SEARCH

SEARCH BY CITATION

Keywords:

  • Acclimation;
  • beech;
  • birch;
  • canopy photosynthesis model;
  • leaf mass per unit area;
  • leaf nitrogen;
  • oak;
  • photosynthetic capacity;
  • Sitka spruce;
  • tropical rain forest

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

The observation of acclimation in leaf photosynthetic capacity to differences in growth irradiance has been widely used as support for a hypothesis that enables a simplification of some soil-vegetation-atmosphere transfer (SVAT) photosynthesis models. The acclimation hypothesis requires that relative leaf nitrogen concentration declines with relative irradiance from the top of a canopy to the bottom, in 1 : 1 proportion. In combination with a light transmission model it enables a simple estimate of the vertical profile in leaf nitrogen concentration (which is assumed to determine maximum carboxylation capacity), and in combination with estimates of the fraction of absorbed radiation it also leads to simple ‘big-leaf’ analytical solutions for canopy photosynthesis. We tested how forests deviate from this condition in five tree canopies, including four broadleaf stands, and one needle-leaf stand: a mixed-species tropical rain forest, oak (Quercus petraea (Matt.) Liebl), birch (Betula pendula Roth), beech (Fagus sylvatica L.) and Sitka spruce (Picea sitchensis (Bong.) Carr). Each canopy was studied when fully developed (mid-to-late summer for temperate stands). Irradiance (Q, µmol m−2 s−1) was measured for 20 d using quantum sensors placed throughout the vertical canopy profile. Measurements were made to obtain parameters from leaves adjacent to the radiation sensors: maximum carboxylation and electron transfer capacity (Va, Ja, µmol m−2 s−1), day respiration (Rda, µmol m−2 s−1), leaf nitrogen concentration (Nm, mg g−1) and leaf mass per unit area (La, g m−2).

Relative to upper-canopy values, Va declined linearly in 1 : 1 proportion with Na. Relative Va also declined linearly with relative Q, but with a significant intercept at zero irradiance (P < 0·01). This intercept was strongly related to La of the lowest leaves in each canopy (P < 0·01, r2 = 0·98, n= 5). For each canopy, daily lnQ was also linearly related with lnVa(P < 0·05), and the intercept was correlated with the value for photosynthetic capacity per unit nitrogen (PUN: Va/Na, µmol g−1 s−1) of the lowest leaves in each canopy (P < 0·05). Va was linearly related with La and Na(P < 0·01), but the slope of the Va : Na relationship varied widely among sites. Hence, whilst there was a unique Va : Na ratio in each stand, acclimation in Va to Q varied predictably with La of the lowest leaves in each canopy. The specific leaf area, Lm(cm2 g−1), of the canopy-bottom foliage was also found to predict carboxylation capacity (expressed on a mass basis; Vm, µmol g−1 s−1) at all sites (P < 0·01). These results invalidate the hypothesis of full acclimation to irradiance, but suggest that La and Lm of the most light-limited leaves in a canopy are widely applicable indicators of the distribution of photosynthetic capacity with height in forests.


Abbreviations
Amax

photon saturated leaf photosynthetic rate at ambient CO2 concentration (µmol m−2 s−1)

Acan

canopy photosynthetic rate (µmol m−2 s−1)

H

mean canopy height (m)

Hr

height in canopy relative to H

Ja

maximum electron transfer rate (conventionally Jmax) at 25 °C, on an area basis (µmol m−2 s−1)

k

canopy irradiance extinction coefficient, in Eqn 3 (dimensionless)

La

Lal, leaf mass per unit area (g m−2), La of lowest layer of leaves in a canopy

Lm, Lml

specific leaf area (cm2 g−1), Lm of lowest layer of leaves in a canopy

Nar

leafnitrogen concentration (g m−2), and Na relative to Na of highest measured leaves

Nm

leaf nitrogen concentration (g g−1)

PUN, PUNl

photosynthetic capacity per unit nitrogen (Va/Na (µmol C g−1 N s−1), and PUN of lowest leaves in a canopy

Q, Qi, Qr

Daily photosynthetic photon flux density averaged over 20 d (µmol m−2 s−1), above-canopy incident Q (µmol m−2 s−1), Q at height z in a canopy relative to Q received by the highest measured leaves, Qs, daily sum of photon flux density, averaged over 20 d (mol m−2 d−1)

Qdr

the diffuse radiation component of Qr, obtained from hemispherical photographs

Rda

daytime leaf respiration rate at 25 °C (µmol m−2 s−1)

Va, Vm, Var, Vml

maximum carboxylation rate (conventionally Vcmax) at 25 °C on an area basis (µmol m−2 s−1), and a mass basis (nmol g−1 s−1), Va relative to Va of the highest measured leaves, Vm of the lowest leaves in the canopy

α

is the apparent quantum efficiency, or initial slope of the J/Q response curve (mol mol−1)

θ

the convexity of the J/Q response curve (dimensionless).

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

The acclimation hypothesis

The combination of the infra-red gas analyser and controlled environment rooms in the early 1960s led rapidly to the demonstration by experiment that photosynthetic capacity of leaves acclimates to the light environment in which individual plants are growing (e.g. Bjorkman & Holmgren 1963; Milner & Hiesey 1964; Gauhl 1976). Furthermore, it was shown that the degree of acclimation, and whether the efficiency of one or both of the carboxylation or photochemical processes was affected, depends on the ecotype, reflecting the environment to which the genotype had become adapted (Björkman 1981).

Subsequently, it was demonstrated both in situ and using excised branches that the photosynthetic capacities of leaves in canopies also acclimate to the light environment in which the leaves are growing. In spruce forest, for example, the photon-saturated rate of photosynthesis of leaves low in canopies was shown to be much less than that of leaves receiving much more irradiance higher up in the canopy (Jarvis, James & Landsberg 1976; Jarvis & Sandford 1986) and the shoots and leaves differed in a number of photosynthetic and structural properties, including leaf mass per unit area, leaf chlorophyll and RUBP carboxylase-oxygenase (Rubisco) activity (e.g. Lewandowska & Jarvis 1977). Indeed, there seemed to be some proportionality between the light received and the photosynthetic capacity at a level in a canopy, of a similar kind to that seen with plants grown in particular light environments in growth rooms (e.g. Leverenz & Jarvis 1980). Similar observations on canopies of other tree and herbaceous species led to several model-based studies to determine whether the composition of particular canopies was optimum with respect to the absorption of photons by leaves and their utilization in photosynthesis, taking into account features of canopy structure such as spatial distribution of leaf area density, leaf age and leaf inclination angles (e.g. Field 1983; Hirose et al. 1988).

Meanwhile, use of the photosynthesis model developed by Farquhar et al. (1980) to analyse response functions of photosynthesis of both plants grown in different light environments and leaves growing at different levels in canopies led to the conclusion that acclimation of photosynthetic capacity to irradiance was primarily through a shift in the parameter for maximum carboxylation, Vcmax (referred to henceforth as Va, µmol m−2 s−1), although close stochiometry between Va and the parameter for maximum electron transport, Jmax (referred to henceforth as Ja, µmol m−2 s−1) is usually found (Wullschleger 1993).

The last piece of evidence is the many observations made over the past 15 years of a linear relationship between photosynthetic capacity and leaf nitrogen concentration (e.g. Field & Mooney 1986; Chazdon & Field 1987; Anten, Schieving & Werger 1995; Reich et al. 1998). This relationship has proved to be robust, for example embracing a wide range of species grown with or without the addition of fertiliser and in ambient or elevated atmospheric CO2 concentrations (Medlyn et al. 1999; Peterson et al. 1999). The relationship arises because the enzyme responsible for carboxylation, Rubisco, may comprise over 30% of the protein in photosynthesizing leaves (Evans 1989; Lawlor 1993). However, these studies have tended to focus on variation among sun leaves from different species or sites, rather than on leaves within a single canopy. Although variations in leaf nitrogen concentration have been shown to correlate with changes in irradiance (Q) in that nitrogen and photosynthetic capacities are low in leaves growing at low Q (e.g. De Jong & Doyle 1985, Kull & Niinemets 1998; Carswell et al. 2000), there are fewer observations of gradients of leaf nitrogen concentration with respect to Q in tree canopies (e.g. Hollinger 1996; Dang et al. 1997; Bond et al. 1999). Nonetheless, the association between Q, Va and leaf nitrogen has been developed through studies in a number of canopies and it has been shown that a theoretically optimal distribution of nitrogen concentration maximizes canopy photosynthesis when the nitrogen concentration closely follows the distribution of Q, approaching zero when Q does (Field 1983; Chen et al. 1993, Kull & Jarvis 1995; Kruijt et al. 1997).

This association between Q, Va and leaf nitrogen has been perceived by ecosystem modellers as a parameter-efficient way to model canopy photosynthesis in soil-vegetation-atmosphere transfer (SVAT) models intended for incorporation into General Circulation Models (e.g. Sellers et al. 1992). In principle, provided that acclimation to Q is complete, if the photon-saturated rate of photosynthesis, or the leaf nitrogen concentration, is known at one level in the canopy, conventionally at the top, the rate of photosynthesis can readily be found at any level, given an appropriate light transmission model, and thus integrated canopy photosynthesis may be derived (e.g. Thornley & Johnson 1990; Sellers et al. 1992; Kull & Jarvis 1995; Kruijt et al. 1997; De Pury & Farquhar 1997). This idea has also been used to show that in such an ‘optimised’ canopy, photosynthesis may be predicted from Q absorption alone, as can leaf area index, given nitrogen availability (Haxeltine & Prentice 1996; Dewar 1996). It has become known as the acclimation hypothesis.

The initial objective of the investigation presented here was to test the assumption that acclimation could be regarded as complete, or ‘perfect’, in several different forest canopies, because there were indications that in some canopies this may be far from the case (e.g. Hollinger 1996; Kull & Niinemets 1998). The acclimation hypothesis requires that Va, determined by leaf nitrogen concentration, linearly covaries with the Q absorbed by a leaf, such that Va for any leaf in the canopy declines relative to the canopy-top value of Va in direct proportion to the Q absorbed by that leaf, also expressed as relative to the canopy-top value of absorbed Q. For complete acclimation, the relationship between relative Va and relative Q is therefore expected to be 1 : 1, and, crucially, to pass through the origin. If the relationship is other than this, additional parameters are required to calculate integrated canopy photosynthetic capacity, or site-specific descriptions must be employed (e.g. Sinoquet et al. 2001).

In this study, data sets from five forest canopies are analysed to test these predictions, and to assess to what extent photosynthetic capacity is uniquely related to Q or to other key variables such as leaf mass per unit area, across widely differing species, sites and vertical positions within each canopy. The five forest canopies comprise four broadleaf and one conifer stand from three geographically separate regions (central Scotland, southern England and the Amazon basin, Brazil). Relationships among vertical profiles in leaf Va, leaf nitrogen concentration and leaf mass per unit area are considered, together with the relationships between these parameters and Q within the canopies. The linearity of the in-canopy decline in both relative and absolute values of Va with Q is then tested, and the extent to which this decline can be predicted in different stands from simpler measurements of leaves at the bottom of the canopy is examined.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

Sites

Oak (Quercus petraea (Matt.) Liebl), birch (Betula pendula Roth), beech (Fagus sylvatica L.) and Sitka spruce (Picea sitchensis (Bong.) Carr.) were studied at sites in the UK. The oak stand was in England near Alice Holt, Surrey (52°09′ N, 0°52′ W); it was 65 years old, and the canopy was 20–25 m tall, with a leaf area index (LAI, m2 m−2) ≈ 6. The other UK sites were in Scotland. The birch and spruce stands were separate 13-year-old experimental plantations at Glencorse Mains, near Penicuik, Lothian (55°31′ N, 3°12′ W). The birch stand was 5–6 m high, with LAI ≈ 2·5 and the spruce stand was 9–10 m high, with an LAI ≈ 7. The beech stand was also planted, in Dalkeith Park, Dalkeith, Lothian (55°31′ N, 3°11′ W). Its age was estimated at 70 years, and the canopy height was between 24 and 28 m, with LAI ≈ 4. The rain forest site (henceforth referred to as ‘tropical’) was in the Cuieiras Reserve, 60 km north of Manaus in Central Amazonia, Brazil (2°35′ S, 60°06′ W). The forest was apparently undisturbed, had a canopy height of 30–40 m, with LAI ≈ 6 (Malhi et al. 1998) and had a high species richness (∼250 species ha−1). Through-canopy towers were used to gain access to leaves in the vertical profile of each canopy. Measurements were made at the tropical site in November 1996, and are described in more detail in Carswell et al. (2000). For the UK sites, measurements were taken in August 1996 (oak) 1997 (birch and spruce) and 1998 (beech), after each deciduous canopy had fully opened and the leaves were mature.

Radiation measurements

Pre-calibrated sensors (Mackam, Livingston, UK) designed to measure photosynthetic photon flux density Q (µmol m−2 s−1) were placed throughout the vertical profile of each canopy. The sensors were attached to poles extending horizontally from the tower and positioned to represent as closely as possible the radiation environment of the leaves used for gas exchange measurement. The sensors were aligned to vertical using a spirit level and, where possible, more than one sensor was used per height, with relatively more sensors placed in the top third of the canopy in order to account for the large point-to-point variation in Q at this height. Each sensor profile was made up of a minimum of five levels, from the bottom of the canopy to the uppermost point where leaves were accessible; the total number of sensors and measurement levels used in each canopy is specified in Table 1. An additional sensor was positioned above the canopy to measure incoming radiation. Q was measured at 1 s intervals using a Campbell CR10 datalogger in combination with an AM32 multiplexer (Campbell, UK) and a Delta-T deltalogger (Delta-T, Cambridge, UK), and recorded as 10 min averages. In all stands data were recorded continuously for 4–6 weeks during the period of measurement and for approximately one week before the gas exchange measurements were made. Additional measurements were made of the radiation environment of a subset of individual leaves or shoots in the spruce, birch and beech canopies: hemispherical photographs were taken of the canopy above specific leaves using a Nikkor 7·5 mm S3HP fisheye lens and Kodachrome 200 ASA positive film. Photographs were taken under uniform diffuse light conditions using standard exposure and analysis procedures (Clearwater et al. 1999). The diffuse radiation component was calculated using image-analysis software (Optimas 5·1; Clearwater et al. 1999) and used to represent the diffuse radiation received by each leaf, Qd.

Table 1.  Characteristics of five forest canopies. k is the extinction coefficient determining the exponential decline in photosynthetic photon flux density with depth into each canopy, obtained by fitting Eqn 3 (see Methods) to relative measures (0–1) of photon flux density and depth. n= number of sensors used; r2 is the coefficient of determination of the model fit; Levels is the number of separate heights at which Q was measured in each canopy, with the measurement heights (m) specified parenthetically; LAI is leaf area index (m2 m−2) and H is the height of each canopy near the access point (m)
CanopyLAIHknLevelsr2
Beech4245·70 87 [1,8,10,14,16,18,20]0·95
Birch2·5 63·27115 [1,3,5·8]0·96
Oak6235·46 96 [2,6,10,14,16,18,20]0·83
Tropical6323·53176 [1,8,16,20,24]0·94
Spruce7 94·08115 [2·5,5,7·8]0·93

Gas exchange measurements and parameter estimation

In the broadleaf canopies, measurements were made on fully expanded and apparently non-senescing leaves. In the spruce canopy, one-year-old shoots were selected for the measurements. Gas exchange measurements were made using portable infra-red gas analysers with proprietory leaf cuvettes and light sources, which were carefully cross-checked for consistency in measurement of both the photosynthetic response to changes in Q and changes in CO2 concentration (CIRAS-1, PP Systems, Hitchin, UK; Licor 6400, Licor, Lincoln, Nebraska, USA; LCA3, ADC, Hoddesdon, UK). The chamber used for measuring spruce shoot gas exchange was lined on the lower side with reflective plastic to facilitate light saturation of all needles at high Q (Barton 1997). Three to 20 leaves (or shoots) were chosen per measurement height in each canopy, with more leaves chosen higher in the canopy where leaf-to-leaf variation was largest. The number of heights at which leaves could physically be accessed varied among the canopies: for the birch and spruce stands three levels were used, whereas in beech, oak and tropical 5–7 levels were used. For the birch, spruce and beech canopies some of the gas exchange measurements were made on the subset of leaves chosen for the measurement of Qd using hemispherical photography. During measurement of leaf gas exchange the humidity of the air entering the cuvette was held close to ambient, and by controlling the incident Q and CO2 concentration within the chamber, the responses of photosynthesis (A, µmol m−2 s−1) to Q (µmol m−2 s−1) and to the inferred CO2 concentration in the leaf air space (Ci, µmol mol−1) were recorded. The parameters for the A-Q response curves of each leaf were determined separately and then used to obtain the parameters of the photosynthetic response to Ci. Parameters of the Farquhar et al. (1980) photosynthesis model, corrected to 25 °C, were fitted to the data, using coefficients given by de Pury & Farquhar (1997). The two principal equations in this photosynthesis model are:

  • image(1)
  • image(2)

where: Rda is the apparent daytime respiration rate (µmol m−2 s−1); Va, the maximum carboxylation rate (µmol m−2 s−1); Ja, the maximum rate of the electron transfer, J (µmol m−2 s−1); θ is the convexity of the J/Q response curve (Eqn 2); α is the apparent quantum efficiency, or initial slope of the J/Q response curve (Eqn 2); Γ* is the CO2 compensation point in the absence of mitochondrial respiration (µmol mol−1); Km is the Michaelis-Menten coefficient for enzyme activity of Rubisco (µmol mol−1).

For the A-Q responses, leaves were subjected to saturating CO2 concentration and semi-steady-state values of A were measured at increasing photon flux densities from ∼10 µmol m−2 s−1 to saturating flux densities of Q (800–2000 µmol m−2 s−1). For each curve, a minimum of three points was obtained at low (< 50 µmol m−2 s−1), medium and saturating Q. In order to estimate α and θ, A was converted to J, using the second term of Eqn 1. The apparent quantum efficiency, α, and day respiration rate, Rda, were obtained by fitting a linear regression through J or A values, respectively, at Q < 50 µmol m−2 s−1 (α= slope of J-Q, Rda= intercept of A-Q). θ was then fitted by least-squares using Eqn 2 to the A-J responses. To determine the A-Ci response, leaves were subjected to a predetermined saturating Q, and the CO2 concentration in the chamber was changed step-wise from ambient concentration (360–400 µmol m−2 s−1) down to ∼25 µmol mol−1 and then, after returning to ambient concentrations, up to a saturating CO2 concentration at 1400–2000 µmol mol−1. In the oak canopy A/Ci responses were not obtained at the 2 and 6 m levels. The parameters Va and Ja were fitted to the data by least squares using Eqns 1 and 2 with values of Rda, θ and α fixed as determined from the A-Q data.

Foliar nitrogen concentration and leaf mass per area

Nitrogen concentration was determined on the leaves selected for photosynthesis measurements. For all broadleaf species, leaf discs of known area were used for nutrient analysis, and for the Sitka spruce, projected area of individual needles was determined using 15 harvested needles per sample (Barton 1997) and calculating shoot needle area by multiplying by the number of needles in each section of measured shoot. Foliage was dried at 70 °C to constant mass, and then finely ground and acid-peroxide digested before determination by colorimetric analysis (Grimshaw et al. 1989). Leaf mass per area (La, g m−2), or its reciprocal, specific leaf area (Lm, cm2 g−1) was also measured for individual leaves, and leaf nitrogen content was expressed on a mass (Nm, g g−1) and area (Na, g m−2) basis for comparison with photosynthesis parameters, also expressed on a mass (Vm, Jm, µmol g−1 s−1) and area (Va, Ja, µmol m−2 s−1) basis. Individual leaf-level gas exchange parameters were paired with their cognate nitrogen and La determinations in all canopies except for the oak, where cognate pairs were not available and mean values (± 1 SE) per level were used.

Data analysis

In an initial intercanopy comparison, an exponential model was fitted to the radiation measurements in each stand (Eqn 3). As each canopy was different in size and experienced different amounts of incident irradiance, in-canopy measurements of Q were expressed relative to incident above-canopy Q, Qi. The mean daily sum of Q, Qs, was obtained from all sensors for all daylight hours over a 20-d period beginning approximately one week before the gas exchange measurements. Equation 3 was then fitted to the data, estimating in-canopy Q relative to Qi, at a depth, D, relative to the maximum depth of the canopy, Dc.

inline image 3

where k is an extinction coefficient, D is depth into the canopy (m), and the subscript ‘r’ refers to a value between 0 and 1 relative to Dc or Qi at each site (i.e. relative depth, Dr, and relative irradiance, Qr)

Relationships among leaf-level measurements in each canopy were then examined using linear regressions, relating photosynthetic capacity (e.g. Va) to La or Na.

inline image 4

where X is La or Na, a is the slope of the regression and b the intercept.

All relationships among photosynthetic capacity (e.g. Va), Na and Q were tested using measured values of Q or Qr, rather than the modelled values from Eqn 3. To account for differences in day length at each site, contrasts were initially analysed between the mean daily sum of Q (Qs) and measured foliar variables using natural log-transformed values for each level within each canopy, and for the combined-canopy dataset. We then examined the relationships between these functions and the photosynthetic capacity per unit nitrogen (Va/Na: µmol g−1 s−1) of leaves at the base of each canopy, PUNl.

Crown geometry prevented access to true canopy-top leaves in some of the stands (tropical, beech and spruce), so leaves slightly below the top were used as upper reference points for testing the acclimation hypothesis. To obtain photosynthetic parameter values relative to those of the uppermost leaves, the mean parameter value at height z in the canopy (e.g. Vaz) and the mean incident radiation measurement cognate with that height (Qz) were normalized relative to the mean value obtained for the leaves at the top level in each measured profile, at height h (Vah and Qh), such that relative Q, Qr = Qz/Qh and relative Va, Var = Vaz/Vah. The regression model in Eqn 4 was used to test the acclimation hypothesis by examining the relationships between the decline in relative photosynthetic capacity (Var) and the decline in relative irradiance, Qr. Where data were available for more than three canopy levels, a quadratic term was introduced into Eqn 4 to examine the significance of any non-linearity in the response of Var to Qr, where Var = aX + bX2 + c. For the diffuse component radiation measurements (Qd, derived from hemispherical photographs), individual Qd and leaf parameter values were used. In this case, the leaf with the highest Qd was used as a point reference for calculating the relative Qd and leaf parameter values, Qdr and Var. The analysis using Qr and Var was repeated using leaf nitrogen concentrations, and the overall results compared with La determinations in each canopy.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

Variation with height in radiation intensity, nitrogen and physiological capacity

Photosynthetic photon flux density, Q, declined significantly with height in all canopies (P < 0·0001). The light extinction model (Eqn 3) fitted the measured profiles in Q with r2 > 0·8 for all stands and with the extinction coefficient, k, ranging from 3·2 to 5·7, increasing in the following order: birch, tropical, spruce, oak, beech (Table 1).

In all five canopies La was significantly larger at the top of the canopy than at the bottom (t-tests between canopy-top and canopy-bottom leaves, n = 8–23, P < 0·01 for all canopies; Fig. 1a) and increased among the five canopies in the following order: beech, birch, oak, tropical and spruce. Na also increased significantly with height in all canopies (t-tests, n = 8–23, P < 0·001 for all canopies; Fig. 1b). Nm was lower in the spruce canopy (9–11 mg g−1) than in the other four (19–39 mg g−1), and its variation with height was smaller and less consistent than Na in all five stands (Fig. 1c).

image

Figure 1. Variation with height above the ground in (a) leaf mass per area (La, g m−2), (b) area-based leaf nitrogen content (Na, g m−2) and (c) mass-based leaf nitrogen content (Nm, mg g−1) in five tree canopies. Height within in each canopy is expressed relative to the mean maximum height of each canopy. The number of leaves measured per canopy was: birch, 31; oak, 50; beech, 19; tropical, 42; spruce, 23; with 2–8 leaves measured per level in the lower canopy and 6–20 per level in the upper canopy. Error bars are ± 1 SE.

Download figure to PowerPoint

An increase from the lowest to the highest leaves of more than 100% in the maximum photosynthetic rate was found in every stand. The biochemical photosynthetic parameters, Va and Ja were significantly correlated in all canopies (P < 0·01) and the slope of the regression of Ja on Va varied from 1·6 (birch) to 2·8 (oak) with r2 > 0·80 in all cases (Fig. 2). All three fitted photosynthetic parameters, Va, Ja and Rda, increased significantly with height (t-tests between canopy-top and canopy-bottom leaves, n= 8–23, P < 0·01; Fig. 3a–c), but the pattern was not always significant when they were expressed on a mass basis (Vm, Jm and Rdm, respectively; Fig. 3d, Vm data shown). The relationship between Va and La or Na was different in each canopy (Fig. 4a & b, Table 2), with lower slopes in spruce, oak and tropical, and higher slopes in beech and birch. Na was a significant predictor of Va for all canopies combined (P < 0·01) but the coefficient of determination was low relative to the individual canopy relationships (Table 2). Both specific leaf area (Lm, m2 g−1, the reciprocal of La) and Nm were significant predictors of Vm for the combined-site data, but the relationships were weaker than with area-based expressions, and only in the tropical canopy was Nm significantly related to Vm (Fig. 4c & d; regression results not shown).

image

Figure 2. The relationship between Va (maximum carboxylation capacity, µmol m−2 s−1) and Ja (maximum electron transport capacity, µmol m−2 s−1) with its slope (± 1 SE) for all five canopies, and the combined dataset. Data are for individual leaves, except for oak, where means (± 1 SE) per canopy level are given.

Download figure to PowerPoint

image

Figure 3. Variation with height above the ground in photosynthesis parameters for five tree canopies. (a) Va (maximum carboxylation capacity, µmol m−2 s−1), (b) Ja (maximum electron transport capacity, µmol m−2 s−1), (c) Rda (leaf respiration, µmol m−2 s−1) and (d) Vm (maximum carboxylation capacity, µmol g−1 s−1). Error bars are ± 1 SE.

Download figure to PowerPoint

image

Figure 4. Area- and mass-based relationships for five tree canopies between leaf carboxylation capacity (Va, µmol m−2 s−1, Vm, µmol g−1 s−1) and leaf mass per area (La, g m−2) or specific leaf area (Lm, cm2 g−1), or leaf nitrogen content (Na, g m−2, Nm, mg g−1). (a) The dashed lines are fitted regressions of Va on La for each canopy. (b) The dashed lines are fitted regressions of Va on Na for each canopy; the solid line is the regression for the combined dataset (see Table 3 for results). (c) Variation in Vm with Lm. (d) Variation in Vm with Nm.

Download figure to PowerPoint

Table 2.  Regression results for the relationships between leaf mass per area (La, g m−2) and maximum carboxylation capacity (Va, µmol m−2 s−1), and between leaf nitrogen per unit area (Na, g m−2) and Va. Regression model: y= aX + b, where y=Va, and X=La or Na; n= total number of leaves at all heights, except Va for oak, where mean values for each of five canopy heights were recorded
CanopyXa (± SE)b (± SE)r2Pn
BeechLa 0·91 (0·12)  7·20 (7·62)0·77<0·0001 19
BirchLa 1·46 (0·20)−42·5 (19·06)0·66<0·0001 31
OakLa 0·57 (0·11)−22·82 (12·04)0·91  0·013  5
TropicalLa 0·30 (0·08) 2·31 (7·98)0·24  0·001 42
SpruceLa 0·27 (0·04)−20·48 (0·27)0·72<0·0001 23
BeechNa31·02 (5·06) 13·13 (8·42)0·69<0·0001 19
BirchNa43·33 (6·43)−20·53 (18·03)0·61<0·0001 31
OakNa18·87 (2·64)−0·13 (6·02)0·94  0·006  5
TropicalNa11·41 (2·20)  0·98 (5·81)0·40<0·0001 42
SpruceNa14·54 (5·12)  5·51 (11·26)0·28  0·009 23
CombinedNa22·60 (3·65)  1·18 (8·94)0·25<0·0001120

Relative leaf photosynthetic capacity, nitrogen and relative radiation environment

Given the consistent relationship between Va and Ja (Fig. 2), Va was used as the measure of photosynthetic capacity for subsequent analyses with the radiation data. For the leaf-scale measurements of Qd in the beech, birch and spruce stands, the variation in Var with Qdr was curvilinear (Fig. 5a). The increase in Var with Qdr tended to be approximately linear for Qdr < 0·2, and to saturate above that value, giving a non-zero intercept between 0·1 and 0·3 at Qdr = 0.

image

Figure 5. The decline in relative carboxylation capacity (Va, µmol m−2 s−1) and relative nitrogen concentration (Na, g m−2) with relative incident photon flux density. All values are expressed relative to (i.e. a fraction of) the upper-canopy measurements, and are denoted: Var, Nar, Qr and Qdr; the 1 : 1 line is drawn in (dotted). (a) Individual leaf variation in Var with the relative diffuse component of incident radiation, Qdr. Variation in (b) Var with Qr, (c) Nar with Qr; (d) Var with Nar; the solid line is the combined-site regression. (a) Shows individual leaf data in birch, spruce and beech, expressed relative to the uppermost leaf; (b–d) show mean values at each measurement height relative to the mean for the uppermost level, in all five canopies.

Download figure to PowerPoint

For the broadleaf stands, Qr was positively correlated with Var. The linear regression of Var on Qr for the combined dataset was significant, with individual intercepts significantly greater than zero (P < 0·01, Table 3). Strong correlations were obtained for all canopies, although low n reduced regression significance in the birch and spruce (r2 > 0·86 for all; Fig. 5b, Table 3). The birch and spruce data could not be analysed using the quadratic regression formula, as n= 3, but in the three other canopy datasets the non-linear term was non-significant (P > 0·2; n = 5–7), and the coefficient of determination of the regression was no higher than in the linear regressions. These results show that among the broadleaf canopies acclimation in photosynthetic capacity to Q did not exhibit 1 : 1 proportionality. Instead, although the relative decline in capacity with Q tended to be linear, a significant positive Var intercept (P < 0·05) was found for all four broadleaf canopies at Qr = 0, with values between 56 and 72% of Var for the uppermost leaves, and 64% for the combined-canopy dataset. Acclimation occurred to less than 50% of the maximum for the spruce stand and the decline in Var with respect to Qr appeared slightly non-linear.

Table 3.  Regression results for the relationships between relative photon flux density, Qr, and relative carboxylation capacity, Var, and relative leaf nitrogen concentration, Nar. Regression model: y= aX + b, where y=Var or Nar and X=Qr; n is the number of levels (heights) within each canopy at which cognate measurements of Q and Va were made. See Methods for full definitions of Qr, Var, Nar and canopy identity. All linear regressions passed through or very close (<1 SE) to the point where Var and Qr = 1. A quadratic regression model (y = aX + bX2 + c was used to test for linearity in the beech, oak and tropical canopies (where n > 3), but all results yielded a non-significant non-linear term
Canopyya (± SE)b (± SE)r2Pn
BeechVar0·70 (0·07)0·28 (0·03)0·96  0·0010 7
BirchVar0·69 (0·11)0·32 (0·06)0·98  0·1000 3
OakVar0·62 (0·13)0·40 (0·06)0·89  0·0170 5
TropicalVar0·62 (0·03)0·37 (0·02)0·99  0·0003 5
SpruceVar0·53 (0·21)0·52 (0·13)0·86  0·2400 3
CombinedVar0·65 (0·05)0·36 (0·23)0·89<0·000123
BeechNar0·65 (0·13)0·27 (0·06)0·83  0·0040 7
BirchNar0·53 (0·18)0·49 (0·11)0·89  0·2100 3
OakNar0·59 (0·06)0·41 (0·03)0·97<0·0001 5
TropicalNar0·51 (0·09)0·46 (0·03)0·97  0·0020 5
SpruceNar0·23 (0·15)0·73 (0·09)0·72  0·3500 3
CombinedNar0·55 (0·07)0·42 (0·04)0·73<0·000123

The relative nitrogen content, Nar, also declined with Qr in all the broadleaf canopies, although the relationships were slightly weaker than for Var (Table 3, Fig. 5c). In the spruce stand, Na was higher at the bottom of the canopy than in the middle (cf. Fig. 1c), resulting in a high Nar intercept (0·73) at Qr = 0, and a relatively shallow slope. Overall, Nar was linearly related to Qr for the combined site dataset, with an intercept at Qr = 0 significantly larger than zero (0·41; Table 3). Na declined relative to Q by between 27 and 73% of the maximum value, with a combined-canopy value of 59%. Both Var and Nar declined with Qr at apparently similar rates, and hence the overall relationship between them was almost 1 : 1, with a slight tendency for the retention of nitrogen over photosynthetic capacity at lower Q in some individual canopies, e.g. spruce, and the reverse tendency in others, e.g. beech (Fig. 5d).

Leaf photosynthetic capacity, average radiation environment, photosynthetic capacity per unit nitrogen and leaf mass per area

Va was also strongly related to absolute daily sums of incident radiation (Qs) using a logarithmic transform: lnVa was significantly related to lnQs for the combined-site dataset (P < 0·001; r2 = 0·43; Table 4, Fig. 6a). Significant regressions were also obtained for individual canopies, and more variance was explained in these individual regressions than for the combined dataset (P = 0·04–0·0001; r2 = 0·78–0·99; Table 4, Fig. 6a). The slope of the regression for birch was higher than for spruce, whereas for the beech, oak and tropical canopies the slopes were similar and intermediate between that found for birch and spruce (Table 4, Fig. 6a). However, the intercepts did not group in the same way: they were similar for the birch, oak and tropical canopies, but higher for beech and spruce (Table 4). These intercepts were significantly related to the photosynthetic capacity per unit nitrogen for foliage at the bottom of each canopy (PUNl: Va/Na, µmol C s−1 g−1 N; P < 0·05, r2 = 0·88; Fig. 6b). In two further analyses on canopy-bottom foliage, mass-based photosynthetic capacity (Vm) was strongly related to specific leaf area, Lm (i.e. Vml on Lml, Fig. 7a), and the zero intercepts on the regressions of Var on Qr (Fig. 5b, Table 3) were strongly related to the leaf mass per area of the canopy-bottom foliage (those leaves lowest in the canopy, Lal, Fig. 7b), but not to Nm. Neither relationship in Fig. 7 changed significantly if the spruce datum was removed, or when values were added from other studies of temperate and tropical forests where comparable measurements have been made on both needle- and broadleaf species (Meir 1996; Bond et al. 1999; Raulier et al. 1999; DeLucia & Thomas 2000; Wilson, Baldocchi & Hanson 2000; Warren & Adams 2001) (Fig. 7a & b; r2 = 0·95 and 0·98, respectively, this study, P < 0·01; including additional data, r2 = 0·90 and 0·92, P < 0·001). There was a strong relationship between Na (but not Nm) of the canopy-bottom leaves (Nal) and Vml (r2 = 0·82; data not shown), but a stepwise multiple linear regression analysis of Vml on Nal and Lml showed Lml to be the only significant coefficient (P < 0·001), explaining more of the variation in Vml alone (r2 = 0·95) than in combination with Nal (r2 = 0·90)

Table 4.  Regression analysis results for the relationship between the daily sum of photosynthetic photon flux density averaged for the 20 d measurement period, Qs(mol  m−2 d−1), and carboxylation capacity, Vamol m−2 s−1). Regression model: lnVa= a lnQs+ b; n is the number of levels (heights) in each canopy at which cognate measurements of Q and Va were made
Canopya (± SE)b (± SE)r2Pn
Beech0·34 (0·03)3·81 (0·04)0·95<0·001 7
Birch0·45 (0·02)3·36 (0·04)0·99  0·029 3
Oak0·33 (0·10)3·05 (0·20)0·78  0·046 5
Tropical0·33 (0·05)2·99 (0·07)0·93  0·007 5
Spruce0·19 (0·00)3·26 (0·00)0·99  0·007 3
Combined0·23 (0·06)3·45 (0·09)0·43<0·00123
image

Figure 6. Variation in carboxylation capacity (Va, µmol m−2 s−1) with incident photon flux density (Qs, mol m−2 d−1) and photosynthetic capacity per unit nitrogen of the lowest leaves in the canopy (PUNl, Va/Na, µmol g−1 s−1). (a) Variation in Va with the mean daily sum of Q, Qs, for the 20 d measurement period: data are natural log-transformed values for each measurement height in each of the five canopies. All the regression lines are significant (see Table 4). Thick solid line, combined-canopy dataset; thin solid line, birch; short-dashed, oak; medium-dashed, beech; dash-dot, tropical; and long-dashed, spruce. (b) Relationship for all five canopies between the regression intercept in (a) for each canopy (lnVa at lnQs= 0; Table 4), and PUNl.

Download figure to PowerPoint

image

Figure 7. (a) Relationship between carboxylation capacity and specific leaf area of the lowest leaves in each canopy (Vml, µmol g−1 s−1, on Lml, cm2 g−1). Vml for oak estimated from nitrogen concentration; also plotted are mean species or site values for Pinus pinaster (Warren & Adams 2001), a semideciduous tropical rain forest understorey (Meir 1996), broadleaf species in a loblolly plantation understorey (DeLucia & Thomas 2000) and a deciduous hardwood forest understorey (Wilson et al. 2000). (b) Relationship between the maximum potential (relative) acclimation in each canopy (Var at Qr = 0), and the leaf mass per area of the lowest leaves (Lal, g m−2) in each canopy. Also plotted are values derived from Bond et al. (1999) for Pseudotsuga menziesii, Pinus ponderosa and Tsuga heterophylla (+), and from Raulier et al. (1999) for Acer saccharum where relative irradiance < 0·8 (▿). Solid regression lines are fitted for the five canopies in this study and the dotted regression lines are fitted through the data from all; the regressions are statistically identical, results are given with standard errors in parentheses.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

Changes in photosynthetic capacity with height and leaf nitrogen concentration

The wide variation in La and Na among the different canopies (Fig. 1) is consistent with the overall ranges described by Schulze et al. (1994) and Reich et al. (1998); this in-canopy variation in Na can often be as great as across a whole biome. The values for Va and Ja are also similar to reported values for these parameters when differences in Na and temperature are considered (Fig. 2; Rey 1996; Barton 1997; Medlyn et al. 1999), and the slope of the regression between Ja and Va (1·65 ± 0·1 SE, Fig. 3) is very close to the cross-species fitted value of 1·64 previously observed in a survey of 109 species (Wullschleger 1993), indicating the maintenance of tight photosynthetic stoichiometry in all stands.

Our results, as for several previous studies, show that variation in foliage properties may be explained to a large extent by differences in incident photon flux density (Björkman 1981). Height within a canopy is strongly correlated with Q, and La, Na and the fitted photosynthetic parameters (Va, Ja, Rda) increase significantly with height in all stands. Variation in Nm and Vm with height, however, is much smaller, and is not significant in some canopies. This is consistent with the well-studied outcome that variability in the composition and construction of foliage, such as in La, mediates a significant part of the photosynthetic acclimation response to changes in Q within natural canopies (e.g. Field 1983; De Jong & Doyle 1985; Chazdon & Field 1987; Hirose et al. 1988; Ellsworth & Reich 1993; Hollinger 1996; Mitchell et al. 1999; Bond et al. 1999; Meir et al. 2001).

Although the changes in La and Va are strongly related in each stand, there is no universal interspecific pattern. Changes in La within each canopy dominate the signal in Na and in consequence Na tends to be a better predictor of photosynthetic capacity than Nm for individual stands, and for the whole dataset. The lack of a universal nitrogen-capacity relationship, and the poor predictive power of Nm, contrasts with studies highlighting strong interspecific relationships using Nm, La or Lm, in combination or alone, to predict Amax or Va (e.g. Field & Mooney 1986; Reich et al. 1998; Medlyn et al. 1999). However, these reports have tended to focus on sun leaves (at high or atmospheric CO2 concentration) and/or to consider a wide range of growth forms (e.g. herbs, shrubs, conifers, broadleaf trees). Variations in leaf structure, nitrogen concentration and photosynthetic capacity in these reports have been principally derived from differences in species, biome or ambient CO2 concentration, and therefore differ from patterns of variation in the same foliar variables for a single canopy resulting from in-canopy differences in the radiation environment. However, for canopy-bottom leaves only, we do observe a species- and biome-wide mass-based relationship between specific leaf area and photosynthetic capacity (i.e. Vml on Lml; Fig. 7a). This widely applicable relationship emerges because constraints on leaf construction and nitrogen use at the canopy-bottom are likely to be strongly influenced by the effects of the need to capture a key limiting resource, irradiance, on the foliar carbon balance. The weaker relationships between photosynthetic capacity and leaf nitrogen in these canopy-bottom leaves suggests that species differences in nitrogen use in deep shade are less important to net carbon assimilation than are differences in leaf mass per unit area.

Despite differences amongst the canopies in the Va-Na relationship, the relative decline in Na with Va (Var versus Nar) is strongly linear and close to 1 : 1 across all stands (Fig. 5b), albeit with a tendency in spruce for the foliage to favour the retention of nitrogen at the expense of capacity in the lower canopy. Proportionality between relative values of Na and Amax has been observed in other forest canopy profiles (Dang et al. 1997; Bond et al. 1999) and this indicates that a dominant parameter separating different stands is the absolute ratio between Na and Amax (or Va) of the sun leaves. With this ratio established, changes within a canopy in Na relative to Amax (or Va) closely follow a 1 : 1 relationship down to a critical Amax (Va) value below which leaf construction and maintenance cannot be achieved.

Is there a unique relationship between incident Q and Va?

The strong linear regressions between lnVa and lnQs in all canopies indicate that in-canopy changes with height in both variables are approximately exponential, as implied by simple radiative transfer schemes (e.g. Sellers et al. 1992). The combined-canopy regression is significant, but would perform poorly if used to model individual canopies. In general, variation among canopies in the slope of this relationship could reflect differences in the proportions of the diffuse and direct components of incident radiation at different sites, as well as differences in species and site ecology. The variation observed in our results is consistent with species ecology, although more data are needed to confirm the differences (Fig. 6a; Table 4; r2 = 0·78–0·99). Birch, a pioneer species, may be expected to invest more heavily in Va per unit (daily) radiation load, and consistent with this it had the largest slope, whereas the beech, oak, tropical and spruce stands were all climax-stage species with smaller increases in Va per unit radiation load, the spruce having a smaller slope than the three broadleaves (Table 4). We might also expect to see differences in this relationship between sites with widely varying proportions of direct and diffuse light.

Changes in photosynthetic capacity per unit nitrogen (PUN, Va/Na) are influenced by changes in La and the differential allocation of nitrogen to carboxylation, electron transfer and light harvesting processes, and at low Q this allocation is influenced by both mass- and area-based considerations relating to light capture and metabolic chemistry (e.g. Niinemets & Tenhunen 1997). PUN of the lowest leaves in each canopy (PUNl) is a significant predictor of the intercept of the regression of lnVa on lnQ for all five stands (Fig. 6b). Although PUN in sun leaves may differ widely among species (Poorter & Evans 1998), our data suggest that PUN of the most light-limited leaves in a canopy is more strongly determined by the minimum achievable La associated with the extinction of irradiance within the canopy than by the minimum Va for a particular species or stand.

Can we predict acclimation in Va to variation in Q with height in the canopy?

Partial acclimation in Na and Va to Q occurs in all stands in this study: using the plots of relative values, a significant non-zero intercept in Nar and Var was found at Qr= 0 (Table 3). There was a curvilinear acclimation response in Var to changes in the diffuse radiation component (Qdr; Fig. 5a). This could have resulted from deeper canopy penetration of diffuse over direct beam radiation. For the same incident flux (Qi), diffuse radiation will penetrate to the middle canopy better than direct radiation. This could have led to an underestimate of the radiant flux within the canopy. The effect of using measured all-beam radiation would therefore be to straighten the curvature in Fig. 5a. This is what was observed when Qr (i.e. diffuse and direct radiation) was used instead of Qdr: a linear response was found for four stands, although the spruce data appeared marginally non-linear (Fig. 5b; Table 3).

The non-zero intercept in this relationship (Fig. 5b, Table 3) contradicts the prediction by canopy optimization theories of full photosynthetic acclimation to irradiance (e.g. Sellers et al. 1992; Kruijt et al. 1997). Instead, acclimation is proportional to Qr in this dataset, but it varies among canopies in the extent to which each can acclimate. The extent of acclimation to irradiance is not strongly related to LAI, although for some canopies with higher LAI, the maximum degree of acclimation is slightly lower (cf. Fig. 7). The failure to fully acclimate to irradiance reflects differences in carbon requirements per unit nitrogen among the canopies, which are themselves likely to be a complex function of non-photosynthetic nitrogen requirements and physical or ecological interactions, including the effects of wind or herbivory (Mooney & Gulmon 1979). However, although full acclimation does not occur in any canopy, the result in Fig. 5(b) suggests that an improved, simplified estimate of canopy photosynthesis is feasible, and depends on how well the Var intercept at Qr = 0 can be estimated. That is, how well it is possible to predict the ‘maximum potential acclimation’ in Va to irradiance in a canopy. La of the lowest leaves in the canopy (Lal) is a strong predictor of this intercept for all stands, the regression remaining significant and almost identical after removal of the spruce canopy or inclusion of values from four other broad- and needle-leaf stands (Fig. 7b).

This apparently robust relationship suggests the basis for a widely applicable simple whole-canopy photosynthesis scheme based on leaf measurements at the bottom of a canopy, assuming a linear response in Var to Qr, as observed up to the highest measurable leaves in our broadleaf species (r2 for the linear function = 0·86–0·99, Table 3). Most comparable studies have used Amax as an estimate of Va, which is valid if light-saturation of photosynthesis occurs in the studied species. Raulier et al. (1999), studying an Acer saccharum canopy and using an indirect estimate of relative irradiance with Amax, also showed a linear response, with some saturation in Amax for the small amount of foliage (∼5%) experiencing a relative irradiance greater than 0·7–0·8. Bond et al. (1999) also encountered weak saturation in Amax at high relative irradiances in conifers, as found here for spruce. Apparent saturation may be an artefact of steep leaf angle distributions near the canopy top constraining the absorption of incident radiation by upper-canopy leaves (and hence effectively reducing Qr). A radiation-transfer model including detailed individual descriptions of canopy structure and incident radiation quality could be used to quantify this effect (e.g. Wang & Jarvis 1990), but in the context of simplified SVAT models which use the assumption of full acclimation, the effect could be represented empirically, if necessary, by modifying the linear model to a bi-modal one (linear, then constant) or by using a rectangular hyperbola.

Significant changes over time in the environment may cause the relationships described here to change, as La and Va are known to vary across strong seasonal cycles and differences in temperature (e.g. Dang et al. 1998; Wilson et al. 2000; Wirtz 2000). However, because La can change with photosynthetic activity it is also possible that Lal predicts maximum relative acclimation potential of a mature canopy throughout the growing season. Canopies almost certainly organize their vertical allocation of nitrogen and photosynthetic capacity in response to a series of metabolic requirements (Kull & Niinemets 1998) and ecological factors (Mooney & Gulmon 1979; Hollinger 1996), rather than solely to radiation gradients, but despite this complexity, the relative allocation of capacity within the canopy appears to be strongly predictable at the point where radiation is most limiting to leaf construction. This empirical result will require refinement, but suggests that the huge variation in photosynthetic characteristics amongst different forest stands is predictable based on easily obtainable measurements made from the ground.

Conclusions and implications

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

Our principal result is that acclimation to irradiance occurs in all the studied canopies, but that full acclimation does not. Although relative leaf photosynthetic capacity and nitrogen concentration (Var and Nar) tend to decline in 1 : 1 proportion with each other, Var and relative irradiance (Qr) do not. Instead, a significant positive intercept in Var at Qr = 0 is observed for all canopies, as implied for some other broad- and needle-leaf stands (Dang et al. 1997; Bond et al. 1999; Raulier et al. 1999). We show here that this intercept is strongly predictable among forests from the leaf mass per unit area of the lowest leaves in the canopy, Lal. We also show for canopy-bottom leaves that, when expressed on a mass basis, photosynthetic capacity (Vml) is strongly predictable from specific leaf area, Lml, and that the intercept of the response in Va to Q can be estimated from the photosynthetic capacity per unit nitrogen, PUNl (Figs 6 & 7). What are the implications of this for modelling canopy photosynthesis?

Simplified mechanistic models of acclimation along canopy light gradients predict similar patterns to those found in our data (Kull & Kruijt 1999; Dewar et al. 1998). These models provide a useful framework to explain why full acclimation does not usually occur, and although they do not make explicit predictions from La, the observation that Lal is a measure of canopy acclimation is also consistent with them. However, these models do not predict higher absolute photosynthetic capacity (at low Q) for a shade-tolerant species such as beech over that of a shallower pioneer species canopy, such as birch (Fig. 6). If variation in PUN as defined here (Va/Na) is included in such models, the higher Va in beech may be explicable in terms of greater nitrogen investment into Rubisco. Generalizing, it is likely that requirements to invest in non-photosynthetic material, including support structure within and outside the leaves, are the key, and these investments are likely to be related to ecological strategy. Mechanistic models of acclimation should be developed further, but the partitioning of nitrogen over structure and photosynthesis should be further incorporated into these models.

For simpler SVAT models, a more empirical approach is appropriate. Depending on canopy LAI and the proportion of the diffuse component in incident radiation, it may be possible to improve the predictive accuracy of a SVAT that assumes full (or no) acclimation. To do this, two approaches can be followed. In the first we propose that absolute capacity in canopy-bottom leaves (Vml) is estimated from Lml, or that Va in canopy-top sun leaves is obtained from literature-based regressions using Nm(or Nm and Lm). La of the lowest leaves, Lal, can then be used to predict the ‘maximum potential acclimation’ in Va, and hence the relative change in Va with irradiance throughout the canopy. This method assumes a linear Var-Qr relationship, as supported by our data, but slight saturation in Var at high Qr can be accounted for simply, by using a bimodal function of Var (linear, then constant), or alternatively by using a more complex canopy model and calculating differences in absorbed Q with height.

Bond et al. (1999) proposed the basis of a second method using similarity in the gradients of the log-linear relationship between Amax and incident radiation for three conifer canopies. However, this gradient is likely to change with a wider range of species or needle ages (Bond et al. 1999; Schoettle & Smith 1999; Fig. 6a), and canopy-top measurements are required to estimate the intercept for new stands. Our results suggest that PUNl may be used instead to estimate the absolute value of this intercept (in Va), but this relationship (Fig. 6b) is less well-defined than that based on normalized data (Fig. 7), and suffers from the errors associated with using log-transform regressions and the additional need to make in-situ physiological measurements.

One way to test whether Lal can be used to predict the integrated photosynthetic behaviour and composition of whole canopies is to examine the convexity in canopy photosynthetic light response curves (Acan-Qi), using canopy-scale measurements of gas exchange. In a completely acclimated canopy, light saturation of leaf photosynthesis at all heights theoretically occurs simultaneously at a particular value of Qi, hence causing the integrated canopy response curve to saturate at the same Qi and therefore to be highly convex. In a non-acclimated stand, photosynthetic light saturation within the canopy will occur at different incident irradiances, for example, lower leaves may remain light-limited at a Qi that is saturating for upper-canopy leaves. This results in the integrated Acan-Qi curve saturating more slowly, over a wider range of Qi, reflecting the different values of Qi at which different parts of the canopy saturate (Berry et al. 1997). Consequently, if acclimation is poor, represented by a large intercept in the Var-Qr relationship and by large Lal, the convexity in the Acan-Qi response should be low. Conversely, convexity should be high if the degree of acclimation is high and Lal is small. If Lal proves to be a robust predictor of the convexity of the Acan-Qi response, it will have useful implications for understanding and modelling terrestrial photosynthesis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References

We thank the following organizations for providing research facilities and support: INPA, Manaus, Brazil, including logistical support made available by the UK ODA project, ‘Abracos’; Forestry Commission, Alice Holt, UK; CEH, Edinburgh, UK, and Buccleuch Estates, Dalkeith, UK. We also thank L. Kruuk, J. Grace and two anonymous reviewers for helpful comments on an earlier draft of this paper, and Samantha Jackson and Veronica Finlayson for technical assistance. Funding was provided by the UK Natural Environment Research Council, grant number GR3/09732, and the Estonian Sciences Foundation, grant number 3775.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions and implications
  8. Acknowledgements
  9. References
  • Anten N.P.R., Schieving F. & Werger M.J.A. (1995) Patterns of light and nitrogen distribution in relation to whole canopy carbon gain in C-3 and C-4 monocotyledonous and dicotyledonous species. Oecologia 101, 504513.
  • Barton C.V.M. (1997) Effects of elevated atmospheric carbon dioxide concentration on growth and physiology of Sitka spruce (Picea sitchensis (Bong.) Carr.). PhD Thesis, University of Edinburgh, Edinburgh.
  • Berry J.A., Collatz G.J., Denning A.S., Colello G.D., Fu W., Grivet C., Randall D.A. & Sellers P.J. (1997) SiB2, a model for simulation of biological processes within a climate model. In Scaling Up from Cell to Landscape (eds P. van Gardingen, G. Foody & P. Curran), pp. 347369. SEB seminar series, 63. Cambridge University Press, Cambridge.
  • Björkman O. (1981) Responses to different quantum flux densities. In Encyclopaedia of Plant Physiology (New Series), Vol. 12A (eds O. L. Lange, P. S. Nobel, C. B. Osmond & H. Ziegler), pp. 57107. Springer-Verlag, Berlin.
  • Bjorkman O. & Holmgren P. (1963) Adaptability of the photosynthetic apparatus to light intensity in ecotypes from exposed and shaded habitats. Physiologia Plantarum 16, 889914.
  • Bond B.J., Farnsworth B.T., Coulombe R.A. & Winner W.E. (1999) Foliage physiology and biochemistry in response to light gradients in conifers with varying shade tolerance. Oecologia 120, 183192.
  • Carswell F.E., Meir P., Wandelli E.V., Bonates L.C.M., Kruijt B., Barbosa E.M., Nobre A.D., Grace J. & Jarvis P.G. (2000) Photosynthetic capacity in a central Amazonian rain forest. Tree Physiology 20, 179186.
  • Chazdon R.L. & Field C.B. (1987) Determinants of photosynthetic capacity in 6 rain forest species. Oecologia 73, 222230.
  • Chen, J.J., Reynolds P., Harley, X. & Tenhunen T. (1993) Coordination theory of leaf nitrogen distribution in a canopy. Oecologia 93, 6369.
  • Clearwater M.J., Nifinluri T. & Van Gardingen P.R. (1999) Forest fire smoke and a test of hemispherical photography for predicting understorey light in Bornean tropical rain forest. Agricultural and Forest Meteorology 97, 129139.DOI: 10.1016/s0168-1923(99)00058-1
  • Dang Q.L., Margolis H.A. & Collatz G.J. (1998) Parameterization and testing of a coupled photosynthesis stomatal conductance model for boreal trees. Tree Physiology 18, 141153.
  • Dang Q.L., Margolis H.A., Sy M., Coyea M.R., Collatz G.J. & Walthall C.L. (1997) Profiles of photosynthetically active radiation, nitrogen and photosynthetic capacity in the boreal forest: Implications for scaling from leaf to canopy. Journal of Geophysical Research – Atmospheres 102, 2884528859.
  • De Jong T.M. & Doyle J.F. (1985) Seasonal relationships between leaf nitrogen content (photosynthetic capacity) and leaf canopy light exposure in peach (Prunus persica). Plant, Cell and Environment 8, 701706.
  • De Pury D.G.G. & Farquhar G.D. (1997) Simple scaling of photosynthesis from leaves to canopies without the errors of big-leaf models. Plant Cell and Environment 20, 537557.
  • DeLucia E.H. & Thomas R.B. (2000) Photosynthetic responses to CO2 enrichment of four hardwood species in a forest understory. Oecologia 122, 1119.
  • Dewar R.C. (1996) The correlation between plant growth and intercepted radiation: an interpretation in terms of optimal plant nitrogen content. Annals of Botany 78, 125136.
  • Dewar R.C., Medlyn B.E. & McMurtrie R.E. (1998) A mechanistic analysis of light and carbon use efficiencies. Plant, Cell and Environment 21, 573588.
  • Ellsworth D.S. & Reich P.B. (1993) Canopy structure and vertical patterns of photosynthesis and related leaf traits in a deciduous forest. Oecologia 96, 169178.
  • Evans J.R. (1989) Photosynthesis and nitrogen relationships in leaves of C-3 plants. Oecologia 78, 919.
  • Farquhar G.D., Von Cammerer S. & Berry J.A. (1980) A biochemical model of photosynthetic CO2 assimilation in leaves of C3 species. Planta 149, 7890.
  • Field C.B. (1983) Allocating leaf nitrogen for the maximisation of carbon gain: leaf age as a control on the allocation program. Oecologia 56, 341347.
  • Field C.B. & Mooney H.A. (1986) The photosynthesis-nitrogen relationship in wild plants. In The Economy of Plant Form and Function (ed. T. J. Givnish), pp. 2555. Cambridge University Press, Cambridge.
  • Gauhl E. (1976) Photosynthetic response to varying light intensity in ecotypes of Solanum dulcamara L. from shaded and exposed habitats. Oecologia 22, 275286.
  • Grimshaw H.M., Allen S.E. & Parkinson J.A. (1989) Nutrient elements. In Chemical Analysis of Ecological Materials(ed. S. E. Allen), pp. 81159. Blackwell Scientific, Oxford.
  • Haxeltine A. & Prentice I.C. (1996) A general model for the light-use efficiency of primary production. Functional Ecology 10, 551561.
  • Hirose T., Werger M.J.A., Pons T.L. & Van Rheenen J.W.A. (1988) Canopy structure and leaf nitrogen distribution in a stand of Lysimachia vulgaris L. as influenced by stand density. Oecologia 77, 145150.
  • Hollinger D.Y. (1996) Optimality and nitrogen allocation in a tree canopy. Tree Physiology 16, 627634.
  • Jarvis P.G., James B.G. & Landsberg J.J. (1976) Coniferous forest. In Vegetation and Atmosphere (ed. J. L. Monteith), pp. 171204. Academic Press, London.
  • Jarvis P.G. & Sandford A.P. (1986) Temperate forests. In Photosynthesis in Contrasting Environments (eds N. R. Baker, S. P. Long), pp. 200228. Elsevier, Oxford.
  • Kruijt B., Ongeri S. & Jarvis P.G. (1997) Scaling PAR absorption, photosynthesis and transpiration from leaves to canopy. In Scaling Up from Cell to Landscape (eds P. van Gardingen, G. Foody & P. Curran), pp. 79104. SEB seminar series, 63. Cambridge University Press, Cambridge.
  • Kull O. & Jarvis P.G. (1995) The role of nitrogen in a simple scheme to scale-up photosynthesis from leaf to canopy. Plant, Cell and Environment 18, 11741182.
  • Kull O. & Kruijt B. (1999) Acclimation of photosynthesis to light: a mechanistic approach. Functional Ecology 13, 2436.DOI: 10.1046/j.1365-2435.1999.00292.x
  • Kull O. & Niinemets U. (1998) Distribution of leaf photosynthetic properties in tree canopies: comparison of species with different shade tolerance. Functional Ecology 12, 472479.DOI: 10.1046/j.1365-2435.1998.00214.x
  • Lawlor D.W. (1993) Photosynthesis: Molecular, Physiological and Environmental Processes. Longman Scientific, London.
  • Leverenz J.W. & Jarvis P.G. (1980) Photosynthesis in Sitka spruce (Picea sitchensis (BONG.) CARR.). X. Acclimation to quantum flux density within and between trees. Journal of Applied Ecology 17, 697708.
  • Lewandowska M. & Jarvis P.G. (1977) Changes in chlorophyll and carotenoid content, specific leaf area and dry weight fraction in Sitka spruce, in response to shading and season. New Phytologist 79, 247256.
  • Malhi Y., Nobre A.D., Grace J., Kruijt B., Pereira M.G.P., Culf A. & Scott S. (1998) Carbon dioxide transfer over a central Amazonian rain forest. Journal of Geophysical Research – Atmospheres 103, 31, 59331, 612.
  • Medlyn B.E., Badeck F.W., De Pury D.G.G., et al. (1999) Effects of elevated [CO2] on photosynthesis in European forest species: a meta-analysis of model parameters. Plant, Cell and Environment 22, 14751495.DOI: 10.1046/j.1365-3040.1999.00523.x
  • Meir P. (1996) The exchange of carbon dioxide in tropical forest. PhD Thesis, University of Edinburgh, Edinburgh.
  • Meir P., Grace J. & Miranda A.C. (2001) Leaf respiration in two tropical rain forests: constraints on physiology by phosphorus, nitrogen, and temperature. Functional Ecology 15, 378387.
  • Milner H.W. & Hiesey W.M. (1964) Photosynthesis in climatic races of Mimulus. 1. Effect of light intensity and temperature on rate. Plant Physiology 39, 208213.
  • Mitchell K.A., Bolstad P.V. & Vose J.M. (1999) Interspecific and environmentally induced variation in foliar dark respiration among eighteen southeastern deciduous tree species. Tree Physiology 19, 861870.
  • Mooney H.A. & Gulmon S.L. (1979) Environmental and evolutionary constraints on photosynthetic characteristics of higher plants. In Topics in Plant Population Biology (eds O. T. Solbrig, S. Jain, G. B. Johnson & P. H. Raven), pp. 316337. Columbia University Press, New York.
  • Niinemets U. & Tenhunen J.D. (1997) A model separating leaf structural and physiological effects on carbon gain along light gradients for the shade-tolerant species Acer saccharum. Plant Cell and Environment 20, 845866.
  • Peterson A.G., Ball J.T., Luo Y.Q., et al. (1999) The photosynthesis leaf nitrogen relationship at ambient and elevated atmospheric carbon dioxide: a meta-analysis. Global Change Biology 5, 331346.DOI: 10.1046/j.1365-2486.1999.00234.x
  • Poorter H. & Evans J.R. (1998) Photosynthetic nitrogen-use efficiency of species that differ inherently in specific leaf area. Oecologia 116, 2637.DOI: 10.1007/s004420050560
  • Raulier F., Bernier P.Y. & Ung C.H. (1999) Canopy photosynthesis of sugar maple (Acer saccharum): comparing big-leaf and multilayer extrapolations of leaf-level measurements. Tree Physiology 19, 407420.
  • Reich P.B., Ellsworth D.S. & Walters M.B. (1998) Leaf structure (specific leaf area) modulates photosynthesis-nitrogen relations: evidence from within and across species and functional groups. Functional Ecology 12, 948958.
  • Rey A. (1996) Response of young birch trees to increasing atmospheric carbon dioxide concentrations. Phd Thesis, University of Barcelona, Barcelona.
  • Schoettle A.W. & Smith W.K. (1999) Interrelationships among light, photosynthesis and nitrogen in the crown of mature Pinus contorta ssp. latifolia. Tree Physiology 19, 1322.
  • Schulze E.-D., Kelliher F.M., Korner C., Lloyd J. & Leuning R. (1994) Relationships among maximum stomatal conductance, ecosystem surface conductance, carbon assimilation rate and plant nitrogen nutrition: a global ecology scaling exercise. Annual Review of Ecology and Systematics 25, 629660.
  • Sellers P.J., Berry J.A., Collatz G.J., Field C.B. & Hall F.G. (1992) Canopy reflectance, photosynthesis, and transpiration. 3. A reanalysis using improved leaf models and a new canopy integration scheme. Remote Sensing of Environment 42, 187216.
  • Sinoquet H., Le Roux X., Adam B., Ameglio T. & Daudet F.A. (2001) RATP: a model for simulating the spatial distribution of radiation absorption, transpiration and photosynthesis within canopies: application to an isolated tree crown. Plant, Cell and Environment 24, 395406.DOI: 10.1046/j.1365-3040.2001.00694.x
  • Thornley J.H.M. & Johnson I.R. (1990) Plant and Crop Modelling. Clarendon Press, Oxford.
  • Wang Y.P. & Jarvis P.G. (1990) Influence of crown structural properties on PAR absorption, photosynthesis and transpiration in Sitka spruce – application of a model (Maestro). Tree Physiology 7, 297316.
  • Warren C.R. & Adams M.A. (2001) Distribution of N, Rubisco and photosynthesis in Pinus pinaster and acclimation to light. Plant, Cell and Environment 24, 597609.
  • Wilson K.B., Baldocchi D.D. & Hanson P.J. (2000) Spatial and seasonal variability of photosynthetic parameters and their relationship to leaf nitrogen in a deciduous forest. Tree Physiology 20, 565578.
  • Wirtz K.W. (2000) Simulating the dynamics of leaf physiology and morphology with an extended optimality approach. Annals of Botany 86, 753764.DOI: 10.1006/anbo.2000.1230
  • Wullschleger S.D. (1993) Biochemical limitations to carbon assimilation in C-3 plants – a retrospective analysis of the A/Ci curves from 109 species. Journal of Experimental Botany 44, 907920.

Received 11 June 2001; received in revised form 10 September 2001; accepted for publication 10 September 2001