SCIAMACHY Water Vapour Retrieval using AMC-DOAS

1
SCIAMACHY Water Vapour Retrieval using AMC-DOAS
S. Noël, M. Buchwitz, H. Bovensmann, J. P.
Burrows Institute of Environmental
Physics/Remote SensingUniversity of Bremen,
Germany
2
The AMC-DOAS Retrieval Method

• Air Mass Corrected (AMC-)DOAS based on
  well-known DOAS method
• Uses only differential structures of
  sun-normalised radiances
• Numerically fast algorithm
• Main differences to standard DOAS
• Parameterisation of saturation effectNon-linear
  dependence of absorber amount from absorption
  depth
• Air Mass Factor (AMF) correctionfrom O2
  absorption in same fitting window
• Inherent data quality check to mask out too
  cloudy ground pixels, etc.
• Has been applied successfully to GOME and
  SCIAMACHY measurements

3
SCIAMACHY and GOME H2O Columns

• SCIAMACHY has higher spatial resolution than GOME
  ( 30 km x 60 km)
• Advantage of VIS spectral regionRetrievals over
  land and ocean possible (unlike MW sensors)
• AMC-DOAS method requires no calibration with
  external sources
• Independent data source

4
AMC-DOAS Results

• Analysis of all available SCIAMACHY nadir data
  for the year 2003 (Level 1 NRT and consolidated
  data)
• Automatic retrieval for all 2004 SCIAMACHY Level
  1 NRT data(see also http//www.iup.physik.uni-br
  emen.de)
• Remarks
• Not all data are available larger gaps
  especially in November 2003
• Inclusion of unconsolidated data may influence
  weighting of individual measurements
• Insufficient radiometric calibration may have an
  influence on the data quality (although expected
  to be small)
• Always the same (specially calibrated) solar
  reference spectrum used for SCIAMACHY retrieval
  (provided by J. Frerick, ESA)
• No correction for surface elevation
• All data have been gridded to 0.5 x 0.5 for the
  comparison with SSM/I and ECMWF results

5
SSM/I H2O Columns(27 January 2003)

• SSM/I gridded Integrated Water Vapour data
  provided by GHRC
• Only descending part of DMSP F-14 orbit (equator
  crossing at 0800 LT)
• Only data over ocean available

6
ECMWF H2O Columns(27 January 2003)

• Operational daily analysis data provided by
  ECMWF
• Not independent from SSM/I data
• Daily averages derived from 6-hourly values
  (integrated over height)

7
SCIAMACHY AMC-DOAS H2O Columns(27 January 2003)

• Regular gaps from alternating limb- nadir
  measurement mode
• Additional gaps from AMC-DOAS quality check
• Max. SZA 88
• AMF correction factor has to be larger than 0.8
• (mainly because of clouds)

(swath data)
8
Correlation (27 January 2003)
SCIA vs. SSM/I
SCIAvs. ECMWF

• Good correlation with both SSM/I and ECMWF
  columns
• On average good agreement (better with ECMWF
  data)
• Smaller SCIA columns seem to be lower, higher
  larger than correlative data
• Deviations difficult to quantify because of large
  scatter

9
Scatter of Water Vapour Data

• Scatter is mainly due to high spatial and
  temporal variability of water vapour
• Difficult to compare individual measurements
  which are (initially) on different temporal
  and/or spatial scales
• Scatter can not be significantly reduced by
  averaging more data (but correlation and mean
  values may improve)
• General problem for validation/verification of
  water vapour products
• Concentrate on long-term analysis of correlation
  and mean values

10
Long-Term Deviations
SCIA vs. SSM/I
SCIA vs. ECMWF

• Mean deviation with SSM/I - 0.2 g/cm2

• Mean deviation with ECMWF - 0.05 g/cm2

11
ECMWF Monthly Mean October 2003
12
SCIAMACHY Monthly Mean October 2003
Preliminary data!
13
Difference SCIAMACHY - ECMWF
Preliminary data!
14
Comparisons with other ENVISAT Sensors

• Other ENVISAT instruments providing water vapour
  column data
• MERIS
• AATSR
• MWR
• Here First comparisons with AATSR and MWR water
  vapour data provided by I. Barton, CSIRO, Hobart,
  Australia
• Advantage of intercomparison Minimum temporal
  offset
• Disadvantages Different spatial resolution,
  ENVISAT products not fully validated yet
• Current limitations
• AATSR and MWR data not independent
• Only sub-satellite track data over ocean (cloud
  free), only few days

15
First Comparisons with AATSR and MWR
AATSR
MWR
Preliminary data!
Preliminary data!

• First preliminary results, only 4 days analysed
  up to now (partly limited by availability of
  SCIAMACHY Level 1b data)
• Agreement with MWR data slightly better than with
  AATSR

16
Summary Conclusions

• SCIAMACHY visible H2O columns agree well with
  correlative data
• High scatter ( 0.5 g/cm2 ), mainly due to
  atmospheric variability
• Validation of water vapour columns difficult
• Mean SCIAMACHY AMC-DOAS water vapour columns
  typically lower than ECMWF and SSM/I data
• SCIAMACHY monthly means look reasonable some
  features need further investigation
• Quite good agreement with first AATSR and MWR
  water vapour results
• SCIAMACHY can provide a new independent global
  water vapour data set

17
Acknowledgements

• SCIAMACHY data have been provided by ESA.
• SSM/I data have been provided by the Global
  Hydrology Resource Center (GHRC) at the Global
  Hydrology and Climate Center, Huntsville,
  Alabama.
• We thank the European Center for Medium Range
  Weather Forecasting (ECMWF) for providing us with
  analysed meteorological fields and our colleagues
  J. Meyer-Arnek and S. Dhomse for assistance in
  handling these data.
• MWR and AATSR water vapour data have been
  provided by I. Barton, Marine Research,
  Commonwealth Scientific and Industrial Research
  Organisation, Hobart, Tasmania, Australia.
• This work has been funded by the BMBF via
  GSF/PT-UKF and DLR-Bonn and by the University of
  Bremen.

18
Importance of Water Vapour

• One of the most abundant atmospheric gases
• More than 99 located in troposphere
• 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 data
• In-situ measurements radio sondes, ground-based
  airborne measurements
• Space borne (N)IR and MW sensors (TOVS, SSM/I,
  MODIS, MERIS)
• GPS observations
• Additional data source Measurements in visible
  spectral region(GOME SCIAMACHY nadir data)

19
Correlation for 2003

• In general good correlation over the whole year
• Lower correlations for SSM/I during the first
  months mainly due to low number of coincidences
  (missing data)
• Reduced correlation with ECMWF data in summer

20
Deviation SCIAMACHY SSM/I
- 0.2 g/cm2
21
Deviation SCIAMACHY ECMWF
- 0.05 g/cm2