Skip to content

Latest commit

 

History

465 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Script to run cycling DA using GDASApp in cube sphere space, and offline Noah-MP model in vector space.

Clara Draper, Nov, 2021.

History Apr, 2022. Draper: Moved to PSL repo, restructuring and renaming of repos.

May, 2023. Draper: Updated intel modules, added MPI for model.

Mar, 2024. Draper: Updated to Rocky8, switched to using GDASApp for JEDI executables.

February, 2026. TZ Gichamo: add jcb-gdas, wxflow, and pygfs to generate the config/yaml files and handle workflow executions.

April, 2026: TZ Gichamo: first pass 2DVar and LETKF soil DA for T2m and SMAP smc.

#############################

COMPILING and TESTING.

  1. Fetch sub-modules.

git submodule update --init --recursive

  1. Compile sub-modules.

2a. load relevant modules for Ursa or GAEA C6

source land_mods_ursa or source land_mods_gaeac6 (these are the modules needed for steps 2b, 2c).

2b.

cd ufs-land-driver

configure select ursa parallel or gaeac6 parallel

make

cd ..

2c.

cd vector2tile

configure select ursa parallel or gaeac6 parallel

make

cd ..

2d.

cd DA_update (We no longer build GDASApp. Instead it is installed at directory shown in make_links)

./make_links.sh

source env_GDASApp

cd ..

  • Note: DA_update does not include the GDASApp submodule. Its linked above, see README in DA_update

2e. If you run NoahMP in ensemble mode (open loop or for LETKF DA)

cd land_ensemble_gen

source ../stochy_mods_ursa OR source ../stochy_mods_gaeac6

./compile_ens_intel

  1. Run the test.

In settings_cycle_test check BASEDIR, WORKDIR and OUTDIR are OK (comment/uncomment for URSA-vs-GAEAC6)

Change the other settings (Make sure the obs dir, forcing dir, etc do exit--most of these are in Land DATA dir).

In submit_cycle.sh make sure #SBATCH --account= points to your own account (& (comment/uncomment for URSA-vs-GAEAC6). Also the number of processes, threads and time can be changed and synced with those set in settings. Note that for bigger ensemble sizes (>4), you need to use large number of procs for the experiments to finish. Ideally the number of procs is set to be a multiple of 6 times the ensemble size (e.g., 120 for enssize=20, or bigger for high resolution experiments).

do_submit_test.sh

A suite of regression tests can be found in the reg_tests directory

RUNNING YOUR OWN EXPERIMENTS

  1. Prepare a settings file, using settings_template. Must fill in all variables, unless otherwise commented.

  2. If running more than 6 tasks with JEDI, need to set layout = [x,y] in settings_cycle file. Make sure to modify submit_cycle.sh to ensure nnodestasks_per_node=NPROC_JEDI=6x*y

  3. Make sure there is a restart in your ICSDIR.

restart filename example:ufs_land_restart.2015-09-02_18-00-00.nc ICSDIR points to the experiment directory with the restart. If creating a new dircetory, the structure is: $ICSDIR/output/mem000/restarts/vector/ufs_land_restart.2015-09-02_18-00-00.nc

  1. in submit_cycle.sh make sure #SBATCH --account= points to your own account.

  2. Submit your job

do_submit_cycle.sh your-settings-filename

About

No description, website, or topics provided.

Resources

Stars

2 stars

Watchers

2 watching

Forks

Releases

Packages

Contributors

Languages