, and so on), but the possibility of adding streams of other valuable data (such as salinity and air-sea fluxes of heat and water) should be taken into account.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Carefully developed subsurface observations in midlatitude oceans will be useful in understanding and possibly predicting global climate variations. GOALS science objectives require greatly enhanced monitoring of the atmosphere above the ocean surface to characterize and understand boundary-layer processes such as moisture convergence and vertical fluxes of momentum and heat. At present, there are no long-term time series of boundary-layer profiles for these variables over the ocean that can be used to evaluate and improve atmospheric models. Therefore, deployment of cost-effective systems capable of making such measurements would be useful. With the development of unmanned land-based profilers and stabilized shipboard profilers, several of the key technological obstacles to deploying sounding systems on buoys have been overcome. For data transmission, GOALS must rely, at least in its early years, on the Global Telecommunication System (GTS) for the transmission of standard meteorological data in real-time. The limitations of the GTS are well known. The program must also rely to a large extent on the Argos system platforms for transmission of GOALS-specific data. Actions have been taken to modernize and increase the efficiency of the Argos system and its capacity to handle a future increase in the numbers of platforms and in data rates. A continuing problem, however, is the charging mechanism for the Argos system. Regular upgrading of the technical capability of Argos and the examination (by the National Oceanographic and Atmospheric Administration [NOAA] and CNES) of possible different charging mechanisms should be encouraged. At the same time, GOALS should also be ready to exploit new technological developments, which might conceivably offer more efficient operational services in the long term. EMPIRICAL STUDIES Empirical diagnostic studies of observational data would be carried out in conjunction with the other elements in the GOALS program. Some of these studies would be exploratory in nature. Some would capitalize on new components of the observing system, while others would rely on long records of historical or paleoclimatic data. In some cases, the observational results would be compared with results based on model simulations. Results of such empirical studies would be of interest in their own right, but they also would provide motivation for numerical modeling experiments and would serve as justification for process studies. Empirical studies could also point toward needed new components of the observing systems.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Scientists participating in GOALS would have access to a number of valuable new data sets. With the availability of comprehensive data sets, there can be more comprehensive diagnostic studies of processes, circulation statistics, and budgets of heat, momentum, mass, and moisture. Improved estimates of atmosphere–ocean and/or land-surface–atmosphere exchanges should be possible. In addition, empirical studies of observed relationships are invaluable for providing insights and delineating phenomena that need to be modeled and understood. Reanalysis projects carried out at the National Center for Atmospheric Research, NMC, and other operational numerical weather prediction centers promise to provide greatly improved, global, gridded data sets for empirical studies of short-term climate variability (Kalnay and Jenne, 1991). These reanalyzed data sets will extend further back into the past than currently available data sets of this type do, and they will be much more reliable, particularly in the tropics. Significant improvements are expected for the irrotational component of the wind, which is of critical importance for diagnosing variations in the intensity and structure of the tropical heat sources. Assuming that modeling and data assimilation continue to improve, it will be desirable to repeat this process at regular intervals to improve and lengthen the data sets. One of the significant accomplishments of TOGA has been the development of an operational tropical-ocean data-assimilation effort (Leetmaa and Ji, 1989). The first gridded ocean data sets derived from this effort recently became available to researchers. During GOALS, similar activity would expand to encompass the global ocean, the land surface, and the cryosphere. Another important development that has taken place during the TOGA period has been the increasing availability of climate data in near real-time through monthly publications and by on-line data services provided by various government agencies. It is important that these agencies continue to develop this infrastructure so that the scientific community will be able to respond to the increasing pressures for real-time assessments of the status of ENSO and regional climate anomalies throughout the world. Gridded satellite data sets are playing an increasingly important role in research on short-term climate variability. NASA's "Pathfinder" project offers the hope of reanalyzed versions of some of the more important, older, satellite data sets. It would be important that GOALS scientists be involved in the planning of these projects so that the resulting data sets would be available in formats suitable for the study of short-term climate variability. In a similar manner, GOALS

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis scientists would have an interest in the maintenance and upgrading of data sets from ocean-based and land-based measurements such as the Comprehensive Ocean-Atmosphere Data Set (COADS) and the archive of XBT data. Hence, it is important that the implementation plan for data management in GOALS include provisions, not only for the observations taken as part of the program, but also for the maintenance and upgrading of the historical data sets that are needed for empirical studies. Questions of midlatitude variability, correlations of SST variations with atmospheric variations, effects of mean flows on tropical–extratropical interactions, and variability of the Hadley Cell have been and can be approached further by analyses of both existing data and future data as they become available. An intensive study of the COARE data set should also be undertaken. The COARE program acquired an enormous amount of data concerning the interaction between the western Pacific and the overlying atmosphere during COARE's intensive operation period (November 1992 to February 1993). These data sets will have information important to the GOALS program, and it is both prudent and necessary that GOALS exploit existing data sets fully for use in planning future process study experiments. It will undoubtedly take many years past the formal end of TOGA to fully exploit the COARE data set. To support empirical studies of the climatological annual cycle and seasonal-to-interannual variability (40-to-60-day oscillations, quasi-biennial oscillations, and so on), GOALS would assemble and analyze a 30-year (1971–2000) climatic data set. Central to this effort would be the development of techniques for assimilating heterogeneous and sparse data into coupled models of the atmosphere, upper ocean, and land. In addition to examinations of the exchanges of energy, momentum, and water between the various components of the climate system, special emphasis would be placed on the determination of upper-ocean circulation and density structure; rates of change of the internal atmospheric structure due to water phase change and radiative heating; the spatial distribution of soil moisture, snow and ice; and measures of the distribution and properties of vegetation. PROCESS STUDIES The objective of GOALS process research is the improvement of the observations and understanding of the climatically important processes that are poorly modeled or parameterized in the models used to predict short-term climate variations. Process studies conducted

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis in the field would provide research-quality data sets for detailed diagnosis of the most serious deficiencies in the models. These data sets would also be available to improve parameterization schemes designed to remedy these deficiencies in the models. Process studies should exploit the combination of numerical models and new observations in a synergistic and efficient manner. Because additional observations will be needed to make progress in this area, process research would guide the development and expansion of the emerging global climate observing system and would therefore also contribute to planning the observational component of GOALS. If more process studies were proposed than could be conducted, the primary scientific considerations guiding the selection process would be the likelihood that a proposed study would (1) make an important contribution to advancing the state of the art of modeling, especially insofar as it influences short-term climate prediction, and/or (2) elucidate the processes that govern the behavior of the coupled atmosphere-ocean system. Logistical considerations such as the feasibility of conducting operations in various proposed sites, the ability of more than one group of investigators to effectively share the same observing platforms, and the prospects for cost sharing with other nations would also inevitably play a role in the selection process. Following are some potential candidates for process studies, grouped by latitude belt. Process studies in the tropics could examine: oceanic heat balance in regions where the models have difficulty simulating SST; maintenance of the oceanic mixed layer in the presence of equatorial upwelling; the reflection of Rossby waves off the western boundary and related flow through the straits surrounding Indonesia; parameterization of the strong backing of the wind with height in the planetary boundary layer, which is not well-simulated in current numerical weather-prediction models; the role of the stratus cloud decks in the oceanic heat balance; the role of the western boundary currents in the oceanic heat balance and the transport of heat to higher latitudes; and atmosphere–ocean interaction in association with the 40-to-60-day waves. Process studies in the higher latitudes could examine: the roles of coastally trapped Kelvin waves and coastal current systems in accounting for the effects of ENSO along the coasts of Mexico, California and Chile; the processes that modulate the production of subtropical-mode water

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis in the Sargasso sea and to the north of New Zealand, and the associated overturning circulations; the origin of subpolar salinity anomalies and their role in modulating sea–air energy fluxes; and the causes of interannual variations in sea-ice extent and their feedback on the atmosphere. The precise form of the process studies (or field programs consisting of regional sets of process studies) that would be performed as part of GOALS cannot be specified more precisely at this time. It would depend on many things, including results of the COARE experiment, progress in coupled atmosphere–ocean modeling, results of empirical studies of tropical and midlatitude variability, possibilities of logistical support, funding, and international cooperation. Before process studies for GOALS got under way, it would be important to hold a series of implementation meetings during which the shape of process research and its interaction with the rest of the GOALS program could be better defined. CONSORTIA AND PRINCIPAL INVESTIGATOR GROUPS The four elements of GOALS—modeling, observations, empirical studies, and process studies—are highly interrelated, and the structure of the program must reflect those interrelationships (see Figure 4-1). Predictions cannot be performed without initializing observations; long-term observations cannot be justified without the predictions; models cannot be improved without parameterizations derived from process studies; the planning of process studies depends on the knowledge gained from empirical studies; and so forth. In the interests of the strength of the program as a whole, no part of GOALS can be carried out at the expense of any other part. Conversely, all elements must be adequately funded and organized if major progress toward the program's objectives is to be made. Finally, it is important that the elements interact strongly so that advances in one element are available to the others. To foster cooperation among scientists with interests in similar phenomena or in phenomena in a particular geographical region and to strengthen the links among the four elements of GOALS, consortia or principal investigator groups should be established, as illustrated in Figure 4-2. Consortia become essential for considering disparate but interlinked processes and for gathering and analyzing the data that relate to the feedback between processes. For example, one consortium might focus on the suite of phenomena associated with short-term climate variations in the vicinity

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis Figure 4-1 GOALS program architecture, showing the interrelated components of the program. Figure 4-2 Illustration of the partitioning of GOALS research among the various program elements (rows) and the various consortia and principal investigators (columns). Examples of hypothetical consortium themes are given in the text.

OCR for page 37
GOALS Global Ocean-Atmosphere-Land System for Predicting Seasonal-to-Interannual Climate: A Program of Observation, Modeling, and Analysis of Australasia, the Indian Ocean, and the extreme western Pacific. A focus of the consortium might be the organization of a series of coordinated studies of relationships between the monsoon heat sources and the underlying SST gradients, flow through the straits separating the two oceans and its effect on the reflection of waves at the western boundary, the role of boundary currents in the western Pacific in the heat balance, and so on. Particular aspects of the transient variability in this region might be emphasized, such as 40-to-60-day waves, or cold surges, or the tendency for phase locking between ENSO and the annual cycle, which gives rise to enhanced quasi-biennial variability. Another consortium possibly might focus on the suite of problems associated with atmosphere–ocean interaction on the cold side of the tropics (that is, the Atlantic and the eastern Pacific). The continental heat sources over northern South America and Africa do move, with considerable consequences for human populations. The cold-hemisphere oceans have a disproportionate influence on the variability of the large-scale zonal SST gradients. Substantial dividends could be gained from making the SST forecast correctly for this region. Phenomena of interest might encompass boundary-layer processes in the equatorial-cold-tongue/ITCZ complexes, stratiform cloud decks, and the American monsoon. Still another possible consortium theme might be the local oceanic heat balances in contrasting regions of the tropical and extratropical oceans, with emphasis on understanding and modeling the SST variations. Activities organized and coordinated by consortia may involve any combination of empirical studies, process studies, and model simulations. In some cases, long-term observations and/or field experiments might also be proposed. Regardless of the consortia organized, the GOALS program would reserve part of its resources to fund the efforts of individual principal investigators.