Fraction of frozen precipitation with RUC LSM

[tanyasmirnova]

Use fraction of frozen precipitation computed in Thompson or GFDL microphysics to fill array SRFLAG. This change will affect only the combination of RUC LSM with Thompson or GFDL microphysics, and will not affect any physics combinations not using RUC LSM. The reason for this change is to exclude temperature-based treatment of convective precipitation from the computation of SRFLAG, as is implemented in the WRF-based RAP and HRRR.

[climbfuji]

@mzhangw @ligiabernardet @llpcarson @grantfirl

[climbfuji]

This PR contains #336

[climbfuji]

Great work @tanyasmirnova, thanks very much for finding and fixing the Thompson MP bug. This change is completely safe from my point of view, but I'll give other GMTB folks a few minutes to look at it, too. I assumed you compiled this combined code change and ran a few tests?

[tanyasmirnova]

Dom,
Thank you very much for taking a look at this from Osaka. Yes, I have compiled both changes and ran a test on C96 resolution. I also printed the values of SRFLAG in my tests, and they look reasonable. Of course, I ran only the GSD suite tests.

[climbfuji]

Dom,
Thank you very much for taking a look at this from Osaka. Yes, I have compiled booth changes and ran a test on C96 resolution. I also printed the values of SRFLAG in my tests, and they look reasonable. Of course, I ran only the GSD suite tests.

That's great, thanks. Since the code changes only affect the CCPP Thompson MP and RUC LSM, this is good enough for merging into gsd/develop.

[ligiabernardet]

Thanks for finding the bug in Thompson pcp. I confirm that the changes in GFS_MP_generic.F90 will only affect RUC. Therefore, there is no risk of it inadvertently affecting other suites. Looks good.

[climbfuji]

All, we need to merge this update for gsd/develop into gmtb/develop and bring it to the master branches before the UFS release code freeze!