The flux site simulations have a constant CO2 concentration. This is because it is read in with the wrong number of dimensions. The ground water work has changed the code for the CO2 read to inquire the number of dimensions from the file and use that to read either a 3D or 4D variable.

The problem is clearly seen by compariing the original CO2 concentrations in the met forcing files and the output from CABLE.
Original CO2 concentrations for AU-ASM in the met forcing file are:

 CO2air =
  609.6982,
  613.0284,
  614.9662,
  607.8287,
...

The output from CABLE main is:

 CO2air =
  609.6982,
  609.6982,
  609.6982,
  609.6982,
...