Forest disturbances and tree growth are major drivers of dynamics and long-term carbon storage in forests. Combining forest inventory data and tree-ring analysis allows for a retrospective estimate of aboveground forest biomass (AGB) and disturbance dynamics at decadal to centennial scales, both being directly related to net primary productivity (NPP). We used this combined approach to estimate precisely-dated ages of aboveground carbon in over 20 temperate old-growth forests in Western Europe and Eastern North America, where trees exceed 400 years of age. We also estimated carbon turnover times as the ratio of stand AGB to AGB annual increment. In some forests, the combination of no logging and natural disturbance regimes dominated by small, frequent disturbances over centuries resulted in high forest stability, long carbon turnover times, and high mean carbon ages. Carbon turnover times ranged between 140 and 250 years whereas mean carbon ages ranged between 100 and 170 years. Up to 20% of total AGB was stored for 300 years or longer while more than 50% of total AGB was stored for more than 100 years. In contrast, more recently disturbed forests or forests with dynamics dominated by more intense disturbances appear to have reduced ages of mean carbon (<100 years) and turnover rates (<125 years). In general, annual forest aboveground productivity (i.e. faster growth rates) was inversely correlated with mean AGB age and thus long-term carbon storage potential. The large amounts of biomass stored at long time scales in uncut old-growth forests highlight the importance of forest management, ecosystem stability, and natural disturbance as a part of the global carbon cycle.