Land vegetation currently acts as a significant carbon sink, absorbing substantial amounts of atmospheric CO2. This carbon sequestration is attributed to several factors, including afforestation, secondary forest expansion, and changes in forest dynamics influenced by nitrogen deposition, elevated CO2 levels, and rising temperatures. Increased CO2 and temperature, particularly in colder regions, are thought to stimulate tree growth, creating an imbalance between growth and mortality rates, resulting in net carbon uptake even in primary forests. This carbon sink is crucial in mitigating the effects of fossil fuel combustion and cement production. Earth System Models (ESMs) predict this forest carbon sink will persist well into the future. However, the longevity of this carbon sink depends not only on growth responses to environmental changes but also on mortality rates, which release carbon back into the atmosphere. Increased growth, regardless of its cause, must eventually result in increased mortality, creating a negative feedback on carbon storage. Our incomplete understanding of this feedback mechanism presents a significant uncertainty in ESM predictions of future forest carbon uptake. Existing forest plot monitoring data reveal widespread increases in mortality, potentially linked to increased growth. However, establishing a direct link between these trends using monitoring data alone is challenging due to the long lifespans of trees. Tree-ring analysis offers a practical method for assessing the lifespan response to growth. Previous tree-ring studies have demonstrated a trade-off between growth and lifespan, with faster-growing species generally having shorter lifespans. This trade-off is explained by the allocation of resources: fast growth prioritizes resource allocation to growth at the expense of survival mechanisms. However, the universality and underlying mechanisms of this trade-off, particularly within species, remain insufficiently understood, hindering accurate incorporation into ESMs.

Previous research has extensively explored the relationship between tree growth and lifespan. Studies across species have consistently shown a trade-off, with fast-growing pioneer species exhibiting shorter lifespans compared to slow-growing, shade-tolerant species. This trade-off is a well-established axis of plant strategies. More recent studies have investigated this trade-off within species, demonstrating that faster-growing individuals often have shorter lifespans. However, these studies often focused on specific species (mostly conifers at high elevations) or limited geographic regions, leaving uncertainty regarding the universality of this phenomenon. Some studies even reported a lack of such trade-offs, emphasizing the need for a broader investigation across diverse species and biomes to establish the pervasiveness and underlying mechanisms driving this relationship. The current incomplete knowledge about the trade-off between growth and lifespan hinders its accurate representation in Earth System Models (ESMs), which are crucial tools for predicting future forest carbon dynamics under global change scenarios. Existing ESMs often overestimate the forest carbon sink's persistence, highlighting the urgent need to improve the understanding and modeling of this crucial feedback mechanism.

This study uses a comprehensive dataset of tree-ring records from over 210,000 trees of 110 species globally, spanning diverse climates and habitats from the tropics to the Arctic. The primary data sources include the International Tree-Ring Data Bank (ITRDB) and the National Forestry Inventory data from Quebec, Canada. Data quality control measures were implemented to address potential issues such as duplicate records and pith offsets. Species selection criteria ensured sufficient sample sizes for robust analysis. The analysis focused on assessing the relationship between early growth (mean ring width over the first 10 years) and lifespan. The strength of the trade-offs was evaluated using 95th quantile regression. To assess potential artifacts influencing the observed relationship, the study considered several factors: the use of living trees to estimate lifespans, the influence of recent growth increases, effects of pith offsets and wood decay, and sampling biases favoring larger trees. Various statistical tests and comparisons were used to address these potential biases. To further investigate the environmental drivers of the trade-off, the study examined the correlations between lifespan and environmental variables such as temperature and precipitation for nine species. Mixed effect models were employed to analyze the independent effects of growth rate and temperature on lifespan. Data-driven forest simulations, using tree-ring data from *Picea mariana* in Quebec, were performed to examine the effect of the observed growth-lifespan trade-off on forest dynamics. These simulations incorporated size-related mortality rates, which realistically reflect the trade-off. The simulations were compared to scenarios without the trade-off (age-dependent mortality) to evaluate its impact on biomass and mortality changes over time. The growth stimulation was modeled based on observed temperature-driven growth increases in boreal forests. The simulations considered various factors like competition effects and the stability of size-related mortality curves. The results were compared with ESM predictions and existing literature on forest dynamics.

The study reveals a remarkably consistent negative relationship between early growth rate and maximum lifespan across nearly all species and climates examined. Faster early growth is directly linked to shorter lifespans. This relationship is not primarily driven by environmental factors like temperature, as the trade-off persists even after controlling for environmental variables. The average lifespan reduction was approximately 23% for a 50% increase in early growth. The analysis confirms that the observed trade-offs are not merely artifacts of sampling biases or the use of living trees. Simulations using data from *Picea mariana* demonstrate that growth stimulation, mimicking temperature-driven growth increases, leads to initial increases in biomass but eventually results in mortality increases that negate these gains. The increase in biomass is transient; eventually the biomass returns to baseline levels due to faster growth reducing tree lifespan. Simulations without the growth-lifespan trade-off, using age-dependent mortality instead, show sustained biomass increases after the growth stimulation period. This suggests that the growth-lifespan trade-off is a crucial factor influencing forest dynamics and carbon storage. The findings align with observations of simultaneous increases in growth and mortality rates in various forest biomes worldwide. Observed increases in mortality might be explained by the pervasive growth-lifespan trade-offs, alongside the effects of climate variability.

The findings of this study directly address the research question of the universality and impact of growth-lifespan trade-offs in trees. The near-universal nature of this trade-off across various species and climates strongly suggests that it's a fundamental biological constraint affecting forest carbon dynamics. The significance of these results lies in their implications for predicting future forest carbon sequestration. Current ESMs tend to overestimate the long-term capacity of forests to absorb atmospheric CO2. The results highlight the critical need to incorporate the growth-lifespan trade-off into ESMs to improve the accuracy of future projections. The study's findings have broad implications for forest management and carbon cycle modeling. Understanding this feedback mechanism is essential for developing effective strategies for mitigating climate change and managing forest resources sustainably.

This research provides robust evidence for the widespread existence of a growth-lifespan trade-off in trees, significantly impacting forest carbon dynamics. Faster growth, regardless of the underlying drivers, directly reduces lifespan, leading to increased mortality and negating the initial carbon sequestration benefits of growth stimulation. Current ESMs should be revised to incorporate this crucial feedback mechanism for accurate predictions of forest carbon sinks. Future research could focus on exploring the mechanistic basis of the trade-off and investigating the interactive effects of CO2, temperature, and other environmental factors on this relationship across a wider range of species and ecosystems.

The study's simulations rely on simplified assumptions, such as neglecting competition effects and changes in tree recruitment. The size-related mortality curves are assumed to be independent of environmental changes, which may not entirely hold true in reality. Species distribution shifts due to climate change are also not explicitly considered in the simulations. Despite these simplifications, the simulation results align with findings from more complex models, suggesting that the overall conclusions are robust. Further research incorporating these complexities would strengthen the findings and improve predictive capabilities.