Impacts of Virtual Train-based Atmospheric Vertical Profile Data Assimilation on the Forecast of the "21.7" Zhengzhou Rainstorm

A record-breaking prolonged and extreme rainstorm occurred in Henan province, China during 18–23 July 2021. Global and regional numerical weather prediction (NWP) models consistently underpredicted both the 24-h accumulated rainfall amount and the 1-h extreme precipitation in Zhengzhou city. This study examines the potential impacts of data assimilation (DA) of atmospheric vertical profiles based on the train-based mobile observation (MO) platforms on precipitation forecasts. The research involved assimilating virtual train-based air temperature (Ta), relative humidity (RH), U and V components of wind profile data based on the ERA5 reanalysis datasets into the Weather Research and Forecasting (WRF) model using three-dimensional variational (3DVar) method. Analysis confirms the reliability of Ta, RH, and wind speed (WS) profiles from ERA5 reanalysis datasets. The assimilation of virtual train-based moisture profiles enhanced the RH analysis field. Furthermore, the forecasts more accurately represented the coverage and intensity of the 6- hour and 24-hour accumulated precipitation, as well as areas with maximum rainfall durations exceeding 20 hours. The threat score(TS) and bias metrics for 6-h, 12-h and 24-h accumulated precipitation forecasts showed marked improvement for heavy to torrential rain in Henan province, particularly in the Central and Northern regions (hereinafter referred to region CNH). The TS for 24-h accumulated precipitation forecasts at 50 and 100 mm rainfall levels increased by 0.17 and 0.18 in Henan province, and by 0.13 and 0.18 in region CNH. During the rainstorm period, water vapor content increased substantially, with enhanced moisture transport from south of Henan province to region CNH driven by southwesterly winds, accompanied by significantly strengthened updrafts. These improvement in water vapor and upward motion ultimately enhanced the forecasts of this extreme rainstorm event.