STC-Report September, 10 th, 2009, Offenbach
STC-Report n Priority Projects/Tasks p Conservative Dynamical Core” (CDC) approved by STC o Decision tree (Deliverables: Dec. 2009) on “how or if to continue the project” considered to be very important p VERSUS 2 (SDO) approved by STC o SDA has to be adapted with resources and milestones accordingly (Deliverables: Nov. 2009) p Priority Task Postprocessing approved by STC o Will be based on Field. Extra o Still some resources missing (operators to calculate metric terms ) n Termination of Priority Project Sat. Cloud p Tasks: use of the NWC cloud analysis, AMSU-A over land p Very little contributions from participating members p Scientists had too little technical and scientific support at their local centers COSMO General Meeting, 2009, Offenbach
STC-report: Enhancing the internal communication n Some issues with collaboration and communication inside COSMO became apparent during last COSMO year p Lack of communication inside Priority Projects and also inside the STC p Project leaders complained about reporting effort and effort to support team members abroad p FTE concept seemed not to be helpful in some cases n Initial proposal of Chairman and SPM p Try to replace FTEs by work-package contracts p Introduce a financial contribution COSMO General Meeting, 2009, Offenbach
STC-report: Enhancing the internal communication n STC evaluated the situation p It showed to be difficult to amend the agreement o Especially regarding direct monetary contributions to COSMO n Possible ways ahead p Enhancing collaboration needed p Video. Conferencing o DWD has new tool whichwas presented to the STC o Test between DWD and MCH next week o Plan to have monthly PP compulsory (video)conferences o In conjunction with less reporting p Instead use license money, contributions from donating partners to o Support research visits, PP kick-offs/meetings, training, education p The STC promotes that the PL actually requests “Educated FTEs” COSMO General Meeting, 2009, Offenbach
STC-report n Preparation of COSMO for future HPC architectures. n Activities in Switzerland implication for COSMO. n Activities in Italy (Cooperation with CASPUR) n Coordination indispensable n STC proposes to set-up a new Priority task n Roshydromet was accepted as a full member by the STC n Roshydromet invites the COSMO Community to MOSCOW for the next General Meeting COSMO General Meeting, 2009, Offenbach
Meteo. Suisse Some activities at Meteo. Swiss outside of the priority projects n Done p Additional 06 UTC run of COSMO-7 p On demand mode (hourly 6 h runs of COSMO-2) p Operationalization of SSO (COSMO-7 and COSMO-2) p Tuning of turbulence (for convection) and horizontal diffusion p Improvement of Post-processing (Kalman Filter) p Clean-up and documentation of Fieldextra n Planned: p Tuning of SSO parameters (values of GME used now) p Assimilation of precipitation in terra s. a. (MSMA) p Redefinition of vertical layers (thinner in PBL) p Use of COSMO-ART for pollen forecasts COSMO General Meeting, 2009, Offenbach
Strategic issues at Meteo. Swiss n Start of operations for project CN Met (COSMO-2 for NPP) n C 2 SM (competence center for climate modelling) in place p with ETHZ and other research institutes p COSMO, COSMO-CLM, COSMO-ART, (ECHAM) p 1 Director, 3 Ph. D, 3 scientific programmers n Swiss national HPCN strategy approved by government p CSCS to build a 4 PFlops computer until 2012 o Probably massive parallel machine (<1 Mio cores) p HP 2 C initiative to support the code adaptation o Proposal of C 2 SM/CSCS/Meteo. Swiss (+COSMO? ) p Support of the modelling activities of Meteo. Swiss o Toward COSMO-1 in 5 years time COSMO General Meeting, 2009, Offenbach
What’s new in Italy n. New computer installed at CNMCA (HP Linux Cluster – 128 nodes (1024 cores) -12. 8 peak Tera. Flops)- final test for acceptance starts on 15 September. n. Contract with CINECA (Bologna) renewed for 3 more years. n. ARPA Piemonte has opened a 3 years contract with CINECA which will provide the computer power for developing an environmental modeling chain based on COSMO model outputs. n. Wave models are now in operation both at CNMCA and ARPA-SIMC (based on WAM and SWAN) with the Italian Navy and the National Civil Protection as main users. n. In the framework of a cooperation with the EUMETSAT MIST, COSMO -IT output has been provided to serve as proxi data-set to simulate IRS radiance at the top of atmosphere. n. After a formal request by the Israeli PR, COSMO-ME output is daily provided to the Israeli NMS since 31 st August. COSMO General Meeting, 2009, Offenbach
Plans and cooperation for 2010 in Italy n Porting of the 3 d-var - COSMO ME - COSMO IT chain on the new computer at CNMCA. n Developing of a RUC cycle for COSMO I 2 at ARPA SIMC. n Start of developing of an environmental modeling chain based on COSMO-Model output at ARPA Piemonte. n A meteo-hydrological coupled system for the Po river basin, driven by COSMO LEPS and COSMO I 7 – COSMO I 2 suite, will be implemented at ARPA-SIMC n In cooperation with the Italian Navy, development at CNMCA of a deployable SWAN model to be used for near coast operations. n In the framework of the program My. Ocean, COSMO ME will provide the atmospheric forcing to the oceanographic system MFS for the Mediterranean. Studies aimed at improving interactions between the two system will be pursued. n In order to improve the cooperation between USAM/CNMCA, ARPA SIMC and ARPA Piemonte, discussion is started on the elaboration of a common strategic plan on the development and operational use of NWP models. COSMO General Meeting, 2009, Offenbach
HNMS - Greece n Implementation of 2. 5 Km COSMO-GR and parallel running for Verification purposes, especially for some important cases affected Greek area. n Development of special applications to assist (more) forecaster’s work (TAF etc). n Operational Version of METEO-FR’s model MOTHY (Modèle Océanique de Transport d'HYdrocarbures) for the Eastern Mediterranean, with 10 m winds and MSLP of COSMO-GR. n Operational Use of COSMO-GR for issuing marine Bulletins for Eastern Mediterranean and Black Sea. n Operational Use of COSMSO-GR for running WAM-4 in the Mediterranean and Black Sea (next slide).
HNMS - Greece n What has changed since last year? The forecasters in our operational meteorological center make continuous use of the COSMO-GR (7 Km) products during their shifts. Consequently, COSMO-GR is now our main operational local area model. The fact that two years ago the situation was quite different, shows that COSMO-GR model became more suitable for the people that make operational forecasts, especially by raising the degree of confidence on their decisions. COSMO General Meeting, 2009, Offenbach
INSTITUTE OF METEOROLOGY AND WATER MANAGEMENT Poland n IMGW strategy for COSMO (continuation of the one presented in 2008) p increase the engagement of IMGW in the area of the NWP development p increase and concentrate resources on well defined areas of model development (esp. dynamics, also assimilation) p strengthen the cooperation between IMGW and academia p continue and develop international cooperation with important NWP research centres, especially NCAR p increase computational capabilities based on own machine and cooperation with academic computational centre in Gdansk (TASK) COSMO General Meeting, 2009, Offenbach
INSTITUTE OF METEOROLOGY AND WATER MANAGEMENT Poland n a formal bid for a new computer is open p expected peak performance 3 TB, p expected delivery: autumn 2009 n expected operational implementation of 7 -km COSMO: p end of 2009/beginning 2010 n expected semi-operational implementation of 2. 8 -km COSMO: p beginning of 2010 n Work on data assimilation started (reconfiguration of initialization system, first successful test trials of analysis cycle) p expected implementation of assimilation data base in 2010 p expected semi-operational implementation of analysis cycle in 2010 n Successful start of the new group working on dynamical problems for COSMO CDC PP COSMO General Meeting, 2009, Offenbach
INM - Romania Operational runs Hardware system: IBM cluster: 14 Server Blades – each blade with 2 processors Intel Xeon Quad Core 3. 00 GHz, 8 GB RAM 2009: 161 x 145 grid points - 7 km resolution: Ø nudging data assimilation for synop; Ø two runs per day (00 UTC and 12 UTC). 301 x 231 grid points – 2. 8 km resolution: Ø no data assimilation; Ø two runs per day (00 UTC and 12 UTC). 2010: Ø 161 x 145 grid points - 7 km resolution: Ø data assimilation for radar data; Øfour runs per day (00, 06, 12 and 18 UTC). Ø 301 x 231 grid points - 2, 8 km resolution: Ø nudging data assimilation for synop; Ø four runs per day (00, 06, 12 UTC 18 UTC). COSMO General Meeting, 2009, Offenbach Domains:
INM - Romania n Development in the frame of COSMO Consortium Ø Further participation on “VERification System Unified Survey” (VERSUS) Priority Project: 0. 3 FTE Ø Further participation on Priority Tasks for Support Activities: 0. 2 FTE (COSMO Newsletter) + 0. 3 FTE ? ? (Contents Management) Ø Participation on “Towards Unified Turbulence-Shallow Convection Scheme” (UTCS) Priority Project: 0, 5 FTE Ø Contributions to other Priority Projects, if it is required Ø Romania may organize and partially support a training course, using the infrastructure of his National Meteorological School, in Bucharest or Constanta (Black Sea border). n COSMO Romanian Group p Dr. Gheorghe STANCALIE (Dr. I. V. PESCARU – retired) p Mrs. Rodica DUMITRACHE, senior researcher, Ph. D student p Mr. Cosmin BARBU, researcher, Ph. D student p Mr. Bogdan MACO M. Sc student p Mrs. Liliana VELEA, Mrs. Aurelia LUPASCU, Mrs. Iulia IBANESCU (currently working abroad) COSMO General Meeting, 2009, Offenbach
ROSHYDROMET COSMO-RU 07 Initial data for 00 и 12 UTC, 78 h forecasts h = 7 km Grid: 701 * 621 * 40 SGI Altix 4700 (1664 cores) (832 processors Itanium, 64 -bit, 3, 3 Tb memory) Run time 33 min: 512 cores (16 x 32) 19 min: 1024 cores (32 x 32) h = 7 км h = 2. 8 km
ROSHYDROMET’s Strategy n Data assimilation p 3 D-var n Global models: p p Spectral models T 169 L 31, T 339 L 63. SL-AB n Mesomodels: p p COSMO-RU (5000 km x 4300 km, 701 x 621, h=7 km; ARW –nowcasting (research) n Ensembles p Breeding method with T 85 L 31 (reseach T 169 L 31) n Visualization p Isograf , Gr. ADS n Sochy 2014 p COSMO-RU 01 (h = 1 km) 2. 8 km)
DWD – Strategy revisited n The NWP system strategy p Local scale ensemble forecasting system p Global forecasting system that must be expanded into an ensemble prediction system providing the necessary BCs on the regional scale p Ensemble data assimilation systems p Maximum use of remote sensing data n Strengthen cooperation with academia p DWD-research program for 3 years (budget of 500 k€ per year) o DWD defined areas of research o Contracts with research institutes/universities concluded o Topics: physics, ensemble forecasting dynamics p HEr. Z o Topics: dynamics and predictability, data assimilation, model development (physics), climate monitoring and diagnostics, exploiting the potential of weather and climate simulation for the society o Budget draft contains 2 Mio€ per year for the next 4 years o Support of 4 -5 research groups at the universities/research institutes o Groups will be lead co-operatively by one scientist from the institute and one from DWD COSMO General Meeting, 2009, Offenbach
DWD – Future development of NWP-system n New Supercomputer@DWD since late summer of 2008 p p 2 x 3 TFlop front-end Q 1 2009 NEC SX 9: 30*LMRAPS within 1400” in parallel Q 3 2010 NEC SX 9: 90*LMRAPS within 1400” in parallel Still not operational o several issues with SX 9 and database prevented numerical experimentation to work as needed o Hopefully SX 9 will take over the operational suite in October (this year!) n Late 2009: GME 30 km, 60 levels n Hybrid data assimilation scheme for GME and COSMO p 3 D-Var and LETKF p 2011 introduction of LETKF in ICON n COSMO-DE-Ensemble p Same configuration as deterministic model p Pre-operational 2010, 20 members p Operational 2011/2012, 40 members n ICON p Non-hydrostatic, triangular mesh p Operational 2011 p Will replace GME and COSMO-EU COSMO General Meeting, 2009, Offenbach
Скачать презентацию STC-Report September 10 th 2009 Offenbach STC-Report
6ed7748f90eba23b442c7014a760b453.ppt