[tapios]

Retroactive compat cap.

ClimaParams 1.3.3 changed the default of KK2000_autoconversion_coeff_A from 7.42e13 to 1350 and added reference values (KK2000_reference_cloud_droplet_number_concentration, KK2000_reference_air_density) that the Khairoutdinov and Kogan (2000) kernels in CloudMicrophysics must read to produce the same rates (CliMA/ClimaParams.jl#335). CloudMicrophysics versions 0.42 - 0.45 use the old form of the kernel, so with ClimaParams 1.3.3 they compute KK2000 autoconversion rates about 10^11 too small (CloudMicrophysics' own test suite fails on main with 1.3.3). CloudMicrophysics 0.46.0, with the updated kernels and ClimaParams = "1.3.3" compat (CliMA/CloudMicrophysics.jl#814), is being released separately.

This caps ClimaParams at 1.3.2 for CloudMicrophysics 0.42 - 0.45; the entry's version range was 0.42 - 0, which only covers those four released versions today.

[tapios]

Rebased onto current master: CloudMicrophysics 0.46.0 was registered after this PR was opened, which split the ["0.42 - 0"] entry into ["0.42 - 0.45"] and ["0.46 - 0"] and left this PR conflicting. The change is now the single line ClimaParams = "1.2.0 - 1.3.2" under ["0.42 - 0.45"]; the new ["0.46 - 0"] entry already carries the 1.3.3 floor, so the two are consistent.

Context: ClimaParams 1.3.3 changed the default of KK2000_autoconversion_coeff_A and added reference values that only CloudMicrophysics >= 0.46 reads, so 0.42 - 0.45 compute Khairoutdinov-Kogan autoconversion rates about 10^11 too small with 1.3.3 (their own test suites fail). The RegistryCI checks passed on the previous revision. Could a registry maintainer merge this once the checks rerun? Thank you.