The story this graph tells me is that of the increasing ability of income groups other than high-income to generate CO2, rather than a decrease in CO2 generation of high-income people.

But it's all still rather ambiguous.

As overall CO2 emissions have changed so much (and non-linearly) over the graph's timespan, this graph doesn't provide much useful information on its own.

Also, without information on the income divisions themselves (are they quartiles by income or by population, or some other division?), it's difficult to interpret.