wiki:DevelopmentActivities/MergeOCN

Merge OCN

This page describes the work done to include the nitrogen cycle into ORCHIDEE. The nitrogen cycle was introduced into ORCHIDEE by Sonke Zaehle a few years ago. This version called O-CN was not parallelized and is based on a obsolete version, so it will be progressively merged into a copy of trunk of ORCHIDEE with regular updates (synchronization with the trunk). The release is expected for the end of 2013.

Preliminary work : add extra dimensions for some carbon variables (October 2012)

Work done by Didier : the following variables (global) have now an extra dimension :

    turnover daily
    turnover_littercalc
    turnover_longterm
    biomass
    litter
    bm_to_litter
    bm_to_littercalc
    bm_sapl (parameter)
    bm_alloc

Some local variables need an extra dimension too :

lpj_cover dilu_lit
lpj_establish total_bm_sapl
stomate_lcchange bm_new, biomass_loss, dilu_lit
stomate_litter qd, litter_inc_PFT, litter_inc, litter_pft
stomate_lpj tot_bm_to_litter, tot_live_biomass, bm_alloc, tot_turnover
stomate_soilcarbon fluxtot, flux

Moreover, I need to add the following parameters :

INTEGER(i_std), PARAMETER :: nelements = 1    !! Number of isotopes considered 
INTEGER(i_std), PARAMETER :: icarbon = 1      !! Index for carbon  

About stomate_io : I didn't add the suffix '_n' like Sonke Zaehle did for the carbon variables. I think the restart of this version need to be compatible by the current ones. We should not add any suffix for carbon variables, only for nitrogen to keep the compatibility.

Test done : A simple binary comparison test was performed on Obelix. The tests are OK.

This work has been commited see [1025].

Meeting 24/04/2013

Present: Bertrand Guenet, Sebastiaan Luyssaert, Philippe Peylin, Ben Poulter, Nicolas Viovy, Nicolas Vuichard

Objective: To present and discuss the different steps and issues related to the merge of the OCN version into the Trunk.

NVui prepares some slides on the schedule (Merge-OCN.pdf). Overall agreement on the actions listed.
During the meeting, the following scheme has been defined in order to clarify the history of the OCN version at LSCE within the last years

the different versions of OCN at LSCE

It is still unclear :

  • what are the differences between OCN 0.70 and OCN 0.71 ?


  • how the OCN 0.71 can work without the corrections made on the 0.70 version to produce OCN NV2 ?

Ben will try to know more on the differences between OCN 0.70 and 0.71. On his side, Nicolas Viovy will try to run OCN 0.71 with his own driver, for comparison with 0.70.

We also agree on using this wiki for exchanging information on the OCN version. Consequently, I have created, 4 wiki pages (for Bertrand, Nicolas Viovy, Ben and Nicolas Vuichard). Here are the links for each of them. Please, do not hesitate to use these pages to document your former changes on the code of OCN and those that you will do in the future.

Meeting 30/01/2015

Present: Philippe Ciais, Daniel Goll, Philippe Peylin, Shushi Peng, Nicolas Vuichard

Status of the work of N. Vuichard

The implementation of the N-cycle into the "hybrid" version of ORCHIDEE is now finised. The hybrid version is the combination of the sechiba component from the trunk with the stomate component from the dofoco branch. Nicolas is still trying to get satisfying results at site level with this version and to debug it.

Future work

Into the trunk: Once the hybrid version gives good results, Nicolas will move on the trunk version, taking only the allocation scheme revised by Sebastiaan in the DOFOCO version. Nicolas Vuichard takes the lead on this action with the help of Sebastiaan (for the allocation scheme)

Into ORCHIDEE-CAN: The merge of the N cycle into ORCHIDEE-CAN can be done now, based on the differences between the current status of the hybrid version and the initial version of this branch. The improvements performed in the hybrid version in a near future (in order to fix the current bugs) will be reintroduced later in the ORCHIDEE-CAN version. Sebastiaan Luyssaert takes the lead on this action with the help of Nicolas.

Into MICT: If the integration of the N cycle into the MICT version starts now, it will be done based on the differences between the current status of the hybrid branch and the initial version of this branch. If the integration into the MICT version is done after the integration into the trunk, it will benefit of this later action and in particular of the differences between the trunk version with and without the N-cycle. Shushi Peng and Yan Sun take the lead on this action with the help of Nicolas.

People:

  • Nicolas Vuichard will
    • Finalize the development of the N-cycle in the trunk version of ORCHIDEE.
    • Evaluate this version at site level and at global scale
    • Coordinate the use of this C-N version of ORCHIDEE within the frame of CMIP6
    • Help for the inclusion of the N specificities into ORCHIDEE-CAN and ORCHIDEE-MICT
  • Sebastiaan Luyssaert will
    • Integrate the N-specificities into ORCHIDEE-CAN
  • Shushi Peng and Daniel Goll will
    • Integrate the N-specificities into ORCHIDEE-MICT
    • Include the P-cycle into this version
  • Yan Sun and Bertrand Guenet will
    • work on the multi-layering of the SOM module and later of the N-mineral module. This work will benefit to all branches.

Meeting 06/05/2015

Present: Philippe Ciais, Daniel Goll, Sebastiaan Luyssaert, Shushi Peng, Nicolas Vuichard

The tests done with the N-version in the "hybrid" version of orchidee give now satisfying results. Nicolas will commit his changes/modifications done the last months on the svn repository. Shushi and Daniel will report for the bugs they found in their respective wiki page and check for possible redundancies with the updated MERGE_OCN version, especially in stomate_growth_fun_all.

We agree on the fact that :

  • mass conservation should be checked in each subroutine as it is done already in ORCHIDEE-CAN.
  • a flag should be used for activating/deactivationg the C-N interactions into ORCHIDEE. One proposes to use as much as possible the same set of equations whatever the CN interactions are considered or not. In case the CN interactions are not activated, the CN ratios will be considered fixed along the simulations.
  • when the N-related process are N-specific (not a simple allometry to Carbon), to develop specific subroutines for these processes, separated of those developed for carbon.

Actions:

  • Nicolas will now develop a version based on the trunk of ORCHIDEE, that considers the N cycle. This implies to include all the developments coded in the "Hybrid" version + to include all the diameter class-related variables and the labile biomass pool variable from the ORCHIDEE-CAN version into the trunk.
  • In parallel, additional tests of the Hybrid version will be performed by Shushi, Daniel and Nicolas

Branches/MergeOCN/Poulter
Branches/MergeOCN/Vuichard

Last modified 5 years ago Last modified on 01/22/16 09:41:16

Attachments (2)

Download all attachments as: .zip