Overview..
The links between seasonal cycles of vegetation activity and atmospheric CO2 concentration demonstrate that the atmosphere and terrestrial vegetation are tightly coupled. These links suggest that there will be significant changes in the productivity of terrestrial vegetation with future changes in atmospheric CO2 concentration. Ecosystem simulation models predict the greatest changes are likely to occur in the forests of the northern hemisphere, primarily the boreal forests. Data from the global CO2 flask sampling network and recent work with isotope tracers in general circulation models also identify the higher latitudes of the northern hemisphere as the location of the largest net CO2 uptake. This work has prompted a large interdisciplinary experiment to assess the importance of boreal forests to the global carbon budget (BOREAS).
In recent years a number of approaches to quantify net photosynthetic CO2 exchange through time, or net primary production (NPP), have been developed. One of the most promising uses satellite remote sensing to monitor vegetation light capture and utilization for NPP (see pages describing such a model). Several variables related to NPP estimation, including leaf area index (LAI), vegetation type, and absorbed photosynthetically active radiation (APAR), have been related to remotely sensed observations of canopy reflectance, whereas others (e.g., biomass and foliage nitrogen) have proven to be more problematic. An important concept that underpins remote sensing of NPP is that natural selection has resulted in a narrow range of light use efficiency among plant functional types (i.e., plants with a related suite of life history traits) (see an abstract of a paper that evaluates this hypothesis).
Objectives and Approach..
The objective of the work reported here was to evaluate the sources of variability in the values for stand to regional scale NPP and the efficiency of light utilization for NPP (PAR utilization). We examined gross and net carbon fluxes using the Terrestrial Carbon Exchange (TCX) model, a mechanistic ecophysiological model, parameterized with data from 63 boreal forest stands in the Superior National Forest of northeastern Minnesota. The "TCX" model was selected for these analyses because it was designed for operation at the stand level, and because it was developed for and has been extensively tested in boreal forest ecosystems.
Components of the approach include evaluation of model simulations of net primary production (NPP) and carbon fluxes, light utilization in relation to respiratory carbon costs, respiratory costs in relation to biomass, and comparisons of the gross versus net carbon yield of absorbed light.
Net Primary Production and Carbon Fluxes
The sensitivity of NPP to model variables, including LAI, foliage N, and incident irradiance, had a strong effect on total tree NPP for both aspen and spruce. LAI had the largest effect on NPP but it elicited a different response in spruce and aspen stands. In spruce, the effect of LAI was curvilinear whereas the effect on aspen NPP was nearly linear. Spruce production was highly dependent on soil temperature, which took longer to increase to levels that permitted photosynthesis when canopy LAI was high. This effect, when combined with increased foliage respiration relative to photosynthesis, resulted in an asymptotic response of spruce NPP to LAI. In aspen stands the mineral soil warmed at nearly the same rate under low or high LAI, and both photosynthesis and respiration increased with LAI at a constant rate.
NPP was consistently lower in spruce than aspen at the same foliage N concentration. Foliage N increased NPP for both spruce and aspen at constant LAI, but was nonlinear. Spruce NPP varied more than aspen with foliage N, particularly at the lower N values (0.5 - 1.0%), however, the full range of values may be wider than those that occur in nature (see below).
Total tree NPP was weakly dependent on variables such as soil color and moderately dependent on variables such as sapwood biomass, soil type (e.g., texture, bulk density and water holding capacity), and litter and moss N content. Biomass of the litter and moss mostly affected heterotrophic respiration and had little effect on tree NPP. Moss biomass affected only moss production. Tree above : below-ground biomass ratio had a moderately strong effect on NPP because it modified respiration costs.
Moderate to strong positive relationships between simulated and observed NPP emerged with constant Nitrogen concentation (N) (explaining 80% of observed NPP in spruce, 92% in young aspen, and 38% in mature aspen). Allowing foliage N to vary for each stand gave values of foliage N between 0.4 - 1.2% (mean = 0.83%) for spruce stands and 0.7 - 2.8% (mean = 1.95%) for aspen stands.
Stand averaged fluxes from the fitted simulations are shown in Figure 1 for all stands of each species in a climatically typical year for the study area. Variability in assimilation and respiration on a daily basis was large, but followed a clear pattern through the growing season in both spruce and aspen stands..
PAR Utilization in Relation to Respiratory Costs
The relationship between simulated annual canopy PAR absorption (APAR) and NPP, that is, PAR utilization for net primary production averaged 0.33 g above-ground production per megajoule (MJ) APAR in spruce stands, whereas aspen stands averaged 0.91 g MJ-1. The range in both species was 0.15 - 0.52 g MJ-1 among spruce stands and 0.35 - 1.44 g MJ-1 among aspen stands.
APAR was a better predictor of spruce NPP (90% explained variance) than aspen NPP (66% explained variance), and the difference between spruce and aspen was significant. APAR was also a better predictor of NPP in young aspen stands (76%) than in mature stands (60%), although this distinction between age classes was not statisitcally significant. The four young aspen stands with low NPP were clearly distinct from all other aspen stands, including both young and mature.
Carbon respiration costs relative to assimilation gains, that is, the R:A ratio, are shown in the next figure. The generalized spruce simulations had a minimum R:A of ~0.6 and a maximum of ~0.9, so at least 60% and as much as 90% of the carbon assimilated annually by spruce stands was consumed in respiration. The range in the R:A ratio for the fitted spruce simulations was smaller, between ~0.7 - 0.8. The range in aspen stands was greater than those for spruce, with a minimum R:A ratio of ~0.30 and a maximum of 0.75. The fitted aspen simulations were also restricted to a narrower range in R:A, between ~0.35 - 0.65. The difference in R:A ratio between fitted spruce and aspen simulations was significant. That is, spruce stands required a significantly greater proportion of assimilated carbon for respiratory processes. Young aspen stands had higher R:A values than the mature stands, and this difference between age classes of aspen was also significant. When foliage N was held constant a comparable range in the R:A ratio emerged (58 - 82% in spruce; 31 - 66% in aspen), a result that was supported by the mostly low correlation between foliage N and the R:A ratio.
Variations in the R:A ratio were inversely related to those in PAR utlization. This was true for both the fitted and the generalized simulations. In spruce stands, 82% of the variability in PAR utilization was explained by the R:A ratio, and in aspen stands the explained variance ranged from 52% in mature stands to 93% in the young stands. The linear regression relationships of R:A on PAR utilization were all significant. The lower explained variance in mature stands was a result of the small range in observed values of PAR utilization among these stands. For the generalized simulations with all variables except LAI and foliage N held constant, 92% of variability in PAR utilization among spruce stands was explained by the R:A ratio. In aspen stands, the relationship was non-linear.
Gross versus Net Carbon Yield of APAR
The amount of gross carbon assimilation (GPP) in relation to APAR (gross PAR utlization) was restricted to a narrower range of values than net PAR utlization. The relationship between APAR and assimilation was highly significant among both the spruce (94% explained variance) and aspen (85% explained variance) stands, and the coefficients of variation on gross PAR utlization were lower in all cases than those on net PAR utlization. The slope of a linear regression of APAR on gross carbon assimilation for all "unstressed" stands (i.e., excluding the young aspen stands on poor soils) was 2.56 g C MJ-1 (89% explained variance), which is equivalent to the unstressed quantum yield of photosynthesis for C3 plants at a temperature of ~22 degrees C. Mean air temperature during the growing season was 14.2 degrees C, which provides a C3 quantum yield value of 3.12 g C MJ-1. Thus, the slope estimate of gross PAR utilization was lower than the comparable quantum yield value, as would be expected due to environmental contraints on photosynthetic rates.
Unlike net PAR utilization, the difference in gross PAR utilization between spruce and aspen stands was not significant, however, the difference between young and mature aspen stands remained significant. Stand-averaged daily gross PAR utilization was very similar for both species (next figure). Those points off the 1:1 line were periods early in the growing season when the organic soils of the spruce stands were still frozen but the aspen stands were photosynthetically active. Thus, whereas average net daily carbon flux during the growing season differed significantly between the two species (3.44 versus 5.41 g C m-2 day-1), mean differences in average daily gross C-flux per unit APAR (2.08 versus 2.14 g C MJ-1) were insignificant.
Summary and Implications of Results
The journal publication discusses these results in some detail - a brief summary is provided here.
These findings strongly support previous work which showed that measurement of light harvesting alone is insufficient to predict NPP in boreal forest stands (see the results of another paper). Despite the fact that much of the variability in NPP is driven by variables that are linked to light harvesting (e.g., leaf area and duration), we have shown that respiratory costs explained a large proportion of the variability in net PAR utilization among stands. The results have thus identified the R:A ratio as an important determinant of NPP, but have also shown that biomass is a poor estimator of the R:A ratio. Additional efforts are therefore needed to characterize respiratory costs in relation to the amount of actively respiring biomass. Estimating unstressed PAR utilization from the quantum yield of photosynthesis is another important step in simplifying the estimation of NPP for large areas using remote sensing. Our results sugggest that the use of remote sensing in monitoring NPP is simplified with respect to gross PAR utilization, but made more difficult by the need to consider differences in R:A between plant functional types.
Methods to incorporate these results in the remote sensing of global terrestrial NPP have been demonstrated to be technically feasible (Global Production Efficiency Model).
These issues are being addressed by related work in Central Alaska and the Boreal Ecosystem Atmosphere Study (BOREAS) among others.
The research discussed here has been published in the following publications:
S.J.Goetz and S.D.Prince, 1998, Variability in carbon exchange and light utilization among boreal forest stands: implications for remote sensing of net primary production, Canadian Journal of Forest Research, 28 (3) : 375-389.
Goetz, S. J. and S. D. Prince. 1996. Remote sensing of net primary production in boreal forest stands, Agricultural and Forest Meteorology 78 (3): 149-179.
Goetz, S. J. and S. D. Prince. 1998. Modeling terrestrial carbon exchange and storage: evidence and implications of functional convergence in light use efficiency,, Advances in Ecological Research 28 : 57-92.