488 merge canopy developments from am3 #489
Conversation
…e rest of the model
which is the only place it is used
cable_surface_type_namelist - which will be implemented here as well at a later date
off and ICE density used that has changed as well
ws format
Co-authored-by: Claire Carouge <ccarouge@users.noreply.github.com>
Co-authored-by: Claire Carouge <ccarouge@users.noreply.github.com>
…lopments-from-am3
capitalized as automatically triggers -fpp compilation flag AND every other files is F90 AND every JULES file is F90
CMakeLists.txt
Outdated
There was a problem hiding this comment.
AM3 implementation of init_wetfac crept in. Although f90 should be
capitalized as automatically triggers -fpp compilation flag AND every
other files is F90 AND every JULES file is F90
src/science/canopy/cable_canopy.F90
Outdated
There was a problem hiding this comment.
AM3 version treatment of conversion (assuming 1e3 corresponds to units)
! Calculate dewfall: from negative lh wet canopy + neg. lh dry canopy:
canopy%dewmm = - (MIN(0.0,canopy%fevw) + MIN(0.0_r_2,canopy%fevc)) * &
dels / air%rlam
and cansto
There was a problem hiding this comment.
AM3 had neater interface. The explicit REAL typecasting is perhaps unnecessary as it is handled here that was used in AM3 - HOWEVER continued removal of r_2 is advantageous. As argued a million times, REAL(selected_real_kind) does not change CPU hardeware. it's either a 32,64,128 bit machine. All values of selected_real_kind fall into their nearest bin (being either 32,64,128 bit etc). Here r_2 defaults to double. In the UM everything is compiled as -i8, -r8 so r_2 is meaningless and only causes complications in development/refactoring especially when TYPES/kinds dont match up. Additionally the UKMO discouraged us from using r_2, special KINDs. However they have since adopted them themselves. Having many different KINDs on either side of the interface is problematic. Last I spoke with them they were uncertain of what this would all look like with LFRic. At any rate it is easier if we just lay off the r_2 for the moment and re-implement a consistent approach at a later date when we know better what is happening.
| @@ -369,15 +381,8 @@ SUBROUTINE dryLeaf( dels, rad, rough, air, met, & | |||
| rdx(i,2) = rdx(i,2) * & | |||
| (0.5 - 0.05*LOG(jtomol*1.0e6*rad%qcan(i,1,2))) | |||
|
|
|||
| !$ xleuning(i,1) = ( fwsoil(i) / ( csx(i,1) - co2cp3 ) ) & | |||
| !$ * ( veg%a1gs(i) / ( 1.0 + dsx(i)/veg%d0gs(i))) | |||
| !$ xleuning(i,2) = ( fwsoil(i) / ( csx(i,2) - co2cp3 ) ) & | |||
There was a problem hiding this comment.
This commenting style has lead to issues in some UM builds. Its redundant code anyway
There was a problem hiding this comment.
This has just had the f90 capitalised
|
@ccarouge - There is still the soilsnow stuff to go in. Fortunately it looks pretty good porting back to AM3. There is an issue around dewmm & canto that I mentioned here, and 1 other, all in cable_canopy |
|
@JhanSrbinovsky Have you tested these changes in benchcab? What's the effect on the results? |
2024-11-27 13:32:16,679 - INFO - benchcab.benchcab.py:391 - 0 failed, 20 passed |
CABLE
Thank you for submitting a pull request to the CABLE Project.
Description
Please include a brief summary of the change and list the issues that are fixed.
Please also include relevant motivation and context.
You can link issues by using a supported keyword in the pull request's description or in a commit message:
Fixes #(issue)
Type of change
Please delete options that are not relevant.
Checklist
Please add a reviewer when ready for review.
📚 Documentation preview 📚: https://cable--489.org.readthedocs.build/en/489/