Reviewing differences and uncertainties in land-use CO2 flux estimates

CO2 fluxes from land use and land-use change (FLUC) are a major source of carbon to the atmosphere. They are composed of gross emissions, mainly from deforestation, peat burning, and peat drainage, and gross removals, mainly from re- and afforestation. The importance of FLUC for climate change mitigation strategies is increasing due to the potential of storing large carbon amounts via re- and afforestation, harvested wood products, and other vegetation-based carbon dioxide removal methods, such as bioenergy with carbon capture and storage. Yet, FLUC estimates remain largely uncertain and show substantial discrepancies between different quantification methods, which makes it challenging to provide reliable projections of their potential future evolution.   Here, we review the main characteristics, uncertainties, and discrepancies of individual methods used to estimate FLUC, and we highlight promising steps to reduce FLUC uncertainties and to harmonize the various FLUC estimates. Differences between the approaches are mainly due to differing definitions and assumptions, such as the definition of anthropogenic fluxes and managed land (leading to a gap in FLUC of ~1.8 GtC/yr in 2000-2020 between FLUC estimates by bookkeeping models used in the Global Carbon Project and inventory-based estimates reported by countries to the United Nations Framework Convention on Climate Change) and the inclusion of environmental effects on carbon stocks (leading to a gap of ~0.4 GtC/yr in 2000-2020 between FLUC estimates from dynamic global vegetation models and bookkeeping models). Furthermore, the individual estimation methods have large uncertainties, mainly arising from the usage of differing land-use forcing data, missing observational constraints, differences in how models implement individual processes, and the degree of implementation of land use practices in models.   To improve the confidence in the individual FLUC estimates, we argue for a systematic model evaluation and an improved parametrization of models, in particular regarding land-use forcing data, carbon densities of vegetation and soils, and the represented processes. Alongside, remaining framework inconsistencies, such as a precise and consistent definition of FLUC and the consideration of transient C densities need to be resolved. This undertaking requires developments in several directions. Earth observations may provide data on carbon densities in vegetation and soil at high spatial resolution, improved estimates of forest regrowth rates as well as impacts of forest management. Models need to be further improved to consider all relevant land-use processes and provide more fine-granular output to guarantee that the different estimates are comparable and/or translatable into each other.   Providing harmonized and more accurate FLUC estimates is essential to improve the stocktake of countries' land use-related CO2 emissions, to provide an accurate budget of the global carbon cycle, and to effectively plan and monitor land-based carbon dioxide removal methods.