The Problem of Tree Senescence in the Role of Elevated CO2 and the Carbon Cycle

Climate change affects the terrestrial carbon cycle through many pathways. In particular, CO₂ fertilization can shape tree growth, death, and tolerance or resilience to climate change (Walker et al., 2021). Titrating the role of the different influences of elevated atmospheric CO₂ (eCO₂) on the carbon cycle through forest ecology has encouraged large experiments (Free-Air CO₂ Enrichment [FACE] experiments), analyses of dendrochronological and inventory data sets (Brienen et al., 2020), and vegetation model simulations to identify and quantify potential effects at different scales (Needham et al., 2020). In this issue, Marquès et al. (2023) tackle a potentially critical yet empirically challenging indirect consequence of eCO₂ fertilization: although faster growth in response to eCO₂ might accelerate carbon fixation, this acceleration could be offset by an increase in large tree mortality. Termed the Grow Fast-Die Young hypothesis (GFDY), this reflects a role of tree size thresholds associated with increased mortality, or “size-driven senescence” in shaping the population-wide consequences of accelerated tree growth.

GFDY requires that size, and not age, is the primary determinant of the mortality of mature trees. Were senescence (late-life increase in mortality) due solely to age, growing fast would be decoupled from dying young, and any increase in life-time productivity would lead to a direct increase in forest biomass. This would offset atmospheric CO₂ concentration increases as the fertilization effect of eCO₂ would lead to a larger terrestrial carbon sink. Although certainly age plays a role in tree mortality (e.g., the advance of pathogens, damage accrued over life) the types of processes that lead to animal death due to age (telomere shortening, mutation accumulation, etc.) appear not to be prevalent in the meristems or distal tissues and organs of even very old trees (Klimešová et al., 2015; Mencuccini et al., 2005; Thomas, 2013); but see (Cannon et al., 2022). Size-based tree senescence supports the premise that the GFDY would subtract from any potential sink gained from eCO₂ growth stimulation. This raises critical questions: how quickly and how universally might the GFDY operate?

Any potential GFDY responses will be entangled with other trends in climate change, and complex physiological and ecological tradeoffs from the scale of the cell to the community. This complicates distinguishing the signal from noise. Marquès et al. (2023), recognizing that multiple mechanisms contribute to stand-level growth, conceived an elegant way of testing hypotheses about the eCO₂ effect, acknowledging contingencies of stand dynamics, and identifying key sensitivities that determine stand responses to eCO₂. They use forest inventory data and simulations to demonstrate potential and observed changes in the stand thinning line (STL, Marques et al., 2023 Figure 1a). STL is a line fit to the upper boundary of the relationship between stand-level biomass and stem number--a relationship that, within a stand, describes the distribution of biomass distributed across stems (Figure 1a).

A strength of using the STL to identify GFDY effects is that it emerges from a simple bivariate distribution easily acquired from inventory data that captures whole-forest dynamics. Essentially, an increase in the intercept of the STL indicates an increase in carbon carrying capacity in the stand across size-classes (i.e., if the STL shifts up, all stems are larger). By applying this analysis only to data from mature, undisturbed forests of similar composition over time, Marquès et al. were able to account for many of the confounding factors in GFDY analyses. A similar pattern was also found in McMahon et al., 2010 (Figure 1b). Both observational data analyses and simulations demonstrated that eCO₂ may not be completely offset by earlier mortality and that some forests have been gaining biomass even as mortality has increased due to the GFDY. They qualify some early work that found a complete offset of fertilization response (Brienen et al., 2020; Bugmann & Bigler, 2010), and support model analyses that also found incomplete offset (Needham et al., 2020). More importantly, the Marquès et al. finding advances the discussion as it indicates that the critical determinant of the strength of the GFDY (or its existence at all), and therefore the extent to which eCO₂ might lead to a change in the carbon sink, is through the shape of the curve relating size to mortality (Figure 1a). This advances the eCO₂ discussion from the physiology of growth to the relative importance of mechanisms of size-based senescence.

Research into the GFDY impacts on the carbon cycle requires we better understand the mechanisms of tree senescence, essentially asking: how do mature individuals die as a function of size? Three categories of mortality processes may help organize efforts to answer this question: (a) large size can leave an individual more vulnerable to exogenous mortality agents, such as wind damage or hydraulic failure, or the high ratio of maintenance costs to crown area (McDowell et al., 2018); (b) an interaction between genetic programs and exogenous factors may accelerate death at maturity through the reallocation of resources that maintain vigor to reproductive organs at the expense of maintenance (Thomas, 2013); and (c) an extension of 2, where genetic programs triggered at a size threshold lead to internal shifts in function that lead to death, such as is observed in monocarpic trees that die when flowering (Batalova & Krutovsky, 2023; Read et al., 2021; Thomas, 2013). We know these categories to be non-mutually exclusive and the processes to be complex.

How we better identify the ways in which these mechanisms shape the mortality curves of species globally requires continued collections of inventory data, experiments, and simulations. In order to properly scale observations, the use of remote sensing technologies such as lidar and hyperspectral imaging might improve observations of canopy stress and structural damage (a) from flowering patterns and canopy function (b). Gene expression data and better annotated genomes, especially of important tropical species which are poorly represented in gene databases, would also better identify internal mechanisms of stress response and senescence pathways (1, 2, and 3). The exploration of tree senescence marks not only a fascinating area of interdisciplinary research in understanding eCO₂ and climate change influences on tree performance, but also one critical to our better understanding and predicting the future of the Earth System.

The authors declare no conflicts of interest relevant to this study.