Numerical simulations of flow-topography interaction using unstructured grids

Eddies and jets are important components of global ocean momentum and heat budgets but are typically unresolved in low resolution global climate models. Herein, they are evaluated with an idealised model set–up that incorporates barotropic flow, past a cylinder on a β –plane. The flow dynamics are a...

Full description

Bibliographic Details
Main Author: McVicar, Alistair J.
Other Authors: Allison, Peter ; Piggott, Matthew ; Czaja, Arnaud
Published: Imperial College London 2012
Subjects:
Online Access:http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.560723
id ndltd-bl.uk-oai-ethos.bl.uk-560723
record_format oai_dc
spelling ndltd-bl.uk-oai-ethos.bl.uk-5607232017-08-30T03:16:38ZNumerical simulations of flow-topography interaction using unstructured gridsMcVicar, Alistair J.Allison, Peter ; Piggott, Matthew ; Czaja, Arnaud2012Eddies and jets are important components of global ocean momentum and heat budgets but are typically unresolved in low resolution global climate models. Herein, they are evaluated with an idealised model set–up that incorporates barotropic flow, past a cylinder on a β –plane. The flow dynamics are a function of two non–dimensional numbers: the Reynolds number and the [Symbol appears here. To view, please open pdf attachment] –parameter. The model used, Fluidity–ICOM, utilises unstructured meshes and a new stable mixed discontinuous/continuous finite element pair (P1DGP2). Unstructured meshes decrease the computational cost; the simulations using a non–uniform unstructured mesh had approximately 40% fewer nodes and ran at twice the speed of a uniform structured mesh for a comparable drag coefficient (Cd). The validation of Fluidity–ICOM was performed for a range of Reynolds numbers (0:0 < Re [Mathematical symbol appears here. To view, please open pdf attachment] 3 x 10[to the power of six]) and the percentage difference between published and Fluidity–ICOM values of Cd was found to be less than 10% for the regimes where the dynamics are essentially two–dimensional. The validation highlighted two important considerations: the position of the lateral domain boundary and the boundary mesh resolution. The wake structure for a moderate Reynolds number (1000) and [Symbol appears here. To view, please open pdf attachment] –parameter (75) changed significantly between coarse and fine boundary resolutions. The former was comprised of a double jet structure and the latter a single jet in the lee of the cylinder. This study demonstrated that resolving the frictional boundary layer dynamics is crucially important, as they substantially impact on the downstream flow. Evaluation of the single jet structure for a large parameter space [Mathematical formula appears here. To view, please open pdf attachment] revealed the presence of interfacial Rossby waves with both eastward and westward propagation with respect to the mean flow. The Rossby wave occurred due to the presence of a strong staircase gradient in absolute vorticity. As the Reynolds number increased for a fixed [Symbol appears here. To view, please open pdf attachment] –parameter, the presence of a stronger shear resulted in a faster phase speed of the Rossby wave and a stronger mean–flow. This parameter–space also showed a large dependence on drag to the [Symbol appears here. To view, please open pdf attachment] –parameter. Overall, this study has implications for the Gulf Stream separation and for understanding the interaction of the Antarctic Circumpolar Current (ACC) with topography.551.462Imperial College Londonhttp://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.560723http://hdl.handle.net/10044/1/10014Electronic Thesis or Dissertation
collection NDLTD
sources NDLTD
topic 551.462
spellingShingle 551.462
McVicar, Alistair J.
Numerical simulations of flow-topography interaction using unstructured grids
description Eddies and jets are important components of global ocean momentum and heat budgets but are typically unresolved in low resolution global climate models. Herein, they are evaluated with an idealised model set–up that incorporates barotropic flow, past a cylinder on a β –plane. The flow dynamics are a function of two non–dimensional numbers: the Reynolds number and the [Symbol appears here. To view, please open pdf attachment] –parameter. The model used, Fluidity–ICOM, utilises unstructured meshes and a new stable mixed discontinuous/continuous finite element pair (P1DGP2). Unstructured meshes decrease the computational cost; the simulations using a non–uniform unstructured mesh had approximately 40% fewer nodes and ran at twice the speed of a uniform structured mesh for a comparable drag coefficient (Cd). The validation of Fluidity–ICOM was performed for a range of Reynolds numbers (0:0 < Re [Mathematical symbol appears here. To view, please open pdf attachment] 3 x 10[to the power of six]) and the percentage difference between published and Fluidity–ICOM values of Cd was found to be less than 10% for the regimes where the dynamics are essentially two–dimensional. The validation highlighted two important considerations: the position of the lateral domain boundary and the boundary mesh resolution. The wake structure for a moderate Reynolds number (1000) and [Symbol appears here. To view, please open pdf attachment] –parameter (75) changed significantly between coarse and fine boundary resolutions. The former was comprised of a double jet structure and the latter a single jet in the lee of the cylinder. This study demonstrated that resolving the frictional boundary layer dynamics is crucially important, as they substantially impact on the downstream flow. Evaluation of the single jet structure for a large parameter space [Mathematical formula appears here. To view, please open pdf attachment] revealed the presence of interfacial Rossby waves with both eastward and westward propagation with respect to the mean flow. The Rossby wave occurred due to the presence of a strong staircase gradient in absolute vorticity. As the Reynolds number increased for a fixed [Symbol appears here. To view, please open pdf attachment] –parameter, the presence of a stronger shear resulted in a faster phase speed of the Rossby wave and a stronger mean–flow. This parameter–space also showed a large dependence on drag to the [Symbol appears here. To view, please open pdf attachment] –parameter. Overall, this study has implications for the Gulf Stream separation and for understanding the interaction of the Antarctic Circumpolar Current (ACC) with topography.
author2 Allison, Peter ; Piggott, Matthew ; Czaja, Arnaud
author_facet Allison, Peter ; Piggott, Matthew ; Czaja, Arnaud
McVicar, Alistair J.
author McVicar, Alistair J.
author_sort McVicar, Alistair J.
title Numerical simulations of flow-topography interaction using unstructured grids
title_short Numerical simulations of flow-topography interaction using unstructured grids
title_full Numerical simulations of flow-topography interaction using unstructured grids
title_fullStr Numerical simulations of flow-topography interaction using unstructured grids
title_full_unstemmed Numerical simulations of flow-topography interaction using unstructured grids
title_sort numerical simulations of flow-topography interaction using unstructured grids
publisher Imperial College London
publishDate 2012
url http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.560723
work_keys_str_mv AT mcvicaralistairj numericalsimulationsofflowtopographyinteractionusingunstructuredgrids
_version_ 1718521688083136512