Research on PWV fusion based on ERA5 PWV and FY-4A PWV in Chinese mainland

High precision and high spatiotemporal resolution of precipitable water vapor (PWV) serve as fundamental support for meteorological research, severe weather forecasting and warning systems, as well as long-term climate change trend analysis. The FY-4A PWV has been widely utilized in PWV research due to its high observational accuracy and superior spatial resolution. However, there are many missing values in the FY-4A PWV due to the influence of weather and cloud cover. The ERA5 PWV has the advantages of high temporal resolution, complete data, rapid updates, and a multitude of parameters, but its spatial resolution is relatively low at 0.25° × 0.25°. Therefore, the focus of this study is to integrate the advantages of multiple datasets to obtain high-precision PWV data with superior spatiotemporal resolution and completeness. This study takes the Chinese Mainland as research area, the ERA5 PWV was interpolated to the same spatial resolution as the FY-4A PWV (4km × 4 km) using the cubic convolution interpolation, and the interpolated ERA5 PWV was used to fill in the missing values of FY-4A PWV. Based on their respective precision weights, the datasets were integrated through the fusion model to generate a complete and high-accuracy PWV product. The results show that more than 55 % of the interpolated ERA5 PWV has a correlation coefficient of more than 0.9 with the GNSS PWV, and the difference in the ERA5 PWV accuracy before and after interpolation is within 3 mm in each region. By establishing regional and seasonal data fusion model, the RMSE of the PWV after fusion can be relatively improved by 0.34 %–24.96 %,0.36 %–29.39 %,0.78 %–27.94 %,0.78 %–34.31 % for the four seasons, respectively. The fusion effect is better in spring and winter, and the fusion PWV error is basically within 4 mm. This study achieves the acquisition of high-precision PWV data with a spatial resolution of 4km × 4 km on an hourly basis in Chinese mainland. The fused PWV can provide an important database for further applications and research.