Harmonizing direct and indirect anthropogenic land carbon fluxes indicates a substantial missing sink in the global carbon budget since the early 20th century

Societal Impact Statement The global carbon budget provides annual updates to society on the main cause of climate change—CO 2 emissions—and quantifies carbon‐uptake ecosystem services provisioned by the biosphere. We show that more consistent assumptions in the estimates of land‐atmosphere carbon exchange results in a global carbon budget that is imbalanced (gains do not equal losses). This imbalance implies that key processes causing land carbon fluxes, especially processes associated with human land management and recovery following abandonment in anthropogenic biomes (anthromes), have been misquantified. This impacts policy for land carbon management across scales and calls for better understanding of carbon cycling in anthromes. Summary Inconsistencies in the calculation of the two anthropogenic land flux terms of the global carbon cycle are investigated. The two terms—the direct anthropogenic flux (caused by direct human disturbance in anthromes, currently a carbon source to the atmosphere) and the indirect anthropogenic flux (caused indirectly by human activities that lead to global change and affecting all biomes, currently an atmospheric carbon sink)—are typically calculated independently, resulting in inconsistent underlying assumptions. We harmonize the estimation of the two anthropogenic land flux terms by incorporating previous estimates of these inconsistencies. We recalculate the global carbon budget (GCB) and apply change‐point analysis to the cumulative budget imbalance. Cumulative over 1850–2018 (1959–2018), harmonization results in a 13% lesser (4% greater) land use source from anthromes and a 20% (23%) lesser land sink. This recalculation yields a greater non‐closure of the GCB, indicating a missing carbon sink averaging 0.65 Pg C year −1 since the early 20th century. The imbalance likely results from a combination of method discontinuity and structural errors in the assessment of the direct anthropogenic land use flux, greater ocean carbon uptake, structural errors in land models, and in how these land terms are quantified for the budget. We caution against overconfidence in considering the GCB a solved problem and recommend further study of methodological discontinuities in budget terms. We strongly recommend studies that quantify the direct and indirect anthropogenic land fluxes simultaneously to ensure consistency, with a deeper understanding of human disturbance and legacy effects in anthromes.