Comprehensive Evaluation on Atmospheric Motion Vectors from Fengyun-4B Geostationary Satellite and Their Application in the South China Sea Monsoon Onset

Since the Chinese new-generation Fengyun-4B (FY-4B) geostationary satellite Atmospheric Motion Vector (AMV) products became operational in June 2022, there has been limited research on their accuracy and usability. This study comprehensively evaluates FY-4B AMV products for August and October 2023, and January and April 2024, exploring their application in monitoring the South China Sea summer monsoon (SCSSM) onset. The results indicate that AMV products from the upper-level water vapor absorption channel (AMV_WV) and the infrared channel (AMV_IR) demonstrate high accuracy when matched with ERA-5 data, with RMSE (root mean square error) mostly between 4.5 m/s and 6.4 m/s, and R2 (coefficient of determination) values ranging from 0.7 to 0.8, indicating the overall reliability of FY-4B AMVs. The observation errors of AMVs exhibit significant vertical structure characteristics. Specifically, the AMV_WV has significantly lower observation errors above 350 hPa, while the AMV_IR shows lower observation errors between 200 hPa and 500 hPa, and between 700 hPa and 950 hPa. Spatially, most areas exhibit low observation errors for AMVs; however, cloudless weather and deep convective cloud systems can increase errors. A lack of clouds or water vapor may reduce the number of observation samples in some areas, leading to unstable RMSE performance, particularly evident for AMV_WV RMSE around 25°-30°N in January and near 25°S in August. Deep convective cloud systems can influence AMV retrieval results, leading to systematic observation errors, especially for the infrared channel. Additionally, AMV_WV is more reliable during the daytime, with a lower RMSE compared to nighttime, while AMV_WV exhibits a diverging diurnal variation pattern. Finally, the FY-4B AMV_WV products were applied to monitor the SCSSM event in 2024. Significant zonal wind direction conversion characteristics were observed in key regions around the onset date, indicating that AMVs can serve as indicators for monitoring the SCSSM onset.