Concerning "spAer" timeslice simulation : 
========================================
The Data Request only expected up to 2 members (r1 and r2) for piClim-spAer-aer and piClim-spAer-anthro. 
So, the files dr2xml for r3 and r4 were created by S. Denvil's scripts, but they are empty.
Therefore, the r3 and r4 members were run with "Member=r2i1p1f1" in config.card, 
  and the output was renamed by G. Levavasseur (r2 -> r3, r2 -> r4).


General info : 
=============
dr2xml files for the RFMIP experiments were produced by S. Denvil, then by A. Caubel, here :
/ccc/work/cont003/igcmg/igcmg/IGCM/CMIP6/IPSLCM6.1.6-LR/IPSL-CM6A-LR

For each simulation : 
     config.card specifies the member : Member=r2i1p1f1"
     and lmdz.card specifies the path to dr2xml as follows:
  ${R_IN}/CMIP6/${config_UserChoices_LongName}/${config_UserChoices_ModelName}/${config_UserChoices_ExperimentName}/${config_UserChoices_Member}/${year}/dr2xml_LMDZ.xml