ISIMIP2b: ORCHIDEE bug on soil carbon change (new runs)

Posted by Matthias Büchner on Dec. 13, 2018

This change applies to the sectors Biomes, Global Water and Permafrost.

As announced here on Sept. 26, 2018 the ORCHIDEE experiments from historical towards extended future period were now replaced with updated runs. A few experiments could not be repeated due to limited capacities (see below).

In order to distinguish old and new data please look at the global NetCDF attribute "ismip_qc_version" which is set to "2" for the new data version.

Additional notes from the modeler:

  1. The uploaded output is downscaled from 1° to 0.5 degree. However, a different method was used for old and new runs. An R program was used for old runs including an additional land-sea mask, while nco+cdo is used to directly downscale output from new runs, which does not follow the 0.5 degree land-sea mask. Hence the grids that can be used for calculation/aggregation depends on the land-sea mask.
  2. The inconsistency of masks will not affect any conclusion derived from the simulations. If you have difficulties in using those files, please let the modeler know. His contact information is listed in each file and in the Impact Model Settings & Characteristics of the ISIMIP homepage.
  3. Although the variables rh, clitter, csoil, and ecoatmflux were most affected by the bug, all variables have been replaced because of the complicated connections within the model between carbon, energy and water. That's why none of the old data from the replaced runs should be used anymore.
  4. Because of the limitations for re-running, not all experiments could be repeated. Hence, some files which existed as part of the old runs do not exist anymore and will not be uploaded.

Periods with replaced runs:

  • historical:
  • future:
  • future_extended (only IPSL):

This translates into the following replaced experiments:

I (1861-2299)

II (1861-2299)

IIa (2006-2099)

III (2006-2099)

IIIa (2006-2099)

VI (2006-2299)

VII (2006-2099)

Experiments retracted but not replaced:

IV and V