881
Short term monsoon forecasts using ICON
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Thursday, 8 January 2015
ICOsahedral Non-hydrostatic (ICON) modeling framework is a joint project developed at the German Weather Service and the Max Planck Institute for Meteorology, targeting a unified modeling system for general circulation models. This framework can be used in a unified way for simulations ranging from synoptic scales to meso-scales to eddy resolving scales. ICON uses geodesic grids derived from an Icosahedron. Among others ICON uses density as a prognostic variable, this leads to an exact natural conservation of mass and tracers. ICON hosts a flexible two-way or one-way nesting capability, with gradually refined grids over one or more regions of interest, overlayed over a coarser global or limited area grids. The non-hydrostatic dynamical cores ensures a smooth transition from resolved scale convection to the parametrized convection. ICON achieves a better skill then its predecessor at the German weather service and at a cheaper computational cost. The goal of current study is to use ICON for short term Indian monsoon forecasts to evaluate if its superior numerical capabilities show any improvements against current, soon to be operational Climate Forecast System (CFS) at T124 (100km) and its high resolution atmospheric component Global Forecast System (GFS) at T574 (27km).
For this study global simulations using ICON are made at a resolution of 100km and gradually refined to 27km over Indian region using several nests. This is mainly intended to match the high resolution version of the GFS. Additional effort is made to ensure that the ICON and GFS are initialized with a similar state. With this setup we perform short term forecasts for specific extreme monsoon events and assess fidelity of ICON in simulating these weather events as compared to the CFS and the GFS. For instance, the steep topography over western coast of India poses a challenge to CFS, signs of spectral ringing typical to any spectral dynamical core are visible in precipitation. Such improvements among others can further lead to an overall improvement in the simulation of precipitation. Work on adapting ICON over Indian monsoon region is currently underway.