Improving the Low-Level QC Procedure for GNSS-Radio Occultation Assimilation and Its Impact on Heavy Rainfall Prediction

The new generation of GNSS-RO data has a high penetration rate in the low-level atmosphere and is expected to have broad applications through understanding the role of moisture transport in the planetary boundary layer. However, the RO data possesses a negative bias below the 1 km height, hindering RO assimilation's impact in the lower atmosphere. The RO refractivity bias is mainly attributed to ducting, which is commonly caused by the existence of an inversion layer, such as the areas of cold SST.

This study examines the occurrence and characteristics of RO refractivity bias from a mesoscale perspective. We improve the low-level QC procedure for assimilating the refractivity atm profiles by removing the biased data. The newly proposed QC procedure considers ducting using the information from the ECMWF forecast, and the data below the ducting layer is removed. We conducted three data assimilation experiments using the WRF-LETKF system to evaluate the impact of RO refractivity assimilation. In addition to an innovation-based QC procedure, the REFALL, DQC, and LSWQC experiments assimilate the RO refractivity data without QC, ducting-associated QC and LSW-based QC procedures. The impact of different RO QC procedures is investigated based on a heavy rainfall event in Taiwan on 22 May 2021.

Focusing on the RO profiles over the South China Sea (SCS), our results show that the new QC procedure can reduce the dry moisture correction derived from assimilating the biased RO data and thus, the moisture transport by southwesterly can be enhanced. As a result, the forecast initialized from the DQC analysis generates more heavy rain in SCS. In comparison, the LSWQC tends to remove more RO data, and the impact of RO assimilation on moisture and rainfall prediction is dimmed.