Characteristics of initial perturbations in the ensemble prediction system of the Korea meteorological administration

Jun Kyung Kay, Hyun Mee Kim

Research output: Contribution to journalArticlepeer-review

16 Citations (Scopus)


In this study, the initial ensemble perturbation characteristics of the new Korea Meteorological Administration (KMA) ensemble prediction system (EPS), a version of the Met Office Global and Regional Ensemble Prediction System, were analyzed over two periods: from 1 June to 31 August 2011, and from 1 December 2011 to 29 February 2012. TheKMAEPS generated the initial perturbations using the ensemble transform Kalman filter (ETKF). The observation effect was reflected in both the transform matrix and the inflation factor of the ETKF; it reduced (increased) uncertainties in the initial perturbations in regions with dense observations via the transform matrix (inflation factor). The reduction in uncertainties is generally governed by the transform matrix but locally modulated by the inflation factor. The sea ice significantly affects the initial perturbations near the lower boundary layer. The large perturbation energy in the lower stratosphere of the tropics was related to the dominant zonal wind, whereas the perturbation energy in the upper stratosphere of the winter hemispheres was related to the dominant polar night jet. In the early timeintegration stage, the initial perturbations decayed in the lower troposphere but grew rapidly in the mid- to upper troposphere. In the meridional direction, the initial perturbations grew greatest in the northern polar region and smallest in the tropics. The initial perturbations maintained a hydrostatic balance, especially during the summer in both hemispheres and during both the summer and winter in the tropics, associated with the smallest growth rates of the initial perturbations. The initial perturbations in theKMAEPS appropriately describe the uncertainties associated with atmospheric features.

Original languageEnglish
Pages (from-to)563-581
Number of pages19
JournalWeather and Forecasting
Issue number3
Publication statusPublished - 2014

All Science Journal Classification (ASJC) codes

  • Atmospheric Science


Dive into the research topics of 'Characteristics of initial perturbations in the ensemble prediction system of the Korea meteorological administration'. Together they form a unique fingerprint.

Cite this