v3.1.1 on 2026-01-07: Uptick GMI and CHEM

GMI includes minor fixes to a few mechanisms (zero diff for Standard, Original and Pyro).
GMI now uses FeedBack_QV = FALSE (very small non-zero-diff)

v3.1.0 on 2025-11-12: new GMI mechanisms

Uptick to GMI v1.5.0, which includes Lyman-alpha, and a few new chemical mechanisms. (Non-zero diff)
NOTE: After setup, be sure to change photolysis from 7.4 to 6.5, since CTM lacks RI and RL.

Modified the reading of MERRA2 wind vectors (10m and 2m), regridding U and V at the same time. (This is non-zero diff).

CTM based on KROK

The repositories for GOCART, GMI, FV3, etc have been updated to match those in GCM v11.7.0 .
This advances CTM from Jason to KROK.
Aerosols are now taken from GOCART2G in Data-Driven mode.
ExtData2G is required, with yaml file specifications.
All testing has been conducted with MERRA2 met-fields.

Before compiling, be sure to review the README under the Applications/GEOSctm_App directory.

This release is non-zero-diff with the previous version.
It has been a few years since the last release, so a lot has changed.

What's Changed

• Update ctm_setup for MPT and MAPL 2.0 by @mathomp4 in #26
• Add mepo and CircleCI by @mathomp4 in #23
• Made change to use ESMF_ConfigGetAttribute instead of rcEsmfReadLogical by @WilliamJamieson in #29
• Add CONTRIBUTING and CODE_OF_CONDUCT by @mathomp4 in #31
• Updates for CTM to Bring in-line with GCM by @mathomp4 in #30
• Updates for MAPL 2.7.0 by @mathomp4 in #32
• Sync with GCM 10.19.2 by @mathomp4 in #33
• Update LICENSE.md by @mathomp4 in #35
• Update LICENSE.md Again by @mathomp4 in #36
• Sync with GCM 10.19.3 by @mathomp4 in #34
• Update LICENSE to Apache, other license changes by @mathomp4 in #37
• Fix license reference in CONTRIBUTING.md by @mathomp4 in #38
• Update CTM to be in line with GEOSgcm 10.19.4 by @mathomp4 in #39
• Compress CircleCI Artifacts by @mathomp4 in #43
• V2.2.0 jk by @JulesKouatchou in #44
• Move to use GitHub Action for Label Enforcement by @mathomp4 in #45
• Extend the transition date for the last batch of MERRA2 met-fields by @mmanyin in #46
• Updates to the setup script and the run script by @mmanyin in #48
• Updates needed to run on SLES15 at NCCS by @mathomp4 in #49
• Check EGRESS to avoid infinite loop by @mmanyin in #51
• Uptick FVdycoreCubed to get AdvCore fix by @mmanyin in #52
• Fixed clone bug in ctm_setup by @mmanyin in #53
• Update to match GEOSgcm main as of 2025-Mar-11 by @mathomp4 in #54
• Enable LOPEZ lightning scheme by @mmanyin in #55
• Always compute LFC etc when running CTM by @mmanyin in #56
• Final updates prior to the v11 release by @mmanyin in #57

New Contributors

• @WilliamJamieson made their first contribution in #29
• @JulesKouatchou made their first contribution in #44

Full Changelog: v1.1.0...v3.0.0

Fixed setup when cloning

The post-processing for a cloned CTM experiment was broken; it is now fixed.

Fixed advection

Two fixes:
AdvCore had occasionally returned small negative values, due to an intialization issue; this is now fixed. (Non zero diff)
Also, the run script now properly checks success/failure of the simulation, so failures no longer result in a recursive SLURM loop.

v2.2.3 on 2024-08-26: Runs on SLES15

This release supports SLES15 (Milan nodes) as well as SLES12 (CascadeLake nodes) at NCCS. The SLES12 version is zero-diff with the prior release. The SLES15 version is non-zero-diff, but looks to be within round-off.

jan2023 minor update

This release is zero-diff from the previous release (v2.2.1).
Debugged the run script issue that prevented using MERRA2 met fields for 2010, or for dates after 2020.
The setup script is better aligned with the GCM setup script, which should make updates in GCM easier to replicate in CTM.
References to REPLAY have been removed, since CTM does not run in REPLAY mode.
GitHub Actions are now used for Label Enforcement.

mar2022 - refactored convection & enhancements

Zero diff to previous release, if PLE is read from file; roundoff differences if PLE is computed.
New CTM options:

• CTM can generate courant numbers and mass fluxes needed for advection, or read them in.
• PLE is now computed from AK and BK; option to read in as before.
• CTM Hist component now handles output of forcing data (optional).

dec2021 - keep pace with GCM

Non-zero-diff but within roundoff of previous release, due to skylake nodes now being the default (NCCS).
Zero diff with the previous release when run on the same nodes, and without I_MPI_CBWR defined in the run script.

This release changes the default choice of nodes from haswell to skylake (NCCS), and introduces the I_MPI_CBWR env var which provides consistent MPI behavior regardless of "cpu per node" count.

All repositories are now in line with GEOSgcm v10.19.4. All chemistry modules are handled properly in the setup script, except GEOS-Chem which needs a bit of customization in the experiment (both here and in GCM as well).

Complete set of notes, for the GCM components: https://github.com/GEOS-ESM/GEOSgcm/releases/tag/v10.19.4

Update 2021-08-30

Non-zero-diff, generally within round-off of the previous release.
(The new version of MAPL includes a bug-fix that is non-zero-diff for OPS GOCART.)

All repositories are now in line with GEOSgcm v10.19.3, with the exception of MAPL which required a bug fix. All chemistry modules are handled properly in the setup script, except GEOS-Chem which needs a bit of customization in the experiment (both here and in GCM as well).

Complete set of notes, for the GCM components: https://github.com/GEOS-ESM/GEOSgcm/releases/tag/v10.19.3