de.mpg.escidoc.pubman.appbase.FacesBean
English
 
Help Guide Disclaimer Contact us Login
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Journal Article

Simulation of tropical-cyclone-like vortices in shallow-water ICON-hex using goal-oriented r-adaptivity

MPS-Authors
http://pubman.mpdl.mpg.de/cone/persons/resource/persons37098

Bauer,  Werner
Director’s Research Group OES, The Ocean in the Earth System, MPI for Meteorology, Max Planck Society;

Locator
There are no locators available
Fulltext (public)
There are no public fulltexts available
Supplementary Material (public)
There is no public supplementary material available
Citation

Bauer, W., Baumann, M., Scheck, L., Gassmann, A., Heuveline, V., & Jones, S. (2014). Simulation of tropical-cyclone-like vortices in shallow-water ICON-hex using goal-oriented r-adaptivity. Theoretical and Computational Fluid Dynamics, 28, 107-128. doi:10.1007/s00162-013-0303-4.


Cite as: http://hdl.handle.net/11858/00-001M-0000-0013-78F7-2
Abstract
We demonstrate how efficient r-adapted grids for the prediction of tropical cyclone (TC) tracks can be constructed with the help of goal-oriented error estimates. The binary interaction of TCs in a barotropic model is used as a test case. We perform a linear sensitivity analysis for this problem to evaluate the contribution of each grid cell to an error measure correlated with the cyclone positions. This information allows us to estimate the local grid resolution required to minimize the TC position error. An algorithm involving the solution of a Poisson problem is employed to compute how grid points should be moved such that the desired local resolution is achieved. A hexagonal shallow-water version of the next-generation numerical weather prediction and climate model ICON is used to perform model runs on these adapted grids. The results show that for adequately chosen grid adaptation parameters, the accuracy of the track prediction can be maintained even when a coarser grid is used in regions for which the estimated error contribution is low. Accurate track predictions are obtained only when a grid with high resolution consisting of cells with nearly constant size and regular shape covers the part of the domain where the estimated error contribution is large. The number of grid points required to achieve a certain accuracy in the track prediction can be decreased substantially with our approach. © 2013 Springer-Verlag Berlin Heidelberg.