Graphics: ESA Graphics: ESA Graphics: ESA Long-Term ...earth.esa.int/workshops/atmos2009/participants/425/[email protected] Atmospheric Science Conference

Institut für Umweltphysik/Fernerkundung Physik/ElektrotechnikFachbereich 1

[email protected]

Atmospheric Science Conference 2009

Long-Term Time Series of Water Vapour Total Columns from GOME,

SCIAMACHY and GOME-2S. Noël, S. Mieruch, H. Bovensmann, J. P. Burrows

Institute of Environmental Physics/Remote SensingUniversity of Bremen, Germany

Graphics: ESA Graphics: ESA Graphics: ESA


[email protected]


Importance of Water Vapour• Strongest natural greenhouse gas.• Significant contributions to atmospheric chemistry, weather and climate.• High spatial and temporal variability.• Global water vapour data especially required for global models.• Current sources for global water vapour data:

– In-situ measurements: Radio sondes, ground-based & airborne / balloon measurements.

– Space borne (N)IR and MW sensors (like TOVS, SSM/I, MODIS, MERIS).– GPS observations.

• Especially only limited information in polar regions (small columns, ice).• Additional data source: GOME, SCIAMACHY & GOME-2 nadir

measurements in the visible spectral region (~700 nm) derived with AMC-DOAS (Air Mass Corrected DOAS).


[email protected]


The GOME-type InstrumentsGOME SCIAMACHY GOME-2

Launch date 21 Apr 1995 1 Mar 2002

ENVISAT

sun-synchronous10:00 LT @ equator

typically 60 km x 30 km(depending on orbital position and wavelength)

960 km

Spectral range ca. 240 – 790 nm ca. 220 – 2380 nm ca. 240 – 790 nm

Measurement geometry

Nadir Nadir, Limb, Occultation Nadir

Data distribution(Level 1)

June 1995 – now(since June 2003 reduced coverage)

August 2002 – now March 2007 – now

Expected mission duration

until at least 2011 until at least 2013 with MetOp series until at least 2020

19 Oct 2006

Platform ERS-2 MetOp

Orbit sun-synchronous10:30 LT @ equator

sun-synchronous09:30 LT @ equator

Spatial resolution 320 km x 40 km 80 km x 40 km

Swath 960 km 1920 km

Graphics: ESA Graphics: ESA Graphics: ESA


[email protected]


Characteristics of the AMC-DOAS Products• Limitations:

- Only measurements on the dayside can be used- No total column data for too cloudy scenes

(removed by AMF correction factor check)- Limited spatial and temporal resolution- Currently no data for high mountain areas

(masked out by quality check; can be avoided by use of external surface elevation data base)

• Advantages:- Retrievals possible over land and ocean

(with same algorithm)- No external calibration sources required

Completely independent data set!


[email protected]


Quality of AMC-DOASWater Vapour Products

• Main results of validation:– Good general agreement– Large scatter (~0.5 g/cm2) due to spatial/temporal variability

• Precision estimate:– 3-5% for columns > 0.6 g/cm2

– Lower columns: up to 10%– Absolute precision always better than 0.2 g/cm2

• Accuracy estimate:– Difficult to determine because of large variability of water vapour– Systematic deviations to e.g. ECMWF and SSM/I data ~ 0.1 – 0.3 g/cm2

– SCIAMACHY data typically lower (cloud free bias?)


[email protected]


Combining GOME & SCIAMACHY Data• GOME and

SCIAMACHY instruments are very similar

• GOME measures (only) ~30 min after SCIAMACHY at same geolocation

• SCIAMACHY and GOME measurements have different spatial resolution

• How good do the data match in the overlap time interval?


[email protected]


GOME vs. SCIAMACHY: Example 27 Jan 2003

• Very good correlation• Scatter of data

somewhat smaller than e.g. in comparisonSCIAMACHY - SSM/I

Overall good agreement within the “usual” scatter

GOME & SCIAMACHY AMC-DOAS data V1.0 (spatially gridded to 0.5° x 0.5°)


[email protected]


GOME vs. SCIAMACHY: Long-Term Comparison

• Time interval: Aug 2002 – Dec 2003

• Very good agreement especially for 2003 data

• On average no offset, but some scatter

• Notes:– Reduced GOME

coverage after June 2003

– No actual solar reference spectrum for SCIAMACHY data in 2002

GOME & SCIAMACHY AMC-DOAS data V1.0 (spatially gridded to 0.5° x 0.5°)


[email protected]


GOME-2 Results• MetOp was

launched in October 2006; Level 1 data distribution since March 2007

• AMC-DOAS retrieval method could be easily adapted to GOME-2 (only spectral resolution of RTM data base had to be adapted to slit function)

• Regular retrieval running at IUP Bremen

• Enhanced daily coverage of GOME-2 at similar spatial resolution as SCIAMACHY


[email protected]


GOME-2 vs. SCIAMACHY (1)• Scatter plot shows

good agreement between GOME-2 and SCIAMACHY data

• Results confirmed by statistical analysis based on larger data set

Noël et al., ACP, 2008


[email protected]


GOME-2 vs. SCIAMACHY (2)

• Avg. deviation small

• Small sinusoidal variation (period ~ 9 days; amplitude ~ 6%, some seasonal change)

• Related to the relative position of orbits.

• Possible reason: Scan angle dependencies.

GOME-2 & SCIAMACHY AMC-DOAS data V1.0 (spatially gridded to 0.5° x 0.5°)


[email protected]


The Combined GOME & SCIAMACHY Water Vapour Data Set 1996-2008

GOME:1996 – 2002

SCIAMACHY:2003 – 2008

Time series to be continued with SCIAMACHY and GOME-2 until 2020.

Annual Means Monthly Means


[email protected]


Example Application: Trend analysis• The combined

GOME-SCIAMACHY data set has been used to determine global water vapour changes

• Several regions with significant positive and negative changes over 12 years could be detected

• See poster by Mieruch et al.

Figure and data by S. Mieruch; see also Mieruch et al., ACP, 2008


[email protected]


Conclusions• The AMC-DOAS retrieval method has been successfully

applied to GOME, SCIAMACHY and GOME-2 measurements.

• The resulting AMC-DOAS H2O columns agree well with correlative data and each other.

Potential to extend AMC-DOAS data set until 2020 with GOME-2.

GOME-type instruments can provide a new independentglobal water vapour data set suitable for climatological studies.

• All data are available on request, details on AMC-DOAS web site (http://www.iup.uni-bremen.de/amcdoas/).


[email protected]


Acknowledgements• GOME &

SCIAMACHY data have been provided by ESA.

• GOME-2 data have been provided by EUMETSAT.

• This work has been funded by DLR-Bonn and by the University of Bremen.

Thank you for

your attention!


[email protected]


Backup Slides


[email protected]


Air Mass Corrected DOAS (AMC-DOAS)• Similar to standard DOAS:

– Does not rely on absolute radiometric calibration– Uses differential structures to derive trace gas column

• Main differences:– Considers saturation effect for water vapour (non-resolved lines):

Non-linear relation between absorption depth and absorber amount

– Air mass factor (AMF) correction (from O2 absorption):Considers uncertainties in radiative transfer calculation caused by insufficient knowledge of atmospheric conditions (esp. cloudiness)

Retrieval also possible for partly cloudy pixels

– AMF correction also used as quality check

DOAS = Differential Optical Absorption Spectroscopy


[email protected]


Fitting Window

• 688 to 700 nm• Absorptions of both water vapour

and O2 of same magnitude (required for air mass correction)

• Further advantage:„Harmless“ region w.r.t. calibration of GOME, SCIAMACHY & GOME-2

very stable0

0.01

0.02

0.03

0.04

0.05

0.06

600 700 800

Sun

-nor

mal

ized

Rad

ianc

e

Wavelength, nm

GOME Channel 4

O2

O2

H2O

H2OH2O

O2

Fitti

ngW

indo

w

Documents

Graphics: ESA Graphics: ESA Graphics: ESA Long-Term ...earth.esa.int/workshops/atmos2009/participants/425/[email protected] Atmospheric Science Conference