Version 3 (modified by peylin, 12 years ago) (diff)



20/09/2012 Developers meeting

Present: the meeting was held in Jussieu (~10 persons) with video links to the LSCE (~8 persons), Barcelona-Spain (Jan Polcher), Roma-Italy (Frederique), UA-Belgium (2 persons) and UG-Belgium (2 persons)

1. News

  • The latest trunk includes the documentation and the 11-layer hydrology scheme
  • The Lib IGCM meeting will be held on Octber 9th (Jussieu) and 10th (LSCE).
  • The structure of the new website is available, more work is needed on its content

2. The presentation

Camille Risi presented the inclusion of isotopes and tracers in the hydrological fluxes in ORCHIDEE.(see Presentation_isotopes)

3. Questions and Discussion

  • Although fractionation occurs for all phase changes, the current isotope implementation does not fully account for fractionation during freezing, thawing, snowing and smelting. The current models has some simple assumptions adressing this issue.
  • Individual molecules can be tagged and their path can be tracked. Tagging is available for geographical regions (e.g. the different oceans) or processes (e.g. transpiration).
  • It is hoped that the 11-layer hydrology could resolve the long standing issue of vertical distribution of isotopes in soil profiles.
  • For the moment, the istope scheme neglect molecular diffusion, including this process is not high on the priority list.
  • Isotopic fractionation could be applied on the AR5 simulations to predict continental recycling under future climate.
  • Isotopic fractionation hasn't been used for parameter optimisation yet but it seems feasible and worth trying.


05/07/2012 Developers meeting

Present: Camille Risi, Chao Yue, Matthieu Guimberteau, Kim Naudts, Aude valade, Sebastiaan Luyssaert, Juliette Lathiere, Anne Cozic, Josephine Ghattas, Niclas Vuichard, Didier Solyga, James Ryder, Jan Polcher (video), Patricia Cadule.

1. News

  • The latest trunk now includes diagnostic VOC emissions, this development is not tagged yet.
  • The documentation is now being merged to the hydrology branch which in turn is merged to the trunk. However, before this merge can be released, the water fluxes simulated the old hydrology scheme 'choinel' in the merge need to be identical to those simulated with the last tag (1.9.6). Following this release the new hydrology scheme will require some further testing, especially its interaction with stomate through plant water stress. A post-doc will be hired on the EMBRACE project to improve the soil-plant-atmosphere coupling of this merge.
  • Developers meetings will start again in September (date to be communicated).
  • In September a LibIGCM users meeting will be organised to discuss the current possibilities of this library of scripts that make up the run environment of ORCHIDEE (and the other components of the IPSL-ESM).

2. The presentation

The semi-analytic spin-up by Didier Solyga and Nicolas Vuichard (see Presentation_spinup)

3. Questions and discussion

  • The current approach relies on the matrix that describes the transfer coefficients between pools and their temperature/moisture sensitivity. Hence the approach should not limit the introduction of new carbon pools such as discretizing the active, slow and passive pools with soil depth. Similar, litter pools i.e. coarse woody debris could be added (as required for forest management).
  • The semi-analytic approach is integrated in the code and it is not straightforward to make the code more generic. This implies that the branch will have to move towards the trunk if they want to make use of this approach. The issue of branch divergence was discussed but did not result in an action.
  • The methods relies on the input to the litter pools to reach equilibrium soil C-pools. Fire, harvest, DOC or other lateral C-fluxes are accounted for in the soil C-pools. As a diagnostic NBP (net biome production) is stored in the restart files. NBP is not required to restart the model, nevertheless, it is stored in the restart to give the user a diagnostic variable to test the initial state of the simulation when the semi-analytic spin-up is used. NBP can be considered as some documentation of the spin-up.
  • The semi-analytic spin-up uses an objective criteria (epsilon) to stop the spin-up. Hence the spin-up does no longer depend on the user and a set of commonly used spin-ups could be stored and shared among users to save computation time.
  • Once the semi-analytic solution is merged in the trunk, forcesoil can be dumped and will no longer be maintained. The number of iterations required by the semi-analytic spin-up depends on the biomass. Hence, there is still room to use testomate. Given the possible instabilities of testomate, full ORCHIDEE could be used as an alternative. This will be further investigated.


DEV-Meeting May 10, 2012

Place: LSCE PRESENT: around 20 people from ORCHIDEE-DEV mailing list

The objective of the meeting was to present and discuss the results of the "ORCHIDEE - Hydrology 11 layers branch". This branch was created from the "AR5" main version, including all past developments on the new hydrology, including river routing, flood plains and some modifications due to soil freezing (LGGE work). The major contributions to the meeting were: (All presentations are accessible below)

1) Presentation of the 11 layer hydrology

Although prepared by Agnes (not present), the presentation was given by Matthieu G., with the exposition of the concept of this new hydrology scheme. A little discussion/precision was made on the changes compared to the initial scheme developped by Patricia De-Rosnay. It concern mainly the number of water budget per grid point. Initially a water budget was made depending on the soil characteristics (3 budget for three major types of soil) while now there is 3 budget for the three main vegetation classes (trees, grasses, crop).

Overall presentation prepared by Agnès Ducharne. See her Presentation.

2) Results from Matthieu for the Amazon basin

Matthieu compared several configurations (with/without STOMATE, different forcing files, and new hydro scheme versus old choisnel scheme) for the the Amazon basin. He looked at the river flow of the amazon and concluded that the current version with the new hydrology was leading to similar results than those obtained during previous work (with a different ORC version). Evaluation of the Hydro-Merge version over the Amazon River Basin by Matthieu Guimberteau. See his Presentation.

3) Result on the SIRTA (Aurelien)

(transparencies at the end of the general presentation of the new hydrology)

Aurelien Campoy presented few slides on a comparison made at the SIRTA site (Saclay Plateau) on a force mode (SAFRAN), without STOMATE. He compared for bare soil the evaporation from the "branch" version with that from an older version with the new hydrology (IGPBAC) and concluded on no significant differences

4) Results on the Mediterranean basin (Jan Polcher)

Jan compared simulations with new hydro between "the merge BRANCH" and his old working version, with WATCH forcing (without STOMATE). He obtained for July increased evapotranspiration with the merge branch, which seems to lead to a better agreement for the river discharge of major basins (Rhone).

Evaluation of the Hydro-Merge version over the Mediterranean Basin by Jan Polcher. See his Presentation.

4) Results at Fluxnet sites (Nicolas Vuichard)

Nicolas presentation a first comparison at several fluxnet site of the evapotranspiration between the new scheme and the old one WITH STOMATE. It turns out that the new hydrology leads to large day to day variations in the evapotranspiration that are probably unrealistics. This points to the strong need of "recalibrating the water stress functions with the new scheme and STOMATE"

Evaluation of the Hydro-Merge version on some Fluxnet sites by Nicolas Vuichard. See his Presentation.


DEV-Meeting March 15, 2012

Place: Jussieu PRESENT: around 20 people from ORCHIDEE-DEV mailing list

  • Presentation by Juliane Otto of a "New Radiative Transfert scheme for ORCHIDEE" to compute the Albedo. See her Presentation. An active discussion followed the presentation of Juliane and in particular several points were mentionned:
    • Need to change the current handling of bare soil : "gather all PFT fractions not covered by leaves into PFT1 (bare soil)" with a new treatment : "consider separately the fraction covered or not by leaves for each PFT independantly". The Radiative transfert scheme will directly benefit from this improvement.
    • The proposed RT scheme uses Monte Carlo simulations to compute the path of the photon within the canopy: This need to be change to reduce the computing time and to have "reproductive" results.
    • The proposed RT scheme uses the output of "Forestry Management Module" to compute the geometry of the stand from from the tree diameters classes, height,... This need to be generalized so that the RT scheme works also with ORCHIDEE standard (without FM). Juliane plans to use some "standard geometry" following some work by B. Pinty.
    • The proposed RT scheme uses three main parameters for the final albedo calculation (back ground albedo, single vegetation scattering albedo, preferential direction of scattering); these parameters are computed from MODIS satellite observations. Juliane presented values per PFTs for each month. This need to be generalized given that the seasonal variation are not the same btw North and South hemisphere for instance.
    • Valentin B. raise the issue that the computation of an "Effective LAI" for the new RT scheme (compare to the "TRUE LAI") might change significantly the gross carbon production and that this need to be check carefully.
    • The inclusion of the "snow" in this new RT scheme needs to be though more deeply. For that we propose to have soon a meeting with C. Ottle for a summary on the current developments of a new snow module for ORC and to discuss the coupling with the RT scheme.
    • The treatment of Grasses still needs to be done.

Note finally that Juliane is planning to first validate and publish the new RT scheme with FM module in a stand alone mode.

Attachments (69)