The article does not give much in the way of methods, but it would make the most sense that CO2 output was estimated using other data, not measured directly. So this is an "estimate", and to that I'd say, show me what data you relied on to estimate and whether the same quality of data exists for all cities. My guess is, the latter does not.