LIBTOOLS/lib/vis5d/src/v5d.h

+
+/* Vis5D version 5.1 */
+
+/*
+Vis5D system for visualizing five dimensional gridded data sets
+Copyright (C) 1990 - 1996 Bill Hibbard, Brian Paul, Dave Santek,
+and Andre Battaiola.
+
+This program is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 1, or (at your option)
+any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; if not, write to the Free Software
+Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+
+
+#ifndef V5D_H
+#define V5D_H
+
+
+/*
+ * A numeric version number which we can test for in utility programs which
+ * use the v5d functions.  For example, we can do tests like this:
+ * #if V5D_VERSION > 42
+ *      do something
+ * #else
+ *      do something else
+ * #endif
+ *
+ * If V5D_VERSION is not defined, then its value is considered to be zero.
+ */
+
+#define V5D_VERSION 42
+
+
+/*
+ * Define our own 1 and 2-byte data types.  We use these names to avoid
+ * collisions with types defined by the OS include files.
+ */
+typedef unsigned char V5Dubyte;     /* Must be 1 byte, except for cray */
+typedef unsigned short V5Dushort;   /* Must be 2 byte, except for cray */
+
+
+
+#define MISSING 1.0e35
+#define IS_MISSING(X)  ( (X) >= 1.0e30 )
+
+
+/* Limits on 5-D grid size:  (must match those in v5df.h!!!) */
+#define MAXVARS     100
+#define MAXTIMES    400
+#define MAXROWS     800
+#define MAXCOLUMNS  800 
+#define MAXLEVELS   100
+
+
+#ifdef VPP
+#define FLOAT double
+#else
+#define FLOAT float
+#endif
+
+/************************************************************************/
+/***                                                                  ***/
+/*** Functions for writing v5d files.  See README file for details.   ***/
+/*** These are the functions user's will want for writing file        ***/
+/*** converters, etc.                                                 ***/
+/***                                                                  ***/
+/************************************************************************/
+
+extern int v5dCreateSimple( const char *name,
+                            int numtimes, int numvars,
+                            int nr, int nc, int nl,
+                            const char varname[MAXVARS][10],
+                            const int timestamp[],
+                            const int datestamp[],
+                            float northlat, float latinc,
+                            float westlon, float loninc,
+                            float bottomhgt, float hgtinc );
+
+
+extern int v5dCreate( const char *name,
+                      int numtimes, int numvars,
+                      int nr, int nc, const int nl[],
+                      const char varname[MAXVARS][10],
+                      const int timestamp[],
+                      const int datestamp[],
+                      int compressmode,
+                      int projection,
+                      const FLOAT proj_args[],
+                      int vertical,
+                      const FLOAT vert_args[] );
+
+
+extern int v5dWrite( int time, int var, const FLOAT data[] );
+
+extern int v5dClose( void );
+
+
+extern int v5dSetLowLev( int lowlev[] );
+
+extern int v5dSetUnits( int var, const char *units );
+
+
+
+/************************************************************************/
+/***                                                                  ***/
+/*** Definition of v5d struct and function prototypes.                ***/
+/*** These functions are used by vis5d and advanced v5d utilities.    ***/
+/***                                                                  ***/
+/************************************************************************/
+
+#define MAXPROJARGS 100
+#define MAXVERTARGS (MAXLEVELS+1)
+
+/*
+ * This struct describes the structure of a .v5d file.
+ */
+typedef struct {
+    /* PUBLIC (user can freely read, sometimes write, these fields) */
+        int NumTimes;                   /* Number of time steps */
+        int NumVars;                    /* Number of variables */
+        int Nr;                         /* Number of rows */
+        int Nc;                         /* Number of columns */
+        int Nl[MAXVARS];                /* Number of levels per variable */
+        int LowLev[MAXVARS];            /* Lowest level per variable */
+        char VarName[MAXVARS][10];      /* 9-character variable names */
+        char Units[MAXVARS][20];        /* 19-character units for variables */
+        int TimeStamp[MAXTIMES];        /* Time in HHMMSS format */
+        int DateStamp[MAXTIMES];        /* Date in YYDDD format */
+        float MinVal[MAXVARS];          /* Minimum variable data values */
+        float MaxVal[MAXVARS];          /* Maximum variable data values */
+
+        /* This info is used for external function computation */
+        short McFile[MAXTIMES][MAXVARS];/* McIDAS file number in 1..9999 */
+        short McGrid[MAXTIMES][MAXVARS];/* McIDAS grid number in 1..? */
+
+        int VerticalSystem;             /* Which vertical coordinate system */
+        float VertArgs[MAXVERTARGS];    /* Vert. Coord. Sys. arguments... */
+
+        /*
+        IF VerticalSystem==0 THEN
+                -- Linear scale, equally-spaced levels in generic units
+                VertArgs[0] = Height of bottom-most grid level in generic units
+                VertArgs[1] = Increment between levels in generic units
+        ELSE IF VerticalSystem==1 THEN
+                -- Linear scale, equally-spaced levels in km
+                VertArgs[0] = Height of bottom grid level in km
+                VertArgs[1] = Increment between levels in km
+        ELSE IF VerticalSystem==2 THEN
+                -- Linear scale, Unequally spaced levels in km
+                VertArgs[0] = Height of grid level 0 (bottom) in km
+                ...                ...
+                VertArgs[n] = Height of grid level n in km
+        ELSE IF VerticalSystem==3 THEN
+                -- Linear scale, Unequally spaced levels in mb
+                VertArgs[0] = Pressure of grid level 0 (bottom) in mb
+                ...             ...
+                VertArgs[n] = Pressure of grid level n in mb
+        ENDIF
+        */
+
+        int Projection;                     /* Which map projection */
+        float ProjArgs[MAXPROJARGS];        /* Map projection arguments... */
+
+        /*
+        IF Projection==0 THEN
+                -- Rectilinear grid, generic units
+                ProjArgs[0] = North bound, Y coordinate of grid row 0
+                ProjArgs[1] = West bound, X coordiante of grid column 0
+                ProjArgs[2] = Increment between rows
+                ProjArgs[3] = Increment between colums
+                NOTES: X coordinates increase to the right, Y increase upward.
+                NOTES: Coordinate system is right-handed.
+        ELSE IF Projection==1 THEN
+                -- Cylindrical equidistant (Old VIS-5D)
+                -- Rectilinear grid in lat/lon
+                ProjArgs[0] = Latitude of grid row 0, north bound, in degrees
+                ProjArgs[1] = Longitude of grid column 0, west bound, in deg.
+                ProjArgs[2] = Increment between rows in degrees
+                ProjArgs[3] = Increment between rows in degrees
+                NOTES: Coordinates (degrees) increase to the left and upward.
+        ELSE IF Projection==2 THEN
+                -- Lambert conformal
+                ProjArgs[0] = Standared Latitude 1 of conic projection
+                ProjArgs[1] = Standared Latitude 2 of conic projection
+                ProjArgs[2] = Row of North/South pole
+                ProjArgs[3] = Column of North/South pole
+                ProjArgs[4] = Longitude which is parallel to columns
+                ProjArgs[5] = Increment between grid columns in km
+        ELSE IF Projection==3 THEN
+                -- Polar Stereographic
+                ProjArgs[0] = Latitude of center of projection
+                ProjArgs[1] = Longitude of center of projection
+                ProjArgs[2] = Grid row of center of projection
+                ProjArgs[3] = Grid column of center of projection
+                ProjArgs[4] = Increment between grid columns at center in km
+        ELSE IF Projection==4 THEN
+                -- Rotated
+                ProjArgs[0] = Latitude on rotated globe of grid row 0
+                ProjArgs[1] = Longitude on rotated globe of grid column 0
+                ProjArgs[2] = Degrees of latitude on rotated globe between
+                                grid rows
+                ProjArgs[3] = Degrees of longitude on rotated globe between
+                                grid columns
+                ProjArgs[4] = Earth latitude of (0, 0) on rotated globe
+                ProjArgs[5] = Earth longitude of (0, 0) on rotated globe
+                ProjArgs[6] = Clockwise rotation of rotated globe in degrees
+        ENDIF
+        */
+
+        int CompressMode;        /* 1, 2 or 4 = # bytes per grid point */
+        char FileVersion[10];    /* 9-character version number */
+
+    /* PRIVATE (not to be touched by user code) */
+        unsigned int FileFormat; /* COMP5D file version or 0 if .v5d */
+        int FileDesc;            /* Unix file descriptor */
+        char Mode;               /* 'r' = read, 'w' = write */
+        int CurPos;              /* current position of file pointer */
+        int FirstGridPos;        /* position of first grid in file */
+        int GridSize[MAXVARS];   /* size of each grid */
+        int SumGridSizes;        /* sum of GridSize[0..NumVars-1] */
+} v5dstruct;
+
+
+
+extern float pressure_to_height( float pressure);
+
+extern float height_to_pressure( float height );
+
+
+
+
+extern int v5dYYDDDtoDays( int yyddd );
+
+extern int v5dHHMMSStoSeconds( int hhmmss );
+
+extern int v5dDaysToYYDDD( int days );
+
+extern int v5dSecondsToHHMMSS( int seconds );
+
+
+extern void v5dPrintStruct( const v5dstruct *v );
+
+
+extern v5dstruct *v5dNewStruct( void );
+
+extern void v5dFreeStruct( v5dstruct* v );
+
+extern void v5dInitStruct( v5dstruct *v );
+
+extern int v5dVerifyStruct( const v5dstruct *v );
+
+
+extern void v5dCompressGrid( int nr, int nc, int nl, int compressmode,
+                             const float data[], void *compdata,
+                             float ga[], float gb[],
+                             float *minval, float *maxval );
+
+
+extern void v5dDecompressGrid( int nr, int nc, int nl, int compressmode,
+                               void *compdata,
+                               float ga[], float gb[],
+                               float data[] );
+
+
+extern int v5dSizeofGrid( const v5dstruct *v, int time, int var );
+
+
+extern v5dstruct *v5dOpenFile( const char *filename, v5dstruct *v );
+
+
+extern int v5dCreateFile( const char *filename, v5dstruct *v );
+
+
+extern v5dstruct *v5dUpdateFile( const char *filename, v5dstruct *v );
+
+
+extern int v5dCloseFile( v5dstruct *v );
+
+
+extern int v5dReadCompressedGrid( v5dstruct *v,
+                                  int time, int var,
+                                  float *ga, float *gb,
+                                  void *compdata );
+
+
+extern int v5dReadGrid( v5dstruct *v, int time, int var, float data[] );
+
+
+extern int v5dWriteCompressedGrid( const v5dstruct *v,
+                                   int time, int var,
+                                   const float *ga, const float *gb,
+                                   const void *compdata );
+
+
+extern int v5dWriteGrid( v5dstruct *v, int time, int var, const float data[] );
+
+
+
+#endif

LIBTOOLS/lib/vis5d/src/vis5d.h

+
+/* Vis5D version 5.1 */
+
+/*
+Vis5D system for visualizing five dimensional gridded data sets
+Copyright (C) 1990-1997 Bill Hibbard, Brian Paul, Dave Santek,
+and Andre Battaiola.
+
+This program is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 1, or (at your option)
+any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; if not, write to the Free Software
+Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+
+/*
+ * This configuration file contains options which can be safely
+ * changed by the user.
+ */
+
+
+
+#ifndef VIS5D_H
+#define VIS5D_H
+
+
+/*
+ * Amount of physical RAM in megabytes:
+ * vis5d normally uses a bounded amount of memory to avoid swapping.
+ * When the limit is reached, the least-recently-viewed graphics will
+ * be deallocated.  If MBS is set to 0, however, vis5d will use ordinary
+ * malloc/free and not deallocate graphics (ok for systems with a lot
+ * of memory (>=128MB)).
+ */
+/* Default Value: 32 */
+#define MBS 128
+
+
+
+/* Default topography file: */
+#define TOPOFILE "/home/chajpmnt/chajp/ADD/data/EARTH.TOPO"
+
+
+/* Default map lines files: */
+#define WORLDFILE "/home/chajpmnt/chajp/ADD/data/OUTLSUPW"
+#define USAFILE "/home/chajpmnt/chajp/ADD/data/OUTLUSAM"
+
+
+/* Default filename of Tcl startup commands: */
+#define TCL_STARTUP_FILE "vis5d.tcl"
+
+
+/* Default directory to search for user functions: */
+#define FUNCTION_PATH "userfuncs"
+
+
+/* Default animation rate in milliseconds: */
+#define ANIMRATE 100
+
+
+/* Default scale and exponent values for logrithmic vertical coordinate system: */
+#define DEFAULT_LOG_SCALE  1012.5
+#define DEFAULT_LOG_EXP  -7.2
+
+
+#define DEFAULT_SOUNDFONTNAME "6x12"
+
+/**********************************************************************/
+/**********************************************************************/
+/***          USERS:  DON'T CHANGE ANYTHING BEYOND THIS POINT       ***/
+/**********************************************************************/
+/**********************************************************************/
+
+/*
+ * Define BIG_GFX to allow larger isosurfaces, contour slices, etc. if
+ * there's enough memory.
+#if MBS==0 || MBS>=128
+#  define BIG_GFX
+#endif
+ */
+
+#define BIG_GFX
+
+
+/*
+ * Shared by code above and below API:
+ */
+#define MAX_LABEL   1000
+#define MAX_FUNCS   100
+
+
+
+#endif

LIBTOOLS/readme/LATEX/Makefile

+tools.ps : tools.dvi
+	dvips -o $@ $<
+
+tools.dvi : tools.tex
+	latex tools.tex
+	latex tools.tex
+	latex tools.tex
+
+clean:
+	rm -f *.aux *.log *.toc *.dvi
+
+realclean: clean
+	rm -f *.ps

LIBTOOLS/readme/LATEX/conv2dia.tex

+\section{Conversion of FM synchronous file to diachronic format}
+Short description is given here, readers must refer to the original documentation on the Meso-NH web site:
+``{\sc traitement graphique des fichiers synchrones produits par le mod\`ele
+mesonh}, J. Duron''. 
+
+\subsection{Synchronous and diachronic formats} \label{diachro_file}
+The Meso-NH graphic utility ({\tt diaprog}) works on FM files which are on
+diachronic format. A diachronic FM file is either
+\begin{itemize}
+\item
+a file produced during the simulation 
+which contain time series of self-documented informations
+(e.g. file with name CEXP.1.CSEG.000).
+An information is one of the following: 
+\subitem - a
+3-dimensional, 2-dimensional, 1-dimensional or 0-dimensional field (eventually
+time-averaged, or compressed in one direction): type {\sc cart}, 
+\subitem - a set of vertical profiles at points checking some criteria:
+type {\sc mask}, 
+\subitem - spectral coefficients obtained by FFT along the X or Y direction:
+type {\sc spxy},
+\subitem - pseudo-observations (ground station: type {\sc ssol};
+dropsonde: type {\sc drst}; radiosonde: type {\sc rspl};
+airborne radar: type {\sc rapl}).
+ \\
+A diachronic file can contains informations of one or several previous types
+stored at different time frequency.
+For a whole description about the diachronic file type, reader must refer
+to the original documentation on the Meso-NH web site:
+``{\sc cr\'eation et exploitation de fichiers diachroniques}, J. Duron''. 
+\end{itemize}
+or
+\begin{itemize}
+\item a `pseudo'-diachronic file resulting of the conversion of a synchronous
+file (e.g. with name CEXP.1.CSEG.00n where n$>$0).
+Recall that such a file contains all the pronostic fields of the model at one 
+instant (initial or during the simulation).
+When converted it is a 'pseudo'-diachronic file, because it contains only one 
+instant and one type of diachronic information ({\sc cart}).
+The next subsection presents the conversion tool (named \texttt{conv2dia})
+to apply to synchronous files, necessary step to use \texttt{diaprog} graphic
+tool.
+\end{itemize}
+
+\subsection{{\tt conv2dia} tool}
+The conversion tool works on files produced by
+the initialisation programs ({\sc prep\_pgd, prep\_ideal\_case,
+prep\_real\_case}), the model simulation, or the post-processing program
+({\tt\sc diag}). It allows to convert one synchronous file onto one diachronic 
+file, as well as merge several synchronous files with chronological times
+(outputs of one run, or files initialised from large-scale model)
+onto one diachronic file.
+
+With {\tt conv2dia.elim} tool, you can choose not to convert all the fields of
+the input file(s). The pronostic fields at $t-dt$ instant, or at $t$ instant,
+or any other fields can be eliminated.
+With {\tt conv2dia.select} tool, you have to indicate the fields to select
+for conversion.
+This is done to reduce the size of the output file.
+
+The output file contains informations whose type is {\sc cart} stored in arrays
+with size of {\tt (IIU*IJU*IKU), (IIU*IJU), (IIU*IKU),} or 1.
+
+
+\subsection{Example}
+
+Only the binary (\textsc{LFI}) part of the input FM files is required
+in the current directory (split the FM file with the {\tt fm2deslfi} 
+script if not).
+
+All characters typed on keyboard are saved in {\tt dirconv.elim} or 
+{\tt dirconv.select} file, it can be appended and used as input (after being 
+renamed) for the next call of the tool
+\newline (e.g.  {\tt conv2dia.elim < dirconv.elim.ex}).
+
+Below is the example of questions when {\tt conv2dia.elim} is invoked.
+
+\small
+\begin{tabular}{l}
+\\
+\\
+{\tt ENTER NUMBER OF INPUT FM FILES}  \\
+{\tt\it 2 }  \\
+{\tt ENTER FM FILE NAME}  \\
+{\tt\it CEXP.1.CSEG.001}  \\
+{\tt ENTER FM FILE NAME}  \\
+{\tt\it CEXP.1.CSEG.002}  \\
+{\tt ENTER DIACHRONIC FILE NAME}  \\
+{\tt\it CEXP.1.CSEG.1-2.dia}  \\
+{\tt DELETION OF PARAMETERS AT TIME t-dt ? (enter 1) } \\
+{\tt DELETION OF PARAMETERS AT TIME t    ? (enter 2) } \\
+{\tt NO DELETION                         ? (enter 0) } \\
+{\tt\it 2 }  \\
+{\tt Do you want to suppress others parameters ? (y/n) }\\
+{\tt\it y }  \\
+{\tt Enter their names in UPPERCASE  (1/1 line) }\\
+{\tt End by END}\\
+{\tt\it DTHCONV }  \\
+{\tt\it DRVCONV }  \\
+{\tt\it END }  \\
+\end{tabular}
+\normalsize

LIBTOOLS/readme/LATEX/intro.tex

+\section{Introduction}
+
+After initialisation, run of the model or computation of diagnostics, 
+output Meso-NH files can be convert into other formats of files. 
+The present documentation aims at describ the differents tools which can be
+applied to the binary part of FM files (their suffix is {\bf .lfi}).
+Most of these tools can be run on the user local
+computer (Linux PC or HP workstation).
+\\
+
+First, the compression tool \texttt{lfiz} and the conversion
+tool \texttt{conv2dia} dealing with FM files (synchronous and diachronic)
+as input and output, are described. 
+The next sections concern tools dealing with other formats than
+FM: conversions with \texttt{lfi2cdf}, \texttt{lfi2grb} and \texttt{lfi2v5d}.
+A set of tools for reading diachronic FM files and dealing with diachronic 
+informations is presented: \texttt{extractdia}, \texttt{mesonh2obs} and 
+\texttt{obs2mesonh} (the 2 latest aim at help users to compare MesoNH outputs to
+observations).
+\\
+
+The figure \ref{fig:fic1} shows when a FM file is either \underline{synchronous}
+(contains the values of all the fields corresponding to the same instant of the
+simulation) or \underline{diachronic} (contains time series of some fields
+obtained during the run of the model). 
+Then the figure \ref{fig:toolstab} resumes the tools which can be applied to a
+FM file according its type, one of the two previous ones. \\
+
+\begin{figure}[htb]
+\psfig{file=fic1.eps,width=17cm} 
+\caption{Type of FM files after a MesoNH program\label{fig:fic1}}
+\end{figure}
+
+\begin{figure}[htb]
+\centerline{\psfig{file=toolstab.eps,width=17cm} }
+\caption{Which tools on FM files? \label{fig:toolstab}}
+\end{figure}
+

LIBTOOLS/readme/LATEX/lfi2cdf.tex

+\section{Conversion to NetCDF files}
+
+\subsection{{\tt lfi2cdf} tool}
+
+The \texttt{lfi2cdf} tool converts the binary part (or LFI file) of a
+FM file (synchronous or diachronic) into a NetCDF file.  All the fields
+(or more precisely all the LFI articles) contained in the input LFI file
+are copied to the NetCDF output file with their values unchanged. As
+a LFI article does not hold any information on the variable, the tool
+tries to describe the corresponding NetCDF variable by using~:
+
+\begin{itemize}
+\item 3 LFI articles: \texttt{IMAX, JMAX,} and \texttt{KMAX}
+  if they are available in the LFI input file. These articles may
+  provide the NetCDF dimensions \texttt{DIMX, DIMY,} and \texttt{DIMZ}
+  of an array variable. If these variables are not available in the
+  input file, the tool treats each array variable as a 1D array.
+\item a small database implemented as a structure array in the
+  \texttt{lfi2cdf} source file \texttt{fieldtype.f90}. This array
+  holds the type (\texttt{REAL, INTEGER, LOGICAL}\ldots) of every
+  common LFI article. When an article is not present in this database,
+  its name is displayed on \texttt{stdout} by the running tool, and
+  the corresponding values are always considered as \texttt{REAL}
+  values. A new LFI article type description can be easily added in
+  the \texttt{fieldtype.f90} source file and the tool must be then
+  recompiled.
+\end{itemize}
+
+\subsubsection{Usage}
+The binary part of the FM file is required in the current directory.
+The following commands convert a file \texttt{myfile.lfi} from LFI to NetCDF:
+
+\begin{verbatim}
+lfi2cdf myfile.lfi
+\end{verbatim} 
+or
+\begin{verbatim}
+lfi2cdf myfile
+\end{verbatim}
+
+\noindent The output NetCDF file is named:
+\texttt{myfile.cdf}. 
+%myfile{\bf .cdf}. 
+It can easily be manipulated by NetCDF tools\footnote{see
+freely available NetCDF software at  http://www.unidata.ucar.edu/packages/netcdf/software.html} like
+\texttt{ncdump}, \texttt{ncview}, or \texttt{NCO} operators.\\
+
+\noindent In the same way, you will convert a NetCDF
+file \texttt{myfile.cdf} back to LFI format by typing:
+
+\begin{verbatim}
+cdf2lfi myfile.cdf
+\end{verbatim}
+or
+\begin{verbatim}
+cdf2lfi myfile
+\end{verbatim}
+The output LFI file is then named: \texttt{myfile.lfi}
+
+
+\subsection{{\tt extractdia} tool}
+The \texttt{extractdia} tool converts a diachronic FM file into a NetCDF file after an extraction of a list of fields and an optional extraction of a sub-domain.  See the section \ref{extractdia}.
+
+
+%%% Local Variables: 
+%%% mode: latex
+%%% TeX-master: "tools"
+%%% End: 

LIBTOOLS/readme/LATEX/lfiz.tex

+\section{Compression of FM files} 
+
+A specific compression tool has been developed for FM files.  This
+tool, called {\tt lfiz}, was first devoted for files that will be
+explored by the graphic utility {\tt diaprog}. In fact, it is also
+used for files used during a simulation (initial and coupling files)
+to reduce the data storage.  Some information of how the compression
+works is given here, its execution is particularly easy.
+
+\subsection{{\tt lfiz} tool}
+
+The \texttt{lfiz} tool works on the binary part (LFI file) of a FM
+file, synchronous or diachronic.  It is a lossy compression tool.
+The compressed articles are exclusively the 2-dimensional or
+3-dimensional \texttt{REAL} fields. When dealing with 3D fields the tool works
+with each 2D plane on every vertical level. The initial values stored
+with 64-bit \texttt{REAL} precision are first converted into 32-bit
+\texttt{REAL} precision and then compressed by mapping the 32-bit
+real values upon 16-bit integer values (with a possible isolation of
+extrema values).  The better compression is
+achieved for fields with small value range.  For fields with missing
+value (e.g.  2-dimensional fields with land-sea mask), the extremum
+value is excluded and the compression is done on significant values of
+the field. The minimum compression ratio is 4 for each 2D or 3D
+\texttt{REAL} compressed field.
+
+\subsection{{\tt unlfiz} tool}
+The \texttt{unlfiz} tool will restore the 64-bit \texttt{REAL} value size to all
+the compressed LFI articles. However, each previously compressed article
+will gain no more than a 32-bit \texttt{REAL} precision because of the lossy
+technique involved above.
+
+
+\subsection{Usage}
+The binary part of the FM file is required in the current
+directory. To compress the file \texttt{myfile.lfi}, you can type:
+
+\begin{verbatim}
+lfiz myfile.lfi
+\end{verbatim}
+
+\noindent This will produce the compressed file \texttt{myfile.Z.lfi}\\
+
+
+\noindent In the same way, to uncompress the file \texttt{myfile.Z.lfi}, you can
+type:
+\begin{verbatim}
+unlfiz myfile.Z.lfi
+\end{verbatim}
+
+\noindent The output file \texttt{myfile.lfi} is a valid LFI file but the LFI
+articles previously compressed are 64-bit \texttt{REAL} with no more than 32-bit
+\texttt{REAL} precision.
+
+
+
+
+%%% Local Variables: 
+%%% mode: latex
+%%% TeX-master: "tools"
+%%% End: 

LIBTOOLS/readme/compute_r00.nam

+&NAM_STO_FILE CFILES(1)='AR40_mc2_19990921.00d.Z',
+              CFILES(2)='AR40_mc2_19990920.12d.Z',
+              CFILES(3)='AR40_mc2_19990920.00d.Z',
+              CFILES(4)='AR40_mc2_19990919.12d.Z', 
+              CFILES(5)='AR40_mc2_19990919.00d.Z' /
+&NAM_FIELD  CFIELD_LAG(1)='THM',
+            CFIELD_LAG(2)='RVM' /