Meeting 13 January 2015 : Proposition of an agenda for CMIP6

Philippe P. presented few slides to i) resume the agenda of CMIP6, ii) make a proposition of possibles developements that would be included in the CMIP6 versions, and iii) to suggest how the group could be organized and could function. See slide ​CMIP6_roadmap_suggestions

Agenda of CMIP6

We propose to have two versions of ORCHIDEE for CMIP6:

• version 1 that needs to be ready by mid 2015 for first tests and could be updated/corrected until the end of 2015 for a full test early 2016. Note that LMDZ group will start with some LMDZOR test in April.
• version 2 that would be coupled with a version 2 of LMDZ (ie, with a different improved physic): to be ready Begining of 2017

Preparation of the Validation / Evaluation tools

For the group project, we propose to concentrate until March 2015 on the Validation chain for ORCHIDEE. This chain will be serve for both i) the ongoing ORCHIDEE paper and ii) the successive validation of any new ORCHIDEE version. Two mains actions are ongoing:

• Gathering and formatting all data sources that we want to use to validate ORC. Fabienne is responsible for that action. The following file (​Datasets for evaluation) describe all data streams that will be potentially used.

• Gathering and preparing different tools that will provide a range of diagnostics: from 2D maps, to time series of integrated quantities and cumulative histograms. Nicolas Vuichard coordinate this action with Josefine.

Version 1 of CMIP6 : scientific content

We propose to include the following model features (compared to the previous version used for CMIP5). Some options are sure and some are most likely too optimistic, but we keep them as a potential and they will anyhow be included in the version 2.

Version 1 is based on the current TRUNC (while version 2 will be based on ORC-CAN). Version 1 will rely and make use as much as possible of the developements made in ORC-MICT. All developments should be done trying to include them in parallele in ORCHIDEE-CAN for version 2

For sure or likely:

• 11 layers hydrology
• Uniform discretization for the soil hydrology and soil Thermics; plus improvement of soil heat capacity and conductivity
• Multi-layer snow module ?
• Improvement of the current albedo (Tuning of the fixed values for each PFT and bare soil)
• bug corrections in general..
• A new land cover map that will be common with JULES and JSBACH (from ECV-Lcover)
• SPITFIRE ?
• Representation of Winter vs Spring crops for temperate regions (adding a new PFT with few specific parameters)

Would be ideal but unlikely (it will depend on the progress of the group and especially the efforts in MICT)

• Nitrogen cycle
• Grassland management (GM)

Version 2 of CMIP6 : scientific content

This version will be based on ORCHIDEE-CAN, assuming that enough people will invest on the tuning and calibration of the version.

• 2 streams RT & Albedo
• Forest management & new carbon allocation
• Hydraulic lift of water within plants
• Nitrogen cycle
• New soil Carbon verticaly discretized model ?
• Global crop model (wheat, maize, rice, …) ?
• Phosphorus cycle and impact on C cycle ?

NOTES on the Functionning of the group to reach CMIP6 objectives =

We discussed a lot on the best way to integrate existing developements and how to agree accross the ORCHIDEE group. The following points emerged:

• We need first to define a "set of minimum criteria" to commit a new developement in the TRUNC: Conservation of Energy, Water and Carbon (to a reasonable order); suitability for parrallelisation, and appropriate documentation. Check / Implementation of the conservation diagnostics should be done as soon as possible.

• Each time a new model developement is proposed to the TRUNC, it needs to receive as much as possible the "agreement" of the "WHOLE (or MOST OF)" the ORCHIDEE group! This effort of "education" should not be avoided.

• We should work in parallele in small groups with: clear objectives, a precise roadmap, one person that is responsible (or a binome). Each group should report to the rest of the project team regularly (each month or two).

• We have to benefit as much as possible from the ongoing "faster" branch developement that are following specific scientific objectives (ex: MICT with the IMBALANCE_P and few others projects)

Initial comments on the MERGE

page gathering initial proposition early 2015: initial Proposition