Opened 4 years ago

Closed 4 years ago

#269 closed enhancement (fixed)

Merge with IPSLCM6.v1

Reported by: luyssaert Owned by: luyssaert
Priority: major Milestone: IPSLCM6.v2
Component: Physical processes Version:
Keywords: Cc:


(1) Merge with IPSLCM6.v1

Change History (3)

comment:1 Changed 4 years ago by luyssaert

ORCHIDEE-CN-CAN started from the branch ORCHIDEE-CN r3238 the start of the ORCHIDEE-CN-CAN branch was recorded as revision 3252. Since the start of the ORCHIDEE-CN-CAN branch the following major developments (originally made in ORCHIDEE-DOFOCO) were implemented and tested in ORCHIDEE-CN-CAN:

  • The code was extended and tested for > 1 circumference class. This change was already anticipated in ORCHIDEE-CN. This resulted in changes in almost all modules in src_stomate.
  • Mortality of vegetation now distinguished natural and anthropogenic processes. This resulted in the addition of new modules: sapiens_agriculture.f90, sapiens_forestry.f90, sapiens_kill.f90, sapiens_lcchange.f90 and sapiens_product_use.f90.
  • Mass balance and surface area conservation was checked for in all appropriate modules and subroutines. This resulted in changes in almost all modules in src_stomate. The mass balance check itself was coded as a subroutine in src_parameters/function_library.f90
  • Stand structure is now calculated in support of the albedo calculations. This resulted in the addition the new module stomate_stand_structure.f90.
  • Multi-layer albedo became the sole and thus default albedo scheme. Reflectance of the multi-layer albedo scheme is used to calculate NIR and VIS albedo. As a result src_parameters/structures.f90 and src_sechiba/albedo.f90 were added.
  • Transmittance from the multi-layer albedo scheme is used to calculate veget. This resulted in relatively few changes in src_stomate and src_sechiba
  • Absorbance of the multi-layer albedo scheme is used to calculate photosynthesis. Changes are constrainted to diffuco.f90. Note that because of this photosynthesis is now calculated for layers with an actual thickness (in meter) rather than for LAI-layers as in the trunk.
  • Hydraulic architecture was implemented and used in the calculation of water stress. Water stress is applied to stomatal conductance. Contrary to ORCHIDEE-DOFOCO the old scheme for water stress was kept as an alternative. This resulted in a new module src_sechiba/sechiba_hydrol_arch.f90 and changes in stomate.f90 where water stress is calculated. Applying the stress to stomatal conductance resulted in changes in diffuco.f90.
  • The new roughness scheme by Niocolas Vuichard has been implemented. Changes are more or less restricted to condveg.f90.
  • The multi-layer energy budget scheme from ORCHIDEE-DOFOCO has been implemented followed by some cleaning. The multi-layer energy budget resulted in a overhaul of enerbil.90, diffuco.f90 and many changes in sechiba.f90.

At the same time ORCHIDEE-CN was further developed with the most notable changes:

  • Dynamic SLA
  • Update of the code with the trunk up to r3623. This is committed in ORCHIDEE-CN in revision r3714. As a result several changes were made to the architecture and driver files. The following new routines were added to src_parameters: vertical_soil.f90, vertical_soil_var.f90. To src_sechiba: chemistry.f90 and thermosoilc.f90. Also many changes were made to condveg.f90, diffuco.f90, enerbil.f90, sechiba.f90, slowproc.f90 and stomate.f90.
  • After this update to the trunk relatively well constrainted bug-fixes and enhancements were made (r3715 to r4109).

OBJECTIVE: update ORCHIDEE-CN-CAN with all appropriate changes made to ORCHIDEE-CN between r3238 and r4109.

WORKPLAN: svn merge -r 3238:4109 svn:// --dry-run. That didn't work so ediff and other related tools and approaches were used. In a first phase all stomate and parameter files were merged. All conflicts were resolved to make the code compile. These changes were committed in r4154. In r4154 all calls in stomate_lpj were still commented. Most of the calls (except to stomate_turnover.f90) were uncommented, the code compiles, small issues were solved and committed as r4161. A basic version of all CN-CAN functionality for stomate (including turnover) has been committed as r4163. Subsequently, the CN-CAN changes to sechiba and related modules were merged with ORCHIDEE-CN r3714. After fixing several minor issues the changes were committed as r4206. This version was tested for a single pixel with all PFTs and a 50 year run was completed with a crash.

At March 30th 2017, r4206 contains the functionality of ORCHIDEE-CN r4109 and ORCHIDEE-CN-CAN r4106. Next: (1) update this version with the latest commits to ORCHIDEE-CN since 4109 and (2) update this version with the latest commits if not already included in ORCHIDEE-CN.

KNOWN ISSUES for r4206:

  • ORCHIDEE-CN-CAN makes use of the multi-level albedo scheme. Although this scheme uses an explicit background albedo the current optimized background albedo map should not be used (because the vegetation albedo is very different). References to this background albedo map can be removed from the merge
  • ORCHIDEE-CN-CAN was developed for CWRR only. The trunk still has the option to use either CWRR and CHOINEL. This is a good time to remove CHOINEL.
  • The explicit snow scheme does not simulate a VIS and NIR albedo. Both bands are required for the albedo scheme.
  • assim_param was removed from ORCHIDEE-CN-CAN. It was put back in through the merge with ORCHIDEE-CN.
  • veget, lai and height were removed from call arguments in ORCHIDEE-DOFOCO. They were (partly) put back in through the merge with ORCHIDEE-CN.
  • tot_bare_soil is partly disabled. There may be several scientific problems left.
  • Many redundant parameters are now defined. This requires a major cleaning effort.
  • New variables related to veget_max were introduced (or old variables renamed) in slowproc.f90. This should be carefully checked.
  • XIOS has not been tested yet.
  • In ORCHIDEE-CN ext_coeff_N is used to calculate photosynthesis as well as to calculate the lai from biomass. The dynamics of photosynthesis and lai are consistent. In r4206 of CN-CAN photosynthesis calculates the N-profile from the absorbed light profile which in turn is calculated from the 2-stream albedo module. The lai dynamics follow the CN approach and are calculated from ext_coeff_N. Lai and photosynthesis now have different dynamics. A similar inconsistency in present in the calculation of leafN_top in diffuco.f90. Implementing a scientifically sound and technically clean solution will require some thought and time. At that time, it can be explore how the dynamic sla can be fully integrated in the allocation scheme.
Last edited 4 years ago by luyssaert (previous) (diff)

comment:2 Changed 4 years ago by luyssaert

  • Owner changed from somebody to luyssaert
  • Status changed from new to accepted

comment:3 Changed 4 years ago by luyssaert

  • Resolution set to fixed
  • Status changed from accepted to closed
Note: See TracTickets for help on using tickets.