Analyzing precipitable water vapor with GNSS radio occultation wet profiles, radiosonde observations and ECMWF ERA5 reanalysis data over the Türkiye region

Abstract The Global Navigation Satellite System (GNSS)-Radio Occultation (RO) is a satellite-based technique providing significant advantages for improving weather forecasts and investigating climatic variations. In this study, precipitable water vapor (PWV) derived from FORMOSAT-7/COSMIC-2 GNSS-RO wet profiles was compared with that calculated from radiosonde (RS) observations and the fifth-generation global reanalysis dataset (ERA5). For this purpose, RS stations situated in Samsun, Istanbul, Izmir, and Adana over Türkiye were selected as reference locations. Two horizontal collocation distances, 300 km and 150 km, were applied to examine the sensitivity of the results to spatial matching. The performance of COSMIC-2 RO PWV was evaluated using root mean square error (RMSE), linear regression, Pearson correlation, and the R-squared coefficient ( R 2 ). For both collocation distances, RMSE values are about 3 mm for RO-RS and about 2 mm for RO-ERA5, and correlation coefficients are generally above 0.8. Regression slopes are lower for RO-RS, ranging between 0.6 and 0.8, and closer to unity for RO-ERA5, ranging between 0.8 and 1.0. Reducing the collocation distance leads to a small increase in correlation. However, it does not result in a consistent improvement in RMSE, indicating the influence of vertical sampling characteristics. Overall, GNSS-RO shows reasonable consistency with RS observations. When used in combination with reanalysis products, it can provide reliable atmospheric moisture estimates over Türkiye.