The covariation of forest age shifts and net carbon balance over the period 2010 to 2020

Understanding the relationship between forest age, an indicator of successional stages, and net carbon fluxes is crucial for effective forest management and climate mitigation. Using the satellite-based Global Age Mapping Integration (GAMI) v1.0 dataset, we analyzed forest age shifts from 2010 to 2020 and their correlation with net carbon dioxide (CO₂) flux changes from independent atmospheric inversions. Globally, we do not report substantial forest age shifts during this period. The total area of young (1-20 years old), intermediate (21-60 years old), mature (61-150 years old), and old-growth (>150 years old) forests in 2020 compared to 2010 changed by approximately -0.07 (-7.7% compared to 2010), +0.03 (+6.0%), +0.03 (+2.1%), +0.01 (+1.1%) billion hectares, respectively. Despite these relatively stable global trends in forest age classes, we observe substantial changes at the regional scales. The Amazon, Congo basin, and Southeast Asia regionally experienced significant forest age decreases with local changes of up to 30% compared to 2010, attributed to deforestation and degradation. Siberian forests maintained their older age structure; however, large areas are transitioning to younger ages (0.09 billion hectares, 7.2% of Eurasia Boreal region), likely driven by increased fire frequency, logging activities, or climate-induced changes. Most European and North American forests trended toward older ages. However, those changes were heterogeneous at the sub-pixel level, revealing a complex mix of stand-replacement and aging dynamics across the different forest age spectrums. Stand-replaced forests, followed by regrowth, constitute a relatively minor fraction (6%) of the overall forested ecosystems, primarily dominated by aging forests (64%) and "stable" old-growth tropical forests (30%). Stand-replaced forests were prominent in young forests (0.1 billion hectares, 54.3% of total stand-replaced forests), while intermediate, mature, and old-growth forests accounted for 13.2%, 17.9%, and 14.6% of the total area of stand-replaced forests. Conversely, aging forests (excluding old-growth "stable" tropical forests) were primarily observed in the mature age classes, encompassing 1.2 billion hectares and constituting around 53% of the total aging forests. When coupling GAMI data with CO₂ flux estimates, we observe a significant correlation between the spatial patterns of the stand-replaced forest fraction and net CO₂ flux changes (R² = 0.37, slope = 118.7 gC m^-2 year^-1 [a positive slope indicates increased carbon released], p-val = 0.05) across the eleven TRANSCOM-land regions. This correlation surpasses the correlation with aging forests (R² = 0.02, slope = -3.7 gC m^-2 year^-1, p-val = 0.69). We attribute this significant correlation to the net above-ground biomass (AGB) losses in stand-replaced forests per unit area, substantially exceeding the magnitude of the net AGB gains observed in aging and old-growth "stable" tropical forests throughout 2010-2020. Our study highlights the importance of rapid forest turnover through stand-replacement, despite its limited spatial extent, on regional net carbon balance, especially when contrasted with the more gradual process of forest maturation.