bullet  EUMETSAT   bullet  SAFs   bullet   NWP SAF   bullet  Monitoring reports    bullet   Scatterometer Monitoring Report

NWP SAF Scatterometer Monitoring

Introduction   NWP   NRT Monitoring Monthly Monitoring Delay Statistics News & Links

Use in Numerical Weather Prediction (NWP)

Scatterometer use in NWP at
the Met Office

Last updated Sep 2015

Back to use in NWP table

Physical characteristics
The atmospheric models used at the Met Office for operational NWP and climate modelling are all derived from the Unified Model (UM) System (Davies et al. 2005).
Global Model
  • Grid-point model
  • Horizontal resolution: N768 (17 km in mid-latitudes)
  • Vertical resolution: 70 vertical levels, hybrid-eta configuration, lid ~80 km
UKV Model
  • Grid-point model (744 E-W x 928 N-S)
  • Horizontal resolution: variable.
    • The inner part of the domain (622 x 810 grid points) has a resolution of 0.0135 degrees (~1.5 km)
    • The outer parts of the domain vary up to 0.036 degrees (~4 km)
  • Vertical resolution: 70 vertical levels, lid ~40 km (different levels from global)
  • Forecasts to T+36
Data assimilation method
Global Model
  • 4D-Var, 6 hr time window
  • Analysis times (T): 00, 06, 12, 18 Z
  • Time window: T ± 3 hr
  • Time constraints (model runtime):
    • Main forecast run: 15 min before time window ends
    • Update run: 3 hr 15 min after time window ends
UKV Model
  • 3D-Var, 3 hr time window
  • Analysis times (T): 00, 03, 06, 09, 12, 15, 18, 21 Z
  • Time window: T ± 1.5 hr
  • Time constraints (model runtime):
    • All cycles: T+75 mins
Scatterometers assimilated
WindSat (if assimilated) should also be included here
Scat name Product Models assimilated
ASCAT-A

OSI-SAF Level 2 BUFR 25-km equivalent-neutral wind product produced by KNMI.

Global

OSI-SAF Level 2 BUFR Coastal 12.5-km equivalent-neutral wind product produced by KNMI.

UKV

OSI-SAF Level 2 BUFR EARS 12.5-km equivalent-neutral wind product produced by KNMI.

UKV
ASCAT-B

OSI-SAF Level 2 BUFR 25-km equivalent-neutral wind product produced by KNMI.

Global

OSI-SAF Level 2 BUFR Coastal 12.5-km equivalent-neutral wind product produced by KNMI.

UKV

OSI-SAF Level 2 BUFR EARS 12.5-km equivalent-neutral wind product produced by KNMI.

UKV
WindSat

Environmental Data Record (EDR) files received from NRL and processed in-house to produce a level 2 BUFR product. Version 2.4.5 low resolution EDRs at approx 50 km resolution.

 Global
RapidScat

OSI-SAF Level 2 BUFR 25-km equivalent-neutral wind product produced by KNMI.

Global
 
Monitoring
External monitoring web pages
  • This site: no password required
Generic Quality Control
Blacklisting
  • All wind speeds outside range 2-25 m/s
  • All observations made over ice according to the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA)
  • All winds in the Southern Ocean below 72S
  • All observations with SST less than 273.15 K according to OSTIA
Ambiguity removal
  • No prior ambiguity removal
  • Scatterometer cost function contains a term for each ambiguous wind so no dealiasing is required prior to assimilation. Essentially the ambiguity removal is performed within the analysis and so uses other independent observations and the latest model state to choose the correct wind (Candy, 2001).
Bias correction
Speed dependent polynomial correction (Cotton, 2009)
  • Correction is of the form: Y = C0 + C1*X + C2*X2 + C3*X3
    • X is the observed wind speed
    • Y is the bias corrected wind speed
    • Ci are the bias coefficients
Thinning
In the global model observations are thinned separately for each instrument
  • Thinning distance 80-km in global model
  • Chosen wind is that closest to centre of grid box
  • Max 1 wind per instrument in each thinning box
In the UKV ASCAT observations are thinned together
  • Thinning distance 46-km in UKV model
  • Chosen wind is that closest to centre of grid box
  • Thinning priority set so EARS data are used to fill in the gaps where no global data receieved in time.
Background check
  • Background check: none
  • Variational quality control applied to scatterometer cost function during assimilation. Initially the a-priori probability of gross error is set to zero then, after about half the iterations in the analysis, this is set non-zero which activates the quality control.
Specific Quality Control
Global model
Scat name ASCAT-A/B OSCAT ERS-2 WindSat RapidScat
Operational since Nov 2007 (Metop-A) and April 2013 (Metop-B) Jan 2013 - Feb 2014 1998-1999 and Mar 2006*- Jul 2011 Nov 2008 Sep 2015
Observation error U/V 2.0 m/s 2.0 m/s 2.0 m/s 2.5 m/s 2.25 m/s
Wind speed range 2-25 m/s 2-25 m/s 2-25 m/s 6-20 m/s 2-25 m/s
Bias corrected? Yes. 0.2 m/s is subtracted from retrieved wind speeds to account for average difference between equivalent-neutral and real wind speeds. Yes. 0.2 m/s is subtracted from retrieved wind speeds to account for average difference between equivalent-neutral and real wind speeds. Speed increased by 5% in retrieval Yes Yes. 0.2 m/s is subtracted from retrieved wind speeds to account for average difference between equivalent-neutral and real wind speeds.
Crosstrack cells used All Nodes 1-21 and 22-42 Nodes 5-14 and 23-32 All Nodes 1-19 n/a Nodes 3-17 and 23-39
QC thresholds
  • Check supplied wind vector QC flag
  • MLE-based QC method. Rank 1 MLE values are normalised by speed and observations are rejected if MLEnorm > 7 or MLEnorm < -20
  • Check supplied wind vector QC flag
  • Check modified PCD product confidence flag during retrieval
  • Rank 1 normalised distance to cone < 2.5 (expect unity)
  • Check supplied EDR QC flag 1
  • Rank 1 chi-squared probability > 0.3
  • Check rain in field of view via retrieved rain rate
  • Retrieved cloud liquid water < 0.2 mm
  • Check supplied wind vector QC flag
*see Keogh and Offiler (2006)
 
Observation Operator
Lowest model level
The UM uses a terrain following height coordinate with the wind (and density) levels vertically staggered from the potential temperature (theta) levels. Whilst the Global and UKV models both have 70 levels, the UKV uses a different set designed to give most benefit to short-range forecasts, i.e. a lower model lid allowing greater resolution in the boundary layer and troposphere. The heights of the lowest horizontal wind levels over sea (zero orography) are
  • Global: 10 m
  • UKV: lowest wind level is 2.5m, second lowest is 13.33 m
Interpolation
The UM surface exchange scheme assumes that Monin-Obukhov similarity theory is valid for the surface layer, with surface variables interpolated to standard observation heights.
  • 10m real wind components are calculated using the stability functions of Beljaars and Holtslag (1991) in stable conditions and Dyers and Hicks (Dyer, 1974) in unstable conditions.
  • Model wind fields are horizontally interpolated to the observation location using standard linear interpolation of the surrounding 4 grid points. In the global model, if any of the 4 grid points are coast then the model value at the nearest grid point is used instead.
History of Changes
The list includes the main scatterometer or model changes implemented operationally in the Met Office models.

15/09/15 Assimilate RapidScat in global model.
17/02/15 EARS 12.5-km data from Lannion station for Metop-A and Metop-B assimilated in UKV model
03/02/15 PS35: migration from 12.5-km hamming window to 12.5 -km coastal winds in UKV model. Now also assimilating Metop-B in UKV and small change to switch on VarQC earlier.
15/07/14 PS34: a) major model upgrade including increased resolution to 17-km, ENDGame dynamics and physics changes. b) VAR grid resolution increased from 60km to 40km. c) Scatwind thinning changed so all data thinned seperately for each instrument in global model
21/02/14 End of Oceansat-2 mission.
30/04/13 Assimilate ASCAT on Metop-B in Global (PS32).
16/01/13 Assimilate OSCAT in Global (PS31).
28/03/12 Update QC for PS29. ASCAT: new MLE-based QC, wind speed bias correction and remove blacklist on cells 21 and 22. WindSat CLW threshold relaxed to 0.2.
20/07/11 Assimilate 12.5 km ASCAT (global and EARS) in UK4/V. Reduce thinning distance to 80 km in global.
05/07/11 End of ERS-2 mission.
02/11/10 WindSat assimilated in NAE model
07/07/10 WindSat EDR version 2.2.1 (no numerical changes)
15/06/10 WindSat EDR version 2.1.4 and new bias correction
09/03/10 Global horizontal resolution upgrade from 40km to 25km. NAE vertical upgrade from 38 levels to 70 levels
 
References

Beljaars, A. C. M. and Holtslag, A. A. M., 1991: Flux parameterisation over land surfaces for atmospheric models. J. Appl. Meteor., 30, 327-341.
Candy, B., 2001: The Assimilation of Ambiguous Scatterometer Winds Using a Variational Technique: Method and Forecast Impact. Met Office Forecasting Research Technical Report 349.
Cotton, J., 2009: A comparison of QuikSCAT with buoy, ship and radar altimeter wind speeds and evaluating the need for a new bias correction. Met Office Forecasting Research Technical Report 538.
Davies, T., M.J.P. Cullen, A.J. Malcolm, M.H. Mawson, A. Staniforth, A.A. White and N. Wood, 2005: A new dynamical core for the Met Office's global and regional modelling of the atmosphere Q. J. R. Met. Soc., 131, Pages 1759-1782.
Dyer, A. J., 1974: A review of flux-profile relationships. Bound. Layer Meteor., 7, 363-372.
Keogh, S.J. and D. Offiler, 2006: ERS-2 scatterometer - reintroduction into the Met Office global model. Met Office Forecasting Research Technical Report 473.