c
.9 ..
v is. It
is l. 33:.

€11.75. SIN“:
..._ ,1.

g n.
.1... gang! _
. .92.»...QWMM

I vgl

. n. J

161in:
l h: .l
I: n I:

.4. ink

s
“402%”
4.3. A
it!!!
. .. .5
3.... as.
. .. {fairs
. ‘12:...3'lln‘li l‘:‘ u 111'
II... tha’¥th\—I‘tsg’i O) Ivl‘!, ‘DL
1.. . ; 2......333. If
.. 3.2...- .11.
latesiile.
.IJ:

llBRARY
Michigan State
University

This is to certify that the
dissertation entitled

MANIPULATING LIGHT AND TEMPERATURE FOR
ENERGY-EFFICIENT GREENHOUSE PRODUCTION OF
ORNAMENTAL CROPS

presented by

MATTHEW GEORGE BLANCHARD

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Horticulture

 

 

Major Professor’s Signature

/7 W200?

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 ICIProj/AocaPreleIRC/Dateoueindd

 

ABSTRACT
MANIPULATIN G LIGHT AND TEMPERATURE FOR ENERGY-EFFICIENT
GREENHOUSE PRODUCTION OF ORNAMENTAL CROPS
By
Matthew George Blanchard

The cost of heating fuel for greenhouse crop production is a significant expense for
growers in temperate climates. With the recent volatility in energy prices, some
ornamental plant growers have adjusted their production temperatures without knowledge
of its impact on crop timing or plant quality. The objectives of this research were to
quantify and model the influence of mean daily temperature (MDT) and photosynthetic
daily light integral (DLI) on flowering and plant quality of approximately 30 annual
bedding plants commonly grown in controlled environments. During one experiment, 18
species of bedding plants were grown in environmental growth chambers at constant air
temperature set points of 5, 7.5, 10, 15, 25, or 30 °C and under a photosynthetically active
radiation intensity of 180 umol’mmz's~l using a 16-h photoperiod. Nonlinear mathematical
equations were developed for each species to predict the effect of constant temperatures on
flowering rate (reciprocal of days to flower) and to estimate the base temperature (Tmin) at
which flowering rates were zero. The estimated Tmin ranged from 1.1 °C in Tagetes
patula L. to 9.9 °C in Angelonia angustifolia Benth. In separate experiments, the same
species were grown in glass-glazed greenhouses at constant air temperature set points of
14, 17, 20, 23, or 26 °C and under a mean DLI of3 to 19 mol-m“2'd'l using a 16-h
photoperiod. Flower development rates were predicted using a model that included a
linear MDT function with the Tmin multiplied by an exponential DLI saturation function.

Within the temperature range studied, flower development rate increased as MDT

increased, and in some species, development rate began to decrease at higher MDTs. For
example, under a mean DLI of 12 mol-m‘2°d_‘, as MDT increased from 14 to 23 °C, time
to flower of Petunia thbrida Vilm.-Andr. ‘Easy Wave Coral Reef” and ‘Wave Purple’
decreased from 51 to 22 d and 62 to 30 d, respectively. The estimated saturation DLI for
flower development rate in most species studied ranged from 8 to 15 mol°m‘2-d"'.

An additional study was performed with three species to validate models at
day/night (16 h photoperiod) temperature set points of 20/ 14, 18/18, 16/22 (mean of 18
°C), 24/18, 22/22, or 20/26 °C (mean of 22 °C). Flowering times were similar among
treatments with the same MDT but all species grown at 20/ 14 °C were 10 to 41% taller
than those grown at l6/22 °C. Using computer software that estimates energy
consumption for greenhouse heating (Virtual Grower version 2.51), energy inputs to
produce these species for spring market dates were estimated to be 3 to 42% lower at a +6
°C day/night temperature difference compared with a constant temperature.

In a final study, Impatiens hawkeri Bull. shoot-tip temperature was quantified
under several retractable greenhouse shade/energy screens during winter. An energy
balance model was developed that predicted shoot-tip temperature using cover (glazing or
screen) emissivity and five environmental parameters including dry-bulb, wet-bulb, cover
temperature, transmitted shortwave radiation (300 to 3,000 nm), and greenhouse air
velocity. At night and under an extended screen, the effective cover and shoot-tip
temperature were 0.8 to 6.9 °C and 0.5 to 2.3 0C higher, respectively, than without a
screen. Thus, screens extended overhead during cold nights can increase plant temperature

and accelerate development.

DEDICATION
I dedicate this dissertation to my loving grandparents, Eleanor and Harold George who

helped plant my first garden and started me down the horticulture path.

iv

ACKNOWLEDGMENTS

I would like to first acknowledge my major professor, Dr. Erik Runkle for his
encouragement, guidance, and support throughout these projects. His mentorship and
insight have been greatly appreciated during the past six years. He has provided many
opportunities for learning and professional development. I would also like to
acknowledge the members of my doctorate guidance committee, Drs. Arthur Cameron,
A]. Both, and Steve Harsh for their valuable knowledge, advice, and support during
these experiments. Thanks also to Drs. Paul Fisher and Jonathan F rantz for their
assistance and expertise.

Thank you to our floriculture greenhouse technician Mike Olrich for his
assistance in experimental setup, problem solving, support, and humor. I would also like
to thank the rest of the floriculture group and our undergraduate students for their support
and help in performing these experiments. Special thanks to past and present MSU
colleagues: Sandy Allen, Dr. Bridget Behe, Dr. Bert Cregg, Nate Durussel, Dr. Beth
Fausey, Grant Jones, Chris Herrmann, Dr. Roberto Lopez, Linsey Newton, Dr. Wook Oh,
Dr. Sonali Padhye, Lisa Voldeck, Dr. Ryan Warner, Aaron Warsaw, and Cathy Whitman.
All of these people and many others in the Department of Horticulture have made MSU
an enjoyable place to work.

Thank you to my family for all their love, support, and guidance. I would
especially like to thank my loving wife Laura, for her friendship, understanding, and
encouragement as I worked many long hours to earn this Ph.D. This would not have

been possible without her support.

PREFACE

In the US, annual bedding and garden plant production is the largest segment of
the floriculture industry with a USDA-reported wholesale value of $1.3 billion in 2008.
The majority of these crops are produced in heated greenhouses from January through
May so that flowering plants are available to consumers for purchasing in the spring.
During this time of year in temperate climates, high energy inputs can be required to
maintain a desirable greenhouse temperature, making fuel for heating costs one of the
largest floriculture production expenses (after labor costs). Rising and volatile energy
prices and shrinking profit margins have motivated many ornamental plant growers to
improve energy conservation and to minimize energy inputs for crop production in
controlled environments.

Energy-efficient and predictable commercial production of greenhouse crops
requires information on how species respond to the environment so that they can be
accurately scheduled for predetermined market dates. Plant growth and development in
response to the environment can be described quantitatively using mathematical
equations or models. Models that describe a biological process can be used to either
facilitate the understanding of a system or to predict a future condition. In controlled
environment experiments, it is often a challenge to deliver all possible levels of an
environmental factor. Therefore, models can be used to predict a response under
conditions that were not tested, but are within the typical range(s) of the parameter(s)
studied.

This research project quantified and modeled the influence of mean daily

vi

temperature (MDT) and mean photosynthetic daily light integral (DLI) on flowering and
plant quality of approximately 30 annual bedding plants grown in controlled greenhouse
environments. Nonlinear mathematical equations were developed for each species to
predict the effect of MDT on flowering rate (reciprocal of days to flower), the estimated
base temperature (Tmin; the temperature at which flowering development rate is zero),
and the estimated saturation DLI for highest flower development rate DLIsat, when
maximum development rate was 99%. During some experiments, models were validated
with independent data and the predicted responses were compared with observed data.
Flower development responses to temperature indicated that there is considerable
variability in thermal tolerance among genera. For example, the estimated Tmin ranged
from 1.1 °C in French marigold (T agetes patula L.) to 9.9 °C in angelonia (Angelom'a
angustifolia Benth.). Under a DLI of 10 mol-m‘Q'd‘l , some species such as cosmos
(Cosmos sulphureus L.) and dahlia (Dahlia thbrida Cav.) had a narrow temperature
range (14 to 17 °C) between Tmin and the temperature at which flower development rate
was greatest (Tom). Other species such as French marigold and black-eyed Susan
(Rudbeckia hirta L.), had a wide temperature range between Tmin and Topt, which was
estimated to exceed 25 °C. The different temperature responses among species could be
attributed to their indigenous habitat or criteria used for breeding selection. For example,
black-eyed Susan has a native distribution throughout temperate and semi-tropical
regions of North America. Black-eyed Susan had a Tmin for flower development of 4.6
0C and is apparently adapted to flower during the summer in habitats with variable
temperature conditions. In contrast, cosmos is native to semi-tropical regions of North,

Central, and South America, Africa, and Asia. Therefore, is not surprising that cosmos

Vii

had a higher Tmin for flower development of 7.2 °C.

A correlation was found between Tmin and the relative delay in flowering as
temperature decreased from 20 to 15 °C: species with a higher Tmin had a greater delay
than those with a lower Tmin- Greenhouse growers could use this information to group
species with similar environmental responses. Crops with only a slight flowering delay
when greenhouse temperature was lowered could be grouped together and grown at a
cool temperature set point without a considerable increase in production time (e. g., more
than 1 week).

The mathematical models presented in this dissertation also describe the influence
of DLI on flower development rate, which was modeled as a multiplier of the
temperature response. Flowering rate increased exponentially as DLI increased and
approached saturation (DLIsat) between 8 to 15 mol-m'Z-d‘l for most species tested.
Increasing the DLI above DLlsat did not accelerate flower development rate. This
information reinforces that supplemental lighting can have the largest effect on reducing
crop time when the ambient DLI is low. In many species, the DLI for the greatest crop
quality (e. g., maximum flower bud or inflorescence number) was higher than DLIsat. For
example, zinnia (Zinnia elegans Jacq.) had a DLIsat of 12.5 mol-m‘Z-d" for flowering
rate, but inflorescence number continued to increase under the DLI range studied (3 to 19
mol-m‘Z-d“). Commercial growers that are able to obtain a higher price for a higher
quality crop may consider using supplemental light to increase the DLI above DLlsat.

This research also compared different hybrids within a genus to determine if
models developed for one cultivar could be used to predict flowering responses in

another cultivar of the same species. Although hybrids within a species had similar

viii

growth responses to temperature and DLI, the models generated for one cultivar did not
accurately predict flower development rate, flower number, or plant height for the other
cultivars. The different environmental responses among cultivars could be caused by
genetic differences and/or by different breeding selection criteria such as grth form,
production time, and heat tolerance. Although it was necessary to develop unique models
for each crop, cultivars of the same species generally had a similar Tmin and DLlsat for
flower development rates. For example, petunia (Petunia thbrida Vilm.-Andr.) ‘Easy
Wave Coral Reef and ‘Wave Purple’ had an estimated Tmin and DLlsat within 1.8 °C
and 0.3 mol°m‘2-d“1, respectively. However, these cultivars had different cumulative
degree-day (°C-d) requirements to reach flowering. Therefore, in future research, the
main cultivar-specific parameter for flower development rate could focus on estimating
the thermal time for flowering. This could be determined by growing plants at a single
MDT, which would allow for rapid adaptation of these models to new cultivars. Further
research is warranted to test this approach.

The developed mathematical models were based on crops grown at constant
temperature set points. However, greenhouse growers often utilize different day and
night temperature set points during crop production. A study was performed with three
species to compare growth and flowering times at constant and a plus 6 °C differential
day/night temperature set points. This research quantified similar flowering times at
different day/night treatments with the same MDT and indicated that these crop models
could also be used to predict flowering at fluctuating temperature set points. This
response reinforces the paradigm that flowering rate is a function of the MDT and, within

limits, the effects of day and night temperature on progress towards flowering are equal.

ix

One strategy that is used by some greenhouses to save energy for heating is to
lower temperature set points during periods when the greenhouse heat loss is high
(typically when the temperature differential between inside and outside is high) and raise
set points when the heat loss is low. This environmental control strategy typically
delivers a higher day than night temperature, a higher temperature on sunny days and
lower on cloudy day, but maintains a target mean temperature during a pre-determined
period (e. g., 5 days). Additional research could investigate delivering a low MDT after
transplant for a specific duration (e. g., 2 weeks) and then growing at a higher MDT until
the plants are in flower. The opposite temperature strategy of beginning production at a
high MDT and ending at a low MDT could also be studied.

Others methods to reduce energy costs for greenhouse heating include energy
conservation methods, such as the installation of retractable thermal screens. Thermal
screens can be extended over a greenhouse crop from sunset to sunrise and reduce the
heat loss to the outside environment. This dissertation presents research information on
the influence of retractable greenhouse shade/energy screens on plant shoot-tip
temperature of New Guinea impatiens (Impatiens hawkeri Bull.). At night and under an
extended screen, the effective cover (glazing or screen) and shoot-tip temperature were
0.8 to 6.9 °C and 0.5 to 2.3 °C higher, respectively, than without a screen. An energy
balance model was developed that can be used to predict shoot-tip temperature under
different screen materials and environmental conditions. The results from this
experiment indicated that a retractable greenhouse screen has the potential to decrease
energy costs for heating and also to increase plant temperature and accelerate

development.

To facilitate the application of this research information by the greenhouse
industry, these crop development models will be integrated into a free computer software
program, Virtual Grower, created by Jonathan Frantz and colleagues of the USDA-ARS
Greenhouse Production Group in Toledo, Ohio. The outcomes of this interactive
software will include crop timing and the predicted energy consumption and cost based
on greenhouse location and user inputs. This tool can be used to identify the target crop
production temperature that results in the least amount of energy consumed on a per-crop
basis.

Collectively, the scientific information presented in this dissertation has added to
the understanding of how temperature and DLI influence plant growth and development
of many popular bedding plants. This new information can be used by the greenhouse
industry and educators to improve the predictability of flowering time of these
ornamental crops and to assist growers in determining energy-efficient production

practices.

xi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... xiv

LIST OF FIGURES ........................................................................................................ xvii

SECTION I

QUANTIFYING THE THERMAL F LOWERING RATES OF EIGHTEEN SPECIES

OF ANNUAL BEDDING PLANTS ......................................................................... 1

Abstract .......................................................................................................... 3
Introduction .................................................................................................... 4
Materials and Methods ................................................................................... 7
Results .......................................................................................................... 12
Discussion .................................................................................................... 14
Literature Cited ............................................................................................ 28

SECTION II

MODELING PLANT GROWTH AND DEVELOPMENT OF PETUNIA IN
RESPONSE TO TEMPERATURE AND PHOTOSYNTHETIC DAILY LIGHT

INTEGRAL .............................................................................................................. 32
Abstract ........................................................................................................ 34
Introduction .................................................................................................. 35
Materials and Methods ................................................................................. 39
Results .......................................................................................................... 45
Discussion .................................................................................................... 47
Literature Cited ............................................................................................ 58
SECTION III

THE INFLUENCE OF DAY AND NIGHT TEMPERATURE FLUCTUATIONS ON
GROWTH AND F LOWERING OF ANNUAL BEDDING PLANTS AND

GREENHOUSE HEATING COST PREDICTIONS .............................................. 62
Abstract ........................................................................................................ 64
Introduction .................................................................................................. 65
Materials and Methods ................................................................................. 68
Results .......................................................................................................... 72
Discussion .................................................................................................... 74
Literature Cited ............................................................................................ 83
SECTION IV
MODELING THE INFLUENCE OF GREENHOUSE SCREENS ON PLANT SHOOT-
TIP TEMPERATURE OF NEW GUINEA IMPATIENS ....................................... 87
Abstract ........................................................................................................ 89
Introduction .................................................................................................. 89
Materials and Methods ................................................................................. 93
Results and Discussion .............................................................................. 103
Literature Cited .......................................................................................... 120

xii

APPENDIX A
MODELING PLANT GROWTH AND DEVELOPMENT OF 29 BEDDING PLANT
SPECIES AND CULTIVARS IN RESPONSE TO TEMPERATURE AND

PHOTOSYNTHETIC DAILY LIGHT INTEGRAL ............................................. 125
Research Objective .................................................................................... 126
Materials and Methods ............................................................................... 126
Literature Cited .......................................................................................... 150
APPENDIX B
COMPARISON OF DIFFERENT LIGHT SOURCES ON GROWTH AND
F LOWERING OF BEDDING PLANTS ............................................................... 152
Research Objective .................................................................................... 153
Materials and Methods ............................................................................... 153
Results ........................................................................................................ 157

xiii

LIST OF TABLES

Table 1.1. Time from seed sow to transplant (TP), mean node number at TP, and
characteristics used to determine flowering date for 18 species of bedding plants in
two experimental replicates ................................................................. 21

Table 1.2. Parameter estimates for nonlinear models (Eqs. [1] and [4]) relating flowering
rate to mean daily air temperature in 18 bedding plant species. Parameter estimates
were used to generate Figs. 1 and 2. Base (Tmin) and maximum temperatures
(Tmax), are the temperatures at which flowering rates are zero (low and high
temperature, respectively), and the optimum temperature (Topt) is the temperature
where the maximum development rate occurs (Rmax). The “B” value defines the
skew of the function. Tmin, Topt, and Tmax are in °C. CI = confidence interval....22

Table 1.3. The effect of temperature on the number of nodes on the primary flowering
shoot at first open flower in 18 species of bedding plants. Plants were grown in
controlled environmental growth chambers under a 16-h photoperiod and a daily
light integral of 10.4 mol-m_2-d_’. Data were pooled between replications. L =
linear; Q =
quadratic ........................................................................................ 25

Table 2.1. Mean daily air temperature (MDT) at the indicated set points and mean
photosynthetic daily light integral (DLI) above benches, from transplant to
flowering of each cultivar, in glass-glazed greenhouse compartments during
experiments ..................................................................................... 54

Table 2.2. Parameters of nonlinear model relating rate of flowering of petunia ‘Easy
Wave Coral Reef” and ‘Wave Purple’ to mean daily air temperature [MDT (°C)]
and daily light integral [DLI (mol-m‘z'd‘lfl. Models (Eq. [6]) are in the form of:
l/d to flower = (-1 X Tmin X b; + b1 X MDT) X (l - exp(-e >< DLI)). Coefficients
for each petunia model were used to generate Figs. HA and 1.2A. ACI =
Asymptotic 95% confidence interval ....................................................... 54

Table 2.3. Parameters of stepwise regression analysis relating flower number, plant
height, leaf number increase or lateral stem length for petunia ‘Easy Wave Coral
Reef’ and ‘Wave Purple’ to mean daily air temperature [MDT (°C)] and daily light
integral [DLI (mol-m‘Z-d‘lfl. All models are in the form of: y = yo + aMDT +
bMDT2 + cDLI + dDLI2 + gMDT >< DLI. Models for ‘Easy Wave Coral Reef and
‘Wave Purple’ were generated using 159 and 200 observations, respectively.
Coefficients for each petunia model were used to generate Figs. 1.1B-D and 1.2B-
D ................................................................................................. 55

xiv

Table 3.1. Mean daily air temperature and daily light integral during experiments in Year
1 and 2. The day and night were 16 and 8 h, respectively .............................. 79

Table 3.2. Predicted relative amount of energy used for greenhouse heating to produce
three annual species grown at different day and night temperature set points in
different locations and finish dates. Heating inputs were estimated using Virtual
Grower software (US. Department of Agriculture, 2009a) and include time from
transplant to first flowering on 15 Mar., 15 Apr., or 15 May. Production time for
each species was calculated using the mean time to flower at 18/18, 20/14, and
16/22 °C or 22/22, 24/18, and 20/26 °C. Percentages were calculated by dividing
heating input by the highest input for each location and species. See materials and
methods for greenhouse and heating parameter inputs .................................. 79

Table 4.1. Characteristics of greenhouse screens used in experiment. Information
obtained from manufacturer (Ludvig Svensson, Inc., Charlotte, NC). The bottom
side of each screen was orientated towards the crop. Control plants were grown in a
greenhouse without a screen and were exposed to glass-glazing (emissivity: 0.9).
SWR= shortwave radiation, 300 to 3,000 nm; AL= aluminum; BL= black; E=
Empty space crossed with polyester threads; TP= transparent polyethylene ....... 1 12

Table 4.2. Abbreviations, symbols, descriptions, and units of parameters in the New
Guinea impatiens shoot-tip temperature model described in Eqs. [1-18] ........... 113

Table 4.3. Constants used with Eq. [13] (Gates, 2003)]14

Table 4.4. The effects of different screens on measured net longwave radiation (L WRne,)
exchange between New Guinea impatiens shoot-tip and the cover (glazing and
superstructure or screen) and the difference between measured shoot-tip
temperature (T517009 and dry-bulb temperature (Tdb) in glass-glazed greenhouses
during the night (1700 to 0800 HR) in winter in East Lansing, MI (lat. 43 °N) under
different air vapor-pressure deficits (VPDm-r). Each experimental period was 4 d.
The mean T db during the experiments was 19.5 to 20.2 °C. See Table 4.1 for a
description of screen materials. Talus“; outside air temperature; Tcover= cover
(glazing and superstructure or screen temperature) ..................................... 115

Table 5.1. Time from seed sow to transplant (TP), mean node number at TP, and
characteristics used to determine flowering date of bedding plants used in modeling

experiments .................................................................................. 1 34

Table 5.2. Mean daily air temperature (MDT) at the indicated set points and mean
photosynthetic daily light integral (DLI) above benches, from transplant to

XV

flowering of each species, in glass-glazed greenhouse compartments during
experiments .................................................................................. 1 3 6

Table 5.3. Parameter estimates for nonlinear model (Eq. [6] relating flowering rate to
mean daily air temperature [MDT (°C)] and daily light integral [DLI
(mol°m‘2-d‘1)]. Models are in the form of: l/d to flower = (-1 X Tmin X b; + b; X
MDT) X (1 - exp(-e X DLI)). CI= confidence interval ................................ 139

Table 5.4. Parameter estimates for the nonlinear model (Eq. [5] X Eq. [7]) relating
flowering rate in Dahlia thbrida ‘Figaro Mix’ to mean daily air temperature
[MDT (°C)] and daily light integral [DLI (mol-m‘Z-d‘lfl. CI: confidence
interval ......................................................................................... 140

Table 5.5. Parameters of stepwise regression analysis relating flower bud or
inflorescence number to mean daily air temperature [MDT (°C)] and daily light
integral [DLI (mol-m‘Z-d‘lfl. All models are in the form of: y = yo + aMDT +
bMDT2 + cDLI + dDLI2 + gMDT X DLI ............................................................ 141

Table 5.6. Validation of the flowering rate models, comparing the observed time to
flower with those predicted. Coefficients for the models are presented in Tables
5.3.and 5.4. r2 values and the slope and intercept were determined by performing
linear regression analysis on the predicted versus observed flowering
rates ............................................................................................ 144

Table 6.1. Influence of light source on time to flower, shoot dry weight, height, and on
the number of flowers, nodes, and branches at flowering in petunia, snapdragon,
and verbena. Plants were grown under a daily light integral (DLI) of 8 mol-m‘z-d'l
delivered either (1) entirely from high-pressure sodium (HPS) lamps (100% HPS),
(2) 4 mol-m‘z-d“ each from sunlight and HPS lamps, delivered separately (50%
HPS), and (3) entirely from natural sunlight (0% HPS) ................................ 158

xvi

LIST OF FIGURES

Figure 1.1. Observed and predicted flowering rates in 9 species of bedding plants as a
function of mean daily temperature based on Eq. [1] (panel B, D-H) and Eq. [4]
(panel A, C, and I) and parameter estimates from Table 2.2. Circles represent the
means of replication 1 and 2. Dashed lines represent predictions outside of the
observed data range. Data points represent treatment means and error bars
represent 95% confidence intervals. Tmax in panels B and H could not be estimated
from observed data and was fixed at 35.0 °C so Eq. [1] could be solved ................ 26

Figure 1.2. Observed and predicted flowering rates in 9 species of bedding plants as a
function of mean daily temperature based on Eq. [1] (panel A-C, F-H) and Eq. [4]
(panel D, E, and I) and parameter estimates from Table 2.2. Circles represent the
means of replication 1 and 2. Dashed lines represent predictions outside of the
observed data range. Data points represent treatment means and error bars
represent 95% confidence intervals. Tmax in panel A, C, and H could not be
estimated from observed data and was fixed at 35.0 °C so Eq. [1] could be
solved ........................................................................................... 27

Figure 2.1. The influence of mean daily temperature (MDT) and daily light integral
(DLI) on petunia ‘Easy Wave Coral Reef’ predicted flowering rate (l/d to
flowering) and time to flower (A), flower bud number (B), plant height (C), and
leaf number increase at flowering (D). The response surfaces were generated using
Eqs. [6] and [7] and coefficients in Tables 1.2 and 1.3. Each model was generated
using 159 observations ........................................................................ 56

Figure 2.2. The influence of mean daily temperature (MDT) and daily light integral
(DLI) on petunia ‘Wave Purple’ predicted flowering rate (1/d to flowering) and
time to flower (A), flower bud number (B), plant height (C), and lateral stem length
at flowering (D). The response surfaces were generated using Eqs. [6] and [7] and
coefficients in Tables 1.2 and 1.3. The models were generated using 200
observations .................................................................................... 57

Figure 2.3. Validation of the flowering rate models, comparing the observed time to
flower of petunia ‘Easy Wave Coral Reef’ and ‘Wave Purple’ grown at a mean
daily temperature of 17.3 to 22.1 0C and under a mean daily light integral of 14 to
19 mol-m‘Z-d‘l with those predicted by Eq. [6]. The dashed lines represent the 7-d
lower and upper boundary. Forty-five observations for each hybrid were used for
the validation study, where each symbol represents an individual plant. Numbers
represent the quantity of observations at each symbol. Coefficients for the

xvii

flowering rate models are presented in Table 2.2. r2 values were generated by
performing linear regression analysis on the predicted versus observed data ........ 58

Figure 3.1. The influence of temperature on time to flower (A), and inflorescence
number (B) and height (C) at flowering, in dahlia ‘Figaro Mix’ at constant and
fluctuating day/night (16 h/8 h) temperature set points. Vertical bars indicate
standard errors of treatment means. Mean separation by Tukey’s honestly
significant difference test at P S 0.01 ...................................................... 80

Figure 3.2. The influence of temperature on time to flower (A), and inflorescence
number (B) and height (C) at flowering, in French marigold ‘Janie Flame’ at
constant and fluctuating day/night (l6 h/8 h) temperature set points. Vertical bars
indicate standard errors of treatment means. Mean separation by Tukey’s honestly
significant difference test at P S 0.01 ...................................................... 81

Figure 3.3. The influence of temperature on time to flower (A), and inflorescence
number (B) and height (C) at flowering, in zinnia ‘Magellan Pink’ at constant and
fluctuating day/night (16 h/8 h) temperature set points. Vertical bars indicate '
standard errors of treatment means. Mean separation by Tukey’s honestly
significant difference test at P S 0.01 ...................................................... 82

Figure 4.1. Schematic illustration of the components of a leaf energy balance which
include shortwave radiation (SWR), longwave radiation (LWR), convective heat
transfer (Conv), and evaporative heat loss through transpiration (XE) ............... 116

Figure 4.2. Frequency of the difference between simulated New Guinea impatiens shoot-
tip temperature and measured shoot-tip temperature for day and night using data
from 24 d (13,824 observations). Plants were grown in glass-glazed greenhouses
during winter in East Lansing, MI (lat. 43 0N) under different vapor-pressure
deficits and with different screens extended over plants during the night (1700 to
0800 HR). Control plants were grown in a greenhouse without a screen ............ 116

Figure 4.3. Validation of the energy balance model, comparing simulated New Guinea
impatiens shoot-tip temperature with measured shoot-tip temperature for the day
[0800 to 1700 HR (A)] and night [1700 to 0800 (B)]. The validation included
measured data from 24 d and consisted of 5,184 observations for the day and 8,640
observations for night. Plants were grown in glass-glazed greenhouses during
winter in East Lansing, MI (lat. 43 °N) under different vapor-pressure deficits and
with different screens extended over plants during the night. Control plants were
grown without a screen above. r2 values were generated by performing linear
regression analysis on the simulated versus measured data ........................... 117

xviii

Figure 4.4. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East
Lansing, MI (lat. 43 °N) under different air vapor-pressure deficits (VPD) and with
different screens extended over plants during the night (1700 to 0800 HR). Control
plants were grown without a screen above. The air temperature set point was 20
°C. See Table 4.1 for a description of screen materials. Data in panels A, C, E, and
G and panels B, D, F, and H were collected during 20 to 24 Jan. and 25 to 29 Jan.,
respectively. SWR= shortwave radiation (300 to 3,000 nm) .......................... 118

Figure 4.5. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East
Lansing, MI (lat. 43 0N) under different air vapor-pressure deficits (VPD) and with
different screens extended over plants during the night (1700 to 0800 HR). Control
plants were grown without a screen above. The air temperature set point was 20
°C. See Table 4.1 for a description of screen materials. Data in panels A, C, E, and
G and panels B, D, F, and H were collected during 29 Jan. to 2 Feb. and 2 to 6 Feb.,
respectively. SWR= shortwave radiation (300 to 3,000 nm) ......................... 119

Figure 4.6. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East
Lansing, MI (lat. 43 °N) under different air vapor-pressure deficits (VPD) and with
different screens extended over plants during the night (1700 to 0800 HR). Control
plants were grown. without a screen above. The air temperature set point was 20
°C. See Table 4.1 for a description of screen materials. Data in panels A, C, E, and
G and panels B, D, F, and H were collected during 15 to 19 Feb. and 11 to 15 Feb.,
respectively. SWR= shortwave radiation (300 to 3,000 nm) ......................... 120

Figure 5.1. Time to flower from transplant of 150 plants each of Petunia thbrida
‘Dreams Neon Rose’ and Antirrhinum majus ‘Montego Burgundy Bicolor’ grown
under the same environmental conditions. Plants were grown in a glass greenhouse
at a mean daily temperature of 21 oC and under a mean daily light integral of 21
mol-m—z-d‘l with a 16-h photoperiod. Petunia was considered flowering when
individual plants had one flower with a fully open corolla. Antirrhinum majus was
considered flowering when individual plants had 2 flowers open on an
inflorescence. Dashed vertical lines indicate the predicted time to flower for each
species using the flower development rate models (Table 5.3) ........................ 146

Figure 5.2. The increase in flower development rate as the daily light integral (DLI)
increases in 12 species of bedding plants grown under a 16-h photoperiod. Circles
represent the estimated saturation DLI (light factor 20.99) for the shortest time to
flower for each species. Symbols are not presented if DLlsat is predicted to occur

above of the DLI range for which the models were developed ........................ 147

xix

Figure 5.3. The increase in flower development rate as the daily light integral (DLI)
increases in 12 species of bedding plants grown under a 16-h photoperiod. Circles
represent the estimated saturation DLI (light factor 20.99) for the shortest time to
flower for each species ...................................................................... 148

Figure 5.4. The increase in flower bud or inflorescence number as the daily light integral
(DLI) increases in 12 species of bedding plants grown at a constant temperature set
point of 20 °C and under a 16-h photoperiod. Predictions were determined using
the polynomial response surface equations presented in Table 5.5. Circles represent
the estimated maximum DLI (DLImax) for the greatest flower bud or inflorescence
number for each species. Circles are not presented if DLImax is predicted to occur
above of the DLI range for which the models were developed ........................ 149

Figure 5.5. The increase in flower bud or inflorescence number as the daily light integral
(DLI) increases in 12 species of bedding plants grown at a constant temperature set
point of 20 °C and under a 16-h photoperiod. Predictions were determined using
the polynomial response surface equations presented in Table 5.5. Circles represent
the estimated maximum DLI (DLImax) for the greatest flower bud or inflorescence
number for each species. Circles are not presented if DLImax is predicted to occur
above of the DLI range for which the models were developed ........................ 150

XX

SECTION I

QUANTIFYING THE THERMAL F LOWERING RATES OF EIGHTEEN SPECIES

OF ANNUAL BEDDING PLANTS

Quantifying the Thermal Flowering Rates of Eighteen Species of Annual Bedding

Plants

Matthew G. Blanchard1 and Erik S. Runkle2

Department of Horticulture, Michigan State University, East Lansing, MI 48824

Received for publication . Accepted for publication . We gratefully
acknowledge funding by Michigan’s plant agriculture initiative at Michigan State
University (Project GREEEN), the Michigan Agricultural Experiment Station, the
American Floral Endowment, the Fred C. Gloeckner Foundation, the USDA-ARS

F loriculture and Nursery Research Initiative, and greenhouse growers providing support
for Michigan State University floriculture research. We thank C. Raker & Sons for

donation of plant material and Mike Olrich for his greenhouse assistance.

1Ph.D. candidate.
2Corressponding author and to whom reprint requests should be addressed. Email

address: runkleer@msu.edu.

Abstract
The effect mean daily air temperature (MDT) on flowering rate (the reciprocal of

days to flower) was quantified for 18 species of annual bedding plants. Plants were
grown in environmental growth chambers at constant air temperature set points of 5, 7.5,
10. 15, 25, or 30 °C and under 180 umol-m‘Z-s"’ of light with a 16-h photoperiod.
Nonlinear mathematical equations were developed to predict the effect of MDT on
flowering rate and to estimate the base, optimum, and maximum temperatures (Tmin,
Topt, and Tmax), which are the temperatures at which flowering rates are zero (low
temperature), maximal, and zero once again (high temperature), respectively. The
estimated Tmin varied among species and ranged from 1.1 °C in French marigold
(Tagetes patula L.) to 9.9 °C in angelonia (Angelonia angustifolia Benth.). Topt and
Tmax could only be estimated for 8 to 10 species with the temperature range tested. TOpt
ranged from 19.1 °C in dahlia (Dahlia thbrida Cav.) to 28.0 °C in blue salvia (Salvia
farinacea Benth.), while Tmax ranged from 30.3 °C in snapdragon (A ntirrhinum majus
L.) to 31.7 °C in moss rose (Portulaca grandiflora Hook). Angelonia, browallia
(Browallia speciosa Hook), cosmos (Cosmos sulphureus Cav.), dahlia, and snapdragon
grown at 25 or 30 °C developed a mean of 2 to 7 more nodes before flowering compared
with plants grown at 515 °C. The results indicate that in many species, flowering rate in
response to MDT is asymmetrical around Tom and the temperature range between Tmin
and Top. is wider than that between Top, and Tmax. This information could be used to
improve the predictability of flowering time of these ornamental crops and to assist

growers in determining energy-efficient production temperatures.

Introduction

Scheduling greenhouse crops for specific market dates requires information on
how the environment influences plant growth and development (Heins et al., 2000).
Empirical models have been developed for several economically important floriculture
crops such as Chrysanthemum [Chrysanthemum X grandiflorum (Ramat.) Kitam.; Larsen
and Persson, 1999], Easter lily (Lilium longiflorum Thunb.; Erwin and Heins, 1990),
poinsettia (Euphorbia pulcherrima Willd. ex Klotz; Lui and Heins, 2002), petunia
(Petunia thbrida Vilm.-Andr.; Adams et al., 1998), and potted rose (Rosa L.;
Steininger et al., 2002) that predict crop time or quality under various environmental
conditions.

Plant developmental responses to temperature, such as flowering or leaf unfolding
time, are primarily influenced by the mean daily temperature (MDT) (Roberts and
Summerfield, 1987). The time required for the completion of a developmental stage can
be converted to a rate by calculating the reciprocal of time (e. g., 1/d). The rate of plant
development in response to MDT increases between the base and optimum temperature.
The base temperature (Tmin) is the species-specific temperature at or below which the
rate of progress towards a developmental stage is zero. Tmin has been estimated for
different developmental stages in several floriculture crops. For example, Tmin for leaf
unfolding and flower bud development rates (the reciprocals of days to unfold one leaf or
days to flower) in Easter lily were calculated to be 1.1 °C and 3.5 °C, respectively (Erwin
and Heins, 1990; Karlsson et al., 1988). Tmin for the flowering rate from visible flower

bud to open flower in campanula (C ampanula carpatica Jacq.) was calculated to be -1.8

°C, while potted rose had an estimated Tmin of 8.1 to 9.5 °C from budbreak to open
flower (Niu et al., 2001; Steininger et al., 2002).

As MDT increases above Tmin, development rate increases until a maximum rate
at the species-specific optimum temperature (Tom). For example, Tom for the flowering
rate of pansy (Viola Xwittrockiana Gams.) and geranium (Pelargonium Xhortorum
Bailey) was calculated to be 21.7 °C and 28.3 °C, respectively (Adams et al., 1997;
Armitage et al., 1981). When MDT >Topt, development rate decreases as MDT increases
and the rate becomes zero at the maximum temperature (Tmax). Estimation of Tmin» Topt,
and Tmax requires quantification of a developmental event at a wide range of MDTs and
therefore, these values have been estimated on a small number of floriculture crops
including Chrysanthemum (Larsen and Persson, 1999), dahlia (Dahlia pinnata Cav.;
Brandum and Heins, 1993), cineraria (Pericallis thbrida R. Nordenstam; Yeh et al.,
1999), and African violet (Saintpaulia ionantha Wendl.; Faust and Heins, 1993). For
example, a model developed for African violet predicted Tmin, Topt, and Tmax for leaf
unfolding rate to be 8.0 °C, 23.0 to 25.5 °C, and 30.8 °C, respectively (Faust and Heins,
1993)

Relationships between MDT and plant development rates have been modeled
using linear, quadratic, cubic, and exponential functions (Landsberg, 1977; Larson,
1990). For example, a linear model predicted that flowering rate in tickseed (Coreopsis
grandiflora Hogg ex Sweet. ‘Sunray’) increased from 0.013 to 0.028 as MDT increased
from 15 to 25 °C (Yuan et al., 1998). In Rieger begonia (Begonia Xhiemalis Fotsch), a
polynomial model predicted that as MDT increased from 13 to 21 °C, leaf unfolding rate

increased from 0.072 to 0.116 (Karlsson, 1992).

The response of a development rate to MDT has been described as either a
symmetrical (Pearson et al., 1993; Volk and Bugbee, 1991) or asymmetrical (Brondum
and Heins, 1993; Faust and Heins, 1993, 1994) peak-shape around Topt. For example, a
model generated for Chrysanthemum predicted that flowering rate had a symmetrical
response to MDT; development rate increased linearly as MDT increased from Tmin to
Topt, and then decreased at the same, but opposite slope from Topt to Tmax (Pearson et al.,
1993). In contrast, Brondum and Heins (1993) supposed that most biological responses
to temperature were asymmetrical and developed a model to predict flowering rate in
dahlia that increased from Tmin to Top. and then decreased from Tom to Tmx with a
greater slope. Scientific studies to determine flowering rates at MDTs above Topt have
been performed on few crops, and it is unknown if other species display a similar
asymmetrical temperature response around Topt (Smnmerfield and Roberts, 1991).

A useful outcome in the generation of crop models is the estimation of Tmin and
TOpt for flowering rate; thermal time is only accumulated at temperatures >Tmin and
from (Roberts and Summerfield, 1987; Wang, 1960). Therefore, determination of Tmin
and Topt are important for accurate thermal time predictions (Arnold, 1959; Wang, 1960;
Yeh et al., 1999). For example, calculation of thermal time in maize (Zea mays L.)
grown at 18.3 °C with a Tmin that was $5.6 °C different from the estimate of 7.2 °C
caused an error of $900 degree days (°C°d; Arnold, 1959). The estimated Tmin is also
important when quantifying the photothermal ratio (PTR) to predict plant growth and
quality. PTR is the ratio of radiant energy to thermal energy and is calculated as the
product of daily light integral (DLI, mol-m‘Z-d‘l) and °C-d above Tmin (Liu and Heins,

2002).

Tmin for flowering rate has been estimated for several flowering potted plants and
temperate herbaceous perennials, but estimates for ornamental annual species is lacking.
Notable exceptions include vinca (Catharanthus roseus L.), celosia (C elosia argentea L.
var. plumosa Voss), impatiens (Impatiens walleriana Hook.f.), geranium, petunia, red
salvia (Salvia splendens F. Sello ex Roem & Schult.), French marigold (Tagetes patula
L.), and pansy (Adams et al., 1996, 1998; Armitage, 1981; Mattson and Erwin, 2003;
Moccaldi and Runkle, 2007; Pietsch et al., 1995; Pramuk and Runkle, 2005). The
estimation of Tmin for additional annual species could be useful in the development of
crop models that predict flowering rates under different environment conditions. In
addition, estimates of Tmin could be used to determine which annual species tolerate low
production temperatures and identify energy-efficient growing strategies.

The objective of this study was to quantify the influence of MDT on flowering
time during the finish stage of 18 species of annual bedding plants, and from that data, to

develop mathematical models that estimate Tmin and Topt for flowering rates.

Materials and Methods

Seeds of African marigold (T agetes erecta L. ‘Antigua Primrose’), angelonia
(Angelonia angustifolia Benth. ‘Serena Purple’), black-eyed Susan (Rudbeckia hirta L.
‘Toto Rustic’), blue salvia (Salviafarinacea Benth. ‘Victoria Blue’), browallia
(Browallia speciosa Hook. ‘Bells Marine’), cosmos (Cosmos sulphureus Cav. ‘Cosmic
Orange’), dahlia (Dahlia thbrida ‘Figaro Mix’), dianthus (Dianthus chinensis L. ‘Super
Parfait Raspberry’), French marigold ‘Janie Flame’, gazania [Gazania rigens (L.) Gaertn.

‘Daybreak Bronze’], moss rose (Portulaca grandiflora Hook. ‘Margarita Apricot’),

pentas [Pentas lanceolata (F orssk.) Deflers ‘Graffiti Lavender’], petunia ‘Dreams Neon
Rose’ and ‘Wave Purple’, snapdragon (Antirrhinum majus L. ‘Montego Orange
Bicolor’), verbena (Verbena thbrida Groenl. & Ruempl. ‘Quartz Waterfall Mix’), viola
(Viola cornuta L. ‘Sorbet Plum Velvet’), and zinnia (Zinnia elegans Jacq. ‘Dreamland
Coral’) were sown in plug trays [288-cell size (6-ml volume)] by a commercial
greenhouse (C. Raker & Sons, Litchfield, MI). After germination, plugs were received at
Michigan State University (MSU) and were grown in a controlled environmental growth
chamber (TC-2; Environmental Growth Chambers, Chagrin Falls, OH) at a temperature
set point of 20 °C. A 16-h photoperiod was provided by 215-W cool-white fluorescent
(CWF; F96T12CWVHO; Philips, Somerset, NJ) and 60-W incandescent lamps (INC;
Philips), at a CWF:INC (by W) of 3.6, and at an intensity of 180 p.mol-m"2-s‘l at plant
height. All plugs were thinned to one seedling per cell. During the plug stage, plants
were irrigated as necessary with well water acidified with H2804 to a titratable alkalinity
of 140 mg-L‘l CaCO; and containing 95, 34, and 29 mg-L‘l Ca, Mg, and S, respectively.
The water was supplemented with a water-soluble fertilizer providing (mg-L") 62 N, 6
P, 62 K, 7 Ca, 0.5 Fe, 0.3 Cu, Mn, and Zn, 0.1 B and Mo (MSU Well Water Special;
GreenCare Fertilizers, Inc., Kankakee, IL).

When seedlings were ready for transplant [16 to 44 d after seed sow, depending
on species (Table 1.1)], plugs were transplanted into 10-cm round plastic containers (480-
ml volume) filled with a commercial soilless peat-based medium (Suremix; Michigan
Grower Products, Inc., Galesburg, MI). The mean node number at transplant for each
species is presented in Table 1.1. Eight plants of each species were randomly assigned to

treatments and grown in controlled environmental growth chambers (TC-2;

Environmental Growth Chambers) at constant air temperature set points of 5, 7.5, 10, 15,
25, or 30 °C and under the light parameters previously described. Before plants were
transferred to 5, 7.5, or 10 °C, they were grown for 1 week at 15 °C followed by 1 week
at 10 °C to acclimate plants to the low temperatures.

The experiment was performed twice with each species and the time from seed
sow to transplant was the same or $7 d between replications (Table 1.1). Species in
which a treatment elicited 250% plant death or plants required 2170 d to flower in the
first replication were not grown at those temperatures during the second replication.
Plants were top irrigated as necessary with well water that was acidified as described
previously. The water was supplemented with a water-soluble fertilizer providing
(mg-L4) 125 N, 11 P, 126 K, 13 Ca, 1 Fe, 0.5 Cu, Mn, and Zn, 0.1 B and Mo (MSU Well

Water Special; GreenCare Fertilizers, Inc.).

Environmental Monitoring

Air temperature was independently measured in each chamber by an aspirated,
shielded thermocouple (0.13-mm type E; Omega Engineering, Stamford, CT) positioned
at bench height. At two temperature treatments, the PPF was measured by a quantum
sensor (LI-1908A; LI-COR, Inc., Lincoln, NE) positioned 16 cm above the height of the
containers. The height was determined to be representative of the canopy height for the
species grown. For treatments that did not contain a quantum sensor, the PPF was
measured weekly at 25 cm above the bench with a line quantum sensor (Apogee
Instruments, Inc., Logan, UT). Bulbs were replaced or the height of the lamp loft was

adjusted to maintain a PPF of 180 mol'm‘z'd‘l. In each temperature treatment, a

thermocouple (0.13-mm type E; Omega Engineering) was inserted 0.5 cm below the
shoot tip of five different plants and the actual plant temperature was recorded.
Thermocouples were repositioned weekly as plants developed.

Environmental measurements were collected every 10 s and 10-min means were
recorded by data loggers (CR10X; Campbell Scientific, Logan, UT). Mean daily plant
temperature for both replications at 5, 7.5, 10, 15, 20, 25, and 30 °C was +1.9, +1.2, +0.9,
+1.5, +0.3, +0.3, and —0.2 0C relative to mean daily air temperature, respectively. In
each treatment, water vapor was injected into the air if the vapor-pressure deficit (VPD)
was >08 kPa. The actual mean VPD at 5, 7.5, 10, 15, 20, 25, and 30 °C for both

replications was 0.5, 0.4, 0.4, 0.9, 0.7, 0.8, and 1.7 kPa, respectively.

Data Collection and Analysis

The date of first open flower was recorded and time to flower was calculated for
each plant. Plants were considered in flower according to individual flowering
characteristics for each species (Table 1.1). When each plant flowered, the number of
nodes on the primary shoot below the first open flower was recorded. Data were
analyzed using the calculated MDT for each plant from transplant to the date of
flowering. Flowering time data were converted to flowering rates.

A nonlinear model was used to describe the relationship between the flowering

rate and MDT for each species (Landsberg, 1977; Reed et al., 1976):

1/d to flower = A X (MDT — Tmin) X (Tmax - MDT)B [1]

10

Where A : Rmax / ((Topt — Tmin) x (Tmax _ Topt) B) [2]

and B = (Tmax _ Topt) / (Topt — Tmin) [3]

where MDT = mean daily air temperature (°C), Tmin and Tmax are the minimum and
maximum temperatures, respectively. When MDT is ST min or 2Tmax, development rate
is zero. TOpt is the temperature where the maximum development rate occurs (Rmax) and
the “B” value defines the skew of the function. This asymmetrical model was chosen
because it describes a biological response to temperature, such as net photosynthesis
(Neilson et al., 1972; Reed, 1976). With this function, a temperature-dependent
promotion of development occurs when Tmin < MDT 5 TOpt and a temperature-dependent
inhibition of development occurs when Topt < MDT < Tmax (Larsen, 1990). This
function has also been used to model the influence of temperature on flowering rate in
dahlia (Brandum and Heins, 1993) and leaf unfolding and leaf expansion rate in African
violet (Faust and Heins, 1993, 1994). In angelonia, dianthus, gazania, pentas, and viola,
Tmax could not be estimated from the observed data and was fixed at 35.0 °C so that the
nonlinear model could be solved.

In African mari gold, black-eyed Susan, French marigold, petunia, and zinnia, Topt
and Tmax could not be estimated from the observed data using Eq. [1] because there was
not enough data points for the nonlinear model to converge. Therefore, an exponential
function was used to describe the relationship between flowering rate and MDT (Larsen,

1990):

11

l/d to flower = Rmax X (1 - exp(—C X (MDT - Tmin» [4]

where Tmin is the temperature at or below which the development rate is zero and Rmax is
the maximum development rate. This function has been used to model leaf unfolding and
flowering rates in Chrysanthemum (Hidén and Larsen, 1994; Larsen and Hidén, 1995;
Larsen and Persson, 1999) and cineraria (Larsen, 1988, 1989).

Parameter estimates (Tmin, Topt, Tmax, Rmax, and C) for the nonlinear functions
(Eqs. [1] and [4]) were estimated with the nonlinear regression procedure (N LIN) of SAS
(version 9.1; SAS Institute, Cary, NC). Initial parameter estimates were obtained from
graphs of the observed data. Models were generated using 70 to 109 observations for
each species. After the nonlinear models were generated, R2 values were determined by
performing linear regression analysis on the predicted versus observed data as
recommended by Maceina and Pereira (2007). Data for the number of nodes at flower
were pooled between replications and were analyzed using SAS mixed-model procedure
(PROC MIXED), and pairwise comparisons between treatments were performed using

Tukey’s honestly significant difference test at P $0.05.

Results

At least 50% of plants died when African marigold, black-eyed Susan, and dahlia
were grown at 5 °C; blue salvia, browallia, cosmos, pentas, and zinnia were grown at
57.5 °C; and angelonia and rose moss were grown at 510 °C. At 30 °C, plants of
browallia, dahlia, and verbena had 250% death. Petunia ‘Wave Purple’ and verbena

grown at 5 °C continued to develop new leaves, but had a low flowering rate and plants

12

were removed from the treatment after 258 d. Cosmos grown at 30 oC had 13% death
and, although the remaining plants unfolded new leaves, only 38% of plants had a visible
inflorescence after 65 d.

The coefficients of determination (R2) for the nonlinear flowering rate models
ranged from 0.74 to 0.94 in the 18 species studied (Table 1.2). In some species,
variability in flowering time was high when plants were grown at an MDT near Tmin or
>Topt. For example, in black-eyed Susan (Tmin = 4.0 °C), flowering rate in plants grown
at 5 or 7.5 °C ranged from 0.0043 to 0.01], while flowering rate at 25 °C varied by only
0.0085. In all species, the rate of flowering increased as MDT increased until Topt (Fig.
1.1 and 1.2). For example, flowering rate in dianthus increased from 0.0048 at 6.0 0C to
0.033 at 26.0 °C.

The estimated Tmin where flowering rate is zero varied among species and ranged
from 1.1 °C in French marigold to 9.9 °C in angelonia (Table 1.2). Species that had a
Tmin 55.0 °C were African marigold, black-eyed Susan, dianthus, French marigold,
gazania, petunia ‘Dreams Neon Rose’, snapdragon, and viola. Those in which Tmin >5.0
°C were angelonia, blue salvia, browallia, cosmos, dahlia, moss rose, petunia ‘Wave
Purple’, verbena, and zinnia.

Among the species in which TOpt could be estimated, it ranged from 19.1 °C in
dahlia to 28.0 °C in blue salvia (Table 1.2). For the species in which the Topt could be
estimated, dahlia had the lowest Rmax at 0.0204, while viola had the highest Rmax at
0.0831. At the MDT range used in this study, sufficient data existed to model the
response of flowering rate to MDT at >TOpt for only seven species. The estimated Tmax

for blue salvia, browallia, cosmos, dahlia. moss rose, snapdragon, and verbena ranged

l3

from 30.3 to 31.7 °C. In these species, the rate of flowering decreased rapidly from TOpt
to Tmax.

As temperature decreased, node number at first flowering in African marigold,
angelonia, blue salvia, browallia, cosmos, dahlia, dianthus, moss rose, petunia ‘Wave
purple’, snapdragon, verbena, and zinnia decreased linearly, quadratically, or both (Table
1.3). Plants of angelonia, browallia, cosmos, dahlia, and snapdragon grown at 25 or 30
°C developed a mean of 2 to 7 more nodes before flowering compared with plants grown
at 315 °C. There were no significant differences in node number among treatments for

black-eyed Susan, French marigold, gazania, petunia ‘Dreams Neon Rose’, and viola.

Discussion

The nonlinear models used to predict Tmin’ Topt, and Tmax were selected because
they described a biological response to MDT and had relatively high coefficients of
determination. Previous studies have used a linear model to describe the relationship
between MDT and development rate (Karlsson, 1988; Niu et al., 2001; Pietsch, 1995).
Exponential models were used in this study because plots of the observed data for each
species indicated that at higher temperatures, flower development rate approached
saturation, and in some species became maximal. In addition, many previous studies that
used linear models were not performed at high temperatures to quantify development rate
near Topt, and therefore a linear function was appropriate.

Our flowering rate models were generated with data from plants grown at a
relatively wide temperature range, from 5 to 30 °C. Although these crops are rarely

grown at <10 °C during commercial greenhouse production, the low temperatures used in

14

our experiments were included to improve the predictions of Tmin- Our estimates for
Tmin are comparable with previous published flowering models on the same species. For
example, we estimated that dahlia had a Tmin of 5.6 °C, which is only 0.4 °C higher than
the Tmin reported by Brandum and Heins (1993) for development rate from visible flower
bud to open flower. Mattson and Erwin (2003) predicted that petunia ‘Dreams Rose’ and
‘Wave Purple’ had a Tmin 2.4 and 0.4 °C lower, respectively, than our estimates.

Among the 18 species investigated, the estimated Tmin for flowering ranged from
1.1 to 9.9 °C, which indicates the variability in thermal tolerance among genera. Tmin
can be used to categorize species according to their temperature response; crops can be
considered tolerant of or sensitive to low temperature. For example, species such as
French marigold and snapdragon had a Tmin <5.0 °C and could be described as low
temperature-tolerant. A Tmin <5.0 °C for flowering rate has also been calculated for
other ornamental crops such as blanket flower [(Gaillardia X grandiflora Van Houtte), 3.3
°C; Yuan et al., 1998], Shasta daisy ([Leucanthemum Xsuperbum Bergman ex .1. Ingram),
—3.4 °C; Yuan et al., 1998], cineraria (1.7 °C; Yeh et al., 1999), and black-eyed Susan
[(Rudbeckiafulgida Ait.), -1.3; Yuan et al., 1998].

We can categorize low temperature-sensitive crops as those that had a Tmin >5.0
°C, which includes angelonia and blue salvia. If these crops are grown at <5.0 °C for an
extended period of time, plant development ceases and chilling injury or death could
occur. Examples of additional ornamental crops that had a reported Tmin >5.0 °C for
flowering rate include tickseed (6.8 °C; Yuan et al., 1998), rose mallow [(Hibiscus

moscheutos L.), 12.2 °C; Wang et al., 1998], geranium (8.7; Armitage et al., 1981), red

15

salvia (7.0 °C; Moccaldi and Runkle, 2007), and potted rose (8.1 to 9.5 °C; Steininger et
al., 2002).

The estimated Tmin, Topt, and Tmax indicate the variation among species in the
temperature range between where flowering rate is zero and maximum. Species that had
a calculated difference between Tmin and Tom of 14 to 17 °C include angelonia, cosmos,
dahlia, and moss rose; 18 to 20 oC include blue salvia, browallia, pentas, and verbena;
and 22 to 24 °C include dianthus, gazania, snapdragon, and viola. In African marigold,
black-eyed Susan, French marigold, petunia ‘Dreams Neon Rose’, and petunia ‘Wave
Purple’, Topt could not be estimated, and therefore, the temperature range between Tmin
and Tom is >25 °C. The different temperature ranges among species could be related to
their indigenous habitat (Jones, 1992) or criteria used for breeding selection. The
identification of crops that develop at a wide temperature range could be used by breeders
to improve low and high temperature tolerance (Summerfield et al., 1991).

Among the species in which TmX could be estimated, the calculated difference
between Topt and Tmax was 3 to 5 °C in blue salvia, browallia, and snapdragon; 7 to 8 °C
in cosmos, moss rose, and verbena; and 11 0C in dahlia. These results indicate that in
these species, flowering rate in response to MDT is asymmetrical around Topt and the
temperature range between Tmin and Topt is considerably wider than the range between
Topt and Tmax. Flowering rate models for other ornamental crops have described a
similar asymmetrical response to temperature. For example, a model developed for 30
Chrysanthemum cultivars predicted Top, to be 14.2 °C >Tmin and 9.2 °C <Tmalx (Larsen

and Persson, 1999). In cineraria, Tom was estimated to be 20.6 °C >Tmin and 14.8 °C

16

<1" max (Yeh et al., 1999), while dahlia had a Topt 19 °C >Tmin and 8.9 °C <Tmax
(Brandum and Heins, 1993).

Our Tmin estimations are for plants grown under a mean DLI of 10.4
mol'm‘z-d“', but in some species, DLI could influence Tmin- For example, Tmin in
celosia and impatiens decreased from 11.7 to 10.2 °C and 7.5 to 4.3 °C, respectively, as
DLI increased from 5 to 15 mol-m‘Z-d“l (Pramuk and Runkle, 2005). Similarly in vinca,
the mean Tmin decreased from 10.2 to 7.2 °C as DLI increased from 18 to 30
mol-m“2-d‘l (Pietsch et al., 1995). In other species such as French marigold and red
salvia, DLI had little or no affect on Tmin (Moccaldi and Runkle, 2007). The Tmin for
some species could decrease as DLI increases because a higher DLI could increase plant
temperature. Faust and Heins (1997) reported that vinca shoot-tip temperature increased
by 1.7 °C as irradiance from high-pressure sodium lamps increased from 0 to 100
pmolm‘Z-s“.

As MDT decreased from Topt to Tmin, flowering rate decreased; however this
response to temperature varied among species. For example, our models predicted that as
temperature decreased from 20 to 15 °C, time to flower increased by 4 to 8 din French
marigold, dahlia, petunia ‘Dreams Neon Rose’, snapdragon, and viola; 11 to 18 d in
African marigold, cosmos, dianthus, gazania, moss rose, petunia ‘Wave Purple’, verbena,
and zinnia; and 20 to 38 d in angelonia, black-eyed Susan, blue salvia, browallia, and
pentas. The relative delay in flowering as temperature decreased from 20 to 15 °C was
significantly correlated (P 5 0.001) with Tmin, and species with a high Tmin had a greater
delay than those with a low Tmin- For example, viola had an estimated Tmin of 4.1 °C

and a 4—d increase in time to flower at 15 °C versus 20 °C, while pentas had a Tmin of 9.3

17

°C and 32-d increase in flowering time at 15 °C versus 20 °C. This information indicates
that during greenhouse production, changing temperature set points can influence the
scheduling of crops differently. If MDT is lowered from 20 to 15 °C, the time required to
produce a crop would increase the most in species with a high Tmin°

The decreased flowering rate at an MDT >Topt can be referred to as heat delay
(Wang et al., 2008). High temperature can delay flowering by inhibiting flower
induction, initiation, and/or development (Warner and Erwin, 2006). In some species that
exhibited heat delay at an MDT >Topt, plants developed more nodes before flowering
compared to plants grown at an MDT <Topt. For example, snapdragon had an estimated
Topt of 25.7 °C and developed a mean of 2.8 more nodes before flowering at 30 °C versus
25 0C. Warner and Erwin (2005) also reported that the number nodes below the first
open flower increased in calendula (Calendula oflicinalis L.), impatiens, and wishbone
flower (T orenia fournieri Linden ex. E. F ourn) as temperature increased from 20 to 32
°C. The higher node number before flowering indicates that in these species, high
temperatures delayed flowering developmentally.

Under the environmental conditions provided in this study, a high percentage of
browallia, dahlia, and verbena died when grown at a constant 30 °C. Semeniuk (1975)
reported that as MDT increased from 20 to 31 °C, flowering rate in browallia decreased
and plants grown at the highest MDT were stunted with abnormal flowers and failed to
develop seeds. African violet grown at a 30 °C day temperature had chlorotic leaves and
no inflorescence development, while geranium grown at 32 °C developed chlorotic leaves
and died (Armitage et al., 1981; Faust and Heins, 1994). Plant stress at high temperature

results from a decline in normal protein synthesis, which is substituted by the synthesis of

18

heat shock and stress proteins (Moseley, 1997). In addition to biochemical changes,
exposure to high temperature can decrease cell membrane therrnostability and result in
electrolyte leakage (Wang et al., 2008).

The temperature that had the highest rate of flowering may not be similar to the
Topt for other physiological or developmental processes. For example, the estimated Topt
for flowering rate is 227.2 °C in angelonia, French marigold, petunia ‘Dreams Neon
Rose’, and petunia ‘Wave Purple’, but Tom for net photosynthesis is 19.8 to 20.8, 15.5
°C. 14.1 °C, and 20.0 °C, respectively (Miller et al., 2001; Niu et al., 2006; van Iersel,
2003). In Easter lily, flowering rate had a Topt of 26.0 °C, but maximum leaf unfolding
rate occurred at >30 oC (Erwin and Heins, 1990; Karlsson et al., 1988). The Tom for
flowering rate also may not correlate with the temperature that elicits the highest plant
quality. Pietsch et al. (1995) calculated that TOpt for flowering rate in vinca was :35 °C,
but flower diameter was 12 to 30% greater at 25 °C versus 30 °C. Similarly, in impatiens
grown at 14 to 26 °C and under 15 mol'm‘Z-d‘l, flowering rate was highest at 26 °C, but
as MDT decreased, flower number, flower diameter, and dry weight at flowering
increased by 141%, 25% and 52%, respectively (Pramuk and Runkle, 2005). These
results collectively indicate that there can be a trade-off between fast cropping time and
plant quality. A growing temperature that elicits the shortest time to flower may result in
a crop that is poor quality and unmarketable.

These experiments were performed at constant temperature set points to allow
modeling of the data without possible interactions between day and night temperatures.
These models may not be valid under conditions when the day or night temperature is

<Tmin or >Tmax. For example, flower inrtratron and development in pomsettra was

19

delayed when night temperature was 27 to 30 °C regardless of MDT (Berghage, 1989).
Fluctuating day/night temperature studies with other crops such as dahlia (Brandum and
Heins, 1993), pansy (N iu et al., 2000), and vinca (Pietsch et al., 1995) indicated that if the
day and night temperatures were between Tmin and Tmax then flowering time is
controlled by MDT. For example, Brendum and Heins (1993) created 25 factorial
day/night treatments by moving plants among temperatures of 10, 15, 20, 25, and 30 °C
and quantified that dahlia development rate was related to MDT. These results
collectively suggest that our calculated Tmin values would be similar for plants grown at

constant and fluctuating temperature regimens.

20

Table 1.1. Time from seed sow to transplant (TP), mean node number at TP, and
characteristics used to determine flowering date for 18 species of bedding plants in two

experimental rQlicates.

 

Time from Mean
seed sow to node no.

 

Species TP (d) at TP Flowering characteristic
. . , . . , 1 inflorescence with
African marigold Antigua Primrose 19 or 23 6.0 250% of petals re flexed

Angelonia ‘Serena Purple’ 40 5.8 :3 flowers open on an
inflorescence
. . , . 1 inflorescence with 1
Black-eyed Susan Toto Rustic 31 5.5 whorl of petals re flexed
Blue salvia ‘Victoria Blue’ 34 4.0 3 flowers open on an
inflorescence
Browallia ‘Bells Marine’ 40 5.9 1 flower open
, . , 1 inflorescence with 1
Cosmos Cosmic Orange 23 2.6 whorl of petals re flexed
- . - - , 1 inflorescence with 1
Dahlia Figaro MIX 26 3'2 whorl of petals reflexed
. ‘ _ . , 1 inflorescence with 1
Dranthus Super Parfait Raspberry 38 5.3 whorl of petals re flexe d
. , . , 1 inflorescence with
French marigold Janre Flame 19 or 23 6.4 250% of petals re flexed
Gazania ‘Daybreak Bronze’ 31 4.7 1 inflorescence wrth
petals reflexed
Moss rose ‘Margarita Apricot’ 44 18.8 1 flower open
Pentas ‘Graffiti Lavender’ 40 3.9 i3 flowers open on an
inflorescence
Petunia ‘Dreams Neon Rose’ 31 10.0 1 flower open
Petunia ‘Wave Purple’ 33 7.7 1 flower open
Snapdragon ‘Montego Orange Bicolor’ 41 3.1 .2 flowers open on an
inflorescence
Verbena ‘Quartz Waterfall Mix’ 32 4.2 8 flowers open on an
inflorescence
Viola ‘Sorbet Plum Velvet’ 38 6.5 1 flower open
Zinnia ‘Dreamland Coral’ 16 or 23 2.2 1 inflorescence With 1

whorl of petals reflexed

 

21

Table 1.2. Parameter estimates for nonlinear models (Eqs. [1] and [4]) relating flowering
rate to mean daily air temperature in 18 bedding plant species. Parameter estimates were
used to generate Fi gs. 1.1 and 1.2. Base (Tmin) and maximum temperatures (Tmax), are
the temperatures at which flowering rates are zero (low and high temperature,
respectively), and the optimum temperature (Tom) is the temperature where the maximum
development rate occurs (Rmax). C defines the skew of the function. Tmin, Tom, and
Tmax are in °C. CI = confidence interval.

 

 

Asymptotic
Eq. Parameter Estimate 95% CI (i) No.Z rZY
African marigold ‘Antigua Primrose’
[4] Tmin 4.4 0.7 86 0.90
Rmax 0.0396 0.0080
C 0.0560 0.0207
Angelonia ‘Serena Purple’X
[1] Tmin 9.9 0.4 80 0.94
TOpt 27.2 0.9
Rmax 0.0354 0.0013
Black-eyed Susan ‘Toto Rustic’
[4] Tmin 4.6 1.2 86 0.92
Rmax 0.0774 0.0564
C 0.0194 0.0182
Blue salvia ‘Victoria Blue’
[1] Tmin 9.4 1.0 70 0.89
Topt 28.0 0.9
Tmax 31.0 0.6
Rmax 0.0294 0.0019
Browallia ‘Bells Marine’
[1] Tmin 8.9 0.6 88 0.91
TOpt 26.7 0.5
Tmax 30.4 0
Rmax 0.0296 0.0014
Cosmos ‘Cosmic Orange’
[1] Tmin 7.2 0.7 83 0.88
Top! 23.7 0.8
Tmax 30.3 30.3
Rmax 0.0354 0.0015
Dahlia ‘Figaro Mix’
[1] Tmin 5.6 0.5 97 0.89
Topt 19.1 0.6
Tmax 30.4 0
Rmax 0.0204 0.0008

 

22

Table 1.2 (cont’d).

 

 

Asymptotic
Eq. Parameter Estimate 95% ngt) No.Z rzy
Dianthus ‘Super Parfait Raspberry’X
[1] Tmin 3.9 0.9 103 0.90
Topt 26.9 1.2
Rmax 0.0333 0.0014
French marigold ‘Janie Flame’
[4] Tmin 1.1 1.6 104 0.94
Rmax 0.104 0.0403
C 0.0282 0.0171
Gazania ‘Daybreak Bronze’
[1] Tmin 4.8 0.8 109 0.90
Topt 27.6 1.0
Tmx 35.0 -
Rmax 0.0303 0.0011
Moss rose ‘Margarita Apricot’
[1] Tmin 8.9 0.7 85 0.83
Topt 23.9 0.9
Tmax 31.7 1.2
Rmax 0.0316 0.0017
Pentas ‘Graffiti Lavender’x
[1] Tmin 9.3 0.5 83 0.94
Topt 27.8 0.9
Rmax 0.0274 0.001 1
Petunia ‘Dreams Neon Rose’
[4] Tmin 2.8 2.0 100 0.89
Rmax 0.3357 0.6607
C 0.00952 0.02162
Petunia ‘Wave Purple’
[4] Tmin 5.5 1.2 77 0.95
Rmax 0.0965 0.0437
C 0.0268 0.0174
Snapdragon ‘Montego Orange Bicolor’
[l] Tmin 2.0 1.6 101 0.74
Topt 25.7 0.9
Tmx 30.3 0
Rmax 0.0428 0.0023
Verbena ‘Quartz Waterfall Mix’
[1] Tmin 5.1 1.1 92 0.75
Topt 24.2 0.9
Tmax 31.0 0
ngx 0.0227 0.0013

 

23

Table 1.2 (cont’d).

 

 

Asymptotic
Eq. Parameter Estimate 95% CI (i) No.2 r2y
C 0.0576 0.0174
Viola ‘Sorbet Plum Velvet’x
[1] Tmin 4.1 1.2 109 0.79
TOpt 26.4 1.3
Rmax 0.0831 0.0034
Zinnia ‘Dreamland Coral’
[4] Tmin 7.8 0.5 96 0.94
Rmax 0.0541 0.0094

 

ZNumber of observations in data set.

YGenerated by performing linear regression analysis on the predicted versus observed
data.

xTmax could not be estimated from observed data and was fixed at 35.0 0C so the
nonlinear model could be solved.

24

Table 1.3. The effect of temperature on the number of nodes on the primary flowering
shoot at first open flower in 18 species of bedding plants. Plants were grown in
controlled environmental growth chambers under a 16-h photoperiod and a daily light
integral of 10.4 mol-m‘z-d‘l. Data were pooled between replications. L = linear; Q =

 

 

 

quadratic.
Temperature set point (°C)

Species 5 7.5 10 15 20 25 30 Trend
:‘fnian mai‘gdd , 10.8 e215.6 ab 15.2 b 17.0 ab 18.3 a 17.9 a 17.8 ab L...
Antrgua Prrmrose Q
Angel?” 8m“ _y — 9.5 b 11.3 b 11.2 b 12.8 b 16.9 a L"‘Q’
Purple

P'aCk'eye‘l Sf‘sa“ 12.5 a 14.6 a 15.5 a 15.4 a 14.3 a 12.8 b 13.5 a LNSQ‘S
Toto Rust1c

33:53]“ ““0“" - — 14.3 ab 15.6 a 10.4 c 10.6 c 11.9 be L‘"Q‘
8’08“!” Bells — — 13.1 c 14.3 bc 15.2 ab 16.1 a — L‘"Q"'S
Marme m
com"? com“ — - 7.7 bc 7.4 be 7.3 be 8.7 b 14.5 a L...
Orange Q
Dahlia .
.Figam Mix, — 9.0 b 8.8 b 8.9 b 12.4 a 13.2 a — L Q
Dianthus ‘Super m NS
Palfait Raspberry, 12.0 a 11.4 b 11.8 b 12.3 ab 12.2 ab 12.3 ab 13.9 a L Q
Pen?“ mang?“ 8.1 7.4 7.6 7.6 7.6 7.8 9.0 NS
Janre Flame

Gaza”? Daybreak 13.6 13.9 13.9 12.7 12.8 12.7 13.2 NS
Bronze

Moss rose NS **
Margarita Apricot. — — — 26.3 b 28.7 ab 30.4 a 27.2 ab L Q
Pen‘as Glam“ — — 5.6 b 7.0 a 6.9 a 6.5 a 6.4 ab LNSQ‘"
Lavender

Pawn” D’Fams 12.8 12.4 13.5 13.7 13.8 13.8 14.3 NS
Neon Rose m
Pam“? wave — 25.6 a 23.0 a 17.2 b 18.1 b 18.5 b 19.8 b L...
Purple Q
Snapdragon m NS
.Momego Orange, 8.9 b 9.1 b 8.9 b 8.9 b 8.4 b 8.3 b 11.1 a L Q
Verbena ‘Quartz NS u
Waterfall Mix’ — 12.4 a 10.3 bc 9.6 c 11.3 ab 11.7 ab 11.8 abcL Q
“01“ ,Sorbe‘ Plum 8.8 8.7 8.3 7.8 8.5 8.4 7.9 NS
Velvet
2mm? Dreamland — — 6.0 b 6.6 ab 6.4 b 6.5 b 7.3 a L""Q“S
Coral

 

ZMeans within rows followed by the same letter are not significantly different by Tukey’s
honestly significant difference test at P 50.05.
YTreatment not included in analysis because 275% of plants died.

.
NS. 1’ t, it.

25

Nonsignificant or significant at P $0.05, 0.01, or 0.001, respectively.

 

0.05 . . ~ .
African marigold
0.04
0.03 1
0.02 1

0.01 1

’Tmm=44°c ]

fi

Angelohia ' . (B)

Black-eyed 'Susan E (C)

 

0.00

0.04 1

.T opt = 230 °C ...........
Tmax = 31.0 °c

0.03 1

0.01 1--

   

002 . ....................................... ':

     

4 4L

  

 

 

L

0.00 4 ’1
0.06 1
005 ‘Tmin = 5.6 °C

°-°4 ‘Tmax = 30.4 °c ‘
003 . .......... ............
002 ....
0.01 1

Flowering rate (1/d to flower)

Dahlia

 

.Topt=19-1°C ................ .
Wm: 3.9°C

 

A
w

.Topt = 25-9 °C. ............... ..

 

 

   

Dianthus ' V (H) .

 

 

0.00 ~ 4 -

0510152025303551015202530355

Mean daily temperature (°C)

10 15 20 25 30 35

Fig. 1.]. Observed and predicted flowering rates in 9 species of bedding plants as a
function of mean daily temperature based on Eq. [1] (panel B, D-H) and Eq. [4] (panel A,
C, and I) and parameter estimates from Table 1.2. Circles represent the means of
replication 1 and 2. Dashed lines represent predictions outside of the observed data
range. Data points represent treatment means and error bars represent 95% confidence
intervals. Tmx in panels B and H could not be estimated from observed data and was
fixed at 35.0 °C so Eq. [1] could be solved.

26

 

0.05 a - . ~
Gazania

004 “Tmin =’ 4.8 OCH ’ ’

003 0Irop‘ = 27.6 °C . ‘
Tm,x = 35.0 °C

002 ll” .

0.01 +~

' (A)

  

I'... Tmax = 31.7 0C

Moss rose

T55: "8§9"C
Topt= 239 °C

1 n

  
 
 

..........

 

7 f

Pentas

 

0.00

0.08 1'

Petunia Wave Purple'

  

(E)

.........

......

 

 

Flowering rate (1/d to flower)

0.08 1

0.06 1

0.041~

 

 

 

’ (H)

...

 

Tmin = 78°C "

= 1

n
‘ v

Zinnia (I)

 

 

 

 

0.02 "‘ m I .
0.00 ‘ 1

0510152025303551015202530355

Mean daily temperature (°C)

10 15 20 25

Fig. 1.2. Observed and predicted flowering rates in 9 species of bedding plants as a
function of mean daily temperature based on Eq. [1] (panel A-C, F-H) and Eq. [4] gianel
D, E, and I) and parameter estimates from Table 1.2. Circles represent the means of
replication 1 and 2. Dashed lines represent predictions outside of the observed data
range. Data points represent treatment means and error bars represent 95% confidence
intervals. Tmax in panel A, C, and H could not be estimated from observed data and was
fixed at 35.0 °C so Eq. [1] could be solved.

27

Literature Cited

Adams, S.R., S. Pearson, and P. Hadley. 1996. Modeling growth and development of
pansy cv. Universal Violet in response to photo-thermal environment: Application
for decision support and scheduling. Acta Hort. 417:23-32.

Adams, S.R., S. Pearson, and P. Hadley. 1997. The effects of temperature, photoperiod
and light integral on the time to flowering of pansy cv. Universal Violet (Viola
Xwittrockiana Gams.). Ann. Bot. 80:107-112.

Adams, S.R., P. Hadley, and S. Pearson. 1998. The effects of temperature, photoperiod,
and photosynthetic photon flux on the time to flowering of Petunia ‘Express
Blush Pink’. J. Amer. Soc. Hort. Sci. 123:577-580.

Armitage, A.M., W.H. Carlson, and J .A. Flore. 1981. The effect of temperature and
quantum flux density on the morphology, physiology, and flowering of hybrid
geraniums. J. Amer. Soc. Hort. Sci. 106:643—647.

Arnold, C. 1959. The determination and significance of the base temperature in a linear
heat unit system. Proc. Amer. Soc. Hort. Sci. 74:430-445.

Berghage, RD. 1989. Modeling stem elongation in the poinsettia. Ph.D. Dissertation,
Dept. of Horticulture, Michigan State Univ., East Lansing.

Berghage, R.D., R.D. Heins, and J.E. Erwin. 1990. Quantifying leaf unfolding in the
poinsettia. Acta Hort. 272:243-247.

Brondum, JJ. and RD. Heins. 1993. Modeling temperature and photoperiod effects on
growth and development of dahlia. J. Amer. Soc. Hort. Sci. 118:36-42.

Erwin, J .E. and RD. Heins. 1990. Temperature effects on lily development rate and
morphology from the visible bud stage until anthesis. J. Amer. Soc. Hort. Sci.
1 15 2644-646.

Faust, J.E. and RD. Heins. 1993. Modeling leaf development of the African violet
(Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 118:747-751.

Faust, J .E. and RD. Heins. 1994. Modeling inflorescence development of the African
violet (Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 119:727-734.

Faust, J .E. and RD. Heins. 1997. Quantifying the influence of high-pressure sodium
lighting on shoot-tip temperature. Acta Hort. 418:85—91.

Heins, R.D., B. Liu, and ES. Runkle. 2000. Regulation of crop grth and development
based on environmental factors. Acta Hort. 514: 1 3—22.

28

Hidén, C. and R.U. Larsen. 1994. Predicting flower development in greenhouse grown
Chrysanthemum. Scientia Hort. 58:123-138.

Jones, H.G. 1992. Plants and microclimate. 2nd ed. Cambridge Univ. Press, New York,
NY.

Karlsson, M.G., R.D. Heins, and J .E. Erwin. 1988. Quantifying temperature controlled
leaf unfolding rates in Lilium longiflorum Thunb. ‘Nellie White’. J. Amer. Soc.
Hort. Sci. 113270—74.

Karlsson, MG. 1992. Leaf unfolding rate in Begonia Xhiemulis [sic]. HortScience
27: 109-1 10.

Landsberg, JJ. 1977. Some useful equations for biological studies. Expt. Agr.
13:273-286.

Larsen, R.U. 1988. A prediction model of the development rate of the pot plant Senecio X
hybridus Hyl. a case study. Acta Hort. 230:381-388.

Larsen, R.U. 1989. A prediction model for floral development of Senecio X hybridus Hyl.
‘Molls Stams’. Acta Hort. 248:247—254.

Larsen, R.U. 1990. Plant growth and modeling by light and temperature. Acta Hort.
272:235-242.

Larsen, R.U. and C. Hidén. 1995. Predicting leaf unfolding in flower induced shoots of
greenhouse grown Chrysanthemum. Scientia Hort. 63:225-239.

Larsen, R.U. and L. Persson. 1999. Modeling flower development in greenhouse
Chrysanthemum cultivars in relation to temperature and response group. Scientia
Hort. 80:73-89.

Liu, B. and RD. Heins. 2002. Photothermal ratio affects plant quality in ‘Freedom’
poinsettia. J. Amer. Soc. Hort. Sci. 127:20—26.

Maceina, M.J. and D.L. Pereira. 2007. Recruitment, p. 121—185. In: Guy, C. and ML.
Brown (eds.). Analysis and interpretation of freshwater fisheries data. Amer.
Fisheries Soc., Bethesda, MD.

Mattson, N. and J. Erwin. 2003. Temperature affects flower initiation and development
rate of Impatiens, Petunia, and Viola. Acta Hort. 624:191—197.

Miller, A.M., M.W. van Iersel, and AM. Armitage. 2001. Whole-plant carbon dioxide

exchange responses of Angelonia angustifolia to temperature and irradiance. J.
Amer. Soc. Hort. Sci. 126:606-610.

29

Moccaldi, LA. and ES. Runkle. 2007. Modeling the effects of temperature and
photosynthetic daily light integral on growth and flowering of Salvia splendens
and T agetes patula. J. Amer. Soc. Hort. Sci. 132:283-288.

Moseley, PL. 1997. Heat shock proteins and heat adaptation of the whole organism. J.
Appl. Plant Physiol. 83:1413—1417.

Neilson, R.E., M.M. Ludlow, and PG. Jarvis. 1972. Photosynthesis in Sitka spruce
(Picea sitchensis (Bong.) Carr.). 11. Response to temperature. J. Appl. Ecol.
9:721-745.

Niu, G., R.D. Heins, A.C. Cameron, and W.H. Carlson. 2000. Day and night
temperatures, daily light integral, and C02 enrichment affect growth and flower
development of pansy (Viola Xwittrockiana). J. Amer. Soc. Hort. Sci.
125:436—441.

Niu, G., R.D. Heins, A.C. Cameron, and W.H. Carlson. 2001. Day and night
temperatures, daily light integral, and C02 enrichment affect growth and flower
development of Campanula carpatica ‘Blue Clips’. Scientia Hort. 87:93-105.

Niu, G., D.S. Rodriguez, and Y.-T. Wang. 2006. Impact of drought and temperature on
growth and leaf gas exchange of six bedding plant species under greenhouse
conditions. HortScience 41 :1408—1 41 1 .

Pearson, S., P. Hadley, and A.E. Wheldon. 1993. A reanalysis of the effects of
temperature and irradiance on time to flowering in Chrysanthemum
(Dendranthema grandiflora). J. Hort. Sci. 68:89—97.

Pietsch, G.M., W.H. Carlson, R.D. Heins, and J.E. Faust. 1995. The effect of day and
night temperature and irradiance on development of Catharanthus roseus (L.)
‘Grape Cooler’. J. Amer. Soc. Hort. Sci. 1202877—881.

Pramuk, LA. and ES. Runkle. 2005. Modeling growth and development of celosia and
impatiens in response to temperature and photosynthetic daily light integral. J.
Amer. Soc. Hort. Sci. 130:813-818.

Reed, K.L., E.R. Hamerly, B.E. Dinger, and PG. Jarvis. 1976. An analytical model for
field measurement of photosynthesis. J. Appl. Ecol. 13:925-942.

Roberts, E.H. and RJ. Summerfield. 1987. Measurement and prediction of flowering in
annual crops, p. 17-50. In: Atherton, J .G. (ed.). Manipulation of flowering.
Butterworths, Kent, UK.

Semeniuk, P. 1975. Effect of temperature on growth, flowering, and fruiting in ‘Blue
Troll’ browallia. HortScience 10:480-481.

30

Steininger, J ., C.C. Pasian, and J .H. Lieth. 2002. Extension of a thermal unit model to
represent nonlinearities in temperature response of miniature rose development. J.
Amer. Soc. Hort. Sci. 127:349—354.

Summerfield, R.J., E.H. Roberts, R.H. Ellis, and RI. Lawn. 1991. Towards the reliable
prediction of time to flowering in six annual crops. 1. The development of simple
models for fluctuating field environments. Expt. Agr. 27:1 1—3 1.

van Iersel, M.W. 2003. Short-term temperature change affects the carbon exchange
characteristics and growth of four bedding plant species. J. Amer. Soc. Hort. Sci.
128:100—106.

Volk, T. and B. Bugbee. 1991. Modeling light and temperature effects on leaf emergence
in wheat and barley. Crop Sci. 31:1218-1224.

Wang, J .Y. 1960. A critique of the heat unit approach to plant response studies. Ecology
41 :785—790.

Wang, C.-H., D.-M. Yeh, and C.-S. Sheu. 2008. Heat tolerance and flowering-heat-delay
sensitivity in relation to cell membrane therrnostability in Chrysanthemum. J.
Amer. Soc. Hort. Sci. 133:754-759.

Wang, S.-Y., R.D. Heins, W.H. Carlson, and AC. Cameron. 1998. Modeling the effect of
temperature on flowering of Hibiscus moscheutos. Acta Hort. 456:161-169.

Warner, RM. and J .E. Erwin. 2005. Prolonged high temperature exposure and daily light
integral impact growth and flowering of five herbaceous ornamental species. J.
Amer. Soc. Hort. Sci. 130:319—325.

Warner, RM. and J .E. Erwin. 2006. Prolonged high-temperature exposure differentially
reduces growth and flowering of 12 Viola Xwittrockiana Gams. cvs. Scientia
Hort. 1082295—302.

Yeh, D.M., J .G. Atherton, and J. Craigon. 1999. A thermal time model of post-initiation
flower development in the shade plant, cineraria. Ann. Appl. Biol. 134:335-340.

Yuan, M., W.H. Carlson, R.D. Heins, and AC. Cameron. 1998. Effect of forcing

temperature on time to flower of Coreopsis grandiflora, Gaillardia X grandiflora,
Leucanthemum Xsuperbum, and Rudbeckiafulgida. HortScience 33:663-667.

31

SECTION II

MODELING PLANT GROWTH AND DEVELOPMENT OF PETUNIA IN

RESPONSE TO TEMPERATURE AND PHOTOSYNTHETIC DAILY LIGHT

INTEGRAL

32

Modeling Plant Growth and Development of Petunia in Response to Temperature

and Photosynthetic Daily Light Integral

Matthew G. Blanchard and Erik S. Runkle1

Department of Horticulture, Michigan State University, East Lansing, MI 48824

Paul R. Fisher

Environmental Horticulture Department, University of Florida, Gainesville, FL 32611

Received for publication . Accepted for publication . We gratefully
acknowledge funding by Michigan’s plant agriculture initiative at Michigan State
University (Project GREEEN), the Michigan Agricultural Experiment Station, the
American Floral Endowment, the Fred C. Gloeckner Foundation, the USDA-ARS
Floriculture and Nursery Research Initiative, and greenhouse growers providing support
for Michigan State University floriculture research. We also thank Raker’s Acres for

their contributions to this project and Mike Olrich for his greenhouse assistance.

1Associate Professor of Horticulture and Extension Specialist and to whom reprint

requests should be addressed (Email: runkleer@msu.edu).

33

Abstract

The effects of mean daily temperature (MDT) and photosynthetic daily light
integral (DLI) on flowering during the finish stage of two petunia (Petunia thbrida
Vilm.-Andr.) hybrids were quantified. Petunia ‘Easy Wave Coral Reef’ and ‘Wave
Purple’ were grown in glass-glazed greenhouses at 14 to 26 °C and under 4 to 19
mol-m“’2-d‘l with a 16-h photoperiod. The flower development rate was predicted using
a model that included a linear MDT function with a base temperature multiplied by an
exponential DLI saturation function. The flower development rate increased and time to
flower decreased as MDT increased within the temperature range studied. For example,
under a mean DLI of 12 mol~m’2-d", as MDT increased from 14 to 23 °C, time to
flower of ‘Easy Wave Coral Reef and ‘Wave Purple’ decreased from 51 to 22 d and 62
to 30 (1, respectively. Flower development rate increased as DLI increased until
saturation at 14.1 to 14.4 mol-m"2-d“'. Polynomial response surfaces were generated for
effects of MDT and DLI on flower bud number, leaf node ntunber, plant height, and
shoot length at flowering. The number of flower buds at flowering increased as MDT
decreased and DLI increased. For example, at an MDT of 14 °C with 18 mol-m’z-d“,
plants had 3.1 to 3.4 times more flower buds than those grown at 23 °C and 4
mol-m‘z-d“. Models were validated with an independent data set and the predicted time
to flower, flower bud number, and plant height were within 7 d, 20 flowers, and 5 cm,
respectively, for 87 to 100% of the observations. The models could be used to predict the

influence of MDT and DLI on crop scheduling and quality of these petunia hybrids.

34

Introduction

Efficient and predictable commercial production of greenhouse crops requires
information on how species respond to the environment so that they can be accurately
scheduled to finish for predetermined market dates. Scientific studies have been
performed on many economically important ornamental greenhouse crops to quantify
how light and temperature influence growth and development (Faust and Heins, 1994;
Karlsson and Heins, 1986; Pramuk and Runkle, 2005; Steininger and Pasian, 2002). Data
from these experiments have been used to generate crop models that predict how
changing environmental factors, such as mean daily temperature (MDT) or
photosynthetic daily light integral (DLI), affect the rate of plant development. Several
analytical approaches have been described to model plant development rate in response to
environmental conditions.

Temperature influences many biochemical, metabolic, and physiological
processes that occur during crop production, including photosynthesis, respiration,
transpiration, and plant development (Jones, 1992). Plant grth is defined as an
irreversible increase in weight, height, or volume of a plant cell, tissue, organ, or whole
plant, whereas development refers to a series of phenological stages that occur during the
life cycle of an organism (Steininger et al., 2002). Although environmental factors
interact and plant species exhibit differences in response, under typical horticultural
conditions, plant development is primarily controlled by temperature whereas growth is
largely influenced by DLI (Jones, 1992).

Plant developmental responses to temperature, such as flowering or leaf unfolding

rate, are primarily controlled by the integrated MDT. The time required for the

35

completion of a developmental event can be converted to a rate by calculating the
reciprocal of time (e. g., 1. /d). The relationship between MDT and development rate has
been described using linear, quadratic, cubic, and exponential models (Landsberg, 1977;

Larson, 1990). In a linear model, development rate is related to MDT as:

Rate=bo+b1XMDT [1]

where rate (e. g., l/d) is equal to the intercept (b0) plus the product of the slope (b1) and
MDT (°C). The parameters of the model are specific to a genotype or a development
stage (Summerfield et al., 1991). In this model, the relationship between MDT and
development rate is linear between the base and optimum temperature. The base
temperature (Tmin) is the temperature at or below which the rate of progress towards a

developmental stage is zero. Tmin can be estimated as:

Tmin = -bo/b1 [2]

and can vary considerably among plant species. For example, Tmin for the rate of flower
development in black-eyed Susan [Rudbeckiafulgida (Ait.) ‘Goldsturm’] was calculated
to be -1.3 °C, compared with 12.2 °C in rose mallow (Hibiscus moscheutos L. ‘Disco
Belle Mixed’) (Wang et al., 1998; Yuan et al., 1998).

An optimum temperature (Tom) is defined as the temperature at which the rate of
progress towards a developmental event is maximal. For example, Tom for flower

development rate in pansy (Viola Xwittrockiana Gams.) and vinca (Catharanthus roseus

36

L.) was calculated to be 21.7 °C and 235 °C, respectively (Adams et al., 1997b; Pietsch et
al., 1995). Linear models relating MDT and the rate of development are only valid when
Tmin S MDT S Topt.

Modeling development rate in different phases has been used to quantify
temperature responses where data indicate non—linearity across the experimental range of
temperatures. A linear equation for temperatures below Topt can be combined with a
negative linear function above Topt. The linear function to describe the response above
Topt may or may not be symmetrical with the slope at temperatures <Topt, or alternatively
a constant development rate can be assumed to occur above Topt (Pearson et al., 1993;
Roberts and Summerfield, 1987).

Polynomial equations have also been used to describe the influence of MDT on
development rate. Although these equations can provide a close empirical fit to data, the
parameters tend to have limited biological meaning (Brondum and Heins, 1993;
Landsberg, 1977). In Rieger begonia (Begonia Xhiemalis F otsch) grown under a 16-h
photoperiod, a quadratic model predicted a maximum rate of 0.12 leaves-d'l at 21 °C
(Karlsson, 1992). A cubic model was developed to describe the relationship between
MDT and the rate of leaf unfolding for cyclamen (Cyclamen persicum Mill.; Karlsson
and Werner, 2001 ), Chinese hibiscus (Hibiscus rosa-sinensis L.), (Karlsson et al., 1991),
and corn (Zea mays L.;Tollenaar et al., 1979).

Various exponential functions have been developed that incorporate parameters
such as Tmin, Topt, and an upper temperature threshold at which development rate is zero
(Tmax) (Reed et al., 1976; Landsberg, 1977; Hidén and Larsen, 1994; Larsen and Hidén,

1995). For example, an asymmetrical exponential function (Reed et al., 1976;

37

Landsberg, 1977) has been used to model the influence of temperature on flower
development rate in dahlia (Dahlia pinnata Cav.; Brondum and Heins, 1993) and leaf
unfolding and leaf expansion rate in African violet (Saintpaulia ionantha Wendl.; Faust
and Heins, 1993, 1994).

The rate of plant development can also be secondarily influenced by other
environmental factors and multiplicative models (Hidén and Larsen, 1994; Larsen and
Persson, 1999) have been developed that combine factors such as MDT, photoperiod, and
DLI (Larsen, 1990; Moccaldi and Runkle, 2007). In some species, factors such as MDT
and DLI can interact to influence development and therefore, models that include more
than one environmental parameter are adaptable to a range of conditions. Multiplicative
models have been published for several ornamental crops including Chrysanthemum
[Chrysanthemum X grandiflorum (Ramat.) Kitam.; Karlsson and Heins, 1986; Larsen and
Persson, 1999], impatiens (Impatiens walleriana Hook.f.; Pramuk and Runkle, 2005),
geranium (Pelargonium Xhortorum Bailey; White and Warrington, 1988), cineraria
(Pericallis thbrida R. Nordenstam; Larson, 1988, 1989), pansy (Adams et al., 1997b),
and red salvia (Salvia splendens F. Sello ex Roem & Schult.) (Moccaldi and Runkle,
2007)

The objectives of this study were to quantify and model the influence of MDT and
DLI on flowering and plant quality during the finish stage of two petunia hybrids
(Petunia X hybrida Vilm.-Andr. ‘Easy Wave Coral Reef and ‘Wave Purple’) under long-
day conditions. Finish stage describes the production period from the time plugs are
transplanted until plants are marketable. Petunia was selected because it is among the top

10 bedding plants produced in the United States, with a reported wholesale value of $120

38

million in 2008 (US. Department of Agriculture, 2009). Petunia flowering responses to
temperature, DLI, and photoperiod have been previously described using response
surface equations by Adams et al. (1997a, 1998, 1999); for temperature and DLI only
using quadratic equations (Kaczperski et al., 1991); and for temperature only using a
linear equation by Mattson and Erwin (2003). Calibration of models with data from
‘Easy Wave Coral Reef and ‘Wave Purple’ was necessary to develop decision-support
tools for crop scheduling of these cultivars, because model parameters are generally

cultivar-specific.

Materials and Methods

On 7 Dec. 2006 and 4 Apr. 2007, seeds of petunia ‘Easy Wave Coral Reef and
on 10 Sept. 2007 and 21 Mar. 2008, seeds of petunia ‘Wave Purple’ were sown in plug
trays [288-cell size (6-ml volume)] by a commercial greenhouse (C. Raker & Sons,
Litchfield, MI). After germination, plugs were received at Michigan State University
(MSU) and were grown in a controlled environmental growth chamber at a constant
temperature set point of 20 0C. A 16-h photoperiod was provided by 215-W cool-white
fluorescent (CWF; F96T12CWVHO; Philips, Somerset, NJ) and 60-W incandescent
lamps (INC; Philips), at a CWleNC (by W) of 3.6, and at an intensity of 180
umol‘m‘VZ-s‘1 at plant height. Plants were irrigated as necessary with well water
acidified with H2804 to a titratable alkalinity of 140 mg-L‘l CaCO3 and containing 95,
34, and 29 mg-L—l Ca, Mg, and S, respectively. The water was supplemented with a

water-soluble fertilizer providing (in mg~L‘1) 62 N, 6 P, 62 K, 7 Ca, 0.5 Fe, 0.3 Cu, Mn,

39

and Zn, 0.1 B and Mo (MSU Well Water Special; GreenCare Fertilizers, Inc., Kankakee,

IL).

Greenhouse Environment

After 27 d and 34 d from seed sow, 6- to 8-leaf ‘Easy Wave Coral Reef and
‘Wave Purple’ seedlings, respectively, were transplanted into 10-cm round plastic
containers (480-ml volume) filled with a commercial soilless peat-based medium
(Suremix; Michigan Grower Products, Inc., Galesburg, MI). At transplant, plugs were
thinned to one seedling per cell. Plants were randomly assigned to treatments and grown
in glass-glazed greenhouses at constant air temperature set points of 14, 17, 20, 23, or 26
°C and under a 16-h photoperiod that consisted of natural photoperiods (43 °N lat.) with
day—extension lighting from 0600 to 2200 HR provided by hi gh-pressure sodium (HPS)
lamps. ‘Easy Wave Coral Reef was not grown at 26 °C. At each temperature, plants
were grown under two DLI treatments provided by ambient light and a combination of
shade curtains (OLS 30, OLS 50; Ludvig Svensson Inc., Charlotte, NC) and different
intensities (25 to 150 umol-m’Z-s“) of supplemental lighting from HPS lamps that were
positioned above the shade curtains. Ten plants of each species were randomly assigned
to each temperature and DLI combination. The HPS lamps were operated by an
environmental computer (Priva Intégro 724; Priva, Vineland Station, Ontario) and were
turned on when the outside light intensity was <290 umol-m‘Z-s“ and turned off at >580
umol-m_2°s_’. Whitewash was applied to the greenhouse glazing during late Mar. each
year, and removed in mid Oct. The experiment was performed twice under mean DLIs

from transplant to flowering that ranged from 3.9 to 18.7 mol-m’Z-d‘l (Table 2.1).

40

Temperature in each greenhouse compartment was controlled by an
environmental computer with steam heating, passive and active ventilation, and fan-and-
pad evaporative cooling as needed. Air temperature was independently measured in each
greenhouse by an aspirated, shielded thermocouple (0.13-mm type E; Omega
Engineering, Stamford, CT) positioned 1.5 m above the floor (at plant level). At 30 cm
above the bench, the photosynthetic photon flux (PPF) was measured by a line quantum
sensor containing 10 photodiodes (Apogee Instruments, Inc., Logan, UT) under six DLI
and temperature combinations. Environmental measurements were collected every 10 s
and hourly means were recorded by a data logger (CR10; Campbell Scientific, Logan,
UT). A vapor-pressure deficit of 1.2 kPa was maintained during the night by the
injection of steam into the air. Horizontal airflow fans positioned 1.4 m above the
growing surface operated if the ridge vent was <90% of the maximum opening and
provided air movement at :0.1 m-s‘l at plant height [as measured with an air velocity
transducer (8475; T81, Inc., St. Paul, MN)]. Plants were irrigated with reverse osmosis
water supplemented with a water-soluble fertilizer providing (in mg-L“1) 125 N, 12 P,
100 K, 65 Ca, 12 Mg, 1.0 Fe and Cu, 0.5 Mn and Zn, 0.3 B, and 0.1 Mo (MSU RO Water

Special; GreenCare Fertilizers, Inc.).

Data Collection and Analysis

The date of first open flower per plant was recorded and time to first open flower
was calculated for each plant. Plants were considered flowering when one flower had a
fully open corolla. When each plant flowered, plant height and the total number of open

flowers and closed flower buds were recorded. In ‘Easy Wave Coral Reef, leaf number

41

on the primary shoot below the first open flower was also recorded. In ‘Wave Purple’,
the first open flower occurred on either a lateral stem or on the primary shoot and
therefore, leaf number below the flower could not be modeled. Plant height was
measured from the soil surface to the tip of the uppermost leaf on the primary stem. In
‘Wave Purple’, the length of the longest lateral stern was measured at flowering by
extending the stem and recording the distance from the axil to the shoot tip.

Flowering data were used to develop mathematical models to predict flower
development rate, flower bud number, height, lateral stem length, and leaf number under
different MDT and DLI conditions. Models for ‘Easy Wave Coral Reef and ‘Wave
Purple’ were generated using 159 and 200 observations (individual plants), respectively.
Data were analyzed using the calculated MDT and DLI for each plant from transplant to
the date of flowering. DLI values for treatments that did not have a line quantum sensor
were determined by calculating the mean irradiance among sensors that were positioned
in other temperature treatments and under similar light conditions. Flowering time data
were converted to developmental rates by calculating the reciprocal of days to flowering
(l/d to flower). A multiplicative model was developed to describe the relationship

between the rate of progress to flowering and MDT and DLI:

l/d to flower = fMDT X f DLI 13]

where f Mm and f Du are temperature and light functions, respectively. Models of this
type have been previously used to describe the rate of flower development in

Chrysanthemum (Hidén and Larsen, 1994; Larsen and Persson, 1999) and cineraria

42

(Larsen, 1989). The response of flower development rate to MDT is described with a
temperature function and can be quantified by algebraically rewriting equation [1] to

include the base temperature:

0 ...ifMDTsT - 4
l/d to flower = mm 1 I

-1. Tm x 7,, +1), x MDT ...imein<MDTSTopt

where Tmin and MDT are measured in °C and b1 is a species-specific temperature
constant. We used a linear function to quantify the MDT response because plots of the
actual data showed that Topt was not observed for both petunia hybrids. The influence of

DLI on flower development rate was described with a light function (Larsen, 1990):

DLI factor = 1 - EXP(-e X DLI) [5]

where the light factor ranges from 0 to 1. The e value is a species-specific light constant
and determines the skew of the curve and DLI is the mean (mol~m‘2°d“) from transplant
to flowering. This function indicates that the rate of progress towards flowering increases
as DLI increases, until some saturating value.

The final model to predict the rate of development towards flowering in petunia

consisted of equations [4] and [5] multiplied together:
0 ...if MDT _<_ Tmin [6]
I/d IO flOWCl’ = (—1 X Tmin X b] + b] X MDT) ”-imein < MDT S TOpt

x (1 — EXP(—e x DLI))

43

Estimates for model coefficients were determined with the nonlinear regression
procedure (NLIN) of SAS (version 9.1; SAS Institute, Cary, NC). Tmin for ‘Easy Wave
Coral Reef was estimated by using the NLIN procedure of SAS. A Tmin of 5.5 0C was
used for ‘Wave Purple’, which was obtained from unpublished experiments performed by
the authors in environmental growth chambers at constant air temperature set points of 5
to 30 °C and under a DLI of 10 mol-m’z-d"l and a constant 16-h photoperiod.

Data for flower bud and leaf number, plant height, and lateral stem length at first
flowering were analyzed using the regression procedure (REG) of SAS to determine the
influence of MDT and DLI. The flower bud and leaf number, plant height, and lateral

stem length response surfaces equations are in the form:

y = yo + aMDT + bMDT2 + cDLI + .11)er + gMDT x DLI [7]

where yo is the y-axis intercept and a, b, c, d, and g are species-specific constants.
Previous published studies that quantified the influence of MDT and DLI on flowering
used a similar polynomial equation (Moccaldi and Runkle, 2007; Pramuk and Runkle,

2005). The terms of the equation were only included if they were significant at P 50.05.

Model Validation
On 15 Jan. 2009, seeds of each petunia hybrid were sown in 288-cell plug trays
by a commercial greenhouse. After germination, trays were received at MSU and grown

in an environmental growth chamber. Environmental conditions inside the chamber,

44

plant culture, and transplant schedules were the same as described for the previous
experiments. Seedlings were transplanted into 10-cm round pots and 15 plants of each
species were grown in glass-glazed greenhouses at constant temperature set points of 17,
20, or 23 °C and under a 16-h photoperiod and a mean DLI of 14 to 19 mol-m‘Z-d‘l. In
each greenhouse, air temperature and PPF were measured on each bench and data was
recorded by a data logger as previously described. Photoperiod control, plant culture, and
data collection were the same a previously described. Data collected from the validation

study were used to test the accuracy and precision of model predictions.

Results

In both petunia hybrids, the rate of flower development increased and time to
flower decreased as MDT increased. For example, under a mean DLI of 12 mol'm"2-d" ,
as MDT increased from 14 to 23 °C, time to flower of ‘Easy Wave Coral Reef and
‘Wave Purple’ decreased from 51 to 22 d and 62 to 30 (1, respectively (Figs. 2.1A and
2.2A). Base temperature (Tmin) for the rate of flower development for ‘Easy Wave Coral
Reef and ‘Wave Purple’ were estimated to be 7.3 °C and 5.5 °C, respectively, although
the 95% asymptotic confidence intervals for Tmin overlapped between cultivars (Table
2.2). The upper temperature at which non-linearity occurred, Topt, was not determined
because the rate of flower development continued to increase within the experimental
range of temperature.

Time to flower decreased as DLI increased in both petunia hybrids. For example,
at an MDT of 20 °C, time to flower of ‘Easy Wave Coral Reef and ‘Wave Purple’

decreased by 10 and 12 (1, respectively, when DLI increased from 4 to 14 mol-m‘z'd‘l.

45

The influence of DLI, which was modeled as a multiplier of the temperature response,
approached saturation (within 99% of maximum development rate) in ‘Easy Wave Coral
Reef and ‘Wave Purple’ at 14.4 and 14.1 mol-m‘z'd‘l, respectively. Values for e (Table
2.2) were not statistically different between cultivars, and estimates of Eq. [5] did not
differ between cultivars by more than 0.01 between 3.9 and 18.7 mol-m‘z'd‘l. The
flowering rate models predicted time to flower within 7 d for 96% and 100% of the ‘Easy
Wave Coral Reef and ‘Wave Purple’ validation data sets, respectively (r2 = 0.73 or 0.93)
(Fig. 2.3). The slope and intercept for the relationship between predicted and observed
‘Easy Wave Coral Reef flower development rate had a 95% confidence interval of 1.2 :I:
0.3 and —0.007 i 0.012, respectively (data not presented). The slope and intercept for the
relationship between predicted and observed ‘Wave Purple’ flower development rate had
a 95% confidence interval of 1.4 i 0.2 and “0.009 :t 0.003, respectively (data not
presented).

Flower number (including open and closed flowers) increased as MDT decreased
and DLI increased (Table 2.3 and Figs. 2.1B and 2.2B). For example, at 14 °C and 18
mol-m“2-d“, plants had 3.1 to 3.4 times more flowers than those grown at 23 °C and 4
mol-m‘z-d". The response surfaces predicted flower number for both hybrids within 20
flowers for 87% of the observations in the validation data sets (data not presented).

Plant height at flowering of both petunia hybrids increased as DLI decreased. In
‘Easy Wave Coral Reef and ‘Wave Purple’, MDT had a quadratic and linear influence,
respectively, on height at flower (Figs. 2.1C and 2.2C). There was also an interaction
between temperature and DLI on plant height; DLI had a greater effect at a high MDT

(>20 °C) compared to a lower MDT (Table 2.3). Under the environmental conditions

46

tested in this study, plants of ‘Easy Wave Coral Reef were tallest (18.2 cm) when grown
at 21.5 °C and 3.9 mol-m‘Z-d‘l, whereas ‘Wave Purple’ were tallest (11.0 cm) when
grown at 26 °C and 4.3 mol-m‘Z-d‘l. Crop models for both hybrids predicted plant
height within 5 cm for all of the observations in the validation data sets (data not
presented).

The number of leaves that developed on the primary shoot before flowering in
‘Easy Wave Coral Reef decreased from a mean of 12 at 14 °C and 12.6 mol-m“2-d‘] to a
mean of 7 at 19.2 °C and 4 mol-m’2~d’I (Fig. 2.1D). However, leaf number was highly
variable and the model had a low coefficient of determination (r2 = 0.17) (Table 2.3).
The length of the longest lateral stern in ‘Wave Purple’ was primary influenced by DLI
and not by MDT. As the DLI increased from 4 to 18 mol-m‘z-d"' , lateral stern length
decreased by 16.7 cm (Fig. 2.2D). The model predicted lateral stem length within 5 and
10 cm for 67 and 93% of the observations in the validation data set, respectively (data not

presented).

Discussion

In both petunia hybrids, a simple linear relationship was adequate to describe the
effect of MDT on development rate, indicating the range of temperature conditions (14.2
to 26.0 °C was between Tmin and Tom. Estimates of base temperature (Tmin) for ‘Easy
Wave Coral Reef and ‘Wave Purple’ of 7.3 °C and 5.5 0C, respectively, did not
statistically differ in terms of 95% asymptotic confidence intervals (Table 2.2). These
estimates of Tmin were also similar to published Tmin values under 22.3 mol-m*2°d“l for

petunia ‘Avalanche Pink’, ‘Dreams Rose’, and ‘Wave Purple’ of 4.5, 4.2, and 5.9 °C,

47

respectively (Mattson and Erwin, 2003), and petunia ‘Sylvana Malve’ of 5.7 0C (Adams
et al., 1997a). Particularly considering that Tmin was well outside the experimental range
in this and other studies, results from different researchers are in close agreement and
average 5.5 i 1.1 °C (mean :1: SD).

Adams et al. (1997a, 1998) reported that under long days, petunia ‘Express Blush
Pink’, ‘Sylvana Malve’, and ‘Sylvana White’ had a predicted Topt of 25.4, 26.0, and 25.2
°C, respectively. Similarly, Topt in petunia ‘Snow Cloud’ grown under 13.5 mol-m‘z-d‘I
was predicted to be 25.0 °C (Kaczperski et al., 1991). Our model should not be
extrapolated beyond the experimental range in temperatures (14 to 26 °C), particularly
given that previous research suggests non—linearity in temperature response in petunia
above 25 °C.

Adams et al. (1998) developed a model for petunia in which DLI had a positive
linear effect on the rate of progress towards flowering (DLI range not reported).
However, we observed that at each experimental temperature, development rate increased
with increasing DLI in a diminishing returns relationship. Therefore, an exponential DLI
function was used in our model based on that reported by Larsen and Persson (1999) for
Chrysanthemum, but with genotype-specific constants that showed saturation above 14
mol-m'Z-d“. Faust et al. (2005) reported that petunia ‘Apple Blossom’ flowered a mean
of 6 d earlier under a DLI of 219 mol~m‘7--d"l versus 12 mol-m‘z-d‘l. In
Chrysanthemum, the estimated saturation DLI for the rate of flowering was 9.8
mol-m"2-d"l (Hidén and Larsen, 1994), whereas geranium had a higher estimated
saturation DLI of 17 mol-m‘Z-d‘l (White and Warrington, 1988). Adams et al. (1999)

hypothesized that petunia flower development rate decreased at a low DLI because of an

48

increase in the duration of the photoperiod-insensitive juvenile phase and photoperiod-
sensitive flower induction phase.

An advantage of using an exponential fimction to quantify the relationship
between DLI and flower development rate is that above the saturation DLI, the equation
predicts that rate does not increase with increasing DLI. Therefore, although the model
was developed with data for plants grown under 4 to 19 mol-m‘Z-d‘l, flowering
predications could theoretically be made for plants grown under 25 or 30 mol'm'Z-d—l of
light because petunia had an estimated saturation DLI of 214 mol-m’Z-d"1. Polynomial
models generated for other crops could inaccurately predict flowering if the crop is
grown outside of the DLI range under which the model was produced. For example, a
polynomial model for celosia (Celosia argentea L. var. plumosa Voss) grown under 8 to
' 26 mol-m‘2~d~l predicted that at 20 °C, as DLI increased from 10 to 25 mol-m‘zd“l
flowering time decreased by 3 d, and as DLI increased from 25 to 35 mol-m"2-d“‘
flowering time increased by 9 d (Pramuk and Runkle, 2005).

Albert et al. (2009) determined that the light compensation and saturation PPF for
photosynthesis in petunia ‘Mitchell’ was 28 and 590 u mol~m‘2~s", respectively, in
plants grown at 22 °C and under a constant irradiance of 600 pmol-m‘z-s"l for 14 h-d‘l.
These results reinforce that supplemental lighting during petunia crop production is most
beneficial when the ambient DLI is low. For example, in ‘Wave Purple’ grown at an
MDT of 20 °C, adding 4 mol-m‘z'd‘l from supplemental lighting when the DLI from
natural sunlight is 4 or 8 mol-m‘Z-d‘l is predicted to accelerate flowering by 10 and 2 d,
respectively. Our flowering time models also indicate that DLI promoted petunia flower

rate more at lower than higher MDT. For example, as DLI increased from 4 to 14

49

mol-m‘z-d‘l, predicted time to flower of ‘Wave Purple’ grown at 14 °C and 26 °C
decreased by 22 d and 9 (1, respectively. This response is in agreement with Kaczperski
et al. (1991), who reported that petunia ‘Snow Cloud’ grown at 14 °C or 26 °C flowered
14 or 4 d earlier, respectively, as DLI increased from 6.5 to 13 mol'm“2-d'1.

Our flowering models assumed that Tmin was not affected by DLI, but crop
models developed for other bedding plants suggest that in some species, Tmin could
decrease with increasing DLI. For example, a polynomial flower rate model generated
for impatiens calculated a Tmin of 7.5 and 4.3 °C under 5 and 15 mol-m’2°d‘l,
respectively (Pramuk and Runkle, 2005). However, in French marigold (T agetes patula
L.) and red salvia, DLI had little or no effect on Tmin (Moccaldi and Runkle, 2007). Tmin
for some species may be lower under a higher DLI because a higher DLI may increase
plant temperature.

The flower development model could be used to predict time‘to flowering using a
thermal time approach. Thermal time, calculated as 1/b. from Eq. [6], describes the
accumulated temperature that is required to reach a certain developmental event, with
units of thermal time of degree-hour (°C°h) or degree-day (°C-d) (Pasian and Lieth, 1994;
Roberts and Summerfield, 1987; Steininger et al., 2002; Wang et al., 1998). Commercial
crop growers can use thermal time to predict the occurrence of a developmental event
(e. g., first flowering) by subtracting Tmin from the MDT and accumulating the amount of
time, in units of °C~h or °C-d (Roberts and Summerfield, 1987; Steininger et al., 2002).
For example, a model developed for miniature rose (Rosa L. ‘Red Sunblaze’) predicted
the development phase from lateral bud break to open flower had a Tmin of 8.1 °C and

required 589 °C-d (Steininger et al., 2002). Using parameter estimates from Table 2.2,

50

thermal time to flowering would be 338 °C'd for ‘Easy Wave Coral Reef and 513 °C-d
for Wave Purple. In our case, the model could be used with a one-day time step where
Tmin is subtracted from the MDT, and the (°C-d) is multiplied by the DLI factor (Eq. [5],
between 0 and 1. For example, at 4, 11, and 18 mol-m‘Z-d", the DLI factor averaged
across both cultivars would be 0.73, 0.97, and 1.00, respectively.

Photoperiod was not included as an experimental factor in our crop models partly
because in many commercial greenhouses during the spring finishing phase, bedding
plants are grown under an inductive photoperiod to reduce production time. Petunia
hybrids have been classified as quantitative or qualitative long-day plants (Erwin, 2007).
For example, Adams et al. (1998) reported that the rate of flower development in petunia
‘Express Blush Pink’ increased linearly as photoperiod increased from 8 h to 14.5 h, and
further increases in photoperiod did not hasten flowering. Therefore, in this study, plants
were grown under a 16-h photoperiod during the plug and finish stages to accelerate
flowering and to make these models applicable for commercial production. Our models
would not be valid for petunia grown under a noninductive photoperiod, but the approach
of Adams et al. (1998) could be incorporated into an expanded version of the model.

In both hybrids studied, flower number at first flowering decreased as the MDT
increased. For example, in ‘Easy Wave Coral Reef grown under a DLI of 4 to 18
molm‘Z-d‘l, as MDT increased from 14 to 23 °C, the predicted flower bud number
decreased by 49 to 62%. Our results are in agreement with Mattson and Erwin (2003)
who reported that mean flower and lateral stem number in petunia ‘Dreams Rose’
decreased by 11 and 4, respectively, as MDT increased from 12 to 24 °C. This

information indicates that there is a trade-off between a short production duration and

51

high plant quality. At a low MDT, the rate of flower development is slow, but plants
have more time to harvest light and accumulate carbon before flowering. For example, in
impatiens grown under 15 mol-m‘z'd“l , as MDT decreased from 26 to 14 °C, time to
flower, flower bud number, and dry weight increased by 15 d, 141%, and 52%,
respectively (Pramuk and Runkle, 2005).

Our models illustrate that if a petunia crop is grown at a high MDT to accelerate
flowering, plant quality can be improved by increasing the DLI. For example, in ‘Easy
Wave Coral Reef grown at 23 °C, as DLI increased from 4 to 18 mol-m‘Z-d“, flower
bud number increased by 72% and plant height decreased by 68%. Lieth et al. (1991)
reported that dry weight accumulation in petunia ‘Snow Cloud’ grown at an MDT of 23
°C increased by a mean of 25 g-m‘z'd‘l as DLI increased from 5 to 25 mol-m”2-d“. The
models in the present study predicted that flower bud number in ‘Easy Wave Coral Reef
and ‘Wave Purple’ would be greatest under a DLI of >1 8.7 and 14.7 mol-m’Z-d‘l,
respectively. These results demonstrate that in some crops, the saturation DLI for the
greatest crop quality can be higher than the DLI that elicits the fastest flower
development rate.

Leaf number below the first flower can be a useful morphological indication of
timing of flower induction. Adams et al. (1999) reported that the duration of the juvenile
phase in petunia increased at both low and high MDTs and was shortest at an optimum
MDT of 21 .3 °C (leaf number not reported). In our study, the number of leaves that
developed on the primary shoot and below the first flower in ‘Easy Wave Coral Reef
decreased by a mean of 3 as MDT increased from 14 to 19.2 °C and then increased by a

mean of 2 as MDT increased to 23 °C. In comparison, Mattson and Erwin (2003) who

52

reported that under 22.3 mol-m‘Z-d“l, leaf number below the first open flower in petunia
‘Avalanche’, ‘Dreams Rose’, and ‘Wave Purple’ decreased by a mean of 6 to 8 as
temperature increased from 12 to 24 °C, suggesting a higher optimum MDT.

Although ‘Easy Wave Coral Reef and ‘Wave Purple’ had similar development
rates and growth responses to temperature and DLI, the models generated for one hybrid
did not accurately predict flower development rate, flower number, or plant height for the
other hybrid. The different models between hybrids could be caused by genetic
differences and results from breeding selection criteria such as growth form, production
time, and heat tolerance. For example, at an MDT of 14 to 23 °C and under a DLI of 4 to
18 mol-m’Z-d‘l, ‘Easy Wave Coral Reef flowered 7 to 12 d earlier, but had <28 fewer
flowers buds than ‘Wave Purple’. It is unknown whether the crop models developed for
these two petunia hybrids could be used with other hybrids within the same Easy Wave
or Wave petunia series. There is estimated to be over 360 petunia hybrids available
commercially (Kelly, et al., 2007) and future research could determine similarities in
flowering responses among other hybrids. Given the similarity in Tmin for petunia
between various studies and between estimates for the DLI factor (e) in the two cultivars
evaluated here, the main cultivar-specific parameter for flower development rate would
focus on estimating the slope (b1). The slope could be estimated by growing plants at a

single MDT, allowing for rapid model calibration to new cultivars.

53

Table 2.1. Mean daily air temperature (MDT) at the indicated set points and mean
photosynthetic daily light integral (DLI) above benches, from transplant to flowering of
each cultivar, in glass-glazed greenhouse compartments during experiments.

 

 

 

MDT (°C) DLI (mol-m"2-d‘l)Z
Replication 14 17 20 23 26 Low High
‘Easy Wave Coral Reef
1 14.2 17.0 19.9 22.7 —y 3.9 10.4
2 17.1 18.7 20.6 23.0 — 6.1 18.7
‘Wave Purple’
1 14.7 17.2 20.3 22.9 25.9 4.3 10.5
2 16.3 18.0 20.4 22.6 26.0 5.9 16.4

 

2Light from natural photoperiods and supplemental light from hi gh-pressure sodium
lamps that were turned on when the outside photosynthetic photon flux was <290
jrmol-m"2-s’l and turned off when >580 umol-m‘zs“.

YTemperature treatment not included.

Table 2.2. Parameters of nonlinear model relating rate of flowering of petunia ‘Easy
Wave Coral Reef and ‘Wave Purple’ to mean daily air temperature [MDT (°C)] and
daily light integral [DLI (mol-m“2-d")]. Models (Eq. [6]) are in the form of: l/d to
flower = (-1 X Tmin X b. + b1 X MDT) X (1 — exp(-—e X DLI)). Coefficients for each
petunia model were used to generate Figs. 2.1A and 2.2A. ACI = Asymptotic 95%
confidence interval.

 

 

Parameter Estimate Lower ACI Upper ACI NZ
‘Easy Wave Coral Reef
Tmin 7.3 °C 5.9 °C 8.7 °C 159
b; 2.96 E—03 2.62 E-03 3.31 E—03
e 3.20 E—01 2.74 E-01 3.67 E—Ol
‘Wave Purple’
Tmin 5.5 °CY 4.3 °C 6.6 °C 200
b; 1.95 E-03 1.91 E—03 1.98 E—03
e 3.28 E-Ol 3.04 E-01 3.53 E—Ol

 

2Number of observations in data set.
YEstimated from unpublished data that included 77 observations.

54

Table 2.3. Parameters of stepwise regression analysis relating flower number, plant
height, leaf number increase or lateral stern length for petunia ‘Easy Wave Coral Reef
and ‘Wave Purple’ to mean daily air temperature [MDT (°C)] and daily light integral
[DLI (mol°m‘2°d’l)]. All models are in the form of: y = yo + aMDT + bMDT2 + cDLI +
alDLI2 + gMDT X DLI. Models for ‘Easy Wave Coral Reef and ‘Wave Purple’ were
generated using 159 and 200 observations, respectively. Coefficients for each petunia
model were used to generate Figs. 2.1 B-D and 2.2B-D.

 

 

 

‘Easy Wave Coral Reef ‘Wave Purple’
Regression Flower Leaf Flower Lateral stern
parameter number Height (cm) number number Height (cm) length (cm)
yo 257.3 -79.5 43.2 21.1 2.98 56.2
(49,1)z (16.7) (8.88) (6.93) (2.12) (1.57)
a -20.6 9.32 —4.00 —X 3.76 E-01 —
(5.26) (1.79) (9.36 E-01) (1.12 E—01)
b 4.44 E—Ol —2.09 E-01 1.04 E—01 -5.38 E-02 — —
(1.39 E—01) (4.78 E—02) (2.46 E-02) (6.88 E-03)
c 1.18 — 6.55 E—01 9.02 — —1.19
(1.90 E-01) (2.03 E-01) (1.35) (1.5 E-Ol)
d — 4.13 E—02 -2.60 E-02 -3.08 E-Ol — —
(1.42 E-02) (8.69 E-03) (6.33 E-02)
g — -7.66 E-02 —- — -1.62 E-02 —
(1.68 E—02) (3.73 E-03)
r2 0.48*** 0.47*** 0.17*** 0.52*** 0.10*** 0.24***

 

2Standard errors are shown in parentheses.

xParameter not significant at P >0.05
***Parameter significant at P 50.001.

55

1ldays to flower

25

A 20

E A.
3 . e
E 15 -' ...
.9 g
#3 a
g 3
E

 

 

Fig. 2.1. The influence of mean daily temperature (MDT) and daily light integral (DLI)
on petunia ‘Easy Wave Coral Reef predicted flowering rate (l/d to flowering) and time
to flower (A), flower bud number (B), plant height (C), and leaf number increase at
flowering (D). The response surfaces were generated using Eqs. [6] and [7] and
coefficients in Tables 2.2 and 2.3. Each model was generated using 159 observations.

56

1/days to flower

 

_L _L
O N

on

Plant height (cm)
A 0)

N

 

 

 

 

    

    

1o 12 1"

 

8

1o 12 1 1
on (moi-m‘z-d' 1

8 -2. '1
DLI tm°“‘“ d i

Fig. 2.2. The influence of mean daily temperature (MDT) and daily light integral (DLI)
on petunia ‘Wave Purple’ predicted flowering rate (l/d to flowering) and time to flower
(A), flower bud number (B), plant height (C), and lateral stem length at flowering (D).
The response surfaces were generated using Eqs. [6] and [7] and coefficients in Tables
2.2 and 2.3. The models were generated using 200 observations.

57

 

50 I I I j I

   

 

 

 

 

/O 1
/
' Petunia 'Easy Wave Coral Reef’ ’2: 0.73 /// 8;

45 ‘ 0 Petunia Wave Purple' r2 = 0.93 // 0 2 -
A 3 /
B 1: / / 4 //
L. 2 / .
3 /
r2 //
.9 1 /, -
a) 4 /

o
g 0‘5’ / / j
/
'0 02 /
‘1’ /
Z /
3
_Q '1
O
35 40 45 50
Predicted time to flower (d)

Fig. 2.3. Validation of the flowering rate models, comparing the observed time to flower
of petunia ‘Easy Wave Coral Reef and ‘Wave Purple’ grown at a mean daily
temperature of 17.3 to 22.1 °C and under a mean daily light integral of 14 to 19
mol-m‘2°d‘1 with those predicted by Eq. [6]. The dashed lines represent the 7-d lower
and upper boundary. Forty-five observations for each hybrid were used for the validation
study, where each symbol represents an individual plant. Numbers represent the quantity
of observations at each symbol. Coefficients for the flowering rate models are presented
in Table 2.2. r2 values were generated by performing linear regression analysis on the
predicted versus observed data.

58

Literature Cited

Adams, S.R., S. Pearson, and P. Hadley. 1997a. The effects of temperature and
photoperiod on the flowering and morphology of trailing petunias. Acta Hort.
435265-75.

Adams, S.R., S. Pearson, and P. Hadley. 1997b. The effects of temperature, photoperiod
and light integral on the time to flowering of pansy cv. Universal Violet (Viola X
wittrockiana Gams.). Ann. Bot. 80:107-112.

Adams, S.R., P. Hadley, and S. Pearson. 1998. The effects of temperature, photoperiod,
and photosynthetic photon flux on the time to flowering of Petunia ‘Express
Blush Pink’. J. Amer. Soc. Hort. Sci. 123:577-580.

Adams, S.R., S. Pearson, P. Hadley, and W.M. Patefield. 1999. The effects of
temperature and light integral on the phases of photoperiod sensitivity in Petunia
X hybrida. Ann. Bot. 83:263—269.

Albert, N.W., D.H. Lewis, H. Zhang, L.J. Irving, P.E. Jameson, and KM. Davies. 2009.
Light-induced vegetative anthocyanin pigmentation in Petunia. 60:2191-2202.

Brondum, J .J . and RD. Heins. 1993. Modeling temperature and photoperiod effects on
growth and development of dahlia. J. Amer. Soc. Hort. Sci. 118:36-42.

Erwin, J. 2007. Factors affecting flowering in ornamental plants, p. 7-48. In: N.O.
Anderson (ed.). Flower breeding and genetics. Springer, The Netherlands.

Faust, J .E. and RD. Heins. 1993. Modeling leaf development of the African violet
(Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 118:747-751.

Faust, J .E. and RD. Heins. 1994. Modeling inflorescence development of the African
violet (Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 119:727-734.

Faust, J .E., V. Holcombe, N.C. Rajapakse, and DR. Layne. 2005. The effect of daily
light integral on bedding plant growth and flowering. HortScience 40:645—649.

Hidén, C. and R.U. Larsen. 1994. Predicting flower development in greenhouse grown
Chrysanthemum. Scientia Hort. 58:123—138.

Jones, H.G. 1992. Plants and microclimate. 2nd ed. Cambridge Univ. Press, New York,
NY.

Kaczperski, M.P., W.H. Carlson, and MG. Karlsson. 1991. Growth and development of

Petunia X hybrids as a function of temperature and irradiance. J. Amer. Soc. Hort.
Sci. 116:232—237.

59

Karlsson, MG. and RD. Heins. 1986. Response surface analysis of flowering in
Chrysanthemum ‘Bright Golden Anne’. J. Amer. Soc. Hort. Sci. 111:253-259.

Karlsson, M.G., R.D. Heins, J .O. Gerberick, and ME. Hackmann. 1991. Temperature
driven leaf unfolding rate in Hibiscus rosa-sinensis. Scientia Hort. 45:323-331.

Karlsson, MG. 1992. Leaf unfolding rate in Begonia X hiemulis [sic]. HortScience
27:109-110.

Karlsson, M. and J. Werner. 2001. Temperature affects leaf unfolding rate and flowering
of cyclamen. HortScience 36:292-294.

Kelly, R.O., Z. Deng, and B. Harbaugh. 2007. Evaluation of 125 petunia cultivars as
bedding plants and establishment of class standards. HortTechnology
1 7:3 86-3 96.

Landsberg, J .J . 1977. Some useful equations for biological studies. Expt. Agr.
13:273—286.

Larsen, R.U. 1988. A prediction model of the development rate of the pot plant Senecio X
hybridus Hyl. a case study. Acta Hort. 230:381-388.

Larsen, R.U. 1989. A prediction model for floral development of Senecio X hybridus Hyl.
‘Molls Stams’. Acta Hort. 248:247-254.

Larsen, R.U. 1990. Plant growth and modeling by light and temperature. Acta Hort.
272:235-242.

Larsen, R.U. and C. Hidén. 1995. Predicting leaf unfolding in flower induced shoots of
greenhouse grown Chrysanthemum. Scientia Hort. 63:225-239.

Larsen, R.U. and L. Persson. 1999. Modeling flower development in greenhouse
Chrysanthemum cultivars in relation to temperature and response group. Scientia ’
Hort. 80:73-89.

Lieth, J .H., RH. Merritt, and HG Kohl Jr. 1991. Crop productivity of petunia in relation
to photosynthetically active radiation and air temperature. J. Amer. Soc. Hort. Sci.
116:623-626.

Mattson, N. and J. Erwin. 2003. Temperature affects flower initiation and development
rate of Impatiens, Petunia, and Viola. Acta Hort. 624: 191-197.

Moccaldi, LA. and ES. Runkle. 2007. Modeling the effects of temperature and

photosynthetic daily light integral on growth and flowering of Salvia splendens
and T agetes patula. J. Amer. Soc. Hort. Sci. 132:283—288.

60

Pasian, CC. and J.H. Lieth. 1994. Prediction of flowering rose shoot development based
on air temperature and thermal units. Scientia Hort. 59:131-145.

Pearson, S., P. Hadley, and A.E. Wheldon. 1993. A reanalysis of the effects of
temperature and irradiance on time to flowering in Chrysanthemum
(Dendranthema grandiflora). J. Hort. Sci. 68:89—97.

Pietsch, G.M., W.H. Carlson, R.D. Heins, and J .E. Faust. 1995. The effect of day and
night temperature and irradiance on development of Catharanthus roseus (L.)
‘Grape Cooler’. J. Amer. Soc. Hort. Sci. 120:877—881.

Pramuk, LA. and ES. Runkle. 2005. Modeling growth and development of celosia and

impatiens in response to temperature and photosynthetic daily light integral. J.
Amer. Soc. Hort. Sci. 130:813-818.

Reed, K.L., E.R. Hamerly, B.E. Dinger, and P.G. Jarvis. 1976. An analytical model for
field measurement of photosynthesis. J. Appl. Ecol. 13:925-942.

Roberts, E.H. and R.J. Summerfield. 1987. Measurement and prediction of flowering in
annual crops, p. 17-50. In: Atherton, J .G. (ed.). Manipulation of flowering.
Butterworths, Kent, UK.

Steininger, J ., C.C. Pasian, and J .H. Lieth. 2002. Extension of a thermal unit model to
represent nonlinearities in temperature response of miniature rose development. J.
Amer. Soc. Hort. Sci. 127:349—354.

Summerfield, R.J., E.H. Roberts, R.H. Ellis, and R.J. Lawn. 1991. Towards the reliable
prediction of time to flowering in six annual crops. 1. The development of simple
models for fluctuating field environments. Expt. Agr. 27:11-31.

Tollenaar, M., T.B. Daynard, and RB. Hunter. 1979. Effect of temperature on rate of leaf
appearance and flowering date in maize. 19:363-366.

US. Department of Agriculture. 2009. Floriculture crops 2008 summary. Agr. Stat.
Board, Washington DC.

Wang, S.-Y., R.D. Heins, W.H. Carlson, and AC. Cameron. 1998. Modeling the effect of
temperature on flowering of Hibiscus moscheutos. Acta Hort. 456:161-169.

White, J .W. and I.J. Warrington. 1988. Temperature and light integral effects on grth
and flowering of hybrid geraniums. J. Amer. Soc. Hort. Sci. 113:354-359.

Yuan, M., W.H. Carlson, R.D. Heins, and AC. Cameron. 1998. Effect of forcing

temperature on time to flower of C oreopsis grandiflora, Gaillardia Xgrandiflora,
Leucanthemum Xsuperbum, and Rudbeckiafulgida. HortScience 33:663—667.

61

SECTION III

THE INFLUENCE OF DAY AND NIGHT TEMPERATURE FLUCTUATIONS ON

GROWTH AND F LOWERING OF ANNUAL BEDDING PLANTS AND

GREENHOUSE HEATING COST PREDICTIONS

62

The Influence of Day and Night Temperature Fluctuations on Growth and

Flowering of Annual Bedding Plants and Greenhouse Heating Cost Predictions

Matthew G. Blanchard1 and Erik S. Runkle2

Department of Horticulture, Michigan State University, East Lansing, MI 48824

Received for publication_ . Accepted for publication . We gratefully
acknowledge funding by Michigan’s plant agriculture initiative at Michigan State
University (Project GREEEN), the Michigan Agricultural Experiment Station, the
American Floral Endowment, the Fred C. Gloeckner Foundation, the USDA-ARS

F loriculture and Nursery Research Initiative, and greenhouse growers providing support
for Michigan State University floriculture research. We also thank C. Raker & Sons for

donation of plant material and Mike Olrich for his greenhouse assistance.

1Ph.D. candidate.

2Associate Professor of Horticulture and Extension Specialist and to whom reprint

requests should be addressed (Email: runkleer@msu.edu).

63

Abstract

Volatile energy costs and lower profit margins have motivated many flower
growers in temperate climates to improve the energy efficiency of crop production. We
performed experiments with dahlia (Dahlia thbrida Cav. ‘Figaro Mix), French
marigold (Tagetes patula L. ‘Janie Flame’), and zinnia (Zinnia elegans Jacq. ‘Magellan
Pink’) to quantify the effects of constant and fluctuating temperatures on growth and
flowering during the finish stage. Plants were grown in glass-glazed greenhouses with a
day/night (16 h/8 h) temperature of 20/14, 18/18, 16/22 (means of 18 °C), 24/18, 22/22,
or 20/26 °C (means of 22 °C) with a 16-h photoperiod and under a photosynthetic daily
light integral of 11 to 19 mol-m‘z'd"1. Flowering times of dahlia, French marigold, and
zinnia (Year 2 only) were similar among treatments with the same mean daily air
temperature (MDT). All species grown at 20/ 14 °C were 10 to 41% taller than those
grown at 16/22 °C. Crop timing data and computer software that estimates energy
consumption for heating (Virtual Grower) were then used to estimate energy
consumption for greenhouse heating on a per-crop basis. Energy costs to produce these
crops in Charlotte, NC, Grand Rapids, MI, and Minneapolis, MN for a finish date of 15
Apr. or 15 May and grown at the same MDT were estimated to be 3 to 42% lower at a +6
°C day/night temperature difference (DIF) compared with a 0 °C DIF and 2 to 90%
higher at a -6 °C DIF versus a 0 °C DIF. This information could be used by greenhouse

growers to reduce energy inputs for heating on a per-crop basis.

64

Introduction

Plant developmental responses to temperature, such as flowering or leaf unfolding
rate, are controlled by the integrated mean daily temperature (MDT) (Roberts and
Summerfield, 1987). The rate of plant development is zero at or below a species-specific
base temperature (Tmin) and increases as MDT increases until a maximum rate at an
optimum temperature (Topt). Development rates between Tmin and Topt for different
phenological stages have been estimated for several floriculture crops such as African
violet (Saintpaulia ionantha Wendl.; Faust and Heins, 1994), Chrysanthemum
[Chrysanthemum X grandiflorum (Ramat.) Kitam.; Hidén and Larsen, 1994; Karlsson et
al., 1989a], Easter lily (Lilium longiflorum Thunb.; Erwin and Heins, 1990; Karlsson et
al., 1988), poinsettia (Euphorbia pulcherrima Willd. ex Klotz; Berghage, 1990), and
potted rose (Rosa L.; Steininger et al., 2002). Models developed for these species can be
used to predict crop time in response to temperature and other environmental factors.

Temperature can also affect plant morphology; in many species, stem elongation
is influenced by the difference between the day and night temperature (DIF) (Myster and
Moe, 1995). Stem elongation is promoted when the day temperature is higher than the
night temperature (+DIF) and suppressed when the day temperature is lower than the
night temperature (-DIF). The effect of DIF on plant height has been studied in many
common greenhouse crops, such as campanula (Campanula isophylla Moretti) (Moe et
al., 1991), Chrysanthemum (Karlsson et al., 1989b), fuchsia (Fuchsia thbrida hort. ex
Sieb. and Voss) (Erwin et al., 1991), Easter lily (Erwin et al., 1989), and poinsettia
(Berghage, 1989; Berghage and Heins, 1991). For example, Erwin et al. (1989) reported

that plant height of Easter lily increased by 129% as DIF increased from ~16 to +16 °C.

65

Similarly, intemode length of Chrysanthemum increased by 133% as DIF increased from
—12 to +12 °C (Karlsson et al., 1989). Stem elongation responses to DIF occur primarily
from changes in cellular elongation rather than cellular division, and gibberellins are
likely involved (Grindal et al., 1998; Jensen et al., 1996; Myster and Moe, 1995). A
-DIF is sometimes used by growers as a height control strategy in controlled
environment production of floriculture crops (Myster and Moe, 1995).

In temperate climates, high energy inputs can be required to maintain a desirable
greenhouse temperature, making fuel for heating one of the largest floriculture production
expenses (Bartok, 2001). In the Netherlands, greenhouses account for 79% of the energy
. used by the agricultural sector and 7% of the country’s total energy consumption
(Lansink and Ondersteijn, 2006). In the US, the mean commercial price of natural gas
increased by 119% from 1998 to 2008 (US. Department of Energy, 2009). Rising and
volatile fuel costs for heating greenhouses have motivated many growers to improve
energy conservation and to minimize energy inputs for crop production in controlled
environments.

Greenhouse growers can reduce energy consumption by managing the greenhouse
environment with dynamic temperature control (DTC) strategies (K'orner et al., 2007;
Lund et al., 2006). In DTC, instead of static temperature set points that are the same each
day, heating set points are lowered during periods when the greenhouse energy-loss
factor is high (e.g., outside temperature and incoming solar radiation are low) and
increased when the energy-loss factor is low (Kbrner et al., 2004). This environmental
control strategy integrates temperature and maintains a target MDT over a l- to 7-d

interval (Ktirner and Challa, 2003; Kbmer et al., 2004). Lund et a1. (2006) reported that a

66

greenhouse in Denmark with DTC had 32 to 79% and 75 to 89% lower energy
consumption for heating during winter and spring months, respectively, compared to a
greenhouse with static temperature set points.

To achieve the greatest potential energy savings with temperature integration, a
greenhouse environmental control computer with sophisticated software (e. g., DTC) is
required (Aaslyng et al., 2003, 2005; Komer and Van Straten, 2008). However, not all
greenhouses utilize environmental control computers, and of those that do, relatively few
utilize DTC strategies. An alternative and simple energy-saving approach is to use a
+DIF with static day and night heating and ventilation set points. With a +DIF, the
heating set point is lowered during the night when energy consumption for heating is
highest (Bartok, 2001). A low night temperature is compensated by increasing the day
temperature so that the target MDT is achieved.

A DTC or DIF strategy to reduce energy consumption assumes that plant
development is influenced by MDT, and crop time is similar at different day and night
temperatures (>Tmin and STopt) that deliver the same MDT. However, studies with
bedding plants that compared flowering times at DIF and constant temperatures regimens
with the same MDT have reported different responses among species. For example,
geranium (Pelargonium Xhortorum Bailey) grown at an MDT of 18 °C flowered
similarly at day/night (12 h/12 h) set points of 18/18 or 27/9 °C (White and Warrington,
1988). In contrast, Mortensen and Moe (1992) reported that petunia ‘Ultra Red’ (Petunia
thbrida Vilm.-Andr.) flowered 5 d earlier at a day/night (16 h/8 h) of 21/1 5 °C
compared with 19/19 °C and 17/23 0C, while red salvia (Salvia splendens F. Sello ex

Roem & Schult.) flowered 5 (1 later at 17/23 °C compared with 19/19 °C and 21/15 0C.

67

Therefore, the benefits of using DIF to reduce energy inputs or to suppress stem
elongation may not be practical for all bedding plant species if crop time is delayed. The
objectives of this research were to (1) quantify the effects of constant and fluctuating
temperatures on growth and flowering during the finish stage of three bedding plant
species and (2) predict greenhouse heating costs for different crop finish dates, at

different locations in the United States, with different DIF regimens.

Materials and Methods

During Sept. 2008 (Year 1) and Mar. 2009 (Year 2), seeds of dahlia (Dahlia
thbrida Cav. ‘Figaro Mix’), French marigold (Tagetes patula L. ‘Janie Flame’), and
zinnia (Zinnia elegans Jacq. ‘Magellan Pink’) were sown in plug trays [288-cell size (6-
ml volrune)] by a commercial greenhouse (C. Raker & Sons, Litchfield, MI). In Year 1,
zinnia received a foliar spray of paclobutrazol (Bonzi; Syngenta Crop Protection, Inc.,
Greensboro, NC) at an unreported rate and volume to suppress hypocotyl elongation.
Ten to 17 d after seed sow, plugs were received at Michigan State University (MSU) and
were grown in a controlled environmental growth chamber at a constant temperature set
point of 20 °C. A 16-h photoperiod was provided by 215-W cool-white fluorescent
(CWF; F96T12CWVHO; Philips, Somerset, NJ) and 60-W incandescent lamps (INC;
Philips), at a CWF2INC (by W) of 3.6, and at an intensity of 180 jrmol-m”2-s‘I at plant
height. Plants were irrigated as necessary with well water acidified with H2804 to a
titratable alkalinity of 140 mg-L‘l CaCO3 and containing 95, 34, and 29 mgL‘1 Ca, Mg,

and S, respectively. The water was supplemented with a water-soluble fertilizer

68

providing (mg'L‘l) 62 N, 6 P, 62 K, 7 Ca, 0.5 Fe, 0.3 Cu, Mn, and Zn, 0.1 B and Mo

(MSU Well Water Special; GreenCare Fertilizers, Inc., Kankakee, IL).

Greenhouse Environment

After 26 d (dahlia), 19 or 23 (1 (French marigold), and 23 or 16 d (zinnia) from
seed sow, seedlings were thinned to one seedling per cell and transplanted into 10-cm
round plastic containers (480-ml volume) filled with a commercial soilless peat-based
medimn (Suremix; Michigan Grower Products, Inc., Galesburg, MI). At transplant,
dahlia, French marigold, and zinnia had a mean of 3, 6, or 6 leaves, respectively. Fifteen
plants of each species were randomly assigned to each of 6 glass-glazed greenhouse
sections with constant temperature set points of 18 or 22 °C and fluctuating day/night (16
h/8 h) temperature set points of 20/14, 16/22, 24/ 18, or 20/26 °C. The temperature set
points were chosen so that 3 treatments each had an MDT of 18 or 22 °C. In each
greenhouse section, temperature set points were maintained by an environmental
computer (Priva Intégro 724; Priva, Vineland Station, Ontario) that controlled steam
heating, passive and active ventilation, and evaporative cooling pads when needed. The
transition period between the day and night temperature set points was 3 min-°C‘1. The
experiment was performed twice with transplant dates beginning on 18 Oct. 2008 (Year
1) and 20 Mar. 2009 (Year 2).

The photoperiod was maintained at 16 h and consisted of natural photoperiods (43
°N lat.) with day-extension lighting from 0600 to 2200 HR provided by high-pressure
sodium (HPS) lamps. The HPS lamps were operated by an environmental computer and

were turned on when the outside light intensity was <290 umol-m‘Z-s‘l and turned off at

69

>5 80 umol-m’z-s‘ 1. The photoperiod and skotoperiod paralleled the day and night
temperature set points, respectively. In Year 2, whitewash was applied to the greenhouse
glazing so that the maximum photosynthetic photon flux (PPF) was 1200 umol-m‘Z-s‘l
at plant height. A maximum vapor-pressure deficit of 1.2 kPa was maintained during the
night by the injection of steam into the air. Horizontal airflow fans positioned 1.4 m
above the growing surface operated if the ridge vent was <90% of the maximum opening
and provided air movement at z0.1 m-s‘l at plant height [as measured with an air

velocity transducer (8475; T81, Inc., St. Paul, MN)].

Environmental Monitoring and Plant Culture

Air temperature was independently measured in each greenhouse by an aspirated,
shielded thermocouple (0.13-mm type E; Omega Engineering, Stamford, CT) positioned
at plant level. In each temperature treatment, the PPF was measured by a line quantum
sensor containing 10 photodiodes (Apogee Instruments, Inc., Logan, UT) positioned at 30
cm above the bench. Environmental measurements were collected every 10 s and hourly
means were recorded by a data logger (CR10; Campbell Scientific, Logan, UT).
Temperature control during the experiment was within 1.3 °C of the greenhouse
temperature set points for all treatments in both years and the actual MDT was 18.0 +0.4
°C or 22.0 +0.2 °C (Table 3.1). The mean photosynthetic daily light integral (DLI) from
transplant to flowering ranged from 10.6 to 12.3 mol-m‘z-d‘I in Year 1 and 15.7 to 19.1
mol'm‘z-d‘l in Year 2.

Plants were irrigated as necessary with reverse osmosis water supplemented with

a water-soluble fertilizer providing (mg-L“) 125 N, 12 P, 100 K, 65 Ca, 12 Mg, 1.0 Fe

70

and Cu, 0.5 Mn and Zn, 0.3 B, and 0.1 Mo (MSU RO Water Special; GreenCare

Fertilizers, Inc.).

Data Collection and Analysis

The date of first open inflorescence (flowering) was recorded and time to flower
was calculated for each plant. Plants were considered flowering when each species had
an inflorescence with at least 50% of the ray petals fully reflexed. When each plant
flowered, plant height and the number of inflorescences were recorded. Plant height was
measured from the soil surface to the base of the first whorl of flowers on an
inflorescence.

A completely randomized block design was used during each year. Data were
analyzed using SAS (SAS Institute, Inc., Cary, NC.) mixed model procedure (PROC
MD(ED), and pairwise comparisons between treatments were performed using Tukey’s
honestly significant difference test at P S 0.01. Data were pooled between replications if

the treatment Xyear interaction was not significant at P S 0.01.

Heating Cost Estimation

The cost to heat a 1,991 m2 greenhouse to produce a flowering crop grown at
day/night (16 h/8 h) temperature set points of 18/18, 20/14, 16/22, 22/22, 24/18, or 20/26
°C for finish dates of 15 Mar., 15 Apr., or 15 May was estimated for Charlotte, NC,
Grand Rapids, MI, and Minneapolis, MN using the Virtual Grower 2.5 software (F rantz
et al., 2007; US. Department of Agriculture, 2009a). Production time for each species

was calculated from the greenhouse experiments using the mean time to flower at an

71

MDT of 18 or 22 °C. Flowering time for zinnia was calculated based on data from Year
2 only. The greenhouse characteristics used to estimate heating costs included: 8 spans
each 34.1 X 7.3 m, arched 3.7-m roof, 2.7-m gutter, polyethylene double layer roof,
polycarbonate bi-wall ends and sides, forced air unit heaters burning natural gas, 50%
heater efficiency, no energy curtain, an air infiltration rate of 1.0-h", and day temperature
set points from 0600 to 2200 HR. These values and characteristics are typical of
commercial greenhouses used to produce floriculture crops in the northern half of the
United States. The heater efficiency was chosen based on the mean recorded values from
several commercial greenhouses (Frantz et al., 2009). Cities were subjectively chosen
from a list of the largest garden plant-producing states in the United States (US.
Department of Agriculture, 2009b) and were selected if they had an outside MDT <10 °C

during Jan., Feb., Mar., and Apr. (US. Department of Energy. 1995).

Results

In dahlia and French marigold, there were no differences in time to flower among
plants grown at temperature treatments with a similar MDT (Fig. 3.1A and 3.2A). For
example, French marigold flowered a mean of 26 d after transplant when grown at 18/18,
20/14, or 16/22 °C and a mean of 21 d when grown at 22/22, 24/ 18, or 20/26 °C. In
zinnia, plants grown at temperature treatments that delivered a similar MDT flowered at
the same time in Year 2, but not Year 1 (Fig. 3.3A). In Year 1, zinnia flowered 3 to 7 d
earlier in plants grown at 20/14 °C compared with plants grown at 18/ 18 or 16/22 °C.

In dahlia, plants grown at 18/18 or 20/14 °C had a mean of 8 more inflorescences

than plants grown at 22/22, 24/18, 20/26, or 16/22 °C (Fig. 3.1B). French marigold had a

72

similar inflorescence number among treatments, although those grown at 16/22 °C had a
mean of 3 more than those at 24/18 °C (Fig. 3.2B). In Year 1, when the mean DLI was
=11 molim'z-d”, zinnia grown at an MDT of 22 °C developed 2 or 3 more
inflorescences than plants grown at an MDT of 18 °C (Fig. 3.3B). In contrast,
inflorescence number was similar among temperature treatments in Year 2, when the
mean DLI was =17 mol'm‘Z-d".

Dahlia grown at a +6 °C DIF (20/ 14 or 24/18 °C) was 4.6 to 5.3 cm taller at
' flowering than plants grown at a -6 OC DIF (16/22 or 20/26 °C; Fig. 3.1C). In French
marigold, plants were 11% taller when grown at 20/14 °C or 24/18 °C versus .16/22 °C.
In Year 1, zinnia grown at 24/18 °C were 17 to 58 % taller than all other treatments,
while in Year 2, plants grown at 20/14 were 13 to 32% taller than all other treatments
(Fig. 3.3C). For all species, there were no differences in height between plants grown at
a 0 oC DIF.

In all species and locations, energy for heating predictions to produce a flowering
crop for 15 Apr. or 15 May were up to 41% lower when grown at a +6 °C DIF compared
with a constant temperature (Table 3.2). As finish date progressed from 15 Mar. to 15
May, the relative difference in heating costs between a +6 °C DIF and 0 °C DIF
increased. Heating costs per crop for all locations and finish dates were estimated to be
greatest when grown at 16/22, 20/26, or 22/22 °C. For example, dahlia grown in
Minneapolis, MN would consume 2 to 29% more energy if grown at 16/22 °C versus
18/18 or 20/14 0C. In nearly all instances, the least amount of energy consumed per crop
of dahlia or French marigold occurred at 20/14 °C. In contrast, the lowest energy input

for a crop of zinnia was 24/18 °C for 15 Mar. finish dates and a +6 °C DIF for the later

73

two finish dates. The estimated energy consumption for heating was greatest for dahlia
grown at 20/26 °C or constant 22 °C, regardless of location or finish date. For French
marigold and zinnia, greenhouse heating was greatest for a crop grown at a +6 °C DIF.
As finish date increased from 15 Mar. to 15 May, heating costs at each temperature
regimen decreased by 52 to 84%. For example, zinnia grown for 15 May at 20/14 °C

would require 77% less energy inputs for heating than the same crop grown for 15 Mar.

Discussion

Flowering time of dahlia, French marigold, and zinnia (Year 2 only) was similar
among temperature treatments with the same MDT. This response reinforces the
paradigm that flowering rate is a function of the MDT and, within limits, the effects of
day and night temperature on progress towards flowering are equal (Karlsson et al., 1988;
Roberts and Summerfield, 1987; Steininger et al., 2002). Our results are in agreement
with Mortensen and Moe (1992) who reported no difference in flowering time of fuchsia,
geranium, impatiens (Impatiens walleriana Hook.f.), pocketbook plant (Calceolaria
Xherbeohybrida Voss), potted rose, and tuberous begonia (Begonia Xtuberhybrida
pendula) grown at day/night (16 h/8 h) temperature set points of 19/19, 21/15, or 17/23
°C. Similarly, flowering time was controlled by MDT and not DIF in pinnate dahlia
(Dahlia pinnata Cav.; Brondum and Heins, 1993), pansy (Viola Xwittrockiana Gams.;
Niu et al., 2000), and vinca (Catharanthus roseus L.; Pietsch et al., 1995).

Crop models that predict flowering time under different environmental conditions
have been developed for several bedding plants including celosia, French marigold,

impatiens, pansy, petunia, pinnate dahlia, and red salvia (Adams et al., 1997, 1998;

74

Brandum and Heins, 1993; Moccaldi and Runkle, 2007; Pramuk and Runkle, 2005). In
many of these experiments, models were generated with data from plants that were grown
at constant temperature set points. For example, flowering models predicted that as
constant temperature set points increased from 14 to 26 °C, time to flower in celosia
(C elosia argentea L. var. plumosa Voss) and impatiens grown under a DLI of 15
mol-m’z-d‘l decreased by 38 and 15 d, respectively (Pramuk and Runkle, 2005). Data
from our study presenting similar flowering times at different day/night treatments with
the same MDT indicates that these crop models could also be used to predict flowering
time at fluctuating temperature set points. Caveats of many crop models that predict
flowering time are that they are only valid if the day and night temperatures are ZTmin
and _<_Topt (Summerfield et al., 1991).

In zinnia during Year 1, flowering time was different among treatments with an
MDT of 18 °C. The actual MDT among these treatments varied by only 0.1 to 0.4 °C
and plants received a DLI within 0.6 mol-m’z-d‘l. Therefore, it is not clear why plants
grown at a 20/14 °C flowered later than those grown at a constant 18 °C or 16/22 °C. At
an MDT of 18 °C, zinnia flowered a mean of 2 to 9 d earlier and had a mean of 3 more
inflorescences in Year 2 compared with Year 1. The differences between years could be
because after transplant, plants in Year 2 received a DLI that was 6 mol-m“2-d‘l higher
than in Year 1. In many species, flowering time decreases and flower number increases
as DLI increases. For example, as DLI increased from 10 to 15 mol-m‘z-d", time to
flower in celosia grown at 20 °C decreased by 3 d and inflorescence number increased by
2 (Pramuk and Runkle, 2005). The flowering delay and reduced inflorescence number in

zinnia during Year 1 could also be at least partially attributed to the paclobutrazol

75

application that was applied during the plug, which has been shown to delay flowering in
some crops (Blanchard and Runkle, 2007).

Plant height at flower in all species generally increased as DIF increased from -6
°C to +6 °C. These results are in agreement with Myster and Moe (1995) that the relative
promotion or suppression of stem elongation is influenced by the magnitude of DIF.
Similar effects of DIF on plant height have been reported in other bedding plants
including geranium (Mortensen and Moe, 1992), pinnate dahlia (Brondum and Heins,
1993), impatiens (Mortensen and Moe, 1992), pansy (Niu et al., 2000), petunia
(Kaczperski et al., 1991), red salvia (Mortensen and Moe, 1992), snapdragon
(Antirrhinum majus L.; Neily et al., 1997), tall verbena (Verbena bonariensis L.; Shimizu
and Heins, 2000), and zinnia (Neily et al., 1997). For example, stem elongation during
the vegetative stage in snapdragon and zinnia increased by 38 and 13%, respectively, as
DIF (l3-h day/l l-h night) increased from -5 to +5 °C (Neily et al., 1997).

Although a +DIF temperature regimen promoted stem elongation, greenhouse
energy inputs to heat these bedding plants at these locations and finish dates were
estimated to be lowest with a +DIF. For example, the estimated energy inputs to produce
these three crops in Charlotte, NC, Grand Rapids, MI, and Minneapolis, MN for a finish
date of 15 Mar. were similar or up to 11% lower if grown at +6 °C DIF instead of a
constant temperature. In contrast, for a finish date of 15 May, energy inputs at the same
MDT were estimated to be 9 to 42% lower at a +6 °C DIF compared with a 0 °C DIF.
Similar results were reported in a simulated greenhouse study in the Netherlands: total
annual energy consumption was 9%, 13%, and 23% lower during Feb., Mar., and Apr.,

respectively, with a +6 °C DIF compared with a —2 °C DIF (Komer et al., 2004).

76

These results collectively indicate that for many locations, a +DIF temperature
regimen is an energy efficient-production strategy and the energy—savings with +DIF
increases with later production dates. Because plants grown at some +DIF treatments
were taller than those grown at a constant or -DIF temperature regimen, the advantages
and disadvantages of DIF should be considered. If a +DIF temperature regimen is used
to save energy, growers may need to utilize an alternative height control strategy to
suppress stem elongation. An example of a height control strategy could be the
application of a chemical plant grth retardant (Blanchard and Runkle, 2007). Plants
grown under a -DIF had suppressed stem elongation, but were estimated to require the
highest energy inputs to produce. An economic analysis could determine if it is more
cost-effective to deliver a +DIF to save energy and use different height control strategies
or to deliver a —DIF with more energy inputs, but less height control requirements.

Dahlia had a mean of 42% more inflorescences when grown at a day/night of
18/18 and 20/14 °C compared with the other treatments. Plants grown at a lower MDT
could have more inflorescences at flowering than those grown at a higher MDT because
they had more time to harvest photosynthetic light and accumulate dry matter before
flowering. For example, Moccaldi and Runkle (2007) reported that as MDT decreased
from 26 to 14 °C, time to flower, inflorescence number, inflorescence diameter, and dry
weight at flowering increased. Dahlia grown at 16/22 °C had a mean of 7 fewer
inflorescences than plants grown at the same MDT but at 18/18 and 20/14 °C. The lower
flower number at a night temperature of 22 °C could be attributed to a higher respiration
rate during the night that could decrease available carbon. van Iersel (2003) reported that

as temperature increased from 6 to 36 °C, whole-plant dark respiration in French

77

marigold, geranium, pansy, and petunia increased exponentially and daily carbon gain
decreased quadratically.

The production of these species at different MDTs would require that crops are
transplanted on different dates so that they are finished on the same market date. For
example, dahlia grown for a market date of 15 May would need to be transplanted on 25
Mar. if grown at a constant 18 °C and on 30 Mar. if grown at a constant 22 °C. In this
example, the crop transplanted earlier and grown at 18 °C would develop 8 more
inflorescences before flowering than the crop grown at 22 "C.

Some greenhouse growers have lowered the MDT in an attempt to conserve
energy for heating. Although this strategy can decrease the heating requirement on a
daily basis, time to flower increases. For some locations and crops, the longer production
time at a low MDT could require more total energy inputs per crop than a shorter
production time at a higher MDT (Blanchard et al., in press; Shimizu et al., 2003). We
estimated that to produce flowering zinnia for 15 Mar. or 15 Apr. in Grand Rapids, M1 or
Minneapolis, MN, energy consumption for heating per crop would be similar or 4 to 14%
lower at 22/22 or 24/18 °C instead of 18/18 or 20/14 °C. However, to produce French
marigold for the same finish date at these locations, energy consumption is estimated to
be 2 to 13% higher at 22/22 or 24/18 °C compared with 18/18 or 20/14 °C. Therefore,
energy-efficient production temperatures depend on the crop grown, time of year, and
location. When selecting a growing temperature, other production factors such as the
number of crop turns, irrigation frequency, and disease and insect control could also be

considered.

78

 

Table 3.1. Mean daily air temperature and daily light integral during experiments in Year
1 and 2. The day and night were 16 and 8 h, respectively.

 

 

 

Day/night Actual day/night Actual mean daily Daily light integral
temperature temperature (°C) temperature (°C) (mol-m—Z-d-l)
set point (°C) Year 1 Year 2 Year 1 Year 2 Year 1 Year 2
18/18 18.4/16.7 18.1/17.8 17.9 18.0 11.6 15.9
20/14 20.4/14.1 19.7/15.0 18.3 18.2 11.0 16.3
16/22 16.5/21.4 16.8/21.8 18.1 18.4 11.0 16.9
22/22 22.3/21.1 22.3/21.5 21.9 22.1 10.7 19.1
24/18 240/186 240/187 22.2 22.2 12.3 15.7
20/26 20.7/25.1 20.3/25.7 22.2 22.1 10.6 18.3

 

Table 3.2. Predicted relative amount of energy used for greenhouse heating to produce
three annual species grown at different day and night temperature set points in different
locations and finish dates. Heating inputs were estimated using Virtual Grower software
(US. Department of Agriculture, 2009a) and include time from transplant to first
flowering on 15 Mar., 15 Apr., or 15 May. Production time for each species was
calculated from greenhouse experiments using the mean time to flower at 18/18, 20/14,
and 16/22 0C or 22/22, 24/18, and 20/26 °C. Percentages were calculated by dividing
heating input by the highest input for each location and species. See materials and
methods for greenhouse and heating parameter inputs.

 

 

 

 

Day/n1 ght Greenhouse location and finish date
temperature Charlotte, NC Grand Rapids, MI Minneapolis, MN
set point (°C) 15 Mar. 15 Apr. 15 May 15 Mar. 15 Apr. 15 May 15 Mar. 15 Apr. 15 May
Dahlia ‘Figaro Mix’
18/18 0.83 0.45 0.22 0.91 0.62 0.35 0.96 0.55 0.26
20/14 0.79 0.39 0.17 0.91 0.60 0.32 0.96 0.53 0.24
16/22 0.90 0.55 0.34 0.93 0.65 0.41 0.98 0.59 0.31
22/22 0.96 0.62 0.38 1.00 0.70 0.44 1.00 0.61 0.35
24/ 18 0.91 0.55 0.29 0.98 0.67 0.40 0.98 0.58 0.31
20/26 1.00 0.67 0.45 1.00 0.71 0.48 1.00 0.63 0.37
French marigold ‘Janie Flame’
18/18 0.76 0.43 0.18 0.97 0.60 0.32 0.96 0.44 0.26
20/14 0.68 0.35 0.11 0.95 0.57 0.27 0.95 0.42 0.21
16/22 0.88 0.58 0.35 1.00 0.65 0.39 1.00 0.50 0.32
22/22 0.94 0.60 0.30 0.95 0.63 0.36 0.99 0.48 0.27
24/18 0.86 0.49 0.18 0.92 0.59 0.30 0.97 0.44 0.23
20/26 1.00 0.67 0.40 0.96 0.65 0.40 1.00 0.50 0.31
Zinnia ‘Magellan Pink’
18/18 0.91 0.51 0.27 0.97 0.64 0.38 0.97 0.55 0.27
20/14 0.86 0.44 0.20 0.97 0.62 0.33 0.97 0.53 0.24
16/22 1.00 0.63 0.40 1.00 0.68 0.43 1.00 0.59 0.33
22/22 0.89 0.61 0.35 0.93 0.64 0.39 0.87 0.49 0.30
24/18 0.83 0.53 0.26 0.91 0.62 0.35 0.85 0.46 0.26
20/26 0.93 0.66 0.43 0.94 0.66 0.42 0.87 0.51 0.33

 

79

60 1 T 1 r . a

 

a a A
. ..... . ... ab. ..... . . .. .. ..
50 . 3. . ., , b b b
g 5 "! >‘ ‘27-; P: [hi i "1
h 40 _ , ‘3. ..... .1 , .. ‘l... "w ‘f . ..... -
0 1 . ‘- .1
C 30 . :2" ..... ’ » ‘7 ........ . 5 ..... E ..... j
.9 1 'i V ‘ _
a) ’ :i ‘ . §
E 20 . » .I .. .7 . ‘ ..... . . .' ‘ ..... . .
i: : 1. 1'“ , - ‘ »- ’ a.
10 _ .......... .

Inflorescence (no.)

Height (cm)

         

 

 

Li

1 . g

. '1 5?

1:» Li »: > .' “’12—. ... u ’r
18/18 20/14 15/22 22/22 24/18 20/26

(17.9) (18.2) (18.3) (22.0) (22.1) (22.2)
Temperature set point (°C. day/night)
(Actual mean daily temperature, °C)

 

Fig. 3.1. The influence of temperature on time to flower (A), and inflorescence number
(B) and height (C) at flowering, in dahlia ‘Figaro Mix’ at constant and fluctuating
day/night (16 h/8 h) temperature set points. Vertical bars indicate standard errors of
treatment means. Mean separation by Tukey’s honestly significant difference test at P S
0.01.

80

 

Time to flower (d)

12 -»

’7
o
c

v
a)
o
E
m
o
‘—
o

c

5

Height (cm)

     

 

 

18/18 20/14 16I22 22/22 24/18 20/26
(18.2) (18.2) (18.6) (22.1) (22.1) (22.3)
Temperature set point (°C, day/niglm
(Actual mean daily temperature, °C)

 

Fig. 3.2. The influence of temperature on time to flower (A), and inflorescence number
(B) and height (C) at flowering, in French marigold ‘Janic Flame’ at constant and
fluctuating day/night (16 h/8 h) temperature set points. Vertical bars indicate standard
errors of treatment means. Mean separation by Tukey’s honestly significant difference
test at P S 0.01.

81

O)
O

 

-Year 1 A
El Year 2

    
   
 
  
   

CI
0

A
0

Time to flower (d)
8 8

_\
O

 

 

 

# O) on O

Inflorescence (no.)

N

 

Height (cm)

 

 

 

 

 

18/18 20/14 16/22 22/22 24/18 20/26

%9\To\(&b2& 90% 53:15 (9,31% (133990 5);?"

Temperature set point (°C, day/nth)

(Actual mean daily temperature, °C)
Fig. 3.3. The influence of temperature on time to flower (A), and inflorescence number
(B) and height (C) at flowering, in zinnia ‘Magellan Pink’ at constant and fluctuating
day/night (l6 h/8 h) temperature set points. Vertical bars indicate standard errors of

treatment means. Mean separation by Tukey’s honestly significant difference test at P S
0.01.

0

82

Literature Cited

Aaslyng, J .M., J .B. Lund, N. Ehler, and E. Rosenqvist. 2003. IntelliGrow: A greenhouse
component-based climate control system. Environ. Model. Software 18:657-666.

Aaslyng, J .M., N. Ehler, and L. J akobsen. 2005. Climate control software integration
with a greenhouse environmental control computer. Environ. Model. Software
20:521-527.

Adams, S.R., S. Pearson, and P. Hadley. 1997. The effects of temperature, photoperiod
and light integral on the time to flowering of pansy cv. Universal Violet (Viola
xwittrockiana Gams.). Ann. Bot. 80:107—112.

Adams, S.R., P. Hadley, and S. Pearson. 1998. The effects of temperature, photoperiod,
and photosynthetic photon flux on the time to flowering of Petunia ‘Express
Blush Pink’. J. Amer. Soc. Hort. Sci. 123:577-580.

Bartok, J .W. Jr. 2001. Energy conservation for commercial greenhouses. Natural
Resource, Agr., and Eng. Serv. Coop. Ext., Ithaca, NY.

Berghage, R.D. Jr. 1989. Modeling stem elongation in the poinsettia. Ph.D. Dissertation,
Dept. of Horticulture, Michigan State Univ., East Lansing, MI.

Berghage, R.D., R.D. Heins, and J .E. Erwin. 1990. Quantifying leaf unfolding in the
poinsettia. Acta Hort. 272:243-247.

Berghage, RD. and RD. Heins. 1991. Quantification of temperature effects on stem
elongation in poinsettia. J. Amer. Soc. Hort. Sci. 116:14-18.

Blanchard, M.G. and ES. Runkle. 2007. Dipping bedding plant liners in paclobutrazol or
uniconazole inhibits subsequent stem extension. HortTechnology 17:178-182.

Blanchard, M.G., E.S. Runkle, and J. Frantz. Energy-efficient greenhouse production of
Petunia and Tagetes by manipulation of temperature and photosynthetic daily
light integral. Acta Hort. (in press).

Brondum, J .J . and RD. Heins. 1993. Modeling temperature and photoperiod effects on
growth and development of dahlia. J. Amer. Soc. Hort. Sci. 118:36-42.

Erwin, J .E., R.D. Heins, and M.G. Karlsson. 1989. Thermomorphogenesis in Lilium
longiflorum. Amer. J. Bot. 76:47-52.

Erwin, J .E. and RD. Heins. 1990. Temperature effects on lily development rate and
morphology from the visible bud stage until anthesis. J. Amer. Soc. Hort. Sci.
115:644-646.

83

Erwin, J .E., R.D. Heins, and R. Moe. 1991. Temperature and photoperiod effects on
Fuchsia >< hybrida morphology. J. Amer. Soc. Hort. Sci. 116:955-960.

Frantz, J ., E. Runkle, M. Blanchard, L. Locke, and C. Krause. 2007. Virtual Grower:
Estimating greenhouse energy costs and plant growth using new computer
software. HortScience 42: 1018 (abstr.).

F rantz, J ., B. Hand, and L. Buckingham. 2009. Virtual Grower software user manual
(Version 2.51). US. Dept. Agr., Agr. Res. Serv., Washington DC.

Faust, J .E. and RD. Heins. 1994. Modeling inflorescence development of the Afi‘ican
violet (Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 119:727—734.

Grindal, G., A. Emstsen, J .B. Reid, 0. Junttila, B. Lindgard, and R. Moe. 1998.
Endogenous gibberellin A] levels control therrnoperiodic stem elongation in
Pisum sativum. Physiol. Plant. 102:523-531.

Hidén, C. and R.U. Larsen. 1994. Predicting flower development in greenhouse grown
Chrysanthemum. Scientia Hort. 58: 123-1 38.

Jensen, E., S. Eilertsen, A. Emsten, O. Juntilla, and R. Moe. 1996. Therm0periodic
control of stem elongation and endogenous gibberellins in Campanula isophylla.
J. Plant Growth Regul. 15:167-171.

Jones, H.G. 1992. Plants and microclimate. 2nd ed. Cambridge Univ. Press, New York,
NY.

Kaczperski, M.P., W.H. Carlson, and M.G. Karlsson. 1991. Growth and development of
Petunia X hybrids as a function of temperature and irradiance. J. Amer. Soc. Hort.
Sci. 116:232-237.

Karlsson, M.G., R.D. Heins, and J .E. Erwin. 1988. Quantifying temperature controlled
leaf unfolding rates in Lilium Iongiflorum Thunb. ‘Nellie White’. J. Amer. Soc.
Hort. Sci. 113:70—74.

Karlsson, M.G., R.D. Heins, J .E. Erwin, and RD. Berghage. 1989a. Development rate
during four phases of Chrysanthemum growth as determined by preceding and
prevailing temperatures. J. Amer. Soc. Hort. Sci. 114:234-240.

Karlsson, M.G., R.D. Heins, J .E. Erwin, R.D. Berghage, W.H. Carlson, and J .A.
Biembaum. 1989b. Temperature and photosynthetic photon flux influence
Chrysanthemum shoot development and flower initiation under short-day
conditions. J. Amer. Soc. Hort. Sci. 114: 158-163.

Kemer, O. and H. Challa. 2003. Temperature integration and DIF in cut Chrysanthemum.
J. Hort. Sci. Biotechnol. 78:335—342.

84

 

Komer, 0., M.J. Bakker, and E. Heuvelink. 2004. Daily temperature integration: A
simulation study to quantify energy consumption. Biosystems Eng. 87:333-343.

Komer, 0., J .M. Aaslyng, A.U. Andreassen, and N. Holst. 2007. Microclimate prediction
for dynamic greenhouse climate control. HortScience 42:272-279.

Komer, O. and G. Van Straten. 2008. Decision support for dynamic greenhouse climate
control strategies. Comput. Electron. Agr. 60:18—30.

Lansink, AD. and C. Ondersteijn. 2006. Energy productivity growth in the Dutch
greenhouse industry. Amer. J. Agr. Econ. 88:124-132.

Lund, J .B., A. Andreassen, C.—O. Ottosen, and J .M. Aaslyng. 2006. Effect of a dynamic
climate on energy consumption and production of Hibiscus rosa-sinensis L. in
greenhouses. HortScience 41:384-388.

Moccaldi, LA. and ES. Runkle. 2007. Modeling the effects of temperature and
photosynthetic daily light integral on growth and flowering of Salvia splendens
and Tagetes patula. J. Amer. Soc. Hort. Sci. 132:283—288.

Moe, R., RD. Heins, and J. Erwin. 1991. Stem elongation and flowering of the long-day
plant Campanula isophylla Moretti in response to day and night temperature
alternations and light quality. Scientia Hort. 48:141—151.

Mortensen, L.M. and R. Moe. 1992. Effects of various day and night temperature
treatments on the morphogenesis and growth of some greenhouse and bedding
plant species. Acta Hort. 327:77-86.

Myster, J. and R. Moe. 1995. Effect of diurnal temperature alternations on plant
morphology in some greenhouse crops—a mini review. Scientia Hort.
62:205-215.

Neily, W.G., P.R. Hicklenton, and D.N. Kristie. 1997. Temperature and developmental
stage influence diurnal rhythms of stem elongation in snapdragon and zinnia. J.
Amer. Soc. Hort. Sci. 122:778-783.

Niu, G., RD. Heins, A.C. Cameron, and W.H. Carlson. 2000. Day and night
temperatures, daily light integral, and C02 enrichment affect growth and flower
development of pansy (Viola xwittrockiana). J. Amer. Soc. Hort. Sci.
125:436-441.

Pietsch, G.M., W.H. Carlson, RD. Heins, and J .E. Faust. 1995. The effect of day and

night temperature and irradiance on development of Catharanthus roseus (L.)
‘Grape Cooler’. J. Amer. Soc. Hort. Sci. 120:877—881.

85

Pramuk, LA. and ES. Runkle. 2005. Modeling growth and development of celosia and
impatiens in response to temperature and photosynthetic daily light integral. J.
Amer. Soc. Hort. Sci. 130:813-818.

Roberts, E.H. and R.J. Summerfield. 1987. Measurement and prediction of flowering in
annual crops, p. 17-50. In: Atherton, J .G. (ed.). Manipulation of flowering.
Butterworths, Kent, UK.

Shimizu, H. and RD. Heins. 2000. Photoperiod and the difference between day and night
temperature influence stem elongation kinetics in Verbena bonariensis. J. Amer.
Soc. Hort. Sci. 125:576—580.

Shimizu, H., RD. Heins, and E. Runkle. 2003. Simulation study of total energy
consumption required to produce a mature plant at different greenhouse
temperatures. J. Soc. High Tech. in Agr. 15:123—129.

Steininger, J ., C.C. Pasian, and J .H. Lieth. 2002. Extension of a thermal unit model to
represent nonlinearities in temperature response of miniature rose development. J.
Amer. Soc. Hort. Sci. 127:349-354.

Summerfield, R.J., EH. Roberts, R.H. Ellis, and R.J. Lawn. 1991. Towards the reliable
prediction of time to flowering in six annual crops. 1. The development of simple
models for fluctuating field environments. Expt. Agr. 27:] 1—3 1.

US. Department of Agriculture. 2009a. Application technology research unit: Virtual
Grower. Agr. Res. Serv., Washington, DC. 23 Sept. 2009.
<http://www.virtualgrower.net>.

US. Department of Agriculture. 2009b. F loriculture crops 2008 summary. Agr. Stat.
Board, Washington, DC.

US. Department of Energy. 1995. National solar radiation data base 1961-1990: Typical
meteorological Year 2. Natl. Renewable Energy Lab., Golden, CO.

US. Department of Energy. 2009. Monthly energy review August 2009. Energy Info.
Admin, Washington, DC.

van Iersel, M.W. 2003. Short-term temperature change affects the carbon exchange
characteristics and growth of four bedding plant species. J. Amer. Soc. Hort. Sci.
128:100—106.

White, J .W. and 1.] . Warrington. 1988. Temperature and light integral effects on growth
and flowering of hybrid geraniums. J. Amer. Soc. Hort. Sci. 113:354-359.

86

 

 

SECTION IV

MODELING THE INFLUENCE OF GREENHOUSE SCREENS ON PLANT SHOOT-

TIP TEMPERATURE OF NEW GUINEA IMPATIENS

87

 

Modeling the Influence of Greenhouse Screens on Plant Shoot-tip Temperature of

New Guinea Impatiens

Matthew G. Blanchard and Erik S. Runklel

Department of Horticulture, Michigan State University, East Lansing, MI 48824

A.J. Both
Bioresource Engineering, Department of Environmental Sciences, Rutgers University,

New Brunswick, NJ 08901

Hiroshi Shimizu
Division of Environmental Science & Technology, Kyoto University, Kyoto 606-8502,

Japan

Received for publication . Accepted for publication . We gratefully
acknowledge funding by the Michigan Agricultural Experiment Station, the USDA-ARS
Floriculture and Nursery Research Initiative, and greenhouse growers providing support

for Michigan State University floriculture research. We also thank Raker’s Acres for

their contributions to this project and Mike Olrich for his greenhouse assistance.

1Associate Professor of Horticulture and Extension Specialist and to whom reprint

requests should be addressed (Email: runkleer@msu.edu).

88

Abstract

The effect of retractable nighttime screens on plant shoot-tip temperature of New
Guinea impatiens (Impatiens hawkeri Bull.) was quantified in glass-glazed greenhouses
during winter. An energy balance model was developed that predicts shoot-tip
temperature using cover (glazing or screen) emissivity and five environmental
measurements: dry-bulb, wet-bulb, and cover temperature; transmitted shortwave
radiation (SWR, 300 to 3,000 nm); and air velocity. Plants under a screen extended at
night had a shoot-tip temperature 0.5 to 2.3 °C higher and absorbed 70 to 125% more net
radiation (250 to 60,000 nm) than plants without a screen. Shoot-tip temperature was 0.9
to 1.4 °C lower under a shading screen with open-weave construction (high air
permeability) compared with closed-weave constructed screens (e. g., blackout). Under
an air vapor-pressure deficit (VPD) of 0.4 to 0.9 kPa, plant shoot-tip temperature was up
to 1.5 °C closer to dry-bulb temperature than plants under a VPD of 1.4 to 1.8 kPa.
During the day, shoot-tip temperature was 1.2 °C lower than dry-bulb temperature when
SWR was <100 W-m‘z, and a mean of 1.6 °C higher than dry-bulb temperature when
SWR was >100 W-m‘z. The model predicted shoot-tip temperature within 1 °C for 62%
and 91% of the actual shoot-tip temperature during the day and night, respectively
(13,824 observations). Thus, a screen extended at night over a crop of New Guinea

impatiens could increase plant shoot-tip temperature and accelerate development.

Introduction
The commercial production of ornamental greenhouse crops demands that plants

are accurately scheduled for predetermined market dates (Heins et al., 2000). The ability

89

to schedule crops requires knowledge of how environmental factors (e.g., temperature
and light) influence plant growth and development. The rate of plant development and
the time required to complete a phenological stage (e. g., unfold leaves or flower) is
controlled by the mean daily temperature (MDT) of the apical meristem (e. g., shoot-tip)
(Faust and Heins, 1993; Niu et al., 2001; Roberts and Summerfield, 1987). For example,
as shoot-tip MDT increased from 16 to 24 °C, flower development rate in campanula
(Campanula carpatica Jacq.) increased from 0.018 to 0.024 (N iu et al., 2001).

Many published crop models use ambient air MDT to predict the completion of a
developmental stage rather than actual shoot-tip temperature (Adams et al., 1998; Erwin
and Heins, 1990; Moccaldi and Runkle, 2007). However, shoot-tip temperature is often
different from air temperature. Faust and Heins (1998) reported that shoot-tip MDT in
vinca (Catharanthus roseus L.) grown at 15 and 20 °C was within 2 °C of the air MDT,
while plants grown at 35 °C had a shoot-tip MDT 4 to 6 °C below the air MDT. Crop
models that are based on actual shoot-tip MDT can be more accurate than those that use
air MDT. For example, in African violet (Saintpaulia ionantha Wendi), the mean
deviation between predicted and observed leaf number was 63% higher when air
temperature was used instead of plant temperature (Faust and Heins, 1993).

Plant temperature is determined by the transfer of energy between the plant and
the surrounding environment and can be calculated using an energy balance equation
(Nobel, 2005). Under steady-state conditions, plant temperature is equal to the sum of
the total radiation absorbed by the plant, emitted longwave radiation (LWR, 3,000 to
100,000 nm), convection, and transpiration (Nobel, 2005). Plant energy balance models

have been developed that predict shoot-tip temperature under different greenhouse

9O

 

 

 

 

 

conditions (Faust and Heins, 1998; Shimizu et al., 2004). During the day, shortwave
radiation (SWR, 300 to 3,000 nm) and transpiration have the largest influence on energy
transfer between the plant and the surrounding environment (Faust and Heins, 1998).
During the night, convection and the transfer of LWR between the plant and the
greenhouse structure often have the greatest influence on plant temperature.

The exchange of LWR between plants and the surrounding environment is
influenced by the temperature and emissivity of radiating objects. In temperate climates
during winter nights, outside temperatures are low and the greenhouse glazing
temperature can be considerably lower than inside air temperature. As glazing
temperature decreases, LWR emitted by the glazing material decreases, and LWR
emitted by the plant can become greater than incoming LWR. This net loss of LWR can
cause plant temperature to be below air temperature. For example, as glazing
temperature at night decreased from 2 to 16 °C below air temperature, vinca shoot-tip
temperature decreased from 1 to 5 °C below air temperature (Faust and Heins, 1998).

In response to increasing and volatile fuel costs, some commercial growers have
made greenhouse structural improvements to reduce energy losses, such as the
installation of retractable thermal screens (Dieleman and Kempkes, 2006; Komer et al.,
2004; Lund et al., 2006). Thermal screens are commonly extended over a greenhouse
crop from sunset to sunrise and retracted during the day (Bailey, 1988; Lund et al., 2006).
A thermal screen that has a closed-weave construction (low air permeability) and seals
tightly with the greenhouse sidewalls creates a barrier between the heated and non-heated
space above and below the screen, respectively (Oztiirk and Bascetincelik, 1997;

Zabeltitz and Meyer, 1984). Energy requirements for greenhouse heating can decrease

91

with a thermal screen because the volume of heated air below the screen is reduced.
Brajeul et al. (2005) and Le Quillec et al. (2005) reported that a greenhouse with a
thermal screen closed at night consumed 22 to 41% less energy than a greenhouse
without a thermal screen.

Another benefit of retractable screens is the potential to increase LWR absorbed
by the crop (Kittas et al., 2003). During the night, the interior surface temperature of a
thermal screen is often higher than glazing temperature. Therefore, when a screen is
extended over a crop, plants receive more incoming LWR than without a screen. The
higher net LWR (LWRnet) exchange can increase plant temperature. For example, a rose
(Rosa hybrida L.) crop grow under a greenhouse thermal screen with 75% SWR
transmission had 100% higher absorbed LWR and 1 to 3 °C higher canopy temperature at
night than a crop grown without a screen (Kittas et al., 2003). Studies with other crops
such as African violet, tomato (Lycopersicon esculentum Mill), and vinca have also
reported higher canopy, leaf, or shoot-tip temperatures under screens compared with no
screens (Bailey, 1977 and 1981a; Faust, 1994). However, to our knowledge, crop models
that predict the influence of thermal screens on plant shoot-tip temperature during the
night have not been published.

The objective of this study was to quantify the effect of different screen materials
on shoot-tip temperature of New Guinea impatiens (Impatiens hawkeri Bull.) during cold
nights. Our goal was to develop a shoot-tip temperature model using environmental
factors that can be measured in commercial greenhouses. New Guinea impatiens was
selected because it is among the top 10 bedding plants produced in the United States,

with a reported wholesale value of $54 million in 2008 (US. Department of Agriculture,

92

 

 

2009). Since plant development of New Guinea impatiens is delayed considerably at a
low temperature (Erwin, 1995), a model that predicts shoot-tip temperature could

improve production scheduling and management of the greenhouse environment.

Materials and Methods
Greenhouse Experiment

Plant material. On 26 Nov. 2008, rooted vegetative cuttings of New Guinea
impatiens ‘Supersonic Flame’ grown 50-cell trays (2.5 X 2.5 cm; 27.2-ml volume) were
received from a commercial greenhouse (Raker’s Acres, Inc., Litchfield, Ml). Plants
were subsequently transplanted into ll-cm round plastic containers (600-m1 volume)
filled with a commercial soilless peat-based medium (Suremix; Michigan Grower
Products, Inc., Galesburg, MI) and were grown at a constant temperature set point of 23
°C. The photoperiod was a constant 16 h that consisted of natural photoperiods (42.7 °N
lat.) with day-extension lighting from 0600 to 2200 HR provided by high-pressure sodium
(HPS) lamps that delivered a photosynthetic photon flux of 75 to 100 umol-m‘z's‘l at
plant height (as measured with a line quantum sensor [Apogee Instruments, Inc., Logan,
Utah]).

Ethephon (Florel; Bayer CropScience LP, Research Triangle Park, NC) with a
surfactant (Capsil; Aquatrols, Paulsboro, NJ) was applied as a foliar spray at 300 mg-L‘l
and 0.2 L'm‘2 every 2 to 3 weeks to abort flower buds. Plants were irrigated as
necessary with reverse osmosis water supplemented with a water-soluble fertilizer
providing (mg-L“) 125 N, 12 P, 100 K, 65 Ca, 12 Mg, 1.0 Fe and Cu, 0.5 Mn and Zn,

0.3 B, and 0.] Mo (MSU RO Water Special; GreenCare Fertilizers, Inc., Kankakee, IL).

93

 

After 7.5 weeks, leaf canopy exceeded the container diameter and plants were transferred
to different greenhouses for the screen study.

Greenhouse environment. The research facility for the screen experiment
consisted of a row of six connected (4.8 X 4.1 m) glass-glazed greenhouse compartments
oriented east-west. Compartments were separated by a glass wall orientated north-south.
Only the middle four compartments were used for the experiment. In each compartment,
25 plants were grown on a bench that was positioned in the middle of the greenhouse.
Plants were spaced so that leafs of adjacent plants touched and hence formed a full
canopy.

In three compartments, a screen located horizontally at eave height was manually
extended over plants at 1700 HR and retracted at 0800 HR. Screens were closed so that
they contacted the greenhouse sidewalls. Control plants were grown in one compartment
without a screen. Six 4-d experiments were performed during Jan. and Feb. with
different screens and under different air vapor-pressure deficits (VPDair). The screens
used in the study were obtained from Ludvig Svensson, Inc. (Charlotte, NC) and
included: (1) blackout with aluminized (AL) polyester and 0% light transmission
(blackout AL); (2) blackout with black polyester and 0% light transmission (blackout
BL); (3) closed-weave screen with alternating 5-mm wide AL and transparent plastic
(TP) strips and 46% light transmission (energy AL/TP); (4) open-weave screen composed
of alternating 5-mm wide AL, TP, and no strips with 50% light transmission (shade
AL/TP/X); (5) closed-weave screen with 5-mm wide TP strips and 88% light
transmission (energy TPl); and (6) closed-weave screen with S-mm wide TP strips and

83% light transmission (energy TP2) (Table 4.1).

94

Temperature in each greenhouse compartment was controlled by an
environmental computer (Priva Intégro 724; Priva, Vineland Station, Ontario) that
controlled steam heating and passive and active ventilation when needed. The air
temperature set point in each compartment was a constant 20 °C and the actual mean
daily dry-bulb temperature was 19.9 :t 0.7 °C (:t SD). The greenhouse was heated with
steam-heating units located below the benches along the knee wall of each compartment.
During the screen study, plants were grown without HPS lamps and under natural
photoperiods. The calculated time from sunrise to sunset ranged from 9 to 11 h. Plants
were grown with minimal overhead obstructions and were irrigated as previously
described without water stress.

In each compartment, dry-bulb and wet-bulb temperature were independently
measured by aspirated, shielded thermocouples (0.13-mm type B; Omega Engineering,
Stamford, CT) positioned at plant level and adjacent to the canopy. Shoot-tip
temperatures was independently measured on three plants in each compartment by
inserting fine-wire thermocouples (Type E; Omega Engineering) 30.3 cm below the
shoot apex. Thermocouples were reinserted every 4 d as plants unfolded leaves. At 15
cm above the canopy in each compartment, a pyranometer (LI-2008A; LI-COR, Lincoln,
NE) measured shortwave radiation (SWR, 300 to 3,000 nm). At the same height, net
radiation was measured in two compartments using total hemispherical radiometers
(THR) [(250 to 60,000 nm) THRDS7.1; Radiation and Energy Balance Systems, Inc.,
Seattle, WA] and in one compartment using a net radiometer [(300 to 50,000 nm) CNRl;
Kipp & Zonen USA Inc., Bohemia, NY]. Prior to the study, radiation measurements

from the two THRs were collected over 2 days (307 measurements per instrument) and

95

compared against the net radiometer (reference) and a calibration equation was
developed. The equation was subsequently applied to data from the THRs and adjusted
values were calculated.

In each compartment, canopy temperature and cover temperature were recorded
with infrared (IR, 6,500 to 14,000 nm) sensors (Type K, OS36-01; Omega Engineering
Inc.) positioned 15 cm above the canopy and oriented vertically upward (cover) or
downward (canopy). Cover temperature represented the greenhouse glazing and
superstructure temperature when the screen was retracted and the surface temperature of
the screen when it was extended over the crop. In two compartments, air velocity
transducers (8470 and 8475; T81, Inc., St. Paul, MN) positioned 2 cm from the shoot-tip
measured air velocity and was 0 to 0.1 ms“. Active ventilation with fans did not occur
during the study. In the two compartments in which air velocity was not measured
continuously, air velocity was measured instantaneously at the same location and was
found to be $0.05 ms"1 of the other compartments.

Environmental greenhouse measurements were collected every 10 s and 10-min
means were recorded by two data loggers (CR10; Campbell Scientific, Logan, UT).
Outside air temperature was measured every 5 min by a Priva weather station located 2 m
above the greenhouse peak and data was recorded by the greenhouse environmental-
control computer. During three experimental periods, VPD,“r was maintained at 0.4 to
0.8 kPa by the injection of steam into the air. During the other three experimental
periods, VPD,“r was not controlled and ranged from 1.4 to 1.8 kPa.

Data analysis. Data for cover temperature, net LWR, and the difference between

shoot-tip and dry bulb temperature were analyzed for each 4-d experimental period by

96

using the lO-min means for each recorded parameter from each treatment. Data were

analyzed using SAS (SAS Institute, Inc., Cary, NC) mixed model procedure (PROC

 

MIXED), and pairwise comparisons between treatments were performed using Tukey’s

honestly significant difference test at P S 0.01.

Model Development

Plant shoot-tip temperature (Tshow) is determined by the balance of incoming and
outgoing energy. Incoming energy must equal outgoing energy or Tshoo, will increase or
decrease until an equilibrium is obtained. The relationship between the energy gained
and the energy lost from a plant shoot-tip can be described as (Monteith and Unsworth,

2008) (Fig. 4.1):

Shortwave radiation + longwave radiation + convection + transpiration = 0 - [1]

We developed a model to predict New Guinea impatiens T Shoo, using the following

equations to describe the individual components of the energy balance equation (Eq. [1]):

Shortwave radiation. SWR absorbed (S WRabs) by the shoot-tip can be described

as:

SWRabs = “SWR x SWRmcS [2]

97

where “SWR is the shoot-tip absorptivity for SWR and S WR ms is incident SWR on the
shoot-tip. A typical mean absorbance value for plants of 0.5 was used for “SWR (Nobel,
2005)

Longwave radiation. In greenhouses, closely spaced plants (as in this study) form
a horizontal surface and exchange LWR with the greenhouse ceiling (e. g., glazing or
screen) (Jones, 1994). The net flux of LWR (I. WRne,) absorbed by a shoot-tip is the sum

of incoming and outgoing radiation and can be described as (Holman, 1997):

E . - E .
L WRnet = (SFL WRnet) X 1 BC lBShOOl (Wm—2) [3]

__ ___-__-- _ 1

8c 8shoot

 

where —l—+ 1 ——l [4]

8c 8shoot

 

is resistance to LWRne, and is calculated with the emissivity of the cover (ac) and shoot-
tip (asham). The emissivity for the bottom surface (facing downward towards crop) of
each screen used in the experiment is presented in Table 4.1. An emissivity value of 0.9
and 0.96 was used for the glass—glazing and shoot-tip, respectively (N ijskens et al., 1984;
Nobel, 2005). The emissive power of a black body at a cover temperature (E Bc) and at

Tshoo, (£135,100,) is calculated as:

EB = (Ix T4 (Wm-2) [5]

98

where o is the Stephan-Boltzmann constant (5.67 E—8) and T is the surface temperature
of the cover or shoot-tip in Kelvin. The shape factor for L WRne, (SF L WRnet) can be

described as (Shimizu et al., 2004):

d1

SF = —— 6
LWRnet dl +(4x hi) I I

where d1 and hl are the diameter (m) and height (m) of the shoot-tip, respectively. New
Guinea impatiens plants used in this experiment had a d1 and hl of 0.0032 and 0.003 m,
respectively, which were means of ten measurements made with a digital caliper.

Convection. Two types of convection (Conv) can be described as (Gates, 2003):

Free convection: Gr > Re2 [7]

Forced convection:Gr < Re2

The Grashof (Gr) and Reynolds (Re) numbers are calculated as:

Gr: @ngd3)x(lszb —Tshoot I) [8]
V

_ de

— V

Re

 

[9]

where B is the coefficient of volumetric thermal expansion [for air, [3 equals the inverse of
the dry-bulb temperature (Tdb) in Kelvin], g is the acceleration of gravity (9.80 m-s‘z), V
is the air velocity (m-s“), v is the kinematic viscosity of air (mZ-s’l) at the Tdba and d is

the characteristic dimension of the shoot-tip:

99

.=[L..L]“ (m) [10]
d1 hl

Free Conv can be classified as laminar or turbulent flow based on the value of Gr

multiplied by the Prandtl number [(Pr), 0.71] (Holman, 1997):

[11]

l
Laminar flow:104 < erPr <109 -——)h‘ =1.42x LTdb -TSh00‘ I 4(W-m‘z-K—l)
C d

 

I
Turbulent flow:109 < (er Pr)—) he = 1.31 x (I Tdb 45,00, 05 (Wm—2K") [12]

where he is the Conv heat-transfer coefficient.

Forced Conv across a cylinder (e.g., plant shoot-tip) can be described as:

l

N -_
hc=[-3-)X(C)X(an'] xPr3 (W-m"2-K") [13]
v

 

where K is the thermal conductivity of air (0.024 W-m‘z-K_') and C and N are constants

that depend on Re (Table 4.3).
Using the calculated he value for free Conv (Eqs. [11] or [12]) or forced Conv

(Eq. [13]), the total energy exchange by C onv is described as (Gates, 2003):

Conv = hc x (Tdb — Tsh00,) (w-m-Z) [14]

100

 

 

T ranspiration. The conductance of water vapor from the plant to the air (IE) is

described as (Monteith and Unsworth, 2008):

 

1E = [P X CP x (es (TShOOI )- 3(Tdb »

. -2

where p is the density of air at the mean temperature of T db and T Shoo, in Kelvin, C p is the
specific heat of air (1010 J -kg-’-K"), es(Tsh00,) is the saturation vapor pressure at Tshoop
e(Tdb) is the vapor pressure at Tdb, y is the psychrometer value (0.0663 kPa-K"), Rb is
the boundary layer resistance for water vapor, and RC is the cuticle resistance of the shoot
tip. Rb has been reported to be 0.93 times that of convective heat transfer resistance and

can be calculated as (Monteith and Unsworth, 2008):

 

xC
Rb = 0.93%p h P] (s-m“) [16]

cl

Shoot-tip RC is species dependent and must be experimentally calculated (Shimizu
and Heins, 2000). Leaf RC for New Guinea impatiens at a Tdb of 18 to 24 °C and under
S WRmeS of 30 to 450 W-m‘z’1 has been reported to range from 20 to 350 smfl (Al-Faraj et
al., 1994; Mankin et al., 1998). To our knowledge, there is no published information on
shoot-tip RC for New Guinea impatiens. Therefore, measured environmental data and
New Guinea impatiens Tshoot data from the validation study were used to generate
equations that predict RC. Equations were developed by first calculating S WRabs,
LWRne,, and Conv from measured greenhouse data and solving Eq. [1] for IE. RC was

then estimated by rearranging Eq. [15]:

101

 

[17]

 

R [P x C p X (es (T shoot )- 8(Tdb ll]
c = "‘ Rb
yxkE

During the night, RC was found to be related to shoot-tip vapor-pressure deficit
(VPDSh00,) and LWRne,. Therefore, when SWRmes is <5 W'm’z, RC was described with

the following equations:

LWRne, < 0 .- RC = 257.4 + (445.3 x (e, (Tshoo, )— e(Tdb )))+ (84.4 x LWR)

18
LWRne, 20: RC =354.1 + (649.1 x(es(Tsh00,)—e(Tdb)))+(84.8 xLWR) [ 1

where RC is always 20.

The environmental conditions used to generate these equations included 8,694
observations for RC at a T db of 18.3 to 22.3 °C, under a VPDshoot of 0.1 to 1.6 kPa, and
under LWRne, of-10.0 to 3.9 W-m‘z. During the day when SWRmeS was 5 to 530
W'm‘z, LWRne, was -17.9 to 5.2 W-m‘z, Tdb was 15.0 to 23.4 °C, and VPDShOO, was 0.2
to 2.6 kPa, we found no relationship between RC and VPDshoozs VPDair, S WRmes, LWRne,,
Tdb-Tsh00,, or Tshoot- Therefore, a constant RC during the day was assumed and was
calculated from the median of 5,114 observations (373 s-m‘z).

The complete model to predict New Guinea impatiens T Shoo, under different
greenhouse conditions was developed by substituting the components of Eq. [1] with Eqs.
[2], [3], [l4], and [15]. A similar modeling approach was used to predict poinsettia Tshoot
(Shimizu and Heins, 2004). The model for New Guinea impatiens can predict T Shoo,
using five greenhouse environmental factors: T db: wa, Tcover, S WRmes, and V. The

model also requires values for the diameter and height of the plant shoot-tip, SWR

102

absorptivity of the shoot-tip, and emissivity of the glazing or screen and the shoot tip.
Since the model can not be analytically solved, a computer program was written and
compiled in Microsoft Visual Basic 6.5 (Microsoft Corp, Redmond, WA) that used the
root bisection technique (Wainwright and Mulligan, 2005) to find a value for shoot-tip

temperature that satisfied [Eq. 1].

Results and Discussion
Model Validation

Environmental data (dry-bulb, wet-bulb, and cover temperature, measured SWR,
and air velocity) from the greenhouse experiment and the emissivities of the glazing,
screens, and plant were entered into the New Guinea impatiens shoot-tip model and the
simulated shoot—tip temperatures were compared with the measured temperatures. The
validation included measured data from 24 d and consisted of 5,184 observations for the
day (screens open from 0800 to 1700 HR) and 8,640 observations for night (screens
extended over crop). The night period for plants without a screen was considered to be
from 1700 to 0800 HR during which the maximum measured SWR was 20.2 W°m‘2. The
actual sunset ranged from 1736 to 1815 HR and the actual sunrise ranged from 0729 to
0803 HR. During the night under all treatments, the simulated shoot-tip temperature was
within 1 °C and 2 °C of the measured temperature 91% and 100% of the time,
respectively (Fig. 4.2 to 4.3). During the day, the simulated shoot-tip temperature was
within 1 °C and 2 °C of the measured temperature 63% and 89% of the time,

respectively.

103

Influence of the Greenhouse Environment on Shoot-tip Temperature

During each experimental period and under all treatments, the measured mean
shoot-tip temperature during the night was 0.6 to 3.8 °C lower than the mean dry-bulb
temperature (Table 4.4 and Figs. 4.4 to 4.6). This shoot-tip temperature depression at
night is in agreement with a poinsettia model that predicted shoot-tip temperature to be
4.2 °C lower than dry-bulb temperature at an air velocity of 0.05 ms‘1 and a dry-bulb,
wet-bulb, and glazing temperature of 25, 22, and 0 °C, respectively (Shimizu et al.,
2004). During the night, shoot-tip temperature depression occurs because energy losses
from the plant exceed energy gains. Shoot-tip temperature at night can never be greater
than air temperature without an infrared heat source (Rotz and Heins, 1982).

After the screens were retracted at 0800 HR, dry-bulb and shoot-tip temperature
decreased on some days by up to 1.8 and 5.4 °C, respectively (Figs. 4.4 to 4.6). This
temperature drop generally lasted for 30 to 60 min. During the day, plants had a
measured mean shoot—tip temperature 1.2 °C lower than dry-bulb temperature when SWR
was <100 W-m‘z, and a mean of 1.6 °C higher than dry-bulb temperature when SWR
was >100 W-m‘z. Linear regression analysis performed on measured data predicted that
as SWR increased from 5 to 250 W-m‘z, the difference between shoot-tip and dry-bulb
temperature increased linearly from -2.0 to 3.3 °C (data not shown). In African violet,
shoot-tip temperature was up to 4 °C above the dry-bulb temperature on a sunny day (100
to 150 W-m‘z) and 1 to 3 °C below the dry-bulb temperature on a cloudy day (550
W-m‘z) (Faust and Heins, 1993).

Within each experimental period, plants under a screen extended during a cold

night had a shoot-tip temperature 0.5 to 2.3 °C higher than plants without a screen (Table

104

 

 

 

4.4 and Figs. 4.4 to 4.6). This is in agreement with Faust (1994) who reported that at a
nighttime glazing temperature of -1 .5 °C, vinca had a mean shoot-tip temperature 2.9 °C
higher under a screen compared to without a screen. Similarly, tomato grown under a
thermal screen at night had a leaf temperature 0.5 °C higher than unscreened plants
(Bailey, 1977).

The lower shoot-tip temperature in unscreened plants can be at least partially
attributed to a net loss of LWR from the plant to the glazing at night (Bailey, 1977; Kittas
et al., 2003). In a greenhouse, plants exchange LWR with the superstructure and sky
above (Silva et al., 1991). According to the Stephan-Boltzmann law, emission of
radiation from an object decreases as the surface temperature of the object decreases
(Nobel, 2005). During the experimental period, plants that had a screen extended over
them during the night had a cover temperature 0.8 to 6.9 °C higher than plants that were
exposed to the glazing (Table 4.4). As cover temperature at night decreased, emitted
LWR from the screen or glazing decreased and LWRnet absorbed by the plant decreased
(data not shown). For example, LWRnet decreased from 5.7 to -22.9 W-m“2 and shoot-
tip temperature decreased by 1.6 °C as cover temperature decreased from 20.1 to 14.4 °C
during period 1.

Plants under a high air VPD (1.4 to 1.8 kPa) had a similar or up to 1.5 °C lower
shoot-tip temperature at night than plants under a low VPD (0.4 to 0.9 kPa) (Table 4.4).
If substrate moisture is not limiting, plant water uptake and transpiration increase as VPD
increases (Mankin et al., 1998). Since the evaporation of water is an endothermic
process, transpiration represents the loss of latent energy from a plant (Nobel, 2005). The

simulation model predicted that at a T db, Tcover, and Vof 20 °C, 19 °C, and 0.1 ms",

105

 

respectively, energy lost through transpiration at night would increase by 78% and shoot-
tip temperature would decrease by 0.4 0C as air VPD increased from 0.5 to 2.0 kPa.
Under the same environmental conditions, but with a Tcove, of 14 °C, the model
predicted energy lost through transpiration at night would increase by 24% and shoot-tip
temperature would decrease by 0.2 °C as air VPD increased from 0.5 to 2.0 kPa
Similarly, vinca shoot-tip temperature decreased by 1 °C when the VPD at night
increased from 0.5 to 3.0 kPa (Faust and Heins, 1997). Mankin et al. (1998) also
reported that New Guinea impatiens leaf temperature during the day decreased by 1 °C as
air VPD increased from 0.5 to 1.5 kPa.

Shoot-tip temperature depression at night can also be caused by the convection of
energy between the air and the plant. The amount of convection is influenced by the
resistance of the boundary layer around the shoot-tip or leaf (Nobel, 2005). As air
velocity decreases, boundary layer thickness and resistance to convective heat transfer
increase. In this study, boundary layer resistance during the night was probably high
because horizontal airflow fans (HAF) were turned off and mean air velocity was 0
m-s’l. New Guinea impatiens shoot-tip temperature at night could have been closer to
air temperature with air movement. The simulation model predicted that at a Tdb» T 1va
and Tam, of 20 °C, 14 °C, 19 °C, respectively, shoot-tip at night would increase from
18.2 to 19.5 °C if air velocity increased from O to 0.5 m-s‘l. These results reinforce the
importance of operating HAF fans at night in commercial greenhouses to increase
convective heat transfer and thus plant temperature (Bartok, 2005).

Greenhouse screens are available from several manufacturers and are constructed

of materials with different thermal and radiometric transmission properties (Bailey,

106

1981b; Cohen and Fuchs, 1999; Winspear and Bailey, 1978). A characteristic that can

 

vary among screens is the emissivity. Emissivity is the ratio of radiation emitted from a
material to the radiation emitted by a black body at the same temperature and ranges from
0 to 1 (Gates, 2003). As the emissivity of a material decreases, reflectivity and emitted
radiation increase. The screens used in this study had a bottom emissivity that ranged
from 0.5 to 0.88. In the present study, plants under closed-weave screens constructed of

aluminized blackout (blackout AL), blackout with black polyester (blackout BL), and

 

alternating AL and TP strips (energy AL/T P) had a shoot-tip temperature within 0.4 °C of
each other (Table 4.4). Among these screens, there were no consistent trends between
experimental periods in which screen elicited the greatest shoot-tip temperature.
Similarly, vinca shoot-tip temperature under various closed-weave black or aluminized
blackout screens was within 0.7 °C (Faust, 1994). The simulation model predicted that at
a Tdb’ wa, Tcover, and Vof 20 °C, 13 °C, 16 to 20 °C, and 0 to 0.5 m-s‘l, respectively,
shoot-tip temperature would be within 0.3 °C under a screen with an emissivity of 0.1
versus 1.0. Therefore, at the screen surface temperatures observed in this study, screen
emissivity had little influence on shoot-tip temperature.

Although some of the screens had a similar emissivity, they had different bottom
surface temperatures and different effects on plant shoot-tip temperature. For example,
during periods 5 and 6, screens constructed of all transparent strips (energy TPl and
energy TP2; emissivity: 0.5 to 0.54) had a nighttime bottom surface temperature 1.7 to
2.6 °C lower than plants under Blackout AL (emissivity= 0.6) (Table 4.4 and Fig. 4.6).
Under these same screens, plants had a shoot-tip temperature under 0.2 to 1.0 °C lower

than plants under Blackout AL. The lower surface temperatures of some screens could

107

be related to differences in thermal conductivity, permeability, and thickness of the
screen materials (Miguel et al., 1997; Nijskens et al., 1984; Winspear and Bailey, 1978).
Materials that are thick or have low conductivity for heat transfer can elicit a high
temperature gradient between the material surfaces (N ij skens et al., 1984). Therefore, a
screen with high thermal conductivity can have a similar temperature on the upper
surface exposed to the glazing as that of the lower surface exposed to the canopy. One
strategy to increase screen temperature could be to utilize multiple horizontal shading or
blackout screens (Heins and Runkle, 2004). The extension of two screens over a crop at
night could increase the surface temperature of the lower screen and increase LWR
absorbed by the plant.

Some greenhouse growers operate internal shading systems during the day when
solar irradiance and outdoor temperature are high to reflect SWR and to prevent high
temperature stress on crops (Heins and Runkle, 2004; Hoffrnann and Waaijenberg, 2002).
In this study, plants under a 50% shade screen with open-weave construction (shade
AL/TP/X) had a nighttime shoot-tip temperature 0.2 to 0.6 °C higher than plants without
a screen, but 0.9 to 1.4 °C lower than plants under Blackout AL or Shade CW (Table 4.4
and Fig. 4.4). These results indicate that the operation of an open-weave shade screen to
reduce SWR transmission during the day can also increase shoot-tip temperature at night.
However, there is a tradeoff between the shading percentage of a screen and its potential
to reduce energy inputs for greenhouse heating. As the light transmission and air
permeability of a screen increase, its insulative value decreases (Ludvig Svennson, 2009).

If a screen is constructed of a material with a very low emissivity (e. g.,

aluminum) and the heat permeability is high, then its surface temperature could decrease

108

below the dew-point temperature (Bailey, 1981b; Meijer, 1980). At surface temperatures
lower than the dew-point temperature, water can condense on the screen and drip onto the
crop below. In this study, screen temperature was always above the dew-point
temperature and condensation did not occur. To minimize condensation in a greenhouse
environment with high relative humidity, the screen surface with the lowest emissivity
(i.e., high reflectivity of SWR) could be oriented upward and the surface with highest
emissivity (high LWR absorption) could be orientated downward (Bailey, 1981b; Cohen
and Fuchs, 1999).

The New Guinea impatiens shoot-tip model underestimated temperatures by up to
7.7 °C during the day when SWR was >100 W'm’z. The deviation between simulated
and measured shoot-tip temperature during the day could be at least partially explained if
the shoot-tip SWR absorptivity factor for New Guinea impatiens is greater than 50%.
When a simulation was performed with a SWR absorbtivity factor of 80%, the model
accuracy slightly improved between 1200 to 1400 HR, but was worse during the morning
and late afternoon. Alternatively, the lower temperature predictions of the model during
the day could be caused by over estimating the energy lost from transpiration. If
transpiration decreased under high SWR, then shoot-tip temperature would increase.
When the model was simulated with a very high cuticle transpiration resistance of 3,000
s-m-l (i.e., low stomatal conductance), simulated shoot-tip temperatures increased and
became closer to actual temperatures only when SWR was >150 W-m‘z. However,
published research on transpiration in New Guinea impatiens does not indicate that
stomata would close during the day under these environmental conditions (Al-Faraj et al.,

1994; Mankin et al., 1998; Pang, 1992).

109

The energy balance model used in this study is similar to the one developed by
Shimizu et al. (2004) to predict poinsettia shoot-tip temperature. However, our model did
not include a SWR shape factor to describe the ratio of shoot-tip surface area exposed to
direct SWR to the total surface area. When a SWR factor was included in the model,
simulated shoot-tip temperatures throughout the day were often greater than 2 °C below
that measured. The accuracy of the model during the day improved when SWR absorbed
by the shoot-tip was calculated based on the shoot-tip absorptivity for SWR and
measured SWR (Faust and Heins, 1994).

Another difference between our model and the one for poinsettia is that we
assumed that shoot-tip cuticle resistance for transpiration is not constant during the night
(Al-Faraj et al., 1994; Mankin et al., 1998). The New Guinea impatiens model predicted
cuticle resistance at night to increase as LWRnet increased and as the VPD between the
shoot-tip and air increased. This response is in agreement with Mankin (1994) who also
determined a positive relationship between increasing leaf VPD and New Guinea
impatiens cuticle resistance. The shoot-tip model predicted that plants grown under a
screen at night would have higher absorbed LWRnet and greater cuticle resistance than
plants grown without a screen. For example, at a Tdbz 20 °C, wa= 12 °C, Tcover= 19 °C,
egs= 0.9, and V= 0.1 m°s"1, shoot-tip temperature and cuticle resistance at night would be
19.1 °C and 849 s-m", respectively. Under the same environmental conditions, but with
Tamer: 14 °C, shoot-tip temperature and cuticle resistance would be 18.3 °C and 442
s-m-1, respectively.

Several C3 and C4 annual and herbaceous perennial species, including New

Guinea impatiens, maintain stomatal conductance and transpiration at night (Mankin et

110

al., 1998; Pang, 1992; Snyder et al., 2003). Therefore, nighttime transpiration and cuticle
resistance represent important components of the energy balance model for New Guinea
impatiens. The model predicted that shoot-tip transpiration would occur under any VPD.
To prevent a shoot-tip temperature depression at night, the energy lost from transpiration
must be balanced with the energy gained from convection and absorbed LWR. During
winter in temperate climates, VPD at night can increase considerably because of
infiltration of outside air containing little water vapor and the subsequent heating of this
air (Hanan, 1990). These results indicate that during production of New Guinea
impatiens, injecting water vapor into the air to decrease VPD would reduce transpiration
and increase plant temperature.

In this study, plants exposed to the glazing had a measured mean nighttime shoot-
tip temperature 2.4 to 4.4 °C lower than the greenhouse temperature set point of 20 °C.
At temperatures below :17 °C, New Guinea impatiens development is severely delayed
and there is greater variability in flowering (Erwin, 1995; Runkle 2008). Therefore,
increased shoot-tip temperature at night under a screen can accelerate plant development.
For example, time to flower of New Guinea impatiens ‘Celebrette Peach’ decreased by 3
d for every 1 °C increase in mean daily air temperature between 18 to 26 °C (Whitman et
al., 2000). Assuming the cultivar in this study had a similar developmental response to
mean daily shoot-tip temperature, a crop grown under a screen extended at night would
flower up to 5 d earlier than a crop grown without a screen. These results support that a
retractable greenhouse screen has the potential to decrease energy costs for heating

(Brajeul et al., 2005) and also increase plant temperature and reduce production time.

111

Table 4.1. Characteristics of greenhouse screens used in experiment. Information
obtained from manufacturer (Ludvig Svensson, Inc., Charlotte, NC). The bottom side of
each screen was orientated towards the crop. Control plants were grown in a greenhouse
without a screen and were exposed to glass glazing (emissivity= 0.9). SWR= shortwave
radiation, 300 to 3,000 nm; AL= aluminum; BL= black; E= Empty space crossed with
pohlester threads; TP= trangmrent polyethylene.

SWR Energy

 

 

 

 

Screen treatment transmission savings Emissivity
Description (%) (%) bottom Product name
Blackout AL BL woven polyester; 0 75 0.61 XLS Revolux
Top: polished AL FB A/A
coating with TP ‘
threads; Bottom: dull
AL coating with BL
threads
Blackout BL BL woven polyester 0 75 0.88 XLS Obscura
with BL threads on Revolux B/B
both sides
Energy AL/TP AL-TP stripsZ woven 46 57 0.54 XLS 15 PB

with TP threads; Top:
polished AL; Bottom:
dull AL
Shade AL/TP/X AL-TP-AL-E-AL—E 50 20 0.50 XLS 15 F FB
stripsZ woven with TP
threads; Top: polished
AL; Bottom: dull AL

Energy TPl TP-TP stripsZ woven 88 43 0.50 SLS 10 Ultra
with TP threads Plus

Energy TP2 TP-TP stripsZ woven 83 47 0.54 XLS 10 Ultra
with TP threads Revolux

 

Z5-mm wide parallel strips with different materials.

112

Table 4.2. Abbreviations, symbols, descriptions, and units of parameters in the New

Guinea impatiens shoot-mtemperature model described in Eqs. [1-18]

 

 

Abbreviation

or symbol Description Unit

B Coefficient of volumetric thermal expansion K—1

7 Psychrometer value (6.63 E—2) kPa~K’l
71E Transpiration W-m‘2
1c Thermal conductivity of air (2.4 E-2) W-m‘"2'K_l
p Density of air Kg-m’3
o Stefan-Boltzmann constant (5.67 E-8) W-m‘Z-K‘4
cc Greenhouse glazing or screen emissivity

Sshoor Plant shoot-tip emissivity

a Coefficient of thermal expansion rate of air Ki"l
“SWR Plant shoot-tip absorptivity

C Constant for calculating convection heat-transfer coefficient

C onv Convection W-m‘2
Cp Specific heat of air J -kg_’-K'l
d Plant shoot-tip characteristic dimension m

d1 Plant shoot-tip diameter m

e( Tdb) Vapor pressure at dry-bulb temperature kPa

es( Tshm) Saturation vapor pressure at shoot-tip temperature kPa

E B Emmisive power of a black body W-m"2
E Bc E B at cover temperature W-m‘ 2
E Bshoot E B at shoot-tip temperature W-m’2
g Acceleration of gravity m-s‘2
Gr Grashof number

hl Plant shoot—tip height in

he Convection heat-transfer coefficient W-m‘z'K_I
LWRne, Net longwave radiation W-m’2
N Constant for calculating convection heat-transfer coefficient

Pr Prandtl number (0.71)

Rb Boundary layer resistance for mass transfer s-m-l
RC Cuticle resistance of shoot-tip sm—l
Re Reynolds number

SF L WRnet Shape factor for longwave radiation

S WRabs Shortwave radiation absorbed by shoot-tip W-m’2
S WRmes Measured shortwave radiation Wm"?-
T Temperature K

Tdb Dry-bulb temperature °C
Tcove, Cover (glazing and superstructure or screen) temperature °C
Tshoot Shoot-tip temperature °C

T wb Wet-bulb temperature °C

v Kinematic viscosity of air mz-s‘l

 

113

Table 4.2 (cont’d).

 

 

Abbreviation

or symbol Description Unit
V Air velocity m-s"l
VPDShOO, Shoot vapor-pressure deficit [es( Tsh00,)— e( Tdb)] kPa

 

Table 4.3. C and N (unitless) with Eq. [13] (Gates, 2003).

 

 

 

Reynolds number (Re) C N

0.4 to 4 0.989 0.330
4 to 40 0.911 0.385
40 to 4,000 0.683 0.466
4,000 to 40,000 0.193 0.618
40,000 to 400,000 0.0266 0.805

 

 

114

Table 4.4. The effects of different screens on measured net longwave radiation (1. WRne,)
exchange between New Guinea impatiens shoot-tip and the cover (glazing and
superstructure or screen) and the difference between measured shoot-tip temperature
(Tshom) and dry-bulb temperature (T db) in glass-glazed greenhouses during the night
(1700 to 0800 HR) in winter in East Lansing, MI (lat. 43 °N) under different air vapor-
pressure deficits (VPDair). Each experimental period was 4 d. The mean T db during the
experiments was 19.5 to 20.2 °C. See Table 4.1 for a description of screen materials.
Town-aye= outside air temperature; Tcovefi cover (glazing and superstructure or screen
temperature).

 

 

Expt.
period Mean Toutside VPDai, Mean LWRne, Mean
(110') (°C) Screen (kPa) Tcover (°C) (“I'm—2) Tshoot_T db (°C)
1 -6.8 Blackout AL 1.6 20.1 a7- 5.7 a -2.1 b
Energy AL/TP 1.5 18.9 b 1.3 by -1.9 a
Shade AL/TP/X 1.6 17.1 c -2.8 c —2.3 c
None 1.8 14.4 d -22.9 d -3.7 d
2 -9.8 Blackout AL 0.5 19.5 a 2.7 a -1.2 a
Energy AL/TP 0.5 18.5 b -0.8 by -1.3 a
Shade AL/TP/X 0.6 16.4 c -5.2 c -1.8 b
None ' 0.7 14.0 d -23.5 d -2.7 c
3 —4.7 Blackout AL 1.4 19.1 b 1.3 b -1.5 a
Blackout BL 1.5 20.1 a 2.2 a --1.7 a
Energy AL/TP 1.4 18.6 c 0.2 cy -1.9 b
None 1.5 13.2 d -25.4 d -3.8 c
4 -10.8 Blackout AL 0.6 18.4 a —0.1 a -1.6 b
Blackout BL 0.4 18.6 a —2.2 b -1.4 a
Energy AL/TP 0.5 17.8 b -1.5 by —1.3 a
None 0.9 13.0 c —27.7 c —3.2 c
5 —2.9 Blackout AL 1.4 19.4 a 2.1 a -1.3 a
Energy TPl 1.4 16.8 c -4.2 c —2.3 c
Energy TP2 1.4 17.7 b -2.5 by -1.5 b
None 1.4 15.0 d -22.4 d -3.1 d
6 0.5 Blackout AL 0.6 19.3 a -0.3 a —0.6 a
Energy TPl 0.5 17.3 c -4.5 cy -0.9 b
Energy TP2 0.5 17.5 b -3.3 b -1.1 0
None 0.5 16.5 d —l4.8 d -1.6 d

 

zMeans within columns and experimental period followed by the same letter are not
significantly different by Tukey’s honestly significant difference test at P £0.01.

yL WRne, not measured; calculated using Eq. [3] with cover and leaf temperature and
screen and plant emissivity, but omitting the shape factor for LWRne,.

115

SWR LWR LWR
in In out Conv AE

 

Energy balance: SWR + LWR + Conv + AE = 0

Fig. 4.1. Schematic illustration of the components of a leaf energy balance which include
shortwave radiation (SWR), longwave radiation (LWR), convective heat transfer (Conv),
and evaporative heat loss through transpiration (XE).

 

 

 

Frequency

 

 

 

 

 

Temperature difference (°C)

Fig. 4.2. Frequency of the difference between simulated New Guinea impatiens shoot-tip
temperature and measured shoot-tip temperature for day and night using data from 24 (1
(13,824 observations). Plants were grown in glass-glazed greenhouses during winter in
East Lansing, MI (lat. 43 °N) under different vapor-pressure deficits and with different
screens extended over plants during the night (1700 to 0800 HR). Control plants were

grown without a screen above.

116

 

 

 

a, Day Period ° °, ,3 ’ _°. A
‘— 26 .. .......................................... . .....h’... ...... 9“,..11 ........... ...
3 9 o 0‘0 .
at r2-069 - ...,-.,.,.- .
h — ' '.oo {0.}; .0 3." O
a) 24 r ......................... .' ..................... o..”,.:....o2' 6."? """
2' '3‘. ’ £50”.

: . ‘00.
Q 22 0 .....flbzz'. )3“. ...“... ........

...... o
.9- , '. «so ‘ , .
:3 20 _ .O‘:.Qt..‘f5. ' 64.5; 0 g
o 0.. O ”0.? g.‘ ‘1
-c .0 ’0” 00.00 ”
w 18 ...o. .‘............°...° ‘2'? k ......... . .......................
8 °'- :1 ° °
‘3 16 - ............ 3: . . ...... ................... ........ ,. .. .......
8 °"$ ° 0
2 14 ...... . . .. ........ ...

12 I I I I I I I

 

 

 

12 14 16 18 20 22 24 26 28
Simulated shoot-tip temperature (°C)

 

A 22
9 Night Period 3
I; 21. ..... ......... .....9...°......,’..., ...... . ........ a
g 12:0.62 ...-. :: .-° ,-, . .
E 20 . ............ , ......... . .................. .3;. . . .‘.,. .............
8 . if.
E 19 .l.. ..:‘\ 0 $39... ..
9, 0:: ~ ....
.9 131......” ...!" g..:9‘..............-
:3. .0...‘ ... 0.. .0.
(D o z. “‘0'. . ...
U a ........ ’9.' ......3
m 16 ~.
L o
D .0 o
g 15 '1 .............. .. ...... a... .............. . .................................... ..
2’

14 I I I I I I I

 

 

 

14 15 16 17 18 19 20 21 22
Simulated shoot-tip temperature (°C)

Fig. 4.3. Validation of the energy balance model, comparing simulated New Guinea
impatiens shoot-tip temperature with measured shoot-tip temperature for the day [0800 to
1700 HR (A)] and night [1700 to 0800 (B)]. The validation included measured data from
24 d and consisted of 5,184 observations for the day and 8,640 observations for night.
Plants were grown in glass-glazed greenhouses during winter in East Lansing, MI (lat. 43
°N) under different vapor-pressure deficits and with different screens extended over
plants during the night. Control plants were grown without a screen above. r2 values
were generated by performing linear regression analysis on the simulated versus
predicted data.

117

 

600

r I I I I T I r T

VPD = 1.6 kPa Blackout AL A

    
 

4 -1 .
Measured shoot—tip

A -------- Simulated shoot-tip ,_

 

  

-—— Measured SWR ” 40°

300

 

 

-- 200

- 100

- 500

1 400

300

 

4- 200

 

1100

f l 5’ \ 5 k: l l l \ f l
1.6 kPa Shade AUTPIX E 0.6 kPa Shade ALITPIX F

1* 500

 

SWR (W-m'z)

 

2 . .. {k A .. .. 400

Shoot-tip minus dry-bulb temperature (°C)

 

 

 

 

 

 

O O O O O O O O O O O O O O o O o O O O O O O O O O O O O

O O O D O O o O O O O O O O O O O O O O O O O O O O O O O

N 6') ID ‘- r\ n I!) v- N n in ‘- h 0') t0 ('0 to ‘- Is 0") In 1- N 0‘) In ‘- N (V) In

t- N O 1- v- N O r- v- N O ‘- ‘- N O N 0 v- v- N O 1- v- N O ‘- 1- N 0
Time (HR)

Fig. 4.4. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East Lansing, MI
(lat. 43 °N) under different air vapor-pressure deficits (VPD) and with different screens
extended over plants from] 700 to 0800 HR. Control plants were grown without a screen
above. The air temperature set point was 20 °C. See Table 4.1 for a description of screen
materials. Data in panels A, C, E, and G and panels B, D, F, and H were collected during
20 to 24 Jan. and 25 to 29 Jan., respectively. Black bars represent night from actual
sunset to sunrise. SWR= shortwave radiation (300 to 3,000 nm).

118

I I I I fI # I I I I I 600

 

 

  

VPD = 1.4 kPa Blackout AL A 0.6 kPa Blackout AL 3
4 4 —— Measured shoot-tip .. .. 500
-------- Simulated shoot-tip
2 1 - - — Measured SWR 4' . . r'._ ,9, v 400
o .. 37‘ A ‘ €114 ’13:. 300
_, : L... ..... undying/Mk...
J‘ I | Ii [‘1‘
1 I \ I"! ’ 1
| l \ I 1 I 1
l I l I 1 l
1.5 kPa Blackout BL C 0.4 kPa Blackout BL D
4 1 <- 1~ 500

 

   

4 1.4 kPa Energy ALITP E 0.5 kPa Energy ALITP F 500

SWR (W-m'z)

«- 400

300

 

" 4 200

Shoot-tip minus dry-bulb temperature (°C)

1* 100

w 500

~ 400

300

 

~~ 200

<- 100

 

 

 

 

Fig. 4.5. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East Lansing, MI
(lat. 43 °N) under different air vapor-pressure deficits (VPD) and with different screens
extended over plants from 1700 to 0800 HR. Control plants were grown without a screen
above. The air temperature set point was 20 °C. See Table 4.1 for a description of screen
materials. Data in panels A, C, E, and G and panels B, D, F, and H were collected during
29 Jan. to 2 Feb. and 2 to 6 Feb., respectively. Black bars represent night from actual
sunset to sunrise. SWR= shortwave radiation (300 to 3,000 nm).

119

 

11111111600

 

 

     
   

 

 

 

 

I I T I“ I I r z I I I I I I I I I I I I I
VPD = 1.4 kPa Blackout AL A 0.6 kPa Blackout AL B
4 . - 333 4 Measured shoot-tip I 500
‘ g -------- Simulated shoot-tip
2 4 —-—- Measured SWR ‘ .. 400
‘° 3‘1 M fl
o '3 5 r "1. 300
; . l. \»‘_,V_~£ 1¥~HWM#HI ¥-~‘.~
. .. . .......,,..,m: ... ......fl
:1' I r 200
l 1‘
. K i A ..
-4 l/ , 1 A J (11‘ ,...,\ 100
l 1 /L\ I k l \ l \
0.5 kPa Energy TP‘l D
~ 0 500
9 , ., 1....
8 r . "' 200
E 11
£2 \'1 t .. ‘?
'° If, 1‘1 l \\\\ 100 E
5.3 ‘ 5
3‘ Energy TP2 05
3 F 1' 500 E
3 U)
._c_ 400
E
.9-
:5 300
o
.8 '~ 200
(D

n 100

~ 500

1 400

 

300

- 200

-- 100

 

 

 

 

2300
0500
1100
1700
2300
0500
1100
1700
2300
0500

Time (HR)
Fig. 4.6. Measured and simulated difference between New Guinea impatiens shoot-tip
and dry-bulb temperature in glass-glazed greenhouses during winter in East Lansing, MI
(lat. 43 °N) under different air vapor-pressure deficits (VPD) and with different screens
extended over plants from 1700 to 0800 HR. Control plants were grown without a screen
above. The air temperature set point was 20 °C. See Table 4.1 for a description of screen
materials. Data in panels A, C, E, and G and panels B, D, F, and H were collected during
15 to 19 Feb. and 11 to 15 Feb., respectively. Black bars represent night from actual
sunset to sunrise. SWR= shortwave radiation (300 to 3,000 nm).

120

Literature Cited

Adams, S.R., P. Hadley, and S. Pearson. 1998. The effects of temperature, photoperiod,
and photosynthetic photon flux on the time to flowering of Petunia ‘Express
Blush Pink’. J. Amer. Soc. Hort. Sci. 1232577-580.

Al-Faraj, A., GE. Meyer, and JB. Fitzgerald. 1994. Simulated water use and canopy
resistance of New Guinea Impatiens (Impatiens X hb.) in single pots using
infrared heating. Trans. ASAE 37:1973-1980.

Bailey, B.J. 1977. Thermal screens for reducing heat losses from glasshouses. Acta Hort.
70:26-34.

Bailey, B.J. 1981a. The evaluation of thermal screens in glasshouses on commercial
nurseries. Acta Hort. 115:663—670.

Bailey, B.J. 1981b. The reduction of thermal radiation in glasshouses by thermal screens. ‘
J. Agr. Eng. Res. 26:215-224.

Bailey, B.J. 1988. Improved control strategies for greenhouse thermal screens. Acta Hort.
230:485-492.

Bartok Jr., J. 2005. Grower 101: Horizontal air flow. Greenhouse Product News
15(9):26-28.

Brajeul, E., D. Lesourd, and D. Loda. 2005. Thermal screen evaluation in soilless
cucumber crop under glasshouse. Acta Hort. 691 :679-686.

Cohen, S. and M. Fuchs. 1999. Measuring and predicting radiometric properties of
reflective shade nets and thermal screens. J. Agr. Eng. Res. 73:245—255.

Dieleman, J .A. and F.L.K. Kempkes. 2006. Energy screens in tomato: Determining the
optimal opening strategy. Acta Hort. 718:599—606.

Erwin, J .E. and RD. Heins. 1990. Temperature effects on lily development rate and
morphology from the visible bud stage until anthesis. J. Amer. Soc. Hort. Sci.
1 1 5 2644-646.

Erwin, J .E. 1995. Light and temperature, p. 41-54. In: W. Banner and M. Klopmeyer
(eds), New Guinea impatiens: A Ball guide. Ball Publ., Batavia, IL.

Faust, J .E. and RD. Heins. 1993. Modeling leaf development of the African violet
(Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 118:747-751.

121

 

Faust, J.E. 1994. Modeling and managing shoot-tip temperatures in the greenhouse
environment. Ph.D. Dissertation, Dept. of Horticulture, Michigan State Univ.,
East Lansing, MI.

Faust, J.E. and RD. Heins. 1997. Quantifying the influence of high-pressure sodium
lighting on shoot-tip temperature. Acta Hort. 418:85—91.

Faust, J .E. and RD. Heins. 1998. Modeling shoot-tip temperature in the greenhouse
environment. J. Amer. Soc. Hort. Sci. 123:208-214.

Gates, D.M. 2003. Biophysical ecology. Dover Publ. Mineola, NY.

Hanan, J .J . 1990. The influence of greenhouses on internal climate with special reference
to Mediterranean regions. Acta Hort. 287:23-34.

Heins, RD., B. Lin, and ES. Runkle. 2000. Regulation of crop growth and development
based on environmental factors. Acta Hort. 514: 1 3-22.

Heins, RD. and E. Runkle. 2004. Materials and strategies for greenhouse shade, p.
39-42. In: P. Fisher and E. Runkle (eds), Lighting up profits: Understanding
greenhouse lighting. Meister Media Worldwide, Willoughby, OH.

Hoffmann, S. and D. Waaijenberg. 2002. Tropical and subtropical greenhouses — a
challenge for new plastic films. Acta Hort. 578:163-170.

Holman, JP, 1997. Heat transfer. 8th ed. McGraw Hill, New York, NY.

Jones, H.G. 1992. Plants and microclimate. 2nd ed. Cambridge Univ. Press, New York,
NY.

Kittas, C., N. Katsoulas, and A. Baille. 2003. Influence of an aluminized thermal screen

on greenhouse microclimate and canopy energy balance. Trans. ASAE
46: 1653-1663.

Komer, 0., M.J. Bakker, and E. Heuvelink. 2004. Daily temperature integration: a
Simulation study to quantify energy consumption. Biosystems Eng. 87:333—343.

Le Quillec, S., E. Brajeul, D. Lesourd, and D. Loda. 2005. Thermal screen evaluation in
soilless tomato crop under glasshouse. Acta Hort. 691 :709-716.

Ludvig Svensson, Inc. 2009. Cloth, firebreak, screens - Svensson screens. Ludvig
Svensson, Inc. <http://www.ludvigsvensson.com> (last accessed October 29,
2009)

122

 

Lund, J.B., A. Andreassen, C.-0. Ottosen, and J .M. Aaslyng. 2006. Effect of a dynamic
climate on energy consumption and production of Hibiscus rosa-sinensis L. in
greenhouses. HortScience 41 :384-3 88.

Mankin, KR. 1994. Modeling nutrient and water uptake responses to the environment by
New Guinea impatiens. Ph.D. Dissertation, Dept. of Agr. Eng., The Ohio State
Univ., Columbus, OH.

Mankin, K.R., R.P. Fynn, and T.H. Short. 1998. Water uptake and transpiration
characterization of New Guinea impatiens. Trans. ASAE 41 :219—226.

Meijer, J. 1980. Reduction of heat losses from greenhouses by means of internal blinds
with low thermal emissivity. J. Agr. Eng. Res. 25:381—390.

Miguel, A.F., N.J. van de Braak, and GPA. Bot. 1997. Analysis of the airflow
characteristics of greenhouse screening materials. J. Agr. Eng. Res. 67:105—112.

Moccaldi, LA. and ES. Runkle. 2007. Modeling the effects of temperature and
photosynthetic daily light integral on growth and flowering of Salvia splendens
and T agetes patula. J. Amer. Soc. Hort. Sci. 132283-288.

Monteith, J .L. and M.H. Unsworth. 2008. Principles of environmental physics. 3rd ed.
Academic Press, Burlington, MA.

Nijskens, J ., J. Deltour, S. Coutisse, and A. Nisen. 1984. Heat transfer through covering
materials of greenhouses. Agr. Forest Meteorol. 33: 193-214.

Niu, G., RD. Heins, A.C. Cameron, and W.H. Carlson. 2001. Day and night
temperatures, daily light integral, and C02 enrichment affect growth and flower
development of Campanula carpatica ‘Blue Clips’. Scientia Hort. 87:93-105.

Nobel, RS. 2005. Physiochemical and environmental plant physiology. 3rd ed. Academic
Press, Burlington, MA.

Ozttirk, H.H. and A. Bascetincelik. 1997. The nocturnal heat loss and internal
temperatures in plastic tunnel greenhouses. Acta Hort. 443:79—84.

Pang, T. 1992. Dynamic analysis of water and nutrient uptake for New Guinea impatiens.
Ph.D. Dissertation, Dept. of Agr. Eng, The Ohio State Univ., Columbus, OH.

Roberts, E.H. and R.J. Summerfield. 1987. Measurement and prediction of flowering in
annual crops, p. 17-50. In: Atherton, J.G. (ed.). Manipulation of flowering.
Butterworths, Kent, UK.

Rotz, CA. and RD. Heins. 1982. Evaluation of infrared heating in a Michigan
greenhouse. Trans. ASAE 25:402—407.

123

 

Runkle, E. 2008. Technically speaking: Overcoming New Guinea impatiens stall.
Greenhouse Product News 18(2):62.

Shimizu, H. and RD. Heins. 2000. Estimating cuticle resistance of seeding shoot tips
based on the Penman Monteith model, p. 59-62. In: C. Kubota and C. Chun
(eds.). Transplant production in the let century. Kluwer Academic Publ.,
Dordrecht, The Netherlands.

Shimizu, H., E.S. Runkle, and RD. Heins. 2004. A steady-state model for prediction of
poinsettia plant shoot-tip temperature. J. Amer. Soc. Hort. Sci. 129:303—312.

Silva, A.M., A. Miguel, and R. Rosa. 1991. Thermal radiation inside a single span
greenhouse with a thermal screen. J. Agr. Eng. Res. 49:285-298.

Snyder, K.A., J.H. Richards, and LA. Donovan. 2003. Night-time conductance in C3 and
C4 species: Do plants lose water at night? J. Expt. Bot. 54:861-865.

US. Department of Agriculture. 2009. Floriculture crops 2008 summary. Agr. Stat.
Board, Washington DC.

Wainwright, J. and M. Mulligan. 2005. Environmental modeling: Finding simplicity in
complexity. Wiley & Sons, England.

Whitman, C., D. Tschirhart, D. Joeright, and R. Heins. 2000. New Guinea impatiens: .
Flowers on time. Greenhouse Grower 18(10):48-60.

Winspear, K.W. and B.J. Bailey. 1978. Thermal screens for greenhouse energy
effectiveness. Acta Hort. 87:1 11—118.

Zabeltitz, CV. and J. Meyer. 1984. Evaluation of movable thermal screens in commercial
greenhouses. Acta Hort. 148:437-442.

124

 

 

 

 

APPENDIX A

MODELING PLANT GROWTH AND DEVELOPMENT OF 29 BEDDING PLANT
SPECIES AND CULTIVARS IN RESPONSE TO TEMPERATURE AND

PHOTOSYNTHETIC DAILY LIGHT INTEGRAL

 

125

Research Objective

The objectives of this study were to quantify and model the influence of mean
daily temperature (MDT) and photosynthetic daily light integral (DLI) on flowering

during the finish stage of bedding plants under long-day conditions.

Materials and Methods

 

During Dec. 2006, Apr. 2007, Sept. 2007, or Mar. 2008, seeds of 29 bedding plant
species and cultivars were sown in plug trays [288-ce11 size (6-ml volume)] by a
commercial greenhouse (C. Raker & Sons, Litchfield, MI). After germination, plugs
were received at Michigan State Universrty (MSU) and were grown in a controlled
environmental growth chamber at a constant temperature set point of 20 °C. A 16-h
photoperiod was provided by 215-W cool-white fluorescent (CWF; F96T12CWVHO;
Philips, Somerset, NJ) and 60-W incandescent lamps (INC; Philips), at a CWF:INC (by
W) of 3.6, and at an intensity of 180 umol-m“2-s"l at plant height. Plants were irrigated
as necessary with well water acidified with H2804 to a titratable alkalinity of 140 mg-L‘l
CaC03 and containing 95, 34, and 29 mg-L'l Ca, Mg, and S, respectively. The water
was supplemented with a water-soluble fertilizer providing (in mgL") 62 N, 6 P, 62 K,
7 Ca, 0.5 Fe, 0.3 Cu, Mn, and Zn, 0.1 B and Mo (MSU Well Water Special; GreenCare

Fertilizers, Inc., Kankakee, IL).

Greenhouse Environment

126

When seedlings were ready for transplant [16 to 45 d after seed sow, depending
on species (Table 5.1)], they were transplanted into 10-cm round plastic containers (480-
ml volume) filled with a commercial soilless peat-based medium (Suremix; Michigan
Grower Products, Inc., Galesburg, MI). At transplant, plugs were thinned to one seedling
per cell. Plants were randomly assigned to treatments and grown in glass-glazed
greenhouses at constant air temperature set points of 14, 17, 20, 23, or 26 °C and under a J.-
16-h photoperiod that consisted of natural photoperiods (43 °N lat.) with day-extension
lighting from 0600 to2200 HR provided by high-pressure sodium (HPS) lamps. At each ..
temperature, plants were grown under two DLI treatments provided by ambient light and
a combination of shade curtains (OLS 30, OLS 50; Ludvig Svensson Inc., Charlotte, NC)
and different intensities (25 to 150 pmol-m"‘2-s“') of supplemental lighting from HPS
lamps that were positioned above the shade curtains. Ten plants of each species were
randomly assigned to each temperature and DLI combination. The HPS lamps were
operated by an environmental computer (Priva Intégro 724; Priva, Vineland Station,
Ontario) and were turned on when the outside light intensity was <290 umol-m‘z-s‘l and
turned off at >5 80 umol~m_2°s_’. Whitewash was applied to the greenhouse glazing
during late Mar. each year, and removed in mid Oct. The experiment was performed
twice under mean DLIs from transplant to flowering that ranged from 3.2 to 19.6
mol-m‘z-d‘l (Table 5.2).

Temperature in each greenhouse compartment was controlled by an
environmental computer with steam heating, passive and active ventilation, and fan-and-

pad evaporative cooling as needed. Air temperature was independently measured in each

greenhouse by an aspirated, shielded thermocouple (0.13-mm type E; Omega

127

Engineering, Stamford, CT) positioned 1.5 m above the floor (at plant level). At 30 cm
above the bench, the photosynthetic photon flux (PPF) was measured by a line quantum
sensor containing 10 photodiodes (Apogee Instruments, Inc., Logan, UT) under six DLI
and temperature combinations. Environmental measurements were collected every 10 s
and hourly means were recorded by a data logger (CR10; Campbell Scientific, Logan,
UT). A vapor-pressure deficit of 1.2 kPa was maintained during the night by the
injection of steam into the air. Horizontal airflow fans positioned 1.4 m above the
growing surface operated if the ridge vent was <90% of the maximum opening and
provided air movement at 230.1 ms". Plants were irrigated with reverse osmosis water
supplemented with a water-soluble fertilizer providing (in mg-L"') 125 N, 12 P, 100 K,
65 Ca, 12 Mg, 1.0 Fe and Cu, 0.5 Mn and Zn, 0.3 B, and 0.1 Mo (MSU R0 Water

Special; GreenCare Fertilizers, Inc.).

Data Collection and Afllvsis

The date of first open flower per plant was recorded and time to first open flower
was calculated for each plant. The date of first open flower was recorded and time to
flower was calculated for each plant. Plants were considered in flower according to
individual flowering characteristics for each species (Table 5.1). When each plant
flowered, plant height and the total number of open flowers and closed flower buds were
recorded.

Flowering data were used to develop mathematical models to predict flower
development rate under different MDT and DLI conditions. Models were generated

using 155 to 200 observations (individual plants) for each species. Data were analyzed

128

 

 

 

using the calculated MDT and DLI for each plant from transplant to the date of
flowering. DLI values for treatments that did not have a line quantum sensor were
determined by calculating the mean irradiance among sensors that were positioned in
other temperature treatments and under similar light conditions. Flowering time data
were converted to developmental rates by calculating the reciprocal of days to flowering
(l/d to flower). For all species except Dahlia thbrida, the following multiplicative
model was developed to describe the relationship between the rate of progress to

flowering and MDT and DLI:

“(I to flower 3 fMDT X fDLI [3]

where f MDT and f DLI are temperature and light functions, respectively. Models of this
type have been previously used to describe the rate of flower development in
Chrysanthemum (Hidén and Larsen, 1994; Larsen and Persson, 1999) and cineraria
(Larsen, 1989). The response of flower development rate to MDT is described with a
temperature function and can be quantified by algebraically rewriting equation [1] to

include the base temperature:

0 ...ifMDTsT - 4
l/d to flower = mm I I

—] x Tmin X 01+ b] X MDT ...Imein < MDT ST01”

where Tmin and MDT are measured in °C and b; is a species-specific temperature

constant. We used a linear function to quantify the MDT response because plots of the

129

 

 

actual data showed that Topt was not observed for all species except Dahlia thbrida.
The influence of DLI on flower development rate was described with a light function

(Larsen, 1990):

DLI factor = 1 - EXP(—e X DLI) [5]

where the light factor ranges from 0 to 1. The e value is a species-specific light constant
and determines the skew of the curve and DLI is the mean (mol-m‘Z-d“‘) from transplant
to flowering. This function indicates that the rate of progress towards flowering increases
as DLI increases, until some saturating value.

The final model to predict the rate of development towards flowering consisted of

equations [4] and [5] multiplied together:
0 ...if MDT _<_ Tmin [6]
1/d to flower = (-1 X Tmin X b. + b. X MDT) X ...imein < MDT 5 Topt

(1 — EXP(—e x DLI))

A nonlinear model was used to describe the relationship between flowering rate

and MDT for Dahlia X hybrida (Landsberg, 1977; Reed et al., 1976):

1/d to flower = A x (MDT — Tmin) x (Tm,x — MDT)3 [7]

where A = Rmax / ((Topt - Tmin) X (Tmax — Topt)B) [8]

130

 

 

 

and B : (Tmax — Topt) / (Topt _ Tmin) [9]

where MDT = mean daily air temperature (°C), Tmin and Tmax are the minimum and

 

maximum temperatures, respectively. When MDT is ST min or ZTmax, development rate
is zero. TOpt is the temperature where the maximum development rate occurs (Rmax) and
the “B” value defines the skew of the function. This asymmetrical model describes a
temperature-dependent promotion of development when Tmin < MDT S Top, and a
temperature-dependent inhibition of development when Topt < MDT < Tmax (Larsen,
1990). This function has been used to model the influence of temperature on flowering
rate in Dahlia pinnata Cav. (Brondum and Heins, 1993) and leaf unfolding and leaf
expansion rate in Saintpaulia ionantha Wendl. (African violet) (Faust and Heins, 1993,
1994). The complete multiplicative model to describe the flowering rate of Dahlia
thbrida was generated by multiplying Eq. [5] with Eq. [7].

Estimates for model coefficients were determined with the nonlinear regression
procedure (N LIN) of SAS (version 9.1; SAS Institute, Cary, NC). Tmin for Ageratum
houstonianum ‘High Tide Blue’, Begonia semperflorens-cultorum ‘Sprint Blush’,
Catharanthus roseus ‘Viper Grape’, C leome hassleriana ‘Queen Mix’, Lobelia erinus
‘Riviera Midnight Blue’, Osteospermum ecklonis ‘Passion Mix’, Pelargonium
Xhortorum ‘F lorever Violet’, Petunia thbrida ‘Easy Wave Coral Reef’, Petunia
thbrida ‘Fantasy Blue’, Tagetes erecta ‘Moonstruck Orange’, and Verbena thbrida
‘Obsession Lilac’ was estimated by using the NLIN procedure of SAS. Estimates of

Tmin for all other species were obtained from Table 1.2.

131

Data for flower bud or inflorescence number at first flowering were analyzed
using the regression procedure (REG) of SAS to determine the influence of MDT and
DLI. The flower bud and inflorescence number response surfaces equations are in the

form:

 

y = yo + aMDT + mm? + cDLI + dDLI2 + gMDT x DLI [10] ....

where yo is the y-axis intercept and a, b, c, d, and g are species-specific constants.

Previous published studies that quantified the influence of MDT and DLI on flowering

 

used a similar polynomial equation (Moccaldi and Runkle, 2007; Pramuk and Runkle,

2005). The terms of the equation were only included if they were significant at P £0.05.

Model Validation

During fall 2008 and spring 2009, seeds of each species were sown in 288-cell
plug trays by a commercial greenhouse. After germination, trays were received at MSU
and grown in an environmental growth chamber. Environmental conditions inside the
chamber, plant culture, and transplant schedules were the same as described for the
previous experiments. Seedlings were transplanted into lO-cm round pots and 15 plants
of each species were grown in glass-glazed greenhouses at constant temperature set
points of 17, 20, or 23 °C and under a 16-h photoperiod and a mean DLI of 10 to 19
mol-m‘Z-d‘l. In each greenhouse, air temperature and PPF were measured on each
bench and data was recorded by a data logger as previously described. Photoperiod

control, plant culture, and data collection were the same a previously described. Data

132

collected from the validation study were used to test the accuracy and precision of model
predictions. Validation experiments were not performed with Begonia semperflorens-
cultorum ‘Sprint Blush’, Catharanthus roseus ‘Viper Grape’, Osteospermum ecklonis
‘Passion Mix’, Pelargonium Xhortorum ‘Florever Violet’, and T agetes erecta
‘Moonstruck Orange’ because plants were not available from the commercial plug
producer.

A separate validation experiment was performed with Petunia thbrida ‘Dreams
Neon Rose’ and Antirrhinum majus ‘Montego Burgundy Bicolor’ to validate the flower
development rate models and to determine the variation in flowering time among a
population of plants grown under the same environmental conditions. On 1 April and 28
March 2009, seedlings of Petunia and A. majus, respectively, grown in 288-cell plug
trays were transplanted into 10-cm round pots and 150 plants of each species were grown
in a glass-glazed greenhouse at a MDT of 21 °C and under a 16-h photoperiod and a
mean DLI of 21 mol-m"2-d"'. Plants of each species were grown on the same bench.

Air temperature and PPF were measured on each bench and data was recorded by
a data logger as previously described. Photoperiod control, plant culture, and data
collection were the same a previously described. Petunia was considered flowering when
individual plants had one flower with a fully open corolla. A. majus was considered

flowering when individual plants had 2 flowers open on an inflorescence.

133

Table 5.1. Time from seed sow to transplant (TP), mean node number at TP, and
characteristics used to determine flowering date of bedding plants used in modeling

 

 

experiments.
Time from Mean
seed sow to node no.
Species TP ((1) at TP Flowering characteristic
A geratum houstonianum ‘High Tide Blue’ 27 3.2 .2 flowers open on an
inflorescence
. . . , , 3 flowers open on an
Angelonia angustzfolza Serena Purple 40 5.8 inflorescence
Antirrhinum majus ‘Montego Burgundy’ 41 3.1 i2 flowers open on an
inflorescence
Begonia semperflorens-cultorum ‘Sprint 1 flower open on an
. 38 3.9 .
Blush inflorescence
Browallia speciosa ‘Bells Marine’ 40 5.9 1 flower open
Catharanthus roseus -‘Viper Grape’ 38 or 45 2.2 1 flower open
Cleome hassleriana ‘Queen Mix’ 28 4.3 .6 flowers open on an
inflorescence
, . , 1 inflorescence with 1
Cosmos sulphureus Cosmlc Orange 23 2.6 whorl of petals re flexed
. . , . . , 1 inflorescence with I
Dahlia thbrzda Figaro MIX 26 3.2 ‘whorl of petals re flexed
Dianthus chinensis ‘Super Parfait 38 5 3 1 inflorescence with 1
Raspberry’ ' whorl of petals reflexed
Gazania rigens ‘Daybreak Bronze’ 31 4.7 1 inflorescence “nth
petals reflexed
Lobelia erinus ‘Riviera Midnight Blue’ 45 3.1 1 flower open
-. , . . , 1 inflorescence with 1
Osteospermum ecklonis Passwn Mix 28 4.6 whorl of petals re flexed
Pelargonium Xhortorum ‘Florever Violet’ 28 3.8 .5 flowers open on an
inflorescence
Pentas Ianceolata ‘Graffiti Lavender’ 40 3.9 .8 flowers open on an
inflorescence
Petunia thbrida ‘Dreams Neon Rose’ 31 6.2 1 flower wrth fully open
corolla
Petunia thbrida ‘Easy Wave Coral Reef 27 6.7 $013311? wrth fully open
Petunia thbrida ‘Fantasy Blue’ 34 l 1.2 1 flower wuh fully open
corolla
Petunia thbrida ‘Wave Purple’ 33 7.2 1 flower wrth fully open

corolla

 

134

Table 5.1 (cont’d).

 

Portulaca grandiflora ‘Margarita Apricot’

Rudbeckia hirta ‘Becky Cinnamon
Bicolor’

Salviafarinacea ‘Blue Bedder’

Tagetes erecta ‘Antigua Primrose’
Tagetes erecta ‘Moonstruck Orange’
Tagetes patula ‘Janie Flame’

Verbena thbrida ‘Obsession Lilac’
Verbena thbrida ‘Quartz Waterfall Mix’
Viola cornuta ‘Sorbet Plum Velvet’

Zinnia elegans ‘Dreamland Coral’

44
31
29
19 or 23
25
19 or 23
28
32
38

16’

18.8

5.5

4.3

6.0

6.0

6.4

4.5

3.6

6.5

3.5

1 flower open

1 inflorescence with 1
whorl of petals reflexed
3 flowers open on an
inflorescence

1 inflorescence with
250% of petals reflexed
1 inflorescence with
250% of petals reflexed
1 inflorescence with
250% of petals reflexed
6 flowers open on an
inflorescence

6 flowers open on an
inflorescence

l flower open

1 inflorescence with 1
whorl of petals reflexed

 

135

 

 

 

 

 

 

 

we we NeN NNN MEN 2: v.2 N .82
N2 3. SN NNN 2N N5 3: _ cease 8.58%.. $583.85
1 1 1 1 1 1 1 N .85
a: S. 1 wNN of at 2: _ 53:22 825. 858 £33
0 0— o o ~.©N m mN — ~N Q 0~ I N s0Nco._m MNOLDEANQ. 2.5%.: 3.2.5590
_.N~ NH N N oNN ea N: 1 _
Na: o8 1 NNN EN 3: 4.: N Boise
4.2 3. 1 ”N 8.2 N.: n: _ ease 8%. 38:8 .3385
q: S NeN NNN SN 3: _.2 N
2: 3. 3N NNN 2N N: S; _ .52 8%? 83.3.. 83.5
0.: am 1 EN EN 3: ”.8 N .8580
3 NM 1 NNN 8.2 Q: a: _ 2880. $3.35. 8.5.8
N.: Nm SN N.NN EN E: NE N .52 52.0. Egress: 2.883
I: 3. SN NNN NON _: f: _ .
2: 3. 3N NNN 3N qt 1 N .
w.: we 5mm w.mm 0.3 N: 1 _ .0920 53>. .333 Sfizgefieb
8.: 4M NmN CNN c. _N f: ..S N .8582 23m. 8888.. sausage
2: 3. SN wNN SN .2 m: _ .
_.2 Ne NeN NNN ZN N2 1 N .55 Exam.
v.2 NV 9mm w.~m $2 N: I _ szeczoétoxermQENH firewom
3: o8 1 ON 98 3: .5 N .35?m
2: 3. 1 NNN me Q: N: z 83:52. .33.. 2:23:22...
2: 1 1 EN v.2 N: 1 N .295
S; ....m 3N NNN S: _.: 1 _ 228m. 8.54.39... ensues.
2: ca 1 EN EN 3: a: N .85
Z. 2. t NNN NE at 3; _ DE 35. 33522283 seesaw
:9: 33 N N N t E .axm seam
NATpNeaev :o CL 52

 

.mEoEtomxo waist mEoEtmmEoo omzoccooew wo§_w-mm£w E £20on some mo wctoBo: 8 Hem—mamas: Eot .monocon
o>onm QAQV _SwBE Em: EEC 38558859 508 wan Eamon 8m 3639: 05 8 Ch"): 238388 he 36c :82 .N.m 033.

136

 

Lu,

 

 

 

 

52 3 SN NNN SN 2: 92 N .52
v.2 3. 3N wNN 3N N.: S: . :52“? 53:0. 33.3.. 3.35
Na m6 Qmm _.MN fiom o.w_ 1. N of coamom a 2 A 35 .5
Na 3 EN EN 52 at 1 _ . S . so. .... a 5.. a N.
o 2 o 0 CN o _N N 2 _.: _.2 N beam 25:. sagging
2: 3 a NN EN 02 v 2 Z; _ .
N2 3 1 3N :N 52 Q: N

mm m 2: Eco a .5 m a... awn
0.: 3. 1 NNN we N.: ...: _ . c 5.. H 2. . .. N .N
h.0~ 0.0 ~.0N 0 NN 0 ON M.w~ m.0~ N .OmOuE—um NSWSGAQ. grumxm WENMBK
Z: NV 3N NNN 3N N: f; _ . .
q: 3 3N EN EN 2: 1 N

.5 0 03 3996223 33 Q
a: 3. 3N NNN 0.2 m: 1 _ . 2. 5 5. . ..N . Z
«.2 3 1 SN N._N we a: N .585
N.N_ S. 1 w.NN N5 5: ”.2 _ Seesaw 385. SE 33.35
5.2 3 SN 3N N.~N 0.2 1 N 5855
n2 3 EN 3N N2 ...: 1 _ 95.53:. 30338.... 83:25
«.8 3 SN QNN EN 3: 52 N
o E: o m a A
2: 3. 3N mNN 2N N.: S: _ . _ 5 > 3. 33 s. 3.355
S: S 1 EN EN 2: at N 585
E: 3 1 NNN 0.2 9: Ni _ _28 953 Nap 3333.. 352.5
5: 3 SN nNN SN 0.: 02 N
N .N

5.2 2. SN QNN 98 v: 3: _ .85 $.55. 53% S. 5:55
m: 3 EN 98 N5: 92 62 N .085
2: 2. SN EN 2: NS 3; _ :82 2.85. 33:53.. 35:5
3: 3 N.©N NNN 3N oi Q: N .5223
2: S. 3N NNN SN Q: 1 _ 5520. 35833 835
2: .3 SN 3N N._N we 1 N 5035
5.: S. 3N NNN 52 n: 1 _ 555:. 5593? 53:05.55
55 33 5N NN ON 2 E .95 8.8%
~A_1c.N1e._o.5 :5 CL 52

 

6.255 N.m 2.5

137

 

 

 

.Uowuz—odm HOG “coaummb OSHNHQQEUFA

._Im.~IE._o:3 owmA 5:3 to BEE was _Im.mIE._o:3 oomv $3 55 5323 ouoficxmouofi
023:0 05 5:3 co BEE 203 :2: mafia— :568 05329.32 80¢ Em: _Scofiofimsm 28 3253993 333: Set EwEN

 

 

 

9: mm 0.2 EN SN 2: I m

So an Em»: 38%» m 325
3 2 EN, EN a2 9: I _ ._ 0c _ o. N . .N
a: I I is 0.2 w: I N

o 0 Ed 0 a: 96 ~
3: 3 I mam we q: m: _ 3,; 328m. 2% 3:
E: 33 8 MN om : E .axm 83am
NA_Ic.~Ie._oev So Gov E:

 

.5253 mm 2.5

138

 

 

 

 

 

Ed SIm 8N _oIm 9mm moIm o3 SIm m2 2 .52 .3333 $30. $33;
who NoIm v? SIm w; onm o3 moIm SN 2 2.55 83830. anti? Sass;
as _oIm a: :Tm 2% gIm o3 SIm N2 2 .08“; Des. £53 33$
23 SIm o3 SIm m2 voIm 2: BIm _S 3 ,owceo €23,802. 38% as?“
Ed SIm m2 sIm o3 moIm 8.,“ SIm NE E .3255 332%. 225 33$:
:3 SIm 23 _oIm N3 2; OS SIm SN v.0 .85 $22? §§§§E£
one _oIm o3 :Tm m: moIm 86 MAI m2 3 .8305 882:6 38m. SE sfimfié
N2 o sIm 32 mOIm 8.0 SIm SN 3 .82.? «Ems? Eo§g§m 8333;
So SIm o3 _oIm mum moIm Sm SIM. m3 3 .283 233. anti? s§§
NS SIm o3 _oIm own «cum o2 SIm 3m 2 max .88 253 5m 533$? 325$
93 NoIm 3m SIm a: voIm 8a SIm mom in .35 has? a??? 352$
23 SIm 3m SIm 53 moIm o3 SIm com 3 Box :82 £35. asséx 3:35
33 NoIm 2% :Tm 2+ mcIm oi SIm w? mg .5223 was? 333823 35$
23 SIm _3 sIm o3 moIm 8.x SIm _2 oh .563 558E. 253:8? ssgfisfi
So SIm N2 SIm m2 SIm cow moIm 2: E .52 c235, £338 5353826
So SIm mam sIm New voIm o3 SIm Fm ca .85 22:22 225. 25.5 £33
m3 SIm EN sIm v2 moIm 8.0 SIm a: 3 @2on Vegas. Emm: 3:28
86 _oIm :2 SIm Ev moIm 3m SIm m2 3 .Eofimé E5 395. Eésfi £5235
93 :Tm NS sIm o2 moIm o3 moIm SN 2 .cheo 2580. §§~Ea 85.89
:3 SIm a: SIm of voIm of SIm m3 2 ,5: 825. magmas: 2:85
86 o SIm So 2; OS moIm 8a v. : .aeo 59>. 252: £3§Ss§
:3 SIm 3m SIm 32 ml 8.0 SIM. a: 3 .0552 23m. 3,38% 593%
Ed SIm m3 sIm :é voIm SN 25 SN S :35 2am. siéfiéségsa Egg
3o SIm 3m SIm Ra 3m 8.0 moIm NZ 3 .xncstm 0352. $.35 §=E§§
3o SIm v2 SIm o5 moIm o3 SIm _2 3V beam 220m, §§§m§ smfimzw
woo _oIm m2 _oIm mam voIm o3 SIm ”2 3 ,BE 2i swE. §§§§§ 555»?
N1 5 $8 m m 5 x3 5 5 Gov 2:; 83am

 

.3:ch oucowccoo H5 .ACAQ

x arts I c x ABE x 5 + s x sf x TV I 33oz 9 u: we gee 2: 5 em 2252 .2_I§-e._oé 3a 335 £3 S8
was 20% HOS: 338388 hm xzmw :36 9 28 mafia/om meta—2 GE .3: 388 Sufi—co: 8m 388:3 $68me .m.m 2an

139

 

 

.83 cotomno max; 366on 05 co £9298 2282on “3:: wE—Eoctoq 3 cowfiocoON

 

 

Sum m3 Sim :3 m

Elm 8.0 Sum RN 52.x

Sm o. _ M 52:9

Sim o3 mam ES

92 Sum 86 om 55%
me Aumv _U 33 0883mm 85:38;

 

.EfiBE oocoumcoo "_U .R_nn.m-E._oEV 5n: _EmBE Em: 32% was :UoV PQS: 838352 :«N=% :88
3 ,52 Emmi. 333A? EEQQ E 38 mctoioc wean—8 :5 .cm x .3 .53 388 50:25: 2: 8m 38853 3898qu 4mm 2an

.83 33030 m2m5> 386th 2: :o 29:98 568th 80:: wEEhofioq b... 3:30:99

 

 

o3 Sum N3 sum 3m Bum 8m ml :1 M: .58 22585. E933 SEN
Po 81m x: 51m 3: 31m o2 Sum 3m 3 ,§_u> :5: ES. 5528 so:
N1 5 :8 m m _o :3 s 5 6L £5 860%

 

.€.E§ 3 29¢

140

 

 

 

.852 8&8.

 

 

cud Nolm modl I MON SIm SM. SST 963 A5 monsoomfiocfi 223%» 83:238an

.85 $222

I I I I I I I .. A5 €026E 803E. $3.2m 3:234

3:9 8 as .

mud I molm No; I NOIm mod EST 2.? EV moosoomohocfi . mzmMmSmQNMv

. . . .xhunammm :mfimm 89m.

one I SIm :.T of“ SIm ow N s m T Ma :2 E 888882 E§sfi £§§Q
5

23 I I _oIm N3 SIm 8.? I :2 E 82.8382 2%; anééx smfiw

. . .owSEO 05500.

«We I I xv; SIM S.mI mo 2 mm owI A5 moocoomouomfi wxméfiézm SEED

.52 :88.

I I I I I I I Q: mooaoomflomfi ESENEE 3:83

. . ,380 SE?

Ed I _olm we. _I on m molm vofi I ow VI A5 3032.,— agwox “SSEEEEU

. . . . bats—2

$6 I _oIm om mI am e molm vm mI I mo 2 EV m8>>oE mzom 388me aficicxm

. . . . . Ems—m Bram. 533:5

on o moIm S N Sn» 3 T I SIm $ oI I E o 3 32.8882 -EEQEREa 3%QO

. . . . . . ,xugwhsm 03:82.

3 o SIm E ”I I 3 _ SIm 3 m S :I 8 09 E 82.88.55 gas 535.13%

. . . . . bans; «seam.

S o I melm _o v I _olm no NI cm x on mnI EV muocuomeocfi ccgazmnu Srfimwzw
on o _ w_

26 I I I NI _olm m5? afiwm A5 moocuomohocfi Ezzaiwfihwx ”wwuxmww

mg M h a c a o» $6883 86on

.Eo x 52w

+ N393 + 5% + NEE + 533 + i I % “I8 :28 05 5 ea 20on =< 3-3-5.33 :a Bugs 53 3% 2a 2ro Es:
838388 EN 3:8 :38 9 838:: cocoomaocfi 8 95 526$ mats—2 airman .838th $33on mo whoaofifimm .m.m 2an

141

 

 

.52 £333 538.

 

one I _oIm :T 8N SIm NN._ NV? 8% E 888353 333x E35.

. . . . .02: 85330.

mg I I SIm a N SIm E NI I M: S E 8058885 a??? 33$

. . . be“:

NE NoIm 3w NI I :Tm NW N I I E c E 888883 28: SEE §§£
I . I .owSE 2538

N2 SK» 0qu I SN 8 m NE EN VSVNI E 388353 mam»; 83mm“

. .08255

E I I v: SIm 3T 3+ NET E 88082an mamas/N 82S aswfi

. . . . . .5an

coo NoIm we m I SIm 8 NI ... SIm 8 oI : 2 E 888853 35 Suggxaéam

. . . .. . . L285 :25:ch

mm o I NoIm Q NI 2 N :Tm N_ _ mm mI 3 No E 888882 38m SE 3.93332

. . . EOE? magma?

2.0 I I :Tm N2 NOIm 3 o a NI mm mm E 3888.85 Eo§§m 833:8

. . . . . .285

Nm 0 I _oIm No NI No a NoIm NN mI I o. _N E 235a 263 53:3? 326$
00 who 03w

$6 I I N: :Tm 3; SENI oNNmN E 5%: >QO MEN»? wseww

. . . . mam 52$.

3 o NOIm N T I Bk» 3. a I I 3 2 E Ego; a??? 353m

. . . . . .omom Gooz madam?

3 o NOIm % N I I SIm : T $ N NN NT E 336: €23? 35%:

Lo :96 E 8 .

woo NoIm N3 NoIm 0: 8T NoIm QNI I 3. _N E 828$:an swabs 9%»sz

. . . . .335 $583.

0N o I I SIm mo _ NoIm «N NI I on m E 8888.85 5528? £55335

mg m h u a a 0% Sagan museum

 

.9253 nm 23

142

 

.modA & Hm “52.2:me Ho: BBEEMQN

 

. L28 wages.

 

wed moIm m: I NoIm cod! :4 2: _I Ev muocoumuhozfi .28me BEEN
. I . I . . .628, :35
E o moIm mo _ _olm oo VI 5 .2 EV 9032» 698m. 3528 30.3
mg M B a a o» .838ng $625

 

ESE 3 29¢

143

Table 5.6. Validation of the flowering rate models, comparing the observed time to

flower with those predicted. Coefficients for the models are presented in Tables 5.3.and

5.4. r2 values and the slope and intercept were determined by performing linear

regression analysis on the predicted versus observed flowering rates.

Species 3:5 d (%) i7 d (%) r2 Slope= 1 Intercept: 0 NZ

A geratum houstonianum * *

‘High Tide Blue’ 84 98 0'60 49

A ngelonia angustifolia
‘Serena Purple’

Antirrhinum majus
‘Montego Burgundy’

Begonia semperflorens-

cultorum ‘Sprint Blush’

Browallia speciosa ‘Bells

Marine’

C atharanthus roseus

‘Viper Grape’

‘Cleome hasslerzana 73 96 O. 68 * *
Queen MIX

Cosmos sulphureus

‘Cosmic Orange’

Dahlia thbrida ‘Figaro

Mix’

Dianthus chinensis ‘Super

Parfait Raspberry’

Gazania rigens ‘Daybreak

Bronze’

Labelia erinus ‘Riviera

Midnight Blue’

Osteospermum ecklonis
‘Passion Mixed’

Pelargonium Xhortorum

‘F lorever Violet’

Pentas Ianceolata Graffiti 16 23 0.94 NS *

Lavender

Petunia thbrida ‘Dreams

Neon Rose’

Petunia thbrida ‘Fantasy

Blue’

Petunia thbrida rEasy

Wave Coral Reef’

Petunia thbrida ‘Wave

Purple’

Portulaca grandiflora

‘Margarita Apricot’

 

 

78 91 0.83 * * 45

86 93 0.84 * * 57

_y _ _ _ _ _

31 47 0.92 "‘ * 45

 

45

93 96 0.79 * * 45

29 49 0.26 * * 45

82 87 0.96 * NS 45

51 67 0.88 NS NS 45

56 76 0.67 NS NS 45

 

44

89 100 0.67 NS NS 45

84 98 0.32 * "‘ 45

67 96 0.60 * * 45

98 100 0.93 NS * 45

49 60 0.88 NS NS 45

 

144

Table 5.6 (cont’d).
Species 35 d (%) i7 d (%) r2 SIOpe= 1 Intercept: 0 NZ
Rudbeckia hirta Becky 57 76 0.51 * * 45
Cinnamon Bicolor

Salviafarinacea ‘Blue

 

 

 

*
Bedder’ 71 73 0.86 NS 45
Tagetes erecta Antigua 40 69 0.80 ... ... 45
Primrose
Tagetes erecta _ _ _ _ _ _
Moonstruck Orange’
Tagetes patula Jame 91 100 0.60 NS ... 45
Flame
Verbena thbrida~ * ...
‘Obsession Lilac’ 22 31 0'84 45
Verbena thbrida ‘Quartz * *
Waterfall Mix’ 57 66 0'67 44
Viola cornuta Sorbet 96 100 0. 84 NS NS 45
Plum Velvet
Zinnia elegans Dreamland 100 100 0.94 ... ... 45

Coral’

ZNumber of observations in data set.
YValidation experiment not performed.

”3’ *Nonsignificant or significant at P £0.05.

 

 

145

 

     

+ Petunia 'Dreams Neon Rose'
100 —0— Antirrhinum 'Montego Burgundy Bicolor' >
A I
2x: I
a)
9. I
c I
a)
0)
°- I
c»
E I
<3 I I
20-.. .. . _ .. | ..... .....“lfl
I I
| I
O l I T l r l ! l l ! I l l

 

 

 

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Time to flower from transplant (d)

Fig. 5.1. Time to flower from transplant of 150 plants each of Petunia thbrida ‘Dreams
Neon Rose’ and A ntirrhinum majus ‘Montego Burgundy Bicolor’ grown under the same
environmental conditions. Plants were grown in a glass greenhouse at a mean daily
temperature of 21 °C and under a mean daily light integral of 21 mol~m“?--d‘l with a 16-h
photoperiod. Petunia was considered flowering when individual plants had one flower
with a fully open corolla. Antirrhinum majus was considered flowering when individual
plants had 2 flowers open on an inflorescence. Dashed vertical lines indicate the

predicted time to flower for each species using the flower development rate models
(Table 5.3).

146

1.2 - . . . . . j i .
Ageratum houstonianum (A) Angelonia _angustifo/a (B) Antirrhinum majus. (C)
1.0 0' I I f .' '. I
06 - a-8__mo.l:mf?:df1 .. etaimoI-mT’d" ...... ; ,. ....18-2m9ltrn'. rd'.‘. ..
04 f t % t . = t 5 ?
Begonia semperflorens (D) Browallia speciosa (E) Catharanthus roseus (F)
10.. .M‘ .............. i ' 1 3 .H.M2
._ o 6 . ,. . 10:5. moI-m’z-di’ - 48 .m.o':.m72'd"_.
2 f ' I
o
g 0 4 ': r ‘4 i i '5
.5, Cleome hasslen'ana (G) , Dahlia xhybn'da : (I)
5 i I Z :

 

 

  

 
 

 

    
 

 

 

  

 

 

    
   

 

 

   

 

 

 

  

 

 

 

  

 

   
 

 

 

10 .. ...................................
08 4 ..................... DL|sat= __‘
0 5 - ......... . ?20. m9|>m‘?:d“ .. 1.6-.4. met-mils".
0.4 5 5 £ 5 fiL ; 5 + %.
Dianthuschinensis (J) ; Gazania rigens 1 (K) Osteospermum ecklonis (L)
10 4 I i ‘ I ............... i ..................
o 6 - .9,-8m9'm'1-d." 319.5 m9!-.m"rd'.‘. .....1.8:.3.m9|:m'?rdi’ .
0.4 . , . . - - ; .
5 10 15 0 5 10 15 0 5 10 15 20

Daily light integral (moI-m'z-d'1)

Fig. 5.2. The increase in flower development rate as the daily light integral (DLI)
increases in 12 species of bedding plants grown under a 16-h photoperiod. Circles
represent the estimated saturation DLI (DLlsat; light factor 20.99) for the shortest time to
flower for each species. Symbols are not presented if DLIsat is predicted to occur above
of the DLI range for which the models were developed.

147

 

 

 

 

1 .2 ‘7— i f ,f ' ' t ' T
Pelargonium Xhortorum (A) Pentas lanceolata (B) Petunia thbn'da (C)
: 5 ; ; 'Dreams Neon Rose'

 

       
 
  

 

     

     

 

1O ... ..... ‘
|
DLIsat = . DLIsat =
6 6 . 1.7-9 moI-m’Z-d" . . , ...... "-1. m9l-m":d" T ......... f ...... 19:6m9I-m'2'd’.’ ., .
0.4 ‘f ‘% ‘fi 7L 5 :— i— c 6
Petunia thbn'd (D) Petunia thbn'da; (E) Portulaca grandiflora (F)
'Easy Wave Coral Reef Wave Purple' 3 : I -_
1.0 6 .. .. ........... .. ......... :‘ ............. ......
. l. . :. ..... ... ,. .l . ..
0 8 DL'sat ‘ DLIsat = r DL'sat =
14.4 moI-m'Z-d'1 14.1 mol-m'z-d'1 Q 12.0 moI-m'z-d'1
b 0.6 it , ,. . .I . , ., .. .. ll . ............ . ....... .
g ;
0 4 a v e é % fit ‘4; l i
E Rudbeckia hirta : (G) _‘ Salvia farinacea ; (H) Tagetes erecta (I)
9 ll » ' ‘ 1 'Antigua Primrose'
A 1.0 y... . ..... _A i

 

 

  

       

 

1.31.1536: ,. rm...” ”DI-.lg'atfi I. . ..

 

 

 

     
 

     

 

 

 

 

 

' ‘ou66t=” "
66 . 10-5moI-m'Z-d"... . 15-8.m°'-m'?-d".., l ,. _ ...... 8:3‘moltmTz-d" .......
0.4 i i = . i . 1 6 _ _ i i

Tagetes patula ' (J) Verbena thbnda (K) 3 Zinnia elegans ' (L)

' ‘Quartz Waterfall Mix' f : ;
DLIsat ‘ DLIsat ‘ DLIsat "
66 , 8-6 moI-m'Z-d“ I ....19-5m9'm‘23d71..- .12:5m°|:m"6d." ..
0.4 . . . r . . ; .
0 5 10 15 0 5 10 15 O 5 10 15 20

Daily light integral (moI-m'z-d'1)

Fig. 5.3. The increase in flower development rate as the daily light integral (DLI)
increases in 12 species of bedding plants grown under a 16-h photoperiod. Circles
represent the estimated saturation DLI (DLlsat; light factor 20.99) for the shortest time to
flower for each species.

148

 

25 . 4 . . . . . . .
Ageratum houstonianum 1AM Angelonia angustifolia (B) Antirrhinum majus (C)

..........................................................

 

15 4

   

1o «»

-2 d.1

DL'max >19 mol- rn'2 d'1

5 ., ., 1 ...... _ ........................................ ..L.

DLI = Nonsignificant DLImax >16 mol m

 

0 . 1 6 . : 1 1
Begonia semperflorens 1D). Browallia Speciosa . 1E.) ......Cathamnthys means- _..1F1

1 -)
-3 N N 0)
01 O 0' O

[)leax - 14 9 Incl m'2 -d 1

_s
o
i

 

 

v

 

4
v

Cleome hasslen'ana 1G)

(a)
O

 

Flower bud or inflorescence no

 

 

 

   
   

 

 

 

 

 

DLI = Nonsignificant DL'max >18 mol'm'z'd'1
1o .............................................. ... . , -2. -1
DL'max >19 mol in d
0 fi‘ 1 + if 6% 4,
Dianthus chinensis (J) Gazania n'gens 1K) Osteospermum ecklonis (L)
25 4, . ........................ . .L. .. ... ., .. .. ..... .I. ......................
20 .. . ......... . . ...................................... 4
15 Dleax Dleax >20 mol- m 2 d 1 Pleax >19 mol- m 2 d'1
15 5 mol in 2 d 1
1o .. 4..
5 ..
o . . 5 . - . . .
O 5 1O 15 0 5 10 15 0 5 10 15 20

Daily light integral (mol-m'z-d'1)

Fig. 5.4. The increase in flower bud or inflorescence number as the daily light integral
(DLI) increases in 12 species of bedding plants grown at a constant temperature set point
of 20 °C and under a 16-h photoperiod. Predictions were determined using the
polynomial response surface equations presented in Table 5.5. Circles represent the
estimated maximum DLI (DLImax) for the greatest flower bud or inflorescence number
for each species. Circles are not presented if DLImax is predicted to occur above of the
DLI range for which the models were developed.

149

 

30 . . . a . . . - a

 
  
     

 

 

   
  
 

 

 

 

 

    

   

 

    

 

 

 

 

 

Pelargonium Xhortorum (A) Pentas lanceolata (B) Petunia thbn'da (C)
20 4.. . ..L. ,. .. . .. ..... . ., . .4 ........................
15 ..... DL1max ’20 11101111 2 d .1 ............ DUmax >15. ”10'? m'2 ‘1 1 .................. ..
puma,< >I20 moI m 2 d 1
1o .. ........
5 .L. ____ .I .........................................................................
0 t t I 1 . ¢ . + i 1 3
Petunia thbrida (D) 26.33"” :hyb'fqa (E) Portulaca grandiflora (F)
60 'Easy Wave Coral Reef' ave u
A .. ............................................ . .. ........................ . ...................... 1
2 DLImax >19 mol-m'Z-d'1_2_.' 2.22122. >20 mol m 2 d 1
"' __ _ 14. 7 mol In d
o 40 4» .. . ..... .
o
c
3 I
m 20 2 1
o
L
O
a 0 .L 1 1 I I 1 I 1 ¢
.5 Rudbeckia hirta (G) Salvia farinacea (H) Tagetes erecta (I)
'5 ’Antigua Primrose'
g 30 . ........................................................................................... L ............................... ....
{-1 DLImax = 13.5 mol-m'Z-d'1I DLImax >15 mol In 2 -d 1
a 20 4» . .1 . ........................... . ..
3 DL'max >19 moI rn'2 2.21
LI... 1o... ......
0 2 2222222 patula 2 (J) 2Verbena thybn'da2 (K) 2 Zinnia elegans 2 (L)
20 .5 . .. .. .. ., » . 20.11.3112. ..wa..t.e..rfa|.| Mix .....................................
DLIImax >20 mol m'2 d'1II Dleax >19 mol- IIIn'ZI- d'I1
15 4. . .
nLIIImaIIx- — 13 5rn9I m 2 d 1
0 5 10 15 0 5 10 15 0 5 10 15 20

Daily light integral (mol-m'z-d'1)

Fig. 5.5. The increase in flower bud or inflorescence number as the daily light integral
(DLI) increases in 12 species of bedding plants grown at a constant temperature set point
of 20 °C and under a 16-h photoperiod. Predictions were determined using the
polynomial response surface equations presented in Table 5.5. Circles represent the
estimated maximum DLI (DLImaX) for the greatest flower bud or inflorescence number
for each species. Circles are not presented if DLImIx is predicted to occur above of the
DLI range for which the models were developed.

150

Literature Cited

Brondum, J .J . and RD. Heins. 1993. Modeling temperature and photoperiod effects on
growth and development of dahlia. J. Amer. Soc. Hort. Sci. 118236—42.

Faust, J .E. and RD. Heins. 1993. Modeling leaf development of the African violet
(Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 118:747—751.

Faust, J .E. and RD. Heins. 1994. Modeling inflorescence development of the African
violet (Saintpaulia ionantha Wendl.). J. Amer. Soc. Hort. Sci. 119:727-734.

Hidén, C. and R.U. Larsen. 1994. Predicting flower development in greenhouse grown
chrysanthemurn. Scientia Hort. 58: 123-138.

Landsberg, J .J . 1977. Some useful equations for biological studies. Expt. Agr.
13:273-286.

Larsen, R.U. 1989. A prediction model for floral development of Senecio X hybridus Hyl.
‘Molls Stams’. Acta Hort. 248:247-254.

Larsen, R.U. 1990. Plant growth and modeling by light and temperature. Acta Hort.
272:235-242.

Larsen, R.U. and L. Persson. 1999. Modeling flower development in greenhouse
Chrysanthemum cultivars in relation to temperature and response group. Scientia
Hort. 80:73-89.

Moccaldi, LA. and ES. Runkle. 2007. Modeling the effects of temperature and
photosynthetic daily light integral on growth and flowering of Salvia splendens
and Tagetes patula. J. Amer. Soc. Hort. Sci. 132:283-288.

Pramuk, LA. and ES. Runkle. 2005. Modeling growth and development of celosia and

impatiens in response to temperature and photosynthetic daily light integral. J.
Amer. Soc. Hort. Sci. 130:813-818.

Reed, K.L., E.R. Hamerly, B.E. Dinger, and P.G. Jarvis. 1976. An analytical model for
field measurement of photosynthesis. J. Appl. Ecol. 13:925-942.

151

 

 

APPENDIX B

COMPARISON OF DIFFERENT LIGHT SOURCES ON GROWTH AND

FLOWERING OF BEDDING PLANTS

152

 

Comparison of Different Light Sources on Growth and Flowering of Bedding Plants

Research Objective
To compare the effects of high-pressure sodium (HPS) lamps with sunlight on

growth and flowering during the finish stage of three bedding plant species.

Materials and Methods

On 6, 4, and 5 Mar. 2009, seeds of petunia (Petunia thbrida ‘Dreams Neon
Rose’), snapdragon (Antirrhinum majus L. ‘Montego Orange Bicolor’), and verbena
(Verbena thbrida ‘Obsession Lilac’), respectively, were sown in plug trays [288-cell
size (6-ml volume)] by a commercial greenhouse (C. Raker & Sons, Litchfield, MI).
After germination, plugs were received at Michigan State University (MSU) and were
grown in a controlled environmental growth chamber (TC-2; Environmental Growth
Chambers, Chagrin Falls, OH) at a constant temperature set point of 20 °C. A l6-h
photoperiod was provided by 215-W cool-white fluorescent (CWF; F96T12CWVHO;
Philips, Somerset, NJ) and 60-W incandescent lamps (INC; Philips), at a CWF:INC (by
W) of 3.6, and at an intensity of 180 pmol-m’Z-s“l at plant height. On 1 April 2009,
plugs were transferred to a controlled environmental growth chamber at a constant
temperature set point of 10 °C and under a 9-h photoperiod to slow the rate of
development.

All plugs were thinned to one seedling per cell. During the plug stage, plants
were irrigated as necessary with well water acidified with H2804 to a titratable alkalinity

of 140 mg-L"l CaCO3 and containing 95, 34, and 29 mg-L‘l Ca, Mg, and S, respectively.

153

The water was supplemented with a water-soluble fertilizer providing (in mg-L“) 62 N,
6 P, 62 K, 7 Ca, 0.5 Fe, 0.3 Cu, Mn, and Zn, 0.1 B and Mo (MSU Well Water Special;
GreenCare Fertilizers, Inc., Kankakee, IL).

On 9 April 2009, seedlings were transplanted into 10-cm round plastic containers
(480-ml volume) filled with a commercial soilless peat-based medium (Suremix;
Michigan Grower Products, Inc., Galesburg, MI). Petunia, snapdragon, and verbena had

a mean node number at transplant of 12, 7, and 4, respectively. After transplant, plants

 

were randomly assigned to three different photosynthetic daily light integral (DLI)
treatments that each delivered 8 mol-m‘Z-d‘l. Each treatment contained 15 plants of
each species. The DLI treatments consisted of (1) 8 mol-m’z'd‘l delivered entirely from
400-W HPS lamps (100% HPS), (2) 4 mol'm’Z-d’l each from sunlight and HPS lamps,
delivered separately (50% HPS), and (3) 8 mol-.m‘Z-d‘l delivered entirely from natural

sunlight (0% HPS).

 

DLI Treatments l
1. 100% HPS. Plants of each species were grown in an environmental growth l
chamber under HPS lamps delivering 139 umol-m"2's“l for 16 h (8 mol-m‘2°d"). The l
HPS lamps operated continuously from 0600 to 2200 HR. Plants remained in the same
growth chamber throughout the study and were not shaken during the experiment.
2. 50% HPS. Plants of each species were grown under HPS lamps delivering 139
umol-m‘Z-s"1 from 0600 to 0900 HR (1.5 mol'm”7--d“) in a compartment that excluded
natural sunlight (CompENS). Each day at 0900 HR, plants were transferred to a

greenhouse and were grown under natural sunlight with a maximum photosynthetic

154

photon flux (PPF) of 1080 umol~m‘2-s‘1. . After plants received 4 mol'm‘z-d“l of
sunlight (141 to 447 min, depending on irradiance), they were subsequently transferred
back to the CompENS and were grown under HPS lamps delivering 139 pmol-m‘Z-s’1
for 5 h (2.5 mol-m‘z-d“‘). In the CompENS, the photoperiod was extended to 16 h with
2 to 4 umol-m‘Z-s‘l from HPS lamps. The time required to transfer plants between
environments was 55 min.

3. 0% HPS. Every day at 0900 HR, plants of each species were transferred from
an CompENS to a greenhouse and were grown under natural sunlight with a maximum
PPF of 1080 umol-m‘Z-s‘l. After plants received 8 mol-m‘Z-d‘l of sunlight (2234 min,
depending on irradiance), they were subsequently transferred back to the CompENS. In
the CompENS, plants were grown under a 16-h photoperiod provided as day-extension
lighting from 0600 to 2200 HR from HPS lamps delivering 2 to 4 umol-m‘Z-s'l. On
some days during the study, plants did not receive 8 mol-m'z-d‘l of sunlight in the
greenhouse environment because of cloudy conditions. Therefore, plants were
transferred to the CompENS between 1700 to 1900 HR, and on the following day, plants
received more than 8 mol-m‘z-d’l from sunlight to make up the deficit. The time

required to transfer plants between environments was 55 min.

Environmental Monitoring and Plant Culture

In each environment, plants were grown at a constant air temperature set point of
22 °C. Air temperature was independently measured in each environment by aspirated,
shielded thermocouples (0.13-mm type B; Omega Engineering, Stamford, CT) positioned

at plant height. At 10 cm above the bench in each DLI treatment, the PPF was measured

155

by a quantum sensor (LI-1908A; LI-COR, Inc., Lincoln, NE). In the greenhouse
environment, the PPF was measured by a line quantum sensor containing 10 photodiodes
(Apogee Instruments, Inc., Logan, UT). In each treatment, a thermocouple (0.13-mm
type E; Omega Engineering) was inserted 0.5 cm below the shoot tip of one plant of each

species and the actual plant temperature was recorded. Thermocouples were reinserted

 

daily after plants were transferred between environments. In the growth chamber,

thermocouples were repositioned weekly as plants developed.

 

Environmental measurements were collected every 10 s and hourly means were
recorded by data loggers (CR10X; Campbell Scientific, Logan, UT). A vapor-pressure
deficit was not maintained during the study. Plants in the growth chamber were irrigated

as previously described, but with a water-soluble fertilizer providing (mg-L-l) 125 N, 11

 

P, 126 K, 13 Ca, 1 Fe, 0.5 Cu, Mn, and Zn, 0.1 B and Mo (MSU Well Water Special;
GreenCare Fertilizers, Inc.). Plants in the CompENS and greenhouse were irrigated with
reverse osmosis water supplemented with a water-soluble fertilizer providing (in mg- L"')
125 N, 12 P, 100 K, 65 Ca, 12 Mg, 1.0 Fe and Cu, 0.5 Mn and Zn, 0.3 B, and 0.1Mo

(MSU RO Water Special; GreenCare Fertilizers, Inc.).

Data Collection and Analysis

The date of first open flower per plant was recorded and time to first open flower
was calculated for each plant. Petunia were considered flowering when one flower had a
fully open corolla. Snapdragon and verbena were considered flowering when plants had
2 and 6 flowers, respectively, open on an inflorescence. When each plant flowered, the

following data were recorded: total number of flower buds (petunia) or inflorescences

156

(snapdragon and verbena), plant height from the soil surface to the base of the open
flower or inflorescence, number of nodes on the primary shoot below the first open
flower, number of branches, and total shoot dry weight. Data were analyzed using SAS
mixed-model procedure (PROC MIXED), and pairwise comparisons between treatments

were performed using Tukey’s honestly significant difference test at P 50.05.

Results

Plants of petunia and snapdragon flowered a mean of 4 to 5 d earlier under 100%
HPS compared with 0% HPS (Table 6.1). In verbena, there was no difference in time to
flower among treatments. Verbena grown under 100% HPS had a 0.8 to 1.0 g higher
shoot dry weight at flower compared to 50% HPS and 0% HPS. Snapdragon and verbena
grown under 100% HPS were a mean of 4 to 8 cm taller at flower than other treatments.
In all species, there were no differences among treatments in the number of flowers, leaf
nodes, or branches. Mean air temperature was within 0.3 °C among treatments, but mean

plant temperature was 0.4 to 0.9 °C higher under 100% HPS versus 0% HPS.

157

688:8 Ho: San:
.modwm we. 52 855$? 2805:me bags: Maud—3. .3 5:28 use 36on some 5532 5:833 :moEN

 

n v; :2 0.0 a Q: NE m. L m wdm 0mm o.w od mm: $0
9 o; a: mg. a v.2 fimfl mdm N._m 3N oé can mam $3
a YN 02 ad a odm 2: _dm 5.3 _.mm od 0.x mam $2:
.025 commomno. m5€o>
ed 03 md n 2: mm m mdm odm Wmm ow od mam o\oo
wd W? m.@ n no— od an 9.2 mgm Nam o6 o6 mam Xom
ed 02 _d a WE 9m 9 3: m._m ..mm o.o 9w mam $2:
2.285 098.5 03252. cowfiwmmcm
m._ - .02 mi 5.0 m wdm 0.3 YNN o.w od mam AX5
m._ - WC Ym NE on odN film mdm oé o.v mm: °\oom
m._ x- QQ m.m m.o_ Na 5.: «.3 fimm ed o.w mam 82:

.80M :82 3889 «Baum
A3 5325 .0: £055 .oc owo Z A88 .0: 3303 A3 826$ :85 :< Emma—3m mam 8253;
R6 Hoosm Ewfim 3 083. Gov oSHEaEB ATVNIEBEV
:82 30
Ame: $8 szca 532 see 32:8 6 25 $5 sou 3333
32:8 £82 mm: Ed 225; so: :08 73-5.65 a. Q .Ama: £8: 85 6%: 528m 033233 sec 32:5 E
5&6 353—26 Ttdlszoe w mo AEQV _SwBE Em: 33. a $95 :3on 225 355 .2213 use .cowfiwmmcm .323qu E mew—032m
8 8:285 use .83: .Boaoc .«o 898:: 2: no new Jaws: .5963 De Hoonm 530$ 8 08: :0 350m Em: mo 8:963 A6 23%

 

 

 

158