Planets - Recent changes [en]

The IOIPSL Library

Emillour — Mon, 01 Jun 2026 13:29:29 GMT

Quick Install and Run

Emillour — Fri, 15 May 2026 14:33:02 GMT

The IOIPSL Library

Emillour — Fri, 15 May 2026 14:08:20 GMT

Other GCM Configurations worth knowing about

Emillour — Fri, 15 May 2026 09:58:43 GMT

Generic unicolumn rcm1d program

Emillour — Fri, 15 May 2026 09:52:32 GMT

Created page with "It is possible to run the Gneric PCM in a uni-column (aka "single-column") configuration: this can be done via the '''rcm1d''' program; quite useful form some first studies bu..."

New page

It is possible to run the Gneric PCM in a uni-column (aka "single-column") configuration: this can be done via the '''rcm1d''' program; quite useful form some first studies but also when developing and testing parametrizations.

== Compilation ==
The main program '''rcm1d''' is compiled using the same compilation script, [[The makelmdz fcm GCM Compilation Script|makelmdz_fcm]], as for the 3D Generic lon-lat PCM. Nevertheless, there are few modifications:
* the ''-d'' option requires only one argument, the number of vertical levels;
* the main program to compile is ''rcm1d'' rather than ''gcm''.
So for instance to compile a case for 26 vertical levels one would run something like:
<syntaxhighlight lang="bash">
makelmdz_fcm -arch somearch -d 26 -p generic rcm1d
</syntaxhighlight>
Note that it is possible to compile and run with the XIOS library, which will require compiling in "mpi" mode (required by XIOS), even though the run will be serial:
<syntaxhighlight lang="bash">
makelmdz_fcm -arch somearch -d 26 -p generic -parallel mpi -io xios rcm1d
</syntaxhighlight>

== Inputs ==
Just like the 3D GCM, the ''rcm1d'' program needs some inputs to run. The needed files are:
* <code>z2sig.def</code> for the definition of vertical levels;
* <code>traceur.def</code> for the definition of tracers that the user wants the model to run with;
* <code>callphys.def</code> for the definition of parametrizations that the user wants the model to run with;
* <code>rcm1d.def</code> for the run configuration, which is similar to the 3D [[The run.def Input File|run.def]]; see [[The rcm1d.def Input File]]. In practice the <code>rcm1d.def</code> file is in fact copied as <code>run.def</code> by the rcm1d program when it runs.

Unlike the 3D GCMs, the ''rcm1d'' program can run without any start files, which is its default (<code>restart=.false.</code>). In this setup, one can (and often needs) needs to provide initial profiles of each tracer. These consist in files called ''profile_sometracername'' containing column-wise the initial values of the considered tracer. Then, the first line corresponds to the surface tracer and the following lines correspond to the layers. At the end of a 1D simulation, ''rcm1d'' outputs a restart file '''restart1D.nc''' which can be used as an initial condition for a following run.

If <code>restart=.true.</code> in <code>rcm1d.def</code> then the program will look for start files '''start1D.nc''' and '''startfi.nc''' and use these as initial conditions.

== Outputs ==
If compiled without XIOS then ''rcm1d'' will output <code>diagfi.nc</code> and <code>stats.nc</code> files, just like the 3D GCM and the optional <code>diagfi.def</code> and <code>stats.def</code> files can be respectively added to specify which variables need be outputted.
Likewise if compiled with XIOS then ''rcm1d'' will output all files as specified in the relevant xml files, just like the 3D GCM.

As mentioned above, at the end of a simulation <code>rcm1d</code> also outputs a '''restart1D.nc''' file containing the final computed state.

[[Category:Generic-Model]]
[[Category:Generic-1D]]

@@ Line 28: / Line 28: @@
 </syntaxhighlight>
 where the trailing argument ''IOIPSL'' is customizable to whatever (non existing) subdirectory name one wants to have the library put in.
 === The makeioipsl_fcm compilation script and related FCM architecture files ===
@@ Line 76: / Line 81: @@
 As an alternative we have put a frozen version (revision 7509) of the IOIPSL source code online here:
 https://web.lmd.jussieu.fr/~lmdz/planets/alternatives/IOIPSL.r7589.tar.gz
 [[Category:WhatIs]]

@@ Line 42: / Line 42: @@
 svn checkout https://forge.ipsl.jussieu.fr/fcm/svn/PATCHED/FCM_V1.2
 </syntaxhighlight>
 You'll then need to add the resulting FCM_V1.2/bin to your PATH environment variable so that the command "fcm" may be used from anywhere on your machine. e.g. by adding the following line in your .bashrc (the .bashrc file is a hidden configuration script in your home directory (~/.bashrc) that runs whenever you start a new Bash shell):
 <syntaxhighlight lang="bash">

@@ Line 69: / Line 69: @@
 * Recent versions of the IOIPSL fails to work properly if compiled by gfortran version 4.8.5 (see e.g. https://trac.lmd.jussieu.fr/Planeto/ticket/62 ) but works fine with more recent versions of the compiler (tested with versions 7.2+)
 * As of April 2022 the IOISPL distribution requires username/password authentication.
 [[Category:WhatIs]]

@@ Line 679: / Line 679: @@
 TBC ML
-= 1D setup =
+= 1D setups =
-== rcm1d test case ==
+== rcm1d program ==
 Running the model in 1D (i.e. considering simply a column of atmosphere) is a common first step to test a new setup. To do so, you first have to compile the 1D version of the model. The command line is very similar to [[Quick_Install_and_Run#Compiling a test case (early Mars)|the one for the 3D]], except for 2 changes:
@@ Line 688: / Line 688: @@
 It will generate a file called '''rcm1d_XX_phyxxx_seq.e''', where ''XX'' and ''phyxxx'' are the vertical resolution and the physics package, respectively.
-Then, copy the executable in your working directory.
+Check out the [[Generic unicolumn rcm1d program| dedicated page about rcm1d]] for more details.
-Note that the '''.def''' files differ a bit from the 3D case: [[The_run.def_Input_File|'''run.def''']] is replaced with [[The_rcm1d.def_Input_File|'''rcm1d.def''']], which contains more general information. Indeed, the 1D model does not use [[The_start.nc_and_startfi.nc_input_files|'''start.nc''']] or [[The_start.nc_and_startfi.nc_input_files|'''startfi.nc''']] files to initialize. Instead it reads everything from the '''.def''' files. You can find examples of 1D configuration in ''LMDZ.GENERIC/deftank'' (e.g. '''rcm1d.def.earlymars''', '''rcm1d.def.earth'''), the best thing is to have a look at them.
+Note that the '''.def''' files differ a bit from the 3D case: [[The_run.def_Input_File|'''run.def''']] is replaced with [[The_rcm1d.def_Input_File|'''rcm1d.def''']], which contains more general information. Indeed, the 1D model generally does not use [[The_start.nc_and_startfi.nc_input_files|'''start.nc''']] or [[The_start.nc_and_startfi.nc_input_files|'''startfi.nc''']] files to initialize. You can find examples of 1D configuration in ''LMDZ.GENERIC/deftank'' (e.g. '''rcm1d.def.earlymars''', '''rcm1d.def.earth'''), the best thing is to have a look at them.
-== kcm1d test case ==
+== kcm1d program ==
 Our 1-D inverse model