How to output the fractions of different types of cloud

[KeerZ]

Hi all,

I am new to CAM and am confused about the meaning of various cloud fraction outputs. I am running a simulation with the BSSP585 compset using CESM2.1.3. The macrop_scheme and shallow_scheme ='CLUBB_SGS', the deep_scheme='ZM',and the microp_scheme ='MG'.

My first question is: what is the difference between CLOUD and CLOUDCOVER_CLUBB? Why does CLOUDCOVER_CLUBB have 33 vertical layers?

According to the CAM6 document, there are four types of cloud: marine stratocumulus cloud, convective cloud (including shallow and deep convective cloud), and layered cloud. I ran a sensitivity simulation and would like to separately examine how the four types of cloud change. I tried to output the CLOUD, CLDST, DP_CLD, and SH_CLD but the CLDST and SH_CLD are always zero because CLUBB handles the shallow convective. I wonder is there a way to get the cloud fractions of the above four types of cloud at each level?

Any suggestions are appreciated. Thanks!

 

[katec]

Ok, in CESM2 there are two convection and one microphysics parameterizations in CAM that work together to handle almost all of the cloud processes. CLUBB is a high-order closure turbulence parameterization that calculates all of the stratiform and shallow cumulus convection and Zhang-McFarlane is a mass-flux parameterization that handles deep convection. The CLOUD history field contains all of the cloud fractions from all of these parameterizations and gives you the final total cloud fraction. the CLOUDCOVER_CLUBB history variable outputs the cloud fraction used by CLUBB and microphysics (so everything NOT deep convection). The CLOUDCOVER_CLUBB field has 33 levels because CLUBB includes a below-surface ghost point at the lowest level and so has an extra level in its vertical grid.

I'm not sure which CAM6 document you found that describes these "four types" of clouds. I think most of us using the model would generally break clouds into two types: the shallow convection clouds simulated by CLUBB (which includes marine StratoCu and layered clouds) and the deep convective clouds simulated by Zhang-McFarlane. And it sounds like you have the right history fields for these.

 

[KeerZ]

Ok, in CESM2 there are two convection and one microphysics parameterizations in CAM that work together to handle almost all of the cloud processes. CLUBB is a high-order closure turbulence parameterization that calculates all of the stratiform and shallow cumulus convection and Zhang-McFarlane is a mass-flux parameterization that handles deep convection. The CLOUD history field contains all of the cloud fractions from all of these parameterizations and gives you the final total cloud fraction. the CLOUDCOVER_CLUBB history variable outputs the cloud fraction used by CLUBB and microphysics (so everything NOT deep convection). The CLOUDCOVER_CLUBB field has 33 levels because CLUBB includes a below-surface ghost point at the lowest level and so has an extra level in its vertical grid.

I'm not sure which CAM6 document you found that describes these "four types" of clouds. I think most of us using the model would generally break clouds into two types: the shallow convection clouds simulated by CLUBB (which includes marine StratoCu and layered clouds) and the deep convective clouds simulated by Zhang-McFarlane. And it sounds like you have the right history fields for these.

I see. Thank you for the clarification! I will use CLOUDCOVER_CLUBB (marine StratoCu + layered clouds) and DP_CLD (deep convection cloud) then.

The CAM6 document I read is 6. Model Physics — camdoc documentation, but perhaps it is not an up-to-date version because there are no descriptions on CLUBB. I wonder if a more recent document on cloud parameterization of CAM6 is available?

 

[katec]

I see. Thank you for the clarification! I will use CLOUDCOVER_CLUBB (marine StratoCu + layered clouds) and DP_CLD (deep convection cloud) then.

The CAM6 document I read is 6. Model Physics — camdoc documentation, but perhaps it is not an up-to-date version because there are no descriptions on CLUBB. I wonder if a more recent document on cloud parameterization of CAM6 is available?

This document is very strange. It says CAM6 but describes CAM5. I think it is a work in progress or a very early draft. There is no complete CAM 6 scientific description. Just the descriptions in various papers published in the last few years. The Gettelman et al (2019) paper has an in-depth description in the supporting materials. https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL083978

 

[KeerZ]

I see. Thank you for the clarification! I will use CLOUDCOVER_CLUBB (marine StratoCu + layered clouds) and DP_CLD (deep convection cloud) then.

The CAM6 document I read is 6. Model Physics — camdoc documentation, but perhaps it is not an up-to-date version because there are no descriptions on CLUBB. I wonder if a more recent document on cloud parameterization of CAM6 is available?

Hi, I reread the subroutine clubb_tend_cam code and it seems that my previous understandings were inaccurate. I wonder if the following understandings sound correct to you:

[line 2586 of clubb_intr.F90]
cloud_frac(i,k) = min(ast(i,k)+deepcu(i,k),1.0_r8)

Variable CLOUDCOVER_CLUBB (cloud_frac) is the sum of ast (Stratiform cloud) and deepcu (deep convective cloud), so it is similar to Variable 'CLOUD'. I checked that the values of CLOUDCOVER_CLUBB and CLOUD are indeed very close (but somehow not exactly the same) from layers 1 to 32.

To examine the fractions of deep convective and stratiform clouds, I should use DP_CLD (deepcu) and CLOUDFRAC_CLUBB (alst, liquid stratiform cloud fraction). However, the CLOUDFRAC_CLUBB does not include Ice stratiform cloud, so I should also output aist (Ice stratiform cloud fraction) or ast (total Stratiform cloud fraction) by adding some 'outfld' codes.

Besides, can I ask a follow-up question? Is there an easy way to compute the fractions of the convective/stratiform cloud that belongs to the low cloud based on the convective/stratiform cloud fraction at each level? I have read the code of subroutines cldsav and cldovrlap, but I wonder if there are any tools or packages that can be used to calculate the vertically-integrated low cloud fraction offline.

Thank you!