The role of ocean biogeochemical processes in the global carbon cycle and, particularly, the involvement of ocean processes in sequestering anthropogenic carbon dioxide are research topics which are now receiving a great deal of attention. A decade ago there was consensus that the ocean took up a quantity of carbon dioxide equal to about half the amount added annually to the atmosphere by human activities. Recent observations and models have eroded the old consensus and now there are several strongly conflicting hypotheses about where the "missing" atmospheric carbon is sequestered. To help answer these important questions an international research program, the Joint Global Ocean Flux Study (JGOFS), was begun. My research group at Duke has participated in research in the following regions of the world ocean: Equatorial Pacific (1992, 1993 and 1995), Arabian Sea (1995) and Southern Ocean around Antarctica (1996, 1998 and 2002). My research focused on the role of physical conditions in regulating primary production and phytoplankton community structure. After this extensive field research we are currently modeling oceanic productivity relationships to gain a predictive understanding of the role of oceanic processes in the global carbon cycle. This research activity is described in the following project summary of our NASA research project, "Impact of Pacific Climate Variability on Ocean Circulation, Marine Ecosystems and Living Resources: A Multi-Scale Modeling and Data Assimilation Approach to Forecasting." The objective of this project is to simulate and forecast climate variability impacts on ocean circulation, marine ecosystems and living resources of the Pacific basin using a multi-scale modeling and satellite data assimilation approach. This project brings together an interdisciplinary team of meteorologists, physical and biological oceanographers and fishery scientists to develop and demonstrate the use of NASA's real time satellite information in conjunction with multi-scale coupled physical-ecosystem models. Pacific climate variability - ENSO (El Niño Southern Oscillation) and PDO (Pacific Decadal Oscillation) - has a significant impact on oceanic ecosystems and their living resources. In the tropical Pacific, climate variability, dominated by interannual ENSO fluctuations, can be predicted a year in advance. In the mid-latitudes, where variability appears to be dominated by the interaction of the seasonal cycle with the PDO, "reemergence" theory suggests that winter SST anomalies are predictable a year in advance with PDO-dependent accuracy. Real-time information from NASA satellite missions will be assimilated into a multi-scale coupled physical-ecosystem model that consists of a relatively coarse resolution (50-km) Pacific basin model with nested fine resolution (5-km) regional models. Biological and chemical processes are modeled with a well tested 10-component ecosystem model regulated with multiple nutrients and embedded in both basin and regional physical models. For the tropics, an empirical atmospheric model will be coupled to the ocean model; for mid-latitudes, a statistical technique based on "reemergence" theory will be applied. We will perform a series of retrospective analyses from 1948 to the present, refining the different model components and the data assimilation system. The physical and ecosystem model output will provide input for several off-line fish population models to simulate the impact of climate variability on living resources. The final objective of this project is real-time simulating and forecasting of living resources in the Pacific with results accessible through a user-friendly web-based interface.