Exploiting Aeolus winds in a regional numerical weather prediction model

Aeolus was the first Doppler wind lidar (DWL) to measure wind profiles from space. Aeolus has proven to be beneficial for global models, used by meteorological centres around the world. Demonstrating added value for regional models has been less successful so far. Here, we continue with earlier impact assessments with the HARMONIE–AROME limited-area numerical weather prediction (NWP) model over the Nordic region. In this study, two limited-area domains are considered: one includes an Arctic region and the other a region covering part of the Atlantic and most of Western Europe. Statistically significant impact could be demonstrated best for the Atlantic domain and is overall neutral for winds determined from the Rayleigh backscattered signal to positive for winds determined from Mie backscatter. Results presented here demonstrate overall improved large-scale dynamics in the upper part of the atmosphere and improved temperature, which is related to the large-scale dynamics as well. Limitations of the existing data assimilation system were revealed, including the lack of an Aeolus bias-correction scheme and an imperfect implementation of the first-guess check, as part of observation quality control. The latter limits a correct assessment of the observation-error inflation technique. Recommendations for further improved usage of Aeolus winds in regional NWP are given to prepare NWP centres for a potential operational Aeolus follow-on mission.