Assimilation of Nationwide and Global GPS PWV Data for a Heavy Rain Event on 28 July 2008 in Hokuriku and Kinki, Japan

Abstract

On 28 July, 2008, a local heavy rainfall occurred over the Hokuriku and Kinki districts, in central Japan. A stationary front existed near the heavy rainfall areas, and the atmospheric instability was increased by the inflow of the upper-level cold air.
The operational mesoscale model (MSM) of the Japan Meteorological Agency (JMA) could not predict the heavy rainfall. The comparison of distributions of precipitable water vapor (PWV) between the initial condition of the MSM and the Global Positioning System (GPS) revealed that the MSM initial condition underestimated water vapor in the heavy rainfall areas because of a positional lag of a low-level convergence zone.
Data assimilation experiments of GPS-derived PWV were conducted with JMA’s mesoscale 4 dimensional variational assimilation system (Meso 4D-Var). The PWV derived from the nationwide ground GPS network (GPS Earth Observation NETwork: GEONET) improved both the northward position error of the low-level convergence zone and the forecast of the observed rainfall. Moreover, further improvements were obtained when the PWV derived from the GPS stations of the International GNSS Service in East Asia were added.