The first ASCAT Scatterometer Winds

1
The first ASCAT Scatterometer Winds

• Ad Stoffelen, Marcos Portabella, Anton Verhoef,
  Jeroen Verspeek, Jur Vogelzang, Gerrit Burgers

Royal Netherlands Meteorological Institute
(KNMI), De Bilt, The Netherlands
2
ASCAT scatterometer
MetOp-A Launch 19 October 2006
3 antennas aft, fore and mid
3
Continuity of scatterometers
4
First Winds on 27 October 2006
ECMWF FG ASCAT ASCAT QC

• Works fine
• Congratulations from ESA, NOAA and NASA HQ
• Room for improvement

ASCAT demonstration wind service since March
28, 2007
5
Coverage

• Three times as much as ERS
• Comparable to Seawinds smaller nominal area but
  fewer gaps
• because ASCAT C-band sees through rain

Seawinds
ASCAT
20070406 descending
20070406 descending
6
Resolution
Effect of Seawinds resolution

• ASCAT product has 25 km resolution
• 12.5 km product planned
• (needs advanced retrieval techniques)

7
OSI SAF ASCAT product release schedule
Product Coverage Demonstration Operational
25 km Global 2007 2008
12.5 km Global 2008 2009
Coastal 12.5 km Global 2009 2011
Integration of regional EARS products into
global OSI SAF products in 2008
8
Quality

• Accuracy scatterometer 2 m/s (vector
  wind)

• Available L1b commissioning sets
• SS 19/10/06 29/01/07
• ZZ 30/01/07 12/01/07
• ZZZ 13/02/07 today
• New set expected in June,
• after full transponder calibration
• Ocean calibration improves quality

9
Relating backscatter to winds

• CMOD5.5
• F (u,v) ? (sf, sa, sm )
• F maps (u,v) on cone
• in s-space
• F depends on elevation
• and azimuth

10
Ocean Calibration, OCfor improving L2 wind
product

• Purpose
• To correct for lack in absolute calibration
• To make ASCAT compatible with ERS product
• To disseminate a demonstration wind product
• Based on ERS heritage
• CMOD5.5, ice model knowledge in ERS ? range
• Continuity of geophysical data set
• Method 4 mappings that scale s s
  
• Scale such that no bias remains between ASCAT and
  ECMWF winds
• Use of ice model to further verify OC
• Scale such that ASCAT data lie on CMOD5.5 cone

11
More on OC corrections

• Assume all corrections are backscatter (z)
  scaling factors (dB additions) all L1b
  calibration changes so far prove to be so
• Use ECMWF wind speed bias at every WVC as
  absolute wind reference to achieve uniform
  quality
• Use CMOD5.5 wind speed sensitivity to compute
  beam scaling factors at every WVC
• Assume that the CMOD5.5 wind speed sensitivity at
  8 m/s is representative of the full wind domain
  sensitivity (quasi-linear 8 m/s is mean global
  speed)
• This backscatter data scaling would ideally
  provide the same wind PDF for all WVCs after
  CMOD5.5 inversion

12
Mid-axis scaling
Vertical cut for WVC 42 (rightmost)
smid
smid
. Upwind . Downwind
. Upwind . Downwind
sf,a
sf,a
Current Level 1b release
KNMI visual correction

• CMOD5.5 is better fitted due to the down scaling
  of mid beam measurements

13
Aft-fore axis symmetry correction
Horizontal projection for WVC 42
sa
sa
. Upwind . Downwind
. Upwind . Downwind
sf
sf
Level 1b 1st release
KNMI visual correction

• CMOD5.5 is better fitted due to the symmetrical
  scaling of fore and aft measurements

14
Wind speed bias correction
Vertical cut for WVC 42
smid
smid
. Upwind . Downwind
. Upwind . Downwind
sf,a
sf,a
KNMI visualwind correction
KNMI visual correction

• CMOD5.5 is better fitted due to the upward
  scaling of measurements

15
Wind speed bias correction
Horizontal projection for WVC 42
sa
sa
. Upwind . Downwind
. Upwind . Downwind
sf
sf
KNMI visual correction
KNMI visualwind correction

• CMOD5.5 is better fitted due to the upward
  scaling of measurements

16
Validation against ECMWF

• ASCAT inverted winds do not only have lower and
  more uniform bias w.r.t. ECMWF, but also much
  lower wind direction SDE (ECMWF ASCAT error)

17
- Validation against observations -
Higher resolution - Closer to the coast -
Better wind quality
Future SAF ASCAT scatterometer RD
18
Concluding remarks

• ASCAT winds close to perfect
• ASCAT demonstration wind service since 28 March
  2007
• Calibration and ERS scatterometer compatibility
  can be achieved by scaling CMOD5.5 backscatter
• After ASCAT L1b transponder calibration, KNMI
  will reprocess data and incorporate remaining
  biases in CMODx.x
• Continuity of current ASCAT L2 wind quality
  is guaranteed over this L1b calibration

More info at http//www.knmi.nl/scatterometer/