DEVELOPMENT OF A CROP
SIMULATION MODEL FOR CUT-FLOWER ROSES
Environmental Horticulture, University of California, Davis, CA 95616-8587 USA
(currently at Department of Horticulture, The Ohio State University, Columbus Ohio)
A growth model for cut-flower rose shoots was developed and is currently being incorporated
into a crop simulation model. (This information was presented at the meeting of the International
Horticulture Society in 1991.)
The objective of this project is to develop a mathematical model for simulating rose crop growth
and development. The desire is to have a model which accurately predicts the response to
environmental factors. Such a model has a variety of uses including: automated optimization of
rose production using environment control computers, testing of various new horticultural
practices, and production optimization.
Materials and Methods
The model consists of a carbon budget which is responsive to fluctuations in temperature and
light. Various submodels are being developed. These will ultimately form the components of the
crop simulation model. The main component is a model for shoot growth and development (Lieth
and Pasian, 1991). It consists of submodels for leaf photosynthesis, growth, respiration, shoot
development (phenology) and partitioning.
Horticulturally, the above-ground portion of a rose crop consists of the "base" and the flowering
shoots. Dry weights of the leaves, stem, and flower of a typical shoot can be predicted by tracking
the amount of carbon throughout the plant.
Carbon enters the shoot through photosynthesis and translocation of carbohydrates from the base
(see figure below).
||Figure 1: Flow diagram of the shoot model:|
C = carbon pool
B = crop base (old wood)
L, S, F = leaf, stem, flower dry weights
TBup, TBdown = translocation coefficients
PCL, PCS, PCF = partitioning coefficients.
Rg,X= growth respiration of compartment X
Rm = maintenance respiration
PSYN = photosynthesis rate
Some of this carbon is exported while the rest is distributed within the shoot to the leaf, stem, and
flower compartments. The carbohydrate is used for producing new tissue (growth) and for
maintenance of existing tissues. Both of these processes involve respiration (growth and
maintenance) and result in the loss of some carbon from the tissue. In addition to these losses,
some carbon also leaves the shoot for storage in the base and in the roots.
Canopy (shoot) structure
The model simulates daily changes in the dry weight of each leaf on the shoot. New leaves are
simulated as being initiated according to the pattern observed on one shoot in an experimental
rose crop ('Cara Mia' roses, growing in ground beds). The individual leaf ages and relative
position on the shoot are tracked. The amount of light which shines on each leaf is calculated by
estimating how much is intercepted by the leaves higher up on the shoot. The leaf temperature is
currently assumed to be the same for each leaf and equal to the air temperature. Later this will be
replaced with a more accurate description.
Shoot photosynthesis and respiration
The photosynthetic rate of each leaf is computed using the model developed by Lieth and Pasian
(1990) multiplied by the leaf area, and summed to obtain the shoot photosynthesis rate. Losses of
carbon due to respiration are calculated using McCree's (1970) approach. Maintenance
respiration is estimated as a fixed percentage of the total dry matter (in this case 1%) per day.
Growth respiration is approximated to be 25% of the photosynthate assimilated by the shoot
during that day. This approach has been used for many plant species. However, no values have
been reported for roses; we are currently using approximate values until we can obtain more
Exchange of carbohydrate between shoot and base
Mor and Halevy (1979) showed that 80% of the photosynthate of the basal leaf at the base of the
shoot ended up in the growing shoot at the time when it had 4-6 expanded leaves. When the
flowering shoot had 10 expanded leaves and a 4 mm diameter flowering bud, only 2% of that
leaf's photosynthates were found in the flowering shoot. This information is used in this model in
the form of a linear equation for the rate of carbon translocation from the base, declining as the
shoot grows. No information is available on the movement of carbohydrates from other parts of
the base (such as stems); for now we are assuming that a further 80% of the estimated amount
comes from other sources in the base.
There is no information on the rates of carbohydrate movement from the flowering shoot into the
base of the crop. As a temporary approximation, we assume that there is no export to the base
until the day on which the last leaf has begun to unfold. From then on the percentage of the
shoot's assimilates exported is assumed to increase linearly to 80% by the day on which leaf
growth ceases. From then on it is assumed to be 80% until harvest.
Partitioning within the shoot
A non-destructive method (Pasian and Lieth, 1988) was used to estimate the dry matter of rose
leaves, stems, and flowers for three shoots every second day from breaking until flowering. Air
temperature and PAR at the top of the rose crop were recorded hourly. For the data collection,
day one is the day when it is first possible to measure the stem diameter 1 cm from its base
without removing or damaging any leaf material. At this stage the shoot is already 4 to 5 cm long.
Figure 2. Partitioning submodel.
The fractions of the amounts of carbohydrate partitioned to the three shoot components were
approximated by the ratio of the compartment dry weight increment to the shoot dry weight
increment. Figure 2 shows the observed dry matter partitioning data for one of the shoots
(circles). The partitioning strategy of the shoot appears to change at three times during the
flowering shoot growth period: 1) the day on which the flowering bud becomes visible (ca. day
10), 2) day on which the last leaf unfolds (ca. day 19), and 3) the day on which the leaves stop
growing (ca. day 27). The equations of the line segments in figure 2 represent the current dry
matter partitioning model
Results and Discussion
Since the model is currently in an early stage of development, extreme care needs to be taken in
interpreting the resulting simulations (a lot of portions are presently very rough approximations).
Figure 3. Shoot model simulation results
The observed and predicted dry weights of leaves, flower, and stem of the flowering shoot used
for model calibration fit very well (figure 3). Although there are some differences between the
observed and predicted dry weights, the overall performance of the model is satisfactory, given
the model complexity and the many rough approximations in the model.
Figure 4. Shoot model predictions for various
The shoot model allows determination of a wide variety of variables. Figure 4 shows the major
whole-shoot processes. For more detail see Lieth and Pasian (1991).
Crop simulation model
The crop simulation model involves the dynamic simulation of a multitude of shoots of various
ages and how these grow and develop in a canopy subdivided into several layers. Canopy layer
leaf area will be computed from the shoot information. This will be used to compute light
availability. Translocation of carbohydrates among shoots, base and roots will be investigated and
modeled. A mechanism will be built into the model to account for horticultural manipulation of
the canopy (e.g. harvest, pruning, pinching, etc).
Significant progress has been made in the development of a simulation model for rose crop
growth and development. A lot of work remains at developing information which is presently
lacking. Further data collection and model development are needed to make the various
approximate submodels more accurate. As this information is developed, the framework of the
crop simulation model will be developed based on the shoot growth and development model.
Acknowledgements: This research was supported, in part, by a grant from the Joseph Hill
McCree, K.J., 1970. An equation for the rate of respiration of white clover grown under controlled
conditions. In: Predictions and Measurement of photosynthetic productivity, Ed. I. Setlik, pp.
221-229. PUDOC, Wageningen, Netherlands.
Mor, Y. and A.H. Halevy, 1979. Translocation of 14 C-assimilates in roses. I. The effect of the age
of the shoot and the location of the source leaf. Physiol. Plant. 45:177-182.
Lieth, J.H. and C.C. Pasian, 1990. A model for net photosynthesis of rose leaves as a function of
photosynthetically active radiation, leaf temperature, and leaf age. J. Amer. Soc. Hort. Sci.
Lieth, J.H. and C.C. Pasian, 1991. A simulation model for the growth and development of flowering rose
shoots. Scientia Hortic. 46:109-128.
Pasian, C.C. and J.H. Lieth, 1988. Automated optimization of rose production: Nondestructive dry matter
estimation for the analysis of partitioning in rose shoots. Roses Incorporated Bulletin, March