Report 18 -- Strategies A

General Considerations

As JGOFS begins its final process study, it is also beginning a Synthesis and Modeling Program (SMP) that will integrate the JGOFS observations and a suite of numerical models that will capture our knowledge of the oceanic carbon cycle, including the uncertainties in this knowledge. As AESOPS unfolds, it should benefit greatly from the SMP activities, particularly during the analysis phase. In this section, we describe the types of observation and modeling activities that will be especially relevant to SMP. Although GLOBEC has not yet developed an equivalent SMP effort, this discussion should be relevant to it as well.

The types of model studies will depend on the scientific issues that we wish to address. Concerning coupled biological/physical models, most biologists would request information relevant to the horizontal and vertical transport of nutrients and biomass, which would then permit evaluation of the interactions between components (e.g., rate processes). The biological aspects of a coupled model must therefore define the links between components, i.e., the functional forms of the interactions. In addition, we must know how the parameters of the functional forms vary in response to variations in temperature, salinity, depth, etc. What do we need from physics so that the coupled model adequately represents the physical environment in which the biological components function? To first order, the physical aspects of the model should provide the mean and spectrum of variations in horizontal and vertical velocities, and the fluctuations in temperature, salinity, and mixing conditions. Although this appears to be straightforward, considerable scientific thought and understanding is required to quantify the time and space scales that are relevant for the particular model. These scales will depend on the scientific questions under study and will drive the specific components and the functional forms that are included in the model.

To develop these models will require a substantial interaction between the observational and modeling components of JGOFS and GLOBEC. For example, how could measurement and sampling strategies be improved to derive the most return from numerical models? What sorts of measurement programs are needed to improve models and model parameters?