Abstract FORMOSAT-7/COSMIC-2 was launched in June 2019 and provides high spatial resolution radio occultation (RO) profile data. The data can be beneficial for weather and climate studies and increase the accuracy of weather forecasts. This study aims to assess the two-year COSMIC-2 RO-PWV (precipitable water vapor) accuracy. Ten years of GNSS-PWV data inverted from PPP (precise point positioning) are better than those from the DD (double difference) method when taking radiosonde-observed data as the true reference values in Taiwan. Hence, in this study, GNSS-PWV from PPP is compared with COSMIC-2 RO-PWV to verify the accuracy of RO-PWV. Removing the elevation of the lowest point of an RO event above 5 km can prevent an RO event with an error exceeding 40 mm. The mean GNSS-PWV is higher than the mean RO-PWV, which are 45.7 mm and 39.9 mm, respectively, from 2020 to 2021. RO-PWV derived from COSMIC-2 in 2020 and 2021 is underestimated when taking GNSS-PWV as a reference. The mean RO error is 6.8 mm, and the mean error on land is higher than that on the ocean, with magnitudes of 8.0 mm and 6.6 mm, respectively. The higher RO-PWV accuracy on the ocean than on land is one of the reasons for the higher RO-PWV accuracy for RO events matched at the Japan GEONET stations (99.2 % on the ocean; mean error of 6.1 mm) than that at the Taiwan CWB stations (68.2 % on the ocean; mean error of 7.7 mm). In the monthly mean RO-PWV data from 2020 to 2021, the RO-PWV and its error show seasonal variations, with high values in summer (51.0 ± 9.2 mm) and low values in winter (31.9 ± 5.0 mm). The seasonal trend of RO-PWV is consistent with that of temperature. Here, we verify that the accuracy of COSMIC-2 RO-PWV is comparable to that of GNSS-PWV.
Accuracy verification of the precipitable water vapor derived from COSMIC-2 radio occultation using ground-based GNSS
Advances in Space Research ; 73 , 9 ; 4597-4607
2024-01-21
11 pages
Article (Journal)
Electronic Resource
English
Sounding the Atmosphere: Ground Support for GNSS Radio-Occultation Processing
Online Contents | 2006
|