Hydrodynamic Model

SIMA Austral Hydrodynamics

The EMS framework is to be applied to a regional domain within the Chilean shelf to deliver a hydrodynamic – connectivity – sediment – BGC system (Figure 1). In the first instance a hydrodynamic model is applied, operating in near real-time (NRT) to deliver now-cast products, and contribute to a growing archive that may be interrogated for system characterization or scenario analysis. Proven technology for orchestration of the real-time components is utilized (TRIKE). If forecast forcing products are available, it is a simple extension to provide forecast capability.

A finite difference hydrodynamical model is constructed over the Los Lagos region using orthogonal curvilinear coordinates and a wetting and drying geopotential vertical coordinate. Applying EMS to a regional domain smaller than the whole Chilean coastline and shelf allows the model resolution to be more optimally tailored to that domain, and potentially results in greater model throughput. The hydrodynamic model currently nests within a larger regional model, which in turn is nested within the OceanMAPS (10 km) global ocean model. Alternatively, the model may be nested directly within the IFOP ROMS operational model in future. Atmospheric exchanges are provided from the global ACESSS-G (25 km) atmospheric model. Freshwater inputs can be included at all locations where near real-time flow data is available. A benthic-pelagic sediment transport, optical and biogeochemical model may be driven by output from the hydrodynamic model using a conservative offline transport model. This de-coupling of hydrodynamics and BGC is necessary due to the high computational cost of the BGC model; by utilizing a transport model the runtime ratio of the system can be increased by several orders of magnitude.

The use of a smaller domain means that geographic complexity is reduced, and a grid can be constructed that more optimally resolves a greater percentage of the domain. The reduced number of grid cells in the mesh also means that runtime is less, allowing more simulations to be completed to produce a better constrained model against observation. The reduced size of the domain also reduces spatial variability in calibration parameters, likely resulting in higher model skill over a greater percentage of the domain. Mean grid spacing in this model is ~575 m, with a minimum of 390 m and maximum of ~1850 m. There are 31 layer in the vertical, with 1 m grid spacing at the surface extending 2 m above msl. The k-e 2-equation turbulence closure scheme (Burchard et al, 1998) is used, and the ULTIMATE QUICKEST scheme (Leonard, 1991) is used for tracers, which also use the Smagorinsky horizontal diffusion scheme (Smagorinsky, 1963). Heat and freshwater fluxes are included at the surface, the latter using the bulk scheme of Kondo (1975). The model has currently undergone no calibration, and is in the pilot stage.

The obvious limitation of this approach is that the entire region of interest is not modelled, and the model can supply no intelligence on issues occurring beyond its boundaries. However, this approach clearly demonstrates the potential of modelling to deliver environmental information, and serves to supply predictive products and assist management strategies and system characterisation in the Los Lagos region.

References

Burchard, H., K., O. Peterson and T.P. Rippeth (1998) Comparing the performance of the Mellor Yamada and the k-e turbulence models. J. Geophys. Res., 103, 10,543 – 10,554.

Kondo, J. (1975) Air-sea bulk transfer coefficients in diabatic conditions. Boundary-Layer Meteorology, 9, 91-112.

Leonard, B.P. (1991) The ULTIMATE conservative difference scheme applied to unsteady one-dimensional advection. Comp. Methods in Appl. Mech. and Eng., 19, 17 – 74.

Smagorinsky, J. (1963) General circulation experiments with the primitive equations, I. The basic experiment. Mon. Wea. Rev., 91, 99 – 164.