remove maxcohort dependencies in ED Restarts

[rgknox]

One of the key steps towards enabling greater flexibility in numpft_ed is to break the dependence of maxcohort to be > nlevcan_ed x numpft_ed x nclmax, and for non-cohort dimensioned restart arrays to be given their own dimensions and IO offsets.

Note: this update will also enable steps towards the interface, as the offsets and vector pointers have to this point been saved in CLM parts of the code. The migration of the new pointer memory could be staged as a second commit if appropriate.

Here are the specific variables that borrow maxcoohort for restart IO:

numpft_ed x patch:
this%leaf_litter(countPft)
this%root_litter(countPft)
this%leaf_litter_in(countPft)
this%root_litter_in(countPft)
this%seed_bank(countPft)

ncwd x patch:
this%cwd_ag(countNcwd)
this%cwd_bg(countNcwd)

nclmax x patch:
this%spread(countNclmax)

nlevcan_ed x numpft_ed x nclmax x patch:
this%f_sun(countSunZ)
this%fabd_sun_z(countSunZ)
this%fabi_sun_z(countSunZ)
this%fabd_sha_z(countSunZ)
this%fabi_sha_z(countSunZ)

nwatermem x site:
this%watermemory (or something like that, lol)

[bandre-ucar]

The changes sound like a bug fix for max pft / max cohort. I'd prefer that they come in their own pull request from the pointer migration.

[ckoven]

I see a conceptually-related issue that ought to be fixed. currently a few variables in the ed_clm_type are indexed as 2d with the second dimension nlevgrnd. these are:
PFTbiomass
PFTleafbiomass
PFTstorebiomass
PFTnindivs
Really, the second dimension of these variables ought to be numpft. doing this requires adding that as a CLM history file dimension, like how "levscpf" was added already, and reordering the variables. otherwise there is the potential to crash the model or worse if mxpft exceeds nlevgrnd.

[rosiealice]

Yes - they shouldn't be nlevgrnd. I just used that because in my
pre-software engineer assisted days, nlevgrnd was the only instance of
outputting 3D variables... That seems like a long time ago now.

On 29 February 2016 at 10:57, Charlie Koven [email protected]
wrote:

I see a conceptually-related issue that ought to be fixed. currently a few
variables in the ed_clm_type are indexed as 2d with the second dimension
nlevgrnd. these are:
PFTbiomass
PFTleafbiomass
PFTstorebiomass
PFTnindivs
Really, the second dimension of these variables ought to be numpft. doing
this requires adding that as a CLM history file dimension, like how
"levscpf" was added already, and reordering the variables. otherwise there
is the potential to crash the model or worse if mxpft exceeds nlevgrnd.

—
Reply to this email directly or view it on GitHub
#12 (comment).

---

Dr Rosie A. Fisher

Terrestrial Sciences Section
Climate and Global Dynamics
National Center for Atmospheric Research
1850 Table Mesa Drive
Boulder, Colorado, 80305
USA.
+1 303-497-1706

http://www.cgd.ucar.edu/staff/rfisher/