摘要：Since the first estimate of global CO2 emissions waspublished in 1894, important progress has been made in the development ofestimation methods while the number of available datasets has grown. Theexistence of parallel efforts should lead to improved accuracy andunderstanding of emissions estimates, but there remains significantdeviation between estimates and relatively poor understanding of the reasonsfor this. Here I describe the most important global emissions datasetsavailable today and – by way of global, large-emitter, and case examples – quantitatively compare their estimates, exploring the reasons fordifferences. In many cases differences in emissions come down to differencesin system boundaries: which emissions sources are included and which areomitted. With minimal work in harmonising these system boundaries acrossdatasets, the range of estimates of global emissions drops to 5%, andfurther work on harmonisation would likely result in an even lower range,without changing the data. Some potential errors were found, and somediscrepancies remain unexplained, but it is shown to be inappropriate toconclude that uncertainty in emissions is high simply because estimatesexhibit a wide range. While “true” emissions cannot be known, by comparingdifferent datasets methodically, differences that result from systemboundaries and allocation approaches can be highlighted and set aside toenable identification of true differences, and potential errors. This mustbe an important way forward in improving global datasets of CO2emissions. Data used to generate Figs. 3–18 are available athttps://doi.org/10.5281/zenodo.3687042 (Andrew, 2020).
