In a globalised world, the transfer of carbon between regions, either physically or em- bodied in production, represents a substantial fraction of global carbon emissions. The resulting emission transfers are important for balancing regional carbon budgets and 5 for understanding the drivers of regional emissions. In this paper we synthesise cur- rent understanding in two parts: (1) embodied CO2 emissions from the production of goods and services produced in one country but consumed in others, (2) physical car- bon flows in fossil fuels, petroleum-derived products, harvested wood products, crops, and livestock. We describe the key differences between studies and provide a consis-10 tent set of estimates using the same definitions, modelling framework, and consistent data. We find the largest trade flows of carbon in international trade in 2004 were fossil fuels (2673 MtC, 37 % of global emissions), CO2 embodied in traded goods and ser- vices (1661 MtC, 22 % of global emissions), livestock (651 MtC, 20 % of total livestock carbon), crops (522 MtC, 31 % of total harvested crop carbon), petroleum-based prod-15ucts (183 MtC, 50 % of their total production), and harvested wood products (149 MtC, 40 % of total roundwood extraction). We find that for embodied CO2 emissions esti- mates from independent studies are robust. We found that differences between indi- vidual studies is not representative of the uncertainty in consumption-based estimates as different studies use different production-based emission estimates as input and dif-
20ferent definitions of allocating emissions to international trade. After adjusting for these issues, results across independent studies converge to give less uncertainty than pre- viously assumed. For physical carbon flows there are relatively few studies to be syn- thesised, but differences between existing studies are due to the method of allocating to international trade with some studies using “apparent consumption” as opposed to 25 “final consumption” in more comprehensive approaches. While results across studies are robust to be used in further applications, more research is needed to understand the differences between methods and to harmonise definitions for particular applica- tions.