University of Bremen IUP/IFE SCIAMACHY WFM-DOAS: Images (v0.4/0.41)
University of Bremen IUP/IFE SCIAMACHY WFM-DOAS: Images (v0.4/0.41) |
|
Here we have collected a series of global maps of our WFM-DOAS Version 0.4/0.41 data products derived from the SCIAMACHY/ENVISAT nadir spectra.
We have processed all available orbits (using only consolidated orbit files, i.e., full Level 1 products) of the year 2003. Currently, not all fourteen daily orbits are available for each day. No orbits are available for November and December 2003 and there are large data gaps for many time periodes. Nevertheless, we have produced maps of "monthly averages" and "seasonal averages" from the data we have.
This work has been carried out within the EU 5th FP research project EVERGREEN . It would not have been possible without funding from the EU via EVERGREEN and from the German BMBF via DLR and GSF and the University and the State of Bremen, who have been supporting our work over many years.
Below we show seasonal averages. Monthly averages are shown here.
The maps have been generated by averaging all cloud free pixels over land (over ocean the columns are less less precise due to low ocean albedo in the near-infrared spectral region). In addition, only ground pixels have been used where at least a resonable fit has been achieved (estimated column error less than 60%).
Seasonal averages:
;sc.focus();)
;sc.focus();)
;sc.focus();)
;sc.focus();)
The maps have been generated by averaging all cloud free pixels over land (over ocean the columns are less less precise due to low ocean albedo in the near-infrared spectral region). In addition, only ground pixels have been used where at least a resonable fit has been achieved (estimated column error less than 10%).
Seasonal averages:
;sc.focus();)
;sc.focus();)
;sc.focus();)
;sc.focus();)
The maps have been generated by averaging all cloud free pixels over land (over ocean the columns are less less precise due to low ocean albedo in the near-infrared spectral region). In addition, only ground pixels have been used where at least a resonable fit has been achieved (estimated column error less than 10%).
It was not possible to simply average the WFM-DOAS Version 0.4 methane data products because of a time dependent bias caused by time dependent channel 8 transmission changes (due to (reversible) ice build-up on the detector). This "ice issue" affects the precision of the measurements but also introduces a bias. Dr. Hans Schrijver, SRON, The Netherlands, and colleagues have found out by analysing in-flight spectral lamp measurements that the "ice issue" results in changes of the instrument slit function. This time dependence is not yet considered in WFM-DOAS.
To generate the maps shown below we applied a simple first order bias correction to the WFM-DOAS Version 0.4 methane data products. This modified v0.4 product is our basically our version 0.41 methane product. We are using a time dependent (daily) correction factor determined from the in-orbit transmission (loss) measurements (see: SOST-IFE Light Path Analysis. ). Details see figure given below.
;sc.focus();)
The red curve ("Transmission (orig.)") shows the time dependent relative transmission of SCIAMACHY channel 8.
The black curve shows (mainly) the bias of the WFM-DOAS Version 0.4 methane data product.
Both curves are independently generated but show a clear correlation which confirmes that the bias is clearly related to the
"ice issue". The magenta curve ("Transmission (transformed)") is a linear transformation of the red curve and is
in good agreement with the observed bias of the methane data. For each day where we have methane data
we have determined a correction factor from the transmission loss curve. Before the methane data have been averaged
all data for a given day have been divided (!) by the correction factor. The daily correction factors (for all 150 days
of the year 2003 where we have data) can be obtained
from
here
(simple ASCII file: first column: day of year; second column: correction factor).
We also applied an other correction to the methane data: Christian Frankenberg from IUP, University of Heidelberg, Germany, showed that the accuracy of the SCIAMACHY methane measurements can be improved by dividing methane by carbon dioxide rather than by oxygen (for spectroscopic, radiative transfer, and calibration reasons). Our preliminary analysis confirm this. The methane data shown here are not exactly the WFM-DOAS Version 0.4 data products but have been derived from them using the simple formula: (XCH4_shown) = XCH4 / (XCO2 / 370), where XCH4 and XCO2 are the WFM-DOAS Version 0.4 data products. This is not as straight forward as the simple formula suggests because of the different ground pixel sizes of the SCIAMACHY CO2 and CH4 measurements (for the currently used fitting windows the CH4 pixels are twice as large as the CO2 pixels). We have used a simplified approach to deal with this in order to produce the figures given below. We will consider this approach for future version of our methane data product and will also consider retrieving methane from channel 6 (as done by Christian Frankenberg who has done important pioneering work using this channel) as this channel is not affected by the ice issue.
Seasonal averages:
;sc.focus();)
;sc.focus();)
;sc.focus();)
;sc.focus();)
Last modification: 28-June-2006.
Authors: Michael.Buchwitz@iup.physik.uni-bremen.de.
Back to: IUP/IFE SCIAMACHY WFM-DOAS page