In
order to study and understand the behavior of sea ice, numerical sea
ice models have been developed since the early seventies and have
traditionally been based on structured grids and finite difference
schemes. This doctoral research is part of the Second-generation Louvain-la-Neuve Ice-ocean Model (SLIM) project whose
objective is to bring to oceanography modern numerical techniques. The
motivation for this thesis is therefore to investigate the potential of
finite element methods and unstructured meshes for sea ice modeling.
Sea ice modeling is a challenging task as it
involves the treatment of sea ice's rheological behaviour and the
resolution of seasonal heat exchanges that drive its thickness
evolution.
The Canadian Arctic Archipelago (CAA) is a complex area formed by
numerous islands and coastlines and constitutes a nice application for
unstructured meshes. Our model is the first to investigate the effects
of resolving the CAA on the ice cover features and the importance of
the CAA in terms of mass balance is highlighted.
We further develop a Lagrangian and adaptive version of the model
allowing the computational grid to move with the ice. We take advantage
of the locality of the mesh adaptation procedure to update the
discontinuous fields thanks to a local Galerkin projection. This
lagrangian version of the model has
several interesting applications, such as the dynamical mesh refinement
along any region of interest (e.g., the ice edge), buoys tracking, or
the inclusion of material properties in the sea-ice rheology.
Sea ice age patterns and how they change in time provide an integrated
view of the recent evolution of sea ice growth, melt and circulation.
We first study the vertical age profile in sea ice and analyze the
age-thickness relationship in a stand-alone thermodynamic sea ice model
of the Arctic. We then take advantage of the Lagrangian model to
reproduce the algorithm used to compute satellite retrievals of ice age
and compare with different ice age definitions. Several characteristics
consistent with satellite observations are deduced from our numerical
simulations.
|
