Greenhouse gas emissions and their trends over the last three decades across Africa

A key goal of the Paris Agreement (PA) is to reach net-zero greenhouse gas (GHG) emissions by 2050 globally, which requires mitigation efforts from all countries. Africas rapidly growing population and gross domestic product (GDP) make this continent important for GHG emission trends. In this project we study the emissions of carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) in Africa over 3 decades. We compare bottom-up (BU) approaches, including United Nations Convention Framework on Climate Change (UNFCCC) national inventories, FAO, PRIMAP-hist, process-based ecosystem models for CO2 fluxes in the land use, land use change and forestry (LULUCF) sector and global atmospheric inversions. For inversions, we applied different methods to separate anthropogenic CH4 emissions. The BU inventories show that, over the decade 20102018, fewer than 10 countries represented more than 75 % of African fossil CO2 emissions. With a mean of 1373 Mt CO2 yr1, total African fossil CO2 emissions over 20102018 represent only 4 % of global fossil emissions. However, these emissions grew by +34% from 19901999 to 20002009 and by +31% from 20002009 to 20102018, which represents more than a doubling in 30 years. This growth rate is more than 2 times faster than the global growth rate of fossil CO2 emissions. The anthropogenic emissions of CH4 grew by 5 % from 19901999 to 20002009 and by 14.8 % from 20002009 to 20102018. The N2O emissions grew by 19.5 % from 19901999 to 20002009 and by 20.8 % from 20002009 to 20102018. When using the mean of the estimates from UNFCCC reports (including the land use sector) with corrections from outliers, Africa was a mean source of greenhouse gases of 2622 (min: 2186, max: 3239) Mt CO2 eq. yr1 from all BU estimates (the minmax indicate range uncertainties) and of +2637 (min: 1761, max: 5873)Mt CO2 eq. yr1 from top-down (TD) methods during their overlap period from 2001 to 2017. Although the mean values are consistent, the range of TD estimates is larger than the one of the BU estimates, indicating that sparse atmospheric observations and transport model errors do not allow us to use inversions to reduce the uncertainty in BU estimates. The main source of uncertainty comes from CO2 fluxes in the LULUCF sector, for which the spread across inversions is larger than 50 %, especially in central Africa. Moreover, estimates from national UNFCCC communications differ widely depending on whether the large sinks in a few countries are corrected to more plausible values using more recent national sources following the methodology of Grassi et al. (2022). The medians of CH4 emissions from inversions based on satellite retrievals and surface station networks are consistent with each other within 2 % at the continental scale. The inversion ensemble also provides consistent estimates of anthropogenic CH4 emissions with BU inventories such as PRIMAP-hist. For N2O, inversions systematically show higher emissions than inventories, either because natural N2O sources cannot be separated accurately from anthropogenic ones in inversions or because BU estimates ignore indirect emissions and underestimate emission factors.