MEPS dataset
About MEPS
MEPS file overview
MEPS parameters
Archive strategy
Differences between MEPS and AROME-Arctic
Domains
MEPS Changelog
References
The MetCoOp Ensemble Prediction System (MEPS) covers Scandinavia and Finland (the green bounding box in the map below). MEPS provides 30 ensemble members every six hours with leadtimes out to 61 hours and a spatial grid spacing of 2.5 km. The control member is output with forecast leadtime of 66 hours. We store a wealth og information from the model, including ground field, surface fields, and fields throughout the atmosphere.
Files are structured in a MEPS latest catalogue with most recent files, and the MEPS archive.
File name convention:
<model>_<form>_<resolution>_<DTG of forecast>.<type>
<form> is "det" for Control member output, or "lagged_Xh_subset/latest" for X hour lagged ensemble output.
Unless otherwise specified, the file is only available as ncml.
- xxxx_latest.nc
- Link to the latest available output, where applicable. Not compatible with ncml.
- meps_det_2_5km_YYYYMMDDThhZ.ncml
- Output from Control member. Contains most forecast parameters on surface, model level and pressure levels
- meps_lagged_6_h_subset_2_5km_YYYYMMDDThhZ.ncml
- Output from ensemble. Contains several surface parameters, and some pressure level parameters. Updated every 3 hours.
- meps_lagged_6_h_latest_2_5km_YYYYMMDDThhZ.nc
- Output from ensemble. Contains a bare minimum of surface parameters. Updated every hour.
- Available as ncml or nc.
- meps_det_vc_2_5km_YYYYMMDDThhZ.nc
- Output from Control member. Contains vertical profiles and cross sections. NOT output on structured grid.
- Available as ncml or nc.
- meps_lagged_6_h_vc_2_5km_YYYYMMDDThhZ.ncml
- Output from ensemble. Contains vertical profiles and cross sections.
- archive
- Link to the latest files added to MEPS archive, including static url for the latest files.
There is a wide range of parameters available in the output files. Below is an explanation for some of the parameters. For the others we refer to the metadata in the files themselves.
is an index integer value from 1 to 8:
- Drizzle
- Rain
- Sleet
- Snow
- Freezing drizzle
- Freezing rain
- Graupel
- Hail
The forecast reference time in NWP is the "data time", the time of the analysis from which the forecast was made. It is not the time for which the forecast is valid; the standard name of time should be used for that time.
Sometimes also referred to as the model run time, the analysis time or generating time. A valid forecast for a given time can originate from different model runs the reference time together with the forecast time can uniquely identify a forecast step.
air_temperature_2m: temperature 2 meters above ground
air_temperature_pl: pressure levels.
The levels are defined in the variable pressure
MEPS output are produced in the MEPS latest catalogue. In parallell, files are written to MEPS archive. Files are kept under MEPS latest for a few days.
For description of the file structure in the MEPS archive, please consult the MEPS archive strategy which also details the cleaning strategy of old data in the archive.
In general, AA is following MEPS updates after some delay
The following list is explicit changes in AROME-Arctic which are NOT in MEPS:
- (Domain and EPS settings)
- Radar assimilation set to NO
- STATNW = no (statistical cloud scheme) We have seen in 500m experiments that it provides a higher negative bias in low cloud cover in Svalbard.
- SICE scheme
- xrimax = 0 in AA set to 0.4 in MEPS, Maximum allowed Richardson number in the surface layer
- MODESEHS_OBS=0 in AA set to 1 in MEPS, Mode-S EHS
- No MWHS2 in AA, FengYun-3C
Here are shown the domains for the different model outputs:
Framed by the Blue contour is Arome Arctic, Green is MEPS, and Red is the YR-product domain (Accessible as Postproduced).
Introducing cleaning of MEPS archive. More details here
The MEPS Archive regime has changed. No longer archiving meps_det_.nc and meps_subset_.nc NetCDF-files. Rather we produce one file per member, level type and time step, and ncml-files pointing to these. More details here
MEPS will upgrade to cy43 on March 23rd 2021. This will have some minor impacts on forecast quality. More details on the upgrade and meteorological impacts are described here. In case of extreme weather, the upgrade will be postponed.
Fra 4. februar 2020 endret MEPS kjøremønster. Dette medførte at filer ble endret, og det ble endring i hva som arkiveres og hvordan.
Endringer i grove trekk
- Domenet blir 20% større
- Det kommer flere ensemble-medlemmer
- Extracted, full og sfx filer erstattes med:
- En deterministisk fil (det) med alle parametre som oppdateres hver 3. time (arkiveres)
- En “latest” fil med et lite utplukk parametre fra hele ensemblet som oppdateres hver eneste time (arkiveres ikke)
- En “subset” fil med en god del parametre fra hele ensemblet, oppdateres hver 3. time (arkiveres hver 6. time)
- Filer i latest-katalogen vil i stor grad endres til “ncml” filer. Disse kan ikke lenger lastes ned direkte, men må aksesseres via OPeNDAP. Ellers skal de fungere på samme måte som netcdf-filer, men er en overbygning som peker til flere nc-filer i bakgrunnen
- pp-filer utgår. For post prosesserte produkter henviser vi til yr-produksjonen under MET post-processed products Latest/
- Vi arkiverer ikke lenger all informasjon fra alle ensemblemedlemmer. Vi gjør et utplukk i en "subset" fil som arkiveres hver 6. time. Merk spesielt at informasjon på modellflater kun arkiveres for kontrollmedlemmet, ikke for hele ensemblet
- Parametere fra surfex vil få “SFX_” i starten av variabelnavnet
Hvorfor gjøres endringen?
MEPS har frem til nå kjørt med 10 medlemmer med lange terminer fire ganger i døgnet (00,06,12,18). Dette har medført en ujevn fordeling av regneressurser på tungregnemaskinen gjennom døgnet. Derfor har man bestemt å endre til en mer kontinuerlig kjørerutine der det kjøres fem medlemmer hver time. Alle medlemmer kjøres med 66 timers prognoselengde, slik at ensemble-filene vil inneholde 30 medlemmer, der noen er lagget opptil 6 timer. Kontrollmedlemmet vil ikke være lagget. Med det økte antall medlemmer og domene er det ikke lenger mulig å arkivere like mye som vi gjør idag, derfor gjennomfører vi denne reorganiseringen.
Frogner, I‐L, Singleton, AT, Køltzow, MØ, Andrae, U. Convection‐permitting ensembles: Challenges related to their design and use. Q J R Meteorol Soc. 2019; 1– 17. https://doi.org/10.1002/qj.3525 AROME MetCoOp reference:
Müller, M., M. Homleid, K. Ivarsson, M.A. Køltzow, M. Lindskog, K.H. Midtbø, U. Andrae, T. Aspelien, L. Berggren, D. Bjørge, P. Dahlgren, J. Kristiansen, R. Randriamampianina, M. Ridal, and O. Vignes, 2017: AROME-MetCoOp: A Nordic Convective-Scale Operational Weather Prediction Model. Wea. Forecasting, 32, 609–627, https://doi.org/10.1175/WAF-D-16-0099.1