GHG data for CMIP6 are initially coming from 
ftp://data.iac.ethz.ch/CMIP6/input4MIPs/UoM/GHGConc/CMIP
see also http://www.climate-energy-college.net/cmip6
They have been in September replaced by v1-2-0 from ESGF. 
The format of the netcdf has changed with one variable (and sectors) per file.
The data were downloaded in September 2016 and are available 
on ciclad in /prodigfs/project/input4MIPs/UoM

The data are processed to produce annual global mean files 
for the IPSL model. This is currently done on ciclad in 
/home/oboucher/CMIP6/GHG but will be moved in a central repository.

The GHG to be used are 

CO2
CH4
N2O

and either 

CFC11=CFC11eq
CFC12=CFC12

OR 

HFC134a-eq
CFC12eq 

Do not use the original CFC11eq and CFC12eq together as some species 
would be double-counted

At IPSL we chose the second way of doing for CMIP5, and we do likewise for CMIP6.
HFC134-eq are converted into CFC11-eq using AR5 values for their 2011 RF and 
volume mixing ratios (Table 8.2 from Chapter 8, CFC11 0.062 Wm-2 for 238 pptv and 
HFC134eq 0.0100 Wm-2 for 62.7 pptv), which gives the following conversions: 
CFC11eq_CMIP6 <== HFC134a-eq * (0.0100/62.7)/(0.062/238.0)
CFC12eq_CMIP6 <== CFC12eq 

So in the end, we have 5 files for GHG for CMIP6 runs with IPSL-CM6
CO2_CMIP6_0000_2014.txt
CH4_CMIP6_0000_2014.txt
N2O_CMIP6_0000_2014.txt
CFC11eq_CMIP6_0000_2014.txt
CFC12eq_CMIP6_0000_2014.txt

Note that the CFCeq files have non-zero values for pre-industrial conditions 
because they include some natural species.

To produce the data used in CMIP6, just run the extract.sh script

The plots in graphs are from Jean-Louis, blue is CMIP6, green is CMIP5.

Olivier
11/10/2016
updated 13/12/2016