Home | deutsch  | Legals | Data Protection | KIT


The global ICON-EPS: recent developments in the operational system and future plans
Oct. 10, 2019, 11:00 - 12:00
KIT, Campus Süd, IMK-TRO, Wolfgang-Gaede-Str. 1, Geb. 30.23, Seminarraum 13.OG

M. Denhard,  A. Rhodin, H. Anlauf, A. Cress, , F. Fundel, J. Winkler, M. Buchhold, G. Zängl and R. Potthast

Deutscher Wetterdienst, Offenbach, Germany

Since January 2018 DWD runs a global ICON ensemble suite with 40 members and approx. 40km horizontal resolution including a grid refinement for Europe of 20km. Forecasts are generated at 00/12UTC up to +180h and at 06/18UTC up to +120h. Four additional runs at 03/09/15/21 provide boundary conditions for the COSMO-D2-EPS up to +30h. The ICON-EPS initializes from analysis states generated by the Local Ensemble Transform Kalman Filter (LETKF) data assimilation system running at DWD. Random perturbations of some physical model parameters are selected at the beginning of each forecast. We provide operational ensemble products on our open-data servers according to the WMC (World Meteorological Center) requirements. Moreover, a contribution to TIGGE (The International Grand Global Ensemble) is under way.

In the talk we show some verification statistics and give an outlook. There are two major developments under way: first, we will introduce a control member by re-centering the perturbations on the high resolution EnVAR analysis of the deterministic system and second, we work on improving the initial condition perturbations.  In the first 2 days of the forecast the spread/skill relation of the ensemble has almost no skill which leads to poor boundary conditions for our LAM simulations. Since the initial perturbations in the ICON-EPS are provided by the LETKF, they depend to a great extend on the co-variance inflation techniques in the filter. It is well known, see Hamill and Whitaker (2011), that these kind of perturbations lack error growth. To overcome the problem we are introducing a new technique for estimating fast growing Singular Vector (SV) perturbations in the ICON modelling framework. Instead of solving an equivalent eigenvalue problem with the Lanczos algorithm, as it is done at ECMWF (requires linear and adjoint model integrations), we use the Arnoldi method for setting up Krylov Subspaces directly from the non-linear model forecasts. The related paper is submitted to QJRMS and a preprint can be obtained from http://www.mathematik.uni-marburg.de/~numerik/preprints/?lang=DE

This event is part of the eventgroup Sonderkolloquium
Michael Denhard

DWD Offenbach
Institut für Meteorologie und Klimaforschung (IMK-TRO)
Wolfgang-Gaede-Str. 1
76131 Karlsruhe
Tel: 0721 608 43356
Mail:doris stenschkeRwm5∂kit edu
Scientific Staff
Propose your own event