'Next Generation' unstructured mesh ocean global circulation modelling.

Lead Research Organisation: University of Oxford
Department Name: Oxford Physics

Abstract

We will build a next generation ocean global circulation model that is more accurate and has more detailed resolution than existing models. This model will be capable of resolving flows simultaneously on global, basin, regional, and process scales. It will be able to change numerical detail in response to both the structure of the modelled flows and user-defined preferences of regions and structures of specific importance. We shall use the model to carry out ocean research which is not possible with existing models. Despite significant advances over the past decade, numerical ocean global circulation models (OGCMs) are based on essentially the same finite-difference methods employed in the earliest ocean models developed in the 1 960s. Meanwhile, unstructured finite element/volume methods have been deployed to great effect in engineering applications and offer several major advantages for ocean modelling. These include the abilities to: conform accurately to complex basin geometries; focus resolution where it is most needed in response to the evolving flow or regional importance; move the mesh in response to error norms and maintain vertical density structures; incorporate various natural boundary conditions in a straightforward manner; and to make rigorous statements about model errors and numerical convergence. The ability of an unstructured mesh ocean model to change resolution smoothly and dynamically in response to changing ocean dynamics can be viewed as the 'ultimate in grid nesting', but avoiding the various difficulties inherent in matching different dynamical regimes at nest boundaries. Although unstructured mesh modelling has long been a goal of many oceanographers, attempts to apply such methods to model the global circulation have failed due to challenges in treating the Coriolis and buoyancy terms accurately and stably on unstructured meshes. Over the past few years important solutions to this problem have been developed (many by us) and incorporated into the Imperial College Ocean Model (ICOM). This model has the best available parallel mesh adaptivity methods, a suite of options for spatial derivatives (such as high-resolution methods for density/tracer advection), novel and robust treatments of balance, optimised bathymetry and coastline geometries and new large eddy mesh adaptive turbulence models. ICOM will form the foundation of the OGCM built by the proposed consortium. The drivers for it are twofold: (a) a new research tool is a necessity as 'standard' finite-difference codes become harder to tweak to improve their accuracy; and (b) there are significant science problems where the disparity between length scales requires the application of new modelling techniques such as ours (e.g. flow through sills and down slopes, eddies, convection within gyres, etc). These methods have the potential to revolutionise the way in which ocean modelling is done, and thus to secure the long-term future of UK Ocean Modelling at the forefront of the field. The next generation ocean model, with its ability to efficiently resolve a wide range of scales, will have numerous applications to the NERC and wider communities. For ocean modellers, it offers the opportunity to efficiently resolve both basin scale circulation and small-scale processes such as boundary currents, through- and overflows and geostrophic eddies. For Earth system and climate modellers, it offers the opportunity to focus resolution in regions of particular importance, such as boundary currents and overflows, without increasing the computational cost above that of a conventional coarse-resolution model. With a wide range of applications in oceanography, climate change, flood defence, pollution and contaminant dispersal, sustainability of water quality and fisheries, the development of such a model is extremely desirable.

Publications


10 25 50
Cotter C (2008) Diagnostic tools for 3D unstructured oceanographic data in Ocean Modelling
Piggott M (2008) A new computational framework for multi-scale ocean modelling based on adapting unstructured meshes in International Journal for Numerical Methods in Fluids
Piggott MD (2009) Anisotropic mesh adaptivity for multi-scale ocean modelling. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Slingo J (2009) Developing the next-generation climate system models: challenges and achievements. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

Related Projects

Project Reference Relationship Related To Start End Award Value
NE/C521028/1 12/01/2006 11/04/2007 £385,029
NE/C521028/2 Transfer NE/C521028/1 01/04/2007 30/04/2011 £352,606
 
Description How to achieve fundamental balance by the finite element choice (balance between Coriolis, free surface height and buoyancy and pressure gradient). Also how hydrostatic balance can be achieved by separate treatments of hydrostatic pressure. How to enforce balance after interpolating the solution variables from one mesh to another when adapting the mesh to optimally resolve the physics. The first conservative and general mesh to mesh interpolation method was developed - a major general contribution to computational physics. How to form the sensitivities in a relatively simply way of complex environmental flow models. How to ensure that non-hydrostatic (vertical inertia ignored) environmental flow models (considered by many to be the future of ocean and atmospheric modelling) can have a computational cost that scales independently of the aspect ratio - key to a successful model. This was achieved with a new multi-grid method. We have also applied the research to discover the
likely cause of events in an ancient tsunami in the Mediterranean through this geometry conforming modelling ability and developed a number of insights into paleo oceanography such as the state of the life (which will be dictated by mixing) in paleo shallow oceans. This is also helping exploration companies determine where to look for oil and gas.
Exploitation Route The GungHo project lead by the Met. office to develop an unstructured mesh atmospheric model users numerical methods (as well as researchers that developed these) derived from the advanced numerical methods developed by us. The widely used FEniCs finite element model now users an adjoint method for data assimilation and optimization derived by our researchers. The work has lead to the atham-fluidity modelling being developed by Cambridge, Imperial College and the Institute of Atmospheric Physics in China. Fluidity, developed here, is now an open source model used throughout the world.
The Institute of Atmospheric Physics in China have incorporated the adaptive mesh adaptivity approach into their
regional scale atmospheric model to transport pollution and chemistry and are experiencing substantial savings
in CPU time. The MAGIC EPSRC grand challenge consortium is now using these methods and codes to develop its urban air flow models.
Sectors Education,Energy,Environment
 
Description As above by the Met office building on the numerical methods we developed within the GungHo project. The Institute of Atmospheric Physics in China have incorporated the adaptive mesh adaptivity approach into their regional scale atmospheric model to transport pollution and chemistry and are experiencing substantial savings in CPU time (or greater resolution) that will be used operationally to achieve results that we can be more confident in. The adaptive methods (interpolation) have been incorporated into open-Foam - the most well used open source CFD code. The work on ocean modelling is also helping exploration companies determine where to look for oil and gas.
First Year Of Impact 2009
Sector Energy,Environment