Hourly assimilation of different sources of observations including satellite radiances in a mesoscale convective system case during RELAMPAGO campaign

Abstract

This paper evaluates the impact of assimilating high-resolution surface networks and satellite observations using the WRF-GSI-LETKF over central and north eastern Argentina where the surface and upper air observing networks are relatively coarse. A case study corresponding to a huge mesoscale convective system (MCS) that developed during November 22, 2018 was used. The accumulated precipitation associated with this MCS was quite high, exceeding 200 mm over northern Argentina and Paraguay. The MCS developed during the Intense Observing Period (IOP) of the Remote sensing of Electrification, Lightning, And Mesoscale/microscale Processes with Adaptive Ground Observations (RELAMPAGO) field campaign. The GSI-4DLETKF data assimilation package is used to produce analyses by assimilating observations every hour with 10-km horizontal grid spacing and a 60-member multiphysics ensemble. Four assimilation experiments are conducted using different sets of observations: CONV, consisting of conventional observations from NCEP's prepBUFR files; AWS, combining CONV and dense automatic surface weather station networks (AWS), SATWND, combining AWS with satellite-derived winds, and RAD, including SATWND; and satellite radiances from different microwave and infrared sensors. The assimilation of observations with high temporal and spatial frequency generates an important impact on the PBL, primarily on the precipitable water content, that leads to the development of deep convection and heavy precipitation closer to the observed in this case study. The assimilation of radiance observations produces a better development of the convection mainly during the mature state of the MCS leading to an increase in the accumulated precipitation. Ensemble forecasts initialized from each experiment were also simulated to evaluate their skill to predict precipitation. The hourly assimilation of the observations in AWS, SATWND, and RAD helped to improve the precipitation forecast.