The panel considers coordination with other programs, both nationally and internationally, to be a crucial aspect for the successful implementation of GOALS because of the overlapping time scales of interest in climate-related issues and research, and importantly the interfaces that need to be built between the various scientific communities involved. Conceivably, for the first time, all of the physical scientific disciplines are involved. Adding complexity, the social and human dimensions disciplines (endorsed by the USGCRP) are also proposed as important partners. If these objectives are to be achieved by GOALS, extensive coordination will be necessary.

The panel considers it extremely important for internal linkages to be established between the six elements of GOALS. That is, each element of GOALS must be cognizant of activities in all other elements of the program. In particular, there is a need to forge strong links among those making observations, the research and applications users of the observations, those responsible for the preparation and distribution of various derived products, and modelers, empirical and physical alike. Linkages are also necessary between the specialized research areas even within each element of GOALS. Achieving the linkages and coordination needed between the elements and sub-elements of GOALS may not be an easy task. The panel recommends the joint participation of scientists in GOALS

projects and planning activities (including working groups and committees) to address this issue.

For example, the Atmospheric Model Intercomparison Project carries quantitative diagnostic analysis of models and observations in parallel. Similar activities should be encouraged for monsoon simulations and other major regional components of the climate system.

Scientists using diagnostic and empirical techniques should be involved in the planning of process studies and numerical experiments. The results of empirical and diagnostic studies should be used to help evaluate the accuracy and representativeness of model forecast fields, and analysis products. Furthermore, as recommended earlier, a mechanism for an ongoing dialogue between physical scientists (and institutions) who produce climate predictions and the community of users and social scientists should be established. Such activities should be coordinated by a GOALS Project Office when established as proposed in NRC (1995).

As one means of strengthening links among the six elements of GOALS, and in conjunction with an intended focus on particular geographical regions, the panel favors the development of consortia or groups of principal investigators loosely or tightly coordinated but unified by some goal or project. By combining the efforts of investigators with various specialties, consortia provide the ability to consider disparate but interlinked processes, to gather and analyze data that relates to feedbacks between processes, and to move forward with modeling and prediction of coupled phenomena.

For example, one consortium might focus on the suite of phenomena associated with short-term climate variations in the vicinity of Asia, Australia, the Indian Ocean, and the far western Pacific, and their impacts on seasonal-to-interannual climate over North America. A focus for the consortium might initially be on understanding coupled ocean—atmosphere variability in that region. Activities include the organization of a series of coordinated studies of monsoon heat sources and sinks and the underlying SST gradients, flows through straits separating the two ocean basins and their effects on the reflection of waves at the western boundaries, the role of boundary currents in the western Pacific and Indian Oceans, and the heat balance between respective basins. Improved understanding of air—sea coupling in these basins would then be built into models of global climate, with the intent of understanding the impact of the Asian—Australian monsoon system on remote locations such as North America. Another consortium might focus on the suite of problems associated with atmosphere—ocean interactions on the cold side of the tropical ocean basins. Investigators in this consortium could conduct process studies on the seasonal and temporal evolution of the SST field in these regions, with the intent of correcting existing

deficiencies in understanding that lead to poor predictions of the state of the upper ocean and, in turn, of improving prediction over North America. Still another consortium might begin to consider the role of extratropical regions in seasonal-to-interannual climate. It could focus initially on the impact of meridional SST gradients on the atmosphere, conducting process studies on the local heat balance in contrasting regions, and examining the sensitivity of models to such gradients.

Activities organized and coordinated by consortia may involve any combination of empirical studies, process studies, and model simulations. In some cases, long-term observations or field experiments would be the emphasis. In other cases, modeling and consideration of applications might be the central theme. GOALS considers consortia of a number of individual investigators a powerful mechanism for achieving critical mass when addressing scientific objectives. At the same time, the efforts of individual principal investigators are a crucial component, and a significant part of the resources need to be reserved to facilitate individual research efforts.

In recent years it has become apparent that ENSO and the monsoon systems show pronounced variability in the decadal and longer time scales. In order to investigate the modulation of seasonal-to-interannual events occurring in the climate system by features of decadal and longer duration, the panel recommends the establishment of strong links to the DecCen and ACC program components of CLIVAR. These two programs deal with natural variability on long time scales involving the "slow physics" of the system, as well as changes that could occur from increasing concentrations of atmospheric greenhouse gases and aerosols, among others.

The panel also underscores that strong interfaces need to be sustained with GEWEX because of the common interests in global energy and water cycles and their role in seasonal-to-interannual variability. This common interest of GOALS and GEWEX includes overlapping objectives aimed at improving the understanding and modeling of processes and in interpreting predictions in terms of terrestrial water resources.

Collectively, the programs discussed in this section offer a cohesive framework for studying the predictability of climate on seasonal-to-interannual time scales as envisioned by USGCRP. The functional relationship between GOALS and other important components of USGCRP are shown in Figure 12–1.

Specifically, the panel recommends partnerships between GOALS and the following programs:

1. GEWEX-GCIP, PACS, and its international program VAMOS——The former process study focuses on energy and water exchange between the atmo-

5. CLIVAR/ACC——This program deals with anthropogenic climate change and is of interest insofar as such climate change might alter shorter-term climate variability.
6. GCOS, GOOS, and GTOS——These programs are important for the organization of long-term observations involving space-based, surface, and in situ platforms and networks. Included should be coordination with the planning of IGOS.
7. WOCE——Collaboration is required, especially with reference to the upper-ocean and ocean—atmosphere interaction.
8. DOE/ARM (Atmospheric Radiation Measurement program)——ARM has three research sites located in the tropical western Pacific Ocean, in Oklahoma, and on the North Slope of Alaska. At each site, detailed radiation, cloud, and turbulent heat flux measurements are made that are invaluable for model development. Of particular interest to GOALS is a tropical western Pacific Ocean site where measurements pertain directly to processes in the warm pool.
9. NOAAOGP projects, ACCP, IRI, SCPP, IRI, COS, and CCDD——ACCP (Atlantic Climate Change Program) is especially interested in interdecadal changes in the circulation of the Atlantic. Such changes are particularly relevant to GOALS because of the manner in which they can modulate interannual variability and, thus, predictability. The SCPP and the IRI provide predictions on seasonal-to-interannual time scales. The Climate Observing System and Climate Change and Data Detection (CCDD) concentrate on climate observing systems.
10. Other Programs——Other linkages may develop with IHDP under the International Social Science Council. They should be actively pursued within the applications and human dimensions component of GOALS.

1. coordinated organization of scientific meetings (or workshops) and program development meetings and announcements of opportunity, with a focus on topics of common interest between programs, for example, a scientific meeting sponsored jointly by GOALS and GEWEX (Joint NRC panel meetings have proven useful in the past and should be encouraged.);
2. exchange of data between programs and coordinated development of data sets of relevance and value to more than one program (For example, time series of surface forcing variables for atmospheric models should be made universally available for use in sensitivity and predictability tests.);
3. exchange, testing, or coordinated development of process sub-models that are relevant to research undertaken in more than one program (For example, land-atmosphere interaction sub-models might be involved in the case of GOALS and GEWEX, and atmospheric radiation transfer sub-models in the case of GOALS, ARM, and GEWEX.);
4. coordination of GOALS with other components of CLIVAR to optimize the international effort on seasonal-to-interannual research (For example, process studies and long-term observations need to be coordinated.);
5. coordination of the timing of complementary regional studies (For example there might be simultaneous intensive observations such as between PACS and GCIP, PACS and the Land-Biosphere-Atmosphere program, and the GOALS JASMINE and other process studies, and GAME.);
6. coordination with DecCen in the development of observational efforts to understand tropical—extratropical linkages, especially in the oceans, which may be relevant to the decadal modulation of ENSO;
7. joint specification of observational systems to provide data series of value to both GOALS and DecCen (For example, an appropriate balance could be established between continuity and coverage for in situ ocean observations.);
8. coordination of the calibration and evaluation of remotely sensed variables of relevance and value to more than one program (Examples include the calibration and evaluation of TRMM and EOS data by observations from in situ monitoring under GOALS and GEWEX.);
9. interprogram support in the form of advice and guidance in the case of activities that are properly fostered within one program but also address the objectives of another program (For example, GOALS's scientists could participate in workshops that address the development and testing of coupled land—atmosphere sub-models under GEWEX or the development of atmospheric radiation transfer sub-models under ARM.);
10. joint development of coupled ocean—atmosphere—land models (For example, GOALS and GEWEX might undertake the joint development of global-scale, coupled ocean—atmosphere models that include nested mesoscale models in specific regions.);
11. joint activities addressing regional issues and the local application/interpretation of the seasonal-to-interannual predictions fostered under GOALS (For

• example, GOALS and GCIP could work jointly to create a capability to predict and interpret seasonal-to-interannual variability of meteorological fields covering North America.);
• coordinated organization of scientific meetings between GOALS and human dimensions groups to improve the use of seasonal-to-interannual predictions and to reduce human vulnerability to climate variability (Included should be the identification and provision of climate variables and products of most value to decision makers.); and
• joint development (GOALS, IHDP) and provision of guidance in the incorporation of probabilistic climate forecasts into operational decision making processes.

To coordinate and carry out the various activities mentioned in this section, the panel endorses fully the recommendations contained in the GOALS Science Plan (NRC, 1994a), namely: the establishment of a tripartite structure, with a project office (GOALS Project Office), a scientific oversight body (NRC, with its CRC, and the GOALS Panel), and a group of participating federal agencies. The federal agencies would be responsible for implementing GOALS through coordinated funding of research grants. An Interagency GOALS Project Office would serve as a focal point for the implementation of the national research effort, and the GOALS Panel (with oversight from the NRC's CRC) would provide scientific guidance for the program. The Plan also anticipated that principal investigators and ''consortia'' would carry out much of the actual implementation of the GOALS scientific plans. As the needs of the program dictate, the GOALS Project Office would invite groups to prepare coordinated sets of research proposals designed to address specific objectives of GOALS. The above are reiterated in the proposals on a GOALS infrastructure for U.S. participation in GOALS contained in NRC (1995). Close coordination between GOALS and the international CLIVAR is recommended by the panel through a formal link between the project offices for the two programs and an informal liaison between the GOALS Panel and the International CLIVAR Scientific Steering Group.