GEOS-5 Documentation for Jason 3.0
This page lists pages with information on setting up and running GEOS-5 Jason 3.0. Earlier and later versions of GEOS-5 may differ in details that, while minor, will require different steps for building and running, so caveat utilitor.
Back to GEOS-5 User's and Developer's Guides
ChangeLog Entry
In src/Applications/GEOSgcm_App/ChangeLog:
==========================================================================
03May2019, Takacs, CVS Tag: Jason-3_0 (non-zero diff to Jason-2_2)
Module: GEOSagcm
------------------------------------------------------------------------------
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.
Restart Changes: None. However, 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
==========================================================================
25Apr2019, Takacs, CVS Tag: Jason-2_2 (zero diff to Jason-2_1 in Default Mode)
Module: GEOSagcm
------------------------------------------------------------------------------
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.
Restart Changes: None, except for a Long-Name change in TR restart
==========================================================================
17Apr2019, Takacs, CVS Tag: Jason-2_1 (zero diff to Jason-2_0 in Default Mode)
Module: GEOSagcm
------------------------------------------------------------------------------
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.
Restart Changes: turb_import_rst, Simply Bootstrap.
surf_import_rst, Simply Bootstrap.
irrad_internal_rst (passive, for clear-sky diagnostics only), Simply Bootstrap
==========================================================================
04Mar2019, Takacs, CVS Tag: Jason-UNSTABLE/Jason-2_0 (non-zero diff to Jason-1_1)
Module: GEOSagcm
------------------------------------------------------------------------------
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
Restart Changes: solar_internal_rst (for RRTMG). Simply Bootstrap.
==========================================================================
23Jan2019, Takacs, CVS Tag: Jason-1_1 (non-zero diff to Jason-1_0_BETA)
Module: GEOSagcm
------------------------------------------------------------------------------
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.
Restart Changes: YES (but bootstrappable)
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.
==========================================================================
17Sep2018, Takacs, CVS Tag: Jason-1_0_BETA (zero diff to Jason-gitintegration from SI Team)
Module: GEOSagcm (non-zero diff to Icarus-3_4_ESMF7)
------------------------------------------------------------------------------
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.
Restart Changes: Saltwater_internal ==> openwater_internal & seaicethermo_internal
==========================================================================
13Jul2018, Takacs, CVS Tag: Icarus-3_4_ESMF7 (zero-diff to Icarus-3_4)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Modifications needed for use of ESMF7
2) Mods for RRTMG Longwave to enable zero-diff partitioning
Restart Changes: none
==========================================================================
13Jul2018, Takacs, CVS Tag: Icarus-3_4 (zero-diff to Icarus-3_3_p3)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Updates to StratChem using tag: Icarus-3_2_p9_MEM_MERGE_3
and Icarus-3_2_p9_MEM_MERGE_3-r1
2) Updates using tags: GEOSldas_m4-17_UNSTABLE
32_GOSWIM_BUG_FIX
b_LANDUPD-change
3) Updates to regrid.pl (and supporting code) to ALLOW split ocean restarts
4) Updates for Icarus-3_3_p3 (i.e., sync with GEOSadas-5_21_2)
Restart Changes: none (for standard model)
==========================================================================
13Jul2018, Takacs, CVS Tag: Icarus-3_3_p3 (zero-diff to GEOSadas-5_21_2)
Module: GEOSagcm
------------------------------------------------------------------------------
1. This is the model tag used for GEOSadas-5_21_2
2. It is zero-diff to Icarus-3_3_p2, EXCEPT for:
a) GOCART runs on Heartbeat
b) DUST emissions have been retuned (updated DU rc files)
Restart Changes: none
==========================================================================
26Apr2018, Takacs, CVS Tag: Icarus-3_3_p2 (non-zero diff to Icarus-3_3_p1)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Changed fvcore_layout_rc to REMAP using Temperature with DZ Filtering
remap_option = 0
fv_sg_adj = 450
2. Update to plots.
Restart Changes: none
==========================================================================
05Apr2018, Takacs, CVS Tag: Icarus-3_3_p1 (non-zero diff to Icarus-3_3)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Changed MIN_RI in MOIST from 20 to 15 microns
2. Changed ANV_ICEFALL in MOIST from 0.5 to 1.0
3. Changed fvcore_layout_rc to REMAP using Total Energy with NO DZ Filtering
remap_option = 2
fv_sg_adj = -1
4. Modified MAPL_PI to use MAPL_R8 structure. This fixes AREA calculation withing DYCORE.
5. On DISCOVER, Uses: Intel-18.0.1.163, MPT-2.17
6. "FAST" compiler options used only for FV3 (Note: zero-diff to SLOW options)
Restart Changes: none
==========================================================================
27Mar2018, Takacs, CVS Tag: Icarus-3_3 (non-zero diff to Icarus-3_2_p9)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Changed MIN_RI for MOIST from 6 to 20 microns
2. Changed to Intel18 with "FAST" compiler options.
3. Changed FV Dynamics to REAL*4.
4. Modified FV and PHYSICS for better conservation of Dry Mass.
Restart Changes: moist_import_rst (if binary format, simply Bootstrap)
==========================================================================
20Feb2018, Takacs, CVS Tag: Icarus-3_2_p9 (non-zero diff to Icarus-3_2_p8)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Set optics tables Chou-Suarez for SW and and RRTMG for LW.
2. Set HEARTBEAT to a minimum of 450 seconds.
3. Updated Cloud Water and Cloud Ice effective Radii.
4. New Methodology for IAU. Default: Traditional P3C6T6_DF
4DIAU: REPLAY_MODE: Exact_4D
CORRECTOR_DURATION: 10800 (3-hours)
Restart Changes: irrad_internal_rst (for RRTMG). Simply Bootstrap.
==========================================================================
19Jan2018, Takacs, CVS Tag: Icarus-3_2_p8 (non-zero diff to Icarus-3_2_p7)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Enhancement to the diffusion directly in the sponge layer.
2. Set optics tables back to Chou-Suarez for both SW and LW.
Restart Changes: solar_internal_rst (Back to Chou-Suarez). Simply Bootstrap.
==========================================================================
08Jan2018, Takacs, CVS Tag: Icarus-3_2_p7 (non-zero diff to Icarus-3_2_p6)
Module: GEOSagcm
------------------------------------------------------------------------------
1. Modified GEOS_MoistGridComp.F90 to have RAS settings behave like x0027_ctl DAS experiments.
No resolution dependency for AGCM.rc file.
Restart Changes: none
==========================================================================
15Dec2017, Takacs, CVS Tag: Icarus-3_2_p6 (non-zero diff to Icarus-3_2_p5)
Module: GEOSagcm
------------------------------------------------------------------------------
1. A bug fix in RRTMG shortwave found by Max that will fix the SW Albedo issues Gary has seen in x0029 (Land & Radiation Consistency)
2. An update in FV3 that allows us to use the non-monotonic advection schemes for momentum and thermodynamic
variables without any degradation in edge effects, plus the truly linear PPM schemes in the sponge layers.
Restart Changes: none
==========================================================================
06Dec2017, Takacs, CVS Tag: Icarus-3_2_p5 (non-zero diff to Icarus-3_2)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Updates to FV3 dynamics. Resolution dependency is automatic.
2) CBL_TPERT = 0 by default. Brings back PBL heights.
3) Revert REPLAY parameters back to version from GEOSadas-5_18_0 (x0029 configuration)
Restart Changes: none
==========================================================================
__Nov2017, Takacs, CVS Tag: Icarus-3_2 (zero-diff to Icarus-3_1 when NOT running Brown Carbon)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Merge with tag: asd+Icarus-3_0_UNSTABLE-a2 to allow option for Brown Carbon
2) Enable IAU_DIGITAL_FILTER as default for DAS and REPLAY modes. (non-zero diff)
3) Changed BKGAVE default to 0 (from 2/3 * CORRECTOR_DURATION). (non-zero diff)
A better approach is to use: IAU_DIGITAL_FILTER: YES
Restart Changes: none
==========================================================================
__Nov2017, Takacs, CVS Tag: Icarus-3_1 (zero-diff to Icarus-3_0 when NOT running NewLand)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Implement Code to enable option for running NewLand
==========================================================================
04Oct2017, Takacs, CVS Tag: Icarus-3_0 (non-zero diff to Icarus-2_0_p1)
Module: GEOSagcm
------------------------------------------------------------------------------
1) Implement R8 version of FV3 VLAB p3 dynamics. Note, current R4 version loses Dry-Mass. R8 version conserves.
2) Updated Single-Moment Physics (MOIST Processes)
3) AGCM now runs RRTMG Shortwave Radiation and Chou-Suarez Longwave Radiation by default (NOTE: cannot use old solar_internal restarts)
4) Implement cubic time-interpolation option for MKIAU in REPLAY (Default: Linear)
5) Updates to MAPL_Base and MAPL_cfio
6) Updates to QuickPlots
7) Implement bugfix to cloudnew.F90 as in Icarus-1_0_p2
8) Changed Dry-Mass value in rs_scale.F90 from 983.24 mb (MERRA-2 value) to 983.05 mb due to use of GMTED topography.
Mean TOPO difference: 1.58436898227561 meters
Note: 983.05 mb is very similar to value based on ERA-Interim data.
9) Modified GEOS_TurbulenceGridComp.F90 with more REAL*8 code to improve Dry-Mass Conservation
10) Due to updated FVCORE, GWD BGSTRESSMAX Default changed from 1.2 to 0.9 for 72-levels (Note: 132-levels probably requires re-tuning).
NOTE: Regular REPLAY to itself FAILS under certain Predictor/Corrector configurations.
(Reason UNKNOWN) P3_C6 and P0_C1 seem to work properly.
Restart Changes: YES
RRTMG requires new solar_internal restart.
Simply delete old restart, and bootstrap new RRTMG restart.
Running Jason 3.0
- Jason 3.0 Quick Start: This page describes the minimum steps required to build and run GEOS-5 Jason 3.0 on NCCS discover and NAS pleiades. You should successfully complete the steps in these instructions before doing anything more complicated.
Jason 3.0 Documentation
- GEOS 5 Developer's Guide (pdf)
- List of Export/Import/Internal states, generated with
src/GMAO_Shared/MAPL_Base/mapl_vlist.py
- 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>