As suggested [here](https://github.com/iiasa/message_doc/pull/27#pullrequestreview-273586234) in #27: > We may want to refer to the original GDP trajectories (Dellink probably?) and just say that we supplied these trajectories to calibration [routines].
As suggested here in #27: