Home

MARSCHALS Level 2 Software User and Reference Manual

1. amp Z i i nz 1 1 altitude km amp fdird i i 1 nz downward direct flux w m2 um amp fdifd i i 1 nz downward diffuse flux w m2 um amp flxdn i i 1 nz total downward flux w m2 um amp flxup i i 1 nz total upward flux w m2 um enddo one output record per run integrated over wavelength output quantities are integrations by trapezoid rule WLINF WLSUP FFEW TOPDN TOPUP TOPDIR BOTDN BOTUP BOTDIR WLINF lower wavelength limit microns WLSUP upper wavelength limit microns FFEW filter function equivalent width microns TOPDN total downward flux at ZOUT 2 km w m2 TOPUP total upward flux at ZOUT 2 km w m2 TOPDIR direct downward flux at ZOUT 2 km w m2 BOTDN total downward flux at ZOUT 1 km w m2 BOTUP total upward flux at ZOUT 1 km w m2 BOTDIR direct downward flux at ZOUT 1 km w m2 radiant fluxes at each atmospheric layer integrated over wavelength Output format write nz phidw do i 1 nz write zz pp fxdn i fxup i fxdir i dfdz heat enddo Prog Doc N IFAC GA 2007 06 LS Page n 165 169 20 21 22 23 MARSCHALS Level 2 Prog Doc N IF AC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 where nz number of atmospheric layers ffew filter function equivalent width um Zz level altitudes km pp level pressure mb fxdn downward flux direct di
2. RTM_SOURCE Batch file for the compilation of RTM code User defined 3 SBDART 2 settings a a Simulated OCM measurements WORKING_DIR Script file for of RTM User defined amp y RTM INP_FILES RTM 0 me Fues settings_rtm dat settings 3 SBDART apa file legend File SBDART code Directory File RTM code content File OFM code Figure 3 4 The arrangement of the OFM codes data and source files The names of the subdirectories are those of the default arrangement MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 36 169 The OFM directory contains 4 subdirectories the BIN subdirectory the SBDART_SOURCE subdirectory the RTM_SOURCE subdirectory the WORKING_DIR subdirectory The BIN directory contains 2 executable files o the sbdart executable file of the SBDART code o the RTM executable file of the RTM code The SBDART_SOURCE directory contains o the source files of the SBDART code o the makefile batch file for the compilation of the SBDART code The RTM SOURCE directory contains o the source files of the SBDART code o the makefile batch file for the compilation of the RTM code Please notices that the OFM is composed by two subcodes the SBDART code and the RTM code that have to be run in sequence Where possible the related files are
3. MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 16 169 The five modules are integrated in three different codes SAMM code MARC code OFM codes gt the SAMM code whose name stands for Supervising Analyzer of MARSCHALS Measurements reads the L1b data and arranges measurement data in the proper format requested by MARC code it works as a pre processor of MARC providing also some auxiliary data It constitutes the Pre Processor Module gt the MARC code whose name stands for Millimetre wave Atmospheric Retrieval Code is based on the Retrieval module and performs the retrieval procedure Actually the retrieval procedure uses the Forward Model module performed by the MFM MARSCHALS Forward Model routine to generate the synthetic spectra necessary to minimize the residuals in case of retrieval on a cloud contamined measurement the Scattering Source Function module performed by the MSSF Mie Scattering Source Function routine is also called gt the OFM codes whose name stands for OCM Forward Model performs a simulation of the radiance collected by the OCM instrument and constitutes the Optical Sounding module The MARC code is the principal code of the MARSCHALS L2 suite and performs the retrieval procedure on the data pre processed by the SAMM code it uses the Retrieval module and the Forward module When necessary it uses also the
4. Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 5 4 3 How to set the bands to be used in the retrieval procedure Page n 90 169 The MARC code performs a multi band retrieval on the three MARSCHALS band B C and D but it can also performs the retrieval on a whatever selection of the aforementioned bands To select the bands on which the retrieval has to be performed the iband flag has to be set Key words related quantity data type iband Array with three logical values T F specifying the bands on which the retrieval is performed T the related band is considered F the related band is not considered the order of the bands inside the array is the following 1 band B 2 band C 3 band D As an example the setting iband TFT means that bands B and D are considered in the retrieval whereas C is not considered dimless 3xlogical This setting is considered also when the MARC code works as a simple Forward Model 5 4 4 How to set the quantities to be retrieved Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 91 169 The MARC code is able to perform a multi target retrieval of various atmospheric species and temperature at the same time The quantities to be retrieved can be specified by settings the values of some key words in the setting file The specification is made by setting
5. The pre processor can be used to generate some files containing instrument information to be used as auxiliary files for the retrieval code particularly the pre processor can generate the following files e instrument dat e rejection dat e fov dat e ils dat This output is optional because these files contain instrument characterization data can be produced once for each flight Key words related quantity Flag used to create the auxiliary file instrument dat CREATE_INSTRUMENT_DAT 1 the file is created flag 0 the file is NOT created Data are extracted from L1b_instrument lut file dimless int Flag used to create the auxiliary file rejection dat CREATE REJECTION_DAT 1 the file is created flag 0 the file is NOT created Data are extracted from l1b_instrument lut file dimless int Flag used to create the auxiliary CREATE ILS DAT file ils dat flag 1 the file is created 0 the file is NOT created dimless int Flag used to create the auxiliary CREATE FOV_DAT file fov dat flag 1 the file is created 0 the file is NOT created dimless int Name of the file with the list of spectral response in band B This file is generated by the user LIST_ and contains the list of the LIB char SPECTRAL _RESPONS_B spectral response files dimless max name_file_list_B The pre processor analyses only 200 char the files summarized in this list
6. gaussian completely random noise is superimposed 4 gaussian standard random noise is superimposed dimless integer lradius rearad Switch T F to specify the model of the Earth Radius T computed from latitude F read below Local radius of the Earth dimless logical real Gene Key words MARSCHALS Level 2 User Manual related quantity Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 116 169 data type rlat Latitude of the simulations deg real Switch T F to define the export of the atmospheric profiles used for the simulations lref dimless logical T external profiles are exported F external profiles are not exported MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 117 169 5 4 15 The setting of the MSSF module The key words of this section are read only if the Mie Scattering Source function is used Imssf F Key words related quantity im data type Switch polarization index for B C D band ipol 1 polarization not defined dimless integer 4 0 vertical polarization plane 1 horizontal polarization plane size1 Radius of water particles cm real 8 size2 Radius of ice particles cm real 8 chem Ice fraction dimless real 8 rzcloud Al
7. 3 for kdist gt 1 atmospheric profile and gas absorption integrals for kdist 1 atmospheric profile read from CKATM also lists wavenumbers specified in CKTAU idb 3 2 causes sbdart to exit after diagnostics 4 solar ephemeris grid parameters atmospheric profiles 5 cloud parameters extinction efficiency asymmetry factor and single scatter albedo 6 aerosol single scattering albedo assymetry factor optical depth increments of total taua and boundary layer aerosols tauab The total optical depth is displayed on the final line of output 7 gaseous absorption integrals and optical depth 8 Optical depth due to Rayleigh aerosols cloud molecular continuum and line single scattering albedo and asymmetry factor Additional printouts for each term in the k fit are produced if KDIST 1 2 element array specifying BOT and TOP altitude points km for IOUT output For example ZOUT 0 50 specifies output information for 0 and 50 km The surface is always set at zero Note that the actual layers for which output is generated is determined by finding the atmospheric layers nearest the chosen value of ZOUT 1 and ZOUT 2 default 0 100 STANDARD OUTPUT SELECTOR no standard output is produced DISORT subroutine is not called but diagnostics selected by idb in gas absorption or aerosol subroutines are active one output record for each wavelength output quantities are WL FFV TOPDN TOPUP TOPDIR BOTDN BOTUP BOTDIR WL
8. MARSCHALS Level 2 fa Cc User Manual Output data when self standing Forward Model is performed Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 125 169 If the MARC code is used in forward modality i e the flag 1mfm is set ON the MARC code produces in general only the observ dat file as a standard output Other files are produced in the Forward Model modality if a further flag is set on if lref TRUE if lref FALSE the pt_ref dat species _ref dat and cont_band_ _ref dat files are produced together with the observ dat file only the observ dat file is produced Output files of the MARC code when used as Forward Model file data contained flag Spectral measurement and some related parameters in the internal format of observ dat MARC code lmfm For more details see RD 3 pt_ref dat For more details see RD 3 Imfm and lref ispecies _ref dat For more details see RD 3 Imfm and lref cont_band_ _ref dat For more details see RD 3 Imfm and lref Table 5 8 Output files of the MARC code when used in the Forward Model modality MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 126 169 5 5 1 The release string All the output files are characterised by a data row that is preceded by the key word input_key e g input_key Simulation 001 001 001 001 001 This data ro
9. tau tbaer Qext wave_length Qext 0 55um where Qext wave_length QBAER interpolated to wave_length If TBAER is not set then the values of OBAER are interpreted as extinction optical depths at each wavelength WLBAER For example the Multi Filter Rotating Shadowband Radiometer MFRSR installed at the Southern Great Plains ARM site is able to retrieve aerosol optical depth in 6 SW spectral channels This information may be supplied to SBDART by setting wlbaer 414 499 609 665 860 938 qbaer 0 109 0 083 0 062 0 053 0 044 0 041 wbhaer 6 9 gbaer 6 0 8 This spectral information is iterpolated or extrapolated to all wavelengths using logarithmic fitting on QBAER and linear fitting on WBAER and GBAER Many aerosol types display a power law dependence of extinction efficiency on wavelength The logarithmic interpolation extrapolation on QBAER will reproduce this behavior if it exists in the input data Wavelengths points um for user defined aerosol spectral dependence Only used when IAER 5 WLBAER and QBAER need not be specified if a single spectral point is set In this case the aerosol optical depth is extrapolated to other wavelengths using a power law see ABAER Single scattering albedo used with IAER 5 WBAER represents the single scattering albedo of boundary layer aerosols at wavelengths WLBAER Asymmetry factor used with IAER 5 GBAER represents the asymmetry factor of boundary layer aero
10. xy xt _ WA 7 2 r a i where iter is the current iteration index A drawback of this criterion is that small variations of the 7 function and even growth may be caused by non linearity d Fi r r r r 2 convergence criterion maximum relative correction The maximum correction that has to be applied to the elements of the state vector is below a fixed threshold tz i e iter 1 iter x Io x j KA Max lt t Different levels of sophistication can be applied to this criterion A different threshold can be used for the different elements of the state vectors The absolute variations can be considered instead of the relative variations whenever an absolute accuracy requirement is present for the elements of the state vector A distinction can be made between target elements quantities for the determination of which the measurements are performed and non target elements quantities that are determined in order to improve the quality of the fit but are not the objective of the measurement e g the instrument parameters MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 ji Cc User Manual Date 10 04 07 Page n 108 169 3 convergence criterion maximum relative correction referred to error Since expression of the 2 convergence criterion is singular whenever an element of the state vector is equal to zero a modified expression that can be considered is
11. MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 1 169 Study The Scientific Analysis of Limb Sounding Observations of the Upper Troposphere MARSCHALS Level 2 Software User and Reference Manual Issue 3 Revision 1 10 April 2007 Prepared by G Bazzini ASPER B Carli IFAC CNR C Cecchi Pestellini IFAC CNR S Del Bianco ASPER B M Dinelli ISAC CNR M Gai ASPER L Santurri IFAC CNR For the comments of Institute IFAC CNR Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 2 169 This page is intentionally left blank MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 3 169 CHANGE RECORD REASON FOR CHANGE AND ISSUE DATE AFFECTED SECTIONS 0 1 15 01 2004 First internal draft First temporary draft Second internal The whole document has been changed 0a WANUNA draft according to a new software arrangement Revision according to some internal 1 0 10 08 2004 First issue comments Full revision and arrangement according oe WASIA oe to substantial changes in the code Revision according to some changes in 3 0 10 04 2007 Third issue the code Some minor changes and corrections 3 1 10 04 2007 Forth issue following MARC developer team obs
12. Executable file of RTM code li Source files of SBDART code i Source files SBDART_SOURCE Batch file for RTM_SOURCE lt gt the compilation of SBDART code F 8 if zinc Source files EQ of RTM code 3 Batch file for the compilation of RTM code i User defined SBDART settings Simulated OCM measurements atms dat i TENNI Settings files i WORKING_DIR Script file for of RTM atmosphere_rtm dat User defined 3 n RTM INP_FILES settings rim dat RTM a settings g SBDART na aa ys e file legend File SBDART code Directory File RTM code File OFM code Figure 6 2 The default arrangement of the OFM codes data and source files The names of the subdirectories are those of the default arrangement MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 130 169 6 1 How to compile the source files of the SBDART and RTM code The SBDART and RTM code is composed by some source files written in standard Fortran language Please notices that the executable code installed on the computer delivered to ESA has been already compiled by the MARSCHALS L2 team if no change has been made to the source codes to run the executable version a compilation is not required The source files
13. Note that this field is read only if CREATE ILS DAT MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 59 169 Key words related quantity name of the file with the list of spectral response in band C This file is generated by the user LIST_ and contains the list of the L1B char SPECTRAL_RESPONS_C spectral response files dimless max name_file_list_C The pre processor analyses only 200 char the files summarized in this list Note that this field is read only if CREATE ILS DAT name of the file with the list of spectral response in band D This file is generated by the user LIST_ and contains the list of the LIB char SPECTRAL_RESPONS_D spectral response files dimless max name_file_list_D The pre processor analyses only 200 char the files summarized in this list Note that this field is read only if CREATE ILS DAT name of the L1B file containing FILE FOV_B the instrumental Field of View char name_file_fov_B in band B dimless max 7 oO Note that this field is read only if 200 char CREATE FOV_DAT name of the L1B file containing FILE FOV_C the instrumental Field of View char name file fov C in band C dimless max B Note that this field is read only if 200 char CREATE FOV DAT name of the L1B file containing FILE FOV_D the instrumental Field of View l char nane Fil
14. files To U Visualization Tool l l l Sequential Output 7 I fit fies 777777 I O files Figure 2 1 The overall scheme of the retrieval procedure performed by the suite of MARSCHALS L2 codes grey boxes represent single modules whereas red contoured boxes indicate the codes The numbers indicates the section of this document in which the related code module or files are described MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 18 169 As previously noticed the aim of the MARSHALS L2 developed software is to provide a retrieval of the atmospheric profile that is the output of the MARC code thus the simulated measurement produced by the forward model during the retrieval procedure would not be required as an output Anyway in order to better perform the test on the adopted forward model the aforementioned simulated measurement can be obtained as an output and saved on a file by adequately setting the MARC configuration parameters in this case the MARC code acts as a Forward Model executing only the part of the code related to the Forward Model module The setting of the MARC configuration is described in 5 4 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 19 169 2 2 The used programming language SAMM code has been developed in standard C langua
15. nn moma do k 1 huge 0 read 11 end 100 wl k do i nz nnt 1 nz read 11 dtau i k waer i k pmom 1 moma i k enddo enddo 100 continue where nn is the number of atmospheric levels for which aerosol information is specified layers 1 through nz nn are unchanged nn should be less than MXLY moma number of phase function moments wl k is the wavelength wl k lt wl k 1l dtau i k is the optical depth increment within level i at wavelength k information is specified in top down order Page n 157 169 RHAER VIS MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 waer i k is the single scattering albedo pmom m i k are legendre moments of the phase function Note that zeroeth moment is not read it is assumed to be 1 O no boundary layer aerosols all BLA parameters ignored 1 rural 2 urban 3 oceanic 4 tropospheric 5 user defined spectral dependence of BLA The wavelength dependence of the aerosol scattering parameters are replaced by those read in from input parameters wlbaer tbhaer wbaer and gbaer Between 1 and 47 spectral values may be specified NOTE the spectral dependence of the boundary layer aerosol models IAER 1 2 3 4 vary with relative humidity See subroutine AEROSOL for details NOTE Don t be mislead by the term boundary layer aerosol The BLA models IAER 1 2 3 4 were originally developed to describe ae
16. to be used in the FOV simulation inside the finer grid should be an odd number The value 2 rfovamp nfov 1 gives the maximum distance between two pencil beams inside the higher precision region of the FOV dimless integer MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 119 169 Key words related quantity im data type Array with parameters related to the FOV function rfov it is constituted by three values rfov rfovampmax rfovamp rzdfov Number specifying the maximum rfovampmax extension for the FOV convolution deg real Half Width Number specifying the amplitude of the finer description of the FOV function Half Width In the current version rfovampmax and rfovamp cannot be equal rfovamp deg real Number specifying step in altitude rzdfov between pencil beam at tangent Km real point Array containing three logical T F switches to specify if the retrieval procedure has to be performed by taking into account for the continuum the contribute of the species H20 O2 and N respectively lcont dimless logical Switch T F to choose the frequency grid for the fine spectrum Lirrgrd Simulajon dimless logical T irregular grid is used F regular grid is used Switch T F to define the ray tracing computation modality T computation is done f
17. water vapor mixing ratio in clear layers is proportionately reduced to maintain the water vapor path specified by WH This option has no effect if RHCLD is negative or TCLOUD is zero default if 1 the relative humidity in clear layers is unchanged NOTE if KRHCLR 1 and clouds are present the actual water vapor path will differ from that specified by WH On the other hand if KRHCLR 0 the normalization procedure may drive the water vapor in clear layers to zero and still be unable produce a given WVP This parameter has no effect when KDIST lt 0O STRATOSPHERIC AEROSOLS LOWTRAN 7 model 5 element array of stratospheric aerosol types O no aerosol 1 background stratospheric 2 aged volcanic 3 fresh volcanic 4 meteor dust altitudes above the surface of stratospheric aerosol layers km Up to 5 layer altitudes may be specified NOTE even though these models are for stratospheric aerosols the scattering layer may be placed anywhere within the numerical grid See ZCLOUD for a discussion of how aerosol cloud layers are positioned within SBDART s computational grid optical depth at 0 55 microns of each stratospheric aerosol layer Up to 5 layer optical depths may be specified BOUNDARY LAYER AEROSOLS BLA Boundary layer aerosol type selector l read aerosol optical depth and scattering parameters from aerosol dat See subroutine AEREAD the file format is readable by the following Fortran code read 11
18. wavelength microns Prog Doc N IFAC GA 2007 06 LS Page n 163 169 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 FFV filter function value TOPDN total downward flux at ZOUT 2 km w m2 micron TOPUP total upward flux at ZOUT 2 km w m2 micron TOPDIR direct downward flux at ZOUT 2 km w m2 micron BOTDN total downward flux at ZOUT 1 km w m2 micron BOTUP total upward flux at ZOUT 1 km w m2 micron BOTDIR direct downward flux at ZOUT 1 km w m2 micron NOTE When ISAT ne 1 these radiometric quantities are each multiplied by the filter function To get the actual specific irradiance divide by FFV WL one output record per wavelength not available for kdist 1 output quantities are WL TXH20 TXCO2 TXO3 TXN20 TXCO TXCH4 TXO2N2 TXTRC TXTOT TXMOL WL wavelength TXH20 log transmission due to water vapor TXCO2 log transmission due to co2 TXO3 log transmission due to ozone TXN20 log transmission due to n2o TXCO log transmission due to co TXCH4 log transmission due to ch4 TXO2N2 log transmission due to 02 and n2 TXTRC log transmission due to trace gases TXTOT log transmission due to all gases TXMOL optical depth due to rayleigh scattering NOTE if you define the optical depth as transmission exp tau then log transmission tau Averaged gas absorption over solar spectrum and filter funct
19. 3 7 idb 8s 1 E The SBDART code can be run in this way at the prompt of the SBDART working directory OFM WORKING DIR SBDART in the default configuration type BIN sbdart gt sbdart dat and then press the enter button The outputs are written in the default configuration in the sbdart dat output file MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 132 169 6 2 2 How to run the RTM code The user can run RTM code by using the batch file run_rtm At the prompt of the RTM working directory OFM WORKING DIR RTM in the default configuration type run_rtm and then press the enter button The user can change the default arrangements of the subdirectory by editing the batch file run_rtm and inserting the personalized organization MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 133 169 6 3 Input files 6 3 1 The input of the RTM code RTM code reads as input the file atmosphere_rtm dat and the file settings_rtm dat The file atmosphere_rtm dat contains data related to the vertical distribution of the pressure temperature and water vapour more on it contains the data elaborated by SBDART such as the linear extinction coefficient and the radiance at a given altitude MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i c
20. 3 MODTRAN_3 solar spectrum 20 cm 1 resolution 100 49960 cm 1 ISAT FILTER FUNCTION TYPES 4 Guassian filter WLINF 2 WLSUP to WLINF 2 WLSUP 3 Trianglar filter WLINF WLSUP to WLINF WLSUP 2 Flat filter WLINF 5 WLSUP to WLINF 5 WLSUP 1 USER DEFINED read from filter dat 0 WLINF TO WLSUP WITH FILTER FUNCTION 1 default NOTE if ISAT 0 and KDIST 1 then the values of WLINF and WLSUP only have an effect if they have been changed from their default values such that WLINF ne WLSUP Otherwise the wavelength sample points are as specified in the CKTAU file 1 METEO 2 GOES EAST 3 GOES WEST 4 AVHRR1 NOAA8 5 AVHRR2 NOAA8 6 AVHRR1 NOAA9 7 AVHRR2 NOAA9 8 AVHRR1 NOAA10 9 AVHRR2 NOAA10 0 AVHRR1 NOAA11 1 AVHRR2 NOAA11 2 GTR 100 chi 3 GTR 100 ch2 4 GTR 100 410nm channel 5 GTR 100 936nm channel 6 MFRSR 415nm channel 7 MFRSR 500nm channel 8 MFRSR 610nm channel 9 MFRSR 665nm channel 20 MFRSR 862nm channel 21 MFRSR 940nm channel 22 AVHRR3 nominal 23 AVHRR4 nominal 24 AVHRR5 nominal 25 Biological action spectra for DNA damage by UVB radiation 26 AIRS1 380 460nm 27 AIRS2 520 700nm 28 AIRS3 670 975nm 29 AIRS4 415 1110nm NOTE If ISAT 1 a user supplied filter data file filter dat is read from the current working directory This ASCII file is read with the following free format read numbers may be separated by spaces commas or carriage returns read 13 end 100 wlfilt i filt i i 1 huge 0 100 con
21. If SAMM code is used to produce the MARC ancillary files i e fov dat ils dat instrument dat and rejection dat files the following L1b data files are needed e files containing spectral response for each channel and each band e files containing the fov acap for each band MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i Cc Date 10 04 07 Page n 50 169 4 3 7 The settings file The user defined options available for the run of the SAMM code are specified in the settings_samm file These options are described in the following paragraph 4 4 MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 51 169 4 4 How to set the SAMM pre processing SAMM Pre Processor provides many options and possible settings of the Pre Processing procedure that can be specified by the user As an example the user can define e The name of the input output files see 4 4 1 e The quantities to be extracted from the L1b data files see 4 4 2 e The set of MARSCHALS observation to be considered see 4 4 2 The user defined choices are addressed to the code by means of a setting file setting_samm dat in ASCII format that can be easily edited by the user In this file the possible choices are defined by means of parameters value specified in some dedicated rows each row containing parameters is preceded by a row with a key word inside
22. L1b data files Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 53 169 The set of data to be extracted from the L1b data files are specified by defining the following parameters Key words related quantity EXTRACTION CRITERION criterion_number The variable criterion_number is a code number that indicates the criterion that the Pre Processor will use to extract data The possible criteria to select are e criterion number l Scan ordinal number The Pre Processor will select the measurements on the basis of the sequential number of the scan The range of the sequential numbers is specified by the parameter labelled by the key word Scans_range_selection defined in the following e criterion number 2 Aircraft coordinates The Pre Processor will select the measurements on the basis of the aircraft coordinates The range of coordinates is specified by the parameter labelled by the key word Scans_coordinate_selection defined in the following dimless int Gene Key words MARSCHALS Level 2 User Manual related quantity Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 Page n 54 169 SCANS RANGE SELECTION scan number 1 scan number 2 SCANS_COORDINATE _ SELECTION latitude 1 longitude 1 latitude 2 longitude 2 The SAMM Pre Processor will extract those scans whose ordinal num
23. LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 142 169 7 1 How to run the visualization tool The visualization tools must be run by directly using the executable file run stored as a default in the MARSCHALS_L2 IDL subdirectory ina IDL environment To run the visualization tool code please follow the above steps 1 move to the subdirectory in which the run IDL executable code is placed the MARSCHALS_L2 IDL in the default arrangement 2 launch the IDL environment by typing at the system prompt idl and then press the enter button 3 Inside the IDL platform launch the visualization tool by typing at the IDL prompt run and then press the enter button A simple graphic user interface will be shown 4 In the graphic interface by using the toolbar choose the directory in which the MARSCHALS output data to be visualized are placed The visualization tool will produce a series of pictures one for each retrieved target with information about the related retrieved quantity These picture are stored in files named target _col TIFF placed in the same directory of the considered MARSCHALS data The output of the visualization tool is described in a more detailed way the following paragraph MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 143 169 7 2 The visualization tool output The visualization tool is able to displ
24. Revision 1 Cc User Manual Date 10 04 07 Page n 40 169 4 1 How to compile the source files of the SAMM code The SAMM code is composed by some source files written in standard C language Please notices that the executable code installed on the computer delivered to ESA has been already compiled by the MARSCHALS L2 team if no change has been made to the source codes to run the executable version a compilation is not required The source files of the SAMM code can be compiled in the usual way by means of whatever C compiler The Linux operating system residing on the computer delivered to ESA is provided with the standard gcc compiler that is able to perform this compilation task Once the compilation is completed the executable code has to be moved in the adequate directory that is in the proposed standard configuration the BIN one The compilation procedure can be performed by using an adequate makefile file provided by the MARSCHALS L2 team together with the code files this makefile file residing in the SAMM_SOURCE subdirectory can be executed by means of the usual make command Once the compilation is performed the makefile automatically moves the executable file in the right standard subdirectory BIN To use the makefile file please move to the directory in which the code files and the makefile file are stored the SAMM_SOURCE subdirectory in the standard configuration and at the prompt type make and
25. a factor 2 exponential fall off between 0 and 1km altitude and then by a factor of 500 between 1 and 100 km If DBAER is set but ZBAER is not set then the elements of DBAER are used to set the aerosol density for each computational layer starting from the bottom layer For example DBAER 10 0 1 0 puts aerosol in the first and third layer If neither ZBAER or DBAER are set the boundary layer aerosols are assumed to follow a pre defined vertical distribution which drops off exponentially with a scale height between 1 05 and 1 51 km depending visibility see VIS Thus even if visibility is not used to set the vertical optical depth it can affect the result through the vertical profile Note that ZBAER and DBAER do not affect the total vertical optical depth of aerosols See discussion for VIS Vertical optical depth of boundary layer aerosols at 0 55 um TBAER input is significant for all values of IAER When IAER 1 2 3 4 the specified value of TBAER supersedes the aerosol optical depth derived from input parameter VIS but VIS still controls vertical structure model unless DBAER and ZBAER are set QBAER is the extinction efficiency QBAER is only active when IAER 5 When TBAER is set QBAER sets the spectral dependence of the extinction optical depth as Page n 159 169 WLBAER WBAER GBAER PMAER MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 User Manual
26. about the rejection factor for the image band of the MARSCHALS instrument ils dat file with general information about the ILS of the MARSCHALS instrument Furthermore the IVP_FILES directory contains the ATM subdirectory with a set of files containing data for the possible sequential fit MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 33 169 The OUT_FILES subdirectory contains the output files with the results of he retrieval the ATM subdirectory The ATM directory contains e files with output data to be used when the sequential fit is adopted The AUX directory contains the auxiliary and the ancillary data files arranged in the following subdirectory REF IND IN GUESS D_SPECT CONT IG ERROR_SPECTRA 000000 The REF IND directory contains the refind dat file with the refractive index table for ice and water The IN_GUESS directory contains a priori data about the atmosphere different atmospheres can be used and for sake of clarity different subdirectories can be used for atmospheres with different characteristics At the present it contains the Mid_Lat subdirectory with data about a mid latitude atmosphere The MidLat subdirectory contains a set of vmr_ species files one for each species species containing the a priori vmr profiles of the species gt the file in_zpt dat with the a priori values of pressure
27. according to a given files arrangement A default arrangement suggested by the MARSCHALS L2 team is described in paragraph 3 1 The computer delivered to ESA is provided with the entire set of files needed to run the MARC code stored according to the default arrangement presented in paragraph 3 1 Prog Doc N IFAC GA 2007 06 LS 5 3 1 The source files MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 80 169 The MARC code is based on a set of source files f source files inc include files written in standard FORTRAN language The whole list is reported in RD 3 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 81 169 5 3 2 The batch files The makefile file can be used to compile the source files and after the compilation to move the executable in the right directory according to the default arrangement described in chapter 3 1 2 for the compilation procedure of the MARC code see 5 1 Please notices that a marc executable file has been already produced and installed in the computer delivered to ESA thus to run the proposed default code a compilation is not required The batch file run marc is provided for the running of the executable code stored in the marc file see 5 2 The run_marc file the running of the executable code is provided with the other MA
28. arrangement described in Sect 3 1 All the aforementioned files have to be placed in a dedicated root directory adequately arranged in subdirectories The name of the root directory can be chosen by the user hereinafter it will be referred as MARSCHALS_L2 directory MARSCHALS software files can be arranged in a personalized manner on the hard disk also different from the one proposed in Figure 3 2 Figure 3 3 and Figure 3 4 The default arrangement of the files needed to run the three MARSCHALS L2 codes can be changed by the user in this case the run file of the related code have to be changed accordingly see 3 1 4 The name of the subdirectories in which the files are arranged have to be specified by the user in the run files of the related codes see 2 4 1 2 Please notice that all the addresses specified in the default batch files are relative addresses referred to the subdirectory of the MARSCHALS L2 root directory in which the source files are placed Once the source files are stored they have to be compiled with the related compiler according to the procedure described in 4 1 SAMM code 5 1 MARC code and 6 1 OFM codes MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 28 169 3 1 The default arrangement of the files As a default option we propose the following architecture for the arrangement of the files needed to run the three co
29. be Switch T F for the convergence criterium th S Lconv4 T 4 Convergence Criterium dimless EN 5 Convergence Criterium Key words conv_criterium5 related quantity Array with two elements rthreshold5 lconvs5 data type Threshold on Marquardt parameter Convergence is reached only if each is considered F gt 5 Convergence Criterium is not considered Ebr esp ddl Marquardt parameter is less than the cule ii set threshold Switch T F for the convergence criterium Lconv5 T gt 5 Convergence Criterium dmeas eia Gene Stop Criteria MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 112 169 Key words related quantity data type iMaxTterG Maximum number of allowed Gauss dimless int iterations Macro iteration Maximum number of allowed iMaxIterM Marquardt iterations Micro dimless int iteration Combination of the Convergence Criteria Key words related quantity data type OR_AND Switch for an AND or an OR combination of the convergence criteria 0 AND logic operation is performed among the results of the selected Convergence Criteria 1 OR logic operation is performed among the results of the selected Convergence Criteria dimless integer 5 4 12 How to set
30. composed by three codes that have to be run in a sequence this document is devoted to the description of the use of the overall suite and of all the single codes MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 10 169 1 2 How to read this document In the following chapters of this document information about the installation the setting and the execution of the codes composing the MARSCHALS L2 software are given In Chapter 0 the list of the applicable and reference documents is presented together with the acronyms and the conventions used in this document In Chapter 2 an overview of the MARSCHALS L2 software is provided the list of the source files is also presented in Paragraph 2 4 with the whole list of the files the code needs of to perform the retrieval procedure Chapter 3 is devoted to the description of the codes installation procedures with the specific hardware and software requirements A detailed description about the use of the codes is then given in chapter 4 5 and 6 in these chapters the parameters setting opportunities are also explained In chapter 7 a tool provided to visualize the results of the retrieval procedure id described The run of the MARSCHALS L2 software requires also the execution of the external not developed by the MARSCHALS L2 team SBDART code In annex 1 chapter 8 the original documentation of SBDART code is reported soft
31. composed of 100um snow grains may be modeled with the following input file amp INPUT sza 30 idatm 4 spowder t wlinf 3 wlsup 2 iout 1 tcloud 10000 10 zcloud aL 2 nre 100 10 Similarly the scattering properties of the surface and atmosphere may be read from aerosol dat with IAER 1 To model a semi infinite granular surface layer the optical depth of the bottom layer should be made very large e g 10000 as indicated in the example However a smaller optical depth may also be specified in conjuntion with a given value of sub surface albedo selected with ISALB Thus in the previous example the effect of a thin snow layer covering a grass field may be modeled by setting tcloud 100 10 and isalb 6 NOTE At present there is no way to set the surface skin temperature to something other than the atmospheric temperature of the bottom level i e the parameter BTEMP is ineffective when SPOWDER is true nothrm 1 gt Thermal emission turned on only for wavelengths greater than 2 0 um default Note During daylight hours solar radiation is a factor of about 1 e5 greater than thermal radiation at 2 0um nothrm 0 gt Thermal emission turned on for all wavelengths Page n 161 169 KDIST ZGRID1 ZGRID2 NGRID MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 nothrm 1 gt No thermal emission NOTE If thermal emission is desired be su
32. default arrangement of the files some changes have to be made also to the provided batch files the batch files to be used to run the codes i e run_marc see 5 2 stored as a default in the MARC working sub directories have to be changed accordingly by the user If used also the makefile files that are provided together with the source codes of the MARSCHALS L2 codes in order to make easier the compilation of the codes have to be changed accordingly see 4 1 for the SAMM code 5 1 for the MARC code 6 1 for the OFM code MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 39 169 4 SAMM code This chapter aims to describe the use of SAMM Pre Processor code with its implemented features The SAMM Pre Processor is intended as a tool to plan the activity of data analysis and should be used to perform a first selection of the MARSCHALS measurements and transferring them into a suitable format readable by MARC code The SAMM Pre Processor will allow tackling changes in the format of the input data without any changes in the MARC code in such a way a new validation of the Retrieval Code will be not necessary in case of changes in input data format An important functionality of SAMM is to provide to the user an overview of the measurement scans available from the analysed campaign data comprehensive of those information which could help the user in the select
33. distribution model of absorption by atmospheric gases Since a three term exponential fit is used SBDART execution times are up to 3 times longer with KDIST gt 0 compared to KDIST 0 KDIST 2 causes the k fit transmissions to exactly match the LOWTRAN transmission along the solar beam direction This option may be useful when computing surface irradiance under clouds of optical thickness less than about 10 This is because in this thin cloud case much of the radiation which reaches the surface propagates along the direct beam direction KDIST 3 causes the k fit transmissions to exactly match the LOWTRAN transmission along the solar beam direction for parts of the atmosphere above a scattering layer As the scattering optical depth increases above 1 the k fit factors are ramped back to there original LOWTRAN values The effect of the slant path correction is ramped down to zero for wavelengths greater than 4um where solar energy input is less important KDIST 3 is the default These three parameters can be used to change the grid resolution of the model atmosphere ZGRID1 controls the resolution near the bottom of the grid while ZGRID2 sets the maximum permissible step size at the top of the grid NGRID sets the number of grid points For example ZGRID1 5 ZGRID2 30 NGRID 45 specifies a 45 element grid with a resolution of 5 km throughout the lower part of the grid and a largest step of 30 km The regridding is performed aft
34. error is computed These extra feature have not to be used with the normal MARSCHALS LIb data the code is developed for and thus the description of their usage is out of the purpose of this document MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 61 169 4 4 6 The contrast level The contrast level related to a given spectrum is defined as follows peak Contrast _ Level valley that is the ratio between two reference intensities Ipeak and Ivattey Reference intensities are evaluated by averaging the spectral intensity over given spectral intervals each related to a specific band as specified in the following For band B Ipeak is evaluated as the mean value on the interval 301 6 302 0 GHz for a total of three spectral points This interval contains the strongest O line Ivatiey is evaluated as the mean value on the interval 295 6 296 4 GHz for a total of five spectral points see Figure 4 1 exemple of spectrum in band B 250 200 150 K 100 50 0 T T T T T 294 296 298 302 304 306 300 frequency GHz Figure 4 1 The range in which the Iyaney green zone and the Ipeak yellow zone values are evaluated in the band B MARSCHALS Level 2 TT Cc User Manual For band C Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10
35. files 85 5 3 7 The setting file 86 5 4 How to set up the marc code execution 87 5 41 How to set the release of the run of the retrieval code 88 5 4 2 How to run the MARC code as a retrieval procedure or a simple forward model 89 5 4 3 How to set the bands to be used in the retrieval procedure 90 5 4 4 How to set the quantities to be retrieved 91 5 4 5 How to set the retrieval grid 95 5 4 6 How to set the atmospheric parameters 98 5 4 7 How to set up the retrieval modality and techniques 101 5 4 8 How to set up the Gauss and Marquardt iterations 102 5 4 9 How to set up the regularization procedure 104 5 4 10 How to set up the optimal estimation procedure 106 5 4 11 How to set the convergence and stop criteria for retrieval 107 5 4 12 How to set the use of the Variance Covariance Matrix VCM of the Forward Model 113 5 4 13 How to set the sequential fit procedure 114 5 4 14 How to set the internal Forward model of the MARC code 115 5 4 15 The setting of the MSSF module 117 5 4 16 How to define the parameters for the advanced settings 118 5 5 Output files 121 Gene MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 7 169 5 5 1 The release string 126 6 OFM 127 6 1 How to compile the source files of the SBDART and RTM code 130 6 2 How to run the OFM 131 6 2 1 How to run SBDART code 131 6 2 2 How to run the RTM code 132 6 3 Input files 133 6 3 1 The input of the RTM cod
36. horizontal gradient of gas specie _grad dat specified by the string specie and related data For more details see RD 3 lgvmrfit retrieved pointing value and related data lpoifit For more details see RD 3 pointing dat Retrieved profile of un accounted cont_band_ dat continuum in the band specified by the key word and related data For more details see RD 3 lexcfit retrieved instrumental offset and AjIsekadi related data lofsfit For more details see RD 3 sanda retrieved instrumental gain and related data lgainfit For more details see RD 3 Table 5 6 Optional output files of the MARC code when used in Retrieval Modality MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 124 169 Output data when sequential fit is performed If the 1seq flag is set 7 the MARC code performs the sequential fit in order to perform the retrieval in a sequential modality some data have to be passed from the previously retrieval to the following one To pass these data for each retrieved quantities a set of files is produced These files are output files of retrieval that begin input files of the following retrieval they are introduced in the following and separately described in RD 3 file contained data VCM matrix of temperature and related data temperature VCM dat For more details see RD 3 VCM matrix
37. in this case to use the make and the make install command the makefile file have to be edited accordingly to the user defined arrangement of the files MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 131 169 6 2 How to run the OFM OFM is not a single self standing code To simulate the OCM measurement two independent codes have to be sequentially used First of all SBDART code have to be run to simulates the source function and the extinction coefficient profile in a plane parallel atmosphere see 6 2 1 Then the data produced by the SBDART code are used as input of the RTM code that have to be run see 6 2 2 on these data to simulate the OCM measurement in a spherical geometry 6 2 1 How to run SBDART code The settings of the SBDART code and the procedure to run it are described in Annex 1 Chapter 0 Please notice that the SBDART code is a general tools and in the MARSCHALS case it have to be set by editing the INPUT file in order to produced the desired source functions and the extinction atmospheric profiles The following suitable default INPUT file is provided with the MARSCHALS codes amp INPUT idatm 0 j wlinf 0 855 j wlsup 0 855 7 sza 39 6483 F isalb 0 albcon 0 0 P zcloud 10 0 12 0 tcloud 0 0 1 0 nre 1020 10 0 iout 22 j nstr 40 j uzen 90 r nphi 10 f phi 0 0 180 0 pa imomc
38. is not necessary to provide input records for layers above the highest cloud In addition a forward slash terminates interpretation of data values Prog Doc N IFAC GA 2007 06 LS Page n 155 169 MARSCHALS Level 2 Issue 3 Revision 1 Cc User Manual Date 10 04 07 in a record For example the following records in usrcld dat specify a cloud that extends from 2 to 4 km assuming idatm gt 0 and no regridding no cloud between 0 1 km no cloud between 1 2 km 30 lwp 30 re 10 between 2 3 km 60 20 5 30 2 lwp 60 re 20 fwp 5 reice 30 cldfrac 2 between 3 4 km Se SE FE EOE Any input quantities that are left unspecified will retain their default values of lwp 0 reff 8 fwp 0 reice 1 and cldfrac 1 The radiative properties of ice are computed from a CCM3 model see subroutine ICEPAR IMOMC Controls the phase function model used in cloud layers 1 isotropic scattering 2 rayleigh scattering phase function 3 henyey_greenstein a function of asymmetry factor g re 4 haze L as specified by garcia siewert 5 cloud c 1 as specified by garcia siewert default 3 LWP The liquid water path or frozen water path if nre le 0 of a cloud is specified in units of g m2 This is another way to specify cloud optical depth A linearly varying opacity distribution can be obtained by setting the second LWP entry to a factor which represents the ratio of the opacity in the highest layer to that in the lowest layer
39. micron TOPUP total upward flux at ZOUT 2 km w m2 micron TOPDIR direct downward flux at ZOUT 2 km w m2 micron BOTDN total downward flux at ZOUT 1 km w m2 micron BOTUP total upward flux at ZOUT 1 km w m2 micron BOTDIR direct downward flux at ZOUT 1 km w m2 micron NPHI number of user azimuth angles NZEN number of user zenith angles PHI user specified azimuth angles degrees UZEN user specified zenith angles degrees VZEN user specified nadir angles degrees UURS radiance at user angles at w m2 um str altitude ZOUT 2 top NOTE The radiance output from SBDART represents scattered radiation It does not include the solar direct beam Also keep in mind that UURS represents the radiance at the user specified sample directions Hence computing the irradiance by an angular integration of UURS will not yield BOTDN because of the neglect of the direct beam and it will probably not yield BOTDN BOTDIR because of under sampling NOTE if IDAY is set then PHI is the actual compass direction in which the radiation in propagating same as IOUT 5 except radiance is for ZOUT 1 altitude bottom radiative flux at each layer for each wavelength This output option can produce a huge amount of output if many wavelength sample points are used write fzw block id used in postprocessors write nz 7 number of z levels write nw number of wavelengths do j 1 nw write wl write
40. no beam source FBEAM 0 0 or 3 there is no scattering SSALB 0 0 for all layers Radiance output Number of user zenith angles If this parameter is specified SBDART will output radiance values at NZEN zenith angles evenly spaced between the first two values of input array UZEN For example nzen 9 uzen 0 80 will cause output at zenith angles 0 10 20 30 40 50 60 70 80 User zenith angles If NZEN is specified then UZEN is interpreted as the limits of the zenith angle range and only the first two elements are required If NZEN is not specified then up to NSTR values of UZEN may be specified If neither NZEN nor UZEN is specified and Page n 167 169 VZEN NPHI PHI PHIO MARSCHALS Level 2 Issue 3 Revision 1 User Manual Date 10 04 07 a radiance calculation is requested IOUT 5 6 20 21 22 23 a default set of zenith angles is used which depends on the value IOUT as follows TOUT 5 or 20 NZEN 18 UZEN 0 85 TOUT 6 or 21 NZEN 18 UZEN 95 180 TOUT 22 or 23 NZEN 18 UZEN 0 180 NOTE UZEN specifies the zenith angle of at which the radiation is propagating UZEN 0 gt radiation propagates directly up UZEN lt 90 gt radiation in upper hemisphere UZEN gt 90 gt radiation in lower hemisphere UZEN 180 gt radiation propagates directly down user nadir angles This is just an alternate way to specify the direction of user radiance angles whereby uzen 180 vzen Number o
41. of a definition of a matricial quantity is reported in the following comments key word ant a12 iin Ain ami Am2 amn Because of the presence of the key word string the order of the definition among different data is not matter of care in any case MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 27 169 3 Installation of the MARSCHALS L2 codes Please notices that the executable version provided to ESA with the delivered computer has been already installed and compiled by the MARSCHALS L2 team and does not require any further installation procedure The reading of this chapter is thus not needed to run the executable codes since the installation have to be performed only to transfer the software on another computer The MARSCHALS L2 software is composed by the SAMM code written in C language the MARC and the OFM codes these last written in FORTRAN language The MARSCHALS L2 codes are provided by means of their source files To run the MARSCHALS L2 codes the user needs of the source files to be compiled but also adequate sets of input data files ancillary files and auxiliary files are requested The files needed to run the MARSCHALS L2 codes has been listed in 2 4 To install the codes it is sufficient to copy the source files the input data files the auxiliary files and the ancillary files on the hard disk according to a given
42. of species specified by the flag species and related data For more details see RD 3 species _VCM dat VCM matrix of retrieved horizontal gradient of species _grad_VCM dat quantity specified by the species string and related data For more details see RD 3 VCM Matrix of continuum of band specified by the flag and related data For more details see RD 3 cont_band_ _VCM dat Initial guess for the atmospheric species with their gradients a dedicated file specified by the string yn species dat species is provided The errors of the VMR and gradient profiles are also reported For more details see RD 3 Initial guess for the atmospheric profiles of temperature and pressure with the related in_zpt dat gradients the error of the temperature profile is also reported For more details see RD 3 Profile with the altitude of the continuum value cont_band_ dat for the band specified by the string For more details see RD 3 Table 5 7 Optional set of output files of the MARC code one set for each retrieved quantity when used in the Retrieval Modality and a sequential fit is adopted Notice that the in_zpt dat vmr_ species dat and cont_band_ dat are replied with the same format of the input files with the same names used for the retrieval procedure and do not complain the key word syntax The other files are produced in order to complain with the key word syntax
43. of the SBDART and RTM code can be compiled in the usual way by means of a whatever FORTRAN compiler The computer delivered to ESA is provided with a FORTRAN compiler from Portland Company that is able to perform this compilation task Once the compilation is completed the executable code has to be moved in the adequate directory that is in the proposed standard configuration the BIN one The compilation procedure can be performed by using an adequate makefile file provided by the MARSCHALS L2 team together with the code files this makefile file residing in the SBDART_SOURCE and RTM_SOURCE subdirectory can be executed by means of the usual make command Once the compilation is performed the make install command can be used to move the executable version To use the makefile file please move to the directory in which the source files and the makefile file are stored the SBDART_SOURCE and RTM_SOURCE directory in the standard configuration and at the prompt type make and then press the enter button To use the make install command after the compilation of the source files please move to the directory in which the compiled files are stored the BIN directory in the standard configuration and at the prompt type make install and then press the enter button Please notice that the make and the make install commands works according to the standard arrangement of the file the user can change this arrangement but
44. results of the OFM code and the auxiliary MSSF module A Visualization Tool is also provided in order to plot the results of the retrieval The Figure 2 1 provides an overview of the developed modules suggesting how these modules have been integrated to generate the MARSCHALS codes A retrieval procedure is performed by running the SAMM code on the L1b data and then the MARC code on the output of the SAMM execution in some conditions also the execution of the OFM code is needed To compile and execute the SAMM code and the MARC code see Chapter 4 and Chapter 5 respectively The compilation and execution of the OFM code is described in Chapter 6 MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 17 169 User defined choices Settings Batch files files settings_samm dat run_samm SAMM code settings_ofm dat Pre Processor module settings_marc dat y run_marc MARC code Pre processed C input data files daid p o gt Retrieval module observation dat OFM code Optical Sounding module Forward Model 6 Module MFM ti User choice Founne l File renaming C2 observation_ dat Auxiliary files observ dat Scattering Source Function Module C MSSF routine Ancillary S af files Library
45. scripts where some indication of file names and paths is addressed More details can be found in 4 2 batch for the SAMM code 5 2 batch for the MARC code and 6 2 batch for the OSM code 2 4 1 3 The executable files Three executable files one for each code are provided They are the results of a compilation of the source files of the code performed by the MARSCHALS L2 team MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 24 169 2 4 2 The data files The data files contain the data processed by the MARSCHALS L2 code together with some user defined settings concerning the run of the codes The set of data files is composed by e the settings files with the user defined set up of the code see 2 4 2 1 e the Llb measurement file with the measurement data coming from the MARSCHALS Llb team see 2 4 2 2 e the Pre processed input data files with the aforementioned L1b measurement file elaborated by the SAMM code see 2 4 2 3 the ancillary files see 2 4 2 4 the auxiliary files see 2 4 2 5 the output files see 2 4 2 6 the I O files for the sequential fit see 2 4 2 7 2 4 2 1 The settings files The MARSCHALS L2 software is built in order to offer to the user the opportunities to set many options and to activate some extra features for the SAMM the MARC and the OFM codes These user defined choices are stated by editing some ASCII settin
46. the order inside this array is that defined in 5 4 4 rlambdavg Initial value for the parameter in rlambdapoi the Marquardt iterations used for the dimless real retrieval of the pointing angle offset Initial value for parameter in the rlambdaexc Marquardt iterations used for the dimless real retrieval of the external continuum Initial value for parameter in the rlambdaoff Marquardt iterations used for the dimless real retrieval of the instrumental offset Gene MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 103 169 Key words related quantity data type Initial value for parameter in the rlambdagain Marquardt iterations used for the dimless real retrieval of the instrumental gain rgaussDumpingFac i i i xg ping Dumping Factor for Gauss iterations Pe eee reai tor macro iteration xMarquardtDumpi Dumping Factor for Marquardt l iterations dimless real ngFactor micro iteration Threshold_on i gt Threshold to activate the Marquardt Weighted_ terati dimless real Chisq_increment C O 1teration MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i c ser Date 10 04 07 Page n 104 169 5 4 9 How to set up the regularization procedure The MARC retrieval code can be performed by considering a Regulariza
47. the use of the Variance Covariance Matrix VCM of the Forward Model The use of the Variance Covariance Matrix can be specified by the user by setting the values related to the lvcmfm key word in the setting file Key words lvcmfm MARSCHALS Level 2 User Manual related quantity Switch T F for the use of FM VCM T FM VCM is used F FM VCM is not used dimless Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 Page n 113 169 data type logical If lvemfm is TRUE MARC computes the VCM of the Forward Model using the files listed in RD 3 The default directory where these files are located in reported in Figure 5 1 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i c ser Date 10 04 07 Page n 114 169 5 4 13 How to set the sequential fit procedure The MARC code is able to perform a sequential fit i e to carries out a retrieval in which the initial guess is constituted by the result of the previous retrieval and a completed VCM of a priori information is used To impose the sequential fit procedure the user has to set to T the value related to the lseq key word of the setting file Key words related quantity im data type Switch T F to select the sequential fit procedure T the sequential fit is performed F the usual not sequential fitting procedure is used lseq dimless logica
48. to ESA is provided with the entire set of files needed to run the SAMM code stored according to the default arrangement presented in paragraph 3 1 MARSCHALS Level 2 iA Cc User Manual 4 3 1 The source files Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 44 169 The SAMM code is based on a set of source files c source files h include files written in standard C language The whole list is reported in RD 3 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 45 169 4 3 2 The batch file The makefile batch file can be used to compile the source files and after the compilation of the source files to move the executable in the right directory according to the default arrangement described in chapter 3 1 1 for the compilation procedure of the SAMM code please see 4 1 Please notices that the samm executable file has been already produced and installed in the computer delivered to ESA thus to run the proposed default code a compilation is not required The batch file run_samm is provided for the running of the SAMM executable code stored in the samm file MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 46 169 4 3 3 The executable file The samm executable file obtained by compiling the SAMM code source files accordi
49. with isotopic ratio information vmr_ species dat files with information about VMR profile Wa T file with information about pressure and lt P temperature profile cont band 2 files with information about atmospheric cont_band_C oe aali continuum profile T_band_ dat species _band_ dat species _g_band_ dat files with information about systematic continuum_band_ dat errors pointing_band_ dat Error spectra data files file with information about frequency marschals grd irregular grid Table 5 3 The list of the MARC auxiliary data files Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 86 169 5 3 7 The setting file The user defined set up of the MARC code execution for the retrieval procedure are specified in the settings_marc dat file The set up options offered to the user are detailed in paragraph 5 4 MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 87 169 5 4 How to set up the marc code execution The retrieval performed by the MARC code offers many options that can be set by the user at the same time to run the code the user has to fix many parameters to set the algorithms of the retrieval procedure The selection of the possible options and the setting of the retrieval procedure has to be made by editing the related sett
50. 04 07 Page n 62 169 Ipeak is evaluated as the mean value on the interval 324 9 325 3 GHz for a total of three spectral points This interval contains the strongest H20 line Daley is evaluated as the mean value on the interval 322 5 323 3 GHz for a total of five spectral points see Figure 4 2 exemple of spectrum in band C 0 FE T T T 316 318 320 322 frequency GHz 324 326 values are evaluated in the band C Figure 4 2 The range in which the Iyaney green zone and the Ipeak yellow zone For band D Ipeak is evaluated as the mean value on the interval 343 0 343 4 GHz for a total of three spectral points This interval contains the strongest O line Ivatiey is evaluated as the mean value on the interval 347 6 348 4 GHz for a total of five spectral points exemple of spectrum in band D 20 o n ii Ea Ms sa OO E E 342 343 344 347 345 346 frequency GHz 348 349 values are evaluated in band D Figure 4 3 The range in which the Iyaney green zone and the Ipeak yellow zone MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i Cc Date 10 04 07 Page n 63 169 4 5 Output files According to the user defined setting specified by the parameter EXTRACTION_TYPE in the settings_samm file see 4 4 3 the SAMM Pre Proc
51. 1 I c Date 10 04 07 Page n 26 169 2 5 The key word syntax of the files with the user defined settings The settings files used by the MARSCHALS L2 codes contain data that can be edited or read by the user To make the editing and the reading easier the most part of these files are written in ASCII format and are characterised by the following syntax hereinafter key word syntax In the settings file characterized by the key word syntax data are written in ASCII format and arranged in rows inside a single data row all data in ASCII format are separated by some blanks The data the row is related to is specified in the previous row with a key word Possible commented rows are preceded by a Data and commented rows can be separated by some blank rows that will not be considered A single data is specified in the following way a commented row starting with a introduces the variable under definition then a string hereinafter key word identifying the variable to be defined is written in a not commented row between square brackets the related quantity is reported in the following not commented data rows Notice that the variable could be vectorial i e a single data specification can occupied more than a single row in case of matricial data moreover the format of the data can be different among the row elements e g in the same row it is possible to find integer data together with real data and so on An example
52. AC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 64 169 4 5 1 The flight_view dat file with the general information about the flight campaign General information about the flight campaign are reported in the file named flight_view dat It reports information about the whole set of scans collected during the campaign giving indication about which scans meet the selection criteria defined by the user The flight view dat is an ASCII file in raw format The content and the organization of this file is described in Table 4 3 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 65 169 4 5 2 Thescan_ view dat_ file with the general information about the selected scan General information about the selected scan are stored in the ASCII file named scan_ _view dat where is a string referred to the specific scan This file reports information related to the selected scan giving also the indication about which lines of sight meet the selection criteria defined by the user The content and the organization of the file are detailed in Table 4 4 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 66 169 4 5 3 The OCM_ dat file with the OCM images The OCM data are stored in the file named OCM_ dat where indicates the specific scan A dedicate elaborati
53. ALS Level 2 User Manual 5 4 6 How to set the atmospheric parameters Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 98 169 The retrieval procedure is based on a forward model that uses an atmospheric model the user can define some features and parameters of this atmospheric model such as the used reference system the boundaries of the atmosphere the presence of horizontal gradient the possibility to take into account the continuum of certain species the imposition of the hydrostatic equilibrium the presence of scattering phenomena and some instrument parameters The aforementioned settings have to be specified by defining the values related to the following key words in the setting file Key words related quantity data type hydroeq Switch T F to impose if True the hydrostatic equilibrium T hydrostatic equilibrium is imposed F hydrostatic equilibrium is not imposed dimless logical rzpr Number specifying the reference altitude for imposition of hydrostatic equilibrium Km real lgrad Switch T F for the use of the Horizontal Gradients T horizontal gradients are used 2D atmosphere F horizontal gradients are not used 1D atmosphere dimless logical lextcont Switch T F for the use of external continuum in the Forward Model procedure T the external continuum is used F the external continuum
54. For more details see the discussion of TCLOUD NOTE a 1 mm column of liquid water 1000 g m2 NOTE LWP and TCLOUD cannot be used at the same time NOTE The cloud optical depth is related to LWP by 4 RHO NRE where Q is the scattering efficiency and RHO is the density of liquid water 1 g cm3 The value of Q that applies to a distribution of cloud droplets can be expressed in terms of the extinction cross section at a given wavelength and liquid drop radius Let sigma extinction cross section at a given wavelength and drop radius q sigma pi r 2 dimensionless where pi r 2 is the geometrical cross section of the cloud drop then Q is a weighted average over drop radius given by 2 2 Q lt rqwN r gt lt r N r gt for visible light 0 is typically about 2 dimensionless For example NRE 10um and LWP 200g m2 0 2mm gt tau 30 RHCLD The relative humidity within a cloud layer a floating point value between 0 0 and 1 0 RHCLD lt O disables the adjustment Prog Doc N IFAC GA 2007 06 LS Page n 156 169 KRHCLR JAER ZAER TAERST IAER MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 User Manual of relative humidity in which case the relative humidity in the cloud layer follows solely from the temperature and water vapor density of the initial model atmosphere This parameter has no effect when KDIST lt 0O If zero
55. G_DIR directory export MARC_PT The MARC_PT variable indicates the path of the directory containing the file related to the temperature and pressure profiles of the atmosphere that is the in_zpt dat file The default configuration is export MARC_PT AUX IN_GUESS Mid_Lat and indicates that the files will be placed in the MARC AUX IN_GUESS Mid_Lat directory export MARC_VMR The MARC_VMR variable indicates the path of the directory containing the files related to the atmospheric VMR profiles of the species that is set of vmr_ species dat files As an example export MARC_VMR AUX IN_GUESS Mid_Lat indicates that the atmospheric VMR profile files will be placed directly in the MARC AUX IN_GUESS Mid_Lat directory export MARC_SPECT The MARC_SPECT variable indicates the path of the directory containing the files related to the spectroscopic database As an example export MARC_SPECT AUX DSPECT indicates that the data files will be placed in the MARC AUX DSPECT directory MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 ji Cc User Manual Date 10 04 07 Page n 75 169 line in run_marc file Related setting export MARC_CONT The MARC_CONT variable indicates the path of the directory containing the files related to the continuum used in the MARC retrieval i e cont_band_ dat As an example export MARC_CONT AUX CON
56. Manual Issue 3 Revision 1 1 c Date 10 04 07 Page n 107 169 5 4 11 How to set the convergence and stop criteria for retrieval The MARC code performs a retrieval process based on a iterative procedure five convergence criteria have been defined to check the convergence of the iterations The user can choose which convergence criterion has to be used a combination of more than one convergence criteria can be also defined the selected criteria can be checked according to user defined AND or OR logic operator The user has to choose the convergence criteria that have to be used and the logical combination of these criteria The user has also to set some parameters of the selected convergence criteria In order to avoid endless loop a stop criterion to terminate the iterative procedure after a given number of Gauss or Marquardt iterations is also provided the corresponding maximum number of allowed iterations is user defined When the maximum number of Marquardt iteration is reached the code moves itself to the next Gauss iteration Conversely if the max number of Gauss iteration is reached the iteration procedure is stopped even if the convergence is not reached The five implemented convergence criteria are the following 1 convergence criterion relative variation of 7 The relative variation of the 7 function obtained in the present iteration with respect to the previous iteration is less than a fixed threshold t i e
57. N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 mixing ratio defined in the US62 atmosphere as listed above The volume mixing ratio VMR of a given species is adjusted by specifying the surface value of its VMR in PPM The entire altitude profile is multiplied by the ratio of the user specified VMR and the nominal surface VMR There are no further re normalizations of the VMR Thus the total of all the VMRs may be greater or less than 10 6 By the way the default set of VMRs do not add up to 10 6 because of the exclusion of the noble gases which do not have any radiative effects sensitivity factor for Rayleigh scattering default 1 This factor varies the strength of Rayleigh scattering for sensitivity studies sensitivity factor to adjust strength of absorption by oxygen collisional complexes default 1 see comments in subroutine o4cont surface pressure in millibars If PBAR gt 0 then each pressure is multiplied by the factor PBAR PO where PO is the surface pressure of the original atmosphere At each height in the atmosphere the density of all molecular species sources of absorption and Rayleigh scattering is proportional to pressure f PBAR le 0 the original pressure profile is used Surface altitude in kilometers This parameter is just an alternate way of setting the surface pressure and should not be set when PBAR is specified When ZPRES is set PBAR is obtained by logarithm
58. R DTAUC SSALB PMOM NSTR USRANG NUMU UMU ALBEDO DELTAM PRNT HEADER and the array dimensions NOPLNK LAMBER are assumed TRUE the bottom boundary can have any ALBEDO the sole output is ALBMED TRNMED UMU is interpreted as the array of beam angles in this case If USRANG TRUE they must be positive and in increasing order and will be returned this way internally however the negatives of the UMU s are added so MAXUMU must be at least 2 NUMU If USRANG FALSE UMU is returned as the NSTR 2 positive quadrature angle cosines in increasing order FISOT intensity of top boundary isotropic illumination units w sq m if thermal sources active otherwise arbitrary units corresponding incident flux is pi 3 14159 times FISOT TEMIS emissivity of top layer not used if NOTHRM 1 BTEMP Surface skin temperature in Kelvin If not set the surface temperature is set to the temperature of the bottom layer of the model atmosphere The surface emissivity is calculated from the albedo see ISALB and ALBCON The input value of BTEMP is ignored when NOTHRM 1 or SPOWDER true DISORT specific output options k Ki input variables except PMOM 2 fluxes 3 intensities at user levels and angles 4 planar transmissivity and planar albedo as a function solar zenith angle IBCND 1 5 phase function moments PMOM for each layer only if PRNT 1 TRUE and only for layers with scattering
59. RC files inside the computer delivered to ESA stored in the default configuration in the MARC WORKING_DIR subdirectory MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 Cc Date 10 04 07 Page n 82 169 5 3 3 The executable file The marc executable file obtained by compiling the SAMM code source files according to the default setting is provided inside the computer delivered to ESA stored in the MARC WORKING_DIR subdirectory MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 83 169 5 3 4 The pre processed data files The MARC code operates on the data coming from the L1b measurement preprocessed by the SAMM code As a matter of fact it operates on the data stored in the observ dat file a file that is obtained by the user from the set of the observation_ dat files produced by the SAMM code being the scan by choosing one of these file that related to the desired scan and renaming it with the name observ dat For more details on these files see RD 3 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 MARSCHALS Level 2 Fa c User Manual Date 10 04 07 Page n 84 169 5 3 5 The ancillary data files The MARC code of the MARSCHALS software needs also some information related to the processed measure i e to some parameters and data related to the acquisition of the data under considera
60. RSCHALS Level 2 User Manual 1 4 Acronyms Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 12 169 Acronyms list ADD Advanced Detailed Description ATBD MARSCHALS Level 2 Algorithm Theoretical Baseline Document ESA European Space Agency IDL Interactive Data Language FORTRAN FORmula TRANslation Lib Level 1 b of MARSCHALS project L2 Level 2 of MARSCHALS project MARC Millimetre wave Atmospheric Retrieval Code MARSCHALS Millimetre wave Airborne Receivers for Spectroscopic CHaracterization of Atmospheric Limb Sounding MFM Marschals Forward Model MSSF Mie Scattering Source Function OCM Optical Cloud Monitor OFM OCM Forward Model OSM Optical Sounding Module RTM Radiative Transfer Model SAMM Supervising Analyzer of MARSCHALS Measurements SBDART Santa Barbara DISORT Atmospheric Radiative Transfer VMR Volume Mixing Ratio MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 13 169 1 5 Conventions General Identifier String identifier indicating a generic MARSCHALS scan number B identifier indicating a generic string among the three band names band_B band_C band_D identifier indicating a generic MARSCHALS band a letter among the band letters B C D identifier indicating a generic flight campaign of the instr
61. T indicates that the continuum data files will be placed in the MARC AUX CONT directory export MARC_Error_Spectra The MARC_Error_Spectra variable indicates the path of the directory containing the files related to the error spectra used in the MARC retrieval As an example export MARC_ Error_Spectra AUX ERROR_SPECTRA indicates that the error data files will be placed in the MARC AUX ERROR_SPECTRA directory export IG_DIR The IG DIR variable indicates the path of the directory containing the file related to the irregular grid to be used when specified in the MARC retrieval i e the marschals grd file As an example export IG_DIR AUX IG indicates that the irregular grid data files will be placed in the MARC AUX IG directory export MARC_REFIND The MARC_REFIND variable indicates the path of the directory containing the file refind dat related to the refractive indexes of water and ice used in the MARC retrieval As an example export MARC_ REFIND AUX REF_IND indicates that the refractive indexes file will be placed in the MARC AUX REF_IND directory export VERSION The VERSION variable indicates the version of the code MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 ji Cc ser Date 10 04 07 Page n 76 169 line in run_marc file Related setting BIN marc gt log_marc txt Th
62. Z Spectroscopic database imaster my2 lt trom MASTER study ID amp ji Spectroscopic dspect_sideband_B dat database for gt eee sideband B C D amp D_SPECT dspect_sideband_C dat Q Selection from E HITRAN by 2 ys z MARSCHALS dspect_sideband_D dat Team y Isotopic_ratio dat Isotopes data i i Files with continuum i CONT cont_band_ dat gt profile for band i Pol Irregular i marschalsgrd trequeney arid i Files with VCM of i ERROR SPECTRA dat Forward Model data H et legend file content Figure 5 1 The default arrangement in subdirectories of the MARSCHALS L2 MARC files as specified in the delivered run file MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 78 169 5 3 The files needed to run the MARC code To run the MARC code some files of different kind are requested These files are listed and classified in The data file needed to run the MARC code File classification Name of files f source files and inc include files standard FORTRAN 0 source files language files The whole list is reported in RD 3 makefile makefile to compile the MARC source files see 5 1 for the MARC compilation procedure and RD batch file 3 for the file format run_marc batch file to run the MARC code see 5 3 2 and RD 3 for the file for
63. a _ x Max z lt t x Stier Dai where Sio __ represents the error associated with the element of state vector x J J at iteration iter The consideration that may prevent the unconditional use of the above iter expression instead of that of the 2 criterion is that S does not really represent p gt y rep iter the error of the element x when the retrieval is still far from convergence x i 2 ae convergence criteria relative difference between WA and linear y The difference between the current chi square y x and the chi square computed in the linear approximation Yun is less than a fixed threshold t4 x x Ki x lt t pA xi 4 Where Yun is computed using the expression I KG y F b x s I KG y F b x oh convergence criteria Marquardt parameter The value of Marquardt parameter Am have to be less than a fixed threshold ts for each retrieved parameter Avic ts i 5 4 1 Convergence is reached only if each Marquardt parameter is less than the set threshold Combination of the convergence criteria The overall convergence criterion is obtained as a suitable combination of the above conditions In some case the convergence criteria can be satisfied also when the iterative process is still far from reaching convergence In order to avoid this problem if the computing time of an extra iteration does not pose problems the requirement of a doub
64. a warning message and a copy of the input file that triggered the warning General options NAMELIST amp INPUT WAVELENGTH LIMITS FILTER FUNCTION SPECIFICATION NF SOLAR SPECTRUM SELECTOR 2 use TOA solar irradiance read from CKTAU file when kdist 1 NF 2 is not a valid input when kdist ne 1 1 read from file solar dat user supplied data file solar dat is read from the current working directory This ASCII file is read with the following free format read statements read 13 end 100 wlsun i sun i i 1 5000 100 continue where wlsun wavelength sample points microns sun direct normal solar irradiance at the top of the atmosphere W m2 micron The number of wavelength sample points read from solar dat should be less than or equal to 5000 This file format is new for the SBDART_2002 version Previous versions of SBDART used a different format for spectral input files albedo dat filter dat and solar dat A perl script newform is available from ftp ftp icess ucsb edu pub esrg sbdart to convert old data files to the new format Issue 3 Revision 1 Page n 146 169 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 User Manual Date 10 04 07 Page n 147 169 spectrally uniform 5s solar spectrum 0 005 micron resolution 25 to 4 micron LOWTRAN_7 solar spectrum default 20 cm 1 resolution 0 to 28780 cm 1 10 cm 1 resolution 28780 to 57490 cm 1
65. ach line of sight LOS the following data are listed in the file the index of LOS the altitude of the instrument the tangent altitude the latitude and longitude of the instruments the latitude and longitude of the tangent point the zenithal pointing angle the index of the line of sight that shares the calibration and finally the measured spectral points are reported in the file for each spectral point the following information are extracted from L1B frequency GHz spectrum intensity K error K gain gain error offset K offset error K gain offset covariance spectral data covariance Please notices that he pre processed data are constituted by the L1b measured spectra rearranged by the pre processor according to a format suitable for the MARC code As a matter of fact the pre processor produces a set of files named observation_ dat see RD 3 by varying the scan number At the present if the sequential fit is not set the MARC code performs the retrieval procedure on a single scan i e on data contained in one file of the set composed by the observation_ dat files once the scan is chosen the related observation_ dat file has to be renamed by the user as observ dat since MARC code in the standard configuration reads the measurement data from such a file MARSCHALS Level 2 TT Cc User Manual This page is intentionally left blank Prog Doc N IFAC GA 2007 06 LS Issu
66. age n 153 169 MARSCHALS Level 2 the cloud and ramps up to 9um at 15km However beware that the actual location of the cloud layers is determined by the resolution and placement of vertical grid points in SBDART as explained below SBDART puts the i th cloud layer at the highest vertical grid point k such that z k le abs ZCLOUD i 001 NOTE A cloud with a nominal altitude equal to that of one of the computational layer altitudes Z K actually extends from Z k to the next higher grid point For example a cloud layer at Z k will not affect the direct beam flux at Z k 1 one layer above but will strongly affect it at 2 k You can check this out your self by setting IOUT 10 and ZCLOUD 1 and messing around with ZOUT to get outputs just above or below the cloud Suppose the bottom of your computational grid looks like k z k 30 2 31 2 32 I 33 ie 34 0 35 0 If you want a cloud to eztend from 0 5 to 1 5 km then set ZCLOUD 5 1 5 Actually the same result would be obtained by setting the second element of ZCLOUD to anything between 1 0 tepsilon and 1 5 Consider TCLOUD 6 0 0 12 0 0 0 ZCLOUD 1 0 6 0 4 0 9 0 D Here two overlapping cloud decks are specified one extending from 1 to 6 km with a total optical thickness of 6 and the other from 4 to 9 km with a total thickness of 10 Since the total optical thickness is spread over the total altitude range we would have 1 optical depth p
67. al depth by l ssa set ssa 0 where ssa single scattering albedo and asym asymmetry factor nosct does not affect the stratospheric aerosol models or the iaer 6 boundary layer model Prog Doc N IFAC GA 2007 06 LS Page n 149 169 ISALB ALBCON SC IDATM MARSCHALS Level 2 Issue 3 Revision 1 User Manual Date 10 04 07 SURFACE REFLECTANCE PROPERTIES SURFACE ALBEDO FEATURE 1 spectral surface albedo read from albedo dat 0 user specified spectrally uniform albedo set with ALBCON 1 snow 2 alear water 3 lake water 4 sea water 5 sand 6 vegetation 7 sunglint off of ocean wind speed set with ALBCON 10 combination of snow seawater sand and vegetation NOTE If ISALB 1 a user supplied spectral reflectance file albedo dat is read from the current working directory This ASCII file is read with the following free format read numbers may be separated by spaces commas or carriage returns read 13 end 100 wlalb i alb i i 1 huge 0 100 continue where wlalb wavelength sample points microns alb spectral albedo unitless The number of wavelength sample points read from albedo dat should be less than or equal to 1000 The user specified reflectance may cover any wavelength range and have arbitrarily high resolution This contrasts with the standard reflectance models sand vegetation lake water and sea water which are only specified in in the range 25 to 4 um at 5
68. alue of WLINC is ignored and the output will be in units of W m2 um for irradiance and W m2 um sr for radiance This parameter specifies the spectral resolution of the SBDART run Though the spectral limits of the calculation are always input in terms of wavelength the spectral step size can be specified in terms of constant increments of wavelength log wavelength same as constant increment of log wavenumber or wavenumber Which one to choose depends on where in the spectral bandpass you want to place the most resolution Since SBDART is based on LOWTRAN7 band models which have a spectral resolution of 20 cm 1 it would be extreme overkill to allow spectral step size less than 1 cm l On the other hand a spectral resolution coarser than 1 um is also pretty useless Therefore the way WLINC is interpreted depends on whether it is less than zero between zero and one or greater than 1 WLINC 0 the default gt wavelength increment is equal to 0 005 um or 1 10 the wavelength range which ever is smaller If the WLINF WLSUP then WLINC 001 WLINC lt 0 gt wavelength increment is a constant fraction of the current wavelength WLINC is interpreted as a specified value of delta lambda lambda and the wavelength steps are adjusted so that wavelength step is approximately the product of the current wavelength and WLINC Specifying the wavelength increment as a fractional step size is useful when the wavelength range
69. and temperature profiles The D_SPECT directory contains the file imaster my2 with the spectroscopic database coming from the MASTER study the dspect_sideband_B dat dspect_sideband_C dat and dspect_sideband_D dat files with the spectroscopic database from HITRAN to be used for the sideband B sideband C and sideband D respectively the isotopic_ratio dat with data about the isotopes The CONT directory contains three files cont band B cont band C and cont band D with the continuum profiles for the three bands MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 34 169 The JG directory contains the marschals grd file with the irregular frequency grid The ERROR_SPECTRA directory contains a set of files with information about error spectra data Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 es M l Issue 3 Revision 1 i Cc User Manua Date 10 04 07 Page n 35 169 3 1 3 The default arrangement of the OFM files The default arrangement of the OFM directory containing the files needed to run the OFM codes is depicted in Figure 3 4 BIN Executable file of SBDART code Executable file of RTM code Source files of SBDART code a files Source SBDART_SOURCE Batch file for the compilation of SBDART code Source files of RTM code files Source
70. ated quantity Displacement of the Reference Vertical with respect to the vertical Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 100 169 data type E ZERE of the instrument in the sense of the degres tal line of sight Switch T F for the use of the Mie Scattering Source function MSSF lmssf dimless logical T MSSF data are used F gt MSSF data are not used This setting is considered also when the MARC code works in Forward modality MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 101 169 5 4 7 How to set up the retrieval modality and techniques The MARC code allows performing the retrieval procedure by using different techniques and modalities in addiction to the Marquardt iteration techniques also Regularization or an Optimal Estimation technique can be set The user can specify the use of these techniques by defining the values related to the following key words in the setting file Key words related quantity im data type Switch to select the retrieval mode available range 1 2 3 1 Marquardt only 2 Regularization Marquardt imode 3 Optimal Estimation dimless integer 4 Marquardt If option 2 is chosen the set up of the regularization has to be performed by defining the related parameters Then the chosen modalities have
71. ay information about all the MARC retrieved quantities For each retrieved target the visualization tool automatically produces a picture containing the retrieved profile its Expected Standard Deviation and the related quality parameters and averaging kernel An example of such a picture is depicted in Figure 7 2 The aforementioned pictures are saved as a files named target _col TIFF in the same directory of the considered MARC output data These pictures are also directly shown in the default output i e generally the monitor of the computer T Campaign 20051205 Sequence 51 time 80963 8 UTC lat 16 40 lon 131 19 alt 19 TEMPERATURE MTR T H O O HNO POINTING CONT_BAND_C OFFS BAND C GAIN BAND C Chi Square 1 52 Trace 30 67 Info Content 85 18 Tot Iter 1 1 M A t 0 100 0 100 Profile SDs 30 aw Biased Err Unbiased Eyr Biased Erro 20 Nh o Altitude km 5 Altitude km 180 200 220 240 260 280 0 1 2 3 4 Quality Pdfameters EEE TT TEFEN Altitude km Altitude km 0 0 0 5 1 0 1 5 2 0 0 5 0 0 0 5 1 0 1 5 Nr ReRatin 10 62 26 0 22 0 20 0 19 0 17 0 15 0 13 0 11 0 9 0 7 0 km Figure 7 2 The output of the visualization tool MARSCHALS Level 2 TT Cc User Manual This page is intentionally left blank Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 144 169 Date 10 04 07 MARSCHALS Level 2 Pro
72. ber lays in the range from scan number 1 to scan number 2 This array defines the numerical range of the scans to be extracted This field is read only if extraction criterion This array defines the coordinate range of the scan to be extracted The SAMM Pre Processor will extract those scans for which aircraft coordinates lay in the indicated range This field is read only if extraction _criterion 2 dimless degrees 2xint 4xdouble BAND_SELECTION selector_first_band selector_second_band selector_third_band This array defines the band to be considered The Pre Processor will select those scans whose measurements are performed inside the indicated band All other measurements are discarded All band combinations are possible the combination is specified by three logical numbers being the first one referred to the first band the second to the second band and the third to the third band The aforementioned order is that used in the L1B data file please notices that the first band could not be the band B If the logical number is set to one the related band is considered otherwise it is not selected dimless 3xint Key words MARSCHALS Level 2 User Manual related quantity Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 Page n 55 169 ILOS RANGE SELECTION altitude los 1 altitude los 2 This arra
73. code 10 11 12 13 15 15 19 40 41 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 l c User Manua Date 10 04 07 Page n 6 169 4 3 The files needed to run the SAMM code 42 4 3 1 The source files 44 4 3 2 The batch file 45 4 3 3 The executable file 46 4 3 4 The measurement files 47 4 3 5 The ancillary data files 48 4 3 6 The auxiliary data files 49 4 3 7 The settings file 50 4 4 How to set the SAMM pre processing 51 4 41 How to set the input output namefiles 52 4 4 2 How to define the quantities to be extracted from the LIb data files 53 4 4 3 How to select the files generated by the pre processor 56 444 How to enable the output of files containing instrumental information 58 4 4 55 SAMM extra features 60 4 4 6 The contrast level 61 4 5 Output files 63 4 5 1 The flight_view dat file with the general information about the flight campaign 64 4 5 2 The scan_ _view dat file with the general information about the selected scan 65 4 5 3 The OCM 3 dat file with the OCM images 66 4 5 4 The observation dat files with the measured data 69 5 MARC CODE 71 5 1 How to compile the source files of the MARC code 72 5 2 How to run MARC code 73 5 2 1 The run_marc file 74 5 3 The files needed to run the MARC code 78 5 3 1 The source files 80 5 3 2 The batch files 81 5 3 3 The executable file 82 5 3 4 The pre processed data files 83 5 3 5 The ancillary data files 84 5 3 6 The auxiliary data
74. cutable file of he SAMM code LIB_MARSCHALS_file measurement files the name of this file changes according to an internal MARSCHALS LIB syntax ancillary data files L1B_instrument lut file The pre processor can be used to generate the files fov dat ils dat instrument dat and rejection dat containing the instrument characterizations needed by the retrieval code auxiliary data files In this case the following Llb auxiliary files are needed e files containing spectral response for each channel and each band e files containing the fov acap for each band settings file settings samm dat 1 These files are needed to create the samm ezecutable code anyway a samm executable file has been already created and installed in the delivered package thus these files are not requested to run the provided default version of the SAMM code 2 Actually this file is not mandatory to run the code since it is only a batch file to run in an easier way the related codes Table 4 1 The data file needed to run the SAMM code MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 Cc ser Date 10 04 07 Page n 43 169 The aforementioned files have to be stored in the hard disk of the computer devoted to run the SAMM code according to a given files arrangement A default arrangement suggested by the MARSCHALS L2 team is described in paragraph 3 1 The computer delivered
75. des of the MARSCHALS L2 suite The user can change the default arrangement of the files needed to run the MARSCHALS L2 software by changing the name of the subdirectories see 3 1 4 According to this default arrangement depicted in Figure 3 1 the files of the SAMM the MARC and the OSM codes have to be stored on the hard disk in three dedicated subdirectory named respectively SAMM MARC and OFM placed in a main MARSCHALS L2 directory named MARSCHALS_L2 MARSCHALS_L2 Visualization tools code files Figure 3 1 The overall default arrangement in three main subdirectories of the files composing the MARSCHALS L2 suite in blue the subdirectories For each of the three proposed subdirectories a default arrangement of the files in a tree structure subdirectories is proposed as reported in the following chapter 3 1 1 SAMM code 3 1 2 MARC code and 3 1 3 OFM codes Please notice that a further IDL subdirectory is present This subdirectory contains the IDL code files of a visualization tool that has been developed in the context of the MARSCHALS project by the MARSCHALS L2 team to make easier the presentation and the analysis of the retrieval results obtained by the MARSCHALS software This tool is provided as it is as an IDL code and requires an IDL licence in the computer in which it is executed The description of this tools is provided in Chapter 7 Prog Doc N IFAC GA 2007 06 LS MARSCHALS Lev
76. e 3 Revision 1 Date 10 04 07 Page n 70 169 MARSCHALS Level 2 Prog Doe N IFAC_G A 2007 06 LS User Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 71 169 5 MARC code This chapter aims to describe the use of the MARC code with its implemented features The MARC code is the core of the MARSCHALS suite and it constitute the retrieval code that performs the retrieval of the specified minor atmospheric constituents from the MARSCHALS L2 previously rearranged in a suitable way by the SAMM code The retrieval is implemented as an iterative procedure based on a Forward Model and together with the MARSCHALS LIb data it requires also some auxiliary and ancillary data The retrieval is characterized by a high flexibility and thus many parameters can be user defined and some feature can be adopted or not according to the user defined settings The marc code is flexible and an exhaustive user defined setting of the retrieval is possible and easily achieved according to user needs and measurement data quality A default setting is in any case provided by means of a suitable setting file In the following paragraphs the input files necessary for MARC retrieval execution see 5 3 the output files resulting from the Pre Processing procedure see 5 5 a list of modalities and functionalities of the code and some indications on the available user defined choices see 5 4 will be addressed together with the instructio
77. e 133 6 4 Settings file of RTM code 134 6 41 How to define the altitude of the receiver 134 6 4 2 How to define the pointing angles 135 6 4 3 How to define the receiver field of view 136 6 4 4 How to define the latitude of the observations 137 6 4 5 How to insert an identifier of the computation 138 6 5 Output files 139 6 5 1 Output of RTM code 139 6 5 2 Log file 140 7 THE VISUALIZATION TOOL 141 7 1 How to run the visualization tool 142 7 2 The visualization tool output 143 8 ANNEX 1 SBDART DOCUMENTATION 145 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 8 169 This page is intentionally left blank MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 l c User Manua Date 10 04 07 Page n 9 169 1 Introduction 1 1 Purpose The MARSCHALS L2 software is a suite of three codes developed for a Linux platform and devoted to the retrieval of the atmospheric constituents from a set of spectroscopic measurements coming from the Level 1b L1b of MARSCHALS project Together with the retrieved profiles also some auxiliary results are provided by the MARSCHALS Level 2 L2 software The aim of this document is to provide the information needed to install and use the software MARSCHALS Level 2 software that has been developed in the framework of MARSCHALS L2 Project Please notices that the MARSCHALS Level 2 software is a suite
78. e file to be used to compile the aforementioned source file The WORKING_DIR directory contains o the run_samm script file to be used to run the SAMM code o the INP_FILES subdirectory o the OUT_FILES subdirectory The JNP_FILES directory contains the settings_samm file to be used for the set up of the preprocessing procedure see 4 3 7 the Lib MARSCHALS file with the data coming from the L1b measurement the Jl b_instrument lut ASCII file containing some instrument characterization needed to the pre processing operations The OUT_FILES directory contains the flight view dat ASCII file reporting general information about the flight campaign under consideration see 4 5 1 the scan view dat ASCII file reporting general information about the scan see 4 5 2 the OCM_ dat ASCII file containing OCM images and additional information about a given scan in the given campaign see 4 5 3 the observations_ dat ASCII file containing all measurement data referred to the scan it is used as input of MARC Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 31 169 3 1 2 The default arrangement of the MARC files MARC T a g Executable file legend A mm fr Directory 4 f inc Source files of MARC code kama gt MARC SOURCE maretie Batch file for the compilation file Mathema
79. e in the model atmosphere Aside from multiplicative factors the vertical profile will be that of the original model atmosphere set by IDATM The original unmodified density profile is used when UW or UO3 is negative integrated ozone concentration ATM CM in troposphere i e for z 1t ZTRP The original tropospheric density is used when O3TRP is negative default 1l The altitude of the tropopause The parameters UO3 and O3TRP sets the total column ozone in the stratosphere and troposphere respectively Note since the default value of ZTRP is zero UO3 normally sets the integrated ozone amount of the entire atmosphere default 0 volume mixing ratio of N2 PPM default 781000 00 volume mixing ratio of 02 PPM default 209000 00 volume mixing ratio of CO2 PPM default 360 00 volume mixing ratio of CH4 PPM default 1 74 volume mixing ratio of N20 PPM default 0 32 volume mixing ratio of CO PPM default 0 15 volume mixing ratio of NH3 PPM default 5 0e 4 volume mixing ratio of S02 PPM default 3 0e 4 volume mixing ratio of NO PPM default 3 0e 4 volume mixing ratio of HNO3 PPM default 5 0e 5 volume mixing ratio of NO2 PPM default 2 3e 5 NOTE Setting any of these factors to 1 causes that atmospheric component to retain its nominal Prog Doc N IFAC GA 2007 06 LS Page n 151 169 XRSC x04 PBAR ZPRES SCLH20 ZCLOUD MARSCHALS Level 2 Prog Doc
80. e list goes on with the species whose gradient is not to be retrieved listed according to the order specified in codefitVMR Trivially if any gradient is not retrieved gvmrfit F the order is that specified by the user with codefitVMR on the contrary if for all the retrieved specie both the VMR and the gradient are retrieved the order is the one specified in by codefitGRAD As an example of a generic case if codefitVMR 12 3 1 5 4 codefitGRAD 1 3 the resulting order will be 5 4 5 How to set the retrieval grid MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 User Manual Date 10 04 07 Page n 95 169 The MARC retrieval can be performed at the tangent altitudes or on a user defined altitude grid In the last case the user has to specify an overall altitude grid that is common for all the retrieved quantities then a dedicated retrieval grid can be specified for each retrieved quantity as a subset of the common grid previously defined by specifying this subset with a mask array The setting of the retrieval grid is made by set the following key words Key words related quantity data type ltangaltret Switch F T to specify the choice of the retrieval altitude grid T retrieval at tangent altitude F retrieval at user defined altitude grid Please notice in Forward Model option when MARC is used in Forward Modality the ltangaltret switch
81. e toy D in band D dimless max a Note that this field is read only if 200 char CREATE FOV_DAT MARSCHALS Level 2 Prog Doe N IF AC GA 2007 06 LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 60 169 4 4 5 SAMM extra features To allow during the first part of the code development the testing of MARC code in presence of temporary input data characterized by some problems in the reported measurements a series of extra feature of the SAMM preprocessor code have been developed such as manual selection of the LOS the user can exclude one or more LOS from the file of observations this manual selection has been added to exclude LOS having large values of chi square computation of the average spectral error for each LOS the preprocessor computes an average rms value for the spectral error the user can decide to use the original error or the averaged one in the retrieval treatment of spectra containing un realistic values some spectra in band B contains channels having un realistic values NULL or very large values in spectral data and or in spectral data error when the file of observations is produced these channels can be replaced with values selected by the user This feature of the preprocessor allows to exclude the un realistic values by introducing a large value in the spectral error The channels marked as un realistic are not taken into account when the average spectral
82. edicated output directory as default this directory is named OUT_FILES Each output file is characterized by a string that gives information about the files used as input in the run of the code by which the output file has been produced see 5 5 1 The set of output files change according to the settings The whole set of output files can be arranged in some classes The retrieved quantities files The VCM files The files with the data used to run the retrieval procedure The results of the performed retrieval procedure In the following the aforementioned output files are introduced a detailed description is presented in RD 3 Standard output data when retrieval is performed without extra features If the code is used as retrieval code without any extra features the produced files are these described in Table 5 5 The retrieved profiles are saved in a set of files one for each retrieved species named target dat being target the chemical symbol of the related molecule e g h20 dat will contain the retrieved profile of the water The spectrum dat file contains the simulated measured spectrum i e the foreseen of the spectrum that will be measured related to the aforementioned retrieved profiles this file also reports the simulated spectrum computed using the adopted initial guess atmosphere and the original measured spectrum The iterationdetails dat file reports information about the details of the iterations of the retrieva
83. el 2 es M l Issue 3 Revision 1 i Cc User Manua Date 10 04 07 Page n 29 169 3 1 1 The default arrangement of the SAMM files The default arrangement of the SAMM directory containing the files needed to run the SAMM code is depicted in Figure 3 2 BIN ant Executable file ji m Eo PAA 5 Source files Eo of SAMM code eee makefile file for the compilation of SAMM cod Script file for run samm gt the running of SAMM code J files Source SAMM_SOURCE no D f sertings_samm dat User defined 8 settings ep WORKING_DIR INP FILES L1B o tie MARSCHALS ft MARSCHALS amp S measurement I Os L1B 2 m instrument data 5 L1B OUT_FILES lt gt preprocessed iy data legend Figure 3 2 The arrangement of the SAMM code data and source files The names of the subdirectories are those of the default arrangement MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 30 169 The SAMM directory presents three subdirectories e the BIN subdirectory e the SAMM_SOURCE subdirectory e the WORKING_DIR subdirectory The BIN directory contains o the samm executable files of the SAMM code The SAMM_SOURCE directory contains o the source files of the SAMM code o the makefil
84. ent of dimless logical one species is retrieved F the VMR gradients are not retrieved Number of species whose gradient is retrieved NfitGRAD Note that the Nf hi tGRAD value is read dimless integer 4 only if lgvmrfit T If lvmrfit is true it has to be NfitGRAD lt NfitVMR Array of integers containing the MARSCHALS MASTER code of the species whose gradients have to be retrieved Note that only the first Nfi tGRAD codes i of the arrays are read The HITRAN code nfitgrad x defitGRAD i pone is that specified in Table 5 4 dimiless integer If lvmrfit is true codefitGRAD has to be a subset of codefitVMR Retrieval tests have been performed for gradient of O and gradient of H2O see RD 4 Key words MARSCHALS Level 2 User Manual related quantity Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 93 169 data type lpoifit Switch logical value F T for pointing bias retrieval T retrieval of pointing bias is performed F retrieval of pointing bias is not performed dimless logical ig_pointing Initial guess of pointing angle bias deg real lexcfit Switch logical value F T for external continuum retrieval T gt retrieval of un accounted continuum is performed F gt retrieval of un accounted continuum is not performed dimless logical loffsfit Switch for instrumen
85. er Manual Date 10 04 07 If IDATM 0 a user supplied atmospheric profile atms dat is read from the current working directory This ASCII file is read with the following free format read statements input values may be separated by spaces commas or carriage returns read 13 nn do 10 i 1 nn read 13 z i p i t i wh 1 wo i 10 continue where nn is the number atmospheric layers nn should be less or equal to than MXLY a parameter used in SBDART see params f to set the maximum number of levels in the vertical grids z is the layer altitude in km z must be monotonically decreasing p is the pressure in millibars t is the temperature is Kelvin wh water vapor density g m3 wo ozone density g m3 If IDATM is set to a negative number in the range 1 to 6 SBDART prints the atmospheric model corresponding to abs idatm to standard out and then quits weighting factor when positive this factor controls how much of the atms dat atmospheric profile to mix in with one of the standard internal profiles selected by IDATM For example IDATM 1 and AMIX 7 specifies a 70 weighting of atms dat and a 30 weighting of profile TROPIC No default 1 integrated water vapor amount G CM2 integrated ozone concentration ATM CM above the level ZTRP The default value of ZTRP 0 so UO3 usally specifies the total ozone column 1 atm cm 1000 Dobson Units Use UW or UO3 to set the integrated amounts of water vapor or ozon
86. er km for the lower cloud deck and 2 optical depths per km for the second The code adds the effects of both cloud decks in the region of overlap So the above specification would yield 1 optical depth per km between 1 and 4 km 3 optical depths per km between 4 and 6 km and 2 optical depths per km between 4 and 9 km for a total optical depth of 18 NOTE When a ZCLOUD range is being specified i e a negative value is used to set the upper end of a range the opacity appears high altitudes in the range layers are set the opacity altitude to the altitude of computational level Thus levels are at 1km intervals following inputs do exactly TCLOUD 2 0 ZCLOUD 1 3 TCLOUD 1 1 ZCLOUD 1 2 only between the low and When individual cloud extends from the named the next higher if the computational starting from zero the the same thing If you have any doubt about where the code is putting the cloud set IDB 8 1 see below print out of cloud optical depth to get a diagnostic TCLOUD MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 User Manual NOTE do not try to put an ice cloud NRE lt 0 ina cloud layer range which includes water cloud 2 le NRE le 128 In other words this specification won t work ZCLOUD 1 4 TCLOUD 1 0 NRE 8 1 Optical thickness of cloud layer up to 5 values TCLOUD specifies the cloud optical depth at a wa
87. er the call to subroutine ATMS This allows regridding of the standard internal atmospheres as well as user specified atmospheres read with IDATM 0 No matter how many grid points were used to specify the original Page n 162 169 IDB ZOUT IOUT value 0 Ta MARSCHALS Level 2 Issue 3 Revision 1 User Manual Date 10 04 07 atmosphere the new regridded atmosphere will contain NGRID vertical array elements The default value of ZGRID1 and ZGRID2 are set to 1 and 30km respectively The default value of NGRID 0 causes the initial un modified atmsopheric model to be used The internal parameter MXLY sets the maximum number of levels allowed Setting NGRID gt MXLY causes NGRID to be set to MXLY If NGRID is negative SBDART terminates execution after printing out the regridded values of Z P T WH WO to standard out This option can be used to preview the effect of a given set of ZGRID1 ZGRID2 and abs NGRID values OUTPUT OPTIONS DIAGNOSTIC OUTPUT SELECTOR integer array The IDB print flag is used to select print diagnostics for a variety of computational parameters Setting IDB n m where m is any non zero integer will produce the diagnostics associated with array index n to be listed For some values of n increasing the value of m e g IDB 8 2 will produce more detailed diagnostics default 0 array index 1 print an explanation of quantities in IOUT output group 2 cloud liquid water profiles
88. ervations Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 User Manual Issue 3 Revision 1 Cc ser Date 10 04 07 Page n 4 169 This page is intentionally left blank Page n 5 169 MARSCHALS Level 2 Prog Doc N IF AC GA 2007 06 LS M l Issue 3 Revision 1 q Cc User Manua Date 10 04 07 TABLE OF CONTENTS 1 INTRODUCTION 11 Purpose 1 2 How to read this document 1 3 Relevant Document 1 4 Acronyms 1 5 Conventions 2 OVERVIEW OF MARSCHALS L2 SOFTWARE 2 1 Software modules and codes 2 2 The used programming language 2 3 Hardware and system requirement 2 4 The MARSCHALS L2 software files 2 4 1 The code files 2 4 1 1 The source files 2 4 1 2 The batch files 2 4 1 3 The executable files 2 4 2 The data files 2 4 2 1 The settings files 2 4 2 2 The L1b measurement files 2 4 2 3 The pre processed data files 2 4 2 4 The ancillary data files 2 4 2 5 The auxiliary data files 2 4 2 6 The output data files 2 4 2 7 The I O files for the sequential fit 2 5 The key word syntax of the files with the user defined settings 3 INSTALLATION OF THE MARSCHALS L2 CODES 3 1 The default arrangement of the files 3 1 1 The default arrangement of the SAMMI files 3 1 2 The default arrangement of the MARC files 3 1 3 The default arrangement of the OFM files 3 1 4 How to change the default names of the subdirectories 4 SAMM CODE 4 1 How to compile the source files of the SAMM code 4 2 How to run SAMM
89. essor code is able to produce a set of output files listed in Table 4 2 together with the related contents and the definition of the special character used inside the filenames This paragraph is devoted to the introduction of these output files Pre processor product ASCII file reporting general information about the SBT View dar flight campaign under consideration see 4 5 1 ASCII file reporting general information about the scan_ _view dat scan see 4 5 2 ASCII file containing OCM images and additional OCM_ dat information about a given scan in the given campaign see 4 5 3 ASCII file containing all measurement data referred to OBS GHOTS Qda the scan it is used as input of MARC see 4 5 4 where Identifier of the scan number of the scan Pre processor auxiliary product ASCII file containing some instrumental parameters it TITHE dar is used as input of MARC ancillary file ASCII file containing some instrumental parameters it rejection dat is used as input of MARC ancillary file ASCII file containing the frequency response of the ils dat channels it is used as input of MARC ancillary file ASCII file containing the antenna Field of View it is fov dat used as input of MARC ancillary file Table 4 2 The output files generated by the SAMM preprocessor MARSCHALS Level 2 Prog Doc N IF
90. extends over more than an decade of wavelength For example if the wavelength range is 0 5 to 20 0 specifying a constant wavelength increment of 01 microns tends to under resolve the low wavelengths and over resolve the long wavelengths Setting WLINC 01 causes the code to use a wavelength increment of about 005 microns in the visible and about 2 micron in the thermal infrared which is a better compromise of resolution and computer time Page n 148 169 SZA CSZA IDAY TIME ALAT ALON SOLFAC NOSCT MARSCHALS Level 2 Issue 3 Revision 1 User Manual Date 10 04 07 1 gt WLINC gt 0 gt WLINC is the wavelength step size um if WLINC gt 1 then WLINC is the step size in inverse centimeters If maximum fidelity is required and gaseous absorption is the primary influence on the output then WLINC should be set to 20 which is the wavenumber resolution of the LOWTRAN7 band models The total number of wavelength steps nwl is given by nwl 1 1n wlsup wlinf wlinc wlinc lt 0 nwl 1 wlsup wlinf wlinc 1 gt wlinc gt 0 nwl 1 10000 1 wlinf 1 wlsup wlinc wlinc gt 1 SOLAR GEOMETRY solar zenith angle degrees default 0 Solar input may be turned off by setting sza gt 90 SZA is ignored if CSZA is non negative or IDAY is non zero Cosine of solar zenith angle If CSZA gt 0 solar zenith angle is set to acos CSZA default 1 If IDAY gt 0 the solar illumi
91. f user azimuth angles If this parameter is specified SBDART will output radiance values at NPHI azimuth angles evenly spaced between the first two values of input array PHI For example nphi 7 phi 0 180 will cause output at zenith angles 0 30 60 90 120 150 180 User relative azimuth angles If NPHI is specified then PHI is interpreted as the limits of the azimuth angle range and only the first two elements are required If NPHI is not specified then up to NSTR values of PHI may be specified If neither NPHI nor PHI is specified and a radiance calculation is requested IOUT 5 6 20 21 22 23 a default set of azimuth angles is used equivalent to the case NPHI 19 PHI 0 180 NOTE Azimuth increases clockwise looking down on the Earth s surface PHI is the relative azimuth angle from the forward scattering direction PHI lt 90 gt forward scattered radiation PHI gt 90 gt backward scattered radiation For example if the sun is setting in the West radiation propagating to the South East has a relative azimuth of 45 degrees NOTE SBDART is currently configured to model radiation with at most 40 computational zenith angles and 40 azimuthal modes While these limits may be expanded be aware that running SBDART with a much larger number will significantly increase running time and memory requirements In tests performed on a DEC Alpha the execution time scaled roughly with NSTR 2 for NSTR less than 40 The code s memory
92. ffuse W m2 fxup upward flux W m2 fxdir downward flux direct beam only W m2 dfdz radiant energy flux divergence mW m3 heat heating rate K day NOTE dfdz i and heat i are defined at the layer centers i e halfway between level i 1 and level i radiance output at ZOUT 2 km Output format write wlinf wlsup ffew topdn topup topdir amp botdn botup botdir write nphi nzen write phi i i 1 nphi write uzen j j 1 nzen write r i j i 1 nphi j 1 nzen The first record of output is the same as format IOUT 10 WLINF WLSUP FFEW TOPDN TOPUP TOPDIR BOTDN BOTUP BOTDIR addition records contain NPHI number of user azimuth angles NZEN number of user zenith angles PHI user relative azimuth angles nphi values UZEN user zenith angles nzen values R radiance array nphi nzen W m2 sr NOTE if IDAY is set then PHI is the actual compass direction in which the radiation in propagating same as IOUT 20 except radiance output at ZOUT 1 km radiance and flux at each atmospheric layer integrated over wavelength Output format write nphi nzen nz ffew write phi i i 1 nphi write uzen j j 1 nzen write z k k nz 1 1 write fxdn k k 1 nz write fxup k k 1 nz write fxdir k k 1 nz write uurl i j k i 1 nphi j 1 nzen k 1 nz where nphi number of user specified azimuth angles nzen numbe
93. files of the MARSCHALS L2 suite They are referred to the provided test data if the codes of the MARSCHALS L2 suite is run with different measurement data the provided ancillary data could not be adequate to the new measurement scenario In this case a new and suitable set of ancillary data files have to be provided by the user 2 4 2 5 The auxiliary data files Some auxiliary files are needed to run MARC retrieval code of the MARSCHALS L2 suite These files are not provided by the pre processor and have to be provided by the user These files are described in RD 3 Please notice that a set of auxiliary data files are provided together with the other files of the MARSCHALS L2 suite They are referred to the provided test data if the codes of the MARSCHALS L2 suite is run with different measurement data the provided auxiliary data could not be adequate to the new measurement scenario In this case a new and suitable set of auxiliary data files have to be provided by the user 2 4 2 6 The output data files The results of the retrieval procedure are stored in some output data files the format of these files are described in RD 3 2 4 2 7 The I O files for the sequential fit If a sequential fit is adopted the output data to be used as input data of the following step are stored in some I O data files These files are described in RD 3 MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision
94. first nretrievalgrid codes in the input line are read e only N itVMR lines corresponding to the fitted VMR are read Key words related quantity MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS M l Issue 3 Revision 1 User Manual Date 10 04 07 Page n 97 169 data type Matrix logical values F T with NfitGRAD rows one for each retrieved gradient in the rearranged order specified in 5 4 4 Each row is composed by nretrievalgrid flag one for each altitude of the retrieval grid specifying the mask for the retrieval of the gradient of the VMRs T the VMR gradient the row is related to is retrieved at the related lvmrgmask altitude dimless logical F the VMR gradient the row is related to is not retrieved at the related altitude Note e only the first nretrievalgrid codes in the input line are read e only Nf itGRAD lines corresponding to the fitted gradient are read Array logical values F T This row is composed by nretrievalgrid flag one for each altitude of the retrieval grid specifying the mask for the retrieval of the un accounted lomask continuum ddes nretrievalgrid T the un accounted x logical continuum is retrieved at the related altitude F the un accounted continuum is not retrieved at the related altitude This setting is considered also when the MARC code works in Forward modality Gene MARSCH
95. g Doc N IF AC GA 2007 06 LS Issue 3 Revision 1 q Cc User Manual 8 ANNEX 1 SBDART documentation Input documentation for SBDART summer 2002 release Paul Ricchiazzi Institute for Computational Earth System Science University of California Santa Barbara This file documents input parameters for SBDART Santa Barbara DISORT Atmospheric Radiative Transfer SBDART is a software tool that computes plane parallel radiative transfer in clear and cloudy conditions within the Earth s atmosphere and at the surface For a general description and review of the program please refer to Ricchiazzi et al 1998 Bulletin of the American Meteorological Society October 1998 SBDART s main input file is called INPUT This file contains a single NAMELIST input block also named INPUT A significant advantage of NAMELIST input is that not all elements of an input block need be specified by the user Since most of the code inputs have been initialized with reasonable default values new users can start by specifying just a few interesting input parameters The default state of input parameters may be determined by removing INPUT from the current working directory When SBDART detects the absense of file INPUT it will print the default settings of all input parameters This output may be redirected to a file for editing The default configuration of INPUT is as follows amp INPUT idatm 4 amix 0 0 isat 0 7 wlinf 0 550 P wl
96. g file MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 140 169 During the process RTM elaborates a log file in which messages about the processing are reported The content of the messages are self explicative and do not require further comments MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 141 169 7 The visualization tool The MARSCHALS suite of codes is provided with a visualization tool that can be used to visualize in a easier way the results of the retrieval perform by the MARSCHALS code The visualization tool is constituted by an IDL code It does not require any installation but an existing IDL license for the computer in which the MARSCHALS code is installed is needed to run the tool Please notice that this license is not provided by the MARSCHALS team for the delivered computer and must be arranged by the user In the default configuration the files of the visualization tools are placed in the subdirectory IDL of the main directory MARSCHALS L2 as depicted in Figure 7 1 MARSCHALS_L2 ce 1 rr a f isualization tools a __ IPL __ code files IA Figure 7 1 The overall default arrangement in three main subdirectories of the files composing the MARSCHALS L2 suite in blue the subdirectories MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06
97. ge whereas MARC and OFM codes are written in FORTRAN 90 About MARC and OFM codes the developers followed a given FORTRAN 90 compatible syntax in order to make possible the use of a tool for the automatic documentation of such codes The visualization tool is written in IDL language MARSCHALS Level 2 ee ere iA c User Manual Date 10 04 07 Page n 20 169 2 3 Hardware and system requirement The MARSCHALS L2 codes have been developed in order to run under a LINUX operating system The MARSCHALS L2 codes have been also tested on a Windows platform To run under a Microsoft Windows Operating System some changes have to be made in the sources of the code in order to take into account for the different system variables The description of these changes is out of the purpose of this document Some requirements have been identified for the computer devoted to run the MARSCHALS 12 codes The only mandatory requirement concerns the RAM memory that has to be not less then 2 Gbyte Such an amount of RAM is the maximum amount of memory that a Linux system on a 32 bit machine can efficiently manage The 2 Gbyte RAM should guarantee that paging system swapping on disk is minimized thus allowing for the maximum speed when running the code An important suggestion concerns the clock speed Due to the massive amount of computations such a speed has to be as high as possible gt 3 GHz The other hardware requirements disk CD writer ne
98. gs files described in RD 3 2 4 2 2 The Llb measurement files The files used as input of the MARSCHALS L2 software i e as input of the SAMM Pre Processor are the L1b data files coming from the MARSCHALS measurements These files listed in paragraph 4 3 4 are processed by the SAMM Pre Processor and thus they can be considered as input files of the overall MARSCHALS L2 software together with the auxiliary and ancillary data files 2 4 2 3 The pre processed data files At the present the retrieval procedure performed by MARC code processes the single scan separately Once the SAMM Pre Processor is executed the file observations_ dat related to the scan to be processed must be manually renamed as observ dat Since the pre processor can be run separately the observ dat file can be considered as the measurement input for MARC code The same procedure has to be followed for the OCM_ dat file related to the scan under consideration that has to be manually renamed as OCM dat MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 25 169 2 4 2 4 The ancillary data files The MARC retrieval code of the MARSCHALS software needs also some ancillary files related to the processed measure These files described in RD 3 contain information provided by level 1 Analysis Please notice that a set of ancillary data files are provided together with the other
99. he arrangement of the MARC code data and source files The names of the subdirectories are those of the default arrangement MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 32 169 The default arrangement of the MARC directory containing the files needed to run the MARC code is depicted in Figure 3 3 It is characterised by 5 subdirectories The BIN subdirectory the MATH subdirectory the MARC_SOURCE subdirectory the WORKING_DIR subdirectory the AUX subdirectory The BIN directory contains o the marc executable files of the MARC code The MATH directory contains o aset of mathematical library files The MARC_SOURCE directory contains o the source files of the MARC code o the makefile file to be used to compile the aforementioned source files The WORKING_DIR directory contains o the run_marc script file to be used to run the MARC code o the INP_FILES subdirectory o the OUT_FILES subdirectory The INP_FILES subdirectory contains the settings_marc dat file with the data to set the retrieval procedure the observ dat file with the pre processed MARSCHALS LIb data and four files with data about the instrument instrument dat file with general information about MARSCHALS instrument fov datfile with general information about the FOV of the MARSCHALS instrument rejection dat file with general information
100. he subdirectories in which the files needed to execute the SAMM code are stored If a change is made the in the default arrangement of the files as reported in 3 1 1 the batch file run_samm stored as a default in the SAMM WORKING_DIR subdirectory have to be changed accordingly by the user Please notice that SAMM code files are arranged in subdirectories These subdirectories are contained in a main directory named as a default SAMM that is user defined all the paths are thus defined in a relative way with respect to the work directory that is the SAMM WORKING_DIR one in the default configuration The run_samm environment file contains information about the name of the setting file to be considered MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 42 169 4 3 The files needed to run the SAMM code To run the SAMM code some files of different kind are requested These files are listed and classified in Table 4 1 The data file needed to run the SAMM code File classification Name of file c source files and h include files standard C language files The whole list is reported in RD 3 source files makefile makefile to compile the SAMM source files see 4 1 for the SAMM compilation batch file procedure and RD 3 for the format file run_samm batch file to run the SAMM code executable file samm exe
101. ic interpolation on the current model s atmosphere pressure and altitude arrays Changing ZPRES does not alter other parameters in the atmospheric model in any way Note that setting a large value of ZPRES may push the tropopause where dT dz 0 to an unrealistically high altitude Water vapor scale height in km If SCLH20 gt 0 then water vapor is vertically distributed as exp z SCLH20 If SCLH20 le 0 then the original vertical profile is used Changing SCLH20 has no effect on the total water vapor amount CLOUD PARAMETERS Altitude of cloud layers km up to 5 values Cloud layers may be specified in two ways To specify separate cloud layers set ZCLOUD to a sequence of monotonically increasing altitudes Each value of ZCLOUD will set the altitude above the surface of the corresponding optical depth in the TCLOUD array To specify a range of altitudes which will be filled by cloud tag the second element of the range with a minus sign Consider zcloud 1 3 10 15 tcloud 4 0 8 0 0 nre 6 6 8 9 10 In this example two continuous cloud layers are defined the lower one extends from 1 to 3 km and has a total optical depth of 4 and an effective radius of 6um The upper cloud layer extends from 10 to 15 km has a total optical thickness of 8 and a sliding value of effective radius which starts 8um at the bottom of Page n 152 169 Prog Doc N IFAC GA 2007 06 LS U M l Issue 3 Revision 1 ea Date 10 04 07 P
102. ied in 5 4 4 T regularization matrix is used in VMR gradient retrieval of the related species F regularization matrix is not used in VMR gradient retrieval of the related species dimless logical UN ACCOUNTED CONTINUUM Key words related quantity data type reg_diagCONT Array with the flags T F related to the use of regularization matrix in un accounted continuum retrieval A value is provided for each band T regularization matrix is used in un accounted continuum retrieval of the related band F regularization matrix is not used in un accounted continuum retrieval of the related band dimless logical MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 106 169 5 4 10 How to set up the optimal estimation procedure To set up the optimal estimation procedure the following parameters have to be defined Key words related quantity im data type b opi Absolute error on a priori opt pote tng information on pointing angle sae ie Absolute error on a priori opt_extCONT information on _Un accounted cm x10 real Continuum lope gaini Absolute error On a priorij Jimless wai information on gain Absolute error on a priori t_of A K lopt_ofs information on offset msi MARSCHALS Level 2 Prog Doc N IF AC GA 2007 06 LS User
103. iew of the program please refer to RD 6 The complete documentation and the original user manual of SBDART code is provided in the Annex 1 Chapter 8 The RTM code is developed by MARSCHALS team in order to use the output from SBDART and some other input data then it computes the simulated radiance as it would received by the OCM receiver In the Table 6 1 a list of all files needed for the execution of RTM code is presented MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 128 169 The data file needed to run the RTM code File classification Name of file batch file run_rim j makefile f source files and inc include files standard source files FORTRAN language the whole list is reported in RD 3 setting files settings_rtm dat auxiliary data files atmosphere_rtm dat This file is not actually mandatory to run the code since it is a batch file to run in an easier way the related code 1 These data are needed by the OCM code but can be obtained by the output of the Pre Processor thus they are not really input data of the MARSCHALS software Table 6 1 The data file needed to run the RTM code Prog Doc N IFAC GA 2007 06 LS MARSCHALS Level 2 U Manual Issue 3 Revision 1 ji Cc Ser Date 10 04 07 Page n 129 169 OFM Executable file of SBDART code t BIN
104. if r is the ratio of the top to bottom and t is the average opacity per level then tau top_level t 2r 1 r tau bot_level t 2 1 r NOTE a linear increase in opacity starting from zero at the cloud bottom is obtained by setting r 1 2 zdiff dz where dz is the grid spacing and zdiff is the total altitude range over which the cloud extends This formula assumes constant grid spacing over the cloud altitude range Thus if dz 1 then ZCLOUD 1 5 and TCLOUD 10 7 yeilds a linear increase from zero Page n 154 169 NRE 100 MARSCHALS Level 2 Issue 3 Revision 1 User Manual Date 10 04 07 Cloud drop effective radius microns up to 5 values Default value of NRE 8 The absolute value of NRE should be a floating point number in the range 2 0 to 128 0 NRE lt 0 selects mie scattering parameters for ice particles NRE gt 0 selects mie scattering parameters for water droplets The drop size distribution is assumed to follow a gamma distribution p 1 r Ro N r C xr Ro e where C is a normalization constant C 1 Ro gamma p p 7 and Ro NRE p 2 The factor p 2 relating Ro to NRE follows from the defining equation of NRE 3 2 NRE lt r N r gt lt r N r gt where the angle brackets indicate integration over all drop radii Another frequently used parameter to describe the size distribution is the mode radius Rm which is defined as the radius at which N
105. ings files in ASCII format with a common text editor In the following paragraphs of this chapter a selection of the most important settings is provided As a general rule it is possible to highlight that in the settings_marc dat file the data are arranged in rows all data are in ASCII format and are separated by a number of blanks Commented rows are preceded by a Blank rows have not to be considered The format of the data is the following a commented row starting with a introduces the variable under definition then a string that identifies the variable that is defined is written in a not commented row between square brackets the related quantity is reported in the following single non commented row Notice that the variable could be vectorial A string end_file ends the file Because of the presence of the key word string the order of the definition is not matter of care in any case for sake of clarity the parameters have been grouped according to their meanings MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i Cc Date 10 04 07 Page n 88 169 5 4 1 Howto set the release of the run of the retrieval code The keyword settings_key is used to define a string containing the file release Key words related quantity data type settings_key String specifying the release of the file dimless char 10 Since the name of the setting file has to be the one specified i
106. ion Not available when kdist 1 Output format write 5x 1lal3 2 airmass h20 co2 o3 amp n2o0 co ch4 o2 n2 trace total do j nz 1 1 write i5 1plle13 5 j z j airmass j amp log epsttrnsgas i j phidw i 1 nta where j is the level index z is the level height km airmass g integral rho dz mu Pzero where g 9 8m s2 pzero 1013 25mb rho is the mass desity of air and mu is the cosine of the solar zenith angle SZA trnsgas is the transmission due to the species listed in the title line the output quantity is the negative log of the transmission which aside from non Beer s law behaviour is like optical depth If the input quantity NF is non zero then the transmission is averaged over the solar spectrum If NF 0 the average is over the filter function Remember to set NF 0 and SZA 0 when dealing with LW radiation nzent 3 records for each wavelength Output format write tbf Block id used in postprocessors do m 1 nw write amp wl ffv topdn topup topdir botdn botup botdir write nphi nzen write phi j j 1 nphi write uzen j j 1 nzen do i nzen 1 1 write uurs i k k 1 nphi enddo enddo where Page n 164 169 10 11 MARSCHALSLeyel2 Issue 3 Revision 1 Date 10 04 07 User Manual WL wavelength microns FFV filter function value TOPDN total downward flux at ZOUT 2 km w m2
107. ion of a specific scan Subsequently the user can indicate which measurement he is interested in evaluating a set of parameters such as quality indicators from Level 1 OCM information and so on In order to reach this goal in addition to the main functionality of selection transcription of data SAMM Pre Processor performs the elaboration of other information not properly necessary for the execution of MARC retrieval but useful to plan the strategy of MARC elaboration Some examples are measurement geolocation from Level 1 and rejection of lines of sight from scan measurements due to low contrast in spectrum The SAMM code can also be used to produce the ancillary data files i e fov dat ils dat instrument dat and rejection dat required by the MARC code by considering some data files that are provided by the L1b team with the related measurement data files The aforementioned ancillary files can be also directly created by the user in any case they must be present to run the MARC code In the following paragraphs the input files necessary for SAMM Pre Processor execution see 4 3 the output files resulting from the Pre Processing procedure see 4 5 a list of modalities and functionalities of the code and some indications on the available user defined choices see 4 3 5 will be addressed together with the instruction to compile see 4 1 and run see 4 2 the code MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3
108. is not used dimless logical MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 99 169 Key words related quantity im data type Altitude above which the atmospheric continuum is forced to 0 It has to be licont_ab aE owe nullcont_above lt rulatm n ie Array defining the Reference Systems Origin izref rzref originref It is composed by one or two values the former being izref the possible latter being rzref both of them are defined in the following Integer specifying the kind of Reference System among some possibilities according to the following code 1 Reference Vertical passing through the instrument location 2 Reference Vertical passing through the lowest tangent point in geometrical approximation greater then 0 3 Reference Vertical at intermediate tangent altitude in geometrical approximation 4 user defined with respect to the instrument location izref dimless integer If the option 4 user defined vertical reference is specified the user defined displacement with respect to the vertical of the instrument has to be defined it is specified in the second element rzref of the array Trivially this element is present only if the option 4 is specified if izref lt 4 rzref is ignored Key words MARSCHALS Level 2 User Manual rel
109. is line specifies the standard output In the case reported on the left the standard output is redirected on log_marc txt ASCII file Table 5 1 The parameters defined in the run file with their environment variables and run of the MARC code MARSCHALS Level 2 User Manual Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 77 169 Executable file Mathematical library files MATH f files Source Batch file for the compilation gt f ince gt Source files of MARC code gt gt Script file for to run MARC e Output files OUT_FILES SS ofj Sequential fit aa settings marc dat lt gt User defined settings files Pre processed observ dat 12 measures r Input data Pre processed Settings INP_FILES instrument dat fov dat MARSCHALS STi lt gt Instrument rejection dat description ancillary files ils dat y Data file for ATM dat Sequential fit Seauential it REF_IND refind dat Refractive index table for ice and water i initial profiles of IN_GUESS atm_namef peher Teat k VMR temperature _zpt ahd pressure G
110. l Notice that the sequential fit performs some actions o it exports each retrieved profile in input format o it exports VCM for each retrieved profile o itreads VCM for each retrieved profile if found The default directory where these files are located in reported in Figure 5 1 The retrieval procedure of the MARC code is based on an internal Forward Model that can be set by the user by adequately specifying the values related to the following key words in the setting file MARSCHALS Level 2 User Manual 5 4 14 How to set the internal Forward model of the MARC code Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 The key words of this section are read only if the Forward Modality is set Imfm F Key words itot related quantity Number of Lines of Sight LOS to be simulated dimless Page n 115 169 data type integer fly_altitude Array containing the altitudes of the instrument for each LOS Note only itot values are read in the array itotxreal fly_angles Array containing the pointing limb view angles for each LOS deg itotxreal kind_of_noise Index to select the Noise Model to be superimposed on the simulated spectrum according to the code above 0 no noise is superimposed 1 uniform completely random noise is superimposed 2 uniform standard random noise is superimposed 3
111. l procedure The effectiveness of the single steps of the iteration is measured by the x value Details about the observed x on each band and each geometry during the retrieval procedure are reported in the chisquare dat file The retrieval procedure performs a check of some convergence criteria at the end of any iteration of the retrieval procedure the convergence dat file contains the values of the quantities that have been evaluated to verify the convergence criteria at the last iteration both the values calculated over the retrieved data and the related thresholds used to evaluate the quality of the retrieval Information about the Averaging Kernel Matrix and the related data are reported in the AK_matrix dat file Finally the total_correl dat file contains information about the total correlation existing between retrieved parameters and related data Gene MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS M l Issue 3 Revision 1 User Manual Date 10 04 07 Page n 122 169 file Data contained AK matrix dat information about the Averaging Kernel Matrix and the related data For more details see RD 3 chisquare dat information about the chi square on each band and each geometry during the retrieval procedure For more details see RD 3 convergence dat values that have been considered to verify the convergence criteria at each iteration For more details see RD 3 target dat set of files o
112. lated quantity The value extraction_type is an integer that indicates the set of data the Pre processor will extract The different modalities of extraction are reported in the following list where indicates the scan inside the considered flight extraction type 1 The Pre processor performs an overview extracting only general data referred to the flight It generates the file Jlight view see 4 5 1 EXTRACTION_TYPE extraction type 2 extraction_type dimless int The Pre processor performs an overview extracting detailed data referred to the scan It generates the file scan_ view see 4 5 2 extraction type 3 The Pre processor extracts MARSCHALS data It generates all the files needed by MARC code that is the observation_ dat file see 4 5 4 it also generates the file scan_ view see 4 5 2 continue in the next page MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 Cc ser Date 10 04 07 Page n 57 169 Key words related quantity extraction type 4 The Pre processor extracts OCM data It generates the file OCM_ dat see 4 5 3 extraction type 5 The Pre processor generates all files MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 58 169 4 4 4 How to enable the output of files containing instrumental information
113. le contiguous occurrence of satisfied criterion may be required for convergence MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 109 169 To avoid endless loop two stop criteria have been considered The adopted stop criteria is the following Stop criteria The code exit from the Marquardt iteration and moves itself to the next Gauss iteration when the maximum number of Marquardt iteration iMaxiterM is reached iterM iMaxiterM 5 4 2 Being iterM the number of Marquardt iteration The code exit from the Gauss iteration and ends the retrieval procedure even if the convergence is not reached when the maximum number of Marquardt iteration iMaxiterm is reached iterG gt iMaxiterG 5 4 3 being iterG the number of Gauss iteration The setting of the aforementioned convergence and stop criteria is made by setting the values related to the following key words in the setting file 1 Convergence Criterium Key words related quantity im data type Array with two elements conv_criterium1 rthreshold1 lconv1 Percent variation of the chisquare dimless h holdl ERB between two subseguent iterations Yo teal Switch T F for the convergence criterium 1conv1 T 1 Convergence Criterium per WA E is considered F gt 1 Convergence Criterium is not considered Gene MARSCHALS Level 2 Use
114. maintained separated in different subdirectories The WORKING _DIR directory contains two files and two subdirectories o the SBDART subdirectory o the RTM subdirectory Please notices two subdirectories with the same name SBDART are present at different level inside the OFM directory The SBDART subdirectory placed in the WORKING_DIR directory contains the atmos dat and the INPUT files with the user defined SBDART settings the sbdart dat file with simulated OCM measurement The RTM subdirectory placed in the WORKING_DIR directory contains the run_rtm batch file for the execution of the RTM code the INP_FILES subdirectory the OUT_FILES subdirectory MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 37 169 The JNP_FILES directory contains gt the atmosphere_rtm dat gt the settings rtm dat file with the user defined RTM settings The OUT_FILES directory contains gt the radiance_rtm dat OFM output file with the simulated OCM radiance produced by the RTM MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 38 169 3 1 4 How to change the default names of the subdirectories The user can change the default arrangements of the subdirectories by moving and changing the name of the subdirectories Please notice that if a change is made in the
115. mat marc executable file of he MARC code see 5 3 3 and RD 3 for the file format executable file observ dat file with the pre processed L1b data see 5 3 4 and pre processed data RD 3 for the file format files files with ancillary data see 5 3 5 and RD 3 for the file format e instrument dat e fov dat e rejection dat e ils dat ancillary data files files with auxiliary data see 5 3 6 and RD 3 for the file format e refind dat imaster my2 dspect_sideband_ isotopic_ratio dat marschals grd vmr_ species dat in_zpt dat cont_band_ T_band_ dat species _band_ dat species _g_band_ dat continuum_band_ dat pointing_band_ dat Error spectra data files see RD 3 auxiliary data files MARSCHALS Level 2 oa M l Issue 3 Revision 1 User Manua Date 10 04 07 Page n 79 169 The data file needed to run the MARC code File classification Name of files setting files settings_marc file to set the MARC code run see 5 3 7 1 These files are needed to create the marc executable file anyway a marc executable file has been already created and installed in the computer delivered to ESA thus these files are not requested to run the provided default version of the MARC code Table 5 2 The data file needed to run the MARC code The aforementioned files have to be stored in the hard disk of the computer devoted to run the MARC code
116. mpute the local earth radius latitude a real value used to define the latitude degrees Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 138 169 6 4 5 How to insert an identifier of the computation The user can introduce a character string to identify the current setting file The string is reported in the output file setting_key a string 3 characters used to identify the file MARSCHALS Level 2 User Manual Date 10 04 07 6 5 Output files 6 5 1 Output of RTM code The RT code returns as output the file radiance_rtm dat This file contains the radiance simulated by RTM code for each pointing angle of the instrument Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Page n 139 169 Data are arranged in rows and columns with a row for each pointing angle of the RTM code In each row the data are arranged in columns labelled by the following key words data Key words related guantity dim type Ind Sequential index of the element int Ang Pointing angle deg real Rztang Value of tangent altitude referred to Km pointing angle ang Angular coordinate of the tangent Rheta point referred to pointing angle deg real ang radiance Value of radiance simulated for the W m sr nm real pointing angle ang Table 6 2 Parameters defined in the radiance_rtm dat file listed with their key words 6 5 2 Lo
117. must be set to F dimless logical nretrievalgrid Number of altitude points of the retrieval grids user defined and tangent altitude grids dimless integer rretrievalgrid Array values with the altitudes of the retrieval grid in decreasing order available range 0 rulatm Note only the first nretrievalgrid values in the array are read nretrievalgrid x real MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 ji Cc User Manual Date 10 04 07 Page n 96 169 Key words related quantity im data type Array T F with nretrievalgrid flag one for each altitude of the retrieval grid specifying the mask for temperature retrieval T the temperature is itmask retrieved at the related dimless nretrievalgrid altitude x logical F the temperature is not retrieved at the related altitude note only the first nretrievalgrid codes in the array are read Matrix with Nfi t VMR rows one for each retrieved species in the rearranged order specified in 5 4 4 Each row is composed by nretrievalgrid flag one for each altitude of the retrieval grid specifying the mask for the VMR retrieval T the VMR the row is related to is retrieved at the related altitude F the VMR the row is Nfitvmr x lvmrmask related to is not dimless nretrievalgrid retrieved at the related x logical altitude Note e only the
118. n the run_marc file it would not be easy to change this name inside a series of executions of the code Thus in order to allow a tracking of different runs of the code with different parameters i e with different settings_marc dat files the key word settings_key has been introduced in the file to distinguish different set up i e different file with the same name settings_marc dat When a run of the code is repeated with a different set up specified in the settings_marc dat file the user can change the string associated to the key word settings_key A settings_key key word is provided for each input file the main code is able to take into account the version of all the files that are read during the run and an array with these version numbers is available in the output data Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 89 169 5 4 2 How to run the MARC code as a retrieval procedure or a simple forward model The MARC code is designed to perform a retrieval procedure but it can be used also as a simple Forward Model The lmfm flag must be set to decide if the code has to be run as a simple Forward Model or a Retrieval Code Key words related quantity im data type Flag F T to define if the code is used as retrieval code or forward model 1mfm T MARC works in forward model dimless logical modality F MARC works in retrieval model modality
119. n to compile see 5 1 and run see 5 2 the code MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 72 169 5 1 How to compile the source files of the MARC code The MARC code is composed by some source files written in standard Fortran language Please notices that the executable code installed on the computer delivered to ESA has been already compiled by the MARSCHALS L2 team if no change has been made to the source codes to run the executable version a compilation is not required The source files of the MARC code can be compiled in the usual way by means of a whatever FORTRAN compiler The computer delivered to ESA is provided with a FORTRAN compiler from Portland company that is able to perform this compilation task Once the compilation is completed the executable code has to be moved in the adequate directory that is in the proposed standard configuration the BIN one The compilation procedure can be performed by using an adequate makefile file provided by the MARSCHALS L2 team together with the code files this makefile file residing in the MARC_SOURCE subdirectory can be executed by means of the usual make command To use the makefile file please move to the directory in which the source files and the makefile file are stored the MARC_SOURCE directory in the standard configuration and at the prompt type make and then press the enter b
120. nation angles SZA PHIO are computed from the specified time and geographic coordinates using an internal solar ephemeris algorithm see subroutine zensun IDAY is the number of days into a standard year with IDAY 1 and the st of January and IDAY 365 on the 31st of December if IDAY gt 365 IDAY is replaced internally by mod IDAY 1 365 1 f IDAY lt 0 the code writes the values of abs iday time alat alon sza azm and solfac to standard output and exits UTC time Grenwich in decimal hours latitude of point on earth s surface east longitude of point on earth s surface NOTE TIME ALAT and ALON are ignored if IDAY eq 0 solar distance factor Use this factor to account for seasonal variations of the earth sun distance If R is the earth sun distance in Astronomical Units then SOLFAC 1 R 2 SOLFAC is set internally when the solar geometry is set through IDAY TIME ALAT and ALON In this case SOLFAC is set to 2 SOLFAC 1 eps cos 2 pi IDAY perh 365 where eps orbital eccentricity 0 01673 and perh day of perihelion 2 jan 2 NOTE seasonal variations in earth sun distance produce a 3 4 perturbation in the TOA solar flux This factor should be included when making detailed comparisons to surface measurements aerosol scattering mode used for boundary layer aerosols 0 normal scattering and absorption treatment 1 reduce optical depth by l ssa asym set ssa 0 2 set ssa 0 3 reduce optic
121. ne for each retrieved target with the retrieved information of the considered targets For more details see RD 3 iterationdetails dat information about the execution of the iterations of the iterative retrieval procedure For more details see RD 3 spectrum dat simulated measured spectrum obtained from the retrieved atmosphere measured spectrum obtained from the initial guess atmosphere and original measured spectrum For more details see RD 3 total_correl dat information about the total correlation between retrieved parameters and related data For more details see RD 3 Table 5 5 Standard output files of the MARC code when used as a retrieval code without any extra features MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 123 169 The MARC code when used in retrieval modality generates a series of output files related to the targets chosen by the user The files the code is able to produce are reported in the following table together with the related flags that have to be set to produce the aforementioned files flag to be set file contained data to produce the file retrieved profile of temperature and temperature dat related data ltempfit For more details see RD 3 retrieved profile of gas specified by the string specie and related data lvmrfit For more details see RD 3 specie dat retrieved
122. ng to the default setting is provided inside the computer delivered to ESA stored in the SAMM BIN directory and in the related package MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 47 169 4 3 4 The measurement files SAMM receives as input the data from Level 1B These data are expected to be stored in the following data file e LIB MARSCHALS file this is a binary file containing the MARSCHALS LIB measurement its name follows an internal L1B syntax see RD 1 Together with the input data coming from the L1b also some auxiliary see 4 3 5 and ancillary see 4 3 6 files are needed to run the SAMM code Trivially to execute the SAMM code the executable file is needed MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i c ser Date 10 04 07 Page n 48 169 4 3 5 The ancillary data files Together with the L1B data file the L1 team provides some auxiliary information about the instrument these information are stored in the following file e LIB_instrument lut file this is an ASCII file containing some instrument characterization needed to the pre processing operations MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 49 169 4 3 6 The auxiliary data files No auxiliary data files are requested for the basic run of the SAMM code
123. nm resolution This file format is new Previous versions of SBDART used a different format for spectral input files albedo dat filter dat and solar dat A perl script newform is available from ftp ftp icess ucsb edu pub esrg sbdart to convert old data files to the new format ISALB 0 a spectrally uniform surface albedo ISALB 7 the wind speed m s Composite albedo fractions applies only when ISALB 10 SC 1 fraction of snow SC 2 fraction of ocean SC 3 fraction of sand SC 4 fraction of vegetation NOTE SC 1 SC 2 SC 3 SC 4 need not sum to 1 Thus it is possible to use the SC factor to boost the overall reflectance of a given surface type For example SC 0 0 2 0 yields results for a surface with spectral reflectivity twice that of sand Beware total reflectance greater than one will produce unphysical results MODEL ATMOSPHERES default ozone atm cm ATMOSPHERIC PROFILE water vapor g cm2 total below 10km 0 User Specified 1 TROPICAL 4 117 0 253 0216 2 MID LATITUDE SUMMER 2 924 0 324 0325 3 MID LATITUDE WINTER 0 854 0 403 0336 4 SUB ARCTIC SUMMER 2 085 0 3 50 0346 5 SUB ARCTIC WINTER 0 418 0 486 0340 6 US62 1 418 0 349 0252 n List to standard out Prog Doc N IFAC GA 2007 06 LS Page n 150 169 AMIX UW UO3 O3TRP ZTRP XN2 KOZI XCO2 XCH4 XN20 XCO XNH3 XSO2 XNO XHNO3 XNO2 NOTE MARSCHALS Level 2 Issue 3 Revision 1 Us
124. nthetic spectra the iterative execution of an inversion algorithm a modeling of scattering contribution due to clouds both in millimetre region and in near infrared spectral range A set of modules have been thus developed and arranged in executable codes 2 1 Software modules and codes The MARSCHALS L2 software is basically constituted by five modules Forward Model module Retrieval module Scattering Source Function module Pre Processor module Optical Sounding module gt The Forward Model module is dedicated to the simulation of synthetic spectra at a given atmospheric state it is devoted to the computation of a whole set of spectroscopic limb measurements as they would be acquired by the MARSCHALS instrument starting from a given profile of atmospheric constituents pressure and Temperature and starting from a model of the instrument itself gt the Retrieval module is devoted to the retrieval of the profile of atmospheric constituents by processing a set of spectroscopic measured or simulated data gt the Scattering Source Function module is dedicated to the computation of additional Source Function due to the presence of aerosols gt the Pre Processor module is dedicated to a preliminary overview of measurement data and to rearrange the data themselves in a suitable format gt an Optical Sounding module is dedicated to the simulation of radiance as it is received by OCM instrument
125. on is necessary for extracting information from OCM images The Pre Processor arranges a wrapped profile of radiance made by segments of profiles provided by each individual OCM image referred to a specific line of sight For more details see RD 3 Page n 67 169 GA_2007_06_LS Prog Doc N IFAC Issue 3 Revision 1 Date 10 04 07 MARSCHALS Level 2 User Manual YSIS JO SOUT O11 aydsoune jo sroquinu 180 progs ay uey Iapa Sie Ayyenb SOTJO Jaquinu program y uey 1342313 PAA 1SLIJSUO9 SOTJO Jagwnu progs ayy uey SSO J AJ srou SOTJO Jaquinu ISIS JO sour T Bis jo soul Ueudsowye jo sJaquunu Jezo ploysesy pejosjes ay UBU 1842916 Jaquunu ul ave sej Aypenb asoym soj Jo Jequuinu asuel apne Spisur quod quosury SOTJO Jagwnu PIOUSa U Pa Dejas JU ULY SS9 SI JAd SIOU SOYM SO JO Jaguwnu PIOUS9 U pa 99 S y UBU 19 e916 si N SV1 SUO9 BSOYM SO JO Jagwnu aBues apnyyye pajoajas y apisu sje JUIOd uoue soym soj Jo Jagwnu uLos y JO SO YBIS Jo Saul pajoajas Jo Jequunu 6ap ueos y Buunp julod jubur pajoeijes ueawW y zo apNyHuo 6ap ueos y Huunp julod juabHue payoesjos ueaw BU JO BpNI E 6ap ueos y Buunp julod juabHue pajoesjos UBS y JO UINWIZe wy ueos y Buunp julod juabue payoesjas au jo BpNyyye WNW wy ueos y Buunp julod juabue pajobuja ay JO pnyye UNWIKBW 6a
126. or each iteration F computation is done only for the first evaluation of the FM using the initial guess atmosphere lray dimless logical MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 Cc Ser Date 10 04 07 Page n 120 169 Key words related quantity im data type Convergence criterion for the C G dimless real integrals deps Switch T F to define the export of the AB matrix A and B being the two matrices used in the retrieval according to the definition of the ATBD document lainvbexp dimless logical T Matrix AB is exported F Matrix A B is not exported It is active only if debug is true debug modality Switch F T to select the exporting of the Jacobian matrix on a file 1jac T Jacobian matrix is dimless logical exported F Jacobian matrix is not exported Switch F T to select the exporting of the histo structure containing all the retrieval details nisto T histo structure is exported aimless logical F histo structure is not exported MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 121 169 5 5 Output files The MARC retrieval procedure generates as output some files that are detailed in the following The MARC code produces as output a set of files that are stored in the d
127. p ueos y Buunp Jalada y jo apnjibuo ue w Bap ueos y Buunp 19419991 y Jo Apne ue w wy ueos y Buunp 1341991 y Jo pnyye ue w dwejys owl ue w yaquinu ueoS yurod mod SIS yod prod wrod u Jur uasue CS i quasuey ques ue quosuey ie ad ie je I3AI39AI IJAI993I JAIIAI F ain en POLI paga paya P sn j ja 3 ayy jo ayy jo ayy jo P o UBU uvour ULI ii ae je apnjiSuo apne apne ayy Jo ayy jo ayy jo r ueu ueour ueour qumu Hy YJ TS i it apnjiSuo apne nuze PNILOUC pon ypu numuru nume JUJWNJISUI qutod u93ue ot JO uoneooj OBeIODAY Si Jo woRBOOT SFWA ight view dat file Table 4 3 The structure of the fl a g S 2 d gt GL Jeng q peyodes se spej AMlijenb aanisod y jo Jagwnu oz a MI 2 2 osiou y Jo anjea 64 5 3 I9A9 ISBIJUO9 y JO NJLA 84 3 3 S y 9A wnnujuo9 y JO njen Z4 g S Bap julod juabue pajoeuja JO apnjibuo 94 A n Bap julod juaHue payoeyas Jo pnyyel G4 a PE wy uonisinboe y Buunp julod juabHue jeoujewo9b y jo Aqyiqeuen oue Yl AAA Wy uiod juaHue paoe jo pnyye El 6ap jBue Bunuiod jeypuaz jo uogeinop pJepueys Zl 69p aj6ue Bunuiod seujiuez LL 6ap yis jo u y Jo ajGue yynw ze Jo uoernap Pyepue s OL 69ap 14518 jo oul y Jo aj6ue YINWIZE jo onjeA UBL 6 a Bap Jalada Jo apnjibuo g 2 Bap Jalada JO apne Z S wy 13419991 JO apne g Nese 38 69p ej6ue Bunuiod seyyuez jeulWwoU G s dwejs e
128. p the users in compiling and running the codes An executable version of the code i e already set up and compiled is also provided trivially the executable version is not strictly needed to run the code since it can be originated by compiling the source codes On the opposite hand the source files are not requested if the executable version is used Each code operates on data stored in different data files Trivially to run the codes the user needs both the source files to be compiled and the adequate input data files The Pre processed input data files are needed to run the MARC code but as for the executable files they are not strictly needed because they can be originated by a run of the SAMM code on the measurement files The files needed to run the codes are briefly presented in the following Paragraphs where files are introduced grouped according to their role in the execution of the codes this functional subdivision of the MARSCHALS L2 files is schematized in Figure 2 2 A detailed list of the files needed to run the code is reported in Chapter 3 As shown in Figure 2 2 the MARSCHALS suite is based on some code files i e on a set of source and batch files these source files are introduced in Paragraph 2 4 1 MARSCHALS L2 codes require some setting and parameters files see 2 4 2 the data coming from the MARSCHALS measurements are trivially requested and provided by the L1b measurement file see 2 4 2 2 Furthermore a
129. r is maximized For our drop size distribution Rm p 1 Ro Using the relation between Ro and NRE we find that Rm p 1 NRE p 2 NOTE If the first element of NRE is zero the values of TCLOUD ZCLOUD LWP and NRE are ignored and cloud specification records are read from file usrcld dat The first record in this file corresponds to the lowest layer in the atmosphere that is between the surface and the lowest cell boundary altitude Each following record sets values for the next higher atmospheric layer in the model atmosphere usrcld dat is read with the following fortran statements do i 1 nz 1 read 13 end 100 lwp i re i fwp i rei i cldfrac i enddo continue where lwp liquid water path in layer i g m2 default 0 re effective radius of liquid water um in layer i default 8um fwp frozen water path in layer i g m2 if fwp lt 0 then scattering parameters are obtained from ccm3 cirrus model see subroutine icepar if fwp gt 0 then scattering parameters are obtained from an internal mie scattering database covering ice spheres with effective radii between 2 and 128 um default 0 rei effective radius of frozen water um in layer i only active when fwp is non zero if fwp 1t 0 and rei le 0 then effective radius of ice is taken from ccm3 cirrus model see subroutine icepar default 1 cldfrac cloud fraction in layer this parameter reduces cloud optical depth by factor cldfrac 1 5 default 1 It
130. r Manual 2 Convergence Criterium Key words conv_criterium2 related quantity Array with two elements rthreshold2 lconv2 Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 Page n 110 169 data type is considered F gt 2 Convergence Criterium is not considered sth esho1a2 Maximum percent correction to state dimless WA vector in percentage Yo Switch T F for the convergence criteria nd P Lconv2 T 2 Convergence Criterium dimless EN 3 Convergence Criterium Key words conv_criterium3 related quantity Array with two elements rthreshold3 lconv3 data type Maximum percent correction to Criterium is considered F gt 3 Convergence Criterium is not considered xthreshold3 modified state array modified in E real percentage E Switch T F for the convergence criteria rd Lconv3 Ta Convergence dimless logical Gene MARSCHALS Level 2 User Manual 4 Convergence Criterium Key words conv_criterium4 related quantity Array with two elements rthreshold4 lconv4 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 111 169 data type is considered Fo 4 Convergence Critertum is not considered Percent variation between chisquare dimless pEhreshotaal and LINEAR chisq in percentage
131. r of user specified zenith angles nz number of atmospheric levels ffew filter function equivalent width um phi user specified anizmuth angles degrees uzen user specified zenith angles degrees Zz altitudes of atmospheric layers km fxdn downward flux direct diffuse W m2 fxup upward flux W m2 fxdir downward flux direct beam only W m2 uurl radiance at each layer W m2 str NOTE if IDAY is set then PHI is the actual compass direction in which the radiation in propagating same as IOUT 20 except lower hemisphere radiance output corresponds to ZOUT 1 upper hemisphere radiance output corresponds to ZOUT 2 Use this output format to determine radiance above and and below a scattering layer For example if ZCLOUD 1 and TCLOUD 10 you can get the scattered radiation field above and below the cloud with IOUT 23 ZOUT 1 2 Page n 166 169 DELTAM NSTR CORINT NZEN UZEN MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 NOTE if IDAY is set then PHI is the actual compass direction in which the radiation in propagating if set to true use delta m method see Wiscombe 1977 This method is essentially a delta Eddington approximation applied to multiple radiation streams In general for a given number of streams intensities and fluxes will be more accurate for phase functions with a large forward peak if DELTAM is set TRUE In
132. re that the temperature steps in the atmospheric model are small enough to resolve changes in the Planck function The original version of the DISORT radiative transfer module issued a warning message if the temperature difference between successive levels in the atmosphere exceeded 20 K All the standard atmospheres violate this condition for at least 1 stratospheric layer This warning message has been disabled to avoid clutter in SBDART s standard output If near IR thermal emission from the stratosphere is important to your application you should supply SBDART with a new model atmosphere with higher resolution in the stratosphere see ZGRID1 ZGRID2 an NGRID KDIST 1 causes correlated k optical depths and weighting factors to be read from files CKATM and CKTAU NOTE KDIST 1 disables the effect of all input parameters that control aspects of the gaseous atmospheric profile Thus KDIST 1 disables input parameters AMIX SCLH20 UW UO3 O3TRP ZTRP XN2 KOZ XCO2 XCH4 XN20 XCO XNO2 XSO2 XNH3 XNO XHNO3 X04 RHCLD KRHCLR NGRID ZGRID1 ZGRID2 PBAR and ZPRES KDIST 0 causes the optical depth due to molecular absorption to be set to the negative log of the LOWTRAN transmission function This approximation is not appropriate for cases in which multiple scattering is important but is not very wrong when the molecular absorption is weak or the scattering optical depth is small KDIST 1 causes SBDART to use the LOWTRAN7 k
133. riable names Page n 145 169 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M c User Manual Date 10 04 07 2 ERROR MESSAGE too many values for NAMELIST variable MEANING you specified too many values for a variable most likely because you separated variables by more than one comma 2 ERROR MESSAGE end of file during read or namelist block INPUT not found MEANING There are two possibilities A you didn t include a NAMELIST block specifier INPUT DINPUT or END or you misspelled it or B You used the wrong character to signify a namelist block name FORTRAN90 expects the namelist blocks to start with amp and end with but most FORTRAN77 compilers used the SEND convention Other input files are sometimes required by SBDART atms dat atmospheric profile see input quantity IDATM aerosol dat aerosol information see input quantity IAER albedo dat spectral surface albedo see input quantity ISALB filter dat sensor filter function see input quantity ISAT solar dat solar spectrum see input quantity NF usrcld dat cloud vertical profile see input quantity TCLOUD SBDART usually lists computational results directly to the standard output device i e the terminal if run interactively However some warnings issued by the DISORT radiative transfer module are written to files named DISORT_WARNING where the question marks indicate a warning message number These files contain
134. rosols in the lower atmosphere However in SBDART the default vertical density of BLA falls off exponentially and affects regions above the normal extent of the boundary layer The vertical influence of these aerosols may be confined to a specified boundary layer altitude with the optional parameters ZBAER and DBAER The spectral dependence of the boundary layer aerosol scattering parameters are sensitive to relative humidity Use input parameter RHAER to set the relative humidity used in the boundary layer aerosol model Set RHAER 1 the default value to use the ambient surface relative humidity RHAER has no effect when IAER 5 Horizontal Path Visibility km at 0 55 microns due to boundary layer aerosols This parameter does not set the optical depth for the user defined aerosol model IAER 5 but does affect that model through the vertical structure see below NOTE unlike the stratospheric aerosols the boundary layer aerosols have predefined vertical density distributions These vertical structure models vary with visibility see discussion of ZBAER and DBAER NOTE The boundary layer aerosol optical depth absorption scattering at 0 55 microns is given by tauaero 0 55um 3 912 integral n z n 0 dz VIS where n z is the vertical profile of aerosol density For the 5 and 23 km visibility models the indicated integral is 1 05 and 1 51 km respectively So tauaero 0 55um 3 912 1 05 w 1 51 1
135. ser Date 10 04 07 Page n 134 169 6 4 Settings file of RTM code The user can set the procedure of the OFM by editing the setting file of RTM called atmosphere_rtm dat 6 4 1 How to define the altitude of the receiver The user can introduce a value to fix the altitude of the receiver receiver_altitude real value Km to identify the altitude of the OCM receiver MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 User Manual 6 4 2 How to define the pointing angles The user can introduce one or more pointing angles at which the simulation is performed n_angles an integer to define the number of pointing angles pointing_angles zenithalpointing angles degrees an array real values used to define the Page n 135 169 MARSCHALS Level 2 fa cC User Manual 6 4 3 How to define the receiver field of view Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 The user can define an uniform field of view by entering an angular value fov a real value which define the half aperture of the antenna Page n 136 169 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS U Manual Issue 3 Revision 1 i c Ser Date 10 04 07 Page n 137 169 6 4 4 How to define the latitude of the observations The user can define the latitude at which the simulation is carried out This information is used by RTM code to co
136. set of adequate data files coming from the pre processing of a L1b data file are requested as input file of the MARC code see 2 4 2 3 A set of auxiliary files with auxiliary data e g meteorological data is also requested see 2 4 2 5 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 23 169 2 4 1 The code files The set of code files is composed by the source files see 2 4 1 1 to be compiled to generate the executable files three makefiles one for each code that can be used to compile the codes see 2 4 1 2 and three executable files that is source codes compiled by the MARSCHALS L2 team see 2 4 1 3 2 4 1 1 The source files The MARSCHALS L2 codes are based on a set of source files written in FORTRAN language MARC code and OFM codes and C language SAMM code The names of these files that have to be compiled to run the related codes are described in RD 3 2 4 1 2 The batch files The batch files are composed by three makefiles and three run files The three makefile files are provided one for each of the three codes of the MARSCHALS suite to compile the source files of the related code More details can be found in 4 1 makefile to compile the SAMM code 5 1 makefile to compile the MARC code and 6 1 makefile to compile the OSM code The three run batch files one for each code are provided to run the codes They are basically shell
137. sols at wavelengths WLBAER Number of values must match the number of WLBAER GBAER is ignored when parameter PMAER is set or when MOMA ne 3 Legendre moments of the scattering phase function of boundary layer aerosols only active for IAER 5 The Legendre moments of the phase function are defined as the following integral over the scattering phase function f 7 pmaer i f mu P i mu d mu mu d mu yA where P i mu is the Legendre polynomial mu is the cosine of the scattering angle and the range of the integrals are from 1 to 1 The Legendre moment for i 0 is always one Hence the zero th moment is assumed by SBDART and should not be specified Unlike the previous boundary layer aerosol parameters you need to specify at least NSTR values for each wavelength point for a total of NSTR NAER values where NAER is the number of wavelength points supplied The order of specification should be such that wavelength variation is most rapid For example here is a case with 4 wavelengths and 6 streams nstr 6 wlbaer 400 500 600 700 pmaer 0 80 0 70 0 60 0 50 0 64 0 49 0 36 0 25 0 51 0 34 0 22 0 12 Page n 160 169 ABAER IMOMA SPOWDER NOTHRM MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 0 41 0 24 0 13 0 06 0 33 0 17 0 08 0 03 0 26 0 12 0 05 0 02 Wavelength Angstrom model exponent used to extrapola
138. square brackets whose appearance is thus key word specifying the quantity defined in the following row The definition of a parameter will be thus as in the following example keyword parameter_value Due to the presence of the key word the order of the couples of rows specifying the parameters is meaning less while the absence of a key word will abort the process In the following the description of all the available user defined choices is reported 4 4 1 MARSCHALS Level 2 User Manual How to set the input output namefiles Prog Doc N IFAC_GA_2007_06_LS Issue 3 Revision 1 Date 10 04 07 Page n 52 169 The name of the input output files can be defined by the user by setting the following parameters Key words related quantity The string input_file_name labelled by the key word char NAME _L1B FILE u t file cei NAME_L1IB_FILE dimless max KRS specifies the name of the binary 200 char file L1B_MARSCHALS file as received from Level 1B The variable NAME OUTPUT DIRECTORY CUCPUt_directory_name char Spee ee eas specifies the directory in which dimless max ie y SAMM will save the output files 200 char path is requested NAME_LUT_FILES name of the LIB file containing char name file LIB LUT the instrument characterization dimless max ee 200 char MARSCHALS Level 2 fa Cc User Manual 4 4 2 How to define the quantities to be extracted from the
139. sup 0 550 ya wlinc 0 0 ji sza 0 0 csza 1 0 p solfac 1 0 A nf 2 F iday 0 time 16 0 m alat 64 7670 alon 64 0670 zpres 1 0 pbar 1 0 sclh20 1 0 UW 1 0 A uo3 1 0 y o3trp p 7 ztrp 0 0 A xrs 1 0 j xn2 1 0 xo2 1 0 7 xco2 1 0 xch4 1 0 xn2o 1 0 r RCO 1 0 pi xno2 1 0 xso2 1 0 r xnh3 i xno 1 0 xhno3 1 0 F xo4 1 0 isalb 0 albcon 0 0 r sc 1 0 3 0 0 zcloud 5 0 0 tcloud 5 0 0 lwp 5 0 0 nre 5 8 0 r rhcld 1 0 krhelr D jaer 5 0 r zaer 5 0 0 A taerst 5 0 0 iaer 0 yig 23 0 rhaer 1 0 Wwlbaer 47 0 0 thaer 47 0 0 r abaer 1 0 wbaer 47 0 950 gbaer 47 0 70 Pi pmaer 940 0 0 zbhaer 50 1 0 Obaer 50 1 0 nothrm 1 nosct 0 kdist 3 ji zgridl 0 0 zgrid2 30 0 ngrid 50 r zout 0 0 100 0 ibut 0 deltam t lamber t z ibend 0 r phil 0 0 A prnt 7 f ipth igot 0 0 1 temis 0 0 z nstr 4 nzen 0 A uzen 20 1 0 7 vzen 20 90 nphi 0 7 phi 20 1 0 imomc 3 imoma 3 ttemp 1 0 btemp 1 0 spowder f A idb 20 0 NOTE Unfortunately many fortran compilers produce rather cryptic error messages in response to improper NAMELIST input files Here are three common NAMELIST error messages and their meaning 1 ERROR MESSAGE invalid reference to variable in NAMELIST input MEANING you misspelled one of the NAMELIST va
140. tal offset retrieval T gt retrieval of offset is performed F retrieval of offset is not performed dimless logical ig_offset initial guess of the instrumental offset for each fitted band 3 x real lgainfit Switch for gain retrieval T retrieval of gain is performed F retrieval of gain is not performed dimless logical ig_gain initial guess of gain for each fitted band dimless 3 x real This setting is considered also when the MARC code works as a simple Forward Model MARSCHALS Level 2 User Manual Date 10 04 07 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Page n 94 169 MARSCHALS MASTER HITRAN code 1 H 0 2 C0 3 0 4 N20 5 CO 6 CH 7202 8 NO 9 SO 10 NO 11 NH 12 HNO 13 OH 14 HF 15 HCI 16 HBr 17 HI 18 ClO 19 OCS 20 H2CO 21 HOCI 22 N 23 HCN 24 CH3C 25 H207 26 CH2 27 C2H6 28 PH 29 COF 30 SF6 31 H2S 32 HCOOH 33 HO 34 0 35 CIONO 36 NO 37 HOBr 38 C2H4 39 CH30H 40 BrO Table 5 4 the MARSCHALS MASTER HITRAN code for the main molecules Please notice that the order of the species to be retrieved is rearranged by the code using the following modality The list starts with the species for which both the VMR and the gradient are to be retrieved among these species the order is that specified by codefitGRAD Then th
141. te BLA extinction efficiency to wavelengths outside the range of WLBAER Qext lambda abaer This parameter is only operative when IAER 5 If ABAER is set to a positive number then that value is used as a power law wavelength dependence to extrapolate the extinction efficiency for wavelengths less than WLBAER 1 or greater than WLBAER nn where nn is the number of specified values If ABAER is not set the wavelength extrapolation is based on the last two specified points wlbaer 1 wlbaer 2 or wlbaer nn 1 wlbaer nn If ABAER is not set anda single wavelength is set then a spectrally constant extinction efficiency is used Controls phase function used for boundary layer aerosol The value of IMOMA is ignored when IAER 5 and PMAER is specified Note that an asymmetry factor must be specified when IMOMA 3 default 3 isotropic scattering rayleigh scattering henyey_greenstein g re lt default haze L as specified by garcia siewert cloud c 1 as specified by garcia siewert OP WNE Setting SPOWDER to true causes an extra sub surface layer extending between 1 and 0 km to be added to the bottom of the atmospheric grid This layer may be used to model the effects of surface reflection and thermal emission caused by a granular surface material e g snow or sand The scattering properties of the surface layer may be specified either with the cloud or aerosol inputs For example a thin water cloud over a snow surface
142. tensities within 10 degrees or so of the forward scattering direction will often be less accurate however so when primary interest centers in this so called aureole region DELTAM should be set FALSE default true number of computational zenith angles used NSTR must be divisible by 2 Using NSTR 4 reduces the time required for flux calculations by about a factor of 5 compared to NSTR 16 with very little penalty in accuracy about 0 5 difference when DELTAM is set true When set TRUE correct intensities for delta M scaling effects see Nakajima and Tanaka 1988 When FALSE intensities are not corrected In general CORINT should be set true when beam source is present FBEAM is not zero and DELTAM is TRUE in a problem including scattering However execution is faster when CORINT is FALSE and intensities outside the aureole may still be accurate enough When CORINT is TRUE it is important to have a sufficiently high order of Legendre approximation of the phase function This is because the intensities are corrected by calculating the single scattered radiation for which an adequate representation of the phase function is crucial In case of a low order Legendre approximation of an otherwise highly anisotropic phase function the intensities might actually be more accurate when CORINT is FALSE Default value false The input value of CORINT is ignored for 1 in irradiance mode i e iout ne 5 6 20 21 22 2 there is
143. the following keyword Key words related quantity data type ltempfit Switch logical value F T for the retrieval of the temperature values T the temperature is retrieved F temperature is not retrieved dimless logical lvmrfit Switch F T for the retrieval of the VMR values T at least the VMR of one species is retrieved F VMRSs are not retrieved dimless logical N itVMR Number of gases whose VMR is retrieved Note that the NfitVMR value is read only if lvmrfit is True dimless integer codefitVMR Array of integers with the MARSCHALS MASTER codes of the molecules to be retrieved Note that only the first nfitvmr codes of the arrays are read The MARSCHALS MASTER code is composed by numbers that indicate the species matching number of the species is specified in the following Table 5 4 Retrieval tests have been performed for H20 03 HNO3 CO N20 CH3Cl see RD 4 dimless Nfitvmr x integer Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 92 169 The MARSCHALS MASTER code corresponds to the HITRAN code and is composed by numbers that indicate the species matching number lt gt species is specified in the Table 5 4 Key words related quantity data type Switch logical value F T for the retrieval of the gradients values lgvmrfit T at least the VMR gradi
144. then press the enter button Please notice that the make commands works according to the standard arrangement of the file as reported in 3 1 1 the user can change this arrangement but in this case to use the make command the makefile file have to be edited accordingly to the new user defined arrangement of the files MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 41 169 4 2 How to run SAMM code Together with the executable MARSCHALS L2 software installed on the computer delivered to ESA a batch file is provided to run the SAMM code The user can run the SAMM code by using the batch file run_samm stored as a default in the SAMM WORKING_DIR subdirectory To run the SAMM code by directly using the samm executable file please move to the work directory the SAMM WORKING_DIR in the default arrangement and at the prompt type BIN samm name file settings where name_file_settings is the name of the settings file with the relative path with respect to the work directory i e INP_FILES settings_samm dat in the default arrangement Alternatively to run the SAMM code by using the batch file please move to the subdirectory in which the run_samm batch file is stored the SAMM WORKING_DIR in the default arrangement and at the prompt type run_samm and then press the enter button The user can change the default arrangements of t
145. tical library files Source files Script file for to run MARC Output files Lo Sequential fit content WORKING_DIR _Output files gt User defined 2 8 ae 0 settings marc dat gt settings wn O B os I m Pre processed 2 8 observ dat 12 measures os Lo sa os INP_FILES a instrument dat gt fov dat MARSCHALS Se a lt gt Instrument e rejection dat description ils dat y is Data file for ATM dat Sequential fit Le 5 Refractive index table BER ING refind dat for ice and water Ymr specie dat initial profiles of IN_GUESS atm_name in pare E K VMR temperature _zpl ahd pressure mn Spectroscopic database imaster my2 trom MASTER study TT n lt lt 2 i Spectroscopic B pi p dspect_sideband_B dat database tor gt SSS sideband B C D amp AUX D_SPECT dspect_sideband_C dat Selection from HITRAN by 2 Te MARSCHALS dspect_sideband_D dat Team 7 Isotopic_ratio dat Isotopes data CONT E oS Files with continuum cont_band_ dat profile for band _ AU wa Irregular marschals grd gt freguency grid g Files with VCM of ERROR_SPECTRA dat Forward Model data Cee Figure 3 3 T
146. tinue where wlfilt wavelength sample points microns AJE filter response value unitless The number of wavelength sample points read from filter dat should be less than or egual to 1000 WLINF WLSUP WLINC MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 This file format is new Previous versions of SBDART used a different format for spectral input files albedo dat filter dat and solar dat A perl script newform is available from ftp ftp icess ucsb edu pub esrg sbdart to convert old data files to the new format lower wavelength limit when ISAT 0 WLINF gt 250 microns central wavelength when ISAT 2 3 4 upper wavelength limit when ISAT 0 WLSUP lt 100 0 microns equivalent width when ISAT 2 3 4 NOTE If ISAT eq 2 a rectangular filter constant with wavelength is used with central wavelength at WLINF and an equivalent width of WLSUP full width WLSUP If ISAT eq 3 a triangular filter function is used with the central wavelength at WLINF and an equivalent width of WLSUP full width 2 WLSUP filter function is zero at end points and one at WLINF If ISAT eq 4 a gaussian filter function is used with the central wavelength at WLINF and an equivalent width of WLSUP full width 4 WLSUP If output is desired at a single wavelength set WLINF WLSUP and ISAT 0 In this case SBDART will set WLINC 1 the user specified v
147. tion This information is stored in four files e instrument dat e fov dat e rejection da e ils dat that in the default configuration are stored in the MARC WORKING DIR INP_FILES subdirectory Please notice that a set of ancillary data files are provided together with the other files of the MARSCHALS L2 suite They are referred to the provided test data if the codes of the MARSCHALS L2 suite are run with different measurement data the provided ancillary data could not be adequate to the new measurement scenario In this case a new and suitable set of ancillary data files have to be provided by the user MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 85 169 5 3 6 The auxiliary data files Some auxiliary files are needed to run MARC retrieval code These files are not provided by the pre processor and have to be provided by the user These files are listed in Table 5 3 and described in RD 3 Please notice that a set of ancillary data files are provided together with the other files of the MARSCHALS L2 suite MARC auxiliary data files File with refractive index table for ice and water refind dat file with the spectroscopic database for the imaster my2 y main bands dspect_sideband_B dspect_sideband_C dspect_sideband_D file with the spectroscopic database for the image bands isotopic_ratio dat file
148. tion procedure The Regularization can be separately applied to single retrieved quantities To set up the Regularization procedure the following key words have to be defined Key words related quantity im data type Parameter for tuning profile t 1 reg_tuning regularization dimless real TEMPERATURE Key words related quantity im data type Switch T F for the use of regularization matrix in temperature retrieval T regularization matrix is used in temperature retrieval F regularization matrix is not used in temperature retrieval reg_diagT dimless logical VMR Key words related quantity im data type Array with the flags T F related to the use of regularization matrix in VMR retrieval A value is provided for each retrieved species in the order specified in sect 5 4 4 reg_diagVMR T regularization matrix is ree logical used in VMR retrieval of the related species F regularization matrix is not used in VMR retrieval of the related species VMR GRADIENT Key words MARSCHALS Level 2 User Manual related quantity Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 105 169 data type reg_diagGRAD Array with the flags T F related to the use of regularization matrix in VMR gradient retrieval A value is provided for each retrieved species in the order specif
149. titude of the top of the clouds Km real 8 MAA Vertical geometrical thickness of the Km real 8 clouds ipdnetoud Numerical density of the particles cm real 8 inside the clouds Parameter used to build the number of direction used to model the solid nag angle surrounding the clouds As a dimes integer 4 matter of fact the solid angle is sampled by means of 2xnag 1 directions 5 4 16 How to define the parameters for the advanced settings Page n 118 169 MARSCHALS Level 2 Prog Doc N IF AC GA 2007 06 LS U M l Issue 3 Revision 1 ser Manya Date 10 04 07 Some advanced settings can be selected in MARC code these settings can be specified by the user by defining the values related to the following key words in the setting file Key words related guantity data type ldebug switch T F to enable the DEBUG option T debug is ENABLED F gt debug is DISABLED dimless logical rulatm Value specifying the altitude of the upper boundary of the atmosphere real rdaup Altitude step to complete the base grid above retrieval altitudes real rdzda Array with two components the former specifies the altitude step to be added to the maximum flight altitude to obtain the altitude under which the grid has to be thickened the latter specifies the maximum step in the thickened grid real nfov Minimum number of pencil beams
150. to be set by defining the related parameters these parameters have to be specified by setting the related key words e To set up the Gauss and Marquardt iterations see 5 4 8 e To set up the Regularization technique see 5 4 9 e To set up the Optimal Estimation technique see 5 4 10 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 102 169 5 4 8 How to set up the Gauss and Marquardt iterations The retrieval procedure is based onto two nested iterations the Gauss one macro iteration and the Marquardt one micro iteration The following key words have to be defined to set the Marquardt and the Gauss iterations Key words related quantity im data type Initial value for lambda parameter in Marquardt iterations used for the lambdat f era ak retrieval of the Temperature dimless real Array containing the initial values for parameters in Marquardt iterations used for the VMRs retrieval the parameter has been di Nfitvmr x imless considered as gas dependent thus a real values is given for each retrieved VMR the order inside this array is that defined in 5 4 4 rlambdav Array containing the initial values for parameters in Marquardt iterations used for the retrieval of the VMR gradients the parameter has di Nfitgrad x imless been considered as gas dependent real thus a value is given for each retrieved gradient
151. tored as a default in the MARC WORKING DIR subdirectory have to be changed accordingly by the user Please notice that MARC code files are arranged in subdirectories These subdirectories are contained in a main directory named as a default MARC that is user defined all the paths are thus defined in a relative way with respect to the subdirectory in which the executable files are contained that is the MARC WORKING DIR one in the default configuration The run_marc environment file contains information about the internal tree structure of the subdirectories in which the files needed to run the MARC code reside thus if a change is made to the delivered default configuration as reported in 3 1 2 the run_marc file have to be changed accordingly by the user In the following the run_marc file is presented each definition is shown and explained in the order of appearance inside the file A default arrangement of the input files is described in 3 1 2 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 Cc ser Date 10 04 07 Page n 74 169 5 2 1 Therun marc file line in run_marc file Related setting export MARC_IODIR The MARC_IODTR variable indicates the path of the 1 0 files directory The default configuration is export MARC_IODIR and indicates that the INP_FILES and OUT_FILES directory containing the I O files will be placed directly in the MARC WORKIN
152. twork board display graphic card are non critical HARDWARE requirements RAM gt 2 Gbyte Processor Speed gt 3 Ghz Suggested As an example the MARSCHALS codes has been executed successfully on a computer with the following characteristics Processor INTEL Pentium IV Memory 2 Gbyte RAM Operating System Linux DEBIAN 2 4 23 At the present the code works under a LINUX operating system A Windows version is also available the two versions differ only for some commands inside some code routines Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 21 169 2 4 The MARSCHALS L2 software files In this chapter the set of the files needed to run the MARSCHALS L2 codes is introduced The executable version on the computer delivered to ESA is provided with all the files needed to run the codes Code files Source files Data files Settings Pre processed input data files Ancillary files Auxiliary files Figure 2 2 The files needed to run the MARSCHALS L2 codes MARSCHALS Level 2 Prog Doe N IF AC GA 2007 06 LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 22 169 As specified in Section 2 1 the MARSCHALS L2 suite is composed by three codes each of thise is provided in source version by mean of a set of source files Among the code files also some batch files are provided to hel
153. ument target identifier indicating a generic retrieved quantity identifier indicating a generic molecule symbol of a retrieved species species e g H20 identifier indicating a generic namefile without extension This documentation uses the following general conventions Italic Font Denotes the name of files directories and subdirectories codes or modules of codes Bold Font Represents menus windows names and tool buttons Bold Italic Font Denotes book titles Plain Typewriter Font Denotes code fragments command lines contents of files and command names Italic Typewriter Font Represents a variable for which an actual value should be substituted Paths are in UNIX notation Forward Slashes MARSCHALS Level 2 TT Cc User Manual This page is intentionally left blank Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 Date 10 04 07 Page n 14 169 MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 15 169 2 Overview of MARSCHALS L2 software The MARSCHALS L2 software is a suite of codes oriented to the retrieval of the atmospheric constituents from a set of spectroscopic measurements These measurements come from the L1b of MARSCHALS project and different elaborations are necessary in order to retrieve all the requested profiles such as a preliminary treatment of measurement data the simulation of sy
154. usage also scales roughly as NSTR 2 azimuth angle of incident beam Use this parameter to relate the radiance output to fixed navigational headings For example if the sun is positioned at zenith 10 and azimuth 110 degrees measured clockwise from due north then setting PHI0 70 degrees will cause PHI to be interpreted as a compass direction In this example the forward scattering peak will be found at uzen 170 phi 70 Otherwise if PHIO is zero the default value PHI is interpreted as a relative azimuth angle i e relative to the forward scattering direction NOTE When IDAY is set PHIO is automatically set to the correct solar azimuth and the input value of PHIO is ignored radiation boundary conditions Prog Doc N IFAC GA 2007 06 LS Page n 168 169 MARSCHALS Level 2 Prog Doc N IFAC_GA_2007_06_LS U Manual Issue 3 Revision 1 I c Date 10 04 07 Page n 169 169 IBCND 0 general case boundary conditions any combination of beam illumination from the top see FBEAM isotropic illumination from the top see FISOT thermal emission from the top see TEMIS TTEMP internal thermal emission sources see TEMPER reflection at the bottom see LAMBER ALBEDO HL thermal emission from the bottom see BTEMP 1 isotropic illumination from top and bottom in order to get ALBEDO and transmissivity of the entire medium vs incident beam angle The only input variables considered in this case are NLY
155. utton Please notice that the make command works according to the standard arrangement of the file the user can change this arrangement but in this case to use the make and the make install command the makefile file have to be edited accordingly to the user defined arrangement of the files MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 i c User Manual Date 10 04 07 Page n 73 169 5 2 How to run MARC code Together with the executable MARSCHALS L2 software installed on the computer delivered to ESA a batch file is provided to run the MARC code The user must run the MARC code by using the batch file run_marc stored as a default in the MARC WORKING DIR subdirectory Please notice that the MARC code cannot be run directly from command line by means of itse executable file as a matter of fact the use of the run_marc batch file is mandatory since in such a file some user defined parameter needed to run the code are specified To run the MARC code please move to the subdirectory in which the run_marc batch file is stored the MARC WORKING DIR one in the default arrangement and at the prompt type run_marc and then press the enter button The user can change the default arrangements of the subdirectories in which the files needed to execute the MARC code are stored If a change is made in the default arrangement of the files as reported in 3 1 2 the batch file run_marc s
156. velength of 0 55um The cloud optical depth at other wavelengths is computed using the relation tau TCLOUD Q wl Q 0 55um where Q is the extinction efficiency which is a function of effective radius and wavelength see discussion of LWP for a definition of Q The code contains a look up table of Q that covers effective radii in the range 2 to 128um for water clouds and for a single effective radius of 106um for ice clouds The wavelengths range is 0 29 to 333 33 um for water clouds and 29 to 20 um for ice clouds When specifying an optical depth for a range of grid levels the second TCLOUD entry corresponding to the cloud top altitude is usually set to zero This produces a uniform distribution of opacity over the altitude range For example ZCLOUD 1 5 0 0 0 uniformly distributed opacity TCLOUD 10 0 0 0 0 for a cloud of extent 4 km NRE 10 20 2 5 optical depths per km effective radius ramps from 10 to 20 between 1 and 5km 4E 4h GE GR FE A linearly varying opacity distribution can be obtained by setting the second TCLOUD entry to a factor which represents the ratio of the opacity in the highest layer to that in the lowest layer For example tau total 10 tau 4 5km tau 1 2km 4 ZCLOUD 1 5 0 0 0 linearly distributed opacity TCLOUD 10 4 0 0 0 for a cloud of extent 4 km between tau 1 2km 1 between tau 2 3km 2 between tau 3 4km 3 between tau 4 5km 4 NOTE
157. w vis where w is a weighting factor between the two extremes and is given by Page n 158 169 ZBAER DBAER TBAER QBAER MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS Issue 3 Revision 1 User Manual Date 10 04 07 1 vis 1 23 y S ssn P 5 lt vis lt 23 1 5 1 23 w l i vis lt 5 w 0 pi vis gt 23 NOTE Visibility is defined as the horizontal distance in km at which a beam of light at 0 55um is attenuated by a factor of 0 02 n 0 sigma VIS 1n 02 or VIS 3 912 n 0 sigma where sigma is the aerosol absorption scattering cross section at 0 55 microns See Glossary of Meteorology American Meteorology Society 1959 Altitude grid for custom aerosol vertical profile km Up to MXLY altitude points may be specified ZBAER is active for all positive values of IAER Aerosol density at ZBAER altitude grid points active for all positive values of IAER Up to MXLY density values may be specified The number of density values must match the number of ZBAER The units used to specify aerosol density is arbitrary since the overall profile is scaled by the user specified total vertical optical depth The aerosol density at all computational grid points is found through logarithmic interpolation on the ZBAER and DBAER values The normal vertical profile from 5s is used when DBAER is unset For example ZBAER 0 1 100 DBAER 1000 500 1 specifies a aerosol density profile that drops by
158. w gt Y Zz li g 7 U p109941 Y Jo X pU g JUBbis jo aul y jo Kapu Z Jajuno9 iz x pur s3eg qse u s ApS mod Aps q i qeururou 3 j 9 A SSON guo nuog u3uoq EET HAY quuoz ypuoz muze 1 yuuoz 1 pees s3eg Auyend a ee WYBIg Jo oury uo Poo JUIN SUT Suma JoyNUSpPI SOT wj a Peel oe fore i a ae a fn ee oe ei Table 4 4 The structure of the scan_ _view dat file MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 i Cc Date 10 04 07 Page n 69 169 4 5 4 The observation_ dat files with the measured data Measurement data are reported in the file named observation_ dat where is a string referred to the specific scan This file contains all fundamental data necessary for the execution of MARC retrieval process It collects in a unique solution spectral data and geometric information of measurement separated for band and inside a single band arranged for line of sight The file refers to a single measurement scan The following list recalls the structure of data inside the file a detailed description of the format is given in RD 3 At the beginning of the file an auxiliary block of data is placed it provides a technical description of the scan and summarizes the number of bands extracted from L1b data and the number of lines of sight contained in the scan itself Then for each band the file reports the number of spectral points furthermore for e
159. w reports the versions of the input files used in the run of the code that have originated the output data contained in the output file according to the following list input_key kO ni n2 n3 n4 n5 being ko string with the name of the L1b data file ni number with the version ofthe instrument dat file n2 number with the version of the fov dat file n3 number with the version of the ils dat file n4 number with the version of the rejection dat file n5 number with the version of the settings dat file Table 5 9 Columns key word for input_key data In case of use of simulated measurement this kO name is replaced with the string simulation MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS User Manual Issue 3 Revision 1 Cc Date 10 04 07 Page n 127 169 6 OFM This chapter aims to describe the use of OFM with its implemented features The OFM is composed by two independents codes being the first SBDART code and the second the RTM code developed by MARSCHALS team The output generated by SBDART module are used as input of the RTM code The execution thus is performed in a sequence of the two components SBDART code RTM code Figure 6 1 The overall scheme of the OFM with the two codes SBDART is a software tool that computes plane parallel radiative transfer in clear and cloudy conditions within the Earth s atmosphere For a general description and rev
160. ware is provided to ESA in electronic form by means of a computer in which the source files are stored and compiled Thus together with the source files also an executable version of the software the it Please notices that according to the contract the MARSCHALS L2 user can directly run is provided To use this executable version it is not needed any installation and compilation procedure in this case the user can skip the reading of the paragraph 2 2 2 3 and 2 4 and the whole Chapter 3 related to the installation and the paragraph 4 1 5 1 and 6 1 related to the compilation of the codes MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 11 169 1 3 Relevant Document RD 1 MARSCHALS Level Product Format Specification Issue 1 rel 3 29 09 2003 RD2 MARSCHALS Level 2 Algorithm Theoretical Baseline Document IFAC GA 2007 01 LS issue 7 rel 1 S Baronti et al 15 January 2007 RD 3 MARSCHALS Level 2 Architectural Design Document IFAC GA 2007 03 LS issue 4 rel 1 S Baronti et al 25 March 2007 RD4 MARSCHALS Level 2 Theoretical Retrieval Study IFAC GA 2007 05 LS issue 5 rel 1 G Bazzini et al 05 April 2007 RD 5 Characterization of Millimetre Wave Spectroscopic Signatures ESTEC Contract No 16377 02 NL FF RD6 Ricchiazzi et al 1998 Bulletin of the American Meteorological Society October 1998 MA
161. y defines the altitude range of the scans to be extracted The Pre Processor will extract data related to the Lines Of Sight whose tangent altitudes lay in the range from altitude los 1 to altitude los 2 double double CONTRAST LEVEL _ THRESHOLD threshold This value defines the acceptable contrast range The Pre Processor will select those scans whose spectra show a contrast level greater than threshold The contrast level value is defined in the following All the other measurements will be discarded dimless double NOISE LEVEL _ THRESHOLD threshold This value defines the acceptable noise range The Pre Processor will select those scans whose spectra show a noise level greater than threshold All the other measurements will be discarded dimless double QUALITY NUMBER _ THRESHOLD threshold At the moment this settings option is not available dimless int MARSCHALS Level 2 Prog Doc N IFAC GA 2007 06 LS M l Issue 3 Revision 1 Cc User Manual Date 10 04 07 Page n 56 169 4 4 3 How to select the files generated by the pre processor The pre processor can extract some general information about the measurement campaign as well as the measurement data related to a selected scan The different modalities of extraction with the related value of the extraction type variable are specified by the means of the following parameter Key words re

MARSCHALS Level 2 Software User and Reference Manual

Contents

Download Pdf Manuals

Related Search

Related Contents