Bug Fix for GCC 14

This release fixes a bug with GCC 14 (see #104). All testing shows it is zero-diff.

---

What's Changed

• WIP: Fix for GCC 14. Always allocate NH arrays by @mathomp4 in #104

Full Changelog: geos/v2.9.0...geos/v2.9.1

Updates for GEOS-JEDI Work

This release of fvdycore has updates needed for GEOS-JEDI work. Namely, it enables support for optional input.nml handling.

This version of fvdycore should be used with GEOS-ESM/FMS geos/2019.01.02+noaff.9 as that has the concomitant changes to FMS.

All testing with this and the associated FMS release show zero-diff for AMIP and Coupled MOM6 runs.

---

What's Changed

• Update fvdycore for FMS updates (no hardcoded input nml path) by @climbfuji in #100

New Contributors

• @climbfuji made their first contribution in #100

Full Changelog: geos/v2.8.2...geos/v2.9.0

Fix uninitialized integers controlling FMS diagnostic output

This patch release has fixes for use with GCC 13, though found in AIST runs by @atrayano as well; details below. All testing shows this is zero-diff to geos/v2.8.1 as it should be as it only deals with diagnostics through FMS that GEOS does not use in this code.

The issue GCC 13 found was that because variables that control FMS diagnostic use in FV3 were not being initialized, code like this:

           if( idiag%id_te>0 ) then
               used = send_data(idiag%id_te, teq, fv_time)
           endif

was being triggered. It seems Intel must set uninitialized integers to zero, but GCC 13 does not. So, when memory space is shifted enough, things like idiag%id_te could be randomly greater than zero. This then triggers the if-block and send_data will only work if diag_manager_init is called. This is not done in FV3 and you end up with:

FATAL from PE    31: diag_manager_mod::send_data_3d: diag_manager NOT initialized

---

What's Changed

• Fix uninitialized variables for GCC 13 by @mathomp4 in #96

Full Changelog: geos/v2.8.1...geos/v2.8.2

Fix for Schmidt transform with locally generated coordinates

This patch release of v2.8 has a bugfix adding a missing protection to run Schmidt transform only when needed with locally generated coordinates.

This does not affect runs using the default global coordinate construction.

---

What's Changed

• Added a missing protection to run Schmidt transform only when needed by @atrayano in #93

New Contributors

• @atrayano made their first contribution in #93

Full Changelog: geos/v2.8.0...geos/v2.8.1

Fix for layout non-reproducibilty

This release of GFDL_atmos_cubed_sphere has a fix for non-layout-reproducibility at C180 and higher that was introduced in v2.

All tests show it is zero-diff in the cases that didn't hit the issue (C90 and lower).

---

What's Changed

• GitFlow: Merge hotfix from main into develop by @mathomp4 in #89
• Reverted fv_tracer2d.F90 to version geos/1.5.0 to get around the laoyut regression caused by dyncore by @aoloso in #87
• GitFlow: Merge develop into main for release by @mathomp4 in #92

New Contributors

• @aoloso made their first contribution in #87

Full Changelog: geos/v2.7.1...geos/v2.8.0

Fix for stretched grid with locally computed coordinates

This is a patch fix for geos/v2.7 with a fix for the stretched grid when using the locally computed coordinates.

---

What's Changed

• Update fv_grid_tools.F90 by @tclune in #88

Full Changelog: geos/v2.7.0...geos/v2.7.1

Add new algorithm for computing coordinates

This release introduces new algorithms for computing lat-lon corners of cubed sphere. These new routines use far less memory, as no global domain arrays are involved. They are also arguably more accurate, and run faster not that the latter are significant concerns.

These have not been validated in GEOS, and the default behavior is to use the older coordinates for backward compatibility. This can be toggled by setting compute_coords_locally = .true. in the fv_core_nml namelist. For best use, a corresponding calculation should be used by ESMF. Efforts are underway to add this change there.

---

What's Changed

• Feature/tclune/#81 compute coords locally by @tclune in #83
• GitFlow: Merge geos/develop into geos/main for release by @mathomp4 in #85

Full Changelog: geos/v2.6.0...geos/v2.7.0

Update to mapz (non-zero-diff)

This release has non-zero-diff changes to mapz from HWT experiments.

As noted, this is non-zero-diff to geos/v2.5.0

---

What's Changed

• pull in mapz implementation for T/PT by @sdrabenh in #78
• GitFlow: Merge geos/develop into geos/main for release by @mathomp4 in #80

Full Changelog: geos/v2.5.0...geos/v2.6.0

Latest Updates from HWT Experiments (zero-diff)

This release updates the T range check in FV Dynamics.

This is zero-diff to version geos/v2.4.1.

---

What's Changed

• Add push-to-develop to CI for MAPL3 work by @mathomp4 in #73
• Updates from feature/wmputman/hwt_spring_exp by @sdrabenh in #75
• GitFlow: Merge geos/develop into geos/main for release by @mathomp4 in #77

Full Changelog: geos/v2.4.1...geos/v2.5.0

Fixes for adiabatic dynamics (FV3 Standalone)

This PR has fixes needed for adiabatic runs of the GEOS FV3 Standalone from @wmputman. Without these changes, the FV3 standalone lost temperature and essentially ran steady-state at 0 K.

These changes are zero-diff to geos/v2.4.0.

---

What's Changed

• Fixes for adiabatic (aka FV3 Standalone) by @mathomp4 in #71
• GitFlow: Merge geos/develop into geos/main for release by @mathomp4 in #72

Full Changelog: geos/v2.4.0...geos/v2.4.1