General Issues for JGOFS and GLOBEC

The JGOFS Southern Ocean implementation plan identified several regions in which to address scientific issues. The identified regions were sufficiently distinct that no single coupled model could address all regions. In addition, each region possesses considerable temporal and spatial variability over a range of scales, thus complicating the interpretation of a data set which must be collected over a limited time and space domain.

Several specific scientific questions that fit within the JGOFS effort include:

These questions require a range of modeling approaches, none of which is presently available "off-the-shelf." One of the major limitations identified was the lack of resolution within the mixed layer by most existing models. It was suggested that most of the JGOFS questions required a higher resolution of FRAM (Fine Resolution Antarctic Model) with a well-defined mixed layer, as well as implementation of a hierarchy of biogeochemical models.

In contrast to the JGOFS scientific plan, the GLOBEC initiative has a greater emphasis on the response of the marine ecosystem to environmental change. Since the GLOBEC plan focuses on the response of populations of marine zooplankton, fish, mammals, and birds, the regional focus is closer to the continent than is the main JGOFS plan. This regional focus results in a different suite of questions and a different range of modeling issues. Several specific modeling problems were identified within this regional focus.

These issues would require the construction of regional, high-resolution models that could be nested within larger scale, lower-resolution models. The complexity of the Antarctic food web poses additional problems for incorporating the biological components into coupled models. It was suggested that a hierarchy of food web models of increasing complexity be developed. This is a long-term research problem that is not limited to analyses of the Antarctic food web.

All models suffer from sub-grid scale representations of biological and physical processes. The time and space scales of importance for many of the biological processes, and several of the physical processes, are within this sub-grid scale. This creates serious research issues in the development of coupled models in general.

For many issues, regional models are the appropriate tool. In particular, regional models are especially valuable for the study of specific processes. However, regional models in the Southern Ocean are faced with several challenges, including boundary conditions, model initialization, ice dynamics, coverage of the full seasonal cycle, shelf drainage, and bottom water formation. The Antarctic Circumpolar Current (ACC) presents its own set of challenges in terms of modeling. The weak stratification and high latitude lead to short dynamical scales (Inoue, 1985). In addition to requiring higher resolution in both the vertical and horizontal dimensions, these smaller scales in turn lead to a different mix of processes that must be considered compared to mid and low latitude models. The impacts of ice formation must also be included in circulation models of the ACC.

Another issue concerns the relative merits of the isopycnal layer coordinate system versus the fixed depth (level) coordinate system in providing vertical resolution for biological processes within the upper ocean. While it was concluded that the layer approach is probably more appropriate for the representation of isopycnal transport and diapycnal mixing, most existing coupled models rely on a fixed depth coordinate system. However, new capabilities of isopycnal models show promise for future Southern Ocean models. In particular, isopycnal models are superior in their abilities to resolve the thermohaline circulation of the ocean.

The term balances from any coupled model must be evaluated. The relative imbalance between terms in the equations of any given model can provide important information on model parameterization as well as key parameters to be measured in the field with better precision.

Finally, it was agreed that the scientific questions must be articulated (as in the program overview statements for JGOFS and GLOBEC), but there needs to be some consideration of how to break these larger questions into achievable components that can be addressed by simple coupled models. This is not a trivial task, but it is essential that the models developed for Southern Ocean studies focus on specific issues rather than try to be all-encompassing. At this point in time, our knowledge of the basic processes is sparse and direct measurements are few. Performance of a single, complex model may be compromised by either inaccurate forcing functions, poor parameter estimation, incorrect model formulation, or simply a lack of observations with which to compare model output.

homepage contents previous section next section