Global ocean ecosystems research necessarily involves regional scale, basin scale and planetary scale considerations. Given the diversity of marine ecosystems, regional intercomparison and contrast of various types of ecosystems-including banks and shallow seas, eastern boundary currents, and representative open ocean regions-are essential. For example, what are the similarities and differences in the ecosystem dynamics and climate sensitivities of stable high-nutrient/low-productivity regions (northeast Pacific, equatorial Pacific, Southern Ocean) versus regions with seasonally varying nutrients and productivity (central mid-latitude bloom regions of the North Atlantic and North Pacific). Subbasin and basin scale transports of material into and out of regional ecosystems are important to the maintenance of such ecosystems and provide significant sources of variability. If the climate of a regional ecosystem is changing, these transports could be especially important. An example may be found in the appearance of tropical planktonic species in the northern temperate waters off the northwest coast of the U.S. during and following the intense 1982-1983 ENSO event. This, of course, was a temporary (interannual) climatic fluctuation. Basin scale processes also require substantial research. Empirical evidence for a possible Pacific-wide phenomenon is provided by the correlation of the intense variability of the sardine catch in Japan, California and Chile (Kawasaki 1992; Lluch-Belda et al. 1989; Figure 1). Another example might be the apparent connection between salmon stocks in the Pacific Northwest (Oregon, Washington, British Columbia, and Alaska) and North Pacific atmospheric pressure and circulation (Beamish and Bouillon, 1993). The mechanism may be related to winds, upwelling and production of the subarctic Pacific gyre and its effects on survival of salmon in their earliest marine phases. Long-term, large-scale changes in production have been hypothesized for the subarctic Pacific gyre (Brodeur and Ware 1992). There has been to our knowledge no quantitative demonstration of planetary scale biological phenomena or variability. It must, however, be expected to occur. Large-scale physical variability occurs on multiannual and decadal scales, e.g., regime shifts, ENSO and the North Atlantic Oscillation (see U.S. GLOBEC Reports No. 2, 7, and 11). These large-scale physical variabilities have planetary scale teleconnections and correlations which must be expected to cause related biological variabilities.

To address global research objectives will require dedicated observational, experimental and modeling studies, retrospective data analysis and data assembly and management. Experiments will need to be coordinated scientifically, methodologically and logistically. New instruments and sampling strategies will be necessary. Nested multiscale (local to global) interdisciplinary models for nowcasts, forecasts and simulations must be developed and made operational. Regional ecosystem simulations need to be nested within existing state-of-the-art global climatic models, and simple global ecosystem models will have to be constructed. Analytical and computational issues in physical, biological and chemical modeling are involved. Global data sets need to be assembled and studied. The identification of critical variables and standard parameters for global change ecosystem research and monitoring is of crucial importance. Factors in selecting these variables include: relevance, feasibility, and efficiency of acquisition. Concentration and distribution of phytoplankton biomass and zooplankton biomass by size class are potential standard variables. Given U.S. GLOBEC's emphasis on populations it will be essential to include zooplankton species composition as a standard variable whenever feasible. Retrospective data analysis can provide initial guidance for the selection of standard variables. Data sets from remote and in situ platforms will need to be acquired. Emphasis should be placed on ecosystem parameters that could be derived by statistical inference (correlations) from relatively easily sensed variables such as acoustic data (active and passive techniques) and color. Surface and near surface phytoplankton biomass from ocean color sensors, and zooplankton biomass by size class from acoustics, are examples. Less obviously correlated ecosystem parameters should also be explored. Perhaps proxy indices for parameters not directly measurable by inexpensive remote sensing could be developed by retrospective analysis. The global coordination, standardization, and intercalibration of national fisheries measurements and the use of fisheries vessels worldwide as ships of opportunity could represent a feasible and important means for acquiring the data sets needed to explore global change. All of these issues relate directly to research and development of a global monitoring and management system.

The fundamental scientific knowledge on biophysical interactions in the sea to be acquired through U.S. GLOBEC research will be immediately applicable to the design of a system for monitoring, prediction and management of the ocean ecosystem. That system will consist of an observational network coupled to interdisciplinary numerical models. For efficiency and flexibility both the network and the model should be modular so that simpler or more complex versions can be utilized. The observational network should be an efficient mix of platforms and sensors and the models should be fully data assimilative, i.e., capable of including physical, biological and chemical data in near-real time. The network is envisaged to have both a sparse global component and a relocatable regional generic system. The former will provide the globally comprehensive long-term observations, but at low spatial resolution. The latter will provide the capability to enhance the spatial resolution in key areas for limited time periods. A successful operational system depends essentially upon 1) the identification of relevant global change ecosystem variables that can be remotely sensed and 2) an optimal allocation of sampling resources selected through analysis of observational system simulation experiments (OSSEs; see GLOBEC International Spec. Contribution No. 2). The U.S. GLOBEC program is designed to move from the deployment and exercise of research systems to the deployment of prototype operational monitoring systems. The ecosystem operational monitoring system designed, demonstrated and validated within U.S. GLOBEC should become an integral element of the Global Ocean Observing System program (GOOS).

The global research activities of U.S. GLOBEC will be implemented in cooperation with, and as a participating national program of, the GLOBEC International program. GLOBEC International was established in 1992 and is sponsored by IOC, SCOR, ICES and PICES. There are presently GLOBEC programs in China, Japan, Canada, Norway and the United States, and programs are being developed in the United Kingdom, Germany, and New Zealand; several other nations are also expected to participate. GLOBEC International will finalize its science plan in 1994. GLOBEC International will provide a forum for 1) focusing global ocean ecosystem research issues; 2) implementing global research activities; and, 3) coordinating regional studies (e.g., Southern Ocean, GLOBEC International Report No. 5, 1993; Cod and Climate Change, GLOBEC International Report No. 4, 1993). Important issues that GLOBEC International has been addressing include technology-transfer (methods, instruments, models) from advanced to developing nations and laboratories; intercalibrations of instruments; the definition of standard variables and techniques; coordination of research; exchange of ideas; and the generation of global and shared data sets.

The global ecosystem is of paramount importance to the United States not only because of our interest in global scale processes, but also because the seas of our own EEZ are almost certainly influenced by global processes. Such influences and interactions are poorly documented. U.S. GLOBEC will advance fundamental knowledge of ecosystem dynamics in the context of changing climate, and will provide input to research on, and management of, living marine resources. At the same time, U.S. GLOBEC will help direct applied oceanographic and fisheries research towards the development of an observational network and data assimilative model system for monitoring and management. Such a system will provide the timely information base critical to informed decision making by environmental policy makers, economists, commercial leaders and resource managers.