Application of Continuous Dynamic Grid Adaption Techniques to Meteorological Modeling. Part II: Efficiency

View More View Less
  • 1 Bureau of Meteorology Research Centre, Melbourne, Australia
© Get Permissions Rent on DeepDyve
Restricted access

Abstract

The continuous dynamic adaptive grid (CDGA) technique has been shown to yield significant improvements in solution accuracy over equivalent fixed-grid methods. In this paper we consider the related question of efficiency; for a given degree of solution accuracy is the CDGA method faster than the fixed-grid approach, and if so, by how much? A new grid-generation algorithm based on discrete equidistribution principles is proposed, which, while it has the disadvantage of producing nonorthogonal grids, is extremely simple to implement and fast to execute. The algorithm is applied to a barotropic model designed to examine the motion of interacting multiple vortices. Numerical tests confirm the efficacy of the proposed algorithm and show that the CDGA method can produce solutions as accurate as those obtained from a fixed-grid model in only one-third the time.

Abstract

The continuous dynamic adaptive grid (CDGA) technique has been shown to yield significant improvements in solution accuracy over equivalent fixed-grid methods. In this paper we consider the related question of efficiency; for a given degree of solution accuracy is the CDGA method faster than the fixed-grid approach, and if so, by how much? A new grid-generation algorithm based on discrete equidistribution principles is proposed, which, while it has the disadvantage of producing nonorthogonal grids, is extremely simple to implement and fast to execute. The algorithm is applied to a barotropic model designed to examine the motion of interacting multiple vortices. Numerical tests confirm the efficacy of the proposed algorithm and show that the CDGA method can produce solutions as accurate as those obtained from a fixed-grid model in only one-third the time.

Save