ENG

Top

R&D

Korea Institute of Atmospheric Prediction Systems

R&D

Model Development
Ensemble-based Data Assimilation

Model Development

Both high-resolution uniform grid and variable resolution grid on the cubed sphere are introduced to better represent weather process over Korean Peninsula and neighboring regions, and will be adopted to the existing non-hydrostatic dynamic core using spectral element method. The sub-grid scale physics will be upgraded to meet with the varying grid spacing ranging from 12km down to 1km, covering the grey zone.
To extend the predictability up to 30 days in advance, the model will be coupled with various earth system components including land surface, ocean, sea ice, and chemical processes.
Software engineering includes parallelization of codes, couplers, workflow management in the supercomputer environment.

Ensemble-based Data Assimilation

The hybrid-4DEnVar system is being developed in a number of directions, including adding a scale-dependent horizontal localization capability and adding further increment variables to support assimilation of "all-sky" satellite radiances and radar reflectivity observations. The LETKF-based ensemble will be enhanced by adding support for model-space vertical localization, and by improving the inflation and stochastic perturbation schemes.
To support medium-range forecasts using the future KIM Earth system model, we are also developing systems with weakly-coupled atmosphere, land surface, sea ice, and ocean data assimilation.
International Cooperation