[CF-metadata] control run base time convention

[Karl E.Taylor]

Dear all,

I have forgotten whether we decided on a recommended way to define
the base time for a control run of a coupled model that is supposed
to be roughly present day, but where individual years are not
expected to correspond with real years.  

One reason they are not expected to correspond is that in the
real world the forcing (greenhouse gases, for example) are
changing whereas in the model they are held constant.

thanks for any advice.

Karl