The study investigates how long it takes for human-driven trends in the land carbon sink to become distinguishable from natural year-to-year variability, using large ensembles of Earth system model simulations. The authors find that anthropogenic signals emerge relatively quickly for gross carbon fluxes but more slowly for the net land carbon sink—particularly at regional scales—and that future mitigation scenarios delay detection by weakening long-term trends. The study contributes to enhancing our ability to identify human impacts on the land carbon sink and evaluate the effectiveness of climate policies.