Fix ak/bk via interp_restarts

This release of FVdycoreCubed_GridComp fixes #309 by correcting the ak/bk written to an fv restart via interp_restarts.x to better match those in m_set_eta. The difference was mainly due to writing out r4 data as r8 leading to slightly different values. See below for more.

This is non-zero-diff in a subtle way. The NZD-ness comes in if you compare restarts remapped by v2.14 and older vs v2.15. So, if a user uses old restarts, then v2.14 and v2.15 are zero-diff. But if a user remaps restarts with v2.15 then the resulting restarts will cause the run to be non-zero-diff due to the changes in ak/bk.

Since our model uses the ak/bk from fvcore_internal_rst when running the model, this will then cause runs to be NZD.

---

The change makes the ak/bk match the ak/bk from m_set_eta.

v2.14:

 AK = 1, 2.00000023841858, 3.27000045776367, 4.75850105285645,
    6.60000133514404, 8.93450164794922, 11.9703016281128, 15.9495029449463,
    21.1349029541016, 27.8526058197021, 36.5041084289551, 47.5806083679199,
    61.6779098510742, 79.5134124755859, 101.944023132324, 130.051025390625,
...
 BK = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8.17541323527848e-09,
    0.00696002459153533, 0.0280100405216217, 0.0637200623750687,
    0.113602079451084, 0.156224086880684, 0.200350105762482,
...

v2.15:

 AK = 1, 2.0000002, 3.2700005, 4.7585009, 6.6000011, 8.9345014, 11.970302,
    15.949503, 21.134903, 27.852606, 36.504108, 47.58061, 61.677911,
    79.513413, 101.94402, 130.05102, 165.07903, 208.49704, 262.02105,
    327.64307, 407.6571, 504.6801, 621.68012, 761.98417, 929.2942, 1127.6902,
...
 BK = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8.175413e-09,
    0.0069600246, 0.028010041, 0.063720063, 0.11360208, 0.15622409,
    0.20035011, 0.24674112, 0.29440312, 0.34338113, 0.39289115, 0.44374018,
...

m_set_eta:

        data a72 / &
         1.0000000,       2.0000002,       3.2700005,       4.7585009,       6.6000011, &
         8.9345014,       11.970302,       15.949503,       21.134903,       27.852606, &
         36.504108,       47.580610,       61.677911,       79.513413,       101.94402, &
         130.05102,       165.07903,       208.49704,       262.02105,       327.64307, &
...

        data b72 / &
         0.0000000,       0.0000000,       0.0000000,       0.0000000,       0.0000000, &
...
         0.0000000,       0.0000000,       0.0000000,       0.0000000,       0.0000000, &
         0.0000000,   8.1754130e-09,    0.0069600246,     0.028010041,     0.063720063, &
        0.11360208,      0.15622409,      0.20035011,      0.24674112,      0.29440312, &
        0.34338113,      0.39289115,      0.44374018,      0.49459020,      0.54630418, &
...

---

What's Changed

• v11: Clean up fv3_setup and fv3.j by @mathomp4 in #295
• fixes #309 by @bena-nasa in #311
• GitFlow: Merge develop into main for release by @mathomp4 in #315

Full Changelog: v2.14.1...v2.15.0

Add MZ Export

This is a patch on v2.14 which adds a new MZ export which is a single-precision analogue to MFZ (which is double-precision). The X and Y already had both variants as an export.

This is brought over from v2.13.1

---

Full Changelog: v2.14.0...v2.14.1

v2.13: Add MZ export

This is a patch on v2.13 which adds a new MZ export which is a single-precision analogue to MFZ (which is double-precision). The X and Y already had both variants as an export.

---

Full Changelog: v2.13.0...v2.13.1

Remove OS Detection at NCCS

This release removes some now-unneeded OS detection at NCCS in fv3_setup as we no longer have the dual OSs.

---

What's Changed

• v11: Remove NCCS OS detection by @mathomp4 in #297
• Gitflow: Merge develop into main for release by @mathomp4 in #299

Full Changelog: v2.13.0...v2.14.0

Add dependency to fms_r4

This release adds a dependency to fms_r4. We need this because of the weird way we build FV3 as r4 but link to the r8 version of FMS. In that case, we still need to point to the include files from r4 FMS.

But, CMake in that case has no idea that to build, say, Moist, that we need to build fms_r4 before building this.

NOTE

This change is NOT needed in GEOSgcm v12 (i.e., v3) because in that case FMS is in Baselibs/Spack and so CMake doesn't need to know.

v1.2: Updates for CTM

This is a patch update to v1.2 with fixes for the AdvCore needed by CTM

---

Full Changelog: v1.2.17...v1.2.18

Fixes for Spack Runs of FV3 Standalone

This release of FVdycoreCubed_GridComp has minor fixes to fv3_setup and fv3.j for spack based runs of GEOSfvdycore

---

What's Changed

• Add GCM run to CI by @mathomp4 in #276
• GitFlow: Merge hotfix from main into develop by @mathomp4 in #282
• Updates for non-Baselibs runs by @mathomp4 in #284
• GitFlow: Merge develop into main for release (2024-08-26) by @mathomp4 in #286

Full Changelog: v2.11.1...v2.12.0

Fix use of variable before set

This patch release of v2.11 fixes a bug where in a variable was evaluated before set.

All testing shows this as zero-diff to v2.11.0

---

What's Changed

• Fix use of npz before being set by @mathomp4 in #280

Full Changelog: v2.11.0...v2.11.1

Use CMake to detect MPI Stack

This release of FVdycoreCubed_GridComp updates fv3_setup to use CMake detection of MPI stack rather than relying on $BASEDIR. This requires ESMA_cmake v3.41.0

---

What's Changed

• Update CircleCI to use Orb v2 by @mathomp4 in #267
• Use CMake to determine MPI Stack by @mathomp4 in #270
• GitFlow: Merge Develop into Main by @mathomp4 in #272

Full Changelog: v2.10.0...v2.11.0

v1: Modify LONG_NAME of DOXDTDYN

This is an update to the v1 release of FVdycoreCubed_GridComp. The release modifies the LONG_NAME of DOXDTDYN

---

What's Changed

• Modify LONG_NAME of DOXDTDYN by @sdrabenh in #266

Full Changelog: v1.13.0...v1.13.1