Assimilating WIVERN windpseudo-observations in WRF model: an application to the outstanding case of the Medicane Ianos

Accurate weather forecasts are important to our daily lives. Wind, cloud and precipitation are key drivers of the Earth's water and energy cycles, and they can also pose weather-related threats, making the task of numerical weather prediction (NWP) models particularly challenging and important.The Wind Velocity Radar Nephoscope (WIVERN) mission will be the first space-based mission to provide global vertical profiles of winds within clouds, and to deliver simultaneous observations of winds, clouds, and precipitation with unprecedented resolution and coverage. The mission has been selected as the European Space Agency's (ESA) Earth Explorer 11 within the Future Earth Observation (FutureEO) programme. Its data could be beneficial to several sectors: improving our knowledge of weather phenomena, validate climate statistics, and enhancing NWP performance. This paper aims to contribute to the last point by analyzing the impact of assimilating WIVERN Line of Sight (LoS) winds on NWP performance for the high-impact case study of Medicane Ianos, which occurred between 15 and 21 September 2020 in the central Mediterranean and made landfall on the west coast of Greece.To this end, we generate WIVERN pseudo-observations, that are assimilated in the Weather Research and Forecasting (WRF) model run at moderate horizontal resolution (4 km).Results show that assimilating WIVERN into the WRF model has a positive impact on the prediction of the Medicane trajectory. Specifically, assimilating WIVERN just once reduces the trajectory forecast error by about 40 %. The data assimilation of WIVERN pseudo-observations affects not only the storm's trajectory but also its physical characteristics. It is also shown that the assimilation improves the prediction of precipitation and surface winds, and has the potential to improve our resilience to severe weather events by enabling better forecasts of storm impacts. Finally, we present the results of two sensitivity experiments in which the background and observation errors were changed. The results show greater sensitivity to changes in the background error matrix.