Satellite observations of anthropogenic carbon dioxide (CO2) emissions within urban settings offer unique potential to understand carbon sources and sinks and evaluate carbon mitigation strategies. Despite availability of column-averaged dry air mole fraction of CO2 ((Formula presented.)) from Orbiting Carbon Observatory-2 (OCO-2), temporal variations of (Formula presented.) and their drivers in cities remain poorly understood due to inconsistent definitions of urban extent, diverse urban forms, and unresolved impacts of urban vegetation on carbon fluxes. To this end, this study revealed that OCO-2 (Formula presented.) measurements from 2014 to 2018 exhibited statistically significant seasonal and trend components for each city. A correlation analysis suggested a weak association between (Formula presented.) trends and fossil fuel CO2 emissions ((Formula presented.)) trends but a close relationship between yearly average (Formula presented.) and (Formula presented.) trends. Vegetation abundance exhibited a negative relationship with the (Formula presented.) seasonality, though it only explained 21% of the variance. No statistically significant relationship between urban morphological factors (areal extent, complexity, and compactness) and temporal (Formula presented.) components was observed. However, urban morphological factors had a close relationship with the total amount of (Formula presented.) aggregated over the study period. Thus, it was speculated that urban morphological factors exerted their influence on (Formula presented.) through fossil fuel consumption. When only cities of high normalized difference vegetation index seasonality were used, statistically significant correlation coefficients between urban morphological factors and winter/summer averaged (Formula presented.) measurements were found. The variations of these correlation coefficients between leaf-on and leaf-off seasons stress the important role that urban trees play in mitigating carbon emissions in cities.