Comparing city fossil fuel CO2 emissions between ODIAC and inventory-based estimates

Measuring emissions of CO2 in large cities is crucial for tracking emission reduction goals at city scales. Inventory-based urban emission estimates provide city total emission with relatively high accuracy, however they rarely correspond to all fossil fuel emission occurred within city administrative boundaries. The recently emerged Open-source Data Inventory for Anthropogenic CO2 (ODIAC), a global gridded carbon emission dataset at 1 km spatial resolution sharing total fossil fuel emission with the Global Carbon Project, also allows for timely and quick emission estimations at city scales, but its accuracy has not been assessed at such scale. This study aims to first derive a reasonably good estimate of total fossil fuel CO2 emissions directly occurred within 14 of world's major cities' boundaries, based on reported sector-specific inventory-based city CO2 emission estimates. These estimates are then compared with ODIAC which distributes national level emission inventories spatially, primarily based on night lights satellite data. Despite of their distinct approaches in urban emission estimates, the two methods show good consistency (within 30% difference in total emission) in half of the cities examined, with very close numbers (<10% difference) in Bangkok, Shanghai, Paris, and Los Angeles. The ODIAC dataset exhibits a much higher emission compared to inventory-based estimates in Cape Town, Delphi and Beijing. On the other hand, ODIAC shows lower estimates in Manhattan, New York City , Washington DC and Toronto, all located in North America. The findings suggest that the ODIAC global emisssion data could serve as a reasonably good prior fossil fuel CO2 field that is valuable to the carbon satellite and other science communities.