Hard to argue with robust monitoring of CO2 concentrations which are now increasing linearly rather than exponentially (i.e emissions are constant instead of increasing).
I am not sure... with caveats about timeseries ofc, plotting the yearly mean with a linear fit shows a curve where values in 1960 and 2010s are above the line and the between years are below. This would fit with a non constant slope that is slowly increasing in steepness.
The size of the increase each year is much higher now than in 1980s even accounting for the increased CO2 value. So the rate of change has increased.
