Incorrect freqsh definition in shallow convection diagnostics

[jimmielin]

What happened?

The freqsh diagnostic in CAM4 shallow convection code is incorrectly defined.

   ! --------------------------------------------------------!
   ! Calculate fractional occurance of shallow convection    !
   ! --------------------------------------------------------!

 ! Modification : I should check whether below computation of freqsh is correct.

   freqsh(:) = 0._r8
   do i = 1, ncol
      if( maxval(cmfmc2(i,:pver)) <= 0._r8 ) then
          freqsh(i) = 1._r8
      end if
   end do

From @adamrher:

That is indeed strange -- a value of 1 should be assigned to the column when there is shallow convection. The comment above it "! Modification : I should check whether below computation of freqsh is correct." suggests we aren't the only one who thinks so (can you git blame this comment?).

I looked through hk_conv.F90 and verified that the array (cmfmc2, which I presume is cmfmc_sh in your code) cannot be negative. There are explicit limiters on the components that are summed and multiplied to give cmfmc2: beta and eta:

CAM/src/physics/cam/hk_conv.F90

Line 697 in 6eb3046

if (eta(i) > tmass*rtdt .or. eta(i) <= 0.0_r8) eta(i) = 0.0_r8

.

And beta is limited using beta(i) = max(0.0_r8,beta(i)) statements anytime it's modified. So I believe this is a bug in this particular diagnostic, and that the logic should be:
if (maxval(cmfmc_sh(i,:)) > 0._kind_phys) then

What are the steps to reproduce the bug?

Observe definition of freqsh in a CAM4 model run

What CAM tag were you using?

cam6_4_055

What machine were you running CAM on?

CISL machine (e.g. cheyenne)

What compiler were you using?

Intel

Path to a case directory, if applicable

No response

Will you be addressing this bug yourself?

Yes

Extra info

Found this during CCPPization work of Hack shallow convection.

Will likely address this in a separate PR after the CCPPization, in order to maintain bit-for-bit of the eventual CCPP PR.