21th North America/Europe Data Exchange Meeting Asheville September 17-19, 2008

Status report

Hervé Roquet (CMS/RetD)

Bruno Lacroix (DPrévi/COMPAS)

With contributions from CNRM/GMAP

Outlines•Operational suite(s)

–current configurations–last changes

•Use of data•Issues under development

–AROME•Future Plans

Computing platform

Configuration NEC phase 1

• NEC SX8 (Q4 2006)

• 32 nodes of 8 processors

• 2 machines:

• Operations 16 nodes since 9th May 2007

• Research 16 nodes

• 9.1 Tflops max

• 4 Terabytes memory

• 36 Terabytes shared disk RAID5 (GFS)

• Next step phase 2 Q4 2008

Global model up to 102H at 00UTC (cut off 2H20), 72H at 06 (3H), 84H at 12UTC(1H50), 60H at 18UTC (3H)

ARPEGE global spectral model TL538 C2.4 L60

60 levels, from 17m to 5Pa, horizontal resolution from 15km to 87km

Linear grid with T360 C2.4 orography (1080x540 pts)

8 processors for ARPEGE forecast (10’ for 24H forecast)

4DVAR assimilation :

2 loops of minimization T107 C1 L46 (25 it.), T224 (30 it.)

18 processors for assimilation (40’ between cut-off and P0)

data used:

SYNOP, SHIP, BUOY, AIREP, AMDAR, ACARS, TEMP, PILOT

CMW winds GOES 11, 12 + Meteosat 7, 9 + MTSAT-1R, Modis

SEVIRI radiances (Meteosat 9)

AMI (ERS2), Seawinds (Quikscat) and ASCAT (Metop) winds

HIRS, AMSU-A, AMSU-B/MHS NOAA15, 16, 17, 18, Metop & AQUA

SSM/I (DMSP F13 and F14), AIRS AQUA, GPS ZTD, GPS RO, IASI (Metop)

SST 0.5 degree from NCEP/NESDIS + SSM/I sea ice mask

Models configuration

Horizontal resolution

Vertical levels

Models configuration (follow-up)

Regional model up to D2 06UTC (at 00, 06, 12 and 12UTC)

ALADIN spectral limited area model 9.5 km resolution on 2740kmx2740km

domain, 60 levels (289x289 pts) 3DVAR data assimilation: same data as

ARPEGE plus SEVIRI HR radiances Many coupling files

Tropical model 72H range at 00 and 12 UTC ARPEGE uniform model (TL539 C1 L60) ~37km No own data assimilation (interpolation of

stretched model analysis) Short Range Ensemble Prediction System 102H

range 11 runs ARPEGE TL358 C2.4 L55 (23 to 133km) Based on singular vector perturbation

Short range ensemble Forecast (ARPEGE)

56 singular vectors computed on the globe: 16 TL95 + 40 TL44

Optimization time window : 0-12h

Evolving perturbations on 24H window (breeding)

Error statistics of the day

18 181260 24

VS VSjour jour

P24

12

i

10…

…

P24

Mean and spread for a 0H forecast – Z500 (dam)

PEARPOPER

PEARP1.5

08/01/2007

Assimilation/Forecast Suites

Operational suites Atmospheric models:– Global stretched, very short cutoff (1H05) with 3DVAR FGAT (00 UTC only)

– Limited-Area ALADIN• La Réunion 3100x4600km with 3DVAR assimilation, • several research, commercial and transportable dynamical adaptation versions

Chemical Transport Model MOCAGE, Forecasts of air quality up to 96H– 3 domains:Global/Europe/France, horizontal resolutions: 4°, 0.5°, 0.1°

– Observations currently only used for validation

Ocean Wave Models, Forecasts up to 102H– Global (2), Europe, France, horizontal resolutions: 1°, 1°, 0.25°, 0.1°

– assimilate Jason-1 and Envisat altimeter wave height data

Changes in ARPEGE and ALADIN-FranceChanges in ARPEGE and ALADIN-France

•09-2007 : data from Metop-A (AMSU-A et MHS), GPS Radio-occultation (COSMIC-1-6, CHAMP, GRACE A/B), AMI (ERS-2) winds

•02-2008 : enhanced horizontal and vertical resolution T538-L60 , variational bias correction + ASCAT (Metop-A) winds

•07-2008 : IASI, SSM/I F14, statistics from ensemble assimilation cycle (6 members 3D-VAR with T359C1L60 forecast)

SIGMAB’s

« CLIMATOLOGY »

SIGMAB’s

« OF THE DAY »

8 dec 2006 r0

Evolution RMSE Z500 Europe

Steady improvement over 20 years

Telecom and data received (files)

links to Toulouse relevant to US/Europe data exchange:

Daily volume of satellite data files received from Exeter :HIRS NOAA16, 17AMSU-A NOAA15, 16, 18, AQUAAMSU-B/MHS from NOAA15, 16, 18: 500 MbytesSSM/I and IS from DMSP F13, F14 and F16: 280 MbytesSeawind from Quikscat 240 MbytesAIRS from Aqua 500 Mbytes

total 1.5 Go from Exeter

RMDCN = ECMWF + GTS

3 Mb/s

Lannion 2 x 4 Mb/s soon 2 x 8 Mb/s

Recent advances in the use of observations in the French NWP

models

L. Auger, N. Fourrié, É. Gérard, V. Guidard,

F. Karbou, P. Moll, T. Montmerle, C. Payan, P. Poli, F. Rabier

Météo-France/CNRS, Toulouse, France

New operational developmentsSeptember 2006:

20 stratospheric AIRS channels

SSM/I F13 and F15

Ground-based GPS data over Europe

September 2007:

GPS radio-occultation

ATOVS on MetOp (AMSU-A, MHS)

ERS scatterometer

10m wind over land in Aladin

February 2008:

Variational Bias Correction for radiances

ASCAT assimilation

June 2008:

HIRS on Metop, SSM/I F14

Emissivity parametrisation over land for micro-wave

CSR Meteosat

IASI

04/20/23 bouhhh

(1) Averaged emissivities (atlas): (1) Averaged emissivities (atlas): using 2 weeks prior to the assimilation using 2 weeks prior to the assimilation

period; period; Ts is taken from the model’ FG.Ts is taken from the model’ FG.averaged emissivities for ch3 and ch16 are given to averaged emissivities for ch3 and ch16 are given to

Temperature and humidity channelsTemperature and humidity channels(2) Dynamically estimated emissivities (instantaneous):(2) Dynamically estimated emissivities (instantaneous):

derived at each pixel using only one channel (or two) of derived at each pixel using only one channel (or two) of each instrument;each instrument;

Ts is taken from the model FG.Ts is taken from the model FG.(3) Averaged emissivities + dynamically estimated skin temperature:(3) Averaged emissivities + dynamically estimated skin temperature:

Ts at each pixel using one (or two) channel of each Ts at each pixel using one (or two) channel of each instrument instrument

All surface parameterizations are handled by the RTTOV model (Eyre 1991; All surface parameterizations are handled by the RTTOV model (Eyre 1991; Saunders et al. 1999; Matricardi et al. 2004)Saunders et al. 1999; Matricardi et al. 2004)

Three land surface parameterizations with increasing complexity (Karbou et al. Three land surface parameterizations with increasing complexity (Karbou et al. 2006)2006)

AMSU frequency (GHz)

Humidity (x3)

Temperature (x11)Surface (x5)

Ch1 Ch15

&16

Ch17Ch2 Ch3

Land schemes for data assimilationLand schemes for data assimilation

Observation operator simulations

50.3 GHzCh 3 AMSU-A

Control ATLAS EMIS-DYN ATLAS+SKIN

89 GHzCh15 AMSU-A

150 GHzCh 2 AMSU-B

Emissivity at 23 GHz is used, bias could be corrected with an emissivity frequency interpolation

First guess departures (obs-guess) global histograms, 15-31 August 2005

RMS error reduction (EXP_DYN): AMSUA ch2 (-54%), ch3 (-29%), ch4 (-12%), ch5 (-2.5%)

Impact studies during summer 2006Impact studies during summer 2006Overview:Overview:- Period: Period: from 15 July to 15 September 2006from 15 July to 15 September 2006- Many Experiments: Many Experiments:

- Different schemes to estimate the land emissivity and/or the skin temperatureDifferent schemes to estimate the land emissivity and/or the skin temperature- AMSU-A, AMSU-B, SSM/IAMSU-A, AMSU-B, SSM/I- Progressive assimilation of surface sensitive channelsProgressive assimilation of surface sensitive channels

- Preliminary Results:Preliminary Results: CTL/ EXP1 CTL/ EXP1 ((dynamic estimation of dynamic estimation of for AMSU, assimilation of same for AMSU, assimilation of same channels as CTL)channels as CTL)

EXP1 EXP1 / CTL, 18 days statistics AMSU-A, N-Hemis, NOAA16/ CTL, 18 days statistics AMSU-A, N-Hemis, NOAA16Similar results for AMSU-BSimilar results for AMSU-BNo degradation of the fit to the other observationsNo degradation of the fit to the other observations

+35%+32 %+22%

04/20/23 bouhhh

19

Number of assimilated data, AMSU-A channel 7, August Number of assimilated data, AMSU-A channel 7, August 20062006

ReferenceReference Emissivity dynamically estimated from Ch3Emissivity dynamically estimated from Ch3

Assimilation of MSG SEVIRI Clear Sky Radiances

10

.8

m c

han

nel

Associated percentage of cloud

free

CSR product from Meteosat-8/-9 (MSG/MSG-2)

Hourly product Assimilation of

– 2 WV channels in 4D-Var

250 km thinning

CSR

Assimilation of MSG SEVIRI Clear Sky Radiances

96h forecast scores (forecast – obs) wrt synop data (precipitation)with SEVIRI CSR

without SEVIRI CSR

rmse

bias

précip [mm] no of obs

2007

large reduction of rms error

04/20/23 bouhhh

IASI assimilation: general features

Level 1C radiances are received via EumetCast in Toulouse(whole BUFR including 8461 channels)

A subset of 314 channels is retained in the Operational Observational Data Base (commonly chosen with other NWP centres)

Radiances are bias corrected using VarBC

Data selection:– Cloud detection is based on a channel ranking method from ECMWF

McNally & Watts (2003)

– First-guess check

– Geographical thinning: average distance between 2 obs. is 250 km

04/20/23 bouhhh

Current IASI operational use (1) The whole subset of 314 channels is monitored in ARPEGE

Assimilation operational since July 2008

Monitoring for 22 October 2007analysis time of 18 UTC

04/20/23 bouhhh

Current IASI operational use (2)

50 channels are actively assimilated, only over sea

(peaking between 100 hPa and 620 hPa)

13 august 2007analysis time of 00Z

obs. minus analysisfor channel #219(699,50 cm-1)peaking at ~200 hPa

brightness temp. (K)

Information content: DFS10

00->

300

3

00-

>9

0

90->

0.1

hP

a

Radios. & ProfAircraftSurfaceGPSRO

AIRS, IASIBright. Temp

SATWIND

Other developments

Cloudy AIRS/IASI– Cloud detection (Polar areas, A. Bouchard)– Assimilation of cloudy radiances (T. Pangaud)

SEVIRI used as images in the assimilation (Y. Michel)

Cloudy/rainy SSM/I data for tropical cyclones (R. Montroty)

Observation operator adapted for fine-scale analysis (F. Duffourg)

Data monitoring

•http://www.meteo.fr/special/minisites/monitoring/menu.html

User/password available upon request herve.benichou@meteo.fr

Increase of observation number

H. Bénichou

Since July 2008, more than 2 million data per day

Evolution in managed/used data ratio

75% satellite data / 25% in situ data used

AMDAR data

109 Air France planes

Wind profilers network

La Ferté Vidame available on GTS

Nice, Marignane, Clermont-Ferrand and Lannemezan available on bilateral agreement basis

GPS surface network

IGN : RGP 162 stations in September 2008

Observations Usage Summary

Datatype Contact Operations

ATOVS E. Gerard N15,16,17, 18, Aqua, Metop-A

Geostat. winds C. Payan GOES, MeteoSat, MTSAT

MODIS winds P. Moll Aqua,Terra

scatterometer winds C. Payan Quikscat, Metop-A, ERS-2

Conv. (RS, AIREP, TEMP) P. Moll

GPS P. Poli ZTD, RO

AIRS F. Rabier Aqua

IASI V. Guidard Metop-A

New NWP models configuration at Météo-France

3 models : ARPEGE (resolution 15km, 60 levels) ALADIN (resolution 10km, 60 levels) AROME (resolution 2.5km, 41 levels)

Application de la Recherche à l’Opérationnel à MésoEchelle:

AROME plans

since 2007: test of the real-size model with data assimilation; evaluation by forecasters

October 2008: start of operational production

in 2009: new computer & continue model upgrades

Temperatures and winds ALADIN Temperatures and winds AROME

Lyon Lyon

Genève Genève

illustration: forecast on orography

3DVAR cycle

AROME: - analysis of U, V, T, q et PS variables - cycling of TKE, NH et microphysical variables

ARPEGE Observations

Hu2m, T2m

V10m

SEVIRI HR

ALADIN (+ SEVIRI HR, Hu2m,T2m,V10m) AROME (+ radar)

GPSRO

GPS sol

IASI

SEVIRI CSR

Data assimilated into NWP models

Doppler radar network

24 radars: 16 in C band (yellow circles) + 8 in S band (green circles). Volumes reflectivities (from 2 to 13 elevations).

19 Doppler radars (red circles), 5 planned (dashed red circles)

3 polarimetric radars (black circles)

2 planned (dashed black circles)

Future plans

End 2008: new data processing environnement SOPRANO

October 2008: AROME France operational

Spring 2009: Nec phase 2

ARPEGE 10km, AROME L60

PEARP with 30-40 members

New data (SSM/IS F16 and F17, NOAA-N’)

Enhanced use of data: high density radiances, more IASI and AIRS channels, extended conditions of use