Version 3 (modified by peylin, 7 years ago) (diff)


Summary of the retreat

How to proceed ?

  • We collectively need to finalize the table for the output variables before Christmas: ie, suggest all changes for the code and XIOS xml files

Table access:

  • Josefine will implement in the TRUNK the different modifications so that we have a new code in January with "cleaner outputs"
  • The suggested changes should follow a set of RULES (summarized below in point 4 or under the wiki)

ACTIONS FOR INDIVIDUALS: cross-check of the table

For a set of module we ask the following person to cross-check and complete the suggestions:

  • lpj_gap / lpj_establish / lpj_light / lpj_fire: Chao Y and Dan Z
  • npp_calc / season : Nicolas viovy and Daniel G
  • routing : Matthieu G and Jan P
  • sechiba_main: Nicolas Vui, Fuxing W, Jan P
  • thermosoil : Fuxing W and Frederique C
  • slowproc_main / stomatelpj : the Carbon group (see below)
  • turn : Nicolas Viovy

==> Any other cross-checked are welcome. Please report any problem to the organizing group

SPECIFIC GROUPS with dedicated tasks

We identified several actions to perform by little groups:

  • SNOW group (new snow scheme):

persons: Fuxing W, Catherine O, Frederique C

Action: identify the necessary variables to output and implement the XIOS output of these variable (in the code + xml file)

Time: Before end of 201

  • ALBEDO group

persons: Philippe P, Vladislav B, Catherine O, Fabienne M

Action: Simplify the output of the albedo variables as they are outputted in 3 places (make them in intersurf); Dont use the "axe2" but 3 different variables (VIS, NIR, TOT)

Time: before end of 2015

  • CARBON-variables group

persons: Fabienne M, Dan Z, Chao Y, Nicolas V, Nicolas V, Jonathan B, Philippe P, Daniel G,...

Actions: Verify all variables linked to carbon fluxes that are outputted in Stomate and Slowproc: drop the one in stomate that are EXACTLY the same as in slowproc, else rename the one in stomate/slowproc so that we see why they are different; check units to uniformize; add total variable (total per PFT) as "xxx_tot" for the variable name and field_id in xios.

Time: before end of 2015

  • TIME variable group

persons: Bertrand G, Josefine G, Jan P, Juliette L., ???

Action: Group in one unique module all temporal variables, change name for more coherence. Maybe include also the work on vertical axes (cleaning)

Time: Beginning 2016

  • CMIP6 variables:

persons: Patricia C, Frederique C, ??

Action: complete a table with all CMIP6 output that would indicate the "field_id" of the xios variable that should be used and the unit conversion.

Time: Beginning 201


persons: Agnes D, Frederique C, Patrica C, Josefine G, Fabienne M, Nicolas Viu, Matthieu G

Actions: Work on the ATLAS and MONITORING to check variables problems; update with the "free to use" observations gathered by FM; possibly include the "inter-comparison diagnostics" proposed by NV;

Time: Early 2016

  • NOBIO group

persons: Catherine O, Sebastiaan L, Nicolas Viovy, Fuxing W, Jan P, ..

Actions: Decide how to get read of the "normalisation by the fraction of "bio" land in stomate; possibly implement a new PFT-0 for Nobio land areas; Check properly what is currently done with nobio; propose a solution that avoid any confusion...

Time: Early 2016

  • Code cleaning and New Tiling

persons: Philippe P, Jan P, Nicolas V, Sebastiaan L, Nicola V, James R, Bertrand G,.....

Actions: Circulate across all a liste of the flag; decide which one (with the associated feature) that we want to drop for CM6_v2 code version; continue the reflexion on the different possibilities for the tiling (nb of energy and water budget given the set of PFT per grid)

Time: During 201

RULES TO FOLLOW to standardize and clean the outputs

All rules are summarize under:

  • A physical quantity (variable) should be outputted only once in the code; Operation like taking the mean/sum over a period, multiplying by contfrac or vegetmax, changing units,.. SHOULD be done through the XIOS xml files (field_def or file_def)
  • Variables that we want to sum over PFT or any other dimension: for now it should be done in the code by creating a new variable XXX_tot (keep this "_tot" convention) that is outputted by a call to xios_send..
  • For all variables we should try to follow SI units and we should output "fluxes" per square meter of the PFT (when PFT dependant) or per square meter of total bio-land (for a total variable). The second case may change after the NOBIO group investigation. Changing the variable to fluxes per square meter of the total grid area (or the total land area) should be done via xios in the xml file
  • All field_id should be in minuscule
  • if a variable is a fraction (no unit) the unit field should be: 1