Sub‐Seasonal Impact of GNSS Radio Occultation Data

Abstract This study investigates for the first time whether the assimilation of the Global Navigation Satellite System Radio occultation (GNSS‐RO) data has a significant impact on the European Centre for Medium‐Range Weather Forecasts sub‐seasonal forecasts. The impact was measured by running two large sets of 32‐day ensemble re‐forecasts over the extended winter periods from 2020 to 2023 which were initialized from an analysis with or without GNSS‐RO assimilated. Results indicate a difference of 4 m/s in zonal wind at the equator in the two analyses and a statistically significant improvement in the reforecast skill scores up to week 4 in the stratosphere, particularly over the Tropics when assimilating GNSS‐RO. The impact in the troposphere is generally negligible. However, the amplitude of the Madden Julian Oscillation (MJO) is significantly stronger during the first 2 weeks when the reforecasts are initialized from the analysis with GNSS‐RO assimilated, suggesting a potential link between MJO prediction and the initialization of the stratosphere.