1 / 19

Implementation of KENDA at ARPA-SIMC and first experiences with an OSSE

Implementation of KENDA at ARPA-SIMC and first experiences with an OSSE. Chiara Marsigli Tiziana Paccagnella ARPA Emilia-Romagna, SIMC. Outline. Implementation of KENDA at ECMWF (sms system) Set-up of the OSSE suite First results Discussion items. KENDA suite at ECMWF.

amethyst-randall
Télécharger la présentation

Implementation of KENDA at ARPA-SIMC and first experiences with an OSSE

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Implementation of KENDA at ARPA-SIMC and first experiences with an OSSE Chiara Marsigli Tiziana Paccagnella ARPA Emilia-Romagna, SIMC

  2. Outline • Implementationof KENDA at ECMWF (sms system) • Set-up of the OSSE suite • First results • Discussionitems

  3. KENDA suite at ECMWF creationof the directories and set-up of the runs MARS retrievalofconventionalobservation (ARPA) bufr2netcdf getBoundaryConditionsfrom ECMWF ENS (and Ics for the cold start) currently 20 members int2lm getanalysisfromprevious KENDA cycle 3-hourly cycles COSMO run KENDA analysisstep processing of output output archiving

  4. OSSE set-up 15 18 21 24 00 03 06 09 12 NATURE RUN optimalconfiguration, generates the pseudo-observations(foffiles) FOF FOF FOF FOF FOF FOF FOF 10 IC originalqv KENDA CYCLE 10 IC halvedqv CONTROL RUN “dried” run, qvreducedbyhalf in the InitialCondition

  5. KENDA OSSE suite at ECMWF getfoffrom nature run (psuedo-observations) modificationof the IC (like in the controlrun) onlyformembers 1-10 • Executionof: • Nature run • Controlrun • KENDA OSSE • isdrivenby the specification in the .def file (filefor the definitionof the sms suite)

  6. Set-up ofthe experiments • KENDA cycle: • 3-hourly cycles, 24 hours • 20 members • BCsfrom ECMWF ENS (alsoICsforcold start) • no modelperturbations • observations: TEMP SYNOP AIREP

  7. qvlev 40-41 +3h NATURE RUN CONTROL RUN

  8. qvlev 40-41 +6h NATURE RUN CONTROL RUN

  9. qvlev 40-41 +12h NATURE RUN CONTROL RUN

  10. KENDA cycle-qvlev 40-41 +3h BACKGROUND (lff) ANALYSIS (laf) MEAN SPREAD

  11. KENDA cycle - qvlev 40-41 +6h BACKGROUND (lff) ANALYSIS (laf) MEAN SPREAD

  12. NATURE RUN KENDA ANALYSIS qvlev 40-41 +6h CONTROL RUN

  13. KENDA cycle - qvlev 40-41 +9h BACKGROUND (lff) ANALYSIS (laf) MEAN SPREAD

  14. KENDA cycle - qvlev 40-41 +12h BACKGROUND (lff) ANALYSIS (laf) MEAN SPREAD

  15. NATURE RUN KENDA ANALYSIS qvlev 40-41 +12h CONTROL RUN

  16. Issues under investigation • “sufficient” spread isneeded in the OSSE ensemble • after the KENDA step: • the ensemble meaniscorrectlydriventowards the pseudo-observations • the ensemble spread isgreatlyreduced • pseudo-observationrejectedif the spread istoosmall • techincalissuestobesolvedrelatedtoqualitycontrolchecks(pseudo-observationsrejectedifthey are toodifferentfromthe background, whichmaybeneededforan OSSE) • furthertesting: localisation, multi-stepanalysis • suggestionsforthisOSSE are welcome!

  17. Thankyou!

  18. Heghtof the levels

  19. Test cases (Hymex SOP)

More Related