Refactoring of cohort and patch subroutines

[adrifoster]

I am working on some refactoring of the FATES derived types (i.e. site, patch, and cohort types). See my WIP branch here.

I am planning on moving some of the methods associated with patches and cohorts into modules for those types. Right now, in the modules EDCohortDynamicsMod and EDPatchDynamicsMod there seem to be subroutines that should belong cohorts or patches (or sites).

@glemieux suggested I put them up here with my thoughts so others can chime in:

@ckoven, @rgknox, @rosiealice, @billsacks, @mpaiao, @JessicaNeedham, @jkshuman

In EDCohortDynamicsMod (for now just focusing on these b/c there are a lot...):

1. create_cohort

• this subroutine allocates memory for a new cohort, sets initial values for all cohort attributes, and then inserts the cohort into the correct place in the patch's cohort list
• this has been modified by me so that the cohort-relevant methods (i.e. setting init values) are in their own cohort methods
• the insertion into the patch's linked list part seems like a patch method to me
• specifically, if we make the linked list parts actual methods, we can make more easily test them

2. InitPRTObject

• this subroutine initializes the PRT object to be associated with each cohort
• there are no cohort objects used inside the subroutine
• this seems like a PRT method to me

3. terminate_cohorts

• this subroutine loops through all the cohorts in a patch, checks to see which ones need to be deleted for various reasons, and then calls terminate_cohort (see below) and deallocates memory for each one marked for deletion
• this seems like a patch method (definitely NOT at cohort method) to me, or perhaps we should put it in some other kind of module

4. terminate_cohort

• this subroutine puts the litter from a cohort in the appropriate places on a site/patch and updates the C flux for the site, removes the cohort from the patch's linked list, and then calls DeallocateCohort on that cohort (see below)
• seems like mostly patch methods, or perhaps in some other module

5. SendCohortToLitter

• this subroutine transfers existing mass in all pools and all elements from a cohort to the appropriate litter pools
• needs access to the patch's litter pointer and the site's flux_diags pointer

6. DeallocateCohort

• deallocates dynamic memory and objects inside the cohort structure
• seems like a cohort method to me

7. fuse_cohorts

• this subroutine joins similar cohorts within a patch's cohort list
• this involves looping through a patch's cohort list (patch methods?) and
• rescaling the attributes of the new fused cohort (cohort methods?)

8. sort_cohorts

• sorts the cohorts within a patch's linked list into the correct order
• seems like patch methods to me

9. insert_cohort

• inserts a cohort into a patch's linked list
• seems like it should be patch methods

10. copy_cohort

• copies all attributes from one cohort to another empty cohort
• seems like cohort methods

11. count_cohorts

• counts the cohorts within a patch's linked list
• seems like patch methods

12. EvaluateAndCorrectDBH

• corrects DBH of a cohort to be allometrically consistent
• cohort methods?

13. DamageRecovery

• @JessicaNeedham perhaps you can weigh in?

[JessicaNeedham]

Thanks @adrifoster this looks great. DamageRecovery allows cohorts with crown damage to regrow some of their crowns. It’s mostly a cohort method as it assesses the fraction of a cohort that could recover, based on available CNP. But the recovered portion of a cohort is then assigned to a new cohort and has to be inserted into the cohort linked list, so more of a patch method? Maybe this routine will need to be split to fit with the new cohort and patch methods?