AOPP Seminar - Exploiting the information from advanced satellite sounders for numerical weather prediction

The Met Office acquire everyday hundreds of millions of observations which undergo strict quality control procedures. A subset of them is selected and then assimilated in the operational numerical weather prediction (NWP) model. A brief theoretical description of the variational data assimilation system used at the Met Office is provided, as well as the results of a data denial experiment that was recently performed to identify the most impactful observation types in the operational system. Also, I will introduce our move to using the Joint Effort for Data assimilation Integration (JEDI) system for observation processing and assimilation. The second part of the talk will compare the results from the assimilation of satellite radiances and of optimal retrievals (Transformed Retrievals) over selected assimilation trials. Our results show for the first time that the two techniques lead to equivalent forecast skill benefits in an operational NWP system. These results pave the way to a significant increase in the number of radiance channels from advanced hyperspectral sounders that can be used for operational assimilation. However, more research is needed on bias correction and optimal uncertainties for Transformed Retrievals.