Performance Analysis of the CMA-BJ Model for the "23·7" Extreme Heavy Rainstorm simulation

Typhoon “Doksuri” weakened to a low-pressure system in the north, and it was blocked by a high-pressure ridge and moved slowly. From 29 July to 2 August 2023, an extremely heavy rainstorm event occurred in Beijing-Tianjin-Hebei region and lasted for a long time. Record-breaking cumulative precipitation occurred at individual stations, causing severe flooding. Based on precipitation observations, this paper comprehensively evaluates the CMA-BJ v2.0 performance on the simulation of the precipitation event. Spatial distribution characteristics of the cumulative precipitation and precipitation period in the CMA-BJ v2.0 forecast are consistent with observations. The rainfall overall was overestimated, and the D02 forecast is closest to the observation. When the GFS (Global Foercast System) global background field is used, the rain belt moved further south, and high precipitation area appeared in central and southern Hebei province, greatly deviating from the observations. In contrast, the ECMWF-driven forecast is closer to the observation. Compared to results of the cold-start run and single-time assimilation run, wind speed in the multiple-assimilation run keeps increasing, and brings in a large amount of water vapor. As a result, the distribution of precipitation is more consistent with observations. The choice of boundary layer parameterization scheme has certain impacts on the forecast of wind field, and affects the distribution and magnitude of precipitation. The YSU scheme performs better than the ACM2 and Boulac schemes in the simulation of this precipitation event.