JXB Advance Access originally published online on November 13, 2007
Journal of Experimental Botany 2008 59(7):1951-1961; doi:10.1093/jxb/erm234
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RESEARCH PAPER |
The role of stomatal acclimation in modelling tree adaptation to high CO2
1School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney NSW 2052, Australia
2Bushfire Cooperative Research Centre, Australia
3Ensis, PO Box E4008, Kingston ACT 2604, Australia
* To whom correspondence should be addressed at Ensis, PO Box E4008, Kingston ACT 2604 Australia. E-mail: tom_buckley{at}alumni.jmu.edu
Received 21 June 2007; Revised 2 September 2007 Accepted 4 September 2007
 |
Abstract |
|---|
 Top Abstract IntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
Carbon dioxide enrichment changes the balance of photosynthetic limitations due to water, nitrogen, and light. This paper examines the role of stomata in these changes by comparing enrichment responses predicted by an optimality-based tree growth model, DESPOT, using three alternative ‘setpoints’ for stomatal acclimation: leaf water potential (
l-setpoint), the ratio of intercellular to ambient CO2 mole fraction (ci/ca-setpoint), and the parameters in a simple model in which stomata are controlled by H2O and CO2 supply and demand (linked feedback). In each scenario, stomatal conductance (gs) and photosynthetic capacity (Vm) declined, productivity and leaf area index (LAI) increased, and ci/ca remained within 5% of its pre-enrichment value. Height growth preceded the LAI response in the l-setpoint and linked feedback scenarios, but not in the ci/ca-setpoint scenario. These trends were explained in terms of photosynthetic resource substitution using the equimarginal principle of production theory, which controls carbon allocation in DESPOT: enrichment initially increased the marginal product for light, driving substitution towards light; height growth also drove substitution towards N in the l and feedback scenarios, but the inflexibility of ci/ca prevented that substitution in the ci/ca scenario, explaining the lack of height response. Each scenario, however, predicted similar behaviour for ci/ca and carbon and water flux. These results suggest that ‘setpoints’ may be robust tools for linking and constraining carbon and water fluxes, but that they should be used more cautiously in predicting or interpreting how those fluxes arise from changes in tree structure and physiology. Key words: Carbon dioxide, climate change, photosynthesis, stomata, transpiration, water potential
|
Introduction |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
A major challenge facing plant biologists today is the need for confident predictions about plant responses to climate change, based on sound theoretical understanding and supported by experiment where possible (Steffen and Canadell, 2005). This need is especially important for trees, because forests dominate the terrestrial biosphere, in terms of both carbon stocks and carbon fluxes. However, it is also especially difficult for trees, because the long lifespans of many species make it impossible to document experimentally the long-term effects of CO2 enrichment above present-day levels. This makes long-term forecasting of tree growth and gas exchange heavily dependent on assumptions about the character of adaptive growth regulation.
One hallmark of plant growth is the strong co-ordination between structural development and physiological functioning, often described as ‘functional balance’, for example, between sapwood and leaf area (Shinozaki et al., 1964a, b) or roots and shoots (Davidson, 1969; Thornley, 1972a, b). Many growth models incorporate functional balance as a central constraint to predict carbon allocation and thus structural adjustment (Valentine, 1985, 1999; Mäkelä, 1986, 1997, 1999; Hilbert and Reynolds, 1991; Deleuze and Houllier, 1995; Luan et al., 1996; Grote, 1998; Lo et al., 2001; Battaglia et al., 2004). The dynamic regulation of leaf gas exchange is also characterized by functional balance: stomatal behaviour co-ordinates CO2 supply with photosynthetic demand and evaporative demand with hydraulic supply, leading to virtually homeostatic control of both ci/ca (intercellular/ambient [CO2]) and leaf water potential, l (Wong et al., 1979, 1985a, b; Farquhar and Sharkey, 1982; Jones, 1990; Tardieu, 1993; Saliendra et al., 1995) (a list of symbols with definitions and dimensions is given in Table 1). These functional balancing acts—of structural adaptation and stomatal regulation—are deeply entangled with one another. Homeostatic control of ci/ca depends as much on the co-ordination of root, stem, and leaf production as it does on stomatal responses to the leaf environment, because those structural features determine the relative supplies of nitrogen, water, and light needed for photosynthesis.
|
Very little is known, however, about how the ‘setpoints’ for
l and ci/ca are controlled, or about their potential to change systematically over long time scales in relation to the environmental and structural effectors of photosynthetic resource balance (Farquhar et al., 1989; Ehleringer and Cerling, 1995; Miller et al., 2001; Miller, 2002). The data that are available paint a varied and often contradictory picture. For example, Barnard and Ryan (2003) found that the hydraulic burdens created by height growth in Eucalyptus saligna were compensated by structural adjustment of sapwood:leaf area ratio and down-regulation of the l-setpoint, which sustained stomatal conductance, gs, with the result that the ci/ca-setpoint was not lower in taller, older trees (it was slightly higher, in fact). In contrast, Phillips et al. (2003) found large reductions in gs, l and ci/ca in 25 m tall Oregon white oak (Quercus garryana Dougl. ex. Hook) as compared with 10 m tall trees. Wide ranges of compensation have been reported for each of these parameters in relation to height growth for various species (Ryan et al., 2006). Carbon dioxide enrichment is similar to height growth in that it drives a complicated series of interacting changes in tree structure and function (Norby et al., 1999; Korner et al., 2005; Hyvonen et al., 2007) that are particularly difficult to disentangle because of the long time scales over which they occur. Most of what we know about how trees will adapt to high CO2 is based on observed responses to a few years of CO2 enrichment (Ainsworth and Long, 2005; Norby et al., 2005). Although trees suddenly exposed to elevated CO2 typically gain more carbon to begin with, they may not distribute the extra carbon among roots, leaves, and other carbon pools in the same proportions as they did before enrichment (Norby et al., 1999; Korner, 2006; Hyvonen et al., 2007). The actual patterns of carbon reinvestment, combined with any changes in the apparent setpoints for stomatal control, will ultimately determine how tree carbon–water balance responds to CO2 enrichment over many decades and centuries.
The purpose of this paper is to examine the effect of alternative hypotheses about the long-term ‘setpoints’ for stomatal acclimation on predictions of adaptive tree responses to CO2 enrichment. Tree growth was simulated at elevated CO2 using a model, DESPOT (Buckley and Roberts, 2006a), in which carbon allocation is govered by the equimarginal principle of production theory, and carbon gain is calculated from a big-leaf Farquhar et al. (1980) model with stomatal conductance prescribed by one of three alternative ‘setpoint’ hypotheses. These simulations were assessed in relation to enrichment responses reported for trees in FACE (free-air CO2 enrichment) experiments, and the patterns predicted for structural and physiological adaptation under each acclimation scenario were analysed in terms of photosynthetic resource substitution using the equimarginal principle.
|
Outline of the approach |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
The response of mature trees to a step increase in ambient CO2 mole fraction (ca) from 370 ppm to 570 ppm was simulated. This was meant to mimic FACE experiments, to allow the early stages of simulated responses to be compared with empirical data. A tree growth model called DESPOT (Buckley and Roberts, 2006a) was used, which was designed to minimize the influence of a priori assumptions about the structural or physiological outcomes of adaptive growth regulation. For example, patterns of height growth, diameter–height relationships, sapwood:leaf area ratios, and root:shoot ratios are not prescribed as empirical constraints. Instead, they emerge from a sequence of carbon allocation decisions that are chosen by a numerical algorithm to satisfy a single goal function, which is to maximize carbon gain in the next time step.
How DESPOT responds to CO2 enrichment
A large step increase of atmospheric CO2 initiates a cascade of adaptive responses in leaf physiology and plant structure. In order for this adaptive cascade to make the best use of the extra CO2, in the sense of maximizing the response of net carbon gain, it must satisfy a rule known as the equimarginal principle (see any microeconomics text). This principle states that the ratio of marginal product (the sensitivity of carbon gain to a given limiting resource, such as water) to marginal cost (the additional carbon investment needed to procure more of that limiting resource) should be the same for each photosynthetic resource, provided the responses of carbon gain to each resource, and of each resource to C investment, are convex. The marginal products are defined by 
Equation 1:
 | (1) |
This leaves two degrees of freedom in the calculations of carbon gain, however: the ratio of electron transport capacity to carboxylation capacity (Jm/Vm) and stomatal conductance. To reduce the analysis presented here to a single degree of freedom, Jm/Vm will be treated as a constant—acknowledging that enrichment often alters this ratio (Ainsworth and Long, 2005), but leaving analysis of those effects for further work and focusing instead on stomatal control. Three alternative assumptions for constraining stomatal conductance will be explored in the present study:
- (i) l-setpoint: leaf water potential is assumed invariant over growth time scales. Transpiration is calculated from soil–leaf water potential gradient, given the leaf-specific hydraulic conductance (KL), which is calculated from structural properties such as sapwood area, fine root biomass, and tree height. gs is calculated from the transpiration rate assuming fixed evaporative demand (D). The published form of DESPOT used this assumption, on the grounds that l is physically bounded in the negative direction by the threshold for runaway xylem embolism. However, the direct stomatal response to CO2 enrichment reduces water loss, potentially shifting l in the positive direction and thus weakening the rationale for using a fixed l to constrain stomatal control in the case of CO2 enrichment responses. Two other alternatives are outlined below.
- (ii) ci/ca-setpoint: the ratio ci/ca is assumed to be constant. The assimilation rate, A, is calculated from ci, given ca and structurally determined values for irradiance [calculated from leaf area index (LAI) using Beer's Law] and photosynthetic capacity (based on canopy N content, which is simulated dynamically based on uptake by fine roots, N requirements for tissues other than leaves, and senescence losses). gs is then calculated from A, ci, and ca.
- (iii) Linked feedback: stomata respond directly to the balance of both CO2 and H2O supply and demand through linked feedback responses, as postulated in the models of Jarvis and Davies (1998) and Buckley et al. (2003). Simple measures of the balance of CO2 and H2O supply and demand are, respectively, the residual photosynthetic capacity, Ar (the amount A would increase if ci equalled ca), and leaf turgor pressure, P. Assuming gs = FPAr, where F is a parameter, and using P=Pm–E/KL=Pm–gsD/KL, where Pm is maximum leaf turgor (osmotic pressure plus soil water potential), this implies
(2)
- In this scenario, the parameters Pm and F are assumed to be conserved by long-term stomatal acclimation, so they are analogous to the ‘setpoints’ in the other scenarios. Note that the parameter F incorporates both stomatal density and the sensitivity of guard cell osmotic pressure to P and Ar.
- (ii) ci/ca-setpoint: the ratio ci/ca is assumed to be constant. The assimilation rate, A, is calculated from ci, given ca and structurally determined values for irradiance [calculated from leaf area index (LAI) using Beer's Law] and photosynthetic capacity (based on canopy N content, which is simulated dynamically based on uptake by fine roots, N requirements for tissues other than leaves, and senescence losses). gs is then calculated from A, ci, and ca.
Simulations were performed in DESPOT using each of these alternatives, with the following values for each setpoint in their respective simulations: for the l-setpoint scenario, l= –1.0 MPa (the value used in the original publication of DESPOT); for the ci/ca-setpoint scenario, ci/ca=0.65 (the value from the l-setpoint scenario at the time of CO2 enrichment); and for the linked-feedback scenario, Pm=1.2 MPa and F=1.0 mol air MPa–1 µmol–1 CO2 (these values were chosen to produce pre-enrichment values of l and ci/ca similar to those in the other two scenarios). l, ci/ca, Pm, and F were calculated in each scenario, to determine how these quantities vary when one of them is held constant by assumption. The marginal products (Eqn 1) were also calculated to illustrate how the economic landscape of tree growth differs under each stomatal acclimation scenario.
|
Results |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
Figures 1–5
show the effects predicted by DESPOT for a step increase in atmospheric CO2 mole fraction (ca) from 370 ppm to 570 ppm. The initial response is dominated by an increase in height growth (Fig. 1a) in the l-setpoint and linked-feedback scenarios, but by an increase in LAI (Fig. 1b) in the ci/ca-setpoint scenario. LAI declines in the first year after enrichment in the l and feedback scenarios, then increases steadily, whereas height growth never responds significantly in the ci/ca scenario. All three scenarios predict decreased allocation to fine roots, whether expressed relative to leaf carbon (Fig. 1c) or sapwood C (Fig. 1d).
|
|
|
|
|
A/All three scenarios also predict declines in stomatal conductance (gs, Fig. 2a) and photosynthetic capacity [carboxylation capacity, Vm, Fig. 2b (electron transport capacity, Jm, is not independent of Vm in this model)]. Incident photosynthetic irradiance per unit leaf area (I, Fig. 2c) declines immediately in the ci/ca scenario, but initially rises in the
l and feedback scenarios as a result of the initial drop in LAI; irradiance eventually drops below pre-enrichment values in the l and feedback cases, but by a smaller amount than gs or Vm. In all cases, net primary productivity (NPP, Fig. 3a) increases strongly at first before declining somewhat and eventually settling at a stimulation of 
18–21%. Stand-level transpiration (Fig. 3b) declines by 25% in all scenarios after 10 years, before diverging somewhat after 20 years and then converging again such that the net decline after 100 years is 21% for the l and ci/ca scenarios, and 24% for the feedback scenario (Fig. 3b).
Trajectories for l, ci/ca, Pm, and F are presented in Fig. 4a–d, respectively; in each panel, one trajectory is simply a horizontal line, representing the simulation in which that parameter was treated as a constant. Leaf water potential increases (moves towards zero) by 5% in the linked-feedback scenario (dashed line in Fig. 4a), or by 25% in the ci/ca-setpoint scenario (dotted line in Fig. 4a). ci/ca fluctuates somewhat in the l-setpoint scenario before increasing by 3% after 50 years (solid line in Fig. 4a); in the linked-feedback scenario, ci/ca declines by 5% after 30 years, but then slowly increases to within 3% of its pre-enrichment value after 50 years. To maintain a fixed setpoint for either l or ci/ca after enrichment, one of the parameters of the linked-feedback model (Eqn 2) would have to decline: the l-setpoint scenario requires either a 6% decline in Pm or a 16% decline in F after 50 years (solid lines in Fig. 4b and c, respectively), whereas the ci/ca-setpoint scenario requires declines of 22–29% in Pm or of 50–60% in F (dotted lines in Fig. 4b and c, respectively).
Figure 5 shows trajectories predicted for the marginal products for water, nitrogen, and light (µe, µn, and µi; solid, dashed, and dotted lines in Fig. 5, respectively; Eqn 1), under each stomatal acclimation scenario. The general trend is identical in each scenario: µi increases strongly at first, then drops to near its initial value, while µn and µe increase less than µi but remain elevated. The eventual relative increase is greatest for µn, followed by µe and µi. In the ci/ca-setpoint scenario (Fig. 5b), µe and µn do not change after the initial response to enrichment, whereas µe and µn both fluctuate after enrichment in the l and feedback scenarios (Fig. 5a and c, respectively), with µn decreasing and µe gradually increasing during height growth.
|
Discussion |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
No significant change in the ratio of ci/ca was predicted by DESPOT after several years of CO2 enrichment, for any of the three alternative ‘setpoint’ assumptions used to constrain stomatal acclimation—ci/ca,
l, or the parameters of the linked hydraulic–biochemical feedback model for gs (Eqn 2). This result is consistent with many data showing that the co-ordination of stomatal conductance and photosynthesis is usually not substantially altered by CO2 enrichment (Ellsworth, 1999; Medlyn et al., 2001; Schafer et al., 2003; Herrick et al., 2004; Ainsworth and Long, 2005). Several other predictions common to all three scenarios are also consistent with most observations from FACE experiments on trees (Ainsworth and Long, 2005): a decline in photosynthetic capacity (Vm), a somewhat larger decline in gs, an eventual increase in LAI, and an enhancement of NPP that is greatest in the first year or two after enrichment, but which persists indefinitely at 14–21% above pre-enrichment values (Table 2).
|
Another prediction common to all three scenarios, that allocation to fine roots decreases relative to allocation to leaves or to sapwood, contradicts some results (Hyvonen et al., 2007). One possible explanation for this is the lack of a mechanism for negative feedback between increased growth and nutrient availability in the model used here. However, reported enrichment responses for below-ground allocation are highly variable, with strong responses typically found in nutrient-limited conditions (Norby et al., 2004) and small and ambiguous responses in the absence of container effects or prior nutrient limitations (Norby et al., 1999; Tingey et al., 2000; Hyvonen et al., 2007). The fact that similar trends were predicted by each acclimation scenario for fine root allocation and NPP does suggest, at any rate, that variation in below-ground allocation responses is not related to variation in the character of stomatal acclimation, and that none of the physiological constraints built into DESPOT should inherently prevent a sustained productivity enhancement of the same order as that observed in several forest FACE experiments (23±2%) (Norby et al., 2005).
The adaptive character of these responses can be interpreted in terms of changes in the sensitivities of carbon gain to water, nitrogen, and light (the marginal products µe, µn, and µi, respectively; Eqn 1 and Fig. 5), by reference to the equimarginal principle of production theory, which governs carbon allocation in the growth model used here. This principle states that the ratio of marginal product to marginal carbon cost should be the same for each photosynthetic resource. CO2 enrichment initially causes a very large increase in µi relative to µe and µn (Fig. 5). Therefore, equimarginality requires either a shift in carbon allocation towards light capture, which would increase the marginal carbon cost of light, or an increase in leaf-level irradiance relative to Vm and gs, which would reduce the marginal product for light relative to those for nitrogen and water. The increase in LAI (Fig. 1b) increases light capture, which works towards the first objective, and the reductions in Vm and gs (Fig. 2a, b) both work towards the second objective. Enrichment also increases µn more than µe; this dictates an increase in Vm relative to gs, which explains why gs declines more than Vm (Fig. 2a, b).
There was one fundamental difference in the enrichment responses predicted for the alternative scenarios for stomatal acclimation. The l-setpoint and linked-feedback scenarios predicted a strong initial response of height growth followed by an increase in LAI, whereas the ci/ca-setpoint scenario predicted an immediate increase in LAI with no associated height growth response (Fig. 1a, b). Both of these responses—height growth and LAI enhancement—are driven by the initial increase in µi, which increases the profitability of light capture. Height growth eventually becomes less profitable due to increasing hydraulic burdens and respiratory costs, so allocation eventually shifts to leaf area production in the l and feedback scenarios. The reason height growth did not respond in the ci/ca-setpoint scenario involves the effect on economic flexibility of treating ci/ca as a rigid setpoint. Height growth preferentially increases the marginal carbon cost of water (Buckley and Roberts, 2006b) so, in order for height growth to be economically efficient, the equimarginal principle requires substitution of light and/or nitrogen for water—or equivalently, of CO2 demand for CO2 supply. Treating ci/ca as a rigid setpoint precludes the required plasticity in CO2 supply and demand. This makes height growth inherently less efficient in the ci/ca-setpoint scenario, so DESPOT's optimization routine chooses to enhance leaf area rather than height growth in that scenario. The economic effect of this constraint is illustrated by the fact that µn and µe remain constant in the ci/ca scenario (Fig. 5b), whereas they change with respect to one another in relation to height growth dynamics in the l and feedback scenarios (Fig. 5a, c).
This does not imply that the ci/ca-setpoint scenario is less representative of how real trees adapt to CO2 enrichment, however. Reported LAI responses to enrichment in tree species are variable, with larger responses typically found for young trees in open stands (Kellomaki and Wang, 1997) or for mature stands with lower LAI (Norby et al., 2005), and little or no response in mature forests with high LAI (Norby et al., 2003; Schafer et al., 2003). The present analysis suggests this variation may relate to plasticity in height growth, such that the initial responses of LAI and height to enrichment are anti-correlated among species. There are reasons to suspect that DESPOT's assumption of total plasticity in carbon allocation fractions overestimates the responsiveness of height growth to environmental change, particularly in mature trees. First, small reductions in either the flexibility or the precision of DESPOT's allocation algorithm cause NPP to decline with age, which the model does not otherwise predict (Buckley and Roberts, 2006a)—hinting that the model may indeed overestimate the plasticity of growth form in older trees. Secondly, McDowell et al. (2005) reported that carbon isotope discrimination was less sensitive to environmental change in tall Douglas-fir (Pseudotsuga menziesii) trees than in short trees, which they suggested may reduce the sensitivity of tall or otherwise hydraulically burdened trees to climate change.
Changes in stomatal function required to maintain l- or ci/ca-setpoints
The alternative setpoints can also be compared on the basis of their implications for stomatal physiology, because specific changes in the anatomical determinants of stomatal function are required to maintain fixed setpoints for either l or ci/ca as tree structure and physiology adapt to CO2 enrichment. A model based on the premise that stomatal control involves direct feedback responses to leaf turgor pressure and residual photosynthetic capacity (the ‘linked-feedback’ scenario, Eqn 2) (Jarvis and Davies, 1998; Buckley and Mott, 2002; Buckley, 2005) was used to infer how the parameters of those putative feedback responses—maximum turgor (Pm) and stomatal sensitivity (F) to the effectors of guard cell osmotic pressure—must vary during enrichment responses to preserve setpoints for l or ci/ca. The parameter F captures changes in both stomatal density and the gain of guard cell signal transduction. The l-setpoint scenario requires a 3–6% decline in Pm or a 9–16% decline in F, whereas the ci/ca-setpoint scenario requires a 25% decline in Pm or a 75% decline in F (Fig. 4b, c).
These predictions are difficult to validate or refute empirically, because little is known about the long-term acclimation in parameters that describe processes or properties directly involved in stomatal control. Stomatal density (embedded in F) is well known to decline in response to increasing ca during growth; however, that response appears to saturate close to present-day ambient levels, and FACE experiments typically show little or no change in stomatal density (Medlyn et al., 2001; Herrick et al., 2004; Tricker et al., 2005). Similarly, Herrick et al. (2004) found no effect of enrichment on the sensitivity of gs to soil moisture or evaporative gradient in sweetgum (Liquidambar styraciflua) trees in the Duke FACE experiment, and Leakey et al. (2006) found that short-term stomatal responses in enriched soybean (Glycine max) plants were accurately described using the Ball–Berry model with pre-enrichment parameter values. These results suggest, but by no means prove, that the reduced parameters of stomatal function are not strongly sensitive to CO2 enrichment above present-day levels.
This supports the linked-feedback scenario, which assumes that Pm and F do not respond to enrichment. In contrast, the l- and ci/ca-setpoint scenarios both require either Pm or F to change after enrichment. The required change is smaller for Pm than for F in both cases, suggesting that either setpoint scenario is more likely to involve osmotic down-regulation (i.e. reductions in maximum leaf turgor, Pm) than reductions in stomatal density or guard cell metabolic sensitivity. Additionally, much smaller changes in Pm and F are required to preserve a constant setpoint for l than for ci/ca. The matter could be clarified by characterizing in greater detail how enrichment affects stomatal properties in trees, ideally in the language of a process-based model. It is interesting to note that the linked-feedback scenario (fixed Pm and F) predicts trajectories for both l and ci/ca that remain within ±5% of their pre-enrichment values (Fig. 5a, b). This prediction is intriguing because it shows that conservative behaviour in the apparent hydraulic and biochemical setpoints of stomatal control can emerge passively from local feedback regulation—without the need for any systemic co-ordination or acclimation of stomatal physiology, and even after a 200 ppm increase in ca.
Recommendations for model-based prediction and analysis
The analysis presented above leads to two conclusions relevant to the choice of stomatal constraints in models used to interpret or simulate tree responses to CO2 enrichment. First, the adaptive landscape on which growth and gas exchange are regulated is strongly affected by a priori constraints on the phenomenological outcomes of structural and physiological adaptation. A constant ratio of ci/ca, for example, precludes resource substitution between nitrogen and water, which in turn leads to very different predictions for both the timing and magnitude of acclimatory adjustments in LAI and tree height to enrichment. Secondly, if constraints are based on the regulatory processes that are hypothesized to give rise to homeostatic setpoints, rather than on the assumption that those setpoints will be preserved, the long-term behaviour that emerges for those setpoints turns out to be quite conservative anyway. The recommendation arising from these conclusions depends on the reason for which one is simulating tree responses to CO2 in the first place. Any process-based model, even the simple conceptual ‘linked-feedback’ model used here, increases computational time and complexity, and the results of this study suggest that the phenomenology that emerges from such models is very similar to that embedded in setpoint assumptions anyway. Therefore, when the purpose of modelling is simply to project gas exchange dynamics under high CO2, this study offers no reason to abandon the simplicity and computational efficiency offered by setpoint assumptions. If, on the other hand, the purpose of modelling is to understand the interplay between physiological and structural responses to CO2 enrichment—particularly from an adaptive or economic perspective—a more wary approach to ‘setpoints’ is warranted. Understanding of these issues could be improved by experiments designed to characterize stomatal acclimation in terms of reduced properties related to the processes involved in stomatal control, and by experiments designed to quantify the plasticity of carbon allocation, particularly height growth, with respect to the drivers of photosynthetic resource balance.
|
Conclusions |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
Adaptive tree responses to a step increase in ambient CO2 from 370 ppm to 570 ppm were simulated with an optimality-based growth model, using three alternative ‘setpoints’ for stomatal acclimation: leaf water potential (
l), the ratio of intercellular to ambient CO2 mole fraction (ci/ca), or the parameters of a simple feedback model for stomatal control (Eqn 2). None of the scenarios predicted large changes in the co-ordination of stomatal conductance with photosynthesis (as ci/ca) in the first decade after enrichment. All scenarios predicted declines in gs and Vm, and increases in NPP and LAI as reported in FACE experiments. The LAI response was initially offset by diversion of carbon for height growth in the l and feedback scenarios, but not in the ci/ca scenario. Both responses (LAI and height growth) served to re-establish equimarginality by shifting photosynthetic resource balance towards light, after the CO2 enrichment-induced jump in the marginal product for light. Height growth also requires substitution of N for water, which is impossible if ci/ca is fixed, explaining the lack of height response in the ci/ca scenario. Conservative behaviour for ci/ca was predicted in the l-setpoint and feedback scenarios, and all scenarios predicted similar long-term behaviour for stand-level carbon and water fluxes and above- versus below-ground partitioning. These results suggest that ci/ca or l could be used as setpoints for constraining flux projections. However, they also highlight the importance of a priori constraints on allocation and co-ordination in determining patterns of structural and physiological adjustment. This suggests that the adaptive or physiological basis of those patterns cannot be fully understood without a better understanding of the process basis of stomatal acclimation itself. |
Appendix |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
Gas exchange model
The calculations described in this paper, including those embedded in the DESPOT model, were based on the photosynthesis model of Farquhar et al. (1980), combined with simplified Fick's Law-based diffusion models for CO2 and water vapour flux between leaves and the atmosphere. Net CO2 assimilation rate is defined by the intersection of mesophyll demand (Ad) and stomatal diffusion (supply, As):
 | (A1) |
 | (A2) |
 | (A3) |
} is defined as the lesser solution X of X2–(y+z)X+yz=0, and Av and Aj are RuBP carboxylation- and regeneration-limited assimilation rates, respectively:
 | (A4) |
 | (A5) |
* is the photorespiratory compensation point, and J is the potential electron transport rate, given by J=minh{Jm, 
I, J}; Jm is the maximum potential electron transport rate, I is incident PPFD, is effective maximum quantum yield of electron transport from incident PPFD, and J is a curvature parameter (0.9). Vm and Jm are considered proportional to leaf N content, N (Vm=
vN and Jm=jN), which in turn is the ratio of canopy N content to leaf area (N=Nt/L). In DESPOT, j=1.9 µmol e– mmol–1 N and v=j/(2.1 e–/CO2). Incident PPFD, I, is annual canopy PPFD capture divided by leaf area and by a time integration factor Y [Y represents, in essence, the number of seconds of positive photosynthesis each year; see Buckley and Roberts (2006a) for details] (I=It/L). g is annual canopy transpiration (Et) divided by leaf area, Y and 1.6xD [g=Et/(1.6xDLY)], where D is leaf–air H2O mole fraction difference and 1.6 converts from H2O to CO2 diffusivity. Note that this assumes negligible boundary layer resistance. Calculating E, N, and I from Et, Nt, and It in this way is equivalent to a ‘big-leaf’ assumption, i.e. the gas exchange model is assumed to be scale-invariant (Farquhar, 1989; Field, 1991).
Note that for A<1, the intersection of Eqns A2 and A3 creates a quartic expression for ci, which can be solved directly as described in any mathematics text. It would be simpler to calculate the intersections of As with Aj and Av separately and then take the minimum; however, this is strictly incorrect if A<1, and it creates discontinuous marginal products if A is assumed equal to 1. In reality, there will always be some co-limitation, however small, between carboxylation and regeneration, and although accommodating A<1 is more difficult computationally, meaningful analysis of photosynthetic resource economics is otherwise impossible. In the present study A was assumed to be equal to 0.9.
Marginal products
The marginal products from water, nitrogen, and light (transpiration, photosynthetic leaf N content, and incident PPFD, to be precise) can be calculated from the gas exchange model described above. Only the final expressions are given here; see Buckley et al. (2002) for complete derivations. The marginal product from water use is
 | (A6) |
 | (A7) |
 | (A8) |
 | (A9) |
 | (A10) |
 | (A11) |
 | (A12) |
 | (A13) |
 | (A14) |
|
Acknowledgements |
|---|
I thank James Morison, Gabriel Cornic, and the organizers of the 14th International Congress on Photosynthesis for the invitation to submit this article. This work was funded by the Highfire Project of Australia's Bushfire Cooperative Research Centre.
|
References |
|---|
|
TopAbstractIntroductionOutline of the approachResultsDiscussionConclusionsAppendixReferences |
|---|
Ainsworth EA, Long S. What have we learned from 15 years of free-air CO2 enrichment (FACE)? A meta-analytic review of the responses of photosynthesis, canopy properties and plant production to rising CO2. New Phytologist (2005) 165:351–372.[CrossRef][Web of Science][Medline]
Barnard HR, Ryan MG. A test of the hydraulic limitation hypothesis in fast-growing Eucalyptus saligna. Plant, Cell and Environment (2003) 26:1235–1245.[CrossRef]
Battaglia M, Sands P, White D, Mummery D. CABALA: a linked carbon, water and nitrogen model of forest growth for silvicultural decision support. Forest Ecology and Management (2004) 193:251–282.[CrossRef][Web of Science]
Buckley TN. The control of stomata by water balance (Tansley Review). New Phytologist (2005) 168:275–292.[CrossRef][Web of Science][Medline]
Buckley TN, Miller JM, Farquhar GD. The mathematics of linked optimisation for nitrogen and water use in a canopy. Silva Fennica (2002) 36:639–669.[Web of Science]
Buckley TN, Mott KA. Stomatal water relations and the control of hydraulic supply and demand. Progress in Botany (2002) 63:309–325.
Buckley TN, Mott KA, Farquhar GD. A hydromechanical and biochemical model of stomatal conductance. Plant, Cell and Environment (2003) 26:1767–1785.[CrossRef]
Buckley TN, Roberts DW. DESPOT, a process-based tree growth model that allocates carbon to maximize carbon gain. Tree Physiology (2006a) 26:129–144.[Web of Science][Medline]
Buckley TN, Roberts DW. How should leaf area, sapwood area and stomatal conductance vary with tree height to maximise growth? Tree Physiology (2006b) 26:145–157.[Web of Science][Medline]
Davidson RL. Effect of root/leaf temperature differentials on root/shoot ratios in some pasture grasses and clover. Annals of Botany (1969) 33:561–569.
Deleuze C, Houllier F. Prediction of stem profile of Picea abies using a process-based tree growth model. Tree Physiology (1995) 15:113–120.[Abstract]
Ehleringer JR, Cerling TE. Atmospheric CO2 and the ratio of intercellular to ambient CO2 concentrations in plants. Tree Physiology (1995) 15:105–111.[Abstract]
Ellsworth DS. CO2 enrichment in a maturing pine forest: are CO2 exchange and water status in the canopy affected? Plant, Cell and Environment (1999) 22:461–472.[CrossRef]
Farquhar GD. Models of integrated photosynthesis of cells and leaves. Philosophical Transactions of the Royal Society B: Biological Sciences (1989) 323:357–367.[CrossRef]
Farquhar GD, Ehleringer JR, Hubick KT. Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology (1989) 40:503–537.[CrossRef][Web of Science]
Farquhar GD, Sharkey TD. Stomatal conductance and photosynthesis. Annual Review of Plant Physiology (1982) 33:317–345.[Web of Science]
Farquhar GD, von Caemmerer S, Berry JA. A biochemical model of photosynthetic CO2 assimilation in leaves of C3 species. Planta (1980) 149:78–90.[CrossRef][Web of Science]
Field CB. Ecological scaling of carbon gain to stress and resource availability. In: Response of plants to multiple stresses—Mooney HA, Winner WE, Pell EJ, eds. (1991) San Diego: Academic Press. 35–65.
Grote R. Integrating dynamic morphological properties into forest growth modeling. II. Allocation and mortality. Forest Ecology and Management (1998) 111:193–210.[CrossRef][Web of Science]
Herrick JD, Maherali H, Thomas RB. Reduced stomatal conductance in sweetgum (Liquidambar styraciflua) sustained over long-term CO2 enrichment. New Phytologist (2004) 162:387–396.[CrossRef][Web of Science]
Hilbert DW, Reynolds JF. A model allocating growth among leaf proteins, shoot structure, and root biomass to produce balanced activity. Annals of Botany (1991) 68:417–425.
Hyvonen R, Agren GI, Linder S, et al. The likely impact of elevated [CO2], nitrogen deposition, increased temperature and management on carbon sequestration in temperate and boreal forest ecosystems: a literature review. New Phytologist (2007) 173:463–480.[CrossRef][Web of Science][Medline]
Jarvis AJ, Davies WJ. The coupled response of stomatal conductance to photosynthesis and transpiration. Journal of Experimental Botany (1998) 49:399–406.[Abstract]
Jones HG. Physiological aspects of the control of water status in horticultural crops. Horticultural Science (1990) 25:19–26.
Kellomaki S, Wang KY. Effects of long-term CO2 and temperature elevation on crown nitrogen distribution and daily photosynthetic performance of Scots pine. Forest Ecology and Management (1997) 99:309–326.[CrossRef][Web of Science]
Korner C. Plant CO2 responses: an issue of definition, time and resource supply. New Phytologist (2006) 172:393–411.[CrossRef][Web of Science][Medline]
Korner C, Asshoff R, Bignucolo O, Hattenschwiler S, Keel SG, Pelaez-Riedl S, Pepin S, Siegwolf RTW, Zotz G. Carbon flux and growth in mature deciduous forest trees exposed to elevated CO2. Science (2005) 309:1360–1362.
Leakey ADB, Bernacchi CJ, Ort DR, Long SP. Long-term growth of soybean at elevated [CO2] does not cause acclimation of stomatal conductance under fully open-air conditions. Plant, Cell and Environment (2006) 29:1794–1800.[CrossRef][Medline]
Lo E, Zhang MW, Lechowicz M, Messier C, Nikinmaa E, Perttunen J, Sievanen RP. Adaptation of the LIGNUM model for simulations of growth and light response in Jack pine. Forest Ecology and Management (2001) 150:279–291.[CrossRef][Web of Science]
Luan J, Muetzelfeldt RI, Grace J. Hierarchical approach to forest ecosystem simulation. Ecological Modelling (1996) 86:37–50.[CrossRef][Web of Science]
Mäkelä A. Implications of the pipe model theory on dry matter partitioning and height growth in trees. Journal of Theoretical Biology (1986) 123:103–120.[CrossRef][Web of Science]
Mäkelä AA. A carbon balance model of growth and self-pruning in trees based on structural relationships. Forest Science (1997) 43:7–24.[Web of Science]
Mäkelä A. Acclimation in dynamic models based on structural relationships. Functional Ecology (1999) 13:145–156.[CrossRef]
McDowell NG, Licata J, Bond BJ. Environmental sensitivity of gas exchange in different-sized trees. Oecologia (2005) 145:9–20.[CrossRef][Web of Science][Medline]
Medlyn BE, Barton CVM, Broadmeadow M, et al. Stomatal conductance of forest species after long-term exposure to elevated CO2 concentration: a synthesis. New Phytologist (2001) 149:247–264.[CrossRef][Web of Science]
Miller JM. Carbon isotope discrimination by Eucalyptus species (2002) The Australian National University. Canberra, 258.
Miller JM, Williams RJ, Farquhar GD. Carbon isotope discrimination by a sequence of Eucalyptus species along a sub-continental rainfall gradient in Australia. Functional Ecology (2001) 15:222–232.[CrossRef]
Norby RJ, DeLucia EH, Gielen B, et al. Forest response to elevated CO2 is conserved across a broad range of productivity. Proceedings of the National Academy of Sciences, USA (2005) 102:18052–18056.
Norby RJ, Ledford J, Reilly CD, Miller NE, O'Neill EG. Fine-root production dominates response of a deciduous forest to atmospheric CO2 enrichment. Proceedings of the National Academy of Sciences, USA (2004) 101:9689–9693.
Norby RJ, Sholtis JD, Gunderson CA, Jawdy SS. Leaf dynamics of a deciduous forest canopy: no response to elevated CO2. Oecologia (2003) 136:574–584.[CrossRef][Web of Science][Medline]
Norby RJ, Wullschleger SD, Gunderson CA, Johnson DW, Ceulemans R. Tree responses to rising CO2 in field experiments: implications for the future forest. Plant, Cell and Environment (1999) 22:683–714.[CrossRef]
Phillips N, Bond BJ, McDowell NG, Ryan MG, Schauer A. Leaf area compounds height-related hydraulic costs of water transport in Oregon White Oak trees. Functional Ecology (2003) 17:832–840.[CrossRef]
Ryan MG, Phillips N, Bond BJ. The hydraulic limitation hypothesis revisited. Plant, Cell and Environment (2006) 29:367–381.[CrossRef][Medline]
Saliendra NZ, Sperry JS, Comstock JP. Influence of leaf water status on stomatal response to humidity, hydraulic conductance, and soil drought in Betula occidentalis. Planta (1995) 196:357–366.[Web of Science]
Schafer KVR, Oren R, Ellsworth DS, Lai C-T, Herrick JD, Finzi AC, Richter DD, Katul GG. Exposure to an enriched CO2 atmosphere alters carbon assimilation and allocation in a pine forest ecosystem. Global Change Biology (2003) 9:1378–1400.[CrossRef][Web of Science]
Shinozaki K, Yoda K, Hozumi K, Kira T. A quantitative analysis of plant form: the pipe model theory. I. Basic analysis. Japanese Journal of Ecology (1964a) 14:97–105.
Shinozaki K, Yoda K, Hozumi K, Kira T. A quantitative analysis of plant form: the pipe model theory. II. Further evidence of the theory and its application in forest ecology. Japanese Journal of Ecology (1964b) 14:133–139.[Medline]
Steffen W, Canadell P. Carbon dioxide fertilisation and climate change policy (2005) Canberra: Australian Greenhouse Office.
Tardieu F. Will increases in our understanding of soil–root relations and root signalling substantially alter water flux models? Philosophical Transactions of the Royal Society B: Biological Sciences (1993) 341:57–66.[CrossRef]
Thornley JHM. A balanced quantitative model for root:shoot ratios in vegetative plants. Annals of Botany (1972a) 36:431–441.
Thornley JHM. A model to describe the partitioning of photosynthate during vegetative growth. Annals of Botany (1972b) 33:419–430.
Tingey DT, Phillips DL, Johnson MG. Elevated CO2 and conifer roots: effects on growth, life span and turnover. New Phytologist (2000) 147:87–103.[CrossRef][Web of Science]
Tricker PJ, Trewin H, Kull O, Clarkson GJJ, Eensalu E, Tallis MJ, Colella A, Doncaster CP, Sabatti M, Taylor G. Stomatal conductance and not stomatal density determines the long-term reduction in leaf transpiration of poplar in elevated CO2. Oecologia (2005) 143:652–660.[CrossRef][Web of Science][Medline]
Valentine H. Tree-growth models: derivations employing the pipe-model theory. Journal of Theoretical Biology (1985) 117:579–585.[CrossRef][Web of Science]
Valentine H. Estimation of the net primary productivity of even-aged stands with a carbon-allocation model. Ecological Modelling (1999) 122:139–149.[CrossRef][Web of Science]
Wong SC, Cowan IR, Farquhar GD. Stomatal conductance correlates with photosynthetic capacity. Nature (1979) 282:424–426.[CrossRef][Web of Science]
Wong SC, Cowan IR, Farquhar GD. Leaf conductance in relation to rate of CO2 assimilation. I. Influence of nitrogen nutrition, phosphorus nutrition, photon flux density, and ambient partial pressure of CO2 during ontogeny. Plant Physiology (1985b) 78:821–825.
Wong SC, Cowan IR, Farquhar GD. Leaf conductance in relation to rate of CO2 assimilation. III. Influences of water stress and photoinhibition. Plant Physiology (1985b) 78:830–834.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. A. Cernusak, K. Winter, J. Aranda, and B. L. Turner Conifers, Angiosperm Trees, and Lianas: Growth, Whole-Plant Water and Nitrogen Use Efficiency, and Stable Isotope Composition ({delta}13C and {delta}18O) of Seedlings Grown in a Tropical Environment Plant Physiology, September 1, 2008; 148(1): 642 - 659. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||