Home

GOME Data Processor Extraction Software User's Manual

1. Write LAFSPI Lvl 1b Dt Level ib Data DLR Figure 1 Data flow diagram of the Extraction software The various calibration steps are indicated by the tags LAFS etc which correspond to the filter used in the c option see Section 3 3 5 Dotted lines indicate input parameters which remain valid for parts of a whole orbit GOME Extraction SUM Re Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 3 Instructions for use 3 1 Installation The extraction software comes on the product CD ROM as a set of C routines in the directory programs gdp01_ex The extraction programme can be compiled from this directory using the statement cc c o gdpOl_ex 1m where cc is an ANSI C compiler and gdp0 _ex is the default name of the extraction program for Level 1 data For full information read the file readme_lst txt in the root directory of the CD ROM If you have already a version of gdp0 _ex installed and want to check if this is the latest version available then check the file with release history docs release_101 txt or run the program using the i option see Section 3 3 If you are using an older version of the extractor or 1f you are using a new version with older Level 1 data you may get the following warning message on screen when running the software KKAKXAAKKAKAKKAKKKKAKAAKAAKAKAKAKAKAAKAAKAAKAKAKAAAAKAKAAAKAKAKAAAAKAAAAKAKAAAAKAAKAAKAAAKAAAA WARNING The software ve
2. satellite height cloud clearing algorithm results fractional cover subpixel trace
3. 04 DLR 6 Instructions for use Contrary to the Level 1 extraction where the extraction process converts pre processed signals on the Level 1 Product into calibrated radiance the Level 2 extraction does not involve computa tional steps In this sense it is more like a Product reader It merely converts the binary format into an easier readable ASCII format with the option to select only a subset of the data see Section 6 3 6 1 Installation The extraction software comes on the product CD ROM as a set of C routines in the directory programs gdp12_ex The extraction programme can be compiled from this directory using the statement cc c o gdpl2_ex lm where cc is an ANSI C compiler and gdp 2_ex is the default name of the extraction program for Level 2 data For full information read the file readme_1st txt in the root directory of the CD ROM If you have already a version of gdp12_ex installed and want to check if this is the latest version available then check the file with release history docs release_112 txt or run the program using the i option see Section 6 3 1 If you are using an older version of the extractor or 1f you are using a new version with older Level 2 data you may get the following warning message on screen when running the software KKAKKAAKKAKAKKKKAKAKAKAKAAKAKAKAAKAAKAKAKAAKAAKAKAAKAAKAKAAKAAAAKAAKAAAAKAAAKAAKAAKAAAAKAKAAKAAAAKAAA WARNING The software version used to generate th ER A Leve
4. 1b stray 2a respecitvely For example the default value in the extractor science bands only is b_filter yyyyyynnnn extracting channels 1 and 2 completely has b_filter yyyymyyyy If you wish to extract PMD signals only set b_filter e extract only ground pixels selected via pixel number gdp01_ex p p first p last input_file output_file This selects only the ground pixels from and including pixel number p_first to and including pixel number p_last If only 1 pixel is to be extracted set p_first equal to p_last e extract only ground pixels selected via coordinates gdp01_ex r lat_ul lon_ul lat_lr lon_lr input_file output_file This extracts all pixels whose centres lie within a rectangle with upper left corner coordinates lat_ul lon_ul and lower right corner coordinates lat_lr lon_1r The coordinates are given as latitude longitude in degrees The range for latitude is from 90 to 90 degrees the range for longitude is from 0 to 360 degrees Example to extract all pixels over the north pole down to 70 degrees latitude use the command gdpOl_ex r 90 0 70 360 input lvl output e extract only ground pixels selected via time gdp01_ex t t_first t_last input_file output_file This is similar to the p option except that pixels are not selected according to pixel number but according to measurement readout time The time tags t_first t_last are charac ter strings formatted as DD MMM YYYY_H
5. Dark Signale 22232 o od re lcd aa ola dos al Ps 28 4 3 Correction for EPA ROSS 30 EA A i 29 dd Apply BERG R Rn DIRI 30 4 5 Apply Spectral Calibration Parameters 2 0 0 eee ee nee 31 4 6 Apply Straylight Correction sis Li a 33 47 Apply BSDP curar a ea wie See bee date A teri lola he db 34 4 8 Apply Radiance Response cvs organo anne eared Ba 35 4 9 Apply polarisation correction rl a Bie 36 4 10 Apply Degradation Parameters or azimuth asymmetry of the BSDF 39 4 11 Calculation ol Frost tada dede 40 Appendix Version history cc cece ccc cece cece wee cece eee eee eeeee 41 GOME Extraction SUM Doc No ER SUM DLR GO 0045 6 4 Issue 3 Date 14 12 04 DLR PART II Level 2 Data 6 Instructions for USE 00 A A nes 46 01 Tala A tinti Gah sciogli A eee As gol 46 6 2 Standard Deyel2 EXPAcuO a a eu 47 6 3 Special Options ns ian 48 6 3 1 Options to obtain product information eseese e eee eee nenn 48 6 3 2 Options to select a subset of data ns nr en 49 Ta Appendix Version MISTER RE 50 GOME Extraction SUM 4 a Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 1 Introduction 1 1 Purpose and Scope The Global Ozone Monitoring Experiment GOME was originally conceived as a scaled down version of the scanning spectrometer SCIAMACHY It was given fast track development status by ESA G1 and was launched on 21 April 1995 on board the second European Rem
6. ERSE DTEX EOPG TN 04 0007 Issue 1 A 15 12 2004 GOME Extraction SUM Doc No ER SUM DLR GO 0045 12 Issue 3 Date 14 12 04 PART I Level 1 Data DLR GOME Extraction SUM 180 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 2 Overview of Extraction Algorithms 2 1 Program Concept There are four basic calibration steps needed to convert the instrument binary data into calibrated physical quantities 1 Signal processing correction for dark signal FPA crosstalk pixel to pixel gain and straylight 2 Wavelength calibration assigning to each detector pixel its associated wavelength 3 Radiance calibration conversion of the corrected detector signals of the earth shine spectra to radiance units this includes polarisation correction 4 Irradiance calibration conversion of the corrected detector signals of the solar spectra to irra diance units this includes correction for the BSDF of the diffuser plate Furthermore the measurements have to be geolocated i e the geographical position of the foot print on the Earth s surface has to be determined from the instrument s scan mirror angle and from the spacecraft data Performing these calibration steps is the task of level 0 to 1 processing In order to keep the data product as small as possible the GOME processing is done in two steps All necessary calibration constants are processed from the calibration measurements in the operational processing from
7. Extractor for the solar spec trum However this correction is not based on an average shift but is calculated on the measured spectrum using on line cross correlation with the Kurucz spectrum This correction is applied as default but can be switched off using the k option in the Extractor Doppler shifts The solar spectra will be doppler shifted due to the motion of the spacecraft in the direction of the light source These shifts vary from maximally 0 06 detector pixel in channels 1 3 to maximally 0 1 detector pixel in channels 2 4 This does not in any way influence the wavelength calibration of the instrument as such the wavelength of each pixel for calibration parameters is given in the instrument reference frame not in the moving solar reference frame Therefore the wavelength calibration on the level 1b product 1s in the instrument rest frame only The user must be aware of this and if necessary account for solar doppler shift himself The velocity of the spacecraft is around 7 6 km s the exact value is provided on the Level 1 prod uct The observation angle to the Sun is around 21 degree from the forward position depending on the time of the year The exact solar azimuth angle w r t spacecraft is provided with the Sun measurements not the Sun Reference spectrum on the Level 1 product as far as 1t contains Sun measurements Directions for use The calibration is always applied automatically The time consuming cross c
8. Jul 18 13 49 06 MET DST 1996 FIRST OPERATIONAL VERSION EN Release 0 90Mon Jul 15 17 12 10 MET DST 1996 ay o New option x extract specific ground pixels EN Release 0 80Thu Jun 20 10 42 58 MET DST 1996 o Support new GOME co adding mode EN Release 0 7Fri Apr 11 10 11 49 MET 1996 7 o Level 1 Data product format changed o Generation of Sun and Moon products o FPA cross talk correction o Polarization correction for pixel with IT gt 1 5 seconds o Reflectivity jumps flagged o Changes on the KeyData error factors Version 0 6Tue Nov 28 14 07 13 MET 1995 ay GOME Extraction SUM Doc No ER SUM DLR GO 0045 44 Issue 3 Date 14 12 04 DLR EN Version 0 5Wed Oct 25 13 35 26 MET 1995 NE E o KeyData version 7 0 o Uses only the 6 5 f2 function o Polarization correction of band la using an averaged p7 o Error values calculated using one average KeyData factor per channel o Adds the flag w write the calibration data to the extracting pro gram I EN Alfa ReleaseMon Jul 31 14 00 00 MET 1995 NA o KeyData version 5 0 amp 6 0 o Uses an interpolated f2 function o Wrong polarization correction for band la o Error values calculated for signal to noise DLR GOME Extraction SUM 45 Doc No ER SUM DLR GO 0045 Issue 3 Date 14 12 04 PART II Level 2 Data GOME Extraction SUM Doc No ER SUM DLR GO 0045 46 Issue 3 Date 14 12
9. character string x_filter which consists of 4 flags which are either y to enable selection or n to disable selection The flags are in sequence of East centre West backscan corresponding to the pixel types 0 1 2 3 on the Level 1 and Level 2 products For example the default value in the extractor all pixel types is x_filter yyyy extracting all but the backscan pixels is via x_filter yyyn GOME Extraction SUM Doc No ER SUM DLR GO 0045 50 Issue 3 Date 14 12 04 DLR 7 Appendix Version history The most recent description of the version history can always be found on the distribution CD in the file docs release_112 txt The majority of software changes reported in that file applies to the operational GDP processor and not to the extraction software Changes in extraction software up to the time of writing are for clarity seperately listed below GDP Level 1 to 2 Processing Release Notes Version 04 00 Wed Dec 1 2004 New operational version based on the UPAS system and the GDOAS algorithm The product format is changed both binary and ascii New option q added in the extractor Version 02 00 Fri Jul 19 14 45 06 MET DST 1996 Version 01 40 Mon Mar 18 11 08 56 MET 1996 Extended level 2 format additional geophysical paramters atmosphere height extended field of view angles at satellite and top of atmoshere earth radius
10. flat fielding the spectra of each detector array This comprises the calcula tion of pixel to pixel gain PPG The PPG calibration is also done only at regular typically monthly intervals Daily observations of the Sun over a diffuser plate provides a solar reference which is used in the calculation of the Earth shine reflectivity spectra The calibrated solar irradiance spectrum is out put as measurement data on the product of the orbit containing the solar measurement for subse quent orbits it is used as calibration data for reflectivity until a new Sun is observed This calibration data is referred to as the Sun Reference spectrum or Sun Mean Reference SMR 4 1 3 Correction for degradation The degradation of the instrument can be monitored by combining measurements from several sources over a longer period in time This is an off line activity The results have to be fed into the data processor as additional calibration constants At the time of writing this research is still an on going activity Preliminary results have been obtained for the Radiance Response function in combination with the BSDF R10 R11 Correc tion factors for these calibration constants can at present be applied in the Extraction Software only see Section 4 10 At present no attempts are being made to correct for the etalon effect a low frequency spatial i e as function of pixel number oscillation in detector response caused by a changing thickness of
11. of the GOME slit over the Lunar surface Note that only a small fraction of all Level 1 product files contains Sun or Moon observations To check whether or not a particular product file contains such observations the contents of the file can be examined output to screen using the command see Section 3 3 1 gdp01_ex i input_file GOME Extraction SUM 19 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 3 3 Special Options A large number of possibilities exists to configure the execution output of the Extractor In the fol lowing we will separate the options into several classes options to obtain product information options to steer the output options to select a subset of data options for scientific post processing options to perform omit calibration steps EA IAS Options from the last 4 classes can be combined with each other Following the UNIX syntax the extraction sofware can be called with flags to implement these options The general syntax is given by gdp01_ex flagl param1_1 param1_2 flag2 param2_1 param2_2 flagn paramn_1 paramn_2 input_file output_file where input_file is a Level 1 product file default extension lv1 and the output file gets an extension el 1 s11 or m11 added depending on the type of observation see Section 3 2 In the following subsections we list the various options with program call using their flag their parameters if any and we provide a short descri
12. spectral channels containing an array detector sometimes an expression like channel la is used for band la etc Extractor a software tool described in this manual to inflate a Level 1 Product into Level 1 Data FPA crosstalk a phenomenon which may cause a variation in detector signal related to the switching of coolers for the Focal Plane Assembly detector housing Ground pixel the footprint on the Earth s surface during one integration time Integration time pattern specifies the integration time of each of the 6 bands Level I Data a data set ususally 1 orbit or a subset thereof which contains the fully calibrated GOME ir radi ance spectra Level I Product the GOME data product from the ERS 2 ground segment it contains the raw detector signals plus all information needed to convert these to calibrated ir radiance this conversion is performed using the extraction software described in this manual Pixel usually one pixel spectral element on the detector is meant depending on the context this may also be shorthand for ground pixel Pixel sub type denotes a certain geometry in the scan pattern pixel type 0 1 2 3 refer to the East Nadir West and Backscan ground pixel respectively Virtual pixel a wavelength interval on the PMD detector which corresponds to the wavelength interval of the corresponding channel array detector pixel Virtual channel boundary the pixel or wavelength which sepa
13. subtracts it from the measurements Apply BSDF OP E Uses measurements of the Sun over the diffuser to calculate a daily Sun Mean Reference spec trum where the irradiance is calibrated via the BSDF function of the diffuser OP The applica tion of BSDF is also performed in the radiometric calibration of science measurements of the Sun Apply Radiance Response E Perform the absolute radiometric calibration of the detector signals including the calculation of the radiometric accuracy Optionally a Sun normalised spectrum the reflectivity is calculated Determine Fractional Polarisation Values OP Calculates the Fractional Polarisation Values at 7 wavelengths over the GOME range from the ratio of PMD to channel signals from the ratio of 2 overlapping channel signals or from single scatter theory at 300 nm Apply Polarisation Correction E The 7 Fractional Polarisation Values are interpolated to wavelength OP E This is partly done using modified spline interpolation Akima partly using a parameterisation of the polarisation curve in the UV From the interpolated polarisation values a polarisation correction factor is derived for and applied to each channel array detector pixel of the Earth shine measurements Quality Flagging OP Several data quality flags are written to the Level 1b product e g for signal detection limit satu rated pixels sun glint conditions Calculation of errors OP E
14. the wavelength of the detector pixel to be calibrated but if pixel to pixel variations in quantum effi ciency are present the quantum efficiency of one pixel would be transferred to that of another pixel if the radiance response function were not corrected for PPG The radiance response function in the GOME calibration Keydata is corrected for PPG and so must be the detector signals before applying that function The PPG calibration data are stored in the form of calibration spectra in the calibration database The PPG spectrum from the calibration database is updated in regular baseline monthly inter vals following a LED calibration measurement The Level 0 to 1b processing takes the most recent PPG values from the database and stores these in the FCD record on the Level 0 to 1b prod uct The application of this calibration parameter consist of dividing each measured detector pixel by the corresponding pixel in the PPG calibration spectrum from the Level 1 product The error on PPG correction is neglected in the calculation of the precision error Directions for use This calibration should always be applied in the Extractor Limitations None GOME Extraction SUM Ds Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 4 5 Apply Spectral Calibration Parameters The main objective of the spectral calibration is to assign a wavelength value to each individual GOME detector pixel In each channel a polynomial has bee
15. the result irrespective of e g a systematic deviation In the GOME level 1 context as precision errors are taken the random errors due to shot noise electronic readout noise and errors due to digitisation of the signal As accuracy errors are taken the precision error augmented for each calibration step with errors on the related calibration Key data The precision error is on the Level product expressed in absolute radiance units on the level 1 ASCH output this is the 3rd column with data the first two columns are wavelength and radi ance The accuracy error is on the Level product expressed in relative units as ratio of accuracy errors in absolute radiance units with the signal in absolute radiance units on the level 1 ASCII out put this is the 4th column with data GOME Extraction SUM 4 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 5 Appendix Version history The most recent description of the version history can always be found on the distribution CD in the file docs release_101 txt The majority of software changes reported in that file applies to the operational GDP processor and not to the extraction software Changes in extraction software are explicitly mentioned as such Changes in extraction software mentioned in the file are up to the time of writing are for clarity seperately listed below EN ER GDP Level 0 to 1 Release Notes AIR Author Diego Loyola DLR DFD
16. to Watt s cm nm For earthshine the PMDs are normalised to the PMD signal of the Solar Reference spectrum and divided by the polarisation fraction to obtain an unpolarised reflectivity See Section 4 8 Unit_Conversion Convert units of Watt to photons GOME Extraction SUM Doc No ER SUM DLR GO 0045 296 2 Issue 3 Date 14 12 04 DLR 4 Background Information on Algorithms 4 1 Overview of Calibration Procedures Calibration of the instrument is performed on 3 different levels 1 Onground Calibration determines the instrument response to calibrated radiance and irradiance sources as function of wavelength and scan mirror angles determines the straylight properties of the instrument provides preliminary calibration of wavelength and dark signal 2 Level 0 to 1b Processing of Calibration Constants Calibration constants which can be directly derived from measurements using on board cal bration sources are determined during the operational Level 0 to 1 processing This comprises dark signal measurements on the night side of each orbit and at regular intervals wavelength calibration using the SLS measurements and PPG calibration using the LED measurements 3 Correction of degradation due to degradation mostly damaging by UV light in optical components or detectors the cal ibration constants for radiance and irradiance will change in time This degradation cannot be derived from on board calibration sources and has
17. 04 DLR GOME Extraction SUM 4 ER Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 Table of Contents IL Introduch n neu 7 1 1 Purpose nd COPE a ea 7 1 2 Document Status and History 0 0 eee eeaeee 8 1 3 Doc ment OVER We ea ee nee a een ae O 8 1 4 Abbreviations and Acronyms td tad ED E 9 lo AOS A Dee te ae AT are Bere 10 156 gt Referente Documents mia ith ooh ind ap 11 PART I Level 1 Data 2 5 Overview of Extraction AlgorithmS ooooooooooooooororrrrrr rro 13 2 1 Program Concept ve A at Ra 13 NS A o 14 Instructions for USE E reiten 17 Sil estalla otro 2 Yee 2 pee he eta add ee IT 17 3 2 Standard Level 1 Extraction u a cisga ein in Bi nn nd 18 3 2 1 Earth shme Radiance ts A A HE 18 3 2 2 Sun Moon Observations 18 o o ladroni 19 3 3 1 Options to obtain product information 0 0 00 cece eee eee 19 3 3 2 Options to steer the output AAA 20 3 3 3 Options to select a subset of dala ais aa 21 3 3 4 Options for scientific post processing 0 cee eee eee eee 22 3 3 5 Options to perform omit calibration steps 0 0 0 eee eee eee 23 Background Information on Algorithms 0 cece cece cece cece ee eees 26 41 Overview of Calibration Proceduresic rl aa oh ees 26 4 1 1 Onground Calibration 0c dada 26 4 1 2 Level 0 to 1 Processing of Calibration Constants 00 26 4 1 3 Correction for degradation veros 27 41 27 Apply
18. AP ER NE ry EX EN Release 2 30 Tue Dec 14 16 31 52 MET 2004 si o New functionality in the Extraction software Two new BSDF correction algorithms have been implemented with smoothed correction parameters using a polynomial in wavelength and with unsmoothed correction parameters read from the degradation file EN Release 2 20 Mon Apr 8 15 51 34 MET DST 2002 N E o New functionality in the Extraction software update spectral calibration of the Sun Mean Reference spectrum using cross correlation with a template spectrum option k disables this feature o Optimize the test to reject Sun calibrations with improper illumination ERS 2 EBM problem starting in 17th January 2001 SF Release 2 10 Thu Mar 1 12 54 19 MET 2001 xy o Correct geolocation of ground pixels in static view mode o Close geolocation gaps Interpolate centre coordinates of ground pixel o Include a test to reject Sun calibrations with improper illumination ERS 2 EBM problem starting in 17th January 2001 GOME Extraction SUM Doc No ER SUM DLR GO 0045 42 Issue 3 Date 14 12 04 DLR JE Release 2 00Wed Jul 21 15 56 04 MET DST 1999 e o New options in the Extraction software e degradation correction f BSDF asymetric azimuth dependancy correction o Include scan mirror correction for albedo computation a
19. Calculates the accuracy errors on Keydata functions and on fractional polarisation and the accu racy and precision errors on the Sun Mean Reference spectrum OP Calculates the precision of the measured spectrum and its accuracy using errors on keydata and fractional polarisation from the Level 1 product GOME Extraction SUM Doc No ER SUM DLR GO 0045 16 Issue 3 Date 14 12 04 Level 1b Product Read Read Pixel Data FCD data IU l Polari Straylight Errors on Sensitivity Coeffs Keydata e fo T T i i I j Sun Leakage PPG Radiance Spectral option s Parameter i Parameter i Sensitivity Parameter i Da a Ge NM 1 A Po Scanning ee i i i i i Incl Geo in esse i Let i l i i i i i Straylight o i O O J i i E i Correction x oh 1 I i Leakage j i Scan Sun i j Scan Sun i LAES i i L is I I pal i i i oe pt ola APP eee ee ee ectral Apply da p i FPA i 1 i l i I i Scanning Sun l Scan Sun i i LAFS HAFS i LA i i Y Y i i i i i I Apply Apply Ryaues Polarisat BSDF I 7 Apply i Y Y I PES Scanning Sun LAFSP LAFSB i N i Scan Sun i Sun Ref LAF Spectrum i i i Scanning LAFSPI Postproc optional Scanning Radiom Calibration Sun Ref Spectrum Calculate Errors
20. Deutsches Zentrum 4 f r Luft und Raumfahrt e V DLR GOME Data Processor Extraction Software User s Manual Doc No ER SUM DLR GO 0045 Issue 3 Date 14 December 2004 GOME Extraction SUM Doc No ER SUM DLR GO 0045 ad Issue 3 Date 14 12 04 DLR Name Function Affiliation Signature Date prepared S Slijkhuis Project Scientist DLR IMF released D Loyola GOME Project Manager DLR IMF Additional Distribution List Function Copies Name Affiliation Internal Development 5 B Aberle Team Y Livschitz L Butenko S Wahl P Kenter Various Available to the public via the internet DLR IMF GOME Extraction SUM 4 o Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 Document Change Log Issue Rev Date Sheet Description of Change Draft 16 04 99 all completly new level 0 gt 1b only 1 08 07 99 all included comments from N Schutgens KNMI updated level 0 gt 1b to version 2 00 level 1 gt 2 completly new 2 05 08 02 1 42 updated level 0 gt 1b to version 2 20 changes mainly in spectral calibration A 08 08 02 15 41 included comments from C Zehner ESRIN 3 14 12 04 7 8 references document history 17 18 46 47 location of products and documents 23 39 new BSDF correction algorithm 47 note on level 2 format change 41 50 release notes GOME Extraction SUM Doc No ER SUM DLR GO 0045 4 Issue 3 Date 14 12
21. H MM SS m where DD is day of the month MMM are the first 3 characters of the name of the month etc GOME Extraction SUM Doc No ER SUM DLR GO 0045 90 Issue 3 Date 14 12 04 DLR Example to extract 20 seconds of data near 11 o clock on April Ist 1996 gdpOl_ex t 01 APR 1996_11 04 35 3 01 APR 1996_11 04 55 3 input lvl output This option cannot be used in combination with the p option e extract only ground pixels selected via pixel type East centre etc gdp01_ex x x_filter input_file output_file This allows to extract only ground pixels at one position in the scan pattern East centre West backscan note that East and West only give the correct compass directions at the descending sun lit part of the orbit on the ascending part it is just the other way around Selection is via a character string x_filter which consists of 4 flags which are either y to enable selection or n to disable selection like in the b option The flags are in sequence of East centre West backscan corresponding to the pixel types 0 1 2 3 on the Level 1 product For example the default value in the extractor all pixel types is x_filter yyyy extracting all but the backscan pixels is via x_filter yyyn 3 3 4 Options for scientific post processing e extract albedo s instead of radiances gdp01_ex a input_file output_file The Earth
22. LR GO 0045 Issue 3 DLR Date 14 12 04 6 3 2 Options to select a subset of data e extract only ground pixels selected via pixel number gdpl2_ex p p_first p last input_file output_file This selects only the ground pixels from and including pixelnumber p_first to and including pixel number p_last If only 1 pixel is to be extracted set p_first equal to p_last e extract only ground pixels selected via time gdp12_ex t t_first t_last input_file output_file This is similar to the p option except that pixels are not selected according to pixel number but according to measurement readout time The time tags t_first t_last are charac ter strings formatted as DD MMM YYYY_HH MM SS m where DD is day of the month MMM are the first 3 characters of the name of the month etc Example to extract 20 seconds of data near 11 o clock on April Ist 1996 gdp12_ex t 01 APR 1996_11 04 35 3 01 APR 1996_11 04 55 3 input lv2 output This option cannot be used in combination with the p option e extract only ground pixels selected via pixel type East centre etc gdp12_ex x x_filter input_file output_file This allows to extract only ground pixels at one position in the scan pattern East centre West backscan note that East and West only give the correct compass directions at the descending part of the orbit on the ascending part it is just the other way around Selection is via a
23. ality assurance purposes and is not further discussed here e write additional polarisation data and calibration data gdp01_ex w input_file output_file In addition to the regular level 1 data in the file output_file ell s11 ml1 this creates two files output_file pcd and output_file fcd The pcd file contains information from the Pixel specific Calibration Data structure see G4 in particular it provides for each ground pixel the following in order of appearance pointer to the spectral parameters and leakage parameters on the fcd file to be used for spectral calibration and dark signal correction of this ground pixel 7 polarisation fractions for PMDs 1 2 3 overlaps 1 2 2 3 3 4 p7 single scatter value 300nm at centre of ground pixel for the integration time of bands 1b 4 plus a value of p7 for the ground pixel centre over the last 1 5 second given as array with effective wavelength p value error on p value In addition the value of the single scatter polarisation angle x see Section 4 9 in radian units is given Note that the numbering of ground pixels on the pcd file starts at pixel 0 whereas the number ing in the ell file starts at pixel 1 the pixel number on the pcd is thus always 1 less than the number of the same ground pixel on the ell file The fcd file is the ASCII output of the Fixed Calibration Data structure see G4 which contains amongst others the following band co
24. and the Extraction algorithms the latter are described further in detail in this document see Section 4 Noted in italics is if the algorithm belongs to the operational Level 0 to 1b processing OP or to the Extraction program E The sequence below also reflects the sequence of signal processing for the Extraction software This is illustrated by the data flow diagram of the Extraction software in Figure 1 Conversion to Engineering units and time correlation OP Calculate orbit and timing information and convert raw housekeeping data into physical quantities e g optical bench temperatures Calculate Geolocaltion OP In the geolocation processing the instrument scan angle and time information is converted to geo graphical coordinates and the solar illumination condition is determined This relies heavily on the use of ESA s Orbit Propagator whose routines will not be explained in detail here Calculate Dark Signal OP The algorithms describe how to use the Dark Measurements to derive calibration constants for FPN and Leakage Current correction Output to the calibration database is a set of dark signal spectra for different integration times Apply Dark Signal OP E Subtract the correct dark signal spectrum from the measured spectrum Correction for FPA noise OP E Applies a correction factor for interference from the Peltier coolers on the Focal Plane Assembly Calculate PPG OP Calculates from the LED calibr
25. ata Directions for use This calibration should always be applied in the Extractor Limitations It is known that in orbit regions affected by cosmic particle bombardments the South Atlantic Anomaly the real background signal will be higher than the usual dark signal This has the effect that within the SAA the dark signal is underestimated whereas it is overestimated on orbits which directly follow an orbit where the SAA was measured at the night side It is also known that due to interference effects from the Peltier coolers the real background signal may differ a little plus or minus from the calibrated one this effect is addressed in the next section on FPA correction GOME Extraction SUM 4 29 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 4 3 Correction for FPA noise Time series of the dark signal show a noise on the dark signal level which seems to be corre lated to the voltage controlling the Peltier coolers on the Focal Plane Assembly R9 The correction is calculated as the Peltier output multiplied by a scaling factor This scaling factor is in fact dependent on the history of cooler switches and thus should ideally be calculated anew for each orbit using all measurements in that orbit However this conflicts with the GDP architec tural requirement that each ground pixel shall be independently processable Therefore the scaling factor from one typical orbit is used the variation with orbit is
26. ation data the PPG detector pixel to pixel gain as the ratio of the measured LED spectrum and the smoothed spectrum smoothing using a triangular convolution kernel of 5 pixels basewidth Apply PPG OP E Applies the calibration constants for PPG to the measurement spectrum Calculate Spectral Calibration Parameters OP All algorithms listed above are applied to detector signals without need of knowledge of the pre cise wavelength of each pixel This changes for the algorithms which follow which need wave length information in order to use the correct calibration constants The algorithms described here provide this wavelength calibration using information from the internal spectral lamp and from the Fraunhofer line structure in the solar calibration spectra GOME Extraction SUM 4 15 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 Apply Spectral Calibration Parameters OP E Calculates for each detector pixel its wavelength nm from the spectral calibration parameters Determine PMD out of band Straylight OP From solar calibration measurements the measured PMD signal is compared to the expected sig nal calculated from the channel array intensity and the ratio of the two is stored as so called Q factor in the calibration database Apply Straylight Correction OP E Uses pre flight straylight characteristics uniform straylight fraction and ghost intensites widths to calculate the straylight spectrum and
27. ault extraction contains header information which calibrations were applied the physical units of the data orbit number orbit propagation data the calibrated Sun Reference Spectrum and for each ground pixel of Earth shine data Sun normalised PMD values and for each spectral Band the spectral radiances and their errors The detailed format description can be found in the Product Specification Document PSD which is contained on the CD ROM or on the ftp server as acrobat pdf file docs PSD pdf 322 Sun Moon Observations Apart from the Earth shine observations obtained during normal scanning mode there are at reg ular intervals also observations made of the Sun and the Moon These observations can be extracted from the data product using the s or m flags Solar observations are extracted using the command gdp01_ex s input_file output_file where input_file isaLevel 1 product file The ASCII output with calibrated irradiances will be written to file output_file sll where the extension s11 denotes Sun shine Level 1 data Lunar observations are extracted using the command gdp01_ex m input_file output_file where input_file isa Level 1 product file The ASCII output with calibrated radiances will be written to file output_file ml1 where the extension mll denotes Moon shine Level 1 data Note that the calibrated radiance in this case is not corrected for the slit filling factor which depends on the Moon s phase and the position
28. ay occur in the science measurement are covered The dark spectra are included in a Calibration Database as function of integration time pattern For processing the most recent applicable dark spectrum from the database is used Spectral calibration is performed at regular intervals typically monthly using the sharp emission lines of an internal Pt Cr Ne hollow cathode lamp the SLS For each channel a 3 order in channels 1 2 or 4 order in channels 3 4 polynomial is fitted to the spectral line positions to obtain wavelength as a function of pixel number Since the SLS shows signs of degradation with possible failure an option is beinig implemented to obtain a wavelength calibration from Solar observations using the Fraunhofer lines in the Sun GOME Extraction SUM 27 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 The wavelength calibration depends on the temperature of the optical bench Therefore the poly nomial calibration constants are stored in the Calibration Database as function of temperature in 0 1 K intervals of temperature For processing the most recent applicable coefficients from the database are used The typical temperature variation over one orbit is 1 K the instrument starts heating up as soon as it enters sunlight and it starts to cool down when entering the dark part of the orbit Calibration measurements from Light Emitting Diodes LEDs which are placed close to the detectors are used for
29. ce version 2 30 of the software the correction method has been changed and a different parametrisation is used Versions of deg_par_file lower than 2 0 are not compatible with the new method but the newer file still contains the old parameters so that older versions of the software still run The f option corrects for a seasonal variation in the Sun spectra due to an asymmetry in the BSDF of the diffuser using a correction which is smooth in wavelength This is the default correction to be applied As experimental option which is currently not recommended one can instead use the same syntax with u instead of f since software version 2 30 This will apply a correction which contains spectral features on the BSDF for channels 1 3 and 4 3 3 5 Options to perform omit calibration steps e omit cross correlation of solar spectrum gdp01_ex k input_file output_file This option is implemented from version 2 20 of the software As default the spectral calibra tion of the SMR is from version 2 20 onwards corrected in the extraction software by cross correlation to a theoretical solar spectrum see G3 for detailed description This time con suming step can be switched off using the k switch if a high spectral accuracy is not needed apply only selected calibration steps These options shall not be used by the normal user of GOME level 1 data Apply the options in this section only if you are an advanced user who wants to study s
30. chronised with detector pixel 1 a correc tion can be applied to re normalise all spectral intensities to the intensity measured at the inte gration time of pixel of the arrays R8 this is also the time for which the geolocation is defined note pixel number 1024 is read 93 75 ms before pixel 1 Note that this option adjusts only the absolute radiance levels it cannot correct for the spectral changes in absorption line depth or Ring effect with changing ground scene Also it can cor rect only a linear trend The scientific usefulness of this option is therefore a matter of debate GOME Extraction SUM Rs Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 e correct degradation in radiance response gdp01_ex e deg_par_file input_file output_file This option is implemented from version 2 00 of the software A description of this correction together with the most recent file with degradation parameters deg_par_file will be put on the DLR DFD WWW server under http wdc dlr de sensors gome follow the link to Data describing Instrument Degradation and Degradation files It is intended to update the file every 3 months correct BSDF asymmetry gdp01_ex f deg_par_file input_file output_file This option has been implemented from version 2 00 of the software The parameters for cor rection are on the file with degradation parameters deg_par_file see above under cor rect degradation in radiance response Sin
31. d directly and is connected instead with zero gradient to the point halfway between and A and a parabolic connection continuous in value and gradient is made from the halfway point to A GOME Extraction SUM Doc No ER SUM DLR GO 0045 38 Issue 3 Date 14 12 04 DLR The actually simulated polarisation shows a bit more complex behaviour e g some residual struc ture of the O3 Huggins bands is visible in the polarisation due to different airmasses in and out side the absorption peaks with different depolarisation by aerosols but the approximation above was found accurate to better than 3 compared to the simulated polarisation values Once the interpolated fractional polarisation values have been calculated the polarisation correc tion factor is given by 1 1 c ae 1 n p n where subscript i denotes detector pixel The errors on the polarisation correction factor are derived from the errors on the seven p values on the Level 1 product and from the errors on the calibration keydata for n Errors due to the interpolation of polarisation values is not taken into account The errors are added to the error on absolute radiometric response not to the error on relative radiometric pixel to pixel response Directions for use This calibration should always be applied in the Extractor if absolutely calibrated radiances are required If a relative calibration is sufficient this calibration is not required However n
32. e 3 DLR Date 14 12 04 The fractional polarisation values p give the fraction of light polarised along the direction of the instrument slit this is the quantity relevant to GOME calibration The relation between p and more widely used atmospheric parameters such as the Degree of Polarisation P and the polarisa tion angle x defined w r t the plane through Nadir instrument and viewing direction is given by p 0 5 1 P cos 2 Before these polarisation values can be used for polarisation correction of the measured intensity they must be interpolated to the wavelengths of all detector pixels The interpolation scheme is based on simulations of the Degree of Polarisation P as function of wavelength But since the algorithm assumes a constant value for the polarisation angle X as function of wavelength the same interpolation may be made on p as well Radiative transfer model calculations including polarisation have been performed by P Stammes KN MI to study the polarisation behaviour in the UV for application to the GOME instrument R5 R7 These simulations show that up to wavelength A 300 nm the degree of polarisation P A is virtually a constant so there we set fractional polarisation equal to the theoretical single scatter value at A P A then changes rapidly between A and a value A 325 nm then shows smooth behaviour into the visible ca 325 400 nm Thereafter P A is a smooth slowly varying function of
33. e see Section 3 3 1 under option w The atmospheric polarisation has been retrieved during the operational Level 0 to 1 processing The Level 1 product contains for each ground pixel in the Pixel specific Calibration Data struc ture PCD see Section 3 3 1 or G4 a small array of fractional polarisation values at certain well spaced wavelengths across the GOME spectral range Up to 7 points can be specified Three points come from the comparison of channel array signals with broad band PMD signals in Figure 2 these are the filled circles at points P with index 1 2 3 the corresponding wave lengths A A A are approximately 360 nm 500 nm and 700 nm depending on the exact spectral shape of the input spectrum Three more points come from the channel overlap regions in Fig ure 2 these are the open circles at points P with index 4 5 6 for overlaps between channels 1 2 2 3 3 4 the corresponding wavelengths Ay As Ag are at 312 5 400 600 nm approximately though these appear to be largely unuseable for various reasons in which case they are set to a number gt 99 at wavelength 1 in the PCD The last theoretical point labelled A P in the figure comes from the Rayleigh single scatter model simulation of polarisation in the Ultra Vio let I Il II IV channel Wavelength nm Figure 2 Wavelength interplation of polarisation values GOME Extraction SUM 4 E Doc No ER SUM DLR GO 0045 Issu
34. e of this document is to provide an extended operation manual for these extraction software tools This follows a request from several users who found the instructions for use as delivered on the GOME product CD ROM too concise This document does not contain a product format description Fot that the user is referred to the Product Specification Document G4 which is also present on the product CD ROM GOME Extraction SUM Doc No ER SUM DLR GO 0045 8 Issue 3 Date 14 12 04 DLR 1 2 Document Status and History The draft version of this document has been prepared in April 1999 following user requests for more extended information on the Level 0 to 1 Extraction software The document is based on the instructions for use on the product CD ROM on the on line help of the software itself and on the algortihm description G3 This draft version of the document describes the status of the extrac tion software up to version 1 50 The first version of the document describes the status of the Level 0 to 1 Extraction software up to version 2 00 In addition a description is added of the Level 1 to 2 Extraction software version 2 70 The second version of the document describes the status of the Level 0 to 1 Extraction software up to version 2 20 the Level 1 to 2 Extraction software remains unchanged The third version of the document describes the status of the Level 0 to 1 Extraction software up to version 2 30 and the status of t
35. effect on the Solar Reference The PMD values of the Sun Reference are corrected for Leakage only If no calibration is to be applied i e extract level O data use c_filter For the real hackers amongst you note that the calibrations are always executed on the Level 1 product For example it is not possible to perform certain calibration steps using the extractor then perform a calibration on your own on the ASCII file and then run the extractor on the ASCII file to perform more calibration steps Therefore if you would wish to change calibrations e g to subtract a different dark signal background you would have to overwrite the cor responding fields in the Fixed Calibration Data structure see G4 on the Level 1 product For a certain ground pixel the pointers to the fields actually used there may be more fields depending on temperature etc can be found from examining the pcd file See option w in Section 3 3 1 An overview of calibration steps follows below the calibration steps are performed in the sequence as given For more information on the background of the calibration the reader is referred to Section 4 e Leakage The measured level 0 spectra are corrected for the dark signal by subtraction of a Dark Spec trum which for each detector pixel is a combination of readout fixed pattern noise and leak age current See Section 4 2 e FPA Correct channel 1 for cross talk from the Peltier coole
36. efore the start of Earthshine data Usually this d option is used in connection with the p option Section 3 3 3 to limit the amount of output files GOME Extraction SUM e Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 e write a separate file for each ground pixel without Solar Reference Spectrum gdp01_ex n input_file output_file This option is similar to the d option i e splits the output for each ground pixel in a separate file output_file_nnnn ell where nnnn is the number of the ground pixel but the Solar Reference Spectrum is omitted on each file Instead the Solar Reference Spectrum is written on a separate file output_file sun e write the output in HDF format gdp01_ex h input_file output_file This option is to be implemented in a future version of the software 3 3 3 Options to select a subset of data e extract only selected spectral bands gdp01_ex b b_filter input_file output_file This provides an option to extract any combination of spectral bands In addition to the regular 6 bands la 1b 2a 2b 3 4 also the blind band the first 49 pixels of channel 1 and the straylight bands 50 pixels directly before band la and 2a and directly after band 1b can be extracted Selection is via a character string b_filter which consits of 10 flags which are either y to enable selection or n to disable selection The sequence is for bands la 1b 2a 2b 3 4 blind la stray la stray
37. expected to be 10 The value of the Peltier output is for each readout calculated as a weighted average of the last 61 Peltier voltage readings This weighting is equivalent to a filtering of the signal The filter coeffi cients and the scaling factor are present in the FCD record on the Level 1b product The magnitude of the effect scales with integration time Since it was regarded negligible for the 1 5 second integration times the correction is only applied to band la Directions for use This calibration should always be applied in the Extractor Limitations The FPA correction is known to be not perfect the typical reduction in FPA noise is a factor of 5 but dark signal errors up to 7 BU remain for 12 seconds integration time the typical uncertainty is 3 BU GOME Extraction SUM Doc No ER SUM DLR GO 0045 30 Issue 3 Date 14 12 04 DLR 4 4 Apply PPG The PPG describes the differences in quantum efficiency between neighbouring detector pixels It is in principle contained in the instrument resonse function but it is artificially separated from the instrument response function during calibration for the following reason If the wavelength cali bration of the instrument changes due to whatever mechanism the radiometric response function of the instrument seen by a single detector pixel changes with it this depends on wavelength dependent transmission of lenses etc We have to re interpolate the radiometric response to
38. f the Third ERS Symposium on Space at the Service of our Environment ESA SP 414 Vol II 669 674 Florence 1997 K Bramsted M Weber J Burrows IFE TUP univ Bremen Radiance Jumps and PMDs in GOME GOME workshop 24 06 1996 C Caspar GOME Dark Signal Characterisation ER TN ESA GO 0473 draft issue 1 0 06 11 1995 Hegels E ERS 2 GOME Support Progress Report ER TN DLR GO 0044 Issue 1 29 1 1999 Aben I C Tanzi E Hegels and M Eisinger GDAQI GOME Data Quality Improve ment TN GDAQI 001SR 99 Midterm report Draft April 1999 HDFS5 Release 1 The National Center of Supercomputing Applications at the University of Illinois Urbana Champaign IL 1998 Slijkhuis S CHEOPS GOME Study on seasonal effects on the ERS 2 GOME Diffuser BSDF DLR report CH TN DLR GO 0001 Issue 1 May 2004 GOME project documentation and study reports G1 G2 G3 G4 G5 GOME Interim Science Report edited by T D Guyenne and C J Readings SP 1151 ESA publications Division ESTEC Noordwijk The Netherlands ISBN 92 9092 041 6 1993 GOME Users Manual ESA SP 1182 ESA ESTEC Noordwijk The Netherlands 1996 GOME Level 0 to 1b Algorithm Description ER TN DLR GO 0022 Issue Rev 5 B July 2002 Product Specification Document of the GOME data processor ER TN DLR GO 0016 Iss Rev 4 B December 2004 the latest version is included on the product CD ROM GOME Level 1 to 2 Algorithm Theoretical Basis Document
39. he Level 1 to 2 Extraction software up to version 4 00 1 3 Document Overview The document is divided into 2 parts The first and major part deals with the Level 1 Data the sec ond part deals with the Level 2 Data Part I Section 2 gives a overview of the program concept and of the calibration steps used in Level 0 to 1 processing and extracting Section 3 is the main body of this document describing the use of the Extraction software Installation and version control is described in Section 3 1 the standard use of the software is described in Section 3 2 A large number of options exists to configure the output or execution of the extractor These options are described in detail in Section 3 3 Background information on the instrument calibration used in the Level O to 1 processing is pro vided in Section 4 Although this information may be of interest to the normal user of GOME Level data it is not necessary for the normal operation of the software General calibration considerations are presented in Section 4 1 Subsequent sections provide detailed background information on the various calibration procedures A short overview of the software version history is given in the appendix Part II Section 6 describes the use of the Extraction software for Level 2 data Since this is basically a binary to ASCII format converter with a few options to obtain product information or to select a subset of data the description can be
40. held brief Installation and version control is described in Section 6 1 the standard use of the software is described in Section 6 2 A few of options exists to configure the output or execution of the extractor These options are described in detail in Section 6 3 A short overview of the software version history is given in the appendix GOME Extraction SUM Doc No ER SUM DLR GO 0045 Issue 3 Date 14 12 04 DLR 14 Abbreviations and Acronyms ADC Analogue to Digital Converter ADD Architectural Design Document BSDF Bi directional Scattering Distribution Function BU Binary Unit DFD Deutsches Fernerkundungsdatenzentrum DLR Deutsches Zentrum f r Luft und Raumfahrt e V DOAS Differential Optical Absorption Spectroscopy D PAC German Processing and Archiving Centre ERS European Remote Sensing Satellite ESTEC European Space Centre of Technology FCD Fixed Calibration Data structure Level 1 product FOV Field of View FPA Focal Plane Assembly FPN Fixed Pattern Noise FWHM Full Width Half Maximum GDF General Distribution Function GDP GOME Data Processor GOME Global Ozone Monitoring Experiment GPPR Ground Processing Performance Requirement GSAG GOME Scientific Advisory Group ICU Instrument Control Unit IFE Institut f r Fernerkundung der Universit t Bremen IFOV Instantaneous Field of View ILOS Instantaneous Line of Sight VO DD Input Output Data Definition IT Integration Time KNMI Koninklijk Nederlands Meteor
41. his sequence differ from each other because they are measured with slightly different elevation and azimuth angles see also Section 3 3 5 The Sun Reference spectrum on the Level 1 product is always obtained from the measurements on a previous orbit not from the solar measurements on the product itself Directions for use This calibration shall be used in the Extractor whenever a conversion from radiance to irradiance 1s required Limitations The calibration can only be performed on Sun measurements i e in combination with the s option in the Extractor It is already applied to the Sun Reference spectrum on the Level 1 prod uct this cannot be undone The BSDF calibration does not take into account the azimuth asymmetry of the BSDF which was found after analysing 4 years of data The correction for azimuth asymmetry is possible from soft ware version 2 00 in the Extractor using option f see Sections 3 3 4 4 10 GOME Extraction SUM 4 35 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 4 8 Apply Radiance Response The application of the radiance response is simply a division of the observational signals after dark current PPG and straylight correction by the radiance response function This function is provided in the FCD structure of the Level 1 product per detector pixel one value is specified for each scan mirror angle and for each channel characterised by the same index as for the polarisa tion sensitiv
42. ice deposits on the detector see e g R4 For level 1 to 2 processing using the DOAS spectral analysis method this effect is not significant since it divides out of the ratio of earth shine to solar spectra GOME Extraction SUM Doc No ER SUM DLR GO 0045 28 A Issue 3 Date 14 12 04 DLR 4 2 Apply Dark Signal The measured spectra are corrected for the dark signal by subtraction of a Dark Spectrum which for each detector pixel is a combination of readout fixed pattern noise FPN and leakage current LC this component increases linearly with detector integration time In the present calibration approach these two components are not separated they used to be in the prototype version of the processor the Dark Current amp Noise Factor on the pcd file still refers to this these data are obsolete The dark spectra measured at each night side of the orbit are stored in the calibration database as function of temperature The Level 0 to 1b processing takes the most recent values from the data base for the temperatures encountered in the orbit and writes these in the FCD record on the Level 0 to 1b product Each measurement on the product has an index which points to the applica ble temperature The temperatures themselves are not visible in the product see also option w in Section 3 3 1 For the PMDs the PMD zero offset is subtracted These offsets are written on the FCD record just above the channel detector d
43. iffuser That correction is made by dividing the detector signals of the measured solar spectrum by the BSDF function of the diffuser more precisely what we calculate as BSDF is actually the reflectivity of the calibration unit which corresponds to one viewing direction out of the BSDF s 2 dim reflectivity space The BSDF is characterised by 11 coefficients cg Cjo Which are present in the FCD structure of the Level 1 product It is calculated from these coefficients as a 7th degree polynomial in wave length with correction terms for the Azimuth and Elevation angles Az Ele of the Sun beam onto the diffuser 7 2 A W k 0 where the angles are given in degree and the wavelength A in nm The angles are for each solar measurement given in the Sun Moon Specific Calibration Data structure on the Level 1 product see G4 The PMDs are corrected for BSDF by dividing the PMD signals by the BSDF evaluated at the effective PMD wavelength specified on the product together with the signals Sun Reference spectrum The Sun Reference spectrum which is included in the FCD structure of the Level 1 product is calculated as the average of all spectra in a solar calibration measurement sequence which are fully inside the field of view of the instrument s irradiance port 1 5 degree in elevation cor rected for the instrument s irradiance response a combination of the radiance response function and BSDF the various spectra of t
44. ions we list the various options with program call using their flag their parameters if any and we provide a short description of the actions taken 6 3 1 Options to obtain product information e general product information gdp12_ex i input_file This provides screen output of data from the Specific Product Header see G4 such as GDP Software Version Operational processor not Extractor Static Parameter Version processor settings and initialisation Time Correlation Information e g orbit number UTC plus as summary of measurements on the file the number of Earthshine Measurement Pixels on the product a complete orbit has in the order of 2000 pixels e geolocation information gdp12_ex g input_file This provides screen output of the geolocation For each ground pixel in the input file it writes the Subset Counter from the Earthshine Ground Pixel structure see G4 this counter is 0 1 2 3 for East Nadir West Backscan pixels and the 4 corner and centre coordinates lati tude and longitude from the Geolocation Record 2 structure see G4 e selected product information gdp12_ex q input_file output_file This prints selected data from the product per pixel geolocation and trace gas columns and their errors into a file output_file qa2 This data is used by DLR DFD internally for quality assurance purposes and is not further discussed here GOME Extraction SUM 49 Doc No ER SUM D
45. ity n see Section 3 3 1 under option w Before the division by the radiance response function is made it is interpolated from the wave length grid of the SMR to the wavelength grid of the Earth shine spectrum The radiance response function on the FCD file is in fact composed of 3 calibration Keydata func tions the radiance response for nadir RadResp the scan mirror dependence w r t nadir SM_DEP and a function f2 which was meant to describe the temperature dependence in the overlap region for the latter now always the function for 6 5 C is used Optionally see Section 3 3 4 under option a the Sun normalised radiance or Reflectivity can be calculated from the calibrated radiance S_cal as Ry S_cal SMR Before the division by the Sun Reference spectrum SMR is made the SMR is interpolated to the wavelength grid of the Earth shine spectrum Directions for use This calibration should always be applied in the Extractor if absolutely calibrated radiances are required If a relative calibration is sufficient this calibration is not required However note that in wavelength regions where the radiance response function has sharp features not correcting for radiance response implies that the uncalibrated spectrum has sharp features as well which may disturb e g a DOAS fit Limitations It is known that especially in channel 3 the on ground calibration determination is not very accu rate due to ou
46. l 1 product 1 60 is different to the EE version of the extraction programme 2 00 KKEKKKKKKKKKKK KK KKK KKK KK KKK KK KK KKK KKK KKK AK A AI FF A AK TA AI AK A AI TFA AK A AI KH A This warning is only important if the format of the Level 2 product has changed If there was no format change between the two versions this warning can be ignored See the release notes for format changes Up to the moment of writing version 2 70 no format change has taken place since version 1 40 of March 1996 GOME Extraction SUM 47 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 6 2 Standard Level 2 Extraction The default command for obtaining extracted Level 2 ASCII data is gdpl2_ex input_file output_file where input_file isa Level 2 product file The ASCII output will be written to file output_file al2 where the extension al2 denotes atmospheric Level 2 data Additionally a file gdp12_ex err is generated which describes the errors if any which occurred during program execution this file is appended not newly created after each run of the program The input Level 2 product files are stored on the CD ROM in the directory products or on the ftp server ftp ops de envisat esa int see file readme_1st txt for fur ther navigation on this site They have as default the extension 1v2 The output ASCII file of this default extraction contains header information extractor version orbit number orbit propagation da
47. level O to 1 The Level 1 Product contains the raw detector signals binary ADC units of the sci ence measurements plus calibration constants The end user has to run an Extraction program which applies the calibration constants to the data this inflates the Level 1 Product to a much larger Level 1 data set containing fully calibrated data the difference is that the Product contains calibration data plus raw signals as 2 byte integers the extracted data contain calibrated signals as 4 byte floats plus three 4 byte floats for wavelength accuracy error and precision error Apart from product size considerations an additional advantage of this procedure is that the user can optionally omit certain calibrations to investigate their influence or perform the calibration himself and that by optionally filtering out only a subset of the data i e geographical coverage or wavelength range the final level 1 data may be kept as small as possible The Extraction program with its different calibration options is the subject of this documentation GOME Extraction SUM Doc No ER SUM DLR GO 0045 as A Issue 3 Date 14 12 04 DLR 2 2 Algorithms Overview In this and the following section summaries of algorithms are given without references these will be given in the detailed background sections For the sake of completeness the summaries given include both the Level 0 to 1b operational processing i e generation of the Level 0 to 1b product
48. n determined which gives wavelength nm as function of pixel number The calibration parameters for spectral calibration are the coefficients of this poly nomial The order of the polynomial is not larger than 4 Calibration polynomial coefficients are derived for different optical bench temperatures for inter vals of 0 1 K The polynomial coefficients are stored in the calibration database The data are updated in regular baseline monthly intervals following a spectral line lamp calibration measurement The Level 0 to 1b processing takes the most recent values from the database for the temperatures encoun tered in the orbit and stores these in the FCD record on the Level 0 to 1b product Each measure ment on the product has an index which points to the applicable temperature The temperatures themselves are not visible in the product In software version 2 00 a fixed correction is added to the polynomial coefficients This correction term a constant for channels 1 2 and a 4th order polynomial for channels 3 4 describes the wavelength shift averaged over the first 4 years in orbit between the GOME solar spectrum cal ibrated with the lamp and corrected for Doppler shift and the high resolution solar spectrum of Kurucz This correction is applied in the operational processing to the spectral coefficients on the Level 1b product and does not influence the Extractor In software version 2 20 a similar correction is now calculated in the
49. nfiguration i e start and end pixel of the 6 science bands and the 4 straylight bands PPG parameters see Section 4 4 dark signal parameters and spectral calibration parameters both are indexed as function of temperature the relation of index to temperature is not visible in the Level product but the index to be used for each measurement is visible in the pcd file Sun Reference spectrum units of BU sr s and corresponding PMD values radiance response function and polarisation sensitivity n on the wavelength grid of the SMR indexed as function of channel and as function of scan angle the relation of index to scan angle is not directly visible in the Level 1 product but for the Earthshine measurements they can be recovered by comparing the viewing angle of the ground pixel centre to the index in the Spectral Band Record structure see G4 however this structure is not written to any ASCII output by the extractor For more detailed inforemation see the Product Specification Document G4 3 3 2 Options to steer the output e write a separate file for each ground pixel gdp01_ex d input_file output_file This splits divides the output for each ground pixel in a separate file output_file_nnnn ell where nnnn is the number of the ground pixel e g 0059 for ground pixel 59 in the level 1 product Each file has the formatting of the standard ASCII output in particular each file has its own Solar Reference Spectrum b
50. nwards also applied to PMD 1 Directions for use It is recommended to apply the e calibration in the Extractor see Section 3 3 4 if a Radiance Response calibration is also applied Section 4 8 For solar spectra if also a BSDF correction is to be applied Section 4 7 it is recommended to use both the e and the f option The use of the experimental u option is not recommended Limitations The degradation is modelled with a polynomial function only Shifts in Etalon are not corrected An additional file with Degradation Parameters is needed see Section 3 3 4 under option e Software versions from 2 30 need a degradation file version of 2 00 or higher For Level 1 products generated before version 2 00 no correction is possible on the fractional polarisation values p from PMD 1 The PMD reflectances are not corrected The influence of degradation is relatively small and zero for unpolarised input since a normalisation with Sun observations is made GOME Extraction SUM Doc No ER SUM DLR GO 0045 40 Issue 3 Date 14 12 04 DLR 4 11 Calculation of Errors A distinction is made between accuracy errors and precision errors e The accuracy of a measurement is a measure of how close its result is to the true value e The precision of a measurement is a measure of how well its result has been determined with out reference to its agreement with the true value Precision is therefore a measure of the repro ducibility of
51. oblem since the total intensity in band 1a is small compared to that in band 1b The error on the signal due to errors in straylight subtraction is mainly a precision error This error is neglected in the precision error on the Level 1b data see Section 4 11 The so called PMD straylight the factor Q in Eq 45 in G3 is not corrected neither for earth shine PMDs nor for the PMDs of the Sun Reference spectrum This is a multiplicative term which falls out of the equation for the earthshine PMD reflectivities the only quantity used in level 1 to 2 processing PMD straylight has been taken into account for the calculation of polarisation frac tion in the operational Level 0 to 1 processing but is nowhere explicitly visible on the Level 1 product Directions for use It is recommended to apply this calibration in the Extractor Limitations The Calibration Keydata for straylight are probably not more accurate than 10 this uncertainty is not well characterised GOME Extraction SUM Doc No ER SUM DLR GO 0045 Hd A Issue 3 Date 14 12 04 DLR 4 7 Apply BSDF GOME takes at regular intervals typically once per day solar calibration spectra in the sunrise part of the orbit An on board diffuser can be switched into the lightpath during sun calibration to scatter the collimated solar light irradiance into a diffuse radiance beam Obviously the solar spectrum thus recorded must be corrected for the characteristics of this d
52. ologisch Instituut LED Light Emitting Diode NRT Near Real Time PCA Polarisation Correction Algorithm PCD Pixel specific Calibration Data structure Level 1 product PMD Polarisation Measurement Device PSD Product Specification Document contains Level 1 2 data formats PPG Pixel to Pixel Gain SAA Southern Atlantic Anomaly SAO Smithsonian Astrophysical Observatory SBT Satellite Binary Time SCIAMACHY Scanning Imaging Absorption Spectrometer for Atmospheric Chartography SES Spectral Light Source SMR Sun Mean Reference SRD Software Requirements Document SRON Space Research Organisation of The Netherlands SVD Singular Value Decomposition SZA Sun Zenith Angle TOA Top of Atmosphere TPD Technisch Physische Dienst UV ultra violet VIS Visible GOME Extraction SUM Doc No ER SUM DLR GO 0045 10 A Issue 3 Date 14 12 04 DLR 1 5 Glossary Band spectral one of 6 or 10 spectral bands referring to parts of an array detector band la and 1b cover the short wavelength and long wavelength part of channel respectively band 2a and 2b cover the short wavelength and long wavelength part of channel 2 band 3 and 4 are identical to channel 3 and 4 In addition there are 4 straylight bands two shortwave of band la one longwave of band Ib and one shortwave of band 2a These straylight bands are not part of the standard level 1 extracted data but they are available on the level 1 product Channel one of the 4
53. option o Reorganising of the Extraction software to improve performance PE EN Release 1 50Fri Jan 23 14 15 44 MET 1998 NA o Put radiance jumps correction as extra option j EN Release 1 40 Mon Oct 13 10 22 20 MET DST 1997 o New options in the Extraction software r extract ground pixels between corners upper left and lower right n creating one file without sun spectrum for each ground pixel Extraction software runs also under Linux operating system EN Release 1 35Mon Nov 13 15 33 32 MET DST 1997 Ae e O x EN Release 1 30Mon Jun 23 14 59 26 MET DST 1997 NA ES o Changes in the Extraction software Not apply FPA correction when less than 61 ground pixels are available New calibration option J for radiance jump correction J Extraction software runs also under ALFA computers Jk Release 1 20Tue Oct 15 16 47 00 MET DST 1996 NA k k GOME Extraction SUM 43 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 AN Release 1 10Fri Aug 23 11 52 03 MET DST 1996 o Changes in the Extraction software Saturation check done only when the co adding mode is not active New option a compute albedos instead of radiances ER Release 1 00Thu
54. orrelation on the solar spectrum may be switched off using the k switch if a high spectral accuracy is not needed GOME Extraction SUM Doc No ER SUM DLR GO 0045 30 Issue 3 Date 14 12 04 DLR Limitations Wavelength calibration errors are typically 0 05 of a detector pixel At the short wavelength side of channel 3 the error may be larger due to the scarcety of lamp lines there Mind the solar dop pler shift mentioned above GOME Extraction SUM 4 3 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 4 6 Apply Straylight Correction This comprises subtracting the calculated amount of straylight per pixel from the measurement The calibration keydata contain for each channel uniform straylight factors straylight scales with total intensity in the channel and spectral ghost characteristics The latter consist of the symme try asymmetry pixel the ghost intensity and the ghost spectral width These data are present in the FCD record on the Level 1b product as direct copy of the calibration keydata The straylight is calculated on a per channel basis In the bands where at a certain readout not the whole channel is available the straylight correction is only performed using the signals from the available datapoints missing data are set to 0 no attempt is made to estimate the intensity in the missing part of the channel This can currently only happen for band 1b when band 1a is still integrating It is not considered a pr
55. ote Sensing Satellite ERS 2 GOME is a nadir viewing spectrometer covering the range 240 790 nm in 4 spectral channels In its normal Earth observation mode it scans across track in three steps The field of view of each step may be varied in size from 40 km x 40 km to 320 km x 40 km The mode with the largest foot print three steps with a total swath width of 960 km x 40 km provides global coverage at the equator within 3 days this is the default mode of operation In addition to the on line components at the ground stations the GOME Data Processor GDP system R6 is the operational off line ground segment for GOME The GDP was developed and implemented at DFD with the help of several scientific institutions G2 and became operational in July 1996 It incorporates a Level 0 to 1 processing chain the complete GOME data archive a DOAS 03 total column retrieval process Level 1 to 2 and an image processing chain for the generation of higher level products The Level 1 and Level 2 data products are generated by DFD on behalf of ESA The Level 1 products calibrated spectra are generated in a sort of compressed form The end user has to apply an extraction software tool to obtain the final Level 1 data The Level 2 products trace gas columns are generated in a binary format The extraction soft ware tool for Level 2 data does not much more than to convert the binary product into an ASCII readable data file The purpos
56. ote that in wavelength regions where n has sharp features not correcting for polarisation may imply that the uncalibrated spectrum has sharp features as well which may disturb e g a DOAS fit Limitations It is known that especially in channel 3 the on ground calibration determination of n was not very accurate due to outgassing problems with a dichroic leading to problems with polarisation cor rection This problem is still present for Keydata version 8 work for improvement is in progress Errors in the determination of p or in the wavelength interpolation of p may lead to undercor rected spectral features in the spectrum In Keydata version 8 noise can be seen on the calibration parameter n This noise will be aliased into the calibrated radiance and could adversely influence the pixel to pixel precision of the measurements for high values of the atmospheric polarisation No overlap point at A4 As Ag is currently being used in the wavelength interpolation as their reliability was judged to be too low due to the radiance jump problem see Section 3 3 4 under option j GOME Extraction SUM 4 39 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 4 10 Apply Degradation Parameters or azimuth asymmetry of the BSDF In software version 2 00 an option e has been added to the Extractor to correct the Radiance Response function of the instrument for degradation of the optical components R11 The reason that thi
57. pecial effects e g check of dark signal and straylight subtraction in the straylight bands or if you have a good reason to omit certain calibrations such as polarisation correction or radiance response cal ibration The generic command for applying calibration is gdp01_ex c c_filter input_file output_file where c_filter isa character string with for each calibration step a single key character L Leakage A FPA F Fixed S Straylight N Normalize B BSDF P Polarization I Inten sity U Unit_Conversion details below The order of these calibrations shall not be changed GOME Extraction SUM Doc No ER SUM DLR GO 0045 24 Issue 3 Date 14 12 04 DLR The default extraction has c_filter LAFSNPIU for Earthshine data c_filter LAFSNBIU for Sunshine data and c_filter LAFSNIU for Moonshine data For the 4 blind and straylight bands only the calibrations LAFSN can be succesfully applied The other calibrations yield zero signals because the corresponding keydata are missing The Solar Reference spectrum on the Level 1 product has already the calibrations LF SNB applied the A calibration is omitted because of the small integration time which makes the FPA compo nent small The extractor always adds the calibration I This implies that even when Earthshine data are extracted with e g c_filter L the accompanying Solar Reference spectrum will still have c_filter LFSNBI Only the U calibration has
58. ption of the actions taken For more information on the calibration steps the reader is referred to Section 4 3 3 1 Options to obtain product information e general product information gdp01_ex i input_file This provides screen output of data from the Specific Product Header see G4 such as GDP Software Version Operational processor not Extractor Pre flight Calibration Key Data Version see also Section 4 1 Time Correlation Information e g orbit number UTC plus a summary of measurements on the file such as Time coverage of measurements Number of Earthshine Measurement Pixels Number of Sun Measurements Number of Moon Measurements the latter two are usually 0 i e no such measurementes performed in this orbit e geolocation information gdp01_ex g input_file This provides screen output of the geolocation For each ground pixel in the input file it writes the Subset Counter from the Earthshine Ground Pixel structure see G4 this counter is 0 1 2 3 for East Nadir West Backscan pixels and the 4 corner and centre coordinates lati tude and longitude from the Geolocation Record 1 structure see G4 e selected product information gdp12_ex q input_ file output_file GOME Extraction SUM Doc No ER SUM DLR GO 0045 20 A Issue 3 Date 14 12 04 DLR This prints selected data from the product into a file output_file gal This data is used by DLR DFD internally for qu
59. rates band la from 1b or 2a from 2b this can be set by macro command from the ground operator 1 6 R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 GOME Extraction SUM 4 11 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 Reference Documents Hiroshi Akima A new method of interpolation and smooth curve fitting based on local procedures J ACM Vol 17 4 1970 589 602 Data Reduction from Experimental Histograms W R Falk University of Manitoba Win nipeg Canada Nuclear Instruments and Methods in Physics Research 220 1984 473 478 William H Press et al Numerical Recipes in C Cambridge University Press 1994 Jochen Stutz and Ulrich Platt Problems in using diode arrays for open path DOAS meas urements of atmospheric species Institut fiir Umweltphysik Universit t Heidelberg P Stammes The seventh point polarisation algorithm Internal Report GOME and SCIAMACHY KNMI De Bilt 1994 D Loyola W Balzer B Aberle M Bittner K Kretschel E Mikusch H Muehle T Ruppert C Schmid S Slijkhuis R Spurr W Thomas T Wieland M Wolfmueller Ground segment for ERS 2 GOME sensor at the German D PAF Proceedings of the Third ERS Sympo sium on Space at the Service of our Environment ESA SP 414 Vol II 591 596 Flor ence 1997 P Stammes I Aben R B A Koelemeijer S Slijkhuis D M Stam GOME polarisation validation study Proceedings o
60. rs in the Focal Plane Assembly FPA See Section 4 3 e Fixed Correct signals for pixel to pixel fixed pattern variations in the instrument response See Sec tion 4 4 e Straylight Correct signals for straylight uniform and spectral ghosts components See Section 4 6 e Normalize Normalise the signals to 1 second detector integration time convert from BU to BU s by dividing the signal in each band by the integration time of that band PMD signals are not affected e BSDF Divide signals from Sun measurements by the BSDF of the diffuser to convert the radiance measured by the spectrometer into irradiance in this step the irradiance is not yet absolutely calibrated that occurs in the Intensity calibration below See Section 4 7 e Polarization Interpolate the polarisation fractions p up to 7 points from theory and from the measured polarisation points of PMDs and overlap regions if available to the wavelength of each detec GOME Extraction SUM 25 Doc No ER SUM DLR GO 0045 Issue 3 DLR Date 14 12 04 tor pixel Multiply the signal of each pixel with the polarisation correction factor derived from this interpolated p value and calibration key data for the instrument s polarisation sensitivity See Section 4 9 This applies to Earthshine data only Intensity Perform the absolute radiance calibration For earthshine and moonshine signals the conver sion is from BU s to Watt s cm sr nm for sunshine from BU sr s
61. rsion used to generate th SR RRE Level 1 product 1 00 is different to the AA version of the extraction programme 1 40 KKAKKAAKKAKAKKKKKAKAKAKAAAKAKAKAAKKAAKAKAKAKAAKAKAKAAKFAKAKAAKAAAKAAAAAKAAAAKAKAAAAAAKAA This warning is only important if the format of the Level 1 product has changed If there was no format change between the two versions this warning can be ignored See the release notes for format changes Up to the moment of writing version 2 20 no format change has taken place GOME Extraction SUM Doc No ER SUM DLR GO 0045 218 Issue 3 Date 14 12 04 DLR 32 Standard Level 1 Extraction 3 2 1 Earth shine Radiance The default command for obtaining extracted Level 1 data in the form of calibrated Earth shine radiance spectra is gdp01_ex input _ file output_ file where input_file is a Level 1 product file The ASCII output with calibrated radiances will be written to file output_file ell where the extension ell denotes Earth shine Level 1 data Additionally a file gdp01_ex err is generated which describes the errors if any which occurred during program execution this file is appended not newly created after each run of the program The input Level 1 product files are stored on the CD ROM in the directory products or on the ftp server ftp ops de envisat esa int see file readme_1st txt for fur ther navigation on this site They have as default the extension 1v1 The output ASCII file of this def
62. s is done in the Extractor and not in the operational processing is twofold no change in product format is needed at the cost of having an additional file with Degradation Parameters and users can reprocess a selection of their old data themselves which avoids operational reproc essing and the distribution of the whole archived dataset to all users For the same reasons product format reprocessing a separate option f corrects the solar spec tra for an azimuth asymmetry of the BSDF which was found after analysing almost 4 years of data This asymmetry is not really a degradation but a seasonal effect which can be seen as a dis turbance of the degradation curve R10 Since software version 2 30 an improved correction method is used which also corrects the sidelobes found on the BSDF R13 Like the previous correction this is a smooth function of wavelength As discussed in R13 there is also spectral structure visible on the BSDF which could be described for channels 1 3 and 4 This has experi mentally been implemented in the software option u instead of f The degradation of optical components does not only affect the Radiance Response function but also the calculation of the fractional polarisation values p see Section 4 9 For PMDs 2 and 3 this degradation was always more or less equalised by a self calibration on the Sun which was initially intended to correct straylight only This self calibration is from version 2 00 o
63. shine radiances are divided by the Solar Reference Spectrum after regridding of the Solar spectrum to the wavelengths of the Earthshine spectra Akima interoplation in wave legth Currently no correction is made for Doppler shift of the solar observation which is below 1 10th of a detector pixel but nevertheless neglecting this can introduce high frequency noise on the few percent level Therefore this option shall only be used in applications where no high spectral resolution is required e smooth out radiance jumps gdp01_ex j input_file output_file The detector arrays of GOME were designed for serial readout such that the last pixel of the array is read out 0 09375 seconds later than the first When GOME is scanning over an inho mogeneous ground scene e g broken clouds the variation of upwelling radiance with time is reflected in a variation of intensity depending on read out time the ground scene inhomogen ity is aliased in the spectrum this effect is denoted as spatial aliasing Such aliasing is often visible as a jump in radiance between two detectors e g the last pixels of channel 3 record the same wavelengths as the first pixels in channel 4 but at an integration time which is shifted by 0 09 seconds although they should record the same intensity over a homogeneous scene there may be a radiance jump if intensity varies with time Using information from the PMDs which are read out every 0 09375 seconds syn
64. ta and for each ground pixel of Earth shine data trace gas ver tical column values and their errors airmass factors cloud cover information and diagnostics The detailed format description can be found in the Product Specification Document PSD which is contained on the CD ROM as acrobat pdf file docs PSD pdf or in the corresponding location on the ftp server ftp ops de envisat esa int Note that with version 4 0 the format of both the Level 2 product file and the extracted Level 2 ASCII data has changed This applies e g to the new fields for DOAS diagnostics and cloud information in the Intermediate Results Record IRR structure GOME Extraction SUM Doc No ER SUM DLR GO 0045 48 Issue 3 Date 14 12 04 DLR 6 3 Special Options Several possibilities exists to configure the execution output of the Extractor In the following we will separate the options into two classes 1 options to obtain product information 2 options to select a subset of data Options from these two classes cannot be combined with each other Following the UNIX syntax the extraction sofware can be called with flags to implement these options The general syntax is given by gdp12_ex flagl param1_1 param1_2 flag2 param2_1 param2_2 flagn paramn_1 paramn_2 input_file output_file where input_file isa Level 2 product file default extension 1v2 and the output file gets an extension al2 In the following subsect
65. tgassing problems with a dichroic This problem is still present for Keydata version 8 work for improvement is in progress Errors due to the etalon effect R4 are not corrected This should not be a problem for sun nor malised spectra but an etalon structure with amplitudes of up to a few percent may be visible in radiance Earth Moon or irradiance Sun spectra In Keydata version 8 noise can be seen on the calibration parameter for radiance response This noise will be aliased into the calibrated radiance and could adversely influence the pixel to pixel precision of the measurements GOME Extraction SUM Doc No ER SUM DLR GO 0045 3G 4 Issue 3 Date 14 12 04 DLR 4 9 Apply polarisation correction GOME is a polarisation sensitive instrument The radiance response function described in the pre vious section calibrates the instrument assuming unpolarised input light Therfore a correction factor must be applied which describes the ratio instrument throughput for the actual input polarisation throughput for unpolarised light This correction needs the polarisation sensitivity of the instrument and a characterisation of the atmospheric polarisation The polarisation sensitivity of the instrument n called Eta_Omega on the fcd file is provided in the FCD structure of the Level 1 product per detector pixel one value is specified for each scan mirror angle characterised by the same index as for the Radiance Respons
66. to be modelled externally from the data processor by scientific analysis of time sequences of solar lunar or earth shine observations 4 1 1 Onground Calibration The onground calibration has been performed by TPD TNO The output of the on ground calibra tion relevant to the operational Level 0 to 1 processing is a data set containing the so called Cal ibration Key Data For various reasons the orginal Key Data set had to be adjusted from the on ground situation to the in flight situation mostly due to air vacuum wavelength shifts and outgassing of optical coat ings This has resulted in several updates of the Key Data set by TPD their most recent work resulting in Version 8 of the Key Data At the moment of writing a contract has been given out to SRON and DLR DFD to improve on this Key Data set using the time sequence of science measurements now available and to charac terise the time dependence of calibration constants in flight 4 1 2 Level 0 to 1 Processing of Calibration Constants Dark signal from the channel array detectors and from the PMDs is measured at each dark side of the orbit Dark signal has a fixed component the so called fixed pattern noise FPN as well as a component due to charge leaking which increases linearly with detector exposure time Therefore a series of dark spectra is taken at each night side of the orbit All integration time patterns i e combinations of IT for the different detectors which m
67. wavelength The interpolation of polarisation goes roughly as follows for details see G3 constant shortward of wavelength A 300 nm pre defined on the initialisation file a parametrisation see below between 2 and a pre defined wavelength 7 325 nm power law continuity in value and in gradient between A and A if P4 present else see below Akima interpolation between A and Az this is basically a spline interpolation but with constraints to reduce overshoot constant longward of Az The critical region is 300 325 nm An analytic function is required that mirrors the shape of sim ulated polarisation for all atmospheric scenarios This function must have a turning point at the theoretical point A7 with parabolic fall off behaviour near this point The function must also show exponential tail behaviour as tends towards A This role is fulfilled by the Generalised Distribu tion Function GDF constructed by R Spurr for the GOME calibration G3 w exp A A P P X P A er 1 exp A 2 B where P w B are parameters that characterise the GDF they must be found to fit the given interpolation points If P4 is missing which ususally is the case then there are more parameters than equations In this case the GDF is connected directly to A with continuity of the gradient derived from Akima interpolation However if that gradient turns out to be positive then the GDF cannot be connecte

GOME Data Processor Extraction Software User's Manual

Contents

Download Pdf Manuals

Related Search

Related Contents