Missing Cloudfree aerosol optical depth diagnostics in noresm_develop

[Ovewh]

The diagnotics of clearsky/cloudfree aerosol optical depth did not make it through the refactoring of Oslo_Aero.

These diagnostics might still be useful for evaluations against satellite AOD and AERONET and was a standard diagnostic in AeroCom Phase III experiment @oyvindseland @MichaelSchulzMETNO ? If these diagnostics are still needed i can make PR to re-introduce them.

The codeblock missing from Oslo_Aero noresm_develop.

       ! SLO_AERO begin
       !Calculate cloud-free fraction assuming random overlap
       !(kind of duplicated from cloud_cover_diags::cldsav)
       cloudfree(1:ncol)    = 1.0_r8
       cloudfreemax(1:ncol) = 1.0_r8

       !Find cloud-free fraction (note this duplicated code and may not be consistent with cldtot calculated elsewhere)
       do k = 1, pver
          do i=1,ncol
             cloudfree(i) = cloudfree(i) * cloudfreemax(i)
             cloudfreemax(i) = min(cloudfreemax(i),1.0_r8-cld(i,k))
          end do
       end do

       !Calculate AOD (visible) for cloud free
       do i = 1, ncol
          clearodvis(i)=cloudfree(i)*aodvis(i)
          clearabsvis(i)=cloudfree(i)*absvis(i)
       end do

       !  clear-sky AOD and absorptive AOD for visible wavelength close to 0.55 um (0.35-0.64)
       !  Note that caodvis and cabsvis output should be devided by dayfoc*cloudfree to give physical (A)AOD values
       call outfld('CAODVIS ',clearodvis,pcols,lchnk)
       call outfld('CABSVIS ',clearabsvis,pcols,lchnk)
       call outfld('CLDFREE ',cloudfree,pcols,lchnk)

#ifdef AEROCOM
       do i = 1, ncol
          clearod440(i)     =cloudfree(i)*dod440(i)
          clearod550(i)     =cloudfree(i)*dod550(i)
          clearod870(i)     =cloudfree(i)*dod870(i)
          clearabs550(i)    =cloudfree(i)*abs550(i)
          clearabs550alt(i) =cloudfree(i)*abs550alt(i)
       end do
       call outfld('CDOD440 ',clearod440  ,pcols,lchnk)
       call outfld('CDOD550 ',clearod550  ,pcols,lchnk)
       call outfld('CDOD870 ',clearod870  ,pcols,lchnk)
       call outfld('CABS550 ',clearabs550  ,pcols,lchnk)
       call outfld('CABS550A',clearabs550alt,pcols,lchnk)
#endif
       ! OSLO_AERO end

[MichaelSchulzMETNO]

Weighting by the cloud fraction does not give the real "clear sky AOD". For a monthly mean it does give higher weight to those days within a month with little cloud cover. I believe, clear sky AOD should be computed using the (lower) humidity in clear skies. Old discussion. I wonder which humidity is used for AOD calculation in the present Oslo-Aer model. Can this be checked.

Clear sky AOD is likely smaller than all sky AOD, likely because humidity is lower in clear sky conditions. If one would use a "clear sky" AOD than a negative bias would become even more negative.

[gold2718]

I think this issue has a lot of overlap with #33. Thoughts?