- Biotic and abiotic controls on whole-ecosystem carbon balance in forest ecosystems
- Dynamics and age of nonstructural carbohydrate reserves in forest trees;
- Controls on spatial and temporal variability in vegetation phenology
- Model-data fusion and multiple constraints
Dynamics and age of nonstructural carbohydrate reserves in forest trees
Like all plants, forest trees accumulate and store nonstructural carbohydrates (NSC) as resources to be used to support future growth. This can be viewed as a conservative investment strategy, providing reserves that the tree can utilize in times of stress. NSC reserves are thus important in the context of forest responses to climate change. Understanding how big these reserves are, and how available they are to support growth and metabolism, is essential for predicting the resilience of forests to environmental and biotic stress factors. However, processes related to within-tree carbon allocation remain poorly understood.
In a recent long-term study, we monitored stemwood NSC reserves of the dominant tree species at three sites along a forest composition gradient in New England: an oak-dominated transition hardwood forest (Harvard Forest), a maple-beech-birch northern hardwood forest (Bartlett Experimental Forest), and a spruce-fir boreal transition forest (Howland Forest). Adding to recent evidence indicating that in mature forest trees the nonstructural carbohydrate pool is surprisingly large, these results further suggest that this pool is also highly dynamic on seasonal time scales.
We have used radiocarbon methods—the “14C bomb spike” approach (see figure, below)—to estimate the mean age of the NSC pool. This approach takes advantage of the 14C label that was produced during the period of atmospheric thermonuclear weapons testing, which doubled the 14C content of CO2 in the northern hemisphere atmosphere by the early 1960s. Since then, the 14C content of atmospheric CO2 has decreased. This can be attributed to dilution through mixing with oceanic and terrestrial biosphere C reservoirs, and by addition of 14C-free CO2 from fossil fuel burning. Because the C in photosynthate (newly made NSCs) reflects the 14C content of the atmosphere in the year assimilation occurred, the mean age of an NSC pool can be determined by measuring its 14C content, and comparing this to the atmospheric record. This age is then a proxy for the mean residence time of C in that pool. Our data indicate that the age of the starch and sugar pools in red maple (Acer rubrum) is ≈ 7-14 y. Together, the concentration and age data provide the information necessary to constrain model estimates of the size of the NSC pool, and how quickly this pool turns over.
Using model-data fusion techniques, we found that a one-pool representation of NSC reserves (similar to the standard approach used in most existing forest C models) fails to reproduce the observed seasonal dynamics, and also predicts a much shorter mean residence time (< 2 y). This suggests that there are probably both fast (most dynamic) and slow (less dynamic) cycling components to the NSC pool that must be modeled separately. We have developed a novel two-pool approach to modeling NSC dynamics that is optimally consistent with the observations (NSC data constraints) and also improves model predictions for other ecosystem attributes (e.g., woody biomass increment). This work (J111) is unique in the way in which it combines measurements of NSC concentrations, NSC pool ages, and an ecosystem model constrained with carbon flux and stock measurements, to study carbon allocation processes and reserve pool dynamics. We are expanding on this work by trying to assess how available the stored reserves are to support tree metabolism, how availability varies seasonally (i.e. dormant vs. growing season), and how much mixing there is of stored reserves across rings, and between stored sugars and starches.
We use radiocarbon methods—the “14C bomb spike”—to estimate the age of stored carbohydrate reserves. The inset plot shows the 14C content of atmospheric CO2 from 1950 through 2010, while the main plot shows a model-based comparison between 14C age in 2010 (y axis) and mean residence time (MRT, x axis) of simulated well-mixed carbohydrate pools.