Ocean Assimilation Kit (OAK) User guide
Contents
1. DiagXXX stddevxa vector of strings Standard deviation of the error of the anal ysis DiagXXX stddevHxa vector of strings Standard deviation of the error of the ob served part of the analysis DiagXXX H strings the observation operator DiagXXX yo vector of strings The observations DiagXXX invsqrtR vector of strings The inverse of the root mean square error of the observations If a scalar observation point has been eliminated out of the model grid for example its weight is zero DiagXXX xa xf vector of strings The analysis increment DiagXXX yo Hxf vector of striygs the observation minus the model forecast at the observation points DiagXXX yo Hxa vector of strings the observation minus the model analysis at the observation points DiagXXX Hxa Hxf vector of strings analysis increment at the observation points DiagxXxx path string The path is prepended to all filenames spec 4 2 5 miscellaneous flags Key Type Description nbnest integer Number of nested grids assimnum integer Number between 1 and nbnest different for each model The model with assimnum does the assimilation runtype integer possible values of runtype are 0 do nothing i e a pure run of the model 1 still do not assimilate but compare model to observations 2 assimilate observations schemetype integer possible values of schemetype are 0 global assimilation default 1 local assimilation Zones need t2. The integer is a number ranging from 1 to the di mension of the error space n n vectors of file names are formed and represent a error mode in the state space Their norm represent the importance of the error mode and thus they are in general not normed Orthogonality is not necessary Optional keys ErrorSpace spacescale vector of strings Key Type Description ErrorSpace path string The path is prepended to all file names spec ified in ErrorSpace The current path is used by default ErrorSpace scale real Each error mode is multiplied by this real number The default is 1 Each error mode is multiplied element by el ement by this vector The default is a vector with all elements equal to 1 4 2 2 Zones When the local version of the assimilation algorithm schemetype 1 is used then the assimilation is performed in a number of zones independently Zones are defined by specifying a partition vector which has the same number of variables as the model state vector and each variable has the same size as the corresponding Model mask This vector contains only integer values starting with one and represent labels all elements in the state vector which have the same partition number belong to the same zone For example for a state vector with 5 elements and the partition vector p Ly 1 v2 1 x z3 p 2 41 v4 2 Xs 3 This partition vector defined three zones the first zone contains eleme
3. column major order as opposed to ncdump For Fortran binary files the order of the dimensions is also longitude latitude depth and time 3 The initialisation file With the module initfile a program can read an integer number floating number or a character string from an initialisation file Each line in this file is composed by a name called key an equal sign and the value For example runtype 2 Geoflow maxU 0 3 logfile assim log When the program search for example the key runtype it gets the integer 2 If a key is present several times in the same initialisation file then the last value found is taken The key can be composed by any alphanumeric character and by periods In par ticular spaces and a equal signs are not allowed within the key name The wild cards symbols and brackets are allowed but have a special meaning see Paragraph below Vectors of integers floats and character strings are also supported The values are sepa rated with commas and enclosed in brackets Model variables ETA TEM SAL Model maxCorrection 0 3 3 2 0 3 3 2 0 3 3 2 Blank lines are ignored and comments begin with the pound sign It is recommended to document the meaning and the possible values by a comments directly in the initiali sation file Entries in this files cannot be split across different lines Before assigning a value to a key you should know with type is
4. gitude of the observations Each string is a file name containing the lat itude of the observations Each string is a file name containing the depth of the observations The observation operator stored in a sparse matrix form per observations The observation operator stored in a sparse matrix form The path is prepended to all file names spec ified in ObsXXX The current path is used by default The optional keys are used to create the observation operator If it is applied to the state vector it extracts the observed variables at the location of the measurements Several ways exist to specify the observation operator 1 ObsXXX operator the observation operator is directly given by the non zero ele ments See also 4 2 3 2 ObsXXX variables and ObsXXX HperObs the non zero elements of the observation operator for each variable are given separately The first column in 9 x x matrix is ignored See also 4 2 3 3 ObsXXX variables ObsXXX gridX ObsXXX gridY and ObsXXX gridZ the obser vation operator is created by a trilinear interpolation using the module grids Note that the individual arrays in ObsXXX value ObsXXX rmse ObsXXX mask ObsXXX gridX ObsXXX gridY and ObsXXX gridZ should have the same size Format of the observation operator Only the non zero elements of the observation operator are specified in the 9 x n matrix in column major order where n is the number of non zero elements Each column has
5. home johndoe observations ctd04 1_04_aug_0747 TEM config TEM ligur SST X ligur SST Y ligur SST Z 32 41 scratch johndoe observtn ligur1999 07 02 SST gz scratch johndoe observtn ligur1999 07 03 SST gz scratch johndoe observtn ligur1999 07 04 SST gz scratch johndoe observtn ligur1999 07 10 SST gz scratch johndoe observtn ligur1999 07 11 SST gz 6 API 6 1 ufileformat uload filename matrix exclusion_value filename character of strings input The filename of the matrix to load with the extensions described in 2 matrix 1D 2D or 3D unallocated real pointer output The allocation of the output matrix is done inside the subroutine exclusion value real output The exclusion value usave filename matrix exclusion_value filename character of strings input The filename of the matrix to save matrix 1D 2D or 3D real matrix input The matrix to save exclusion_value real input The exclusion value References D P Dee and A Silva Data assimilation in the presence of forecast bias Quarterly Journal of the Royal Meteorological Society 124 269 295 1998 I Hoteit D T Pham and J Blum A simplified reduced order Kalman filtering and ap plication to altimetric data assimilation in Tropical Pacific Journal of Marine Systems 36 101 127 2002 18
6. the square root of P P S s 7 39 Based on x and S an ensemble can be reconstructed 7 x7 x4 4 yN 1 92 40 The bias aware analysis scheme of Dee and Silva 1998 is also implemented But the error space S is not computed 4 2 Configuration The initialisation file of the assimilation module is composed mainly by four sections configuration of the model model state vector position of the individual variables error space of the model observations to assimilate observations their position their error eventual diagnostics of the analysis and miscellaneous flags 4 2 1 The model The following code contains the definition of the multivariate state vector The key Model variables is a vector of character strings attributing to each variable a user chosen name The keys Model gridX Model gridY Model gridZ and Model mask are vectors of file names The files in Model gridX and Model gridY are degenerated and give the longitude and latitude of each variable The files in Model gridZ can be plain files and contains the depth The key Model mask is used to determine the sea land mask of each variable The exclusion value or missing value or FillValue in NetCDF terminology marks a land point all other values a sea points Every files assembled into a state vector should have physical values where mask assumes a sea point The shape of the arrays in Model gridX Model gridY Model gridZ and Model mask must
7. I 1 UAU UAU 1 1 Ss 20 Sf I 1 UAU UAU 1 1 UAU 8 21 sf I I I UAU S 22 S L UAU S7 23 sf UU UAUT St 24 SU L A UTS 25 e SSU I A A UTS 26 5 So we found a square root decomposition of P in terms of S U I A But in order to construct an ensemble from the columns of S its mean has to be zero So we will transform S so that the identity 26 is preserved One way to do this is S S U I A PUTO 27 The decomposition 18 can also be used in the computation of the Kalman gain K by K S I HSR HS HS TR 28 S UUT UAUT HS 7R 29 S U I A UT HS TR 30 Q is an orthogonal matrix chosen such that sum of all columns of S is zero This sum can be obtained by multiplying to the right with a vector N x 1 with all elements equal to 1 Inx Since SfIynx is zero S 1y 1 is also zero if U I A UTA Iya Ina 31 Q Inx U I A PUTIyx 32 We defined the normalised vectors w and v w Sali 33 v aU I A UTIyx 34 Q is thus a matrix which rotates w onto v Qw v 35 It can be computed by Q vw H v H w 36 where H v is the N x N 1 Householder matrix associated to the vector v i e all columns of H v are vectors perpendicular to v Hoteit et al 2002 Viv H v i j fori lt N 1 v J J Ivy TE 1 ors 37 vy sgn vy v H v y 38 ua Ivy 1 S is
8. Ocean Assimilation Kit OAK User guide Alexander Barth Luc Vandenbulcke April 9 2014 1 Structure of the software The software is structured in different modules e ufileformat Binary output and input of large 1D 2D or 3D matrices in the GHER or NetCDF e initfile Input of integers floating numbers strings and small vectors of those data types e matoper Basic matrix operating multiplication matrix inversion eigenvalue eigenvectors and singular value decomposition relying on BLAS and LAPACK e date module for conversion between modified Julian day number and Gregorian date e grids interpolation from one grid to another of 1D 2D or 3D data e rrsqrt The analysis equation e assimilation I O of state vector observations error space and observation op erator Analysis routine with input output and computation of diagnostics These modules can be either used for specific task with standalone programs 5 or by a hydrodynamic model in the case of a simulation assimilating observations The GHER hydrodynamic model drives the data assimilation modules trough the following subrou tines e dainit initialises of the data assimilation modules daobs loads of the next observation to assimilate e daanalysis performs the analysis e damoderr propagates the error covariance of the model 2 Module ufileformat This module is used for binary output and input of large 1D 2D or 3D matrices The GHER and a subset of t
9. be the same The string in Model path in prepended to each file names Example Model variables ETA TEM SAL Model gridX ligur X end ligur X ligur X Model gridY ligur Y end ligur Y ligur Y Model gridZ ligur Z end ligur Z ligur zZ Model mask ligur Z end ligur Z ligur zZ Model path u abarth soft Ligur3 Data For nested grids the variables of the same nested must be grouped and the groups must be orders according to the resolution started with the highest resolution one To each model grid is associated a Model gridnum one for the highest resolution one two of the next highest resolution one and so one Model variables TEM SAL TEM gt SAL Model gridX ligur X ligur X med X med X Model gridY ligur Y ligur Y med Y med Y Model gridZz ligur Z ligur Z med Z med Z Model mask ligur Z ligur Z med Z med Z Model gridnum 1 1 2 2 Model path u abarth soft Ligur3 Data Mandatory keys Key Type Description ErrorSpace dimension Errorspace init integer vector of strings The dimension of the error space Each string is a Fortran format containing an integer descriptor The format is converted into a file name with an internal write
10. ced by the time index Mandatory keys ObsXXX value ObsXXX rmse ObsXXX mask vector of strings vector of strings vector of strings Key Type Description ObsXXX time yyyy mm yyyy year minimum 1 digit integer ddTHH MM ss mm month 2 digits integer dd day 2 digits integer HH hour 2 digits integer MM min 2 dig ids integer SS second minimum 1 digit integer or real Each string is a file name containing the ac tual values of the observations Each string is a file name containing the root mean square error of the observations Each string is a file name containing the bi nary mask of the observations Values where the mask is different from 1 are rejected Optional keys Key Type Description ObsXXX ObsXXX ObsXXX ObsXXX ObsXXX ObsXXX ObsXXX ObsXXX variables names gridx gridY gridz HperObs operator path vector of strings vector of strings vector of strings vector of strings vector of strings vector of strings string string The names must correspond to the names chosen in Model variables Unknown names are treated as out of the grid and are not assimilated Each string is a description of the data type of the observations You can choose any name meaningful to you These names are only used for the log file The default names are Var01 Var02 Each string is a file name containing the lon
11. e its weight in DiagXXX invsqrtR is zero The state vector is specified as it is described in 5 1 If the program is called with three arguments applyobsoper lt initfile gt lt start time index gt lt end time index gt The action is repeated for all time indexes between the start and end time index 5 4 Program filteroper The standalone program filteroper generates a sparse matrix witch acts as a spatial filter in the model space filteroper lt initfile gt For each variable the filter is a Gaussian function z w y z f z Y Z x y z Ne ve ry ve 43 N is a normalisation factor taking in to account the land sea mask The parameters Ly L and L may be space dependent and have thus the same dimension as the state vector The required keys are Key Type Description Model mask vector of strings sea land mask of each variable Model gridX vector of strings longitude of each variable degenerated file Model gridY vector of strings latitude of each variable degenerated file Model gridZ vector of strings depth Model path string The path is prepended to all filenames spec ified in Model The current path is used by default Correlation lenx vector of strings parameter L in equation 43 Correlation leny vector of strings parameter L in equation 43 Correlation lenz vector of strings parameter L in equation 43 Correlation path string The path is prepended to all file
12. e current path is used by default If the program is called with three arguments assim lt initfile gt lt start time index gt lt end time index gt All assimilation cycles be between the two time indexes are performed in chronological order 5 2 Program genobsoper The standalone program genobsoper generate the observation matrix genobsoper lt initfile gt lt time index gt The first argument is the initialisation file and the second argument is the time index of the observation for witch the observation operator has to be created All keys described in 4 2 have the same meaning for the program genobsoper But the only diagnostic key used is DiagXXX H If the program is called with three arguments genobsoper lt initfile gt lt start time index gt lt end time index gt The action is repeated for all time indexes between the start and the end time index 5 3 Program applyobsoper The standalone program applyobsoper extracts from a state vector the observations applyobsoper lt initfile gt lt time index gt 14 The first argument is the initialisation file and the second argument is the time index of the observation for witch the observation operator has to be created All keys described in 4 2 have the same meaning for the program applyobsoper But the only diagnostic key used are DiagXXX Hxf and DiagXXX invsqrtR If a scalar observation point has been eliminated out of the model grid for exampl
13. epresent the error modes of the model forecast DiagXXX Ef vector of strings Each string is a Fortran format For the conversion into file names see the key ErrorSpace init The files represent the forecast ensemble DiagXXX diagPf vector of strings The diagonal elements of error covariance of the model forecast DiagXXX diagHPfHT vector of strings The diagonal elements of error covariance of the observed part of the model forecast DiagXXX stddevxf vector of strings Standard deviation of the error of the model forecast DiagXXX stddevHxf vector of strings Standard deviation of the error of the ob served part of the model forecast DiagXXX path string The path is prepended to all file names spec ified in DiagXXX The current path is used by default DiagXXX xa vector of strings the analysis ensemble mean DiagXXX Hxa vector of strings the observed part of the analysis DiagXXX Sa vector of strings Each string is a Fortran format For the conversion into filenames see the key ErrorSpace init The files represent the error modes of the analysis DiagXXX Ea vector of strings Each string is a Fortran format For the conversion into file names see the key ErrorSpace init The files represent the analysis ensemble DiagXXX diagPa vector string The diagonal elements of error covariance of the analysis DiagXXX diagHPaHT vector of strings The diagonal elements of error covariance of the observed part of the analysis
14. expected scalar or vector and number or characters If the type does not correspond the program will be stopped Sometimes a sequence of keys are attributed to the same values Obs001 path Obs002 path Obs003 path u abarth soft Ligur3 0bs u abarth soft Ligur3 Obs u abarth soft Ligur3 0bs In this case one can use wild cards and write the following Obs path u abarth soft Ligur3 0bs The meaning of the wild cards are the same as for file name generation of the Burne Shell see also man page of sh and gmatch 4 Assimilation module 4 1 Reduced order analysis Let N be the ensemble size n the size of the state vector and m the observation space di mension The best linear unbiased estimator BLUE of the model s state vector given the model forecast x with error covariance P and the observation y with error covariance R is given by x x x K y Hx 2 K PSH HP HT R 3 P P KHP 4 where H is the observation operator extracting the observed part of the state vector and P the error covariance of the analysis x From the ensemble of forecast states xf where k 1 N one can compute the ensemble mean eee A xf Jo xf 5 k l z and ensemble covariance f KA sO ZARIL FN e RR 6 We construct the columns of the matrix S by gt lI dOa 5 7 N 1 where Sf is an x N matrix which each column being the differe
15. he NetCDF format is currently supported The matrix can contain exclusion points holes Matrices A where the elements are a linear combination of the indices can also be efficiently represented A i j k ao ai aaj ask 1 Only the coefficient ag ao ag and a are stored These file are called degenerated For example the longitude and latitude of each grid point can often be expressed in this way For the GHER format each file represent a real matrix If the file names ends with gz then the file is uncompressed with gunzip in the user s temporary directory defined by the environment variable TMPDIR or by default in tmp Simple Fortran 90 style extraction can also by performed with the module ufileformat A coma separated list of indices or ranges of indices in parenthesis can be appended to the file name if only a subsection of the matrix should be loaded For example if the file toto TEM is a 10 x 10 x 10 matrix the file toto TEM 6 is 10x10x1 matrix containing all elements with the 3rd indices equal to 6 toto TEM end is 10x1x10 matrix containing all elements with the 2nd indices equal to 10 toto TEM 1 end 1 end is 10x10x10 matrix equal to the original matrix But no arithmetic with the indices for example toto TEM end 1 are allowed If data extraction is used with degenerated matrices the four coefficient are changed ac cordingly to the subsection chosen Data extraction and a
16. names spec ified in Correlation The current path is used by default Filter string file name of the filter 5 5 Program opermul opermul is a general purpose program witch multiply two sparse operators It can be used for example for multiplying a filter operator and a observation operator O O20 15 44 O is a operator mapping from space S to S2 Oz from S to Sz and thus the product from S to S3 opermul lt initfile gt The required keys are Key Type Description Spacel mask Spacel path Space2 mask Space2 path Space3 mask Space3 path Term1 Term2 Product vector of strings string vector of strings string vector of strings string string string string sea land mask of space S1 The path is prepended to all filenames spec ified in Space1 mask The current path is used by default sea land mask of space S The path is prepended to all filenames spec ified in Space2 mask The current path is used by default sea land mask of space S The path is prepended to all filenames spec ified in Space2 mask The current path is used by default file name of operator QO file name of operator Op file name of the product O3 5 6 Matlab utility GenObsFile The utility GenObsFile provides an easy way to save all the observations coming from various sources in a few files with the NetCDF format and creates the INIT file required b
17. nce between each member its ensemble mean Its mean over all columns it thus zero As many other assimilation schemes SEEK RRSQRT ESSE EnKF P is decomposed in terms of this square root matrix Sf pi SiS 8 Typically the number of ensemble members N is much smaller than the state vector size n We rewrite the Kalman Filter analysis by avoiding any matrix of the size n x n 1 K S sf H H s s H R 9 sf Hs HS HS R 10 S I HS TR HS HS R 11 Where the Sherman Morison Woodbury identity has been applied from 10 to 11 This identity can be expressed as ABT C BABT A BTC7B BTC 12 with A I B HS C R That is instead of performing the inverse in space of matrix A the inverse is done in the space of the matrix C We also substitute P in the expression of the analysis covariance error P Pp P KHP 13 sfs 7_Kus sf7 14 sisi Sf I HS R 1HS HS R Hss 15 s I 1 HS RHS HS RHS Ss ue In order to avoid to form P explicitly we need to express P also in terms of the square root matrices S Pass 17 This is possible when the following eigenvalue decomposition is made HS RHS UAU 18 where UTU I and where A is diagonal U and A are both of the size N x N Using the decomposition 18 in equation 16 one obtains po Sf I I UAUT H UAUT Ss 19 S
18. nts z and z2 the second zone x3 and x4 and the third zone x3 There should be no gaps in the partition vector For example the vector 1 1 2 2 4 7 would cause an error In practice the state vector is partitioned along water columns The assimilation is performed independently in each zone using only observations within the search radius given by Zones maxLength The weight of the observations 5 is multiplied by a Gaussian function i R R L exp d L 42 where d is the horizontal distance in m the first point of a zone and a single observa tion and L a length scale in m given by Zones corrLength Zones maxLength and Zones corrLength have the same size as the model state vector In most cases these values are constant can be specified by e g Zones corrLength const 30e3 30e3 Zones maxLength const 2000e3 2000e3 Key Type Description Zones partition Zones corrLength Zones maxLength vector of strings vector of strings vector of strings Each string is a file name containing the par tition file for the given model variable Each string is a file name containing the cor relation length Each string is a file name containing the maximum correlation length 4 2 3 The observations All set of simultaneous observation are ordered chronically and are attributed to a time index starting with 001 written always with three digits In the following keys XXX have to be repla
19. o be de fined moderrtype integer possible values of moderrtype are 0 optimal interpolation Pf constant 1 forgetting factor approximation biastype integer possible values of biastype are 0 standard bias blind analysis 1 A fraction of the error gamma is a sys tematic error and the rest 1 gamma is random Dee and Silva 1998 Bias gamma real fraction of the error with is systematic Bias init vector of string the initial estimation of the bias joinvectors integer If joinvectors is 1 then the variables of the nested grids will be assembled to one multi grid state vector logfile string File contains simple diagnostics such as rmse with observations debugfile string File contains debugging information is the code was compiled with the flag DDEBUG 13 5 Standalone programs 5 1 Program assim The standalone program assim can be used to test the assimilation The program can be called from the command line assim lt initfile gt lt time index gt The first argument is the initialisation file and the second argument is the time index of the observation to assimilate All keys described in 4 2 have the same meaning for the program assim But the forecast has to be specified by the following keys Key Type Description ForecastXXX value ForecastXXX path vector of strings string the forecast The path is prepended to all filenames spec ified in ForecastXXX value Th
20. rary amount of data lines The config line starts with the keyword config and has the following format config VAR X Y Z MJD VAR indicates how the observed data should be named in the INIT file TEM X might be a a complete path file name with the longitude data corre sponding to the observations b the keyword file if the longitude data is written in a file with the same file name as the actual data with extension X Y idem Z idem MJD points to the file containing the MJD time corresponding to the observations and might be a a complete path file name b the keyword file c a datum in the format 1999 12 31 d a datum in the MJD format 51251 e character limits to be found in the actual observations file name For example if the actual file name is home johndoe 51657 TEM MJD could be 15 19 as those are the indexes pointing to 51657 in the file name After each config line an arbitrary amount of observation files may be given The filenames may contain matrix delimiters as in 1 100 2 5 Example listfile 17 config TEM Lion X Lion Y Lion Z 1998 01 01 home johndoe observations Lion00000480 TEM gz end config SAL Lion X Lion Y Lion Z 1998 01 01 home johndoe observations Lion00000480 SAL gz end config TEM file file file file home johndoe observations ctd02 1_03_aug_2241 TEM home johndoe observations ctd03 1_03_aug_1840 TEM
21. the following structure 10 Observations Model var in i index j index k index var in i index j index k index Inter dex dex polation coeffi cient The first integer value is related to the observation The index of the variable is the position where the observed variable appears in ObsXXX value and i j k index are the three spatial indexes of a single scalar observation The integers in column 5 to 8 are related to the model state vector Again the index of the variable is the position where the observed variable appears in Model variables and i j k index are the three spatial indexes of a single scalar model forecast If one of the model indexes is 1 the corresponding observation is treated out of grid and the associated weight will be zero The column 9 is a real value between 0 and 1 in the case of a simple trilinear interpolation The observation operator can be generated offline using a trilinear interpolation with the tool genobsoper 4 2 4 Diagnostics All diagnostics are optional and the corresponding files are output 11 Key Type Description DiagXXX xf vector of strings the model forecast ensemble mean DiagXXX Hxf vector of strings the observed part of the model forecast DiagXXX Sf vector of strings Each string is a Fortran format For the conversion into file names see the key ErrorSpace init The files r
22. utomatic decompression can only be used for loading data A variable in a NetCDF file can be loaded by specifying a file name of the following form NetCDF_filename NetCDF_variable If the NetCDF file name end with gz then the file is uncompressed as with the GHER file format The data extraction follows also the same rules as above For example the following is a valid file name for loading a matrix file nc gz temp 1 The file file nc gz is first decompressed then the slice with the 3rd indices equal to 1 of the variable temp is returned to the calling program The special value for missing data is stored in the variables attribute missing data In the case of degenerated file the attribute shape must be present containing the shape of the matrix The actual value of the variable contains the coefficients a 2 1 Order of the dimensions The reported order of the dimensions depends on the tool that you are using to query and access a file Two types of ordering schemes exists column major order used by Fortran programs such as OAK row major order used by C programs such as ncdump The order of the dimensions for NetCDF follows the recommendation of the CF convention If you query your NetCDF files with ncdump the order of the dimensions should be time depth latitude longitude For a Fortran program reading this file the dimensions with automatically be longitude latitude depth and time since Fortran uses the
23. y the assimilation routines Options for GenObsFile must be specified in the header of the Matlab routine as de scribed below e initheader complete path amp file name of the file that must be copied on top of the INIT file This could be the model part of the INIT file diags complete path amp file of a sample diagnostic part of the INIT file The observation number should be replaced with lt INDEX gt and variable names with lt EXT gt This part will be adapted and copied for each observation set Example Diag lt INDEX gt Hxf xf lt EXT gt Outdir path where to store the new observations and INIT file Outfile prefix of the new observation files maxX minX maxY minMJD observations not within these ranges will be ignored when creating the new observation files 16 e rmse vector containing errors on the observations in the following order TEM SAL ETA other It will only be used by the assimilation routine if no other observation error covari ance R matrix is specified GenObsFile only uses values corresponding to variables present in your observations list e obstime time of the day at which observations should be assimilated e listfile complete path file name for the listfile which contains the original ob servations It is build using sections There must be at least one section in the listfile Each section contains a config line followed by an arbit
Download Pdf Manuals
Related Search