Data assimilation experiments of precipitable water vapour using the LETKF system: intense rainfall event over Japan 28 July 2008

Abstract

We performed data assimilation experiments using data from an intense rainfall event in western Japan on 28 July 2008 with the Local Ensemble Transform Kalman Filter (LETKF) and analysed fields of the mesoscale four-dimensional variational assimilation (4DVAR) system of the Japan Meteorological Agency (JMA). We supplemented conventional observation data with precipitable water vapour (PWV) data derived from the Global Positioning System (GPS) Earth Observation Network of the Geospatial Information Authority of Japan. Because the LETKF system assimilates fewer data than the 4DVAR system, ensemble mean fields of the LETKF cycle experiment were replaced with analysed fields from the 4DVAR system each day at 12 UTC. PWV values were converted to relative humidity profiles for assimilation by the LETKF. The addition of PWV data tended to increase low-level water vapour and improve the precipitation forecast. We attempted to reproduce the intense rainfall band using downscale forecast experiments with the JMA non-hydrostatic model (JMANHM) with grid spacings of 5 and 1.6 km. The experiment with 5-km resolution generated a rainfall band in western Japan that was not reproduced using conventional data, although the rainfall was smaller than observations. The experiment with 1.6-km resolution faithfully reproduced the observed band of intense rainfall.