active techniques for wind and wave observations ... · pdf fileecmwf seminar 2014 ... active...

ECMWF Seminar 2014 - Active techniques for wind and wave observations Slide 1

Active Techniques for Wind and Wave Observations: Scatterometer, Altimeter (& SAR)

Giovanna De Chiara & Saleh Abdalla

ECMWF


Scatterometer Winds

The importance of scatterometer wind observations

Physical mechanism and wind inversion

Data usage at ECMWF and their impact

How we can improve the use and impact

Concluding remarks

Altimeter Wind and Waves Introduction

Verification and Monitoring

Wave Data Assimilation

Assessment of Model Performance (inc. Model Effective Resolution)

Error Estimation

Long Term Studies

Concluding Remarks

Synthetic Aperture Radar (SAR)

Outline


Why is Scatterometer important?The scatterometer measures the ocean surface winds (ocean wind vector).

Ocean surface winds:

affect the full range of ocean movement (from individual surface waves to

complete current systems)

modulate air-sea exchanges of heat, momentum, gases, and particulates

Wind observations below 850 hPa

FSO values relative quantities (in %)

Wide daily coverage of ocean surface winds

[Horanyi et al, 2013]

Ex: 1 day of ASCAT-A data


A Scatterometer is an active microwave instrument

(radar)

Day and night acquisition

Not affected by clouds

The return signal, backscatter (σ0 sigma-nought),

which is sensitive to:

Surface wind (ocean)

Soil moisture (land)

Ice age (ice)

What is a Scatterometer?

Scatterometer was originally designed to measure ocean winds:

Measurements sensitive to the ocean-surface roughness due

to capillary gravity waves generated by local wind conditions

(surface stress)

Observations from different look angles: wind direction

[ASCAT-A from http://www.eumetsat.int]


Dependency of the backscatter on... Wind speed


Dependency of the backscatter on... Wind directionBackscatter response depends on the relative angle between the pulse and capillary wave

direction (wind direction)

Fore Beam

Mid Beam

Aft Beam

Wind direction wrt Mid Beam



Mid

Fore

Aft

C-band SCAT geometry


How can we relate backscatter to wind speed and direction?

The relationship is determined empirically

Ideally collocate with surface stress observations

In practice collocation between buoy winds and 10m model winds

Geophysical model functions (GMF) families

C-band: CMOD

Ku-band: NSCAT, QSCAT

U10N: equivalent neutral wind speed

: wind direction w.r.t. beam pointing

: incidence angle

p : radar beam polarization

: microwave wavelength

𝜎0 = 𝐺𝑀𝐹(𝑈10𝑁 , 𝜙, 𝜃, 𝑝, 𝜆)


How can we relate backscatter to wind speed and direction?

For the C-band observations, σ0 triplets lies on a double conical surface

σ0 triplets with the same wind speed

Wind inversion: search for minimum distances between the σ0 triplets and all the

solutions on the GMF surface.

Noise in the observations can change the position wrt the cone: wind ambiguities

0°

90°

180°

270°

CMOD5

z_

mid

[Stoffelen]


C- band scatterometers

Used on European platforms (1991 onwards):

SCAT on ERS-1, ERS-2 by ESA

ASCAT on Metop-A, Metop-B by EUMETSAT

f~5.3 GHz (λ~5.7 cm)

VV polarization

Three antennae

Pros and cons: Hardly affected by rain

High quality wind direction (especially ASCAT)

Two nearly opposite wind solutions

Rather narrow swath:

ERS-1/2: 500km

ASCAT-A/B: 2x500km


Ku-band scatterometersUsed on US and Japanese platforms, later also India, China

NSCAT, QuikSCAT, SeaWinds by NASA (and Japan)

Oceansat by ISRO

Haiyang-2A by China

Pros and cons:

Up to four wind solutions (rank-1 most often correct one)

Broad swath (1,800km)

Affected by rain

Problems regarding wind direction:

azimuth diversity not good in centre of swath

outer 200km only sensed by one beam.

f~13 GHz (λ ~ 2.2 cm)

VV and HH polarization:

Two rotating pencil-beams (4 look angles)

OSCAT daily coverage


Operational usage of Scatterometer winds at ECMWF

95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14

ERS-1

ERS-2 ERS-2 (Regional Mission Scenario)

QuikSCAT

METOP-A ASCAT

METOP-B ASCAT

Oceansat-2 (OSCAT)

C-Band Ku Band


ASCAT (25km) from EUMETSAT

Wind inversion is performed in-house: Sigma nought bias correction Inversion using CMOD5.N Wind speed bias correction

ASCAT-A & ASCAT-B assimilation strategy

Quality control, thinning: Screening: sea ice check based on SST and sea ice data Thinning: 100 km Threshold: 35 m/s

In the 4D-Var 2 solutions provided: best one chosen dynamically during the minimization

Assimilated as 10m equivalent neutral winds Observation error: 1.5 m/s

ASCAT-A vs ECMWF FG Wind Speed


OCEANSAT-2 assimilation strategy

OCEANSAT-2 (50km)

Use of Level-2 wind products from OSI-SAF (KNMI) Wind speed bias correction (wind-speed dependent) Quality control:

Screening: sea ice check on SST and sea ice model Rain flag No thinning; weight in the assimilation 0.25 Threshold: 25 m/s

Assimilated as 10m equivalent neutral winds Observation error: 2 m/s OSCAT vs ECMWF FG Wind Speed


Impact of scatterometer windsContribution to the reduction of the 24h Forecast Error (total dry energy norm)

[Cardinali, 2009, Q.J.R.Met.Soc. 135]

Global statistics with respect to the total observing network

Dec 2012 / Feb 2013


For each storm the min SLP have been detected from the ECMWF model fields

SLP have been compared to observation values (from NHC and JMA)

Statistics based only on cases where ASCAT-A, ASCAT-B and OSCAT passes were available

Dec 2012/ Feb 2013

Impact of scatterometer winds…on Tropical Cyclone FC


Verified against conventional temperature observations

[P. Laloyaux]

Tropical Atlantic Tropical Indian Tropical East Pacific

Temperature difference Scatt – NoScatt

November 2013

Impact of scatterometer winds…on the ocean parameters


analysis increments (U)

Mo

de

l L

eve

ls

137

100

50

ML PL(hPa)

137 ~ 1013

114 ~ 850

96 ~ 500

79 ~ 250

60 ~ 100

Single Observation Experiment (1 ASCAT-A)

137

100

50

N S NS

Is the model able to propagate the information?

analysis increments (V)


How can we improve usage and impact?

Include dependency from other geophysical quantities such as ocean currents

Stress is related to relative wind

The observation operator should act on relative wind

Accurate ocean current input are needed

Improve QC mostly for extreme events

Because of thinning and QC the strongest winds can be rejected

Test on the observation error, thinning and use of high resolution products


Concluding remarks

ECMWF has a long experience with scatterometry Assimilation since 1995

ERS1/2, QuikSCAT, ASCAT-A/B, Oceansat-2 Future missions: ASCAT-C, Oceansat-3,…

GMF development Monitoring, validation, assimilation, re-calibration

On-going efforts to improve usage and impact Improve QC for tropical cyclones (Huber norm) Adapt observation errors, thinning, super-obbing Include dependency from other geophysical quantities

Use in the Reanalysis ERS1/2 and QuikSCAT in ERA-Interim ASCAT-A reprocessed products will be used in ERA5

Typhoon Haiyan 20131107

ASCAT-A ASCAT-B OSCAT

Scatterometer observations widely used in NWP Ocean wind vectors Positive impact on analysis and the forecast Global scale and extreme events Available continuously from 1991 onwards


Altimeter Wind & Waves


Scatterometer Winds


Scatterometer principle



Concluding remarks





Error Estimation

Long Term Studies

Concluding Remarks


Outline


Introduction


Radar Altimeters

● Radar altimeter is a nadir looking instrument.

● Specular reflection.

● Electromagnetic wave bands used in altimeters:

Primary: Ku-band (~ 2.5 cm) – ERS-1/2, Envisat, Jason-1/2/3, Sentinel-3.Ka-band (~ 0.8 cm) – SARAL/AltiKa (only example).

Secondary: C-band (~ 5.5 cm) – Jason-1/2/3, Topex, Sentinel-3.S-band (~ 9.0 cm) – Envisat.

● Main parameters measured by an altimeter:

1. Sea Surface Height (ocean)

2. Significant wave height

3. Wind speed

4. Ice/land/lakes characteristics,…

ENVISAT

Radar Altimeter-2

Sentinel-3

Radar Altimeter (SRAL)


How Radar Altimeter Works:

surface

Height=∆t/2 c

emitted signal returned signal

a t m o s p h e r e

time

ocean surface

illuminated

area

Power of

illumination

radar signal

time

po

we

r

flat surfacerough surface


Information extracted from a radar echo reflected from ocean surface (after averaging ~100 waveforms)

slope of leading edge

significant wave height

amplitude of

the signal

wind speed

epoch at mid-height

sea surface heighttime

po

we

r

waveform


Impact of the atmosphere on Altimeter signal:

● Delay sea surface height

– Water vapour impact: ~ 10’s cm.– Dry air impact: ~ 2.0 m.

● Attenuation wind speed


Typical Daily Coverage of:

Envisat/SARAL

Jason-1/2


Verification & Monitoringof Altimeter

Surface Wind Speed


Altimeter surface wind speed:

● backscatter is related to water surface mean square slope (mss).

● mss can be related to wind speed.

● Stronger wind higher mss smaller backscatter.

● Errors are mainly due to algorithm assumptions, waveform retracking(algorithm), unaccounted-for attenuation & backscatter.

amplitude of

returned signal

wind speed

time

po

wer

waveform

emitted signal backscatter


Relation between wind speed and altimeter backscatter


Comparison between

ENVISAT NRT wind speed

and ECMWF model AN (top)

and in-situ measurements

(bottom) for all data;

1 Jan. 2011 – 31 Dec. 2011

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E

Typical locations of in-situ measurements

0 5 10 15 20 25

ECMWF Wind Speed (m/s)

0

5

10

15

20

25

EN

VIS

AT

W

ind

Sp

ee

d

(m/s

)

SYMMETRIC SLOPE

CORRELATION

SCATTER INDEX

STANDARD DEVIATION

BIAS (ENVISAT - ECMWF)

MEAN ENVISAT

MEAN ECMWF

ENTRIES

STATISTICS

REGR. CONSTANT

REGR. COEFFICIENT

1.0339

0.9461

0.1437

1.1181

0.2924

8.0723

7.7799

1128826

0.6076

0.9595

1

5

10

50

100

500

1000

5000

10000

100000

0 5 10 15 20 25

Buoy Wind Speed (m/s)

0

5

10

15

20

25

EN

VIS

AT

W

ind

Sp

ee

d

(m/s

)

SYMMETRIC SLOPE

CORRELATION

SCATTER INDEX

STANDARD DEVIATION

BIAS (ENVISAT - BUOY)

MEAN ENVISAT

MEAN BUOY

ENTRIES

STATISTICS

REGR. CONSTANT

REGR. COEFFICIENT

0 .9884

0 .9008

0 .1728

1 .4560

- 0 .1138

8 .3144

8 .4282

14597

0 .7128

0 .9019

1515

3050100300500100050000


Comparison between Jason-

2 NRT wind speed and

ECMWF model AN (top) and

in-situ measurements


1 May 2013 – 30 April 2014

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E



Verification & Monitoringof Altimeter

Significant Wave Height


Altimeter Significant Wave Height (SWH)

time

po

we

rwaveform

slope of leading edge

SWH

SWH is the mean height of highest 1/3 of the surface ocean waves.

Higher SWH smaller slope of waveform leading edge.

Errors are mainly due to waveform retracking (algorithm) and instrument characterisation.


Global comparison between Altimeter and ECMWF wave model (WAM) first-guess SWH values (From 02 February 2010 to 01 February 2011)

Jason-1Jason-2Envisat

0 2 4 6 8 10 12 14

WAM Wave Heights (m)

0

2

4

6

8

10

12

14

EN

VIS

AT

Wa

ve

He

igh

ts

(m

)

SYMMETRIC SLOPE

CORRELATION

SCATTER INDEX

STANDARD DEVIATION

BIAS (ENVISAT - WAM)

MEAN ENVISAT

MEAN WAM

ENTRIES

STATISTICS

REGR. CONSTANT

REGR. COEFFICIENT

1 . 0026

0 . 9786

0 . 1051

0 . 2733

- 0 . 0163

2 . 5851

2 . 6014

1 125908

- 0 . 0587

1 . 0163

1

5

10

50

100

500

1000

5000

10000

100000

0 . 9 983

0 . 9 791

0 . 1 044

0 . 2 826

- 0 . 0 032

2 . 7 041

2 . 7 073

142 50 55

0 . 0 637

0 . 9 753

1 . 0 397

0 . 9 738

0 . 1 200

0 . 3 232

0 . 1 140

2 . 8 078

2 . 6 939

138 29 9 7

0 . 1 072

1 . 0 025


Comparison between

Cryosat-2 NRT SWH and

ECMWF model FG (top) and



1 April 2013 – 17 June 2014

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E



Comparison between

SARAL (Ka) NRT SWH and

ECMWF model FG (top) and


(bottom) for all data;1 May 2013 – 30 April 2014

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E



Assimilation of Altimeter Significant Wave Height


Data Assimilation Method

● Assimilation Method for Altimeter data:

– Data are subjected to a quality control process

(inc. super-obbing).

– Bias correction is applied.

– Simple optimum interpolation (OI) scheme on SWH.

– The SWH analysis increments wave spectrum

adjustments... (several assumptions)


Operational Assimilation of SWH

NRT Altimeter SWH operational assimilation at ECMWF:

- ERS-1 URA, 15 Aug. 1993 – 1 May 1996;

- ERS-2 URA, 1 May 1996 – 23 Jun. 2003;

- ENVISAT RA-2 FDMAR, 22 Oct. 2003 – 7 Apr. 2012;

- Jason-1 OSDR, 1 Feb. 2006 – 1 Apr. 2010;

- Jason-2 OGDR, 10 Mar. 2009 – on-going;

- Cryosat-2 FDM, Coming model change (~Q4, 2014);

- SARAL OGDR, Coming model change (~Q4, 2014).

10 Mar.

2009

Jason-1

Envisat

Jason-2

EnvisatEnvisat

8 Jun.

2009

Jason-1

Jason-2

Envisat

1 Apr.

2010

Jason-2

Envisat

7 Apr.

2012

Jason-2

22 Oct.

2003

ERS-2

1 May

1996

ERS-1

15 Aug.

19931 Feb.

2006


Mean impact of assimilating Cryosat-2 SWH (with BC) on SWH analysis [CS & J2] - [J2 only]


Impact of assimilating altimeter data on SWH random error as verified against Extra-tropical in-situ data, Feb. – April 2014

4

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E



Impact of assimilating altimeter data on SWH random error as verified against Tropical in-situ data, Feb. – April 2014

60°S 60°S

40°S 40°S

20°S 20°S

0° 0°

20°N 20°N

40°N 40°N

60°N 60°N

160°W

160°W

120°W

120°W

80°W

80°W

40°W

40°W

0°

0°

40°E

40°E

80°E

80°E

120°E

120°E

160°E

160°E



Impact of assimilating SARAL data on SWH error as verified against Cryosat -2 SWH,Feb. – March 2014


Mean Impact of using SARAL (with BC) SWH on Geopotentialanomaly correlation (bars @ 95% level)

Northern Hemisphere Southern Hemisphere

1000 hPa

500 hPa


Assessment of model changes&

Monitoring of model performance


Change Assessment: Change of SDD between ENVISAT RA-2 and ECMWF Model Wind Speed

IPF

5.0

2

Mod

el C

32R

2

Mod

el C

35R

2

Jan-

2004

Jan-

2005

Jan-

2006

Jan-

2007

Jan-

2008

Jan-

2009

Jan-

2010

Jan-

2011

Jan-

2012

1.0

1.1

1.2

1.3

1.4

U10

Std

. Dev

. of

Dif

fere

nce

(m

s-1

)

Model changeAlt. algorithm

change Model change


Performance Monitoring: Change of SDD between Envisat RA-2 and ECMWF Operational SWH

IPF

6.0

2L04

Jan

-20

04

Jan

-20

05

Jan

-20

06

Jan

-20

07

Jan

-20

08

Jan

-20

09

Jan

-20

10

Jan

-20

11

Jan

-20

12

0.20

0.25

0.30

0.35

0.40

SW

H S

td. D

ev.

of

Dif

fere

nce

(m

)

Change

of alt.

algorithm


Effective Model Resolution


| | | | | | | |

50

00

km

20

00

km

10

00

km

50

0 k

m

20

0 k

m

10

0 k

m

50

km

20

km

10-6

10-5

10-4

wavenumber (m-1)

102

103

104

105

106

Spectr

al D

ensity

(m

3s

-2)

1024 (~7,100 km)

k-5/3

k-2.5

sampling at 7 km 685 seq. 1 year Envisat

k-2.5

k-5/3

~400km

Envisat RA-2

| | | | | | | |

5000 k

m

2000 k

m

10

00

km

500 k

m

200 k

m

100 k

m

50 k

m

20 k

m

10-6

10-5

10-4

wavenumber (m-1)

102

103

104

105

106

Sp

ectr

al

Den

sit

y

(m3

s-2

)

Env. RA-2, 685sp, 1024x7km

T1279, 592sp, 2011, 512x16km

k -5/3

k -2.5

Eff

ecti

ve R

eso

luti

on

= ~

125 k

m

Envisat RA-2

ECMWF Model

Surface wind speed spectra


T255

T511

T639T799

T1279T2047

T3999

0 10 20 30 40 50 60 70 80Grid Resolution, D , (km)

0

100

200

300

400

500

600

700

Sho

rtest

Resolv

ed S

ca

le

(km

)

Full Resolution

50% Reduction8 x D4 x D

| | | | | | | |

5000 k

m

2000 k

m

10

00

km

500 k

m

200 k

m

100 k

m

50 k

m

20 k

m

10-6

10-5

10-4

wavenumber (m-1)

102

103

104

105

106

Sp

ec

tra

l D

en

sit

y

(m3

s-2

)

Envisat RA-2

Model (T1279)

Eff

ecti

ve R

eso

luti

on

~125 k

m

50%

Reso

luti

on

~65 k

m

Effective Model Resolution


Error estimation


0 1 2 3 4 5 6 7 8 9 10Forecast Days

0

1

2

3

4

10

-m W

ind

Sp

eed

Err

or

(m.s

-1)

Model FC

Buoy

Jason-1

Jason-2ENVISAT

0 1 2 3 4 5 6 7 8 9 10

Forecast Days

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Sig

. W

av

e H

eig

ht E

rro

r

(m

)

Model FCBuoy

Jason-1Jason-2ENVISAT

Random Error Estimation of Wind Speed & SWH using triple collocation technique


Long Term Studies


Trend of “zonal” winds between 1992 and 2010 according to:

Scatterometer

&

Altimeter


Mean “zonal” winds in Tropical Pacific basin according to:

Scatterometer

&

Altimeter


Concluding Remarks


Model – Satellite Data: 2-Way Benefit

● Model wind and wave data are used for:

- Monitoring quality of satellite data.

- Assessing changes in processing of satellite data.

- Detecting anomalies in the data.

- Cal/Val of new instruments/products.

● Satellite wind and wave data are used for:

- Data assimilation.

- Monitoring of model performance (inc. model resolution).

- Assessment of model changes.

- Use in reanalyses (assimilation and validation).

● Error estimation.

● Long term assessments & climate studies.


Scatterometer Winds


Scatterometer principle



Concluding remarks





Error Estimation

Long Term Studies

Concluding Remarks


Outline


Synthetic Aperture Concept

● A radar system with a “synthetic” aperture simulates a “virtually”

long antenna in order to increase azimuth resolution without

increasing the actual antenna.

● The green target remains within the radar beam for the distance

travelled by the satellite. The length of the synthesized antenna is

equivalent to this distance. Synthetic Aperture Radar allows for a

resolution of ~30 meters.Azimuth (flight)

direction



● SAR is a side-pointing radar that provides 2D spectra of ocean surface waves (wavelength and direction of wave systems).

● Measures distances in the “slant range” distortions.

● Water motion introduces further (nonlinear) distortions.

● 5 km x 6 (or 10) km images SAR image spectra.

● SAR inversion: ocean wave spectrum is retrieved from SAR spectrum using nonlinear mapping.

https://earth.esa.int/


ENVISAT Advanced Synthetic Aperture Radar (ASAR)



SAR Wave Mode Image & SAR Image Cross-Spectrum (Level 1b)

● Almost full Resolution in the range direction (30 m ~ 1000 m).

● Limited resolution in the azimuth direction (>~200 m).



Global comparison between ASAR and WAM model (2011)

0 2 4 6 8 10 12

WAM Sig. Wave Height (m)

0

2

4

6

8

10

12

AS

AR

Sig

. W

ave

He

igh

t

(m)

SYMMETRIC SLOPE

CORRELATION

SCATTER INDEX

STANDARD DEVIATION


MEAN ASAR

MEAN WAM

ENTRIES

STATISTICS

REGR. CONSTANT

REGR. COEFFICIENT

0. 9565

0. 9591

0. 1381

0. 3690

- 0. 1525

2. 5195

2. 6720

776756

- 0. 0781

0. 9722

15

153050

100300500100050000

0 1 2 3 4 5 6 7 8 9

WAM Swell Wave Height (m)

0

1

2

3

4

5

6

7

8

9

WV

W S

we

ll W

ave

He

igh

t

(

m)

SYMMETRIC SLOPE

CORRELATION

SCATTER INDEX

STANDARD DEVIATION


MEAN WVW

MEAN WAM

ENTRIES

STATISTICS

REGR. CONSTANT

REGR. COEFFICIENT

0. 7303

0. 7884

0. 3534

0. 3903

- 0. 2530

0. 8514

1. 1044

453973

0. 3545

0. 4499

1

5

10

50

100

500

1000

5000

10000

100000

10’s

100’s

OK

Inverted L1b

(full spectrum)

L2

(within azim. cut-off)


Assimilation of SAR Wave Data - Method

●Assimilation Method:

– L1b SAR spectra are inverted to ocean wave spectra.

– Ocean spectra are partitioned into wave systems which then

paired with the corresponding systems in the model.

– Simple OI scheme on integrated parameters of the systems.

– Each partition is adjusted based on its analysis increment.


Assimilation of Wave Data● NRT Altimeter (Ku) SWH operational assimilation at ECMWF:

- ERS-1 URA, 15 August 1993 – 1 May 1996;

- ERS-2 URA, 1 May 1996 – 23 June 2003;

- ENVISAT RA-2 FDMAR, 22 October 2003 – 7 April 2012;

- Jason-1 OSDR, 1 February 2006 – 1 April 2010;

- Jason-2 OGDR, 10 March 2009 – on-going

● NRT SAR spectra operational assimilation at ECMWF:

- ERS-2 SAR UWA, 13 January 2003 – 31 January 2006;

- ENVISAT ASAR WVS, 1 February 2006 – 21 October 2010.

10 Mar.

2009

Jason-1

Envisat

Jason-2

EnvisatEnvisat

8 Jun.

2009

Jason-1

Jason-2

Envisat

1 Apr.

2010

Jason-2

Envisat

7 Apr.

2012

Jason-2

22 Oct.

2003

ERS-2

1 May

1996

ERS-1

15 Aug.

1993

13 Jan.

2003ERS-2 SAR ENVISAT ASAR

1 Feb.

2006

21 Oct.

2010

Alt.

SAR


Impact of assimilating WM data on SWH bias and SDD in the Tropics

0 1 2 3 4 5

-0.16

-0.14

-0.12

-0.10

-0.08

Bia

s =

Mo

d.-

Alt

. (

m)

L2 (QC) L1b No data

0.10

0.15

0.20

0.25

0.30

St.

Dev.

of

Err

or

(m

)

0 1 2 3 4 5FORECAST DAYS (0 = AN)

01 Aug. – 04 Sep. 2008

Verified against Envisat and Jason-1 altimeter SWH data.


SWH Error reduction Due to Assimilating WM (& RA-2 SWH) Products

Verified against in-situ measurements in Tropics


Conclusions (SAR)

• Fast delivery ENVISAT ASAR Wave Mode products:

- L1b inverted in-house operationally assimilated.

- L2 already inverted difficulties in assimilation.

● Assimilation of ASAR WM L1b product leads to positive

impact on the model forecasts (up to 4% error reduction and

last for 2-5 days).

● Assimilation of ASAR WM L2 leads to rather limited impact

(<2% error reduction).

● Similar work will be carried out for Sentinel-1.


Definition of Difference Measures

● Systematic error (Bias) is:

● Root mean square difference (RMSE):

● Standard deviation of the difference (SDD):

● Scatter index (SI):

● x = altimeter data set, T = reference data set (e.g. in-situ),N = total number of collocations

TxTxNN

i

ii

1

1 )(Bias

N

i

ii TxN1

21 )(RMSE

N

i

ii TxN1

21 )Bias(SDD

N

i

ii TxN1

221 Bias)(SDD

T/SDDSI

active techniques for wind and wave observations ... · pdf fileecmwf seminar 2014 ... active...

Documents