Unify rair constant in CAM calculations

[jimmielin]

What is the feature/what would you like to discuss?

This was discovered in code review for #1271

There are a few instances of a magic number (287.15) being used in CAM code that is not referencing physconst.

@nusbaume wrote:

I am like 95% sure that this 287 number is the gas constant for dry air. I might vote to either just use the actual physical constant, or if that changes answers too much to just create a new parameter here labeling it as the gas constant but keeping the actual value the same.

@jimmielin wrote:
This should be rair but the numbers in shr_const_rdair work out to something closer to ~287.04:

Breakpoint 1, physconst::physconst_readnl (nlfile=..., _nlfile=_nlfile@entry=80) at ...CAMdev64_075/src/utils/physconst.F90:273
273	   end subroutine physconst_readnl
(gdb) p rair
$1 = 287.03999999999996

It is used RK_stratiform to calculate the ice/liquid water content

 1316:          iwc(i,k)   = state1%q(i,k,ixcldice)*state1%pmid(i,k)/(287.15_r8*state1%t(i,k))
 1317:          lwc(i,k)   = state1%q(i,k,ixcldliq)*state1%pmid(i,k)/(287.15_r8*state1%t(i,k))

and a couple of other places for diagnostics:

src/physics/cam/cloud_diagnostics.F90:
  405               icimr(i,k)     = min( allcld_ice(i,k) / max(0.0001_r8,cld(i,k)),0.005_r8 )
  406               icwmr(i,k)     = min( allcld_liq(i,k) / max(0.0001_r8,cld(i,k)),0.005_r8 )
  407:              iwc(i,k)       = allcld_ice(i,k) * state%pmid(i,k) / (287.15_r8*state%t(i,k))
  408:              lwc(i,k)       = allcld_liq(i,k) * state%pmid(i,k) / (287.15_r8*state%t(i,k))
  409               ! Calculate total cloud water paths in each layer
  410               iciwp(i,k)     = icimr(i,k) * state%pdel(i,k) / gravit

src/physics/cam/micro_pumas_cam.F90:
 2642           icwmrst(i,k)   = min( state_loc%q(i,k,ixcldliq) / max(mincld,liqcldf(i,k)),0.005_r8 )
 2643           icinc(i,k)     = state_loc%q(i,k,ixnumice) / max(mincld,icecldf(i,k)) * &
 2644:               state_loc%pmid(i,k) / (287.15_r8*state_loc%t(i,k))
 2645           icwnc(i,k)     = state_loc%q(i,k,ixnumliq) / max(mincld,liqcldf(i,k)) * &
 2646:               state_loc%pmid(i,k) / (287.15_r8*state_loc%t(i,k))
 2647           ! Calculate micro_pumas_cam cloud water paths in each layer
 2648           ! Note: uses stratiform cloud fraction!

So maybe they should all be updated in one go as there will be answer changes.

Is there anyone in particular you want to be part of this conversation?

@nusbaume

Will this change (regression test) answers?

Yes

Will you be implementing this enhancement yourself?

Any CAM SE can do this

[PeterHjortLauritzen]

Agreed ... lets change to cpair in CAM4 and CAM6/7