SCIENTIA
HORTICULTURR
Scientia Horticulturae 63 (1995) 83-100 zyxwvutsrqponmlkjihgfedcbaZYXWVU
Net photosynthesis response to light and air CO2
concentration of Begonia X hiemalis: whole plant
measurements and modelling
Panagiotis Giaglaras a, Maria Baille b, Alain Baille a,*
“Institut National de la Recherche Agronomique, Station de Bioclimatologie, 84143 Montfavet Ceakx, France
YYorniti National Interprofessionel de l’Horticulture, Dipartement Technique, 94152 Rungis Cedex, France
Accepted 28 November 1994 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR
Abstract
Response of net photosynthesis (Pn) of Begonia plants to photosynthetic photon flux density (I,)
and carbon dioxide concentration (C) was measured and modelled. Measurements were performed
on whole plants at different growth stages in either controlled (assimilation chambers) or in situ
conditions (glasshouse). A rectangular hyperbola was chosen as the response function of net photosynthesis to I,, C and leaf area (A). Parameterisation and validation of the model was carried out
using Pn measurements on Begonia plants, cv. “Line”. The model was then tested for its extrapolative
ability with two other data sets: one obtained in assimilation chambers with different cuhivars (cv.
“Heidi”, and shoot cuttings of the Rosalie group) and the other from in situ measurements in
glasshouse with the cv. “Line”, using a portable assimilation chamber. Prediction of Pn was satisfactory under normal conditions of air temperature (T,), vapour pressure deficit (VPD) and I, in
glasshouse production. Discrepancies between measured and estimated rates of Pn occurred under
(over25”C) andVPD (morethan 1.5kPa),and
high I, (more than 700 pm01 photon m -2s-1),T,
can possibly be attributed to effects of stomata1 closure on the net photosynthetic rate.
Keywords: Ornament&; Pot plants; Assimilation chambers; Effective leaf area
* Corresponding author.
Abbreviations: PPFD, photosynthetic photon flux density (pm01 photon me2 s-l); Pn, whole plant net
photosynthesis rate (pg CO* per plant s-l); I,, incident PPFD at the top of the canopy (pm01 photon me2 s-l);
T., air temperature (“C) ; C, air carbon dioxide concentration (~1 l- ’ ) ; VPD, vapour pressure saturation deficit
of the air (kPa) ; A, leaf area ( m* per plant) ; Ab, basal area ( m
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ
- * per plant) ; A,, effective leaf anza ( m* per plant) ;
gc,iand g,j conductance to CO2 and to water vapour (where the. index i may be b (boundary),
(mesophyll), t (total)), (molm-*s-‘);r,,,cropsurfaceresistance
(sm-‘).
0304-4238/95/$09.50
0 1995 Elsevier Science B.V. All rights reserved
SSDIO304-4238(94)00755-l
s (stomatal),
m
84 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83400 zyxwvutsrqponmlkjihgfedcbaZYX
1. Introduction
Models for crop growth simulation are expected to be useful tools in greenhouse horticulture, for planning purposes and climate optimisation (Liebig, 1989). They can give
valuable information on the duration of the cultivation period or the dry matter production
under different climate conditions, therefore contributing to improved decision-making
(Buchwald, 1987; Mann and Krug, 1989; Leutscher, 1990) on both strategic and tactical
levels (Challa and Van Straten, 1991). However, development of decision support systems,
including physical, physiological and economic models, requires a good knowledge of longand short-term plant responses to climate conditions (Challa and Schapendonk, 1986).
Concerning Begonia X El&or cultivars, which are commonly used in greenhouse pot plant
production, few data are available in the literature. The role of carbon dioxide enrichment
and artificial lighting to shorten the duration of the cultivation period and to improve plant
quality has been demonstrated by several workers (Mathijssen, 1985; Bierman, 1988;
Hendriks, 1988; Verberkt, 1990). Experimental data on development and flowering showed
that a shortage of assimilates for plant growth occurs at low light intensities (Powell and
Bunt, 1985; Simmonds and Nelson, 1988). Literature on photosynthetic response is very
scarce. As far as we know, only one work deals with photosynthesis of Begonia (Bierhuizen
et al., 1984).
The aims of this paper are to present ( 1) experimental data on the response of Begonia
net photosynthesis to PPFD and CO* concentration, and (2) a model for predicting the net
photosynthetic rate which could be used in algorithms for short-term greenhouse climate
optimisation or as the photosynthesis module in a general crop growth model for Begonia.
2. Materials and methods
2.1. Plant material
Six groups of plants (Cl-C6) were used in the present study. Plant material and climate
conditions were different for each group:
( 1) for the C l-C3 groups, leaf cuttings of Begonia X hiemalis cv. “Heidi” were grown
in a glasshouse at INRA Montfavet (southern France) from February to May 1992 (Cl ),
Five weeks after planting, 100 plants from the greenhouse were placed in a growth chamber
with constant air temperature (20 f 2°C) and relative air humidity (70 f 10%) and divided
in two groups, C2 and C3, differing in the level of incident PPFD: 200 pm01 photon m-’
s- ’ and 400 pmol photon m -’ s-’ for C2 and C3, respectively.
(2) For the C4 and C5 groups, shoot cuttings of Begonia X hiemulis cv. “Line” were
grown in the same glasshouse during two periods: October 1992-February 1993 (i.e. a
winter culture; C4) and March-June 1993 (i.e. a spring culture; C5).
(3) For the C6 group, shoot cuttings of Rosalie group cultivars were grown during winter
1992-1993, in a production glasshouse, at CNIH (Comite National Interprofessionel de
l’Horticulture), Angers (western France).
P. Giaglaras et al. /Scientia
Horticulturae 63 (1995) 83-100
85
2.2. Cultivation practice
Shoot and leaf cuttings were grown in plastic pots ( 14 cm diameter, 1 1 volume) with a
mixture of 20% perlite and 80% peat, placed on metallic benches inside a 60 m* heated
glasshouse. One hundred and fifty plants was selected at the same developmental stage (five
apparent leaves and no secondary shoots) and used for Groups Cl-C5. During the cultivation period, plant density was progressively decreased, from 45 to 8 plants m-* of bench,
to avoid light and space competition between plants.
The plants were fertilised with a commercial mixture ( l-0.3-1.44)
plus micro-elements.
Electric conductivity and pH of the irrigation water were 1.4 (mS cm-‘) and 6.3, respectively. Before each fertigation, the pots were drained with water for 5 min to avoid
accumulation of salts in the substrate. Hence, the substrate conductivity was maintained
between 1 and 1.5 mS cm-‘.
2.3. Glasshouse equipment and climate
The heating setpoint was 20°C. Air ventilation (roof opening) operated when temperature
was higher than 25°C. Three tungsten lamps (60 W Osram) per bench, located 1.5 m above
the plant level along the middle axis of each bench, were used to ensure a 16 h photoperiod
during the vegetative stage of the plants.
PPF’D (I,, 400-700 nm), dry bulb (T,) and wet bulb (T,,,) air temperature were
measured at the top of the canopy, and water vapour pressure of the air (e,) was estimated
from the T, and T,,, measurements. Hourly mean values of Z,, T,, and e, were recorded by
a data logger (Campbell Scientific, Logan, UT, USA; 21X). Mean saturation vapour
pressure (e,) and vapour pressure deficit (e, - e, = VPD) were calculated from T, and e,.
2.4. Carbon dioxide assimilation rate measurements
Carbon dioxide exchange rates of whole plants were measured as follows:
( 1) in assimilation chambers (semi-open system), for plants from all groups (Cl-C6).
The Pn data measured on plants of the C4 and C5 cultivations (cv. “Line”) were separated
in two data sets; one set was used for identification of the model parameters and the second
set for their validation. The data obtained on plants of Cl, C2 and C3 groups (cv. “Heidi”),
and C6 group (Rosalie) were used to test the extrapolative capacity of the model (different
cultivars and greater leaf areas).
(2) In in situ (greenhouse) conditions, with a portable (U-6200, Li-Cor, Lincoln, NB)
photosynthesis meter, only for plants of the C5 cultivation (cv. “Line”). Those data were
used to test the model for its extrapolative capacity (different environment).
2.4.1. Assimilation chambers
The two chambers, control system and operating process have been described by Longuenesse et al. ( 1982) and Gary ( 1988a). The measurements of Pn were conducted on plants
with different growth stages, taken in the glasshouse. Throughout the cultivation period,
two pot plants per week were taken and placed in the assimilation chambers for 4 days.
86 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
P. Giaglaras et al. /Scientia Horticuhrae 63 (1995) 83- 100 zyxwvutsrqponmlkjihgfedcbaZY
START
0
1 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO
2
3 END
Photoperiod
Assimilation chamber
Tie
(2h)
PPFD
6 levels
co2
3 levels
Fig. 1. Schematic representationof the PPFDand COzsequencesappliedduring measurementsof the wholeplant
netphotosynthesis
rate in the assimilation chambers.
Nutrient solution was supplied by capillarity from a stirred water tank located inside each
chamber.
Air temperature inside the chambers was maintained constant at 20 f 2°C day and night,
and relative air humidity at 75 f 10%. PPFD was measured at the top of the plant by means
of quantum sensors (Li-Cor) in the same way as in the glasshouse.
The CO, exchange rate of the plants was obtained from the air flux measurements (mass
flowmeters, Brooks, Inc.) and from the measurements of CO, concentrations (IRGA analyser, ADC) at the input and output of the chambers. The influence of water concentration
changes and inertia of the system were taken into account, using the approach of Gary
( 1988b).
Pn was recorded continuously during periods while PPFD and CO* concentration followed predefined sequences (Fig. 1) . CO1 concentration varied from one photoperiod to
the other: 900 ~1 l- ‘, 400 ~1 l- ’ and 600 ~1 1-l on the first, second and third day,
respectively. During a photoperiod, PPFD changed every 2 h from high (700 pm01 m-*
s- ‘) to zero levels with five intermediate levels (approximately 50, 100,250,400 and 500
pm01 m-* SK’) . At night, CO2 concentration was maintained at 400 pull-‘. At the end of
the third day, the plant shoot was cut and the respiration of substrate plus root was measured
both in darkness and in light (700 pm01 m-* s-l).
In the following, Pn rates refer to the calculated mean steady-state values corrected for
the substrate CO2 efflux.
2.4.2. In situ measurements
Net photosynthesis of whole plant (cv. “Line”) was also measured in the glasshouse
under natural climatic conditions. For this purpose, the portable photosynthesis system was
modified in the following way: the leaf chamber was replaced by a hemispheric glass
chamber enclosing the plant in a volume of 21.1 1 (30 cm diameter, 35 cm height). A
stainless steel base supported the glass cover and the sensors; a small fan mixed the internal
atmosphere. Tests of the system tightness gave very satisfactory results, even in conditions
of high CO2 concentration gradients.
P. Giaglaras et al. /Scientia
Horticulturae 63 (1995) 83-100
87
PPFD was measured at plant level by a quantum sensor. For Pn measurements, the same
CO* concentrations as in the assimilation chambers were applied. At the end of the Pn
measurements, plant leaf area (A) and shoot dry matter (DM,) were obtained by conventional destructive analysis.
3. The model
At whole plant level, net photosynthesis rate (Pn) can be considered as the difference
between gross photosynthesis (Pg) and dark respiration (Rd). Assuming that respiration
at light is equal to Rd, we have
Pn=Pg-Rd
(1)
(where Pn, Pg and Rd are in pg COP per plant s- ’ ) .
Pg integrates leaf gross photosynthesis (Pgl, pg CO2 m-* s-‘) over the leaf area. A
rectangular hyperbola (Thornley, 1976) was adopted for expressing the dependence of Pg,
vs. 1, (pm01 photon m-* s-l):
pgr=aI,W%m~
+ Pgm,
(2)
where Pgm, is the maximum value of Pgl at light saturation level and a! is the incident PPFD
use efficiency ( pg CO, per pm01 photon).
As suggested by Acock ( 1991), the effect of CO2 concentration on the leaf gross photosynthesis, in the range of CO2 concentrations used, can be expressed as
Pgm, = rC
(3)
where r is the total (stomata plus mesophyll) leaf CO, conductance zyxwvutsrqponmlkjihgfedcbaZYXWVU
( pg CO2 m-* s- ’ 1
PI-‘) and C the CO, concentration (~11~r ).
The measurements at whole plant level showed that, for a given light (I,) and CO2 level
(C), Pg increases with leaf area (A, m* per plant) until it reaches a plateau. The selfshading effect of A was described in the following way: we assumed that only part of the
whole plant leaf area responds to the incident PPFD, the so-called “effective leaf area”
(A,, m* per plant). A, was related to leaf area using the following formula:
A,=A(-)
a+A
where u (m* per plant) is a statistical parameter called “effective leaf area coefficient”.
The effects of leaf area distribution, orientation and age on Pg are included in this parameter.
The dependence of A, on A for several values of cr is shown in Fig. 2.
The basal area of the plants, Ab (i.e. the horizontal projection on the ground of the plant
canopy; m2 per plant) is a function of the whole plant leaf area. The following mathematical
expression was used for fitting:
Ab =plA +pd2
+p+i3 +p4A4 +p#
(5)
88
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83-100
V
I
0.05
I
I
I
0.15
I
0.25
Leaf area, A (In2 plant- I)
Fig. 2. Shape of the relation between the effective (A,) and the total leaf area (A) for various values of the
effective leaf area coefficient CT(0.05,0.10,0.15 and 0.20 m2 per plant).
withp, =0.91,p,= -9.16,p3=56.3,p,=
- 1.54X lO’andps= 1.58X 102.Theestimated
relation is given by the continuous line in Fig. 3 (Z?‘=O.85, RSE=
0.00655 me2 per plant). Eq. (5) was used to estimate the leaf area index (L = A/Ah, m2
mm2), a common variable in light interception laws applied in crop growth models.
The dark respiration (Rd) , under conditions of constant air temperature and growth rate,
can be considered to be mainly dependent on the amount of shoot dry matter (DM,, g per
plant) :
Rd = pDMs
(6)
where p is the specific dark respiration ( pg CO g- ’ s-i) of the plant.
Substituting Eqs. (2)-(4) and (6) in Eq. ( 1) leads to
CYZTC
Pn=&-f--$(-&
+
- PDM,
(7)
which describes the net photosynthesis of the whole plant in relation to Z,, C, A and DM,,
with four parameters ((Y,7, u and /3) to be identified. In the following, the values of these
parameters were derived by means of non-linear least-squares fitting of the experimental
data gathered from the assimilation chambers, using the S-PLUS statistical routine (Statistical Sciences Inc., 1991).
The goodness of fit between the estimated and measured values was evaluated using the
relation
%timated = a + bPnmeasured
(8)
The determination coefficient ( R2) and residual standard error (RSE, g CO2 s- ’ per plant)
were calculated. The residuals were plotted versus the estimated values and the independent
variables, C, IO,A and DM,.
P. Giaglaras et al. /Scientia Horticulturae 63 (199s) 83- 100
89
0.10 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
_
0.08
‘r
B
-z
0.06
3
9
0.04
-:
0.02
0.0
0.05
0.10
0. is
A (m2
0.20
fitted
0.25
0.35
area (A),
for plants of Groups Cl-
Assimilation
groups
h
yB
0.30
plant-l)
Fig. 3. Relationship between the basal plant area (A*) and the whole plant leaf
C6: measured values ( X ) and fitted curve (Eq. (5)).
100
curve
80
“0.
-20
0
chamber measurements
et C5 (cv. “Line”)
/
second data set
0%
0
C4
20
Pnmeasured
40
60
80
(Pg Co2 S-’ Pkmt-‘1
Fig. 4. Estimated vs. measured Pn values: comparison of the modelestimations (Eq. (7) ) with Pn values obtained
in the assimilation chambers using plants from Groups C4 and C5 (cv. “Line”).
90 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83- 100 zyxwvutsrqponmlkjihgfedcb
4. Results
4.1. Modeljitting and parameter estimation
The range of measured dependent and independent variables used for the parameter
identification and the values of the estimated parameters are given in Table 1. Comparison
of the measured and estimated rates of Pn (Fig. 4) indicates a good agreement: a =
0.79 L-0.40 pg CO* SK’ per plant;
b = 0.95 f 0.02,
R2 = 0.94
and
RSE =
4.9 pg CO1 s- ’ per plant. The residuals followed a normal distribution and did not show
any correlation with the independent variables used in the model nor with the model
estimations. The confidence interval of the model predictions, at the 95% confidence level,
is * 8.04 j_kgCO2 s- ’ per plant.
4.2. Model validation
Comparison of measured and estimated values, using the second set of Pn data obtained
in the assimilation chambers on plants of C4 and C5 cultivations, showed a good agreement
(Fig. 5): a= -0.20f0.33
pg CO2 s-r per plant; b = 1.04 f 0.01 and R2 = 0.96.
Table 1
Ranges of the independent variables used from the model parameterisation and estimated values of the parameters
using data obtained in assimilation chambers for BegoniaX hiemalis (cv. “Line”)
plants, with constant air
temperature (20 rt 2°C) and air relative humidity (75 f 10%)
Units
Measured range
I.cg COP s-r per plant
- 13-+72
Dependent variable
Pn, whole plant net
photosynthesis
Independent variables
I,, incident PPFD
pm01 photon m-* s-’
/.&11-I
C, air carbon dioxide
concentration
A, total leaf area
DM ,, shoot dry matter
O-630
300-1000
m* per plant
g per plant
0.029-0.216
1.5-18.3
Parameters
Units
Estimated
values
Standard
error
a, incident PPFD use
efficiency
7, conductance to carbon
dioxide
o, effective leaf area
coefficient
p, specific respiration
pg CO* per pmol photon
2.9
0.17
pg COz m -zs-‘lILl-’
2.85
0.21
m2 per plant
0.15
0.02
/.&gco* g-1 s-’
0.45
0.05
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83- 100
Assimilation
91
chamber measurements
groups C4 et C5 (cv. “Line”)
first data set
‘I
-20
-20
n /.///
I
I
I
I
I
I
0
20
40
60
80
too
Pflmeasured
(Pg co2 s-I plant-‘)
Fig. 5. Estimated vs. measured Pn values in the validation test: comparison of estimated values (E!q. (7)) with
the second data set of Pn (plants cv. “Line” from Groups C4 and C5 measured in the assimilation chambers).
Assimilation
chamber measurements
groups C
/
I, C2, C3 (cv. “Heidi”)
and C6 (cv. “Rosalie”)
/
I
I
I
I
I
J
-20
0
20
40
60
80
IO0
L
Pnmeasured
(Pg WI
s-’ pkint-‘1
Fig. 6. Estimated vs. measured Pn values in the extrapolation test: model ability to predict the Pn values measured
in the assimilation chambers with plants from Groups Cl-C3 (cv. “Heidi”) and C6 (Rosalie).
92
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83400
100
Rosalie group plant
S = 0.44 m2 plant-t
80
k&30
‘In
0” 40
U
on
z
$
20
measuredat :
* !mplI-’
0 6oopIl-I
0 4ooplI-1
0
I
-20
0
IO0
200
3w
400
500
600
IO (pm01 mm2 s-l)
Fig. 7. Measured and estimated response of Pn to PPFD and CO* concentration for a plant from Group C6
(Rosalie) having leaf area (0.44 m* per plant) twice that of the largest plant used for the identification of model
parameters. 0, Measurements at 400 ~1 I-‘; 0, measurements at 600 /.rll-‘; A, measurements at 900 ~1 I-‘.
Continuous lines represent model estimations for the same PPFD and COa levels.
“in- situ” measurements
group C5 (cv. “Line”)
0
20
40
60
80
100
f’nmeasured (Pg CO:! s-’ Plant-‘)
Fig. 8. Estimated vs. measured values for Pn measurements obtained in situ (cv. “Line”).
93 zyxwvutsrq
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83- 100
4.3. Extrapolative tests
First, the predictive value of the model was checked against measurements obtained in
assimilation chambers with the C 1 to C3 (cv. “Heidi’ ’ ) plants. The results (Fig. 6) showed
that the agreement is satisfactory, although we note a slight overestimation for high levels
of CO* and Z, (a = 1.94 + 0.30 pg CO% s- ’ per plant; b = 1.24 + 0.01 and R* = 0.95).
60
measurements
group C5 (cv. “Line”)
+ (a)
++
“in- silt/
++ +
+
+
+ +
P
+
+
+
_.
+
+
20
I
I
I
I
25
30
35
40
Ta W)
“in-situ”measurements +
group C5 (cv. “Line”)
+
+
+
+
t
+
++ + +
+
k
++ +
+
+
+
(b)
+
+
+ ++
+
._.
+
+
0
I
I
I
I
2
3
I zyxwvutsrqponmlkjihgfedcbaZ
4
VfD (kPa)
Fig. 9. Difference ( APn) between estimated and measured values of Pn (for in situ measurements, cv. “Line”)
vs. (a) air temperature ( T,) ; (b) vapour pressure saturation deficit zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ
(VPD) . Horizontal dashed lines represent the
95% confidence interval of the model predictions.
94
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83- 100
The prediction reliability of the model for plants with larger leaf area was tested using
data obtained on plants of the Rosalie group (C6) having a leaf area of 0.44 m* per plant,
e.g. twice the leaf area (0.22 m* per plant) of the largest plant used for the model parameter
identification. The model simulated fairly well the measured values until Z, reaches about
250 pmol photon m -2 s-’ . For higher values of Z,, the model significantly overestimates
the measured Pn rates, especially at the 600 and 900 Z~ll-’ levels (Fig. 7).
Finally, the model was checked against data obtained in situ (Fig. 8). In this case, the
predictions were less satisfactory, with a significant overestimation of Pn by the model:
u= 10.43& 1.40 Z.&gco, s-l per plant; b = 1.26 f 0.09 and R* = 0.40. Analysis of the
residuals showed that the deviations (A Pn = Pnestimted - Pnmeasured)are important under
conditions of high temperature (Fig. 9), which were associated with both elevated PPFD
(more than 700 pm01 m-* s-’ ) and high VPD (over 1.5 kPa) levels. zyxwvutsrqponmlkjihgfedcbaZY
5. Discussion
Data on net photosynthesis usually found in the scientific literature refer to the leaf or
crop level, but studies at the whole plant level are less frequent. Measurements at leaf scale
are currently available, as they are much easier to obtain than measurements at the whole
plant level. Extrapolation from leaf level to whole plant level needs the estimation of the
plant leaf area distribution and light extinction, sometimes combined with a directional
treatment of incident light. Such an approach does not significantly improve the predictions
at the whole plant level if not supported by a detailed description of the leaf photosynthetic
characteristics and by the characterisation of the microclimate inside the plant, i.e. temperature and CO2 distribution (Boote and Loomis, 1991) .
Obtaining these inputs is not straightforward. In contrast, the PPFD level above the plant
canopy is commonly available in climate-controlled
greenhouses, and total leaf area and
shoot dry matter can be predicted from simple growth models. These reasons plead for the
development of a photosynthesis model at the whole plant level, applicable and usable in
production conditions. In fact, the model presented contains some elements relative to the
leaf level (i.e. Pgl estimation), but was parameterised and validated with measurements
performed at the whole plant level. This approach results in the identification of parameters
that are analogous to those at the leaf level.
5.1. Light use eflciency , a
The parameter cr, calculated on the basis of incident radiation, was found equal to 2.9 pg
CO2 per pmol photon (Table zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH
1 ), or 0.066 mol CO2 per mol photon. This parameter is
analogous to the quantum yield (although based on incident rather than absorbed radiation),
for which Ehleringer and Bjijrkman (1977) gave a nearly constant average value for C3
species of 0.073 mol CO2 per mol photon. If we assume that Begonia leaf absorptance is
0.85, the quantum yield for Begonia would be equal to 0.066 mol CO, per mol photon i 0.85,
i.e. 0.078 mol CO2 per mol photon, which is very near the above-mentioned average value
for C3 plants.
P. Giaglaras et al. /Scientia
Horticulturae 63 (1995) 83-100
95
Table 2
Estimated light use efficiency values for COz concentration level
Mean COz
concentration
(CL1
I-‘)
U
(kg CO2 per wool
photon)
Standarderror
Residualstandard error
( P g CO* pm01
per photon)
(pgCO,s-‘perplant)
Degrees of
freedom
380
620
900
2.80
2.91
2.61
0.06
0.05
0.04
3.62
4.14
5.11
216
137
146
Bierhuizen et al. ( 1984) reported lower values for Begonia X hiemalis
zyxwvutsrqponmlkjihgfedcbaZY
cv. “Tiara” plants:
’ and 8.0 pg CO 2 J- ’ for 21% and 1% O,, which correspond respectively
to 0.029 and 0.04 mol CO, per mol photon and 0.04 mol COZ per mol photon (assuming
that 1 Joule me2 s-l is equal to 4.54 pmol photon m P-2s- ‘, Varlet-Grancher et al., 1989).
To verify if (Yis CO,-dependent,
the measured Pn data were separated according to the
COZ concentration level (400,600 and 900 ~11~‘). The model was fitted to each set of
data, with r, p and u constant (Table 1). The light-use efficiency changes slightly with
CO, concentration in the range 300-900 /..~ll- ’ (Table 2). The a value obtained at 900 ~1
1-l is smaller than that obtained at 600 ~1 l- ‘: 0.066 ( ~0.001) mol CO* per mol photon
and 0.059 ( &0.0009) mol CO1 per mol photon, respectively. This result could indicate
that CO2 saturation at low light levels was already achieved at the 600 ~1 1-l level for
Begonia plants and that higher CO* concentrations have a negative effect.
5.8 pg CO 2 J-
5.2. Carbon dioxide conductance, T
The parameter 7 of the model (2.9 pg CO2 s - ’ m - ’ l/.~l- ’ , Table 1) is analogous to the
total CO, conductance usually referred to as g,, (mol mm2 s-‘) in the literature (Ball,
1988). The value of T can be converted to mol m-* s- ’ by dividing it by 44 (e.g. r=
0.068 molm-*s- ‘). This value is significantly higher than the CO2 conductance that can
be deduced from the data of Bierhuizen et al. ( 1984). Replacing the Pgm, value reported
by Bierhuizen et al. in Eq. (3) with C= 300 ~11~‘, results in rvalues of 0.0097 mol m-*
s-’ and 0.0164 mol m-* SK’ for the 21% and 1% O2 levels, respectively. On the other
hand, Acock et al. ( 1978), using the same formulation for leaf net photosynthesis of
chrysanthemum
plants, found r= 0.12( 10.049)
molm-*s-l,
i.e. nearly twice the estimated value of T.
Assuming that mesophyll conductance to CO2 (g,,,) equals 2.5-4 times the total conductance to CO2 (g,,) and that the stomatal (g,,,) and the boundary layer conductances to
water (g,,,+) , are respectively 1.56 and 1.37 times the stomata1 (g,,,) and the boundary layer
conductance to CO, (gc,b), the estimated r value corresponds to g,,r ranging from 0.14
(g,,,,=4g,,)
to 0.17 (gc,nr=2.5gc,,) mol m-* s-‘, respectively. Values of g,,, and g,,, are
expected to vary with light, CO*, VPD and wind velocity near the leaf boundary layer.
Baille et al. ( 1994) reported surface resistances to water vapour transfer ( rw,J of a greenhouse Begonia X hiemalis crop increasing from 270 to 900 s m- ’ with VPD increasing
from 0.2 to 1.5 kPa, which correspond to a decrease from 0.165 to 0.05 mol m-* se1 in
terms of conductance. For the mean VPD value in the assimilation chambers (0.55 zyxwvutsrqponml
kPa) ,
96
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83- 100
their formula will give a surface conductance g’,,c= 0.124 molm-*s-i. This value is
smaller compared with the range of gw,rvalues given above, and this difference can be
explained by the higher boundary conductance inside the assimilation chamber. The estimated value for r did not change significantly when the model was fitted using the CO*dependent (Yvalues given in Table 2.
In Eq. (3) the CO2 compensation point was omitted. This simplification was adopted
because Pn data with saturating light and low CO2 concentration values needed for a precise
estimation of this parameter were not available. The effect of a constant compensation point
value (50 ppm) on the estimation of T was not very significant.
5.3. “Effective leaf area’ ‘, A,
The effective leaf area approach was able to simulate in a simple manner the dependence
of the whole plant gross photosynthesis on light extinction, leaf position and orientation
relative to incident light and leaf age. The results of the model at the whole plant level using
this approach were satisfactory. The value found for u was 0.15 m* per plant (Table 1).
With this value, Eq. (4) gives A, = 0.74A when A = 0.05 m* per plant, and A, = 0.37A
when A =0.25 m* per plant.
5.4. Total specijc dark respiration coefJicient, /3
An attempt was also made to compare our value of the specific respiration coefficient
(/3=0.45 pg CO2 g-‘s-l) with values given in the literature. An estimate of Begonia
dark respiration per unit leaf area was derived from the net photosynthesis model proposed
by Bierhuizen et al. ( 1984), by extrapolating the proposed equation for PPFD = 0. Multiplying this dark respiration value by the mean leaf area of the plants used in their experiments
(0.019 m* per plant) and dividing by the shoot dry matter (0.75 g per plant), resulted in
p = 0.98 pg CO,g- ‘s- ‘. This value is high compared with our value (0.45 pg CO2 g-i
s - ‘) . To explain this difference, we analysed dark respiration using McCree’s model
(McCree, 1970) :
Rd = gGR + mDM ,
(9)
where g is the growth respiration coefficient ( pg CO2 pug- ‘)zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO
, GR the growth rate (hg s- ’ )
and m the specific maintenance respiration rate (pg CO2 g - ’ s- ‘).
Equating Eq. (6) to Eq. (9) and solving for the p coefficient gives
p=gGRIDM ,+m=gRGR+m
(10)
where RGR is the relative growth rate ( pg g-i s-r ). Eq. ( 10) indicates that p includes
both the growth and maintenance respiration, and corresponds in fact to the specific respiration coefficient r, as defined by Thomley ( 1970).
From Eq. ( lo), the value of /3 is expected to vary proportionally to RGR during plant
growth. The RGR of the C4 and C5 (cv. “Line”) plants is high at early growth stages
(approximately 0.58 pg g - ’ s- ’ ) and tends to a constant value as the plants become larger
(approximately 0.23 pg gg ’ s- ‘). Taking into account that Bierhuizen et al. ( 1984)
measured Pn on young Begonia plants (A = 0.019 m* per plant), which are expected to
P. Giaglaras et al. /Scientia
Horticulturae 63 (1995) 83-100
91
have high RGR values, it can be concluded that the observed difference between the
estimated value of p (0.45 pg CO* g - ’ s - ’ ) and that estimated from the Bierhuizen et al.
(1984) model (0.98 pg CO* g-’ s-l), is due to growth stage and RGR differences. The
estimated value of p is a mean value for the overall production period of Begonia plants.
5.5. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Extrapolation tests
The 24% overestimation of the net photosynthesis found with the first extrapolation test
(Fig. 6) corresponds to data obtained with the cultivar “Heidi”. The fitting of the model
to those data suggests that the plants of this cultivar have lower light use efficiency ((Y=
1.64 f0.07 pg CO* per pmol photon), lower conductance to CO2 (T= 1.9 f
0.08 pg CO2 s-’ m-’ 1 ~1~‘), greater effective leaf area coefficient (a=
0.205 m2 per plant)
and
lower
specific
dark
respiration
(p = 0.305 f
0.025 /.Lg co* g-1 s-l>, compared with the values of the cultivar “Line” (Table 1).
Additional data from appropriate experiments are needed to verify the hypothesis that the
parameters of the model are subjected to variations owing to genetic variability of Begonia X hiemalis plants.
The validation test vs. plant leaf area showed that we cannot use the same value of u
(0.145 m* per plant) for leaf areas beyond 0.25 m* per plant. Curve fitting to the data shown
in Fig. 7, corresponding to plants with large leaf area (0.44 m2 per plant), would give
a=0.117 m* per plant.
The comparison between the predictions of the model with the in situ (glasshouse)
measurements (cv. “Line”) shows that air temperatures above 25°C and VPD values zyxwvutsrqpo
measured
-
--
0
100
nyxk.1 estimations
Eivxhuizcn et al. (1984)
200
300
400
cl
0
Pn
v alues
plant I
plant 2
500
600
lo (lmol mm2 s-l)
Fig. 10. Comparison of the predictions from Bierhuizen model (dashed line) and present model (continuous line)
with Pn data obtained in the assimilation chambers in the C range 35OAOO$l-’ for two Begonia X hienrnlis
plants (0, Plant 1 with A = 0.025 m* per plant and D&f,= 1.37 g per plant; 0, Plant 2 with A =
0.216 m* per plant and DM,= 17.94 g per plant).
98
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83400
greater than 1.5 kPa limit the assimilation rates of Begonia X hiemulis plants (Fig. 9).
Stomatal and non-stomatal effects (photorespiration, photochemistry) could explain this
observation (Comic et al. 1992). Recently, Baille et al. ( 1994) found that the transpiration
rate of a Begonia X hierndis crop decreases when VPD rises above 1.2 kPa. Thus, it could
be proposed that stomata1 closure was the main cause of the Pn overestimation by the model.
The available transpiration rate measurements at the whole plant level make the subject of
a study for the investigation of the above hypothesis and for the quantification of its effects
on Pn.
Neglecting the stomata1 response to VPD and the effect of CO? on cr (Acock, 199 1) in
the present net assimilation model could affect the model predictions if the application
ranges of temperature and VPD are large. However, if the model is used under the controlled
climate conditions of production greenhouses (i.e. similar to our assimilation chamber
conditions), the predictive value of the model can be considered as satisfactory.
The model of whole plant net photosynthesis presented here was also compared with the
model of Bierhuizen et al. ( 1984). The predictions of the two models are shown in Fig. 10,
for
two
plants
(A, =0.025 m* per plant,
DM ,, = 1.37 g per plant
and
A2= 0.216 m* per plant, DM s2 = 17.94 g per plant) and for a CO2 concentration of 370
~11~‘. As PPFD rises, the predictions of the presented model follow the measured values
better than the Bierhuizen model, which predicts light saturation at a lower value of PPPD.
This could be a specific characteristic of the “Tiara” cultivar, or the result of low light
acclimation of the plants studied by Bierhuizen et al. ( 1984).
6. Conclusions
The presented model for predicting Begonia net photosynthesis using four explanatory
variables (incident PPFD, CO2 concentration, leaf area and shoot dry matter) gives satisfactory results throughout the entire cultivation period and for a wide range of cultivars
(“Line”, “Heidi” and Rosalie group cultivars). Predictions of Begonia plants net photosynthesis are reliable in the ranges of climatic variables similar to those under which the
model was calibrated: T,E [ 18, 25]“C, Z,,E [0, 7001 pm01 photon m-* s-i, VPDE [0,
1.51 kPa and CE [ 350,950] ~11~‘.
The parameter values estimated from model fitting on Pn measurements from assimilation
chambers were similar to those reported in the literature for C3 plants. Crop history (climate
conditions, cultural practice, origin, etc.) did not seem to influence the net photosynthesis
rate at the whole plant level. Plant structure and leaf area effects were significant and the
“effective leaf area” approach was found convenient to account for these effects. Model
fitting on a wider range of leaf area should improve the estimation of the effective leaf area
and the influence of the “self shading effect” on the light-Pn conversion. Cultivar and
growth stage differences are probably at the origin of the observed discrepancy with the
previously reported Pn model for Begonia X hiemalis plants of Bierhuizen and al. ( 1984).
In conclusion, the present model can be used for the estimation of the carbon balance of
whole Begonia plants throughout their cultivation period in production greenhouses. It can
provide a basis for comparison of the photosynthetic characteristics of different BegoniaX hierndis cultivars at the leaf and/or at the whole plant scale for plant breeding
P. Giaglaras et al. /Scientia Horticulturae 63 (1995) 83400
99
purposes. It can also be applied in algorithms for climate control, to optimise CO2 enrichment
and artificial lighting. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Acknowledgements
The authors thank the ADEME ( Agence de 1’Environement et de la Maitrise de 1’Energie)
for the Ph.D. grant, the CNIH (Comite National Interprofessionel de 1’Horticulture) for
providing the Begonia plants, C. Sarrouy (INRA) for manufacturing the measuring systems,
and M. Tchamitchian and C. Gary (INRA) for their valuable comments on the manuscript.
References
Acock, B ., 1991. Modeling canopy photosynthetic response to carbon dioxide, light interception, temperature,
and leaf traits. In: Modeling Crop Photosynthesis-from Biochemistry to Canopy. CSSA Special Publication
19. American Society of Agronomy and Crop Science Society of America, pp. 41-55.
Acock, B., Charles-Edwards, D.A., Fitter, D.J., Hand, D.W. and Warren-Wilson, J., 1978. Modelling canopy net
photosynthesis by isolated blocks and rows of chrysanthemum plants. Ann. Appl. Biol., 90: 256-263.
Ball, J.T., 1988. An analysis of stomatal conductance. Dissertation Thesis, Stanford University, CA, 88 pp.
Baille, M., Baiile, A. and Laury, J.C., 1994. Canopy surface resistances to water vapourtransfer for nine greenhouse
ornamental (pot plants) crops. Sci. Hortic., 57: 143-155.
Bierhuizen, J.F., Bierhuizen, J.M. and Martakis, G.F.P., 1984. The effect of light and CO* on photosynthesis of
various pot plants. Gartenbauwissenschaft, 49(5-6): 251-257.
Bierman, W., 1988. Zusatzbelichtung bei Begonia-Elatior in Anh%ngigkeitvon der nattirlichen Strahlungsenergie.
Gartnerb. Gartenwelt, 45: 1973-1975.
Boote, K.J. and Loomis, R.S., 1991. The prediction of canopy assimilation. In: Modeling Crop Photosynthesisfrom Biochemistry to Canopy. CSSA Special Publication 19. American Society of Agronomy and Crop Science
Society of America, pp. 100-140.
Buchwald, H.H., 1987. A simulation model for planning and control of the potted plant production. Acta Hortic.,
203: 3949.
Challa, H. and Schapendonk, A.H.C.M., 1986. Dynamic optimization of CO* concentration in relation to climate
control in greenhouses. In: H.Z. Enoch and B .A. Kimball (Editors), Carbon DioxideEnrichment of Greenhouse
Crops. Vol. I, Status and CO2 Sources. CRC Press, Boca Raton, FL, pp. 147-159.
Challa, H. and Van Straten, G., 1991. Reflections about optimal climate control in greenhouse cultivation. In: Y.
Hashimoto and W. Day (Editors), Mathematical and Control Applications in Agriculture and Horticulture.
JFAC Workshop Series, No. 1.
Comic, G., Ghashghaie, J., Genty, B. and Briantais, J.-M., 1992. Leaf photosynthesis is resistant to a mild drought
stress. Photosynthetica, 27(3): 295-309.
Ehleringer, J. and Bjiirkman, O., 1977. Quantum yields for CO* uptake in C3 and C, plants. Dependence on
temperature, CO2 and 0, concentration. Plant Physiol., 59: 86-90.
Gary, C. 1988a. Etude et modelisation des effets a court terme du microclimat sur le bilan carbone de la tomate
en phase vegetative. These de Doctorat, Universite de Paris VI.
Gary, C. 1988b. Prise en compte des differentes sources d’erreurs et estimation de la precision dans la mesure des
&changes de CO1 en systeme ouvert. Photosynthetica. 22( 1): 58-69.
Hendriks, L., 1988. Kultursteuerung von Elatior-Begonien. Massnahmen an Produktionszielen ausrichten. Gartenb. Gartenwelt, 45: 1988-1992.
Leutscher, K.J., 1990. A crop growth simulation model zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO
for operational management support in pot plant production.
Agric. Syst., 33: 101-l 14.
Liebig, H.P., 1989. Models to predict crop growth. Acta Hortic., 248: 55-68.
100 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
P. Giaglaraset al. /Scientia Hom’culturae63 (1995) 83- 100
Longuenesse, J.J., Conus, G. and Sarrouy, C., 1982. Temp6rahue nocturne et photosynthkse. II. Une chambre
d’assimilation climatiske pour la mesure des echanges gazeux des plantes en&es. Agronornie, 2(8): 777781.
Mann, W. and Krug, H., 1989. Production planning-CO2 enrichment. Acta Hortic., 248: 201-206.
Mathijssen, E., 1985. Assimatiebelichting verkort Begoniateelt. Vakbl. Bloemisterij, 26: 24-26.
McCme, K.J., 1970. Anequation forthem~ofrespirationofwhitecloverplantsgrownundercontrolledconditions.
In: I. Setlik (Editor), Prediction and Measurement of Photosynthetic Productivity. Proc. IBP/PP. Technical
Meeting, T&on, 14-2 1 September 1969. Centre for Agricultural Publishing and Documentation, Wageningen,
pp. 221-229.
Powell, MC. and Bunt, A.C., 1985. Seasonal variation in the effect of day length on growth and flowering of
Rieger Begonia Schwabenland Red/. Sci. Hortic., 27: 143-152.
Simmonds, J. and Nelson, SD., 1988. Adventitious bud production on explants of BegoniaX hiemalisdepends
on the developmental state of the donor plant. Physiol. Plant., 73: 360-367.
Statistical Sciences, Inc., 1991. S-PLUS Reference Manual, 1991. Version 3.0. Statistical Sciences, Inc., Seattle,
WA.
Thomley, J.H.M., 1970. Respiration, growth and maintenance in plants. Nature, 227: 304-305.
Thomley, J.H.M., 1976. Mathematical Models in Plant Physiology. A Quantitative Approach to Problems in Plant
and Crop Physiology. Academic Press, London.
Varlet-Grancher. C., Gosse, G., Chattier, M., Sinoquet, H., Bonhomme, R. and Allirand, J.M., 1989. Mise au
point: rayonnement solaire absorb&ou intercepte par un couvert v6g&al. Agronomie, 9: 419-439.
Verberkt, H., 1990. Compactere plant door lange belichtingsduur. Vakbl. Bloemisterij, 35: 57-58.