Update vegetation map (PFTmap) in begging of a new year instead of december

[jgipsl]

If VEGET_UPDATE=1Y and LAND_COVER_CHANGE=y then the vegetation map will be updated every last day in decembre. This makes it difficult to set up simulations. Many simulations have been necessaire to be redone because of mistakes.

Also, when running with TIME_LENGTH=1Y and starting without restart files, it is not possible to use the correct PFTmap for the first year and initialize correctly the restart files using the PFTmap for nextcoming year (using annual PFTmap files).

Proposition : Change to read and update PFTmap in the begging of the year, in Januray.

[jgipsl]

See also, ticket #70

[jgipsl]

Further proposition is done (N. Viovy, J. Ghattas):

• Update the PFTmap at the first time step in sechiba in a new year => the first time slowproc_main is called in a new year
• The variable LAND_COVER_CHANGE will be removed. This variable is currently used to deactivate lcchange_main even if the vegetation has been updated. This will not longer be possible. From now on, each time the vegetation map has been updated => lcchange_main will be called.
• lcchange_main will be called the first time stomate_lpj is done after an update of the vegetation map.
• The flag LAND_USE in run.def is confusing. It will be renamed to MAP_FORMAT_IN_PFT. The corresponding variable land_use in the code will be renamed to map_format_in_pft.
  • MAP_FORMAT_IN_PFT = y => Read map on format PFT
  • MAP_FORMAT_IN_PFT = n => Read Olson map

[jgipsl]

Done in trunk revision [2718]

Details:

• Parameter LAND_USE changed into PFT_FORMAT. Parameter LAND_COVER_CHANGE is removed.

• slowproc : New variables FirstTsYear: used to activate vegetation update if needed and LastTsDay: Used to activate do_slow

• veget_max is updated directly after call to slowproc_update. The previous values are saved and sent to stomate, to be used in stomate_lcchange. Same for frac_nobio.

• Slowproc_veget : is called directly after slowproc_update but also in the end of slowproc if do_slow. Removed redundant check for min values. See question 6.

• Removed frac_bare from slowproc: this variable is only used in hydrol and it is now calculated directly in hydrol using veget and veget_max

• Removed public variable tot_bare_soil in slowproc. This variable is now passed as intent(Out) from slowproc to be used in diffuco, condveg and hydrol and hydrolc as intent(in)

• lcchange is renamed into do_now_stomate_lcchange. The new variable is only true after slowproc_update has been done and until lcchange_main has beed called. Then it is reset to false in stomate_lpj.

• New subroutine stomate_veget_update : moved section 2.3.2.2 Total reduction from stomate_lcchange into new subroutine stomate_veget_update. This subroutine is called from slowproc_veget at each do_slow time step.

[jgipsl]

New opened related tickets

• conserve masse in stomate_veget_udpate #184
• regroupe time variables #185

[jgipsl]

There was a problem with the subroutine stomate_veget_update which was called every day but it should only be done if stomate_lcchange has been done. Following commit removes this subroutine. The update of veget_max is now done in the end of the first day instead of directly after it was read from the file. See commit on the trunk [3094]