AMV usage in the ECMWF NWP model

Use of AMVs in the ECMWF model

Physical characteristics

Global Model

  • Spectral model
  • Horizontal resolution: TCo1279 (~9 km cubic octahedral
    grid), analysis: TL399 (~50 km Gaussian grid)
  • Vertical resolution: 137 vertical levels
  • Analysis times: 00,12 Z

Data assimilation method

Global Model

  • Since 29th June 2004, ECMWF have been running an early delivery stream (DA) and a delayed cut-off stream (DCDA). The DCDA is the equivelent of the Met Office update run and is used to generate the background for the next cycle. The early delivery stream is the equivelent of the Met Office main forecast run and is used to generate the operational forecasts.
    • DA: 4-D VAR, 6-h time window.
    • DCDA: 4-D VAR, 12-h time window.
  • Time window:
    • DA: T ± 3 hr
    • DCDA: T-3 hr – T+9 hr
  • Time constraints (model runtime):
    • DA: 1 hr after time window ends
    • DCDA: 5 hr after time window ends

    AMV types assimilated

    Global Model

    • Meteosat-8/-11 BUFR IR, VIS0.8, cloudy WV6.2 and cloudy WV7.3
    • GOES-15 BUFR IR, VIS, cloudy WV
    • HIMAWARI-8 BUFR IR, VIS, cloudy WV6.95
    • Metop-A/B BUFR IR (single and dual products)
    • NESDIS MODIS Aqua IR, cloudy WV, clear sky WV
    • CIMSS AVHRR NOAA-15/16/18/19
    • S-NPP VIIRS IR

    Global quality control

    Blacklisting in space

    • All VIS winds at 700 hPa and above
    • All geostationary WV winds below 400 hPa except for Meteosat-11 WV7.3 where all below 600 hPa.
    • All goestationary winds over land below 500 hPa and additionally all winds over land (any height) in the following areas:
      • North of 20 N in any region that is west of 20W or east of 100 E.
      • North of 35 N for any longitude that is between 20W and 30E.
      • Over the Himalayas.
    • All MODIS/AVHRR/VIIRS winds equatorwards of ± 60° latitude
    • All Metop-A/B single product winds equatorwards of ± 60° latitude
    • All Metop-A/B dual product winds equatorwards of ± 40° latitude and polewards of ± 60° latitude
    • All MODIS/AVHRR/VIIRS and Metop winds over land below 400 hPa.
    • All MODIS/AVHRR/VIIRS and Metop IR winds over sea below 700 hPa.
    • All MODIS cloudy and clear sky WV winds over sea below 550 hPa.
    • All Meteosat-8/-11 IR and WV winds below 250 hPa in the tropics (25N-25S)
    • All HIMAWARI-8 IR winds below 300 hPa in the tropics (25N-25S)
    • All HIMAWARI-8 and Meteosat-8 winds above 150 hPa
    • Zenith angle > 64° for all geostationary winds
    • All winds below 1000 hPa or above 100 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.

      extratropics tropics
    HL ML LL HL ML LL
    GOES-15 QI2 IR 50 50 50 50 50 50
    VIS 50 50
    WV 50 50
    Meteosat-8/-11 QI2 IR 85 85 85 85
    VIS0.8 85 85
    WV7.3 85 85 85
    WV6.2 85 85
    Himawari-8 QI2 IR 70 70 70 70 70 70
    VIS 70 70
    WV6.95 70 70
    Metop/S-NPP QI2 IR 60 60 60

    tropics: 20S-20N
    extratropics: polewards of 20S/N

    hl: 100-400 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 200 km by 200 km by 50-175 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 min time-slots.

    Background check

    AMV observations undergo:

    • Background check: comparison with short-term forecast from previous model run
    • Variational quality control

    Observation errors

    Similar to Met Office – observation errors calculated individually for each wind using estimates of the error in vector, error in height and variation in the background wind column.