Ocean Modelling

The science about turbulent closure schemes is now fairly mature, but there may still be surprises associated with subtle aspects of vertical mixing in the deep ocean that may have important consequences on long time scales. Vertical mixing is also critically important for biogeochemical cycles, because it controls the return of nutrients to the surface euphotic zone, and therefore the magnitude of primary production. Another area where there still is room for improvement concerns the exchanges of heat, momentum and freshwater across the ocean surface.

Accuracy, resolution, and extent (in time ahead) of wind forecasts are the primary limiting factors for sea-state and surge forecasting. Likewise, sea surface heat exchange is clearly a determining factor in forecasting ocean mixed-layer depth and ice formation. In both cases, the need for dynamically coupled ocean-wave-ice-atmosphere models is an essential element to improve atmospheric forcing.

Coastal ocean modelling and forecasting is a major challenge for the scientific community because of the specific and rich dynamics of those regions, and because of the various couplings with the lower atmosphere and exchanges with the near-shore and offshore regions. These issues, needs and challenges have led to the development of a wide range of models of various types. Phenomena of interest include coastal current interactions, coastal meso-scale, tides and storm surges,

2 CLIVAR=World Climate Research Program (WCRP) project that addresses Climate Variability and Predictability. IMBER=Integrated Marine Biogeochemistry and Ecosystem Research; GEOTRACES = International study of the global marine biogeochemical cycles of trace elements and their isotopes; SOLAS = Surface Ocean Lower Atmosphere Study.

Fig. 1.6 Application of the 2-way grid-refinement software AGRIF to the Bay of Biscay, tested in the framework of the MERSEA project (Cailleau et al. 2008). The large-scale model is a 1/3° (Mercator grid) North Atlantic configuration of the NEMO Ocean general circulation model. The fine-scale model is regional configuration of NEMO at a resolution of 1/15° (Mercator grid). Both models are run simultaneously and interactively for years of simulation on either vector or massively parallel super computers. The computational surcharge induced by the 2-way coupling of the grids is very small (just a few percent). The regional model benefits from the smooth and regular behaviour of the large-scale model at its open boundaries. On longer time-scales, the large-scale model benefits from the local improvements brought by the high resolution to the representation of the dynamics in the Bay of Biscay, especially the slope current. The above figure displays a sea surface temperature snapshot on 22 March 1996. One shall notice the fine-scale and the intense eddy field of the fine-grid model, but also the continuity at the limit between the two grids

Fig. 1.6 Application of the 2-way grid-refinement software AGRIF to the Bay of Biscay, tested in the framework of the MERSEA project (Cailleau et al. 2008). The large-scale model is a 1/3° (Mercator grid) North Atlantic configuration of the NEMO Ocean general circulation model. The fine-scale model is regional configuration of NEMO at a resolution of 1/15° (Mercator grid). Both models are run simultaneously and interactively for years of simulation on either vector or massively parallel super computers. The computational surcharge induced by the 2-way coupling of the grids is very small (just a few percent). The regional model benefits from the smooth and regular behaviour of the large-scale model at its open boundaries. On longer time-scales, the large-scale model benefits from the local improvements brought by the high resolution to the representation of the dynamics in the Bay of Biscay, especially the slope current. The above figure displays a sea surface temperature snapshot on 22 March 1996. One shall notice the fine-scale and the intense eddy field of the fine-grid model, but also the continuity at the limit between the two grids tsunamis, shoreline change, coastal upwelling, river plumes and regions of freshwater influence, atmosphere-driven processes, surface waves, and sea ice (Fig. 1.6). Coastal ocean systems can have very high spatial gradients in both the vertical and horizontal, especially near river mouths, requiring the use in models of sophisticated mixing schemes, and high order numerics. The key constraints on the accuracy of these models now lie with the specification of input data (bathymetry, bottom roughness, lateral and surface forcing). In these shallow systems, and especially along exposed shorelines, wave-current interactions play an important role. Measuring and predicting exchanges between the underlying sediment and the water column is critical for coastal biogeochemistry, and is still a key challenge. Sediment models attempt to represent the effects of re-suspension and deposition of particulate material, and their interaction with the circulation, on suspended concentrations (turbidity, important for optical properties and hence primary production) and bed thickness and composition (geomorphology). Models of these processes are still under active development.

With respect to biological processes, we are faced with the general problem of biogeochemical and ecosystem modelling (used here synonymously); namely, choosing the right level of abstraction and approximation in describing and predicting the structure and function of a complex system with many nested levels of complexity.

0 0

Post a comment