remove constraint on Thermodynamics

[szy21]

Purpose

CliMA/Thermodynamics.jl@1701c9a fixes the definition of virtual DSE for surface fluxes, and bumps a patch version, so the constraint on Thermodynamics is not needed anymore. Also updates tests and bump patch version.

To-do

Content

---

• I have read and checked the items on the review checklist.

[Sbozzolo]

I think it's better be on the safe side and put 0.11.5

[Sbozzolo]

CI is failing :'(

[szy21]

Hmm the recent change in Thermodynamics should cause behavior changes, but I thought it wouldn't break the tests. It might be just that we need to update the tests. Let me check.

[szy21]

@akshaysridhar Given we change the definition of virtual DSE (from cp_m to cp_d), the ΔDSEᵥ in test_cases is not accurate anymore. All the failing cases in test_cases are near neutral conditions, so I guess it is because ΔDSEᵥ changes sign from the definition change. I can just modify the tests or remove them. What do you think?

(Update: I checked the sign of ΔDSEᵥ and it indeed changes for those cases)

[akshaysridhar]

I think I agree that these test cases are linked to specific commits within the ClimaAtmos package (the classification was based on the tolerances and older versions of ClimaAtmos+SF - these are prior to Float32 improvements in SF and other physics changes within Atmos) - it makes little sense to maintain checks for those specific thermodynamic inputs; however I think we should track a separate issue then to catch edge cases that rely on our tolerance definitions.

[szy21]

I think I agree that these test cases are linked to specific commits within the ClimaAtmos package (the classification was based on the tolerances and older versions of ClimaAtmos+SF - these are prior to Float32 improvements in SF and other physics changes within Atmos) - it makes little sense to maintain checks for those specific thermodynamic inputs; however I think we should track a separate issue then to catch edge cases that rely on our tolerance definitions.

These cases are still edge cases, they just change from slightly unstable to slightly stable. So maybe we can keep them but change the sign for the test?

[szy21]

@akshaysridhar I update the tests with the new numbers. Would you like to take a look again? Some of them are not edge cases anymore, so we can also remove them if you want. (Edit: They are either close to neutral, or with very small wind speed, so may still be good to have.)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (f3da5ee) 90.05% compared to head (dbb8ac9) 90.05%.
Report is 1 commits behind head on main.

❗ Current head dbb8ac9 differs from pull request most recent head b7a1b52. Consider uploading reports for the commit b7a1b52 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #155   +/-   ##
=======================================
  Coverage   90.05%   90.05%           
=======================================
  Files           4        4           
  Lines         835      835           
=======================================
  Hits          752      752           
  Misses         83       83           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.