COSMO Scientific Management Committee Meeting 12 -13 September 2019, Rome, Italy

1. COSMO Scientific Management Committee Meeting 12-13 September 2019, Rome, Italy (dmitrii.mironov@dwd.de) COSMO SMC Meeting, 12-13 September 2019, Rome, Italy

2. 1. Agenda, minutes SMC meeting agenda (doc file) Minutes • Anastasia Bundel (day 1) • Flora Gofa (day 2)

3. 2. New COSMO versions Ulrich Schättler, all

4. 2. New COSMO versions (cont’d) Version 6.0, a unified version for NWP and climate modelling, is going to be the latest release of the COSMO model The release of 6.0 is expected in December 2019 (optimistic scenario) or in March 2020 (more realistic scenario) Beyond that point in time • the COSMO code is frozen, • only maintenance is provided (incl. bug fixes), • the current SCA (Uli Schättler) will take care of ICON-LAM These points must be taken into account as future work within COSMO is planned (when drafting new PPs or PTs, etc.)

5. 3. NWP test suite • Hindcast mode, lateral BCs from IFS, SST from IFS, soil from ICON-EU, 7 km DP and SP runs, 2.8 km DP runs • Boundary fields at 3, 9, 15 and 21 UTC introduced, these are short-range forecasts (stepRange=3) [Ines] • COSMO-model version 5.06 (vs. v5.05 as reference): runs completed(hindcastmode, lateral BCs 3 hrsapart) [Ines] • Verification done for SP version (loaded on the shiny server). DP version (both 7km and 2.8km) will be ready soon (expected before GM in Rome). There are still unresolved issues with precipitation. Draft report submitted 5.9.2019, all in all v5.06 passed the test [Amalia]

6. NWP test suite (cont’d) ICON-LAM TS[Ines, Daniel] • IC/BC from IFS, hindcast mode, BCs 3 hrs apart • first technical implementation nearly ready (external data prepared, etc.) • work on the IFS BCs in progress • realization of mixed initialization with the IFS atmosphere and the ICON soil has to be investigated • for NWP TS (and COSMO-LEPS) migration to eccodes foreseen within fall 2019 • ECMWF annual report submitted (63% resources already used, therefore a request for extra SBU will be done later, needed for ICON-LAM) [Amalia] • both COSMO-model last versions and ICON-LAM should be tested, extra billing units at ECMWF are required (c/o WG6, will be requested shortly)

7. 6. Model documentation, DOIs, etc.

8. COSMO model documentation • COSMO User Guide for 5.06 uploaded • INT2LM User Guide for 2.06 uploaded • "Model Output and Data Formats for I/O“ uploaded

9. DOI for COSMO Publications COSMO Technical Reports are now published with DOIs DOI format: 10.5676/dwd_pub/nwv/cosmo-tr_XX (XX is the TR number) Landing page at DWD https://dwd.de/EN/ourservices/cosmo_technical_reports/cosmo_technical_reports.htm DOIs are also provided for all Technical Reports published so far (work in progress, c/o WG6) A few latest TRs with DOI shown at the title page are available at the COSMO web http://www.cosmo-model.org/content/model/documentation/techReports/default.htm

10. DOI for COSMO Publications (cont’d) COSMO Newsletters are published with DOIs (starting with NL No. 19) DOI format: 0.5676/dwd_pub/nwv/cosmo-nl_XX_YY (XX is the NL number, and YY is the number of contribution within NL XX) Landing page at DWD https://www.dwd.de/EN/ourservices/cosmo_newsletter/cosmo_newsletters.html NB 1: NL No. 19 has more contribution than previous NL NB 2: Many Mission Reports were not submitted (but the COSMO money used), Mission Reports should give a brief overview of the meeting attended (not only of the authors’ own contribution)

11. COSMO Publications Great thanks are due to the DWD colleagues Magdalena Bertelmann and Jörg Rapp for their kind efforts in providing DOIs for COSMO publications! Apart from COSMO Technical Reports, contributions can be submitted to the “Reports on ICON” series (DWD editors are Florian Prill, Daniel Reinert and Daniel Rieger) https://www.dwd.de/EN/ourservices/reports_on_icon/reports_on_icon.html

12. 7. WGs, PPs, PTs • 15.1. WG1, KENDA-O • 15.2. WG2, EX-CELO, CCE • 15.3. WG3a, T2(RC)2, ConSAT4 • 15.4. WG3b, CALMO-MAX, AEVUS, SAINT • 15.5. WG4, current activities • 15.6. WG5, CARMA • 15.7. WG6, CEL-ACCEL, IMPACT, SPRT • 15.8. WG7, APSU

13. WG1, KENDA-O • Christoph Schraff

14. Status Report for KENDA-O (+ WG1) • Task 1: further development of LETKF scheme (conventional obs) • DWD: comparison 3-D vs. 4-D LETKF: 3-D is worse but better than expected •  3-D EnVar might be a viable option in the future • (to replace deterministic LETKF analysis) • MCH: regional static B matrix: scientific problems (formulation did not work) • now trying an alternative formulation (to compute velocity potential, streamfunct.) •  task is important for future option of regional 3DVar • HMC: technique to estimate (additive + mulitplicative) model errors • by using COSMO-2.2km (as a model) vs. COSMO-0.22km • (as the 'truth', tendencies started at the same point in phase space) • shows problems (spin-up)  method will probably not work • (however, this is not critical for the KENDA project as a whole) • Task 3: soil moisture analysis using satellite SM data: COMET fellowship ended • (no clear benefit yet, will continue tests in parallel suite, with little (!) FTE) • (however, this is not critical for the KENDA project as a whole)

15. Status Report for KENDA-O (+ WG1) • Task 2: extended use of observations: ongoing • radar radial velocity: data quality issues that have caused negative results in • pre-operational test are addressed by adjusting the obs errors • negative impacts eliminated, • now neutral to positive impact, esp. in convective periods •  will probably re-introduce in pre-operational suite soon • radar reflectivity Z: radar Z in LETKF beats LHN in tests at ARPAE, • operational introduction planned in 2020; • DWD: evaluations with ICON-LAM, and tests soon • GPS STD / ZTD: consistent small positive impact; • technical work towards operationalisation • SEVIRI IR WV: clear-sky data: small benefit in upper-tropospheric humidity; • delay for all-sky (cloudy) data; further work with ICON-LAM • screen-level obs: project since 03/19 at DWD: work on bias correction; • assimilation experiments with ICON-LAM in 2020 • various at MCH: first test cases or implementation ongoing: meteodrones, • Raman lidar, Doppler wind lidars, MW radiometer

16. Status Report for KENDA-O (+ WG1) • Task 4: adaptation to ICON-LAM (DWD): progressing well • technical implementation work (really a lot !!) of whole system • (LETKF; snow / SST analyses; ensemble perturbations of soil moisture, SST; new obs operators; latent heat nudging (LHN), nudging of soil moisture, …) • in BACY experimentation script environment (basically) finished; • observation operators implemented in ICON for online LEKTF • LETKF and model settings (e.g. IAU, upper-boundary relaxation, ..) adapted • LHN adapted to ICON-LAM, works sufficiently well in summer (convection), • needs further evaluation / tuning for winter • experiments (MEC-based LETKF): ICON-DE verifies better than COSMO-DE • (14-day summer convective period, with or w/o LHN; 2-week winter period w/o LHN: • upper-air T + wind better, T2M, RH2M, 10-m wind, much better; precip similar) • first extended experiments with 4-D online LETKF for ICON-D2 started • pre-operational suite ICON-D2 with KENDA starting Oct. 2019 • towards 3DVar / EnVar option (for deterministic run): • preliminary version of MEC-based 3DVAR + EnVar exists • COSMO obs operators in DACE + TL/Adjointt.b.d. (upcoming COSMO year) • later on: testing (regional B-matrix from MeteoSwiss project needed!)

17. WG3a, T2(RC)2, ConSAT • Matthias Raschendorfer

18. T2(RC)2 PPL: Harel Muskatel (IMS) • Work on code finished • new routine CLOUDRAD merged with latest COSMO version • possibly part of next official COSMO release • Results of the current PP: • New derivation of optical properties: including large water droplets and ice particles up to size of snowflakes -> introduced in COSMO (Ritter-Geleyn) and ICON (RRTM) • Implementation of MACv2 aerosols climatology • Implementation of prognostic ICON-ART dust and prognostic CAMS aerosols -> also as input for water and ice nucleation processes • New parameterization of water droplet effective radius, including a revised formulation of the sub-grid cloud effect • Run-time optimisations like Monte-Carlo Spectral Integration (MCSI) • New code tested and tuned against ground based and satellite data -> positive results with respect to radiation fluxes and T2m. • Project extension -> Feb. 2020: further testing and documentation

19. T2(RC)2 • New PPCAIIR: Clouds and Aerosols Improvements in ICON Radiation Harel Muskatel (PPL), Pavel Khain, Daniel Rieger, Alon Shtivelman, Alexei Poliukhov, Natalia Chubarova, Marina Shatunova, Julia Khlestova, Gdali Rivin, Uli Blahak, Barabara Fay • Mar. 2020 -> Feb 2022 • Update of the new ecRAD-RRTM in ICON with new development of T2(RC)2 • Connection of ICON with new aerosol-data similar as done for COSMO: Tanre(1983) or Tegen (1997) -> CAMS prognostic aerosols or ICON-ART or Mac-v2 climatology (Kinne 2013) or simplified 2D-advection of aerosols (G. Zängl) • New aerosol data not only for optical properties but also as condensation nuclei for microphysics (according Segal&Khain) • Verification by means of observations at Moscow State University MetObs and Lindenberg • Investigation of the new stochastic shallow-convection parameterization (Sakradzija et al. 2016) being present in ICON-> sub-grid variability of microphysical rates, precipitation from shallow convection

20. ConSAT: Status / Recent development (s. also WG3a development-page) • ConSAT-tasks delayed due to illness and temporal shift of priorities • ICON development-branch of MR with further developed Surface Coupling: TURBDIFF (further implementation of STIC with respect to near surface processes): • Optional new parameterization of Circulation-term as kata- and ana-batic near-surface circulation dependent on current SSO-parameters (thermal SSO contribution). • More realistic SAT at strong topographic inhomogeneity (mountainous regions); more testing required TURBTRAN/TERRA (improving the physical, conceptual and numerical description of SAT: • Revised TERRA with (partly optional) implicit treatment of surface temperatures (Tsf/Tsn) • Coupled implicit system of linearized heat equations for: the partial snow-free surface, the partial snow-covered surface (including the snow pack) and the soil below. • Extensive adaptation of code that was based on an explicit treatment of Tsf/Tsn by various limitations and explicit restrictions all over the code, including: • TURBTRAN (Tsf/Tsn-dependency of atmospheric transfer resistances) • interfaces and treatment of dynamic snow tiles • empirical extensions related to surface processes (implemented by G. Zängl) • Recent extension towards an implicitly coupled multi-layer (ml) heat-equation for snow: reducible to the single-layer treatment and usable for new ml snow-hydrology with individual layering for each grid point and time step (as being implemented through PT SAINT). • a) No time-step oscillations of Tsf/Tsn (even without a flux limiter); • b) Includes already the thermal part of an adapted implementation of a ml snow scheme; • c) Includes necessary preparation for the physical description of a roughness cover (canopy)

21. ConSAT <-> PT SAINT • Short-term aim: Providing a configuration of the new ICON-code that recovers the now operational results as far as possible. • With regard to current tests, this aim seems already be reached. • But further tests are still pending. • Next steps: • A new ‘commit’ of this code and a ‘merge’ with the official development branch of ICON by US + MR: next few weeks • This code will already be formally adapted to the coding of TERRA, TURBTRAN and TURBDIFF in the recent COSMO version and can just be copied to COSMO. • The code contains the implicit formulation of surface processes including the implicitly coupled ml heat-equation for snow. • Transferring the CALL of multi-layer snow hydrology (based on SNOWPACK) into this version within the framework of COSMO by Sascha Bellair (SB) + US. • Related additional code in TERRA should be active only for COSMO via IFDEF. • Connection of this further extended TERRA with a child branch of MR’s development branch in ICON, where it can be ‘merged’ with possible adaptations resulting from MR’s tests in order to go back to SB’s development branch in COSMO thereafter. • Full implementation of the ml snow-hydrology in all modules of COSMO and (later also of) ICON.

22. ConSAT <-> implicit TERRA and canopy treatment in ICON • Short term aim: It is planned to have MR’s development code in the official development branch as soon as testing is finalized, proving that the code can be configured so as to provide similar results as the operational code: expected by the end of this year. • For that purpose, possibly some parts of the whole development need to be switched off intermediately until they are ready for operations (in terms of improving the verification scores). • The related TERRA code will also be used in COSMO and contains the code for the new ml snow treatment, which, however, is deactivated until its full implementation is ready. • Possibly, the so far operational code remains as an intermediate option as a fall back and in order to run the current ml snow scheme for comparison. • Next steps: • Extension of the implicit treatment to phase transitions liquid<->solid also with regard to precipitation (including interception of frozen water), soil water and snow. • Extension of the included skin-layer formulation of the snow-free surface towards the treatment of a full semi-transparent canopy-layer (consisting of massive roughness elements) based on a proto-type that is already present in an older test-version of TERRA in COSMO, and which already showed an almost perfect mean diurnal cycle of Td_2m for a summer day. • As an intermediate option, the so far implemented real skin-layer resistance of the soil can already now be substituted by a parameterized virtual resistance, as Jan-Peter Schulz is currently dealing with.

23. 13. “Adaptation of SMC” document Current version of the “Adaptation of SMC” document here Decision is to have the ICON Git repositories with the “icon-cosmo-master” branch

24. 15. New PPs and PTs

25. Extension of PP EX-CELO (WG2) and PP CEL-ACCEL (WG6) Titles: “Extension of COSMO-EULAG operationalization” and “COSMO-EULAG on accelerators” Leader: Zbigniew Piotrowski (IMGW) Duration (start and end dates): March 2017 – September 2019, extension until March 2020 Reason for extension: PPs were delayed due to unavailability of human resources, the project teams should finish the work planned FTEs: no extra FTEs Participants: IMGW (EX-CELO), IMGW and MCH (CEL-ACCEL) Status: extension of both PPs approved by the STC 6 September 2019, work in progress

26. Extension of PPs EX-CELO and CEL-ACCEL (cont’d) NB: future contributions (beyond September 2019, beyond March 2020 for EX-CELO and CEL-ACCEL participants) of the EULAG team to the ICON-LAM development, testing, performance analysis, etc., is being extensively discussed. Suggestions are welcome!

27. PT AEVUS2 (WG3b) Title: analysis and evaluation of TERRA_URB scheme 2 Leader: Paola Mercogliano (CIRA) Aim: Consolidate the implementation of the TERRA_URB scheme in the COSMO model, draft a new PT or PP aiming at transferring these developments into the ICON model. Duration (start and end dates): September 2019 – August 2020 FTEs: 1.1 (COSMO) + 0.6 (external partners) Participants: CIRA, ARPA-Piemonte, RHM, DWD, Ruhr University of Bochum, Flemish Institute for Technological Research, Politecnico di Torino Status:decision is expected on 10 September 2019 (STC meeting in Rome) The PT proposal was comprehensively and very vividly discussed at the SMC level.

28. PT AEVUS2 (cont’d) • NB: PT AEVUS2 is very much in line with the COSMO strategy as it appropriately accounts for specific features (and the associated difficulties!) of the transition period, viz., • the reduced rate of development of the COSMO model and its limited life expectancy, and • the contribution to the development of the ICON model into which the consortium is getting increasingly involved.

29. PP AWARE (WG4 and WG5) Title: Appraisal of "Challenging WeAther" FoREcasts Leader: Flora Gofa (HNMS) and Anastasia Bundel (RHM) Aim: to try out a number of forecast methods and evaluation approaches that are linked to high impact weather (not necessarily considered extreme to all users) and to provide COSMO Community with an overview and recommendations as to how Challenging Weather/High Impact Weather situations should be handled. Duration (start and end dates): September 2019 – August 2021 FTEs: 4.15 Participants: HNMS, RHM, ARPAE-SIMC, DWD, IMGW-PIB, MCH, NMA Status: approved by the SMC, final decision is expected on 10 September 2019 (STC meeting in Rome) Note: PP addresses cross-cutting issues, joint effort of WG4 and WG5 with the contribution from WG7

30. PP CAIIR (WG3a) Title: Clouds and Aerosols Improvements in ICON Radiation Scheme Leader: Harel Muskatel (IMS) Aim: to address various issues regarding cloud optics, aerosols-radiation interactions, aerosols-cloud interactions and shallow convection topics with the aim to further improve the ecRAD radiation parameterization scheme recently implemented into ICON; comprehensive testing of the new code will be performed, results for several locations will be compared with observational data. Duration (start and end dates): March 2020 – February 2022 FTEs: 5.95 Participants: IMS, DWD, RHM Status: the PP proposal is discussed by the SMC, STC decision is due in March 2020 NB: beneficial for ICON development [i.a. results of PP T2(RC)2 will be transferred to ecRAD in ICON]; PP T2(RC)2 is extended until March 2020

31. PP CAIIR: Comments by Martin Axel, Sophia Schäfer (our radiation expert at DWD) and me have already commented in detail on the CAIIR project starting March 2020 for 2 years and focusing on clouds and aerosol improvements in ICON - in particularly related to the new ecRad radiation scheme. In general the reception was very positive on the proposed work on model development and comparison to observations. I don't want to repeat the detailed comments but give a broad recommendation of how the proposal draft could be brought to a version leading to a successful project. First of all Sophia made numerous text corrections and suggests on the document that are valuable to consider and include (see attachment). Task 1: cloud ice optics The agreed approach is to use a common form of lookup tables to specify the optical properties within ecRad. That format is developed by Robin Hogan with input of Sophia Schäfer. The CAIIR project should produce a consistent table for there extension to large ice particles. Task 2: cloud liquid optics Another set of liquid optics would be interesting. The lookup table again should be consistent to ecRad developments.

32. PP CAIIR: Comments by Martin (cont’d) Task 3: Segal & Khain effective radius in ICON These concepts were developed for the COSMO cloud cover scheme which adds grid-scale (GS) and sub-grid scale (SGS) clouds. ICON has a unified cloud cover scheme. It is not clear to me how to implement that in ICON, but ideas could develop during this project. Task 4: CAMS prognostic aerosols from ECMWF This requires 4D aerosol from ECMWF Copernicus Services forecasts. Would be very interesting but not sure if feasible. Task 5: 2D aerosol advection scheme This work may be done by Daniel Rieger. The link to CAIIR is as yet unclear. Task 6: Kinne's Macv3 aerosol climatology Barbara Fay will implement CAMS aerosol climatology, which could be mentioned. Task 7: ICON vs surface radiation observations Good idea. We will add comparisons versus the radiation reference stations BSRN.

33. PP CAIIR: Comments by Martin (cont’d) Task 8: shallow cumulus clouds Again, consistency with the ICON clouds scheme will be important. I will be happy to collaborate. Stochastic convection will also be an interesting focus here and useful for the specification of variability in the radiation scheme. Task 9: Tuning The final tuning will be most efficient at DWD with a close to operational setup and model version. But this project to point to sensitivities to parameters in some simple standard simulation. In general this project CAIRR requires careful communication with the ICON cloud and radiation developers at DWD (e.g. Sophia, Daniel, Alberto, Martin, Günther, Uli, Axel, Maike). I would suggest to choose one main contact person.

34. PP CAIIR: Comments by Bernhard Dear Martin, thanks a lot for this comprehensive summary. From KIT side I would like to add a few aspects. As you all know from previous meetings we are focussing on implementing radiative and cloud microphysical processes based on the ICON-ART aerosol. Our work is not directly under the COSMO umbrella but we should avoid double developments in the ICON code. Especially, as KIT is now responsible for the aerosol-chemistry developments and the corresponding interfaces in ICON. Therefore, I see a strong need to consolidate those developments as early as possible saving resources and avoiding unnecessary work. I think we should discuss that during the COSMO GA and the SMC meeting next week. Best wishes Bernhard

35. PP CAIIR: reply by Harel Dear Martin, Dear all, Thanks for your comments and inputs. First - please notice that the version Martin sent today is an older version of the PP. I attached the newest version which includes (most of) the corrections proposed by Sophia (CAIIR_priority_project_plan_05Sep2019.docx). This one can be used for discussions and SMC review. I also attached Sophia's version (CAIIR_priority_project_plan_Sophia_Harel_v1.docx) with answers to her comments (as much as I can in this this time restrictions). Regarding Martin's comments: Task 1: cloud ice optics The agreed approach is to use a common form of lookup tables to specify the optical properties within ecRad. That format is developed by Robin Hogan with input of Sophia Schäfer. The CAIIR project should produce a consistent table for there extension to large ice particles. Sure. I will wait for Sophia to finish the work on the lookup tables and adapt the new parametrizations accordingly. Task 2: cloud liquid optics Another set of liquid optics would be interesting. The lookup table again should be consistent to ecRad developments. As I stated in the comments to Sophia, the results of the new water clouds optics will be presented in the parallel session.

36. PP CAIIR: reply by Harel (cont’d) Task 3: Segal & Khain effective radius in ICON These concepts were developed for the COSMO cloud cover scheme which adds grid-scale (GS) and sub-grid scale (SGS) clouds. ICON has a unified cloud cover scheme. It is not clear to me how to implement that in ICON, but ideas could develop during this project. There are two parts for the work: first is to calculate Reff correctly which I found to be much more important than all other parameters. Second is Pavel's proposed to calculate SGS CLC using some adiabatic Reff but you don't have to accept that for ICON. We do not aim for a revision or for a major development of the cloud cover scheme in ICON. Task 4: CAMS prognostic aerosols from ECMWF This requires 4D aerosol from ECMWF Copernicus Services forecasts. Would be very interesting but not sure if feasible. I hope it will because it was very beneficial for COSMO radiation scheme. Where do you see possible problems? Task 5: 2D aerosol advection scheme This work may be done by Daniel Rieger. The link to CAIIR is as yet unclear. Gunther asked us to include it and it is reasonable in my opinion since half of this project is on aerosols inputs. I think it should be developed and evaluated in one framework as other aerosols inputs.

37. PP CAIIR: reply by Harel (cont’d) Task 6: Kinne's Macv3 aerosol climatology Barbara Fay will implement CAMS aerosol climatology, which could be mentioned. Martin - You write this because you didn't see the later versions (see attached - Task 5) Task 8: shallow cumulus clouds Again, consistency with the ICON clouds scheme will be important. I will be happy to collaborate. Stochastic convection will also be an interesting focus here and useful for the specification of variability in the radiation scheme. Great I will tell this to Pavel. Task 9: Tuning The final tuning will be most efficient at DWD with a close to operational setup and model version. But this project to point to sensitivities to parameters in some simple standard simulation. Yes. That's what I offered in the comments to Sophia. The question is about the ICON-IDEAL setup - is it ready to use? In general this project CAIRR requires careful communication with the ICON cloud and radiation developers at DWD (e.g. Sophia, Daniel, Alberto, Martin, Günther, Uli, Axel, Maike). I would suggest to choose one main contact person. Any volunteer? Best regards and see you soon, Harel.

38. 18. Any other business • Suggestions towards re-organization of COSMO GM … • Consider holding parallel session before plenary session on Tu. This gives all WG, PP and PT leaders a chance to summarize the outcomes of their (parallel session) meetings and adjust their plenary talks. • Avoid to long (40-60 min) talks by one speaker, better give less time to a number of speakers.

39. FTE information See COSMO web page PP FTE Tables (http://www.cosmo-model.org/content/tasks/priorityProjects/taskDetails.priv.htm?sqlPred=0:::0|1:::0) PT FTE Tables (http://www.cosmo-model.org/content/tasks/priorityTasks/ftes.htm?sqlPred=0:::0) Grand Summary (http://www.cosmo-model.org/content/tasks/priorityProjects/taskSummary.priv.htm?pp=all&year=2019) Overall looks OK

40. FTE information (cont’d)

41. COSMO Soft: FIELDEXTRA, EXTPAR, DACE, VERSUS, SNOWE FIELDEXTRA, EXTPAR, DACE, VERSUS, SNOWE VERSUS • Versus version 5.1.9 was released in May 2019 including new BUFR template for sounding observations DACE • Much code development since GM 2018, e.g. for ICON-LAM • Management: new “Model developments” page for DACE is planned, specific solution will likely follow respective ICON-LAM management pages

42. Status of non-official COSMO SW (TERRA standalone, BUFR2NETCDF, Rfdbk, CALMO) BUFR2NETCDF • Software is maintained (Davide Cesare, ARPA-SIMC) • Info at COSMO web page is up-to-date Rfdbk • Rfdbk is not a COSMO software, a version of MEC/Rfdbkis distributed by DWD as part of PPCARMA CALMO • CALMO Meta-Model is available at the COSMO web page

43. WG2 • Open issue:future (beyond September 2019) contribution of the EULAG team to the ICON-LAM development, testing, performance analysis, etc. Are there any fresh ideas or/and elaborated work plan as compared to what was discussed previously? • From the earlier discussions: • Notes • Implementation of the EULAG dynamical core into ICON does not seem to be reasonable since the ICON dynamical core already fulfils the properties of mass- and tracer-mass conservation, overall 2nd order convergence, and satisfactory treatment of steep terrain • If comparison of COSMO-EULAG and ICON is to be made, its outcome must be clearly defined • Some suggestions • Investigation of diffusion and damping properties of the ICON core • Tests of ICON (physics) at very high resolution and a comparison between ICON at high and at very high resolution (sub-km, CLM interest, Uni Cottbus)

44. WG2 << 20. (cont’d) Thoughts by Andrzej W. (email of 8 September 2018 following discussions at the SMC meeting in St. Petersburg) I would like to add some thoughts on the COSMO activities in the frame of time 2019-2020+ in relation to the recent COSMO Science Plan 2015-2020. In the SP we mention on several occasions about the potential capabilities of COSMO model (and implicitly ICON) for a high resolution deterministic weather forecasting going down to 500m spatial resolution, which was then planned for an operational work around 2020. From my knowledge there were so far just several experiments at the resolution below 1000m (even down to ~100m resolution) by using both ICON-LAM and COSMO-EULAG. So I think it is about a time to explore these areas in a more coordinated way. If possible I would like you to discuss possibility for starting such an effort. My idea would be to stir some kind of collaborative project between teams working with ICON-LAM and COSMO-EULAG in the frame of WG2 (maybe also in collaboration with WG3x, WG5, etc.) . As you have noticed all the development for COSMO-EULAG should be finalized around sept 2019 and then more resources should be available to perform additional/new tasks.

45. (cont’d) Thoughts by Andrzej W.  My idea is to prepare at least one year PT project which would explore capabilities of our models (ICON-LAM/COSMO-EULAG) to perform NWP at the spatial resolution below 1000m (500m or even below). The tests may include real time or more idealistic simulations, including areas with steep topography, coastal or maybe even urban environments. I see the outcome from this project would provide us more information about model/s: - NWP verification scores in the winter and summer time conditions - changes in computational performance for the different resolutions - stability limits for different environmental conditions (e.g. max time step, parameters controlling the physical parametrizations, etc.) - ability to represent severe phenomena related to deep moist convection/wind - identification of deficiencies in physics and numerics which would constitute a base for further actions to be planned in the science plan 2020-2025. Of course above are only my personal thoughts, you can feel free to discuss these issues within SMC/WGs and come up with your own plan.

46. (cont’d) Model documentation • “Supplements” to our “core documentation” are currently placed under “miscellaneous” but a better solution is desired (issue raised by Uli Blahak) • Documentation of Matthias and Guenther’s (Zängl) extensions implemented into turbulence scheme (Matthias) • Single-column model documentation (Matthias) • Model documentation in view of the advent of ICON-LAM. Any comments, thoughts, suggestions?

47. All WGs: gentle reminder • WG leaders should make sure that the publication list is up to date and, importantly, that the entries are formatted in a unified way with no information missing • Participants of various meetings (COSMO GM, ICCARUS) should upload their presentations, WG leaders should make sure that it indeed happens (e.g. one COSMO GM 2018 plenary talk and WG2 and WG3a parallel-session talks are missing) • (address Massimo and Theodore for assistance as to the web)

48. Thanks for your efforts! COSMO SMC Meeting, 12-13 September 2019, Rome, Italy

49. …