Use of AMVs in the Meteo France NWP model

Use of AMVs in the Meteo France model

Contents

Physical characteristics

Global Model

  • Streched global model, spectral, semi-implicit semi-Lagrangien temporal scheme
  • Horizontal resolution: T1798C2.2 (between T817 (25 km Gaussian grid) over the South Western Pacific and T3956 over Western Europe (5 km Gaussian grid))
  • Vertical resolution: 105 vertical levels, pressure hybrid coordinate, between 10 meters near the surface and 0.05 hPa (70 km) to the top of the model

Data assimilation method

Global Model

  • Analysis  4D-VAR, 6 hr time window, T224/T499 (40 km Gaussian grid) C1L105, full B-matrix flow dependent
  • Time window: T± 3hr, Analysis times (T):00,06,12,18 Z
  • Time constraints (model runtime):
    • Main forecast run: 50 min before time window ends (TWE) for T=00, at TWE for T=06, 1hr 10 min before TWE for T=12, at TWE for T=18
    • Update run: 5 hr after TWE for T=00, 3 hr 50 min for T=06, 5 hr after TWE for T=12, 4 hr 17 min after TWE for T=18

AMV types assimilated

Global Model

  • Meteosat-9 IR10.8, VIS0.75, VIS0.8, cloudy WV6.2, cloudy WV7.3
  • Meteosat-10 IR10.8, VIS0.75, VIS0.8, cloudy WV6.2, cloudy WV7.3, backup Meteosat-11
  • Himawari-9 IR 10.4, VIS0.6, cloudy WV6.2, 6.9 and 7.3, backup Himawari-8
  • GOES-16 IR11.2, VIS0.65, cloudy WV6.2, backup GOES-17
  • GOES-18 IR11.2, VIS0.65, cloudy WV6.2, backup GOES-17
  • NESDIS and Direct Broadcast MODIS Terra IR10.7, cloudy/clear sky WV6.8
  • NESDIS and Direct Broadcast MODIS Aqua IR10.7, cloudy/clear sky WV6.8
  • NESDIS and Direct Broadcast AVHRR NOAA-15,18,19 IR

Global quality control

Blacklisting in space

  • All geostationary winds beyond 62° satellite zenith angle
  • All geostationary winds over land below 700 hPa
  • All geostationary winds over land north of 30N
  • All geostationary winds between 55W and 100W, north of 30N
  • All geostationary winds between 350 hPa and 800 hPa in the Extra-Tropics (polewards of 20 N/S)
  • All geostationary VIS winds above 700 hPa
  • All geostationary cloudy WV winds below 400 hPa
  • All geostationary clear sky WV winds
  • All MODIS and NOAA winds over land below 700 hPa.
  • All MODIS and NOAA IR winds over sea below 850 hPa.
  • All MODIS cloudy and clear sky WV winds over sea below 700 hPa

QI thresholds

The table below is intended only as a guide to the QI thresholds used. Please see the section on blacklisting in space for specific details on which geographical areas are used.

Note 1: no QI thresholds applied to polar data.

Note 2: +5 added to QI threshold for satellites using QI2 (in 3-10-077 BUFR template, QI1 no longer available): GOES series, GOES-16 onwards

extratropics tropics
HL ML LL HL ML LL
Geo QI1 (+5 if using QI2) IR 85 – 85 85 85 85
VIS – – 85 – – 85
WV 85 – – 90 – –

 

tropics: 20S-20N

extratropics: polewards of 20S/N

hl: 10-400 hPa
ml: 400-700 hPa
ll: 700-1000 hPa

QI1 – EUMETSAT QI with first guess check
QI2 – EUMETSAT QI without first guess check
RFF – Recursive Filter Function (CIMSS/NESDIS)

For information on how the quality indicators are formulated see Holmlund (1998, Weather Forecasting 13 1093-1104) and Hayden and Purser (1995, Journal of Applied Meteorology 34 3-15).

Thinning

  • All winds thinned in 1.8° by 1.8° (on average, 92 % (70 %) of data have a neighbour further away than 0.9° (1.35°)) by 10-175 hPa boxes. Vertical extent varying according to nearest standard pressure level. There are 16 vertical boxes.
  • Wind selected by highest QI.
  • Winds are assimilated in 30 minute time-slots.

Background check

  • Satellite wind observations undergo:
    • Background check: comparison with short-term forecast from previous model run.
    • Asymmetry check: additional penalty applied to winds that under-report wind speed when compared with background field (FG) by more than 4 m/s; the test gets increasingly tougher when the FG speed increases, up to 60 m/s when all winds under-reporting by more than 4 m/s are rejected; the test is however relaxed for low levels (pressure > 700 hPa) and the Tropics.
    • Variational quality control.

Observation errors

Observation errors

Satellite Level (hPa) 1000 925 850 700 500 400 300 250 200 150 100
Meteosat/GOES/NOAA Error(m/s) 2.063 2.063 2.145 2.310 2.888 3.218 3.465 3.630 3.795 3.795 3.795
Himawari Error(m/s) 2.438 2.438 2.535 2.730 3.413 3.803 4.095 4.290 4.485 4.485 4.485
Himawari Error(m/s) 2.813 2.813 2.925 3.150 3.938 4.388 4.725 4.950 5.175 5.175 5.175

 

History of changes

21/03/23

Meteosat-10 replaces Meteosat-11 at Met-0 position (Meteosat-11 as backup).
09/01/23 Start using GOES-18 (GOES-W), GOES-17 stopped on 04/01.
13/12/22 Switch from Himawari-8 to Himawari-9.
29/06/22 New operational NWP configuration with lot of changes regarding the global model (deep convection, air-sea fluxes, solar radiation, 1D sea-ice scheme)
10/06/22 Meteosat-9 activated in replacement of Meteosat-8, Meteosat-8 stopped on 01/07.
16/02/21 Switch to 3-10-077 BUFR template for EUMETSAT product
as it was also informed for NESDIS (on 15/01/20).
05/05/20 Start using GOES-17 (GOES-W position)
02/03/20 GOES-15 (GOES-W position) no longer used (satellite stop)
15/01/20 NOAA-20, GOES-17 winds monitored, switch to 3-10-077 BUFR template for NESDIS product (regarding GOES-16, GOES-17, NPP, NOAA-20)
02/07/19 Increase of model resolution from T1198C2.2L105 to T1798C2.2L105, GOES-16 winds assimilation (GOES-E position)
22/02/18 Start using Meteosat-11 winds in Met-0 position (in replacement of Meteosat-10 winds product stopped on 06/03/18)
08/01/18 GOES-13 stop (GOES-E position)
05/12/17 The geographical mask for geostationary satellites (+/-50deg around the sub-satellite position) is replaced by a satellite zenith angle criteria
10/03/17 Himawari-9 becomes the Himawari-8 backup (in replacement of MTSAT-2)
01/02/17 Start using Meteosat-8 on IODC position (in replacement of Meteosat-8 stopped on 31/03/17)
24/03/16 MTSAT-2 becomes the Himawari-8 backup
08/12/15 Start using Himawari-8
13/04/15 Increase of model resolution from T798C2.4L70 to T1198C2.2L105. Monitoring of EUMETSAT dual-MetOp AMVs (double mode)
02/07/13 Reduction of specified observation error of 17% for « conventional » observations including the AMVs.
24/01/13 Switch from Meteosat-9 to Meteosat-10 winds.
06/12/11 Switch from GOES-11 to GOES-15 winds.
15/12/10 Start using Direct Broadcast MODIS (Sodankylä, Fairbanks) and AVHRR NOAA-16,18,19 winds (Barrow, Rothera) through CIMSS/Eumetcast link.
24/11/10 Start using NESDIS AVHRR NOAA-15,16,18,19 winds.
Monitoring Eumetsat Metop-2 winds
11/08/10 Satellite change  MTSAT-1R to MTSAT-2.
14/04/10 Satellite change GOES-12 to GOES-13 (GOES-E position).
06/04/10 Increase of model resolution from T538C2.4L60 to T798C2.4.
Reduction of specified observation error of 10% for all observations including the AMVs.
Start using Clear Sky Water-Vapor MODIS winds.
04/02/09 Start using Direct Broadcast MODIS winds (Tromso,Mc-Murdo) through CIMSS/Eumetcast link.
06/02/08 Increase of model resolution from T358C2.4L46 to T538C2.4L60.
27/03/07 Satellite change Met-8 to Met-9.
23/01/07 Satellite change Met-5 to Met-7.
22/06/06 Start using Met-8 (after removal of Met-7).
Start using GOES-11 (replaces GOES-10 in GOES-W position).
Start using NESDIS MODIS Terra and Aqua winds.
Switch from GOES-W/E and MTSAT-1R SATOB to BUFR winds.
23/01/06 Start using MTSAT-1R SATOB winds (after removal of GOES-9).
21/05/03 Start using GOES-9 SATOB winds (after removal of GMS-5).
22/04/03 Start using GOES-12 SATOB winds (replaces GOES-8 in GOES-E position).
08/12/03 Switch from Met-5/7 SATOB to BUFR winds using QI1 in the quality control and thinning.