Evaluation of the Assimilation Effectiveness of YunYao GNSS-RO Constellation Data in a Global Numerical Weather Forecasting System

Global Navigation Satellite System (GNSS) radio occultation (RO), which is unaffected by clouds, precipitation, or other weather conditions, can provide uninterrupted vertical temperature and humidity profiles from the troposphere to the stratosphere. This makes it a high-quality data source for meteorological data assimilation and highly effective at reducing errors in numerical weather prediction models. In this study, we incorporate data from Yunyao’s own GNSS RO constellation (22 satellites) into NOAA’s Global Forecast System (GFS) and its three−/four-dimensional variational data assimilation framework, Grid Point Statistical Interpolation (GSI), using existing open-source atmospheric, spaceborne, and ground-based meteorological datasets. We designed sensitivity experiments to assess the incremental impact of the Yunyao RO constellation on GFS assimilation performance. The results show that: (1) After assimilating RO data, the bias in the initial analysis of geopotential height in the mid-to-upper atmosphere (above 600 hPa) is significantly reduced—mean absolute error decreases by approximately 2 gpm compared to the European Centre for Medium-Range Weather Forecasts reanalysis version 5 (ERA5). (2) In the upper troposphere above 600 hectopascals, temperature biases over the Southern Hemisphere and tropics are reduced—mean absolute error compared to ERA5 decreases by roughly 0.01 K. (3) When evaluated by region over both land and ocean throughout the entire atmospheric column, geopotential height errors decrease globally, while temperature errors improve notably in the tropics and Southern Hemisphere. Going forward, we plan to continuously assimilate multiple sources of satellite meteorological data, increase model resolution, and refine parameterization schemes to further enhance weather forecasting accuracy.