Documentation for GEOS GCM v10

This page lists pages with information on setting up and running GEOS GCM version 10. Earlier and later versions of GEOS may differ in details that, while minor, will require different steps for building and running, so caveat utilitor.

Back to GEOS User's and Developer's Guides

---

Running GEOS GCM

• GEOS GCM Quick Start: This page describes the minimum steps required to build and run GEOS GCM v10 on NCCS discover and NAS pleiades. You should successfully complete the steps in these instructions before doing anything more complicated.

• Regridding Restarts for GEOS GCM

• Using the ExtData component

• Running a Suite of Forecasts

• Setting up to create an experiment page on GMAO Intranet

• Using the SatSim Gridded Component in Jason

• GEOS GCM Single Column Model

GEOS GCM Documentation

• GEOS 5 Developer's Guide (pdf)
• List of Export/Import/Internal states, generated with src/GMAO_Shared/MAPL_Base/mapl_vlist.py
  • Sorted by short name
  • Sorted by component name

• GEOS-5 Run-Time Characteristics

• Hierarchical structure: Below is a (partially folded) view of the Jason 3.0 hierarchy where each node is a GridComp. Some of the nodes are folded (marked with a circle at the end) for clarity - clicking the composite node reveals the sub-nodes.

<mm>flash|500px|parameters startCollapsedToLevel=4</mm>

Changelog

[10.12.4] - 2020-05-15

Restart Changes: NO

Major changes are:

1. Brings RRTMGP code up to date with recent Pincus Master changes as of 3 May 2020
2. Add ability to run setup scripts in another's install directory
3. Support for new surface parameter resource file for land
4. Fixes issues with the scm_setup
5. Update CircleCI to use Baselibs 6.0.12
6. Use different ALLREDUCE algorithm for Intel MPI
7. Change Intel MPI GATHERV algorithm so GEOSgcm does not lock up on skylakes at checkpoint
8. Add a commented out setenv for the FOR_IGNORE_EXCEPTIONS in MPT
9. Add option to symlink GEOSgcm.x in experiment directory
10. FSEGMENT bug fix in gcm_forecast.setup

[10.12.3] - 2020-04-21

Restart Changes: NO

Major changes are:

• Update to MAPL 2.1 which includes both fixes from MAPL v2.1.0 and MAPL v2.1.1
  • With this is included updating to latest releases of ESMA_env and ESMA_cmake needed by MAPL 2.1
  • Add monthly mean features
  • Add optional equation of time correction to solar position used to calculate solar zenith angle.

[10.12.2] - 2020-04-16

Restart Changes: NO except moist_import has an additional field: FRACI

Major changes are:

1. Added connectivity from MOIST to CHEM for 2 fields (RI and RL).
2. Updates to the obio component which was in the CVS Icarus-2.0 tag.
3. Bugfix for coupled mode renaming TSKINI to TSKINICE.
4. Bugfix for DTDT_moist which was initially set to temperature but never updated to provide tendency.
5. Compute the lightning flash rate (LFR) online following the approach used by GEOSCHEMchem_GridComp. The LFR is calculated in MOIST and then made available to other components as a new export field (LFR_GCC).
6. Modified the mk_Restarts script to run on SLES-12.
7. Update RRTMGP to Pincus 2020-Feb-12.
8. New Surface parameter resource file (GEOSsurface_GridComp/Shared/GEOS_SurfaceGridComp.rc) to avoid having to add new parameters into AGCM.rc.tmpl and prepare for further science development.
9. Correcting the following issues with the single column model: (1) GF imports from DYN were not filled; (2) scm_setup incompletely adjusted for L132; (3) SST file for cfmip_s11 case was incorrect.
10. Allows regrid.pl to run on SLES-12
11. Enabled compilation of convert_aerosols.x as R8

[10.12.1] - 2020-03-27

Restart Changes: NO except moist_internal has an additional 3 fields: NGRAUPEL, NRAIN, NSNOW

Major changes are:

1. Broader support for various computing environments including SLES12. Adds a new question to gcm_setup (and other setup scripts) that asks a user what type of processor to run on. It then uses the "default" number of cores per those processors to better figure out how to handle the IOserver nodes.
2. Changes that enable MAPL2 GEOS to work with MPT at NCCS.
3. Changed REPLAY_Shutoff alarm from NON-STICKY to STICKY, and used it in GAAS to stop updates in forecasts during REPLAY mode.
4. Updates to moist and GOCART using 2MOMENT microphysics and MG3. These are 0-diff when running the default 1MOMENT microphysics.
5. Modified gcmpost.script to only operate on pressure-level collections.
6. Fixes issues where systems do not have either ImageMagick or F2Py. If F2Py is not found, then F2Py targets are not built.
7. Changes to GMAO_hermes:
   • Split off the independent sections of module m_topo_remap used by FV core into shared_topo_remap
   • write_eta.F90 added to prepare for a config file which eventually will replace m_set_eta module
   • Add option to build HERMES_LIGHT

[10.12.0] - 2020-03-09

Restart Changes: NO

Major changes are:

1. Fixed "zero-diff" issue related to GF evaporation vs total rain, fixed GF cloud-mass-flux scaling issue.
2. Move ShallowCu contribution to mass flux and add scale-dependent stochastics to convection.

[10.11.0] - 2020-03-04

Restart Changes: NO

Major changes are:

1. RRTMG_SW: Updates RRTMG_SW in GEOS to be equivalent to RRTMG_SW v4.10. This fix increases the heating rates near the stratospause region and is a non zero-diff change.
2. RAS_NO_NEG: Provides an optional flag that ensures RAS results are non-negative. This is an important fix for chemical species that was extremely useful in the Icarus but did not make it into Jason. The fix is "off" by default and therefore zero-diff.
3. Plots:

• Enable additional upper levels for forecast stats plots.
• Add QITOT & QLTOT to horizontal plots.
• Add aerosols to time series plots.

Other updates:

• Remove #PBS pragmas at NCCS to use #SBATCH.
• Brought the chem setup scripts in line with gcm_setup.
• Bug fixes for gcm_run and gcm_forecast scripts.

[10.10.0] - 2020-02-20

Restart Changes: NO

This release is equivalent to the NOAA 2019.01.01 Release but with the addition of GEOS constants controlled by -DMAPL_MODE. Additional changes allow building with CMake.

[10.9.0] - 2020-02-13

Restart Changes: NO

This release is equivalent in physics to v10.8.0 but now uses MAPL 2.0. The release is non-zero-diff due to the use of ESMF regridding for History and ExtData.

Also, this version of GEOSgcm now uses the new PFIO IO server to handle History and ExtData file access and writing. By default, these IO servers run on the same resources (i.e., nodes) as the application. However, gcm_setup and gcm_run.j have been altered so that, by default, at high-resolution (C180+) the IO servers are on separate nodes for performance and memory reasons.

[10.8.0] - 2020-02-06

Restart Changes: NO (only if AOIL_COMP_SWITCH=ON)

Recoded atmosphere-ocean interface layer (AOIL). Provides an option to exercise old/new versions of the AOIL.

1. Old version (AOIL_COMP_SWITCH = ON): Default. Using the old interface is still non-zero diff to x0039, however, shown to be statistically insignificant in terms of forecast skill scores. Restarts with old interface can be used as is.
2. New version (AOIL_COMP_SWITCH = OFF): If the new interface is turned on, then the user must provide additional fields in the internal state. Contact Santha Akella for a utility to do that.

[10.7.0] - 2020-02-04

Restart Changes: NO

1. Correction to RRTMG Shortwave 'layreffr' variable to reduce stratospheric cold bias.
2. Stats plots updates to produce stratospheric forecast statistics up to 1mb.
3. Correction to gcm_run.j & gcm_forecast.tmpl related to REPLAY with GAAS aod files.

[10.6.0] - 2020-01-31

Restart Changes: NO

1. Atmospheric constituents (aersols, co, co2, etc) are now transported through moist/GF. This is equivalent to the way transport is handled in f525land_fpp/f525_fp.
2. Added exports for nitrate tendency and other fluxes related to GF convection.
3. Added nitrate tendency to history templates.
4. Upgraded to GEOSchem_GridComp to v1.1.0.
5. Set the default k_split = 2.
6. Fixed gcm_convert.j to work consistent with how gcm_setup and gcm_run.j interact.

[10.5.0] - 2020-01-24

Restart Changes: NO

1. Hotfixed a bug to enable regridding of catchement restarts from MERRA2

2. REPLAY updates:

   The REPLAY algorithm within the GEOSagcm has been modified to enable a 4DIAU methodology consistent with that used in DAS mode. Prior to this release, REPLAY could only incorporate one analysis increment within the CORRECTOR_DURATION window (nominally 6-hours). The 4DIAU update allows for multiple analysis increments to be used within the CORRECTOR window. The number of increments that wiil be used is determined by the settings of the REPLAY_FILE_FREQUENCY (in nnnnnn seconds format) and the REPLAY_FILE_REFERENCE_TIME (in HHMMSS format). Based on these two settings, the code will determine how many analysis increments will properly fit within the CORRECTOR window. Note: the previous setting of PREDICTOR_DURATION is no longer needed since the code computes the required length automatically.

   The updated algorthm allows for 2 types of REPLAY:

   1. Exact (Uses Archived Increment Files)

      Exact REPLAYs identically reproduce (bit-for-bit) a previous GEOS DAS experiment by simply using the archived IAU forcing from the DAS experiment. For these cases, only the CORRECTOR step is employed and is, therefore, as efficient as a stand-alone model run.

   2. Regular (Uses Archived Analysis/Assimilation Files)

      Regular REPLAYs are used to mimic the original IAU methodology, i.e., both the PREDICTOR and the CORRECTOR steps are employed. The key difference being that the Analysis is based on archived datasets. These Analyses may be from the GEOS DAS system, or from any other source that is written using the appropriate GEOS DAS ana.eta data format.

   For more detailed information, please see the comments within the AGCM.rc.tmpl file.

[10.4.0] - 2020-01-13

Restart Changes: YES (land regridding)

• Non-zero-diff updates for:
  1. Regridding land restarts
  2. Running Chou-Suarez longwave radiation.

1. Update FV3 to latest version v1.0.9 which fixed UH25 diagnostic to build in both R4 and R8.
2. Land Updates:
   1. Clarified LONG_NAME of "snow depth" variable.
   2. Added wemin wemout options.
   3. Add PEATMAP data processing (disabled by default).
   4. GEOSsurface_GridComp changed values of pi to MAPL_PI.
   5. Added 8 more climatological (fPAR, VISDF, NIRDF stats for CATCHCN) data files and new clsm/country_and _state_code.data file.
   6. Extra column on soil_param.* files.
3. Non-zero-diif bug fix in Chou-Suarez for erroneous cloud contamination in OLR-OLRNA (aerosol forcing).
4. Fixed MAPL bug with vector regridding of winds in CTM.
5. Workaround fix building FMS on macOS.

[10.3.6] - 2020-01-02

Restart Changes: NO

1. Compute decomposition in MKIAU GridComp
2. Use esma_set_this in GCs
3. Allows GEOS-Chem (v12.4.0) chemistry to be setup and run "out of the box".
4. INTERNAL and IMPORT restarts for GEOS-Chem are now in AGCM.rc.tmpl.
5. GEOS-Chem timestep is set to be 2 x HEARTBEAT by default.
6. Enforces GOCART_DT==HEARTBEAT_DT so all species work correctly.
7. Bug fix for comments using GNU compilers with Ninja.

[10.3.5] - 2019-11-07

Restart Changes: NO

1. Enable reading netCDF increments during EXACT replay.
2. Fixed issues affecting the MOIST aerosol diagnostics:
   1. incorrect mapping of tracer names to diagnostics
   2. sign of the diagnostics by not allowing negative values.
3. New approach for increment diagnostics discussed in GEOSchem_GridComp v1.0.3.
4. OXIT and OXIM replaced with MTRI%PCHEM::OXIM and TRI%PCHEM::OXIT in most HISTORY files. OX increments due to moist and turblence are now computed using BundleIncrementMod.F90 and their names have been changed.
5. Implemented increments to land surface variables in the ADAS.
6. Fixed GUST for AMIP vs no-inc REPLAY difference in HISTORY.
7. Removed shallow.F90 which is no longer used.
8. Updated gcm_regress.j to be able to test both agcm and coupled models.
9. Remove references to gcm_stats.j in setup.

[10.3.4] - 2019-10-01

Restart Changes: NO

1. Fields out of the MTR bundle came in the “wrong” order and affected the Henry’s law constants. Hotfix to fetch field by name instead of by index.
2. Fixed triangular pdf option in SHUW.
3. Update chemistry to v1.0.2 to run GMI and Stratchem. Updated the StratChem and GMIchem setup scripts to better match gcm_setup.
4. Hotfix in fv_tracer2d.F90 to compare to ksplt instead of nsplt.
5. FVdycoreCubed_GridComp moved to separate repository.
6. FMS moved to separate repository and renamed GFDL_fms to fms
7. getco2.F90 moved to @GMAO_Shared.
8. Adapted GuestOcean_GridComp to select DATASEA, MOM5 or MOM6 at runtime.
9. Changes to enable LDAS to start using the develop branch:
   1. ensemble ID width for LDAS
   2. added co2 file for CatchCN support
   3. bug fix that only affects New Land when snow impurity model is active
10. Updates from S2S-3 CVS tag:
    1. Includes dual ocean code
    2. Update to precip correction and additional coupled diagnostics
11. MERRA2_EMISSIONS_DIRECTORY path bugfix for gcm_run.j and gcm_regress.j
12. Updates from CVS GEOSadas-5_25_2 tag.
13. Updated plots.
14. Enhancements to scm_setup script.

[10.3.3] - 2019-10-02

Restart Changes: NO

1. Move to GMAO_Shared 1.0.4 to merge more updates from CVS for plots
2. Updates for Travis CI

[10.3.2 / Jason-3_2] - 2019-08-02

Restart Changes: NO

1. Changes in FV3 related to Horizontal Advection choices. (from Bill: JU_20190722_WMP)
2. Changes in FV3 dddmp (from 0.1 to 0.2, to improve stability)
3. Changes in MOIST related to ICE-FALL. (from Bill: JU_20190722_WMP)
4. Updates to GEOS_OpenWaterGridComp.F90 and GEOS_DataSeaGridComp.F90 to modify the original Coupled/Merged System to include Santha’s reformulated ice/melt. (from Santha: sakella-GEOSadas-5_25_0_DV_in_MIZ_SSTunderIceFix)
5. Many updates related to CMAKE for upcoming GIT repository.
6. The addition of ERA5 to QUICKPLOT.
7. Modified STATS for NPO, SPO, and XPO regions. Modified Montage plots.

Restart Changes: None.

[10.3.1 / Jason-3_1] - 2019-06-13

Model Changes

1. Updates to GuestOcean_GridComp.F90 to re-sync OCEAN clock with ATMOS clock when running Regular REPLAY together with DataOcean. (non-zero diff)
2. Introduced code in MAPL_CFIO for HISTORY which corresponds to the “fix_undef” utility used in MERRA-2 post-processing. This creates consistency between horizontally transformed and vertically interpolated quantities and the target output level (i.e., values below surface are UNDEF).
3. MKIAU has been modified to allow for REPLAYing to different vertical resolution. Results are zero-diff to previous versions if vertical resolution is the same.
4. Updates to mk_restarts for Land/Catch. Zero-diff when regrid.pl points to standard (old-land) Icarus/Jason boundary area.

[10.3.0 / Jason-3_0] - 2019-05-03

Restart Changes: NO (with caveats)

Due to changes in import states and diagnostics, these restarts have changed but are passive to the prognostic restarts:

• irrad_internal_rst
• surf_import_rst
• tr_internal_rst
• turb_import_rst

Model Changes

1. Zero-Diff update to FV3 dynamics to allow FV3+ADV option
2. Updates from Mike Manyin to his tag: Jason-2_2_MEM_2
3. This tag primarily is a Merge between Jason-2_2 and Andrea Molod’s coupled tag. This produces a non-zero diff to Jason-2_2 due to a modified surface layer interface.

[Jason-2_2] - 2019-04-25

Restart Changes: NO (with caveats)

There was a long_name change in TR restart

Model Changes

1. StratChem updated to tag: Jason-2_0_MEM_2, as well as updates to GEOS-Chem and HEMCO from Mike Manyin
2. CARMA updated to tag: prc+asd-AeroChem-Icarus-3_3_p2+a1-v7 from Pete Colarco (with minor mods in DustEmissionMod.F90 because of an interface differences)
3. GIT related updates.

[Jason-2_1] - 2019-04-17

Restart Changes: YES

• turb_import_rst: Simply Bootstrap.
• surf_import_rst: Simply Bootstrap.
• irrad_internal_rst (passive, for clear-sky diagnostics only): Simply Bootstrap

Model Changes

1. Retuning when using NewLand.
2. Addition of Optional RRTMPG (off by Default).
3. Modifications for New RRTMG diagnostics.
4. Updates for regrid.pl and related code.
5. StratChem updated to tag: Jason-2_0_MEM_1
6. GIT related updates.

[Jason-2_0] - 2019-03-04

Restart Changes: YES

• solar_internal_rst (for RRTMG). Simply Bootstrap.

Model Changes

1. This is a MAJOR update to include updated moist processes based on the tag: merge_JDB7p1_retuned

• It uses RRTMG for BOTH Shortwave and Longwave radiation.
• It uses the GF Convection Scheme, as well as the UW Shallow Cumulus scheme.
• It uses dynamics tunings based on the DYAMOND experiments.

2. Option included for running Stochastic Physics.
3. 2nd-Moment Cloud MicroPhysics updated to tag: b_Jason_2_0_donifan

[Jason-1_1] - 2019-01-23

Restart Changes: YES

• GOCART IMPORT was modified by adding LWI and FRACI to be non-skip
• TR IMPORT was modified by removing ALL but CLDTT (from RAD), since TR is run after SURF RUN(1,2)
• Note: When TR adopts a 2-phase approach, additional IMPORT variables from SURF will need to be included.

Model Changes

1. An error was found related to GOCART and TR regarding the TR and GOCART IMPORT restarts. This error was introduced going from Icarus-1_0_p1 to Icarus-2_0 with the splitting of GOCART into RUN1 and RUN2 methods. Essentially, the ordering in PHYSICS is: SURF,CHEM,TURB RUN1, followd by SURF,TURB,CHEM RUN2. Emissions are computed in CHEM RUN1 based on LWI locations. However, LWI is not computed until SURF RUN2. Therefore, LWI is required on the GOCART IMPORT. But this variable was labeled as RESTART_SKIP. The TR IMPORT did contain LWI, thus when TR was Enabled, GOCART saw the proper LWI. But when TR was Disabled, LWI was no longer available for GOCART, and resulted in non-zero differences. The quick solution was to label LWI as non-skip on the GOCART IMPORT restart.
2. An error was found related to the TH_for_CHEM within the SYNC-TQ sequence. The error was fixed, but experiments showed negligible impact due to its second-order influence.
3. Modifications to the STATS plotting package to include Mean-Square-Error decompostion.

[Jason-1_0_BETA] – 2018-09-17

Restart Changes: YES

• saltwater_internal ==> openwater_internal & seaicethermo_internal

Model Changes

1. Saltwater split => openwater and seaicethermo. This means saltwater_internal is now two files: openwater_internal and seaicethermo_internal. saltwater_import remains. Regridding utility (regrid.pl) will always produce the split ocean with this tag. You can, however, provide the single saltwater_internal as input.
2. Sponge Layer fix in FV3 to include damping between top model layer and bottom of sponge.
3. Regridding in MAPL now done in a new abstract/OO paradigm. This leads to the new grid definitions in AGCM.rc
4. New HISTORY.rc grid_label definitions. This HISTORY is designated as VERSION: 1. Older HISTORY files (using resolution attribute) will still work provided VERSION = 0.
5. QUICKPLOT updates to produce automated “closeness” plots.

Previous Versions

For previous changes please see ChangeLog-PreJason.txt

---

Back to GEOS User's and Developer's Guides