Ann Bot 2012 Lac Ha Pelle 1149 57

7/31/2019 Ann Bot 2012 Lac Ha Pelle 1149 57

1/17

Annals of Botany 109: 1149 1157, 2012

doi:10.1093/aob/mcs032, available online at ww w .aob.oxfordjournals.org

Interspecific prediction ofphotosynthetic light response curves using specificleaf mass and leaf nitrogen content: effects of differences in soil fertility and

growth irradiance

Pierre-Philippe Lachapelle and Bill Shipley*

De partement de biologie, Universite de Sherbrooke and Centre detude de la foret (CEF), Sherbrooke (Qc), Canada J1K

2R9* For correspondence. E-mail Bill.S hipley@USherb rooke.ca

Received: 14 October 2011 Returned for revision: 16 December 2011 Accepted: 18 January 2012 Published electronically: 22 March 2012

Background and Aims Previous work has shown that the entire photosynthetic light response curve, based onboth Mitscherlich and Michaelis Menten functions, could be predicted in an interspecific context through allo-metric relations linking the parameters of these functions to two static leaf traits: leaf nitrogen (N) content andleaf mass per area (LMA). This paper describes to what extent these allometric relations are robust to changes insoil fertility and the growth irradiance of the plants.

Methods Plants of 25 herbaceous species were grown under controlled conditions in factorial combinations oflow/high soil fertility and low/high growth irradiance. Net photosynthetic rates per unit dry mass were measured at

light intensities ranging from 0 to 700 mmol m22

s21

photosynthetically active radiation (PAR).

Key Results The differing growth environments induced large changes in N, LMA and in each of the parameterestimates of the Mitscherlich and Michaelis Menten functions. However, the differing growth environmentsinduced only small (although significant) changes in the allometric relationships linking N and LMA to the para-meters of the two functions. As a result, 88 % (Mitcherlich) and 89 % (Michaelis Menten) of the observed net

photosynthetic rates over the full range of light intensities (0 700 mmol m22

s21

PAR) and across all four growthenvironments could be predicted using only N and LMA using the same allometric relations. Conclusions These results suggest the possibility of predicting net photosynthetic rates in nature across speciesover the full range of light intensities using readily available data.

Key words: Quantum yield, Mitscherlich curve, Michaelis Menten curve, leaf respiration rate, maximumphotosynthetic rate, SLA, LMA.

IN TRO D UCTI

O N

The photosynthetic light response (PLR) curve describesthe relationship between leaf net photosynthetic rate and thephotosynthetically active photon flux density arriving at theleaf surface. PLR is important in predicting carbon fixationin nature because variation in the light environment of a leafis one of the most important factors affecting photosyntheticrates. Several mathematical functions have been used to de-scribe the PLR curve; three popular ones are the Michaelis Menten, Mitscherlich and the non-rectangular hyperbolicequations. However, a PLR curve requires the estimation ofseveral free parameters and this limits its usefulness in eco-logical contexts involving many species growing in heteroge-neous environments.

Recently, Marino et al. (2010) showed that the parametersof entire PLR curves of individual leaves could be predictedin an interspecific context from two static leaf traits: leaf nitro-gen content (N, % dry mass) and specific leaf mass (LMA, g

cm22

). As these two static leaf traits are routinelymeasured and have been compiled in large trait databases suchas TRY (Kattge et al., 2011) and GLOPNET (Wright et al.,2004b) these results open the possibility of estimatingsuch PLRcurves for every species in an entire community. This possibil-

ity is important because some models of gas exchange for
http://www.aob.oxfordjournals.org/mailto:[email protected]:[email protected]:[email protected]:y@USherbrookmailto:y@USherbrookmailto:y@USherbrookmailto:[email protected]:y@USherbrookhttp://www.aob.oxfordjournals.org/


2/17

entire forests (e.g. Ollinger et al., 2008; Pilegaard et al.,2011) and individual-based models of plant growth such asLIGNUM (Perttunen et al., 1996, 2001; Lo et al., 2001) andJABOWA (Bugmann, 2001) include such PLR curves.

Such models generally ignore species- and environment-specific variation in the parameter values of PLR curvesbecause the empirical estimation of the free parameters foreach species and environment is too time-consuming. Nonethe less, it is well known that such curves vary greatly bothbetween species and between individuals of the same species

growing in different environments (Lambers et al., 1998)and so ignoring interspecific differences and environmental

effects reduces the precision and realism of these large-scalemodels. If one could obtain empirical estimates of the modelparameters of PLR curves for each species and environmenfrom static measurements of leaf nitrogen content and specificleaf mass, then this obstacle would be overcome.

The results of Marino et al. (2010) were obtained from 25herbaceous species grown under constant and controlledgrowth conditions and from 40 tree species growing in thefield. In the first case there was no variation in the growth con-ditions while in the second case the growth conditions were

variable but unknown. Both N and LMA are known to varywith both the light and the soil fertility environment inwhich the plant grows (Meziane and Shipley, 1999, 2001),

# The Author 2012. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved.

For Permissions, please email:[email protected]
mailto:[email protected]:[email protected]


3/17

126 Lachapelle & Shipley Predicting photosynthetic light response curves

and the shape of the PLR curve also varies with these growthconditions (Lambers et al., 1998). Therefore, although theresults of Marino et al. (2010) suggest a way of predictingPLR curves between species, they are incomplete for thispurpose because the effects of different growth conditions onthe empirical trait parameter relationships are still

unknown. If the slopes and intercepts of the empirical trait parameter relationships are sensitive to the growthenvironment of the plant, and especially if this sensitivityis species-specific, then they cannot be used in practice in acommunity context because quantifying the environment iseven more challenging than measuring the PLR curve itself. If,on the other hand, changing growth conditions simplymoves a species up or down along a common trait parameter regression line, or if the environmentally inducedchanges in slopes and intercepts are small, then one cansimply measure N and LMA without having to explicitlyquantify the environment itself.

The purpose of this paper is therefore to determine thedegree of sensitivity of the empirical trait PLR curve

param-eter relationships in an interspecific context with respect totwo key environmental variables during the growth of theplant: soil fertility and irradiance levels. We ask two specificques- tions: (1) Do changes in soil fertility and irradianceduring growth affect the values of N, LMA and theparameters of two common PLR curves (the Mitscherlichand Michaelis Menten functions)? (2) Do changes in soilfertility and irradiance during growth affect the allometricrelationships linking each parameter to N and LMA and, if so,how sensitive are these relationships to changes in the growthenvironment?

METHODS

Eight plants of each of 25 herbaceous species (Table 1)were grown individually from seed in 1-litre pots placedin a growth chamber in one of four environments. Therewere two replicate plants per species per environment.Temperatures were 22 8C for the 14-h day and 18 8C forthe10-h night. Plants were kept well-watered at all times. Thefour experimental environments consisted of a factorial com-bination of high/low levels of soil fertility and high/lowlevels of instantaneous photosynthetically active photon flux.The low level of soil fertility consisted of 600 mL of sandand 200 mL of a commercial potting soil. The high fertility

level was produced by adding 2 g of a slow-release 14-14-14

(N-P-K) fertilizer (Smartcote#) to the low-fertility soilmixture. Preliminary experiments using a wider range of fertil-izer additions and using three species (Rubdeckia hirta,Chenopodium polyspermum, Chicorium intybus) showed thatthe chosen high fertility level produced maximal growthrates and that the chosen low fertility level produced signifi-cantly reduced growth while still allowing the plants tosurvive. None the less, six species in the low-fertility/high-irradiance and 12 species in the low-fertility/low-irradianceenvironments (Table 1) did not grow sufficiently to allowmeasurements. The two irradiance levels during growth [100

or 400 mmol m22

s21

photosynthetically active radiation(PAR)] were obtained by growing the plants in two differentgrowth chambers, resulting in a four-fold change in daily

quantum flux (5.04 and 20.016 mol m22 d

21). Pots receiving

each level of fertility were randomly placed in each growth


4/17

Lachapelle & Shipley Predicting photosynthetic light response curves 1151TAB LE 1. Species used, their taxonomic affiliation and their

growth habit (annual/biennial/perennial); also shown are the

experimental treatments in which the plants remained too smallto provide measurements

Species Family Habit Missing*

Bidens frondosa L. Asteraceae Annual

Centaurea cyanus L. Asteraceae AnnualRudbeckia hirta L. Asteraceae Annual lf Chenopodium polyspermum L. Chenopodiaceae AnnualGaleopsis tetrahit L. Laminaceae Annual Lf, lf Beckmannia syzigachne (Steud.) Poaceae Annual lf FernaldEchium vulgare L. Boraginaceae BiennialOenothera biennis L. Onagraceae BiennialCichorium intybus L. Asteraceae PerennialLeontodon autumnalis L. Asteraceae PerennialSolidago canadensis L. Asteraceae Perennial Lf, lf Silene cucubalus Wibel Caryophyllaceae Perennial lf Prunella vulgaris L. Lamiaceae Perennial lf Lythrum salicaria L. Lytheraceae Perennial Lf Althaea officinalis L. Malvaceae PerennialEpilobium angustifolium L. Onagraceae Perennial Lf, lf Epilobium glandulosum Lehn. Oragraceae Perennial

Plantago major L. Plantaginaeae Perennial Lf, lf Agrostis gigantea Roth Poaceae PerennialDactylis glomerata L. Poaceae PerennialKoeleria macrantha (Ledeb.) Poaceae Perennial Lf, lf Schult.Phleum pratense L. Poaceae Perennial lf Rumex obtusifolius L. Polygonaceae Perennial lf Potentilla recta L. Rosaceae PerennialHyoscyamus niger L. Solanaceae Perennial lf

* Lf high irradiance, infertile soil, lf low irradiance, infertile soil.

chamber and their positions were randomly changed onceper week. Note that the two irradiance levels are confounded

with the two growth chambers; however, because thesegrowth chambers were identical except for the irradiancelevel, we are confident that any differences can be ascribedto the irradiance treatment and not to other unknownchamber effects.

Measurements were taken before floral buds formed andafter the plant had sufficient leaf mass to allow measurementof foliar nitrogen; this took 42 d on average but variedbetween 30 and 120 d with plants growing in infertile soiltaking longer. Plants were watered to field capacity theevening before taking measurements and were allowed to ha-bituate in the laboratory for 30 min before measurementsbegan.

Gas exchange was measured separately on two leaves per

plant following the protocol of Marino et al. (2010) using aQubit infrared gas analyser system (Qubit Systems, http://qubit s y s tems.com/). CO2 concentration in the air entering theleaf cuvette was maintained at 380 mmol L

21and leaf tem-

perature was constantly monitored using a thermistor. PARwas supplied using red diodes. The leaf was initially kept inthe dark for 15 min and then respiration rate was measuredafter which we began increasing light intensity while allowingthe leaf to habituate to each increasing intensity for 15 minbefore taking measurements. For all but 13 leaves theirradiance levels used to measure the PLR curves were 75,
http://qubitsystems.com/http://qubitsystems.com/http://qubitsystems.com/http://qubitsystems.com/http://qubitsystems.com/


5/17

2

I

150, 175, 500 and 700 mmol m22

s21

. For the remaining 13leaves ( from eight different species), we usedirradiance levels of 75, 150, 175, 250 and 500 mmol m

22

s21

because there was evidence of photoinhibition at 700mmol m

22s21

. We varied the number and intensities ofthese levels in a series of preliminary experiments and found

that these pro- vided the most precise estimates of theparameters of the PLR curves while minimizing theirnumber.

Immediately following the gas exchange measurements thetwo measured leaves per plant were cut and their projectedlamina areas, both within the cuvette and the entire lamina,were measured using the WinFolia image analysis system(ww w .regent.qc.ca). These leaf laminas were then dried for aminimum of 48 h at 70 8C, weighed and LMA was calculated.These two leaves were added to the remaining dried leaves(minus petioles) on the plant, ground using a ball mill and

version 2.12.2 (R Development Core Team, 2008). Table 2lists the parameters and their units; note that gas exchangevalues are expressed per unit dry mass of the leaf, not perunit leaf area, following Wright et al. (2004b) and Marinoet al. (2010). Empirical allometric relationships betweeneach of these parameters and the two leaf traits (N, LMA)

were obtained from multiple linear least-squared regressionsusing ln-transformed values. Statistical tests of the effects ofthe treatments on the leaf traits and parameter estimates werebased on a mixed-model ANOVA using the lme function ofthe nlme package of R with the random effects being individ-ual leaves nested within individual plants nested withinspecies.

qLC P( I LC P )

the leaf nitrogen content was measured using a Macro CN ana-lyser (Elementar Americas, Inc.; http: //www.chnos.com/).

The empirical measurements of net photosynthetic rate,A(I ), and irradiance (I ) of each leaf were then fitted tothe

A(I) = Amax 1 e

Gmax I

Amax (1)

Mitscherlich (eqn 1) and the Michaelis Menten (eqn 2)equations (below), each having three free parameters,using the non-linear least squares regression function(nls) of R

A(I) =K

Rd (2)+

TAB LE 2. Two measured leaf traits (N, LMA) and six

estimated parameters of the photosynthetic light response

curves

V ariable/p a r ameter Symbol Units

Specific leaf mass LMA g cm22

Leaf nitrogen content N % dry mass

RESU LT S

Changes in soil fertility induced significant (P

,

0.01)plastic changes in both leaf traits (LMA, N) and in all PLRpara- meters. Changes in the growth irradiance inducedsignificant plastic changes in all variables except the darkrespiration

Photosynthetically active photonflux

I mmol PAR m22

s21 rate. The effects of soil fertility and growth irradiance were

additive for all variables except for N. Supp lementary Data

Net maximum photosynthetic rate Amax mmol CO2 g21

s21

Table S1 gives the values of these variables for each of the

Light compensation point LCP mmol PAR m22

s21

leaves in the experiment and Table 3 gives the means of

Apparent quantum yield at f qLCP mmol CO2 m2

g21

mmol21

PAR these variables per growth environment.

Gross maximum photosyntheticrate

Irradiance level at 1

Gmax mmol CO2 g21

s21

22 21

There were strong patterns of covariation between LMA, Nand the three free parameters of the Mitscherlich and

Michaelis Menten curves. The first two principal componentsGmax K mmol PAR m s

21 21Leaf respiration rate Rd mmol CO2 g s

The first three parameters (Amax, LCP, qLCP) belong to the Mitscherlichequation while the second set of three parameters (Gmax, K, Rd) belong to theMichaelis Menten equation.

explain 95 % (Mitscherlich, Fig. 1) and 87 % (Michaelis Menten, Fig. 2) of the total variation among the five variablesThe separation of the species mean values measured in thefour environments shows the effect of the environmental

TAB LE 3. Means (back-transformed from logarithms), calculated over all leaves of 25 herbaceous species, of two leaf traits (LMA

N), three parameters of the Mitscherlich photosynthesis light response curve (Amax, LCP, q(LCP)] and three parameters of the

Michaelis Menten photosynthesis light response curve (Gmax, K, Rd)
http://www.regent.qc.ca/http://www.chnos.com/http://aob.oxfordjournals.org/lookup/suppl/doi:10.1093/aob/mcs032/-/DC1http://aob.oxfordjournals.org/lookup/suppl/doi:10.1093/aob/mcs032/-/DC1http://aob.oxfordjournals.org/lookup/suppl/doi:10.1093/aob/mcs032/-/DC1http://www.regent.qc.ca/http://www.chnos.com/http://aob.oxfordjournals.org/lookup/suppl/doi:10.1093/aob/mcs032/-/DC1


6/17

Growth irradiance Soil fertility Amax q(LCP) LCP Gmax K Rd LMA N

100 Low 0.23 0.002 9.0 0.30 93.9 0.02 0.0025 1.5

100 High 0.54 0.003 12.2 0.76 223.2 0.04 0.0017 5.2400 Low 0.18 0.001 13.5 0.24 129.5 0.02 0.0039 1.4400 High 0.45 0.002 22.5 0.68 321.5 0.04 0.0029 4.2100 Low + high 0.40 0.003 10.0 0.54 154.6 0.03 0.0019 3.6400 Low + high 0.27 0.001 16.7 0.40 209.7 0.03 0.0033 2.7100+ 400 Low 0.18 0.001 10.3 0.23 107.0 0.02 0.0033 1.4100+ 400 High 0.46 0.002 15.2 0.67 247.0 0.04 0.0022 4.6

See Table 1 for definitions and measurement units. Growth irradiance (mmol PAR m22

s21

) refers to the irradiance level during growth and soil fertilityrefers to the soil nutrient level during growth.


7/17

PCAcomponent2(28%)

Log(LMA)

PCAcomponent2(36%)

10 05 00 05 10

6Low irradiance, low fertility

Low irradiance, high fertility

High irradiance, low fertility4 High irradiance, high fertility

15

10

Mitscherlich PLR curve

There were strong relationships between the PLR curve par-ameter estimates and N and LMA (Fig. 4). Amax was predictedwell in all growth environments except when irradiance was

2

Log(qLCP)

0

Log(Amax)

Log(N)

2

4

Log(LCP)

Log(LMA)

05

00

05

10

05 00 05 10

6 Low irradiance, low fertility

Low irradiance, high fertility

High irradiance, low fertility

4 High irradiance, high fertility

2

Log(K)

10

05

4 2 0 2 4

PCA component 1 (67 %)

FIG . 1 . Principal component biplot of the first two principal axes showing

0

Log(Gmax)

00

two leaf traits, leaf mass per area (LMA) and leaf nitrogen content (N) 2

along with three parameter estimates of the Mitscherlich photosynthetic lightresponse curve: maximum net photosynthetic rate (Amax) per unit leaf drymass, the light compensation point (LCP) and the quantum yield at the LCP(qLCP). The first and second axes account for 67 and 28 % of the variation, re- 4

Log(N) Log(LMA)

Log(Rd)

05

spectively. Each point is a species mean calculated separately in each of fourgrowth environments and broken arrows show the effect of each growth

variable (soil fertility and growth irradiance). See Table 1 for units.

4 2 0 2 4

PCA component 1 (51 %)

10

variables on plastic changes in these species means.Increasing the growth irradiance primarily increased LMAand decreased the quantum yield at the light compensationpoint and the irradiance producing half of the maximumphotosynthetic rate. Increasing the soil fertility primarilyincreased the leaf nitrogen content and the maximum photo-synthetic rate. The light compensation point and the dark res-

FIG . 2 . Information as in Fig. 1 except that the photosynthetic light responsecurve parameters are those of the Michaelis Menten function: maximum grossphotosynthetic rate (Gmax) per unit dry mass, the daytime respiration rate (Rd)

per unit dry mass, and the photosynthetically active photon flux density atwhich the gross photosynthetic rate is half of its maximum (K, i.e. the half

saturation constant). See Table 1 for units.

piration rate were affected approximately equally by bothenvironmental variables.

There was a negative interspecific correlation between LMAand N within each growth environment based on a mixed-

model regression that takes into account both intraspecificand interspecific variation. The high-irradiance environmentresulted in a higher LMA for a given value of N while thehigh soil fertility resulted in a higher N for a given LMA(Fig. 3). There were also significant differences in the slopesbetween the growth environments with shallower negativeslopes occurring when only one of the two growth environ-ments was limiting.

The allometric slopes linking N and LMA to each of theparameters of the PLR curves generally differed significantly(P , 0.05) between the light and soil environments,as shown by mixed-model ANCOVAs relating each log10-transformed PLR curve parameter to the log10-transformedvalues of N and LMA. Table 4 summarizes these allometric

mixed-model regressions. These regressions were obtained

bothsepa

ratelyforeachgrowthenvironm

en

t(en

vironment-

20

22

24

26

28

30

00


8/17

02

04Log(N)

Low

irra

diance

, low fertilityLowirradiance,high fertilityHighirradiance,low fertilityHighirradiance,high fertility

06 08 10

specific) and after combining the observations of all fourgrowth environments together (general) and ignoring the dif-ferent growth environments.

FIG . 3 . Species mean values of leaf mass per area vs. leaf nitrogen content infour different growth environments. Lines are the regression fits based on a

mixed-model taking into account interspecific and intraspecific variation.


9/17

2

Lachapelle&ShipleyPredictingphotosyntheticlightresponsecurves

1153

TAB LE 4. Partial slopes (+ s.e.) associated with log10-transformed values of leaf nitrogen content (N) and leaf mass per area (LMA) of allometric mixed-model

regressions predicting the parameters of two photosynthetic light response curves (Michaelis Menten, Mitscherlich) based on individual leaves of 25herbaceous species grown in four different growth environments

Environment-speci fical l o metr ic eq u a tion

General allometricequation 10 0 s. e.general s.e.specific

Growth environment Parameter Intercept Log10[N] Log10LMA R2 Residual s.e. R2 Residual s.e.s.e.general

All environments combined Amax 2.14 (0.110) 0.65 (0.032) 0.52 (0.046) 0.82 0.119LCP 2.84 (0.184) 0.55 (0.054) 0.77 (0.077) 0.29 0.200

q(LCP) 5.32 (0.096) 0.06 (0.028) 0.98 (0.040) 0.79 0.104Gmax 1.74 (0.119) 0.77 (0.035) 0.39 (0.050) 0.80 0.131K 3.66 (0.158) 0.89 (0.046) 0.71 (0.066) 0.55 0.174Rd 2.46 (0.170) 0.60 (0.050) 0.23 (0.071) 0.52 0.187

High irradiance, fertile soil Amax 2.96 (0.192) 0.45 (0.116) 0.90 (0.88) 0.72 0.102 0.68 0.122 16.4LCP 2.12 (0.313) 0.65 (0.189) 0.47 (0.142) 0.12 0.166 0.08 0.190 12.6q(LCP) 5.32 (0.159) 0.00 (0.096) 1.00 (0.072) 0.75 0.084 0.75 0.084 0.0Gmax 2.69 (0.205) 0.58 (0.125) 0.83 (0.094) 0.69 0.112 0.64 0.135 17.0K 2.48 (0.269) 0.65 (0.163) 0.16 (0.124) 0.14 0.147 0.05 0.173 15.0Rd 3.11 (0.307) 0.61 (0.187) 0.51 (0.141) 0.36 0.168 0.34 0.188 10.6

High irradiance, infertile soil Amax 2.98 (0.299) 0.53 (0.123) 0.87 (0.128) 0.69 0.128 0.67 0.136 5.9LCP 1.16 (0.452) 0.04 (0.186) 0.03 (0.194) 0.00 0.193 0.00 0.214 9.8

q(LCP) 4.92 (0.300) 0.45 (0.123) 0.79 (0.129) 0.64 0.128 0.59 0.138 7.2Gmax 2.87 (0.303) 0.54 (0.125) 0.87 (0.130) 0.69 0.129 0.64 0.143 9.8K 1.78 (0.320) 0.10 (0.131) 0.12 (0.137) 0.02 0.137 0.00 0.178 23.0

Rd 3.79 (0.437) 0.51 (0.180) 0.79 (0.187) 0.48 0.186 0.43 0.201 7.5Low irradiance, fertile soil Amax 1.63 (0.230) 0.44 (0.139) 0.38 (0.092) 0.32 0.096 0.32 0.108 11.1

LCP 2.22 (0.394) 0.14 (0.238) 0.38 (0.157) 0.07 0.164 0.03 0.183 10.4

q(LCP) 5.51 (0.205) 0.06 (0.124) 1.05 (0.082) 0.69 0.086 0.69 0.087 1.1Gmax 1.08 (0.265) 0.56 (0.159) 0.19 (0.106) 0.20 0.112 0.20 0.126 11.1K 4.82 (0.371) 0.65 (0.221) 1.08 (0.147) 0.35 0.156 0.28 0.169 7.7Rd 3.24 (0.346) 0.04 (0.207) 0.65 (0.138) 0.20 0.145 0.10 0.170 14.7

Low irradiance, infertile soil Amax 1.80 (0.542) 1.00 (0.175) 0.37 (0.214) 0.58 0.120 0.56 0.127 5.5LCP 2.08 (1.068) 0.05 (0.348) 0.46 (0.421) 0.00 0.237 0.00 0.251 5.6

q(LCP) 4.13 (0.466) 0.57 (0.150) 0.51 (0.184) 0.50 0.103 0.35 0.126 18.3Gmax 1.58 (0.558) 1.01 (0.180) 0.32 (0.220) 0.56 0.124 0.56 0.127 2.4K 2.53 (0.835) 0.65 (0.269) 0.27 (0.329) 0.09 0.185 0.07 0.196 5.6

Rd 2.11 (0.959) 0.59 (0.312) 0.09 (0.378) 0.08 0.213 0.07 0.214 0.5

Shown are the estimated values based on combined observations from all growth environments (general allometric equation) and based on values estimated separately for each growth environment(environment-specific allometric equation). Also shown are the proportions of variance accounted by the regressions (R

2), the residual standard error (s.e.) and the percentage increase in the residual s.e.

when using the general equation rather than the environment-specific equation. Amax, maximum net photosynthetic rate; LCP, light compensation point; q(LCP), quantum yield at the LCP; Gmax, maximumgross photosynthetic rate; K, photon flux density at 1 Gmax; Rd, respiration rate.
http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/


10/17

130 Lachapelle & Shipley Predicting photosynthetic light response curvesee

LogGmax

Log(

Rd)

LogLCP

LogAmax

LogqLCP

Log

K

08

10

12

14

16

18

20

22

24

26

08 0604

02

1820

2224

2628

3000

02 3204

30

28

26

24

22

20

18

16

1408

0604 02 00

02

1820

2224

2628

3032

High soil fertility, high irradianceLow soil fertility, high irradianceHigh soil fertility, low irradianceLow soil fertility, low irradiance

04

04

02

00

02

04

06

08

101214

0806

04 02 000204

1820

2224

2628

3032

2022

24

26

28

30

32

34

3608

0604

02 00

1820

2224

2628

3002 3204

20 02

18

16

14

12

10

08

06

04

0208

0604 02 00

1820

2224

2628

30

00

02

04

06

08

10

12

1408 06

0402 00

1820

2224

2628

30

0204

32 0204

32

FIG . 4 . Predicted values of the parameter estimates for two photosynthetic light response curves (Mitscherlich, Michaelis Menten) given leaf mass per area(LMA) and leaf nitrogen content (N). Growth environments are indicated in the key.

low and the soil was fertile. A larger proportion (82 %) of thevariance in Amax was accounted for when combining observa-

tions across growth environments than in any single growthenvironment although the residual standard error was slightlyhigher (approx. 10 %) when ignoring the differences acrossenvironments. The light compensation point was poorly pre-dicted in each growth environment separately and only 29 %of the total variance was explained by the general equation.

However, 79 % of the total variance in the quantum yield

at the light compensation point was accounted for in thegeneral equation and the residual standard error of the


11/17

Lachapelle & Shipley Predicting photosynthetic light response curves 131

general equation (ignoring environmental differences) wasonly about 3 % greater. We then used the general allometricrelationships predicting the three parameters given LMA andN (i.e. ignoring environmental differences) to predict thelog10 net photosynthetic rate of each individual leaf at each

experimental photon flux density and compared these withthe measured net photosynthetic rates (Fig. 5A). There was a

good correlation (r

2

0.88) between these observed andpredicted values and no evidence of systematic bias; thereduced major axis regression between observed and predictedvalues had an intercept of 0.02 and a slope of 1.04.


12/17

Predictedrate(mmolCO2g1s1)

Predictedrate(mm

olCO2g1s1)

08

06

MichaelisMenten curves10

05

MichaelisMenten curves; log10

04

02

00

02

01

00 04 08 12 001 005 020 100

08

06

Mitscherlich curves Mitscherlich curves; log1010

05

04

02

00

02

01

00 04 08 12 0005 0020 0100 0500

Measured rate (mmol CO2 g1 s1) Measured rate (mmol CO2 g

1 s1)

FIG. 5 . Observed photosynthetic rates per leaf (x-axis) vs. predicted values obtained from allometric equations linking the parameters of two photosynthetic lightresponse curves (Michaelis Menten, top; Mitscherlich, bottom). Actual values are shown on the left and log10-transformed ( plus a constant to ensure non-

negativity) values are shown on the right. Also shown is the 1 : 1 line.

Michaelis Menten PLR

curve

The allometric regressions for Gmax showed essentially thesame level of predictive ability as for Amax. The K parameter

was moderately well predicted (R2

0.55) in the generalequation even though it was rather poorly predicted in eachseparate environment-specific regression; the residual standarderror of the general equation was only 15 % larger on averagebut the increase in the residual standard error was much higherin the high-irradiance environment (19 %) than in the low-irradiance environment (6.7 %). The respiration rate wasalso moderately well predicted (R

2 0.52) in the general

equation even though it was rather poorly predicted in eachenvironment-specific regression; the residual standard error

of the general equation was only 8 % larger on average butthe increase in the residual standard error was higher in thefertile soil (13 %) than in the infertile soil (4 %). There wasgood agreement between observed and predicted photosynthet-ic rates (Fig. 5B) with an r

2of 0.89 and a reduced major axis

intercept and slope of 0.01 and 1.02, respectively.Our measurements largely fall within those reported by

Marino et al. (2010) although our light compensation pointsare less variable (Fig. 6). Our measurements also mostly fallwithin the cloud of points reported by Wright et al. (2004a)for herbaceous species of open habitats with the exception ofthe N LMA relationship (Fig. 6). The low-fertility plantsin our study had less nitrogen per mass than expectedgiven their values of LMA while those growing in the high-

fertility soil followed the same trend as those in Wright et al.

(2004a)but with both higher leaf nitrogen and lower LMA.


13/17

DISCUSSION

That leaf thickness and, by implication, LMA are affected bylevels of irradiance has been known at least since Hanson(1917). Meziane and Shipley (1999), using factorial levels ofgrowth irradiance and nitrogen supply, showed that LMAand leaf nitrogen content are independently affected by bothenvironmental factors but with strong interactions betweenthem. The physiological basis of this plasticity is poorly under-stood but Sugiyama and Gotoh (2010), studying species ofFestuca, have shown that both the growth irradiance and thenitrogen supply experienced by the meristem during leaf for-mation affect the rate of cell division and the length of timeduring which the dividing cells elongate. Of course, becausean increased nitrogen supply translates into a larger plantmass, this would eventually increase self-shading and thus in-directly affect LMA via changes in the local light environment;this is not the case in our study because the plants were well-spaced and non-overlapping and remained small.

As both LMA and N are sensitive to the growth environ-ment, and because photosynthetic light response curves arealso sensitive to the growth environment (Lambers et al.,1998), the existence of such environmental plasticity putsinto doubt the generalizability of the results of Marino et al.(2010), who linked the parameters of the PLR curve to LMAand N. If such plastic responses were uncoordinated acrossthe measured leaf traits, then different allometric relationshipswould be needed for each environment. Surprisingly, becauseeach point in Figs 1 and 2 represents a species average in agiven growth environment, the fact that the four growth


14/17

Log(LCP)

Log(qLCP)

Log(Amax)

Log(LMA)

Log(N)

06

02

02

20

30

00

10

20

25

35

2

1

0

1

202 02 06 30 20 20 10 00 35 25 2 1 0 1 2

Log(N) Log(LMA) Log(Amax) Log(qLCP) Log(LCP)

FIG . 6 . Scatterplot matrices of the observed values (closed circles) of leaf mass per area (LMA), leaf nitrogen content (% dry weight) and the parameters of theMitschelich light photosynthetic response curve superimposed over the values from two independent data sets. Above diagonal (solid grey circles): data from

Marino et al. (2010); below diagonal (open grey circles): data from Wright et al. (2004b).

environments are clearly separated shows that plastic changesbetween environments were even larger than the species dif-ferences within a single growth environment. As Fig. 3 illus-

trates, this was equally true for the two independentvariables (N, LMA) used by Marino et al. (2010) to predictthe free parameters of the PLR curves. Furthermore, the im-portance of each environmental treatment on the level of plas-ticity depended on the leaf trait. Plasticity in N and the twomeasures of maximum photosynthetic rate was primarily dueto changes in soil fertility, plasticity in LMA and thequantum yield at the light compensation point were primarilydue to changes in growth irradiance, while plasticity in thelight compensation point and the dark respiration rate wereequally affected by both environmental drivers.

Despite this wide range of plasticity, and despite the factthat different leaf traits responded differently to the two envir-onmental drivers, the plastic responses were relatively but

not completely coordinated across the measured leaf traits.


15/17

We are not suggesting that this is strictly true, only that itis a reasonable approximation when dealing with manyspecies and environments; researchers interested in particularspecies would find this level of approximation too course.As a result, although there were statistically significantdifferences between the slopes and intercepts of theseallometric equations in the different growth environments,

these differences were small relative to the generalinterspecific trends. Consequently, the general allometricrelationships obtained when ignoring these environment-specific differences produced quite good pre- dictions of leafgas exchange across a wide range of light intensities over allfour growth environments with only a modest increase in theamount of residual variation. This is because the changes inthe external environment acted mostly indirectly on the PLRcurve parameters via the effects of the environment on LMAand N. In other words, the PLR curve parameters had quitegeneralizable allometric relationships with LMA and N ir-respective of whether the variation in LMA and N wascaused


16/17

by genetic (interspecific) or environmental ( plastic)differences; knowledge of the growth environment was mostlyredundant once one knew the values of LMA and N. Thismeans that one can largely apply these general allometricequations irre- spective of differences in soil fertility and thelight environment in which the plant is growing. We do not

know if this will con- tinue to be true with respect to otherimportant environmental factors (soil moisture, temperature,etc.) or with respect to a wider range of plant types. If so, thenlarge-scale ecosystem models could potentially use values ofLMA and N that are already available in large trait databasesin order to replace generic PLR curve parameters withvalues specific to each species weighted by the relativeabundance of each species in the community. It isimportant to note that our results are based only onherbaceous species typical of open habitats with rich soils.It is possible that shade-adapted understorey herbs mightrespond differently. Similarly, basal eudicots, basalangiosperms and broadleaved monocots are not repre- sentedin our study.

Givnish et al. (2004) showed that although Amax could bepredicted quite well from N and LMA, it could be predictedeven better if the growth irradiance was included. As weonly grew the plants at two levels of irradiance we cannotquantify the relative importance of different growth irra-diances. We did not measure the actual incoming irradianceof each leaf, but the plants were well spaced and small andso we do not think that there was excessive shading thatwould have greatly changed the actual growth irradiance ofdifferent leaves; none the less, this probably contributed tosome of the unexplained variation.

We do not know which physiological mechanisms producethis co-ordination of leaf morphological and physiologicaltraits but there are some interesting potential consequences

of our results. For instance, the quantum yield at the light com-pensation point measures the amount by which a unit mass ofleaf tissue can increase carbon fixation given a unit increase inthe photon flux density when the leaf is in heavy shade (i.e. thelight compensation point); note that this is not the same as theapparent initial quantum yield. Our general allometric equa-tion states that this efficiency is more sensitive to changes inleaf morphology (LMA) than to changes in leaf chemistry(N): an increase in N by one standard deviation above themean of our data set (i.e. 3.05 %) and without any change inLMA would result in a 3.8 % increase in the quantum yieldbut an increase of LMA by one standard deviation above themean (0.0025 g m

22) without any change in N would decrease

the quantum yield by 6.4 %. One way to do both simultaneous-

ly would be to produce leaves with more or larger mesophyllcells and fewer vascular and support tissues as this would bothincrease the amount of nitrogen per unit dry mass and decreaseLMA; Shipley et al. (2006) have argued that this trade-off between leaf volume occupied by cytoplasm vs. cellwalls isone of two main drivers of the leaf economic spectrum.On the other hand, our results are not based on the rapidlychanging light environment of light flecks that are typical oflow-light environments. It is therefore important to extendour results from PLR curves based on leaves that haveequilibrated to a given photon flux density to ones based onthe rapidly changing photon fluxes that are more typical oflow-light environments (Givnish et al., 2004; Montgomery

and Givnish, 2008).
http://aob.oxfordjournals.org/http://aob.oxfordjournals.org/


17/17

SUPPLEM ENTA RY

DATA

Supplementaray data are available online at ww w .aob.oxfordjourna l s.org and consist of Table S1: Excel spreadsheetof the raw data for all variables measured in the experiment.

AC KNO W LED GE

MENTS

This research was funded by the Natural Sciences andEngineering Research Council (NSERC) of Canada.

LITER AT URE C ITED

Bugmann H. 2001. A review of forest gap models. Climatic Change 51:259 305.

Givnish TJ, Montgomery RA, Goldstein G. 2004. Adaptive radiation of

photosynthetic physiology in the Hawaiian lobeliads: light regimes,static light responses, and whole-plant compensation points. AmericanJournal of Botany 91: 228 246.

Hanson HC. 1917. Leaf-structure as related to environment. AmericanJournal of Botany 4: 533 560.

Kattge J, Daz S, Lavorel S, et al. 2011. TRY a global database of planttraits. Global Change Biology 17: 2905 2935.

Lambers H, Chapin FS, Pons TL. 1998. Plant physiological ecology.New York: Springer.

Lo E, Wang ZM, Lechowicz M, et al. 2001. Adaptation of the LIGNUMmodel for simulations of growth and light response in Jack pine. ForestEcology Management 150: 279 291.

Marino G, Aqil M, Shipley B. 2010. The leaf economics spectrum and theprediction of photosynthetic light-response curves. Functional Ecology24: 263 272.

Meziane D, Shipley B. 1999. Interacting determinants of specific leaf area in

22 herbaceous species: effects of irradiance and nutrient availability.Plant, Cell and Environment 22: 447 459.

Meziane D, Shipley B. 2001. Direct and indirect relationships between specif-ic leaf area, leaf nitrogen and leaf gas exchange. Effects of irradiance andnutrient supply. Annals of Botany 88: 915 927.

Montgomery RA, Givnish TJ. 2008. Adaptive radiation of photosyntheticphysiology in the Hawaiian lobeliads: dynamic photosynthetic responses.Oecologia 155: 455 467.

Ollinger SV, Richardson AD, Martin ME, et al. 2008. Canopy nitrogen,carbon assimilation and albedo in temperate and boreal forests: functionalrelations and potential climate feedbacks. Proceedings of the NationalAcademyof Sciences 105: 19336 19341.

Perttunen J, Sievanen R, Nikinmaa E, Salminen H, Saarenmaa H, VakevaJ. 1996. LIGNUM: a tree model based on simple structural units. Annalsof Botany 77: 87 98.

Perttunen J, Nikinmaa E, Lechowicz MJ, Sievanen R, Messier C. 2001.Application of the functional-structural tree model LIGNUM to sugar

maple saplings (Acer saccharum Marsh) growing in forest gaps. Annalsof Botany 88: 471 481.

Pilegaard K, Ibrom A, Courtney MS, Hummelshj P, Jensen NO. 2011.

Increasing net CO2 uptake by a Danish beech forest during the period

from 1996 to 2009. Agricultural and Forest Meteorology 151: 934 946.Shipley B, Lechowicz MJ, Wright I, Reich PB. 2006. Fundamental trade-

offs generating the worldwide leaf economics spectrum. Ecology 87:535 541.

Sugiyama S-I, Gotoh M. 2010. How meristem plasticity in response to soilnutrients and light affects plant growth in four Festuca grass species.

New Phytologist 185: 747 758.R Development Core Team. 2008. R: A language and environment forstat-

istical computing, Vienna: R Foundation for Statistical Computing. http: //r-p r oject.org.

Wright IJ, Reich PB, Westoby M, et al. 2004a. The worldwide leaf econom-

ics spectrum. Nature 428: 821 827.
http://www.aob.oxfordjournals.org/http://www.aob.oxfordjournals.org/http://www.aob.oxfordjournals.org/http://r-project.org/http://r-project.org/http://r-project.org/http://www.aob.oxfordjournals.org/http://www.aob.oxfordjournals.org/http://r-project.org/http://r-project.org/

Ann Bot 2012 Lac Ha Pelle 1149 57

Documents

Transcript of Ann Bot 2012 Lac Ha Pelle 1149 57