論文ID: 2023-022
Herein, the impact of a hyperspectral sounder on a geostationary satellite (GeoHSS) in a regional numerical weather prediction system is investigated during the Baiu seasons in 2017, 2018, and 2020, including heavy rainfall cases. The reanalysis-based observing system simulation experiment (OSSE) technique uses ERA5 as the pseudo-truth atmospheric profile. Temperature and relative humidity pseudo-observations are generated using the one-dimensional variational retrieval scheme based on the spectral characteristics of the GeoHSS. Verification against radiosondes shows improvements at various altitudes and forecast times (FTs). Although wind pseudo-observations are not assimilated, winds are also impacted through assimilation cycles and forecasts. The precipitation forecasts also show an improving trend with a notable impact to extend the forecasts' lead time. Case studies show impacts on precipitation primarily during longer FTs, accompanied by an improved prediction of depressions on the Baiu front and upper-level troughs. These are due to large-scale impacts from the pseudo-observations with a comprehensive coverage over clear-sky areas, propagating to precipitation areas through the assimilation cycle and forecasts. However, the prediction of an event of small-scale localized heavy rain is insufficient even at short forecast ranges owing to a limited spatial resolution. Experiments show that extracting information in the lower atmosphere is critical, and that the impact on upper-level environments is sensitive to using observations in cloudy areas.
