There are a few freely available MPI libraries around; among the most widely spread are the OpenMPI and MPICH libraries

== Checking if you have an Fortran enabled MPI library already installed and at hand ==
This can easily be done by checking if the MPI wrapper '''mpif90''' is available, i.e. that
<syntaxhighlight lang="bash">
which mpif90
</syntaxhighlight>
returns something meaningful

== using a package manager ==
If you have sudo privileges and are on an Ubuntu or Linux Mint machine you should be able to install the OpenMPI library with
<syntaxhighlight lang="bash">
sudo apt install libopenmpi-dev openmpi-bin
</syntaxhighlight>

== Example of personal installation of the the OpenMPI library ==
To adapt to fit your onw needs and system specificities, obviously.
In the example below the installation is done in the '''$HOME/LMDZ''' directory (for convenience this path is stored in the '''PATH_OPENMPI''' variable.
<syntaxhighlight lang="bash">
PATH_OPENMPI=$HOME/LMDZ/
mkdir -p $PATH_OPENMPI
cd $PATH_OPENMPI
wget https://www.open-mpi.org/software/ompi/v1.10/downloads/openmpi-1.10.3.tar.gz
gunzip openmpi-1.10.3.tar.gz
tar xvf openmpi-1.10.3.tar
cd openmpi-1.10.3
./configure --prefix $PATH_OPENMPI/openmpi-1.10.3
make all install
</syntaxhighlight>
To check that the install went well check that
<syntaxhighlight lang="bash">
$PATH_OPENMPI/openmpi-1.10.3/bin/mpif90 -V
</syntaxhighlight>
returns something meaningful

== Worth knowing about ==
* '''mpif90''' is just a wrapper around a given compiler which could be gfortran, ifort, pgfortran, etc. to know more about that run
<syntaxhighlight lang="bash">
mpif90 --version
</syntaxhighlight>
* in some cases (e.g. with ifort) the Fortran wrapper is called '''mpiifort''' , not '''mpif90'''
* It is vital to use the '''mpirun''' command from the same package than the '''mpif90''' compilation wrapper use to compile. Otherwise one can get some really weird results...

03/01/2022

[[Category:HowTo]]
[[Category:WhatIs]]

Building an MPI library

2023-01-05T14:49:43Z

Emillour :

There are a few freely available MPI libraries around; among the most widely spread are the OpenMPI and MPICH libraries

== Checking if you have an Fortran enabled MPI library already installed and at hand ==
This can easily be done by checking if the MPI wrapper '''mpif90''' is available, i.e. that
<syntaxhighlight lang="bash">
which mpif90
</syntaxhighlight>
returns something meaningful

== using a package manager ==
If you have sudo privileges and are on an Ubuntu or Linux Mint machine you should be able to install the OpenMPI library with
<syntaxhighlight lang="bash">
sudo apt install libopenmpi-dev openmpi-bin
</syntaxhighlight>

== Example of personal installation of the the OpenMPI library ==
To adapt to fit your onw needs and system specificities, obviously.
In the example below the installation is done in the '''$HOME/LMDZ''' directory (for convenience this path is stored in the '''PATH_OPENMPI''' variable.
<syntaxhighlight lang="bash">
PATH_OPENMPI=$HOME/LMDZ/
mkdir -p $PATH_OPENMPI
cd $PATH_OPENMPI
wget https://www.open-mpi.org/software/ompi/v1.10/downloads/openmpi-1.10.3.tar.gz
gunzip openmpi-1.10.3.tar.gz
tar xvf openmpi-1.10.3.tar
cd openmpi-1.10.3
./configure --prefix $PATH_OPENMPI/openmpi-1.10.3
make all install
</syntaxhighlight>
To check that the install went well check that
<syntaxhighlight lang="bash">
$PATH_OPENMPI/openmpi-1.10.3/bin/mpif90 -V
</syntaxhighlight>
returns something meaningful

== Worth knowing about ==
...

03/01/2022

[[Category:HowTo]]
[[Category:WhatIs]]

Building an MPI library

2023-01-05T14:33:03Z

Emillour :

Building an MPI library

2023-01-05T10:42:47Z

Emillour : Page créée avec « There are a few freely available MPI libraries around; among the most widely spread are the OpenMPI and MPICH libraries == Checking if you have an Fortran enabled MPI lib... »

Some pointers about standalone installation of LMDZ

2023-01-04T16:29:07Z

Emillour :

This page is mostly for "do-it-yourself" people interested in installing LMDZ6

== Prerequisites ==
In the most minimal case, where one would only run the serial version of the GCM, then the prerequisites are:
# Have a Fortran compiler at hand
# Have a NetCDF library (compiled with the same compiler) available (see [[The netCDF library|this page]] for more about installing NetCDF)
# Have the IOIPSL library available (see [[Some pointers about standalone installation of the IOIPSL library|this page]] for more about installing IOIPSL)

In a more realistic case, you'll probably want to run the parallel version of the GCM, most likely with XIOS; The prerequisites are then:
# Have a Fortran-enabled MPI library at hand (see [[Building an MPI library|this page]] for more about installing MPI)
# A NetCDF4-HDF5 library, preferably compiled with MPI enabled
# Have the IOIPSL library available
# Have the XIOS library available (see [[Some pointers about standalone installation of the XIOS library|this page]] for more about installing XIOS)

== Downloading LMDZ6 source code ==
This is easily done using svn (subversion). To download the latest version from the trunk as an "LMDZ6" directory:
<syntaxhighlight lang=bash>
svn checkout http://svn.lmd.jussieu.fr/LMDZ/LMDZ6/trunk LMDZ6
</syntaxhighlight>
Note that the recommended way to proceed is to have all models and libraries at the same level in a master directory, so a working example would be where the master directory contents would be:
<pre>
IOIPSL LMDZ6 XIOS
</pre>

To compile LMDZ one would then proceed to the LMDZ6 subdirectory, make some adequate [[WhatIs: The target architecture ("arch") files|architecture files]] (if needed) and use the [[WhatIs: The makelmdz fcm script|makelmdz_fcm.sh]] script

04/01/2023

[[Category:HowTo]]

The netCDF library

2023-01-03T13:50:02Z

Emillour :

== the NetCDF library ==
The model reads and writes input and output files in NetCDF (Network Common Data Form) format (developed and maintained by Unidata: https://www.unidata.ucar.edu/software/netcdf/ ) and therefore a NetCDF library must be at hand when compiling and running LMDZ6.

As this library is not quite standard, chances are that you might need to install it yourself on your system (note that the [[WhatIs: The install lmdz.sh script|install_lmdz.sh]] default behavior is to download and install that library), hence this page with some indications on how to do so.

== Checking if a NetCDF library is already available ==
It is possible that the library, including its Fortran component, is already available. An easy way to check this is to see if the related utilities like '''ncdump''' are available, i.e. that
<syntaxhighlight lang="bash">
which ncdump
</syntaxhighlight>
returns a positive answer.

Note that this does not suffice as the full library, including its Fortran component, and not just the related utilities are required. A neat way to check this is to use the '''nf-config'''
<syntaxhighlight lang="bash">
nf-config --all
</syntaxhighlight>
returns something meaningful. If not, you probably need to take some action along the lines of what is indicated in the following sections.

== Various ways to install the NetCDF library ==
=== Install via a system manager ===
If on Ubuntu (and with admin, i.e. ''sudo'' rights) you can try
<syntaxhighlight lang="bash">
sudo apt install libnetcdff-dev
</syntaxhighlight>

=== Personal installation ===
One can always download the source code and compile the NetCDF library (see https://docs.unidata.ucar.edu/netcdf-c/current/faq.html#HowdoIgetthenetCDFsoftwarepackage ). You can use the following home-made "install_netcdf4_hdf5_seq.bash" script to do so. For this, ensure that you are in your home directory and:
<syntaxhighlight lang="bash">
mkdir netcdf
cd netcdf
wget -nv --no-check-certificate http://www.lmd.jussieu.fr/~lmdz/pub/import/install_netcdf4_hdf5_seq.bash
chmod u=rwx install_netcdf4_hdf5_seq.bash
./install_netcdf4_hdf5_seq.bash > netcdf.log 2>&1
</syntaxhighlight>
Compiling the library and dependencies can take a while (>>15 minutes; be patient).
Once this is done, check file netcdf.log to verify that all went well.
You may want to also add its "bin" directory to your PATH environment variable by adding in your .bashrc a line of:
<syntaxhighlight lang="bash">
export PATH=$PATH:$HOME/netcdf/bin
</syntaxhighlight>
The assumption here is that you have run the "install_netcdf4_hdf5_seq.bash" script in a "netcdf" subdirectory of your home directory. Adapt accordingly if not.

Check that the installation was successfully installed :-)
Again, a neat way to do that is to run and check the output of
<syntaxhighlight lang="bash">
nf-config --all
</syntaxhighlight>

=== Personal installation of an advanced NetCDF4-HDF5 library with MPI enabled ===
This specific version is useful especially if running with the XIOS library.
A prerequisite is to have an MPI library installed and available.
You can use the following home-made "install_netcdf4_hdf5_seq.bash" script to do so. For this, ensure that you are in your home directory and:
<syntaxhighlight lang="bash">
mkdir netcdf4hdf5
cd netcdf4hdf5
wget -nv --no-check-certificate http://www.lmd.jussieu.fr/~lmdz/pub/import/install_netcdf4_hdf5.bash
chmod u=rwx install_netcdf4_hdf5.bash
./install_netcdf4_hdf5.bash -CC mpicc -FC mpif90 -CXX mpiCC -MPI /path/to/your/MPI/install > netcdf.log 2>&1
</syntaxhighlight>
You should of course adapt the command line, especially the ''-MPI /path/to/your/MPI/install'' part to your settings.

Compiling the library and dependencies can take a while (>>15 minutes; be patient).
Once this is done, check file netcdf.log to verify that all went well.
You may want to also add its "bin" directory to your PATH environment variable by adding in your .bashrc a line of:
<syntaxhighlight lang="bash">
export PATH=$PATH:$HOME/netcdf4hdf5/bin
</syntaxhighlight>
The assumption here is that you have run the "install_netcdf4_hdf5.bash" script in a "netcdf4hdf5" subdirectory of your home directory. Adapt accordingly if not.

Check that the installation was successfully installed :-)
Again, a neat way to do that is to run and check the output of
<syntaxhighlight lang="bash">
nf-config --all
</syntaxhighlight>

03/01/2022

[[Category:HowTo]]
[[Category:WhatIs]]

The netCDF library

2023-01-03T12:43:41Z

Emillour :

The netCDF library

2023-01-03T11:40:27Z

Emillour : Page créée avec « == the NetCDF library == The model reads and writes input and output files in NetCDF (Network Common Data Form) format (developed and maintained by Unidata: https://www.un... »

WhatIs: The tracer.def input file

2023-01-03T08:23:36Z

Emillour :

== The tracer.def input file ==
This file contains information on the tracers that will be advected in the dynamics. It should be used instead of the now depreciated [[WhatIs: The traceur.def input file|traceur.def]] input file.

== Simple example of a ''tracer.def'' file ==
In this example there are 4 tracers: 3 for water (one for each of its phases, gas, liquid or solid) and one called ''Aga''
<pre>
&version=1.0
&lmdz
default type=tracer phases=g hadv=10 vadv=10 parent=air
H2O hadv=14 vadv=14
H2O phases=ls
Aga
</pre>
* The mandatory first line ''&version'' is there to handle (potential) versioning of these file and their format
* The mandatory second line ''&lmdz'' is there to specify the scope of the following lines, i.e. that all that follow is relevant for the LMDZ lon-lat dynamical core.
* The mandatory third line starting with ''default'' specifies the default attributes for all the tracers. In this example that they are of type ''tracer'' (as opposed to type ''tag'' for tagging), that there phase is ''g'' (gas), that they are advected using advection schemes hadv and vadv ''10'' (Van Leer advection scheme) and that their parent (i.e. carrier fluid) is ''air''.
* Then one should specify on successive lines the tracer names and optionally their properties such as phases (g: gas, l: liquid, s:solid). Note that one may condense information about phases using a mix of letters g/l/s rather than specifying the information for each on a separate line. In practice the lines
<pre>
H2O hadv=14 vadv=14
H2O phases=ls
</pre>
From the example above which specifies there are 3 H2O phases (gas, liquid and solid), where the gas phase is advected with the hadv and vadv ''14'' dedicated scheme could also be written more explicitly as:
<pre>
H2O phases=g hadv=14 vadv=14
H2O phases=l
H2O phases=s
</pre>

== A more advanced example of a ''tracer.def file'' ==
The following example includes water isotopes (identified by the fact that their ''parent'' is not ''air'' but the tracer ''H2O'')
with two tags (''con'' and ''oce'')
<pre>
&version=1.0
&lmdz
default hadv=10 vadv=10 phases=g parent=air type=tracer
H2O hadv=14 vadv=14
H2O phases=ls
RN,PB
H2[18]O,H[2]HO,H2[16]O phases=gls parent=H2O
con,oce phases=gls parent=H2[18]O,H[2]HO,H2[16]O type=tag
</pre>

As LMDZ runs it outputs the detailed information about all tracers and their dependencies to one another. For the example above one would get:
<pre>
readTracersFiles: RAW CONTENT OF SECTION "lmdz":
readTracersFiles: iq | hadv | vadv | name | parent | phase
readTracersFiles: ----+------+------+------------------------+------------------------+-------
readTracersFiles: 1 | 14 | 14 | H2O | air | g
readTracersFiles: 2 | 10 | 10 | H2O | air | ls
readTracersFiles: 3 | 10 | 10 | RN,PB | air | g
readTracersFiles: 4 | 10 | 10 | H2[18]O,H[2]HO,H2[16]O | H2O | gls
readTracersFiles: 5 | 10 | 10 | con,oce | H2[18]O,H[2]HO,H2[16]O | gls
readTracersFiles:
readTracersFiles: EXPANDED CONTENT OF SECTION "lmdz":
readTracersFiles: iq | hadv | vadv | name | parent | igen | phase
readTracersFiles: ----+------+------+---------------+-----------+------+-------
readTracersFiles: 1 | 14 | 14 | H2O_g | air | 0 | g
readTracersFiles: 2 | 10 | 10 | H2O_l | air | 0 | l
readTracersFiles: 3 | 10 | 10 | H2O_s | air | 0 | s
readTracersFiles: 4 | 10 | 10 | RN | air | 0 | g
readTracersFiles: 5 | 10 | 10 | PB | air | 0 | g
readTracersFiles: 6 | 10 | 10 | H2[18]O_g | H2O_g | 1 | g
readTracersFiles: 7 | 10 | 10 | H2[18]O_l | H2O_l | 1 | l
readTracersFiles: 8 | 10 | 10 | H2[18]O_s | H2O_s | 1 | s
readTracersFiles: 9 | 10 | 10 | H[2]HO_g | H2O_g | 1 | g
readTracersFiles: 10 | 10 | 10 | H[2]HO_l | H2O_l | 1 | l
readTracersFiles: 11 | 10 | 10 | H[2]HO_s | H2O_s | 1 | s
readTracersFiles: 12 | 10 | 10 | H2[16]O_g | H2O_g | 1 | g
readTracersFiles: 13 | 10 | 10 | H2[16]O_l | H2O_l | 1 | l
readTracersFiles: 14 | 10 | 10 | H2[16]O_s | H2O_s | 1 | s
readTracersFiles: 15 | 10 | 10 | H2[18]O_g_con | H2[18]O_g | 2 | g
readTracersFiles: 16 | 10 | 10 | H2[18]O_l_con | H2[18]O_l | 2 | l
readTracersFiles: 17 | 10 | 10 | H2[18]O_s_con | H2[18]O_s | 2 | s
readTracersFiles: 18 | 10 | 10 | H2[18]O_g_oce | H2[18]O_g | 2 | g
readTracersFiles: 19 | 10 | 10 | H2[18]O_l_oce | H2[18]O_l | 2 | l
readTracersFiles: 20 | 10 | 10 | H2[18]O_s_oce | H2[18]O_s | 2 | s
readTracersFiles: 21 | 10 | 10 | H[2]HO_g_con | H[2]HO_g | 2 | g
readTracersFiles: 22 | 10 | 10 | H[2]HO_l_con | H[2]HO_l | 2 | l
readTracersFiles: 23 | 10 | 10 | H[2]HO_s_con | H[2]HO_s | 2 | s
readTracersFiles: 24 | 10 | 10 | H[2]HO_g_oce | H[2]HO_g | 2 | g
readTracersFiles: 25 | 10 | 10 | H[2]HO_l_oce | H[2]HO_l | 2 | l
readTracersFiles: 26 | 10 | 10 | H[2]HO_s_oce | H[2]HO_s | 2 | s
readTracersFiles: 27 | 10 | 10 | H2[16]O_g_con | H2[16]O_g | 2 | g
readTracersFiles: 28 | 10 | 10 | H2[16]O_l_con | H2[16]O_l | 2 | l
readTracersFiles: 29 | 10 | 10 | H2[16]O_s_con | H2[16]O_s | 2 | s
readTracersFiles: 30 | 10 | 10 | H2[16]O_g_oce | H2[16]O_g | 2 | g
readTracersFiles: 31 | 10 | 10 | H2[16]O_l_oce | H2[16]O_l | 2 | l
readTracersFiles: 32 | 10 | 10 | H2[16]O_s_oce | H2[16]O_s | 2 | s
</pre>

03/01/2023

[[Category:WhatIs]]

WhatIs: The tracer.def input file

2023-01-03T08:17:55Z

Emillour :

== The tracer.def input file ==
This file contains information on the tracers that will be advected in the dynamics. It should be used instead of the now depreciated [[WhatIs: The traceur.def input file|traceur.def]] input file.

== Simple example of a ''tracer.def'' file ==
<pre>
&version=1.0
&lmdz
default type=tracer phases=g hadv=10 vadv=10 parent=air
H2O hadv=14 vadv=14
H2O phases=ls
Aga
</pre>
* The mandatory first line ''&version'' is there to handle (potential) versioning of these file and their format
* The mandatory second line ''&lmdz'' is there to specify the scope of the following lines, i.e. that all that follow is relevant for the LMDZ lon-lat dynamical core.
* The mandatory third line starting with ''default'' specifies the default attributes for all the tracers. In this example that they are of type ''tracer'' (as opposed to type ''tag'' for tagging), that there phase is ''g'' (gas), that they are advected using advection schemes hadv and vadv ''10'' (Van Leer advection scheme) and that their parent (i.e. carrier fluid) is ''air''.
* Then one should specify on successive lines the tracer names and optionally their properties such as phases (g: gas, l: liquid, s:solid). Note that one may condense information about phases using a mix of letters g/l/s rather than specifying the information for each on a separate line. In practice the lines
<pre>
H2O hadv=14 vadv=14
H2O phases=ls
</pre>
From the example above which specifies there are 3 H2O phases (gas, liquid and solid), where the gas phase is advected with the hadv and vadv ''14'' dedicated scheme could also be written more explicitly as:
<pre>
H2O phases=g hadv=14 vadv=14
H2O phases=l
H2O phases=s
</pre>

== A more advanced example of a ''tracer.def file'' ==
The following example includes water isotopes (identified by the fact that their ''parent'' is not ''air'' but the tracer ''H2O'')
with two tags (''con'' and ''oce'')
<pre>
&version=1.0
&lmdz
default hadv=10 vadv=10 phases=g parent=air type=tracer
H2O hadv=14 vadv=14
H2O phases=ls
RN,PB
H2[18]O,H[2]HO,H2[16]O phases=gls parent=H2O
con,oce phases=gls parent=H2[18]O,H[2]HO,H2[16]O type=tag
</pre>

As LMDZ runs it outputs the detailed information about all tracers and their dependencies to one another. For the example above one would get:
<pre>
readTracersFiles: RAW CONTENT OF SECTION "lmdz":
readTracersFiles: iq | hadv | vadv | name | parent | phase
readTracersFiles: ----+------+------+------------------------+------------------------+-------
readTracersFiles: 1 | 14 | 14 | H2O | air | g
readTracersFiles: 2 | 10 | 10 | H2O | air | ls
readTracersFiles: 3 | 10 | 10 | RN,PB | air | g
readTracersFiles: 4 | 10 | 10 | H2[18]O,H[2]HO,H2[16]O | H2O | gls
readTracersFiles: 5 | 10 | 10 | con,oce | H2[18]O,H[2]HO,H2[16]O | gls
readTracersFiles:
readTracersFiles: EXPANDED CONTENT OF SECTION "lmdz":
readTracersFiles: iq | hadv | vadv | name | parent | igen | phase
readTracersFiles: ----+------+------+---------------+-----------+------+-------
readTracersFiles: 1 | 14 | 14 | H2O_g | air | 0 | g
readTracersFiles: 2 | 10 | 10 | H2O_l | air | 0 | l
readTracersFiles: 3 | 10 | 10 | H2O_s | air | 0 | s
readTracersFiles: 4 | 10 | 10 | RN | air | 0 | g
readTracersFiles: 5 | 10 | 10 | PB | air | 0 | g
readTracersFiles: 6 | 10 | 10 | H2[18]O_g | H2O_g | 1 | g
readTracersFiles: 7 | 10 | 10 | H2[18]O_l | H2O_l | 1 | l
readTracersFiles: 8 | 10 | 10 | H2[18]O_s | H2O_s | 1 | s
readTracersFiles: 9 | 10 | 10 | H[2]HO_g | H2O_g | 1 | g
readTracersFiles: 10 | 10 | 10 | H[2]HO_l | H2O_l | 1 | l
readTracersFiles: 11 | 10 | 10 | H[2]HO_s | H2O_s | 1 | s
readTracersFiles: 12 | 10 | 10 | H2[16]O_g | H2O_g | 1 | g
readTracersFiles: 13 | 10 | 10 | H2[16]O_l | H2O_l | 1 | l
readTracersFiles: 14 | 10 | 10 | H2[16]O_s | H2O_s | 1 | s
readTracersFiles: 15 | 10 | 10 | H2[18]O_g_con | H2[18]O_g | 2 | g
readTracersFiles: 16 | 10 | 10 | H2[18]O_l_con | H2[18]O_l | 2 | l
readTracersFiles: 17 | 10 | 10 | H2[18]O_s_con | H2[18]O_s | 2 | s
readTracersFiles: 18 | 10 | 10 | H2[18]O_g_oce | H2[18]O_g | 2 | g
readTracersFiles: 19 | 10 | 10 | H2[18]O_l_oce | H2[18]O_l | 2 | l
readTracersFiles: 20 | 10 | 10 | H2[18]O_s_oce | H2[18]O_s | 2 | s
readTracersFiles: 21 | 10 | 10 | H[2]HO_g_con | H[2]HO_g | 2 | g
readTracersFiles: 22 | 10 | 10 | H[2]HO_l_con | H[2]HO_l | 2 | l
readTracersFiles: 23 | 10 | 10 | H[2]HO_s_con | H[2]HO_s | 2 | s
readTracersFiles: 24 | 10 | 10 | H[2]HO_g_oce | H[2]HO_g | 2 | g
readTracersFiles: 25 | 10 | 10 | H[2]HO_l_oce | H[2]HO_l | 2 | l
readTracersFiles: 26 | 10 | 10 | H[2]HO_s_oce | H[2]HO_s | 2 | s
readTracersFiles: 27 | 10 | 10 | H2[16]O_g_con | H2[16]O_g | 2 | g
readTracersFiles: 28 | 10 | 10 | H2[16]O_l_con | H2[16]O_l | 2 | l
readTracersFiles: 29 | 10 | 10 | H2[16]O_s_con | H2[16]O_s | 2 | s
readTracersFiles: 30 | 10 | 10 | H2[16]O_g_oce | H2[16]O_g | 2 | g
readTracersFiles: 31 | 10 | 10 | H2[16]O_l_oce | H2[16]O_l | 2 | l
readTracersFiles: 32 | 10 | 10 | H2[16]O_s_oce | H2[16]O_s | 2 | s
</pre>

03/01/2023

[[Category:WhatIs]]

Nudging in LMDZ

2022-05-16T09:58:08Z

Emillour :

The sponge layer

2022-05-09T10:10:47Z

Emillour :

== in a nutshell ==
The sponge layer is necessary to damp vertically propagating waves, which would otherwise non-physically be reflected downward from the model top.
In practice its application is thus limited to a few of the topmost layers (usually 4), with decreasing impact with decreasing altitudes.
The sponge layer is a relaxation of main variables (winds and/or potential temperature) towards a given value or their zonal mean along a given time scale.

== Sponge layer parameters ==
The parameters controlling the sponge layer are in input file gcm.def (or sometimes in vert_L**.def as they are strongly linked to the vertical grid)
* iflag_top_bound (integer): 0 for no sponge, 1 for a sponge over the last 4 topmost layers, 2 to have a sponge layer extending over topmost layers down to a pressure which is 100 times that of the topmost layer's.
* mode_top_bound (integer): 0 for no relaxation, 1 to relax winds (u,v) to zero, 2 to relax winds (u,v) towards their zonal mean, 3 to relax winds (u,v) and potential temperature towards their zonal mean.
* tau_top_bound (real, Hz): inverse of the relaxation characteristic time scale at the topmost layer; which is then halved at each successive descending layer.

== Application of the sponge ==
The sponge is applied right after dissipation at each dissipation step, but only if the flag '''ok_strato=y'''

== Sponge routine ==
The sponge computations are done in routine '''dyn3d/top_bound.F''' (for the serial dynamical core; equivalently in '''dyn3dmem/top_bound_loc.F''' for the parallel dynamical core)

09/05/2022

[[Category:inputs]]

WhatIs: The ok strato parameter

2022-05-09T10:02:26Z

Emillour :

The '''ok_strato''' parameter is a logical flag, i.e. to be set to "y" (or equivalently ".true.") or "n" (or equivalently ".false.") in the run.def file, in practice most often in the "vert.def" (or sometimes "gcm.def") file, as it is strongly linked to the model vertical resolution

Setting '''ok_strato=y''' is recommended in most cases as it implies:
* Some impact on the vertical discretization (i.e. where the model levels are placed)
* Some switches on physical parametrizations related to the impact of gravity waves (orographic and non-orographic)

Running with '''ok_strato=n''' would only be useful and advised if running with simple setups such as with 19 vertical layers (with old ~CMIP3-like physics packages) or for aquaplanets or without physics (in Held&Suarez configuration).

Note also that if '''ok_strato=n''' then the sponge layer is not used either (see the leapfrog routine).

09/05/2022

[[Category:WhatIs]]
[[Category:Inputs]]