Radiation Meeting 11‐10‐2023

Attendees: Leif Denby, Emily Gleeson, Karl-Ivar Ivarsson, Daniel Martin, Peter Ukkonen, Laura Rontu.

Some Notes:

Peter has been working on ecRad in CY49t0 - this version includes optimised gas optics (Peter and Riad were involved in porting this). Uncertain whether the code will have CAMS nrt aerosols configured (Emily to contact Daniel re CY49h). Peter is interested in using NNs for determining the cloud particle size distributions (currently constants are used and are not optimal - see paper by Sebastián et al.). The size distribution can be used in the calculation of effective radius. Ana and Jan are currently working on externalising the effective radius calculation from radiation (will include graupel, snow etc in the structure of the calling subroutine, as Karl-Ivar previously suggested).

Leif has been carrying out aerosol experiments and writing useful tools to set-up and run experiments in batch. Here is a cheat sheet he has created of a HARMONIE flow chart: https://github.com/leifdenb/HarmonieCheatSheet. He has found some errors in the FA to grib and grib to vfld conversions - will submit a PR on this.

Laura has code yet to be committed to CY46, mainly regarding the use of CAMS MMR climatology - more testing and cleaning to be done before opening a PR. Some slides were shown by Laura on the impact of aerosols (CAMS NRT are configured to the microphysics and radiation; TEGEN and CAMS climatology only configured to radiation). She also has code where the CAMS NRT aerosols are only configured to radiation (thus removing the impact on microphysics) and a CAMS MMR climatology that can be configured to both radiation and microphysics - by default when we use an aerosol climatology, the CDNC used in microphysics is not consistent with that but instead either constants or a profile based on land/sea are used. Slides: https://drive.google.com/file/d/1vgvo-34_PVfDEfMIgpXNv7cPtetDVXRP/view?usp=drive_link Laura highlighted the complexity of our system and the need for us to simplify what we need to do with it. What do we need to get right? What are the sensitivities. Peter shared the following by ECMWF which nicely shows the sensitivity of various parts of the radiation scheme - see page 12 of https://www.ecmwf.int/sites/default/files/elibrary/2016/16901-ecrad-new-radiation-scheme-ifs.pdf.

Daniel showed the following slides on cloud overlap: Slides: https://drive.google.com/file/d/10Ks3MFeUKk7GJdteybW47Khif5RusAvm/view?usp=drive_link. In a bid to address the overeastimation of SWD when NRT CAMS aerosols are used, he tested NOVLP=4 (exponential random) compared to NOVLP=6 (default, max random) - NOVLP=4 gives more clouds and less SWD as required. Laura mentioned that the cloud overlap influences diagnostics and and radiation (via the cloud-radiation interaction). Karl-Ivar will present on this too at the next microphysics meeting in November.

We also discussed the writing of a CY46 paper, including future things related to radiation (NRT aerosols, ecRad etc) and how to have enough new material to ensure the paper is accepted.