Trees are major players in the cycling of carbon (C) between the atmosphere and terrestrial biosphere, representing c. 90% of the world's biomass C and half of global primary productivity (Körner, 2003; Beer et al., 2010). Trees take up C through leaf pores (i.e. stomata) during photosynthesis and convert this C into carbohydrates. These carbohydrates can be stored to support the maintenance of existing tissues and future growth, or be immediately transported through the tree to form new leaves, woody tissues, and roots. Within-tree C cycling processes are poorly understood and are not represented well in current ecosystem C cycling models, resulting in large uncertainties in the response of tree growth and whole ecosystem C cycling to future global change (Le Roux et al., 2001; Körner, 2003; Millard et al., 2007). In this issue of New Phytologist, Richardson et al. (pp. 850–861) make substantial progress in understanding tree growth and within-tree C cycling by constraining simple hypothesis-driven models with intensive field measurements of C fluxes, storage pool ages, and stocks. Results indicated that the average age of carbohydrates was a decade and that modeled year-to-year variations in tree growth substantially improved with the incorporation of carbohydrate pools and reserves. These results have implications for improving tree growth and ecosystem C cycling models, and understanding whether man-made increases in atmospheric CO2 will stimulate tree growth.
‘… the big within-tree C cycling questions are: how long are NSC pools physiologically active, and how important are they for growth and metabolism?’
Nonstructural carbohydrates (NSCs) are the primary class of carbohydrates used for storage (Chapin et al., 1990; Kozlowski, 1992). Storage of NSCs can be passive, resulting from the accumulation of photosynthetic C supply when C demand for growth and maintenance is low, or active, resulting in both accumulation and the formation of a NSC reserve that occurs at the expense of growth. Determining which process dominates has implications for whether trees are C limited under current atmospheric CO2 concentrations. Proponents of the passive model argue that NSCs cannot accumulate when growth is C limited (Körner, 2003; Millard et al., 2007), while proponents of the active model argue that NSC reserve formation can occur at the expense of growth and has evolutionary advantages by acting as an emergency reserve during times of stress (Wiley & Helliker, 2012). However, all carbohydrate pools may not be available for growth (Chapin et al., 1990), and reserve formation may just be a mechanism to avoid down-regulation of photosynthesis when C demand is low (Millard et al., 2007). Although views differ in the functional interpretation of NSC pools, the big within-tree C cycling questions are: how long are NSC pools physiologically active, and how important are they for growth and metabolism?
Field studies have demonstrated the importance of storage reserves for forest C cycling. Eddy covariance-derived gross C uptake (i.e. photosynthesis) and tree-growth measures are poorly related at interannual timescales, and suggest that a large carbohydrate pool is decoupling the two processes (Rocha et al., 2006). Indeed, tree NSC pools have been shown to be large enough to completely refoliate a tree canopy (Würth et al., 2005), while isotopic labeling studies have demonstrated that remobilization of late season carbohydrate storage in deciduous trees provides a substantial amount of C for new leaf growth during the following spring (Kagawa et al., 2006). Studies have also shown that decade old C is metabolically available in belowground organs (Vargas et al., 2009). However, disentangling the relative importance of newer vs older C for tree growth has been difficult because assimilation, storage, and accumulation can occur simultaneously and because there are large uncertainties associated with the degree of isotopic mixing between current photosynthate, old NSC pools, and plant tissues (Chapin et al., 1990; Keel et al., 2007).
Richardson et al. provided new insights into within-tree C cycling processes by directly determining the mean age of the measured NSC pool, which helped to constrain NSC pool and tree growth dynamics in later model hypothesis testing. The mean age of the NSC pool was determined by the carbon-14 (14C) content of sugars and starches extracted from the outer 2 cm of stemwood. Atmospheric 14C has declined, as a result of biospheric uptake and dilution by fossil fuel emissions, since the doubling of 14C during atmospheric nuclear weapons testing in the 1960s. This large pulse of atmospheric 14C created a large-scale pulse and chase experiment in which to track the fate of C through the biosphere (Trumbore, 2006). Determining the 14C content of sugars and starch was not a trivial task because previous methods for NSC extraction required the use of methanol, which would contaminate the 14C concentration in the sample. To avoid this issue, Richardson et al. developed 14C extraction techniques that used water for soluble sugar compounds, and ethanol and repeated rinsing steps for the insoluble starch compounds. The authors clearly demonstrate that the results from these new methods were not influenced by methodological contamination, agreed well with previously established techniques, and had an ability to resolve NSC ages within 2 yr.
Once the mean age of the NSC pool was determined, hypothesis-driven modeling exercises were conducted to determine which model best matched the observational constraints, including year-to-year dynamics in woody biomass increment. Current ecosystem and canopy level models poorly capture year-to-year variations in woody biomass increment, and either ignore or crudely incorporate NSC dynamics in their framework (Le Roux et al., 2001; Misson, 2004). Richardson et al. employed three simple box models to determine which processes were required to capture year-to-year woody biomass increment observations. Models were parameterized with actual measurements of canopy, leaf, and soil CO2 fluxes, soil C, litterfall, leaf area index, and the age and size of the NSC pool using Monte Carlo model–data fusion. The first model had no storage pool and a fixed proportion of photosynthate was immediately used for growth; a representation that is dominant in current ecosystem C cycling models. The second model incorporated one storage reserve and accounted for maintenance and growth respiration. The third model incorporated two interacting storage pools with different turnover times (i.e. a ‘fast’ and ‘slow’ storage pool), where the fast C pool fueled growth and metabolism and the slow pool acted as a storage reserve that filled the fast pool when it was C deficient. Richardson et al. concluded that the incorporation of both a fast and slow NSC storage pool performed best in capturing the size and age of the NSC pool and in predicting woody biomass increment measurements.
Richardson et al. undoubtedly provides novel and important insights into tree growth physiology, but also leaves several unresolved questions for future work. The actual age and size of the NSC pool was likely underestimated in the study because NSC in roots, leaves, and deep in the heartwood were not sampled. The actual mobility of the total NSC pools is also unknown, but anecdotal evidence suggests that recently assimilated C is used before old C reserves are drawn upon. In Richardson et al., starch and sugars were of similar age, which was unexpected as starches are considered an immobile long-term storage of NSC (Chapin et al., 1990; Kozlowski, 1992). Starch accumulates in parenchyma cells whenever high levels of sugars build up, and can be remobilized and transported outward toward the cambium when sugar concentrations are low (Kozlowski, 1992). The age similarity between sugars and starches indicates regular remobilization and recycling of starches within the tree for growth, and supports the view that trees are not C limited under current atmospheric CO2 concentrations.
The implications of the study are far reaching and will result in new questions in tree physiology and evolution, 14C studies, ecosystem ecology, and ecosystem C cycle models. With regards to tree physiology, the study provides a new methodology to understand carbohydrate pools and improve our understanding of C starvation and limitation in an environmentally changing world. These techniques can also be used to improve our understanding of the evolutionary costs and benefits of C storage and reserve formation. The study also complicates the interpretation of previous 14C studies that have assumed that the 14C concentrations in plant or mychorrizal tissues reflect current photosynthate (Trumbore, 2006). Future studies using 14C will have to take into account the relative functional roles of new and stored NSC, intermixing between these old and young pools, and how they are allocated for growth and metabolism within the plant. Lastly, this study demonstrates that incorporating C storage pools in ecosystem models is critical in predicting woody growth increment. It remains to be determined if incorporating storage and reserve pools will better represent tree response to extreme climatic events or mortality. Understanding these fundamental processes in tree physiology will improve C cycle models and help constrain the large uncertainties surrounding the response of the terrestrial biota to future global change (Friedlingstein et al., 2006).