The GOALS Program and Its Scientific Objectives
This chapter presents the scientific rationale for the GOALS research program, enumerating and discussing the specific challenges that the program is expected to undertake, and then presents the scientific objectives of the program.
THE RATIONALE FOR GOALS
As successful as the TOGA program has been, its original goals have been realized only partially. TOGA was based on the assumption that variations in the tropical Pacific Ocean, specifically the warm and cold phases of ENSO, could be predicted without explicit reference to the rest of the world. Furthermore, it was assumed that prediction of the state of the upper ocean in this region, especially SST, could be used to infer seasonal-to-interannual climate variations in remote parts of the world on the basis of past correlations (see Figure 2-1). Predictions have been made as far as a year in advance using only initial wind data from the tropical Pacific. They have shown useful skill in predicting interannual variations in the tropical Pacific, specifically certain aspects of ENSO.
Connections to other tropical oceans have not been made despite tantalizing indications of precursors to ENSO arising over the Indian Ocean (Barnett, 1983) and indications that including the Indian Ocean in coupled models produced results different from those of Pacific-only models (Anderson and McCreary, 1985b). Understanding of the
connection of midlatitude atmospheric anomalies to tropical anomalies has thus far proven elusive. Furthermore, it has not yet been demonstrated how useful the ability to forecast tropical SST anomalies will be in predicting climate anomalies in midlatitudes. The Indian Ocean is connected with the tropical Pacific, and extratropical climate anomalies are correlated with extratropical SST anomalies, but these connections were not investigated by the TOGA program.
TOGA concentrated on understanding and predicting interannual variations in the tropical Pacific Ocean. Although no dynamical phenomenon of the climate system on seasonal-to-interannual time scales has been identified at middle and high latitudes comparable to ENSO, prospects for improving interannual climate prediction are likely to be enhanced with more accurate models of global upper-ocean and land-surface processes. The improved specification of the initial state of the global ocean–atmosphere–land climate system through the assimilation of new types of climate data will also likely enhance predictive capability.
TOGA helped develop skill in predicting aspects of ENSO, especially SST. This skill is seasonally dependent and varies according to the time of year of initialization. At certain times SST can be predicted well and at other times it is predicted poorly. The amount of global interannual variability that can be understood, simulated, and predicted is the focus of the GOALS program.
The GOALS program would endeavor to improve the existing predictive methods and would work to improve climate forecasts by including the effects of ocean areas beyond the tropical Pacific, non-ENSO related phenomena, and land-surface processes. It is expected that GOALS would accomplish several tasks:
The domain of interest would be extended from the tropical Pacific to the entire global tropics to understand seasonal-to-interannual tropical variability and to advance tropical predictive capability on time scales of months to a year or more. This would involve expanding the observing system to the other tropical oceans, using other programs to provide data for the land areas bounding these oceans, and using the combined data for better predictions over the global tropics.
Studies would be undertaken to improve understanding of the connections between the global tropics and higher latitudes for the purpose of defining the existence and extent of predictive skill. This would involve a comprehensive modeling program to develop an understanding
of the mutual interactions of the tropics and higher latitudes and the predictability of high latitude atmospheric anomalies.
GOALS would support CLIVAR research objectives in that program's attempt to observe and understand the mechanisms of global interannual variability, advance global predictability on seasonal-to-interannual time scales, and develop a capability for global prediction on these time scales. This would involve understanding the connections between global surface conditions (SST, land-surface properties, snow, and ice) and global atmospheric anomalies.
The original prediction goals of TOGA in and around the tropical Pacific would be more fully accomplished by developing improved coupled models and by exploiting the data produced by the TOGA observing system, especially the TAO array.
As the domain of the GOALS program expands from the original TOGA focus on interannual variability in and over the tropical Pacific Ocean to include global interannual variability, the range of processes considered would also need to expand. Expansion throughout the global tropics must involve consideration of interactions of the atmosphere with the land masses of India, Africa, and South America, as well as with the Indian and Atlantic oceans. Expansion into higher latitudes brings new considerations of the interactions with land and with the higher-latitude oceans, as well as with ice and snow changes over the land masses and in the high latitude oceans. The examination of the global ocean leads to opportunities to support and enhance the World Ocean Circulation Experiment (WOCE). Although GOALS would retain a focus on atmosphere–ocean interactions, the need to include land processes would also require close coordination and collaboration with other programs, especially with the Global Energy and Water Cycle Experiment (GEWEX), which is designed to improve the understanding of land-surface hydrology and other water-transport processes.
The modulation of the ocean–atmosphere coupling also has important effects on ocean biology and the exchange of carbon across the air–sea interface. The understanding of heat-flux fluctuations on seasonal-to-interannual time scales can provide useful input to ongoing programs such as the Joint Global Ocean Flux Study (JGOFS) and GEWEX for testing ideas about coupling among physical, chemical, and biological systems in the ocean, atmosphere, and on land.
The understanding of natural seasonal-to-interannual (i.e., relatively short-term) climate variability is an important prerequisite to detecting, understanding, and predicting anthropogenic climate change, especially greenhouse warming. Not only do short-term climate variations mask (or enhance) the effects of greenhouse changes, there is a distinct possibility that the slowly changing atmosphere and ocean may induce changes in the short-term climate variability that feed back on slow (for example, decadal) climate changes. Thus, the study of short-term climate variability is expected to make important contributions to the greenhouse problem.
As prediction models inevitably become more global in scope, a logical and needed extension of the TOGA prediction program is the investigation of the initialization and prediction of extratropical SST anomalies and their interaction with the global atmospheric flow. Knowledge of extratropical SST anomalies is of interest in its own right and may be necessary for understanding and predicting seasonal-to-interannual climate variations in midlatitudes.
Anomalies in soil moisture and snow cover have been shown to be important in the genesis and persistence of seasonal climate anomalies. Thus, to understand and predict climate variations, vegetation and land processes must also be considered, initially as boundary conditions and eventually as elements of a coupled system.
In summary, to understand and predict natural climate variations on time scales of seasons to several years, the state of the global upper ocean, atmosphere, and land system must be considered. Such a broad endeavor can be built from the current or anticipated TOGA prediction system in a sequence of measured and well-ordered steps. The study of the tropical Pacific Ocean should be expanded to include the tropical Atlantic, the Indian Ocean, and finally the global upper ocean. For time scales of less than several years, it is probable that only the upper few hundred meters of ocean and wind-driven ocean currents need be considered. Land-surface processes involving soil moisture and albedo need to be considered for observing and predicting the state of the land surface and its interaction with the climate system.
For all of these reasons, it is proposed that the GOALS program should now be launched, with the ultimate objective of observing, understanding, and predicting, to the limits feasible, total climate system variations on seasonal-to-interannual time scales. Useful predictions of climate anomalies on these time scales would have economic and social benefits that could be realized immediately. Droughts, floods, and variations in the monsoons are directly tied to ocean–
atmosphere interactions. Their prediction would result in significant human benefit.
The ultimate scientific objectives of the GOALS program would be to understand global climate variability on seasonal-to-interannual time scales; to determine the extent to which these variations are predictable; to develop the observational, theoretical, and computational means to predict these variations; and to make experimental predictions within the limits proven feasible.
The GOALS program would benefit greatly from the prior efforts of TOGA to establish an ENSO modeling and prediction capability and to establish the TOGA observing system. GOALS would develop a broader scientific scope than TOGA by extending the region of interest to the global climate system, by investigating the feasibility of predicting regional short-term climate variations throughout the world, and by expanding the observational and data transmissions network as appropriate to this investigation. GOALS also would investigate the influence of surface conditions of snow cover, soil moisture, and extratropical sea-surface temperature (SST) for describing or predicting interannual variations of regional climate and the feasibility of developing global ocean–atmosphere–land models for predicting these variations. The scientific objectives of GOALS are:
To observe, describe, and model the variability of the coupled global upper-ocean–atmosphere–land system on seasonal-to-interannual time scales, and to understand the mechanisms and processes underlying this variability and its predictability;
To improve the skill of predicting seasonal-to-interannual variations using coupled models of the global upper ocean–atmosphere–land system and to improve the requisite observing systems; and
To design and implement observing, computing, and data collection systems needed for describing and predicting the state of the global upper-ocean–atmosphere–land system.
These objectives would be met by a program of modeling, observing, process studies, and empirical studies. Ultimately, GOALS would seek to facilitate an orderly transition from an experimental to a permanent observing, computing, and data management system in sup-
port of regular and systematic seasonal-to-interannual climate prediction.
To accomplish the scientific objectives of GOALS, it would be necessary to answer many scientific questions. Following is a partial list of high-priority science questions for the GOALS program:
What are the structure and dynamics of the annual cycle of the coupled ocean–atmosphere–land system, and what are the reasons for its large spatial variability over the globe?
What is the nature of the global, interannual, climate variability and what is its relationship to the annual cycle? What processes give rise to such variability? Can our understanding of this variability be exploited for prediction?
What is the role of the slowly varying conditions at the earth's surface (SST, sea ice, snow cover, and soil moisture) in determining the nature of the interannual variations of the global atmosphere?
What determines low-level convergence of moisture in the tropics over water, land, and coasts? More generally, what determines the location of the thermal sources for the atmosphere?
What is the nature of tropical–extratropical interactions? In particular, what is the role of tropical SST anomalies in perturbing the extratropical atmosphere, and thereby generating extratropical SST anomalies? For what regions of the globe can accurate predictions of tropical SST anomalies be translated into skillful regional climate forecasts one or more seasons in advance?
What improvements in coupled ocean–atmosphere–land models are needed to represent convection, mixing, radiation–cloud–aerosol interactions, and the processes that determine the coupling of the atmosphere and ocean for the purpose of seasonal-to-interannual predictions?
Are there interactions between interannual and interdecadal variability? If so, what is the nature of these interactions? Are there similar or different mechanisms at work on these two time scales?
What is the role of synoptic fluctuations in the tropics and midlatitudes in seasonal-to-interannual climate variability and predictability?
What measurements of the global upper ocean and land surface are required to initialize the coupled models of the global ocean–atmosphere–land system for prediction of seasonal-to-interannual variations?