Enhancing FY-3E/VASS Humidity Profiles During a Cloudy Atmospheric River Event Using Clear-Sky FY-4B/GIIRS Guidance

Abstract

Atmospheric rivers (ARs) are characterized by high water vapor content and are frequently accompanied by extensive cloud cover, creating a need for reliable humidity profiles under cloudy conditions. Fengyun-4B/Geostationary Interferometric Infrared Sounder (FY-4B/GIIRS) products provide relatively accurate clear-sky humidity profiles, whereas Fengyun-3E/Vertical Atmospheric Sounding System (FY-3E/VASS) products offer all-sky coverage but retain appreciable humidity errors. This study develops a clear-sky FY-4B/GIIRS-guided framework to enhance cloudy-sky FY-3E/VASS humidity profiles. Spatiotemporally collocated clear-sky profiles are first used to construct humidity residuals, from which principal component analysis extracts a low-dimensional residual basis. A neural network then predicts the corresponding coefficients to generate a baseline correction. Target-day calibration and cloud-condition adjustment further adapt this clear-sky-derived correction to cloudy-sky samples on the target day. The framework is evaluated against ERA5 reanalysis for a preidentified AR event along the eastern coast of China on 6 July 2023, using cloudy-sky samples from the 09:31–09:41 UTC FY-3E overpass. For all cloudy-sky samples from this overpass, the pressure-level-averaged root mean square error (RMSE) decreases by 16.1% over 300–900 hPa and by 20.8% over 700–900 hPa. For cloudy-sky samples in the high-moisture AR subset, the corresponding reductions are 19.7% and 26.6%, respectively. The overlap between the high-moisture masks defined using integrated water vapor (IWV) from FY-3E/VASS and ERA5 increases from 60.7% to 73.9%. The enhanced profiles therefore show improved agreement with ERA5, particularly over 700–900 hPa, and better represent the high-moisture portion of the examined AR event.