The land and oceans have mitigated climate change by taking up about half of the anthropogenic CO2 emitted since the industrial revolution. However, these ‘sinks’ are predicted to lose their efficiency. Globally, the combined sink strength of the land and ocean can be calculated indirectly as the difference between anthropogenic emissions – from fossil fuel burning and deforestation – and the atmospheric CO2 increase. However, large uncertainty in the deforestation term masks out potential changes in sink strength contained in the better-constrained fossil fuel and atmospheric terms. This creates the need for a new accurate approach to quantify emissions from deforestation and its variability over the past decades.I propose to quantify deforestation emissions from the novel fire perspective. A substantial share of deforestation emissions stems from burning vegetation, and this focus enables validation of emissions by comparing atmospheric enhancements of fire-emitted carbon monoxide (CO) with satellite-derived concentrations of CO. The proposed multidisciplinary work will follow three steps: 1) quantify net emissions from fires and decomposition in deforestation and degradation regions, combining satellite data with biogeochemical modelling, 2) validate these emissions by combining newly measured CO:CO2 ratios and the isotopic signature of CO2 downwind of deforestation regions, atmospheric chemistry transport modelling, and satellite-derived CO concentrations, and 3) use relations between fire emissions and visibility reported at airports as a novel way to extend the new deforestation emissions estimates back in time before high-quality satellite observations were available. The new approach will lead to the first constrained, monthly resolved estimate of deforestation emissions. Applying the global CO2 mass balance equation will then provide a better quantitative understanding of the (changing) sink capacity of the Earth's oceans and land surface.