Time-Lagged Climate and Vegetation Feedbacks Intensify Seasonal CO2 Exchange in High Latitudes

The amplitude of the atmospheric CO₂ seasonal cycle has risen sharply since the 1960s, especially across northern high latitudes, yet the mechanisms driving this amplification remain unresolved. Using a time-lag detection framework applied to four decades of CO₂ records, satellite observa-tions, and dynamic vegetation model simulations, we show that the intensifying seasonal cycle arises from a cascade of immediate and lagged ecosystem responses to climatic variability. Two transitional months—May and October, marking the onset and end of the growing season—dominate interannual variability and long-term trends in CO2 amplitude over northern ecosystems. October respiration and productivity exert the strongest influence via a two-year legacy pathway in which warm, productive autumns dry soils, delay refreezing, alter vegetation structure and ecosys-tem carbon storage, thereby conditioning the sensitivity of the subsequent spring carbon fluxes. Earlier snowmelt and higher May temperatures then translate the stored memory into enhanced ecosystem-atmosphere carbon exchange. Together, our findings reveal a coupled hydroclimate-cryosphere-vegetation feedback linking autumn legacies to spring responses, embedding multi-year memory into northern carbon–climate dynamics.