Light-emitting-diode (LED) lighting for greenhouse tomato
production
Paul Deram
Under the supervision of Dr. Mark Lefsrud
Department of Bioresource Engineering
Macdonald Campus
McGill University, Montréal
January, 2013
A thesis submitted to McGill University in partial fulfillment of the
requirements of the degree of Master of Science.
©Paul Deram 2013
Page 2 of 115
ABSTRACT
The cost of artificial lighting is a major expense in the greenhouse production
industry, especially during the winter where supplemental lighting is required to
maintain production. Current technology uses broad spectrum high pressure
sodium lamps (HPS), which, despite being excellent luminous sources, are not the
most efficient light source for plant production. Specific light frequencies have
been shown to impact photosynthesis more directly than others (especially in the
red and blue ranges); focusing on specific wavelengths, light-emitting diodes
(LEDs) could diminish lighting costs due to their high efficiency and lower
operating temperatures. LEDs can be selected to target the wavelengths absorbed
by plants, enabling the growers to customize the wavelengths of light required to
maximize production and limit wavelengths that do not significantly impact plant
growth. The primary purpose of this experiment was to test tomato plants
(Solanum lycopersicum), in a research greenhouse using a full factorial design
with three light intensities (High: 135 µmol m-2 s-1, Medium: 115 µmol m-2 s-1 and
Low: 100 µmol m-2 s-1) at three red to blue ratio levels (5:1, 10:1 and 19:1)
compared to 100% HPS, and a control (no supplemental lighting). The exact
wavelengths chosen were 449 nm for the blue and 661 nm for the red. Secondary
treatments were also tested using 100% red light supplied from the top, 100% red
light supplied from the bottom, a 50%:50% LED:HPS and a replicate of the 10:1
ratio with High light intensity. The experiment was replicated over two different
seasons (Summer-Fall 2011 and Winter-Spring 2011-2012). During the
experiment, the highest biomass production (excluding fruit) occurred with the
19:1 ratio (red to blue), with increasing intensity resulting in more growth,
whereas a higher fruit production was obtained using the 5:1 ratio. The highest
marketable fruit production (fruit over 90 g, Savoura internal standard) was the
50%:50% LED:HPS, followed by 5:1 High and 19:1 High. From this research,
LEDs have been shown to be superior in fruit production over HPS alone, and
LEDs can improve tomato fruit production with HPS and have the ability to
become the dominant supplemental greenhouse lighting system.
Page 3 of 115
RÉSUMÉ Le coût de l'éclairage artificiel est une dépense importante dans le secteur de la
production en serre, surtout en hiver lorsqu’un éclairage supplémentaire est
nécessaire pour maintenir le niveau de production. La technologie actuelle utilise
des lampes à haute pression de sodium (HPS), qui en dépit d'être d'excellentes
sources lumineuses, ne sont pas les sources lumineuses les plus efficaces pour la
production végétale. Certaines fréquences spécifiques de lumière ont montré avoir
un impact plus direct sur la photosynthèse que d'autres (en particulier dans les
gammes de rouge et de bleu); en mettant l'accent sur certaines longueurs d’onde,
les diodes électroluminescentes (LED) pourraient diminuer les coûts d'éclairage,
en raison du rendement élevé et des températures plus basses de ce type de lampe.
Les LED peuvent cibler les principales fréquences de lumière mieux absorbées
par les plantes, ce qui permettrait aux producteurs de créer une lumière aux
longueurs d'onde adaptées à la production optimale des plantes. Le principal
objectif de cette expérience était de tester les lampes sur des plants de tomate
(Solanum lycopersicum) dans une serre de recherche en utilisant un plan factoriel
complet avec trois intensités lumineuses (Haute: 135 μmol m-2 s-1, Moyenne: 115
μmol m-2 s-1 et Basse: 100 μmol m-2 s-1) et trois proportions de rouge et bleu (5:1,
10:1 et19: 1), et comparer leur performance à celle de 100% HPS, et d’un contrôle
(pas d'éclairage supplémentaire). Les longueurs d'onde choisies sont 449 nm
(bleu) et 661 nm (rouge). Certains traitements secondaires ont également été
testés, dont 100% rouge (éclairage par le haut ou le bas), un 50%:50% LED:HPS
et une reproduction du 10:1 à haute intensité. L'expérience a été menée au cours
de deux saisons différentes (été-automne et hiver-printemps). La production
végétative la plus importante s'est produite avec le rapport 19:1 (rouge : bleu). La
production de fruits était la plus élevé avec le rapport 5:1. La production en fruits
commercialisables la plus importante (fruits de 90 g et plus : étalon interne de
Savoura) a été pour le 50%:50% LED:HPS, suivi du 5:1 et 19:1 à haute intensité.
Les LED se sont montrés supérieures aux HPS quant à la production de tomates.
Page 4 of 115
ACKNOWLEDGEMENTS This thesis is the result of the input of numerous parties. First of all, I would like
to thank the company, General Electric Lighting Solutions Canada (Nabil Jacques
Salem, Jean-François Richard) for proposing the research project, and committing
to the research with funding and development of the LED array prototypes.
I would also like to thank my supervisor: Dr. Mark Lefsrud for providing me
with much needed support throughout the duration of this research project, as well
as opening up his laboratory for me to conduct my research, and for sharing his
personal expertise in the domain of LED lighting for plant research.
I would like to thank Dr. Valérie Orsat, for her always kind and encouraging
advice, as well as for all the help given as part of my advisory committee for the
sake of this project.
To Savoura in Portneuf, QC (David Brault and Claire Boivin), I extend my thanks
for their knowledge in tomato growing and care techniques which was
instrumental in the completion of this research, as well as for teaching me how to
work efficiently and in a timely fashion in a greenhouse operation.
To any others who helped me during the greenhouse experiments, for answering
my questions, for advice on methodology or even for their friendship, thank you
all (Alejandro Jaul, Nicholas Matlashewski, Allison Busgang, Michael Schwalb,
Julie Gagné, Anil Patel, Yvan Gariépy and many others).
Finally, I would like to thank my family, friends and my wonderful fiancée, for all
their moral support, and for making their unseen presence known and felt.
Page 5 of 115
TABLE OF CONTENTS
ABSTRACT ............................................................................................................. 2
RÉSUMÉ ................................................................................................................. 3
ACKNOWLEDGEMENTS ..................................................................................... 4
TABLE OF CONTENTS ......................................................................................... 5
LIST OF TABLES ................................................................................................... 8
LIST OF FIGURES ................................................................................................. 9
LIST OF EQUATIONS ......................................................................................... 11
ABBREVIATIONS ............................................................................................... 12
1. INTRODUCTION .......................................................................................... 13
2. HYPOTHESIS AND OBJECTIVE ................................................................ 15
3. LITERATURE REVIEW ............................................................................... 16
3.1. Tomato and Greenhouses ............................................................................ 16
3.2. Photosynthesis ............................................................................................. 17
3.3. Lighting Systems: ........................................................................................ 20
3.4. LEDs in Plant Research:.............................................................................. 23
3.5. Effects of Different Wavelengths ................................................................ 24
3.6. Effect of Intensity ........................................................................................ 27
4. MATERIALS AND METHODS ................................................................... 29
4.1. Plant Care .................................................................................................... 29
4.2. Experimental Setup ..................................................................................... 31
4.2.1. Greenhouse Setup ................................................................................. 31
4.2.1. LED Setup ............................................................................................ 34
4.3. Instrumentation ............................................................................................ 36
Page 6 of 115
4.4. Statistical Analysis ...................................................................................... 39
5. RESULTS ....................................................................................................... 41
5.1. Fruit ............................................................................................................. 41
5.1.1. Marketable Fruit Number ..................................................................... 41
5.1.2. Marketable Fruit Mass .......................................................................... 42
5.1.3. Total Fruit Number ............................................................................... 44
5.1.4. Total Fruit Mass .................................................................................... 45
5.1.5. Average Fruit Mass............................................................................... 47
5.2. Vegetative Biomass ..................................................................................... 49
5.2.1. Fresh Mass ............................................................................................ 49
5.2.2. Dry Mass ............................................................................................... 50
5.2.3. Dry to Fresh Biomass Ratio .................................................................. 52
5.2.4. Fruit and Flower Counts ....................................................................... 53
5.3. Fruit to Biomass Ratio ................................................................................. 55
5.4. Environmental Data ..................................................................................... 57
5.4.1. Temperature .......................................................................................... 57
5.4.2. Relative Humidity ................................................................................. 57
5.4.3. Light Sensor Map ................................................................................. 58
6. DISCUSSION ................................................................................................. 60
6.1. Ratios chosen ............................................................................................... 60
6.2. Fruit ............................................................................................................. 61
6.2.1. Marketable Fruit ................................................................................... 61
6.2.2. Total Fruit ............................................................................................. 63
6.3. Vegetative Biomass ..................................................................................... 64
Page 7 of 115
6.4. HPS to LED comparison ............................................................................. 66
6.5. Top versus Bottom Lighting Systems ......................................................... 66
6.6. Light Measurement Techniques .................................................................. 67
6.7. Production Issues ......................................................................................... 71
6.7.1. First Run: .............................................................................................. 72
6.7.2. Second Run: .......................................................................................... 74
6.8. Observations for future research ................................................................. 76
7. CONCLUSIONS ............................................................................................ 81
8. REFERENCES ............................................................................................... 83
APPENDIX A: Formatted Data ............................................................................. 89
APPENDIX B: Statistical Data (Box Plots) .......................................................... 94
APPENDIX C: Raw Data .................................................................................... 101
APPENDIX D: Weather and Lighting Data ........................................................ 110
Page 8 of 115
LIST OF TABLES
Table 1: Greenhouse section placement ............................................................... 31
Table 2: Treatment list and description.................................................................. 35
Table 3: End of first run and start of second run light map data............................ 38
Table 4: Fruit and flower count summary .............................................................. 54
Table C1: Total fruit data ..................................................................................... 101
Table C2: Red fruit data ....................................................................................... 103
Table C3: Marketable fruit data (> 90 grams) ..................................................... 104
Table C4: Biomass data ....................................................................................... 105
Table C5: Fruit and flower count data ................................................................. 108
Table D1: Temperature data during the second run ............................................. 110
Table D2: Temperature data during the first run ................................................. 111
Table D3: End of second run light map data ....................................................... 112
Table D4: Initial light map data ........................................................................... 113
Table D5: Summary of initial light map data ...................................................... 114
Table D6: GE spectroradiometer data .................................................................. 115
Page 9 of 115
LIST OF FIGURES
Figure 1: PAR curves based on absorbed light and direct photosynthesis
measurement .......................................................................................................... 18
Figure 2: Cross sectional view of an LED lamp as well as a conventional LED .. 20
Figure 3: Spectral distribution of LEDs versus a dysprosium lamp ...................... 21
Figure 4: High pressure sodium spectrum ............................................................. 22
Figure 5: Diagram of the greenhouse used in the experiment ............................... 32
Figure 6: Light and plant setup for each section .................................................... 36
Figure 7: Total number of marketable fruit per light treatment ............................. 42
Figure 8: Total marketable fruit mass per light treatment ..................................... 43
Figure 9: Total number of fruit per light treatment ................................................ 44
Figure 10: Total fruit mass per light treatment ...................................................... 46
Figure 11: Fruit mass average per light treatment ................................................. 47
Figure 12: Fruit mass average for fruit over 90 g per light treatment .................... 48
Figure 13: Wet mass of plant biomass (excluding fruit) per light treatment ......... 50
Figure 14: Dry mass of plant biomass (excluding fruit) per light treatment ......... 51
Figure 15: Dry to fresh biomass ratio .................................................................... 52
Figure 16: Total fruit mass to plant biomass ratio ................................................. 55
Figure 17: Marketable fruit to plant biomass ratio ................................................ 56
Figure 18: Light direction from the LED arrays .................................................... 68
Figure 19: Typical angular response of the LI-193 ............................................... 70
Figure 20: Two examples of viable clusters turning into secondary stems ........... 77
Figure 21: Leaf burn .............................................................................................. 78
Figure 22: Powdery mildew ................................................................................... 79
Page 10 of 115
Figure A1: Total marketable fruit for the second run for the tomato plants .......... 89
Figure A2: Total marketable fruit for the first run for the tomato plants.............. .89
Figure A3: Total mass of marketable fruit for the second run ............................... 90
Figure A4: Total mass of marketable fruit for the first run ................................... 90
Figure A5: Total fruit count during the second run ............................................... 91
Figure A6: Total fruit count during the first run .................................................... 91
Figure A7: Total fruit mass during the second run ................................................ 92
Figure A8: Total fruit mass during the first run ..................................................... 92
Figure A9: Total plant fresh biomass (excluding fruit) for the second run ........... 93
Figure A10: Total plant fresh biomass (excluding fruit) for the first run .............. 93
Figure A11: Total plant dried biomass (excluding fruit) for the second run ......... 94
Figure A12: Total plant dried biomass (excluding fruit) for the first run .............. 94
Figure B1: Total number of marketable fruit per plant for the 120 day harvest .... 95
Figure B2: Total marketable fruit mass per plant, for the 120 day harvests .......... 96
Figure B3: Total number of fruit per plant, for the 120 day harvests .................... 96
Figure B4: Total mass of fruit per plant, for the 120 day harvests ........................ 97
Figure B5: Total fresh biomass per plant for the 120 day harvests ....................... 97
Figure B6: Total dry biomass per plant for the 120 day harvests .......................... 98
Figure B7: Ratio of dry to fresh biomass production ............................................ 98
Figure B8: Ratio of fruit mass to fresh biomass for the 120 day data ................... 99
Figure B9: Ratio of marketable fruit mass to fresh biomass for the 120 day ........ 99
Figure B10: Total number of red fruit per plant, for the 120 day experiment ..... 100
Page 11 of 115
LIST OF EQUATIONS
Equation 1: Number of fruit over 90 grams ........................................................... 40
Equation 2: Mass of fruit over 90 grams................................................................ 41
Equation 3: Number of fruit ................................................................................... 42
Equation 4: Fruit mass ........................................................................................... 43
Equation 5: Vegetative fresh biomass .................................................................... 44
Equation 6: Vegetative dry biomass ...................................................................... 45
Page 12 of 115
ABBREVIATIONS
ANOVA: Analysis of variance
HPS: High pressure sodium
LED: Light-emitting diode
PAR: Photosynthetically active radiation
R2: Coefficient of determination for statistical model fit
s.e.: Standard error
St. dev.: Standard deviation
5:1: ratio of five times more red light than blue light intensity (μmol m-2 sec-1)
10:1: ratio of ten times more red light than blue light intensity (μmol m-2 sec-1)
19:1: ratio of 19 times more red light than blue light intensity (μmol m-2 sec-1)
Page 13 of 115
1. INTRODUCTION
Tomato is one of the most consumed crops in the world. In northern climates,
tomatoes have a short time period for growing in the field, thus to maintain supply
and reduce shipping costs, tomatoes are grown locally in greenhouses. Northern
climate greenhouses permit year-round tomato production and provide consumers
with local, competitively priced production. Supplemental lighting is a major
expense but permits increased production, especially during the winter months,
when the daylight interval is much shorter. Greenhouse growers’ benefit from
natural sunlight, and with the addition of supplemental lighting, provide the plants
with a longer irradiance period during the day (usually a 16 hour photoperiod as
opposed to the 10 hours of natural sunlight).
Conventional lighting systems for greenhouses utilize broad-spectrum light
sources, such as high pressure sodium (HPS) or fluorescent lamps. These lamps
are excellent luminous sources for the human eye, but are not the most efficient
light sources for plant production, due to their low levels of blue light and other
photosynthesis sensitive wavelengths. Light-emitting diodes (LEDs) are a fifty
year old technology which is showing potential in the greenhouse industry. With
LEDs, specific wavelengths can be produced, creating a custom light spectrum
targeted for maximum plant production. Studies have shown that the most
important wavelengths for photosynthesis are in the blue and red wavelengths;
peaks in photosynthetic efficiency are found at 440 (blue), 620 (red) and 670 (red)
nm (+/- 10 nm) (McCree 1972a) (Figure 1 in section 3.2).
This research consisted of a full factorial design of three intensities (High: 135
µmol m-2 s-1, Medium: 115 µmol m-2 s-1 and Low: 100 µmol m-2 s-1) and three
ratios of red (661 nm) light to blue (449 nm) light (5:1, 10:1 and 19:1), to
determine which combination was best suited for tomato plant growth and
production. A series of secondary light treatments were added to compare the
factorial results to the current industry standard HPS lighting, a control with no
supplemental lighting, a 50%:50% LED:HPS (10:1 and HPS combined), a 100%
red, and a 100% red where the LEDs were placed at the base of the plants, facing
Page 14 of 115
upwards. The production increase due to the light treatments was statistically
significant; the top producing treatments (5:1 High and 5:1 Med) were
significantly different from the control (no supplemental light).
Following this introduction is a description of the objectives and basic hypothesis
of the project (Chapter 2). Chapter 3 is a review of pertinent literature which is
used as a basis for the project set-up, as well as a guide of the experimental
design, performance and analysis of the experiment. The methodologies followed
during the experimental phase of this research are outlined in Chapter 4. The
results are presented in Chapter 5, with a discussion of the pertinent quality
assessments of the results as well as an overview of what is needed for future
research presented in Chapter 6. Chapter 7 completes this thesis with an
assessment of the outcome of the research as well as a final conclusion of the
results obtained. Given the industrial nature of the project, specific
recommendations are made for new and existing greenhouses interested in
increasing their production capacity by switching to a partial or full LED
supplemental lighting system.
Page 15 of 115
2. HYPOTHESIS AND OBJECTIVE
This experiment was conducted within the targeted objectives of General Electric
Lighting Solutions Canada, where the overall goal of the project was to use the
company’s prototype LED arrays to research and determine which ratio and
intensity combinations best suit tomato plants (Solanum lycopersicum), in the
fruiting stage for greenhouse applications. Previous studies performed at McGill
University (unpublished data, Biomass Production Laboratory) have shown that
the ratios of red to blue light used for this experimentation increased plant
photosynthetic rate at the seedling stage, and this research was to confirm if these
same ratios would result in improved biomass production (fruit and vegetative)
for mature plants.
The specific aims were to:
Determine which combination of ratio and intensity of LEDs was
optimum for plant production.
Compare these results to plants grown under HPS lighting, 50%:50%
HPS:LED and a control with no supplemental lighting.
Compare 100% red LEDs in a normal configuration to 100% red LEDs
placed at the bottom of the plants shining upwards.
It was theorized that lighting from below could increase plant uptake of the
energy, thus two treatments were put in place as a side experiment to test this
theory.
Page 16 of 115
3. LITERATURE REVIEW
3.1. Tomato and Greenhouses
Tomatoes are one of the most important crops in the world. According to
Statistics Canada, tomato sales accounted for close to 50% of the total fruit and
vegetable sales in the country in 2011 (Statistics Canada, 2011). Tomato sales in
Canada reached $496 million, an increase of 5.5% since 2010. The next highest
selling crops in Canada are peppers ($300 million), cucumber ($279 million) and
lettuce ($27.6 million) (Statistics Canada, 2011). Tomatoes can be grown in the
field, but increasingly, tomato production occurs in greenhouses. Greenhouse
tomato production started in the late 1980’s for Canada and the late 1990’s for the
USA, and grew more than tenfold from 1992 to 2003 (Cook et al. 2005).
Greenhouse production exceeded field grown tomatoes in Canada in 1997, and in
2003, the greenhouse production was more than nine times that of field grown
tomatoes (Cook et al. 2005). Today there are 23 million square meters allocated
to greenhouse fruit and vegetable production in Canada, with sales exceeding $1.1
billion annually. According to Brazaitytė et al. (2009), in countries of high
latitude (e.g. Canada), tomatoes are almost exclusively grown in greenhouses.
Light irradiance is the limiting factor for increasing production in greenhouses,
when all other factors (temperature, nutrient levels and water availability) are
adequately maintained (Lefsrud et al. 2008). Ohashi-Kanto (2007) states that
artificial lighting permits stable vegetable crop production no matter the
environmental conditions (with favorable temperature in the greenhouse).
Greenhouses in northern latitudes must compensate for the attenuation in total
light availability (from prolonged winter with short daylight hours), and
supplemental artificial lighting is required in order to maintain a consistent crop
yield throughout the Canadian winters. Conventional greenhouse lighting
systems utilize broad-spectrum light sources, such as high pressure sodium (HPS)
or fluorescent lamps. These lamps were tailored for human vision and therefore
are not ideally suited for plant growth (Bula et al. 1991). For example, light
Page 17 of 115
irradiance from HPS lights is neither considered optimal nor efficient for plant
growth (Brazaitytė et al. 2009), because it lacks irradiance in the 400-500 nm
range, and this low level of blue light is detrimental to plant growth. Some HPS
lamps have been modified to provide irradiance in the 400-500 nm range (SON-
T-Agro HPS lamp, Philips, Amsterdam), which significantly increased the crop
production when compared to normal HPS lamps (Brazaitytė et al. 2009). These
modified HPS bulbs did not perform as well as most other light treatments tested:
380 nm, 447 nm, 520 nm, 595 nm, 622 nm, 638 nm, 660 nm, 669 nm, 721 nm
LEDs and combinations (Brazaitytė et al. 2009). Modified HPS bulbs can result
in morphological disorders (lower chlorophyll concentrations, smaller stomata
and lower above-ground biomass) in many species, such as tomato plants
(Menard et al. 2006).
3.2. Photosynthesis
Plants require light for photosynthesis, proper growth (seed germination, organ
elongation and phototropism) and physiological development (Goto 2003). Light
is also needed by plants to mediate many hormonal and morphological changes,
and specific wavelengths have been shown to benefit plants in different ways
(Okamoto 1996). For example, both blue light (~420-450 nm) and red light (660
nm) are known to increase chlorophyll production and play an active role in
photosynthesis (Okamoto 1996). Both wavelengths have also been shown to
improve morphological and physiological aspects of plants (Lu et al. 2012;
Poudel et al. 2008). Light catalyzes chemical reactions within the chloroplast
through photosynthesis. Light absorbance occurs in chlorophyll a and b (green
pigments which allow plants to absorb energy from light), as well as in the
carotenoids, lutein and β-carotene (other organic pigments synthesised by plants),
with each pigment having its own absorbance characteristics. Light excites these
pigments which results in improved growth; however, some light wavelengths
which are not absorbed by these pigments can have beneficial effects. For
Page 18 of 115
example, green light is not absorbed by these pigments, but exposure to only
green light at high photosynthetic photon flux will produce a healthy plant (Lu et
al. 2012).
Figure 1: PAR curves based on absorbed light and direct photosynthesis
measurement (Taiz and Zeiger 1998).
The photosynthetically active radiation (PAR; Figure 1) curve represents the
percentage of light absorbed and utilized by the different pigments in the plant as
a function of wavelength between 400 and 700 nm. The curve has been
developed based on the photosynthetic efficiency (called the action spectrum: the
rate of physiological activity plotted versus light wavelength) and plant pigment
light absorbance curves. Maximum growth occurs in the red and blue spectrums
of light while reduced growth occurs in the green region.
Numerous studies have reported that differences in wavelength produce
significant changes in plant production, photosynthetic efficiency and pigment
accumulation; changes in the irradiance wavelength spectrum have resulted in
changes in biomass production and plant morphology, due to the effect of blue
Page 19 of 115
light (Hoenecke et al. 1992; Gauthier et al. 1997; Taiz and Zeiger 1998), such as
shorter hypocotyl length (2 mm as opposed to 30 mm), stronger concentration of
chlorophyll (up to 20% higher), while with shifts in the ratio of the red/far red
spectrum (Brown et al. 1995; Heraut-Bron et al. 2001; Hoenecke et al. 1992; Taiz
and Zeiger 1998), there were plant elongation, lower overall dry mass (10% when
compared to white light) and smaller leaf area (15% lower when compared to
white light). Traditional methods of wavelength control (such as filtering white
light, blue enriched white light, or far red enriched white light) can cause changes
in the chlorophyll a to b ratio (Walters and Horton 1995), chlorophyll biosynthesis
and action spectrum (Anstis and Northcot 1974; Björn 1967; French 1991; Koski
et al. 1951; Ogawa et al. 1973; Virgin 1993), and the β-carotene biosynthesis
(Ogawa et al. 1973). Under UV-B light exposure, decreases in chlorophyll b have
occurred (Taiz and Zeiger 1998). Virgin (1993) showed that chlorophyll’s
relative accumulation is wavelength dependent for bean (Pinguicula vulgaris L.),
pea (Pisum sativum L.), potato (Solanum tuberosum L.) and wheat (Triticum
spp.), with both species and individual plant tissues responding differently.
Measurement of the absorption spectra has been determined for most plant
pigments (Köst 1988); limited information exists on the effect of wavelength on
the production of vegetable crops. Existing studies have emphasized the action
spectra of chlorophyll in plants (Anstis and Northcot 1974; Björn 1967; French
1991; Koski et al. 1951; Ogawa et al. 1973; Virgin 1993), but very little work has
concentrated on the action spectra or the effect of wavelength on PAR efficiency,
at intensities above the compensation point. The compensation point is the light
intensity at which the rate of photosynthesis is exactly equal to the rate of
respiration for a plant.
Page 20 of 115
3.3. Lighting Systems:
Light-emitting diodes (LEDs) were first invented in 1962, by Nick Holonyak
under the name Ga(As1-xPx) p-n Junctions (Holonyak et al. 1962). LEDs are
semi-conductors, which emit light through electroluminescence (when an electric
current or electric field passing through a material triggers a light response).
LEDs are created using a p-n junction, where a p-type semiconductor (positive)
and an n-type semiconductor (negative) are set side by side (Figure 2).
Figure 2: Cross sectional view of an LED lamp as well as a conventional LED
chip (typically 250 x 250 x 250 μm) (Craford 1992)
Different materials give different colours; red light can be a combination of
aluminium gallium arsenide (AlGaAs), gallium arsenide phosphide (GaAsP),
aluminium gallium indium phosphide (AlGaInP), or gallium (III) phosphide
(GaP) (Craford 1992; Mukai 1999). For blue light, Zinc selenide (ZnSe), or
Indium gallium nitride (InGaN) (Craford 1992; Mukai 1999; Xie 1992) can be
used. LEDs have been shown to last up to 100 000 hours, but early degradation in
output or life expectancy will occur in extreme temperatures and high current
Page 21 of 115
settings (Fu et al. 2011). LEDs life expectancy declines exponentially with
increasing p-n junction temperature (Fu et al. 2011). Conditions of high moisture
and temperature, such as those found in a greenhouse setting may deteriorate the
LEDs faster, but their longevity will still remain higher than that of current
greenhouse lighting technologies (Fu et al. 2011). LEDs emit a short span,
monochromatic light, which permits the creation of a custom light spectrum (by
adding different LEDs), which can more closely resemble the light spectrum
needed by plants for photosynthesis. Figure 3 (below) shows the spectral
distribution of four types of LED lamps used by Xiaoying et al. (2012) with a
dysprosium lamp as a control. The LEDs used were 449 nm (blue), 512 nm
(green), 590 nm (orange) and 632 nm (red). As seen in Figure 3, LEDs promote a
much higher intensity for the wavelengths they emit than a dysprosium lamp
(white light, main difference with HPS is a higher percentage of blue light and
lower red light). Each LED creates a very strong intensity of a targeted
wavelength for photosynthesis, increasing the efficiency. For this experiment, the
LED peaks were centered at 449 nm (blue) and 661 nm (red).
Figure 3: Spectral distribution of different LEDs versus a dysprosium lamp
(Xiaoying et al. 2012)
Page 22 of 115
High Pressure Sodium (HPS) lights are the major source for greenhouse lighting
(Argus 2010) with over 56 million lamps sold worldwide (Heliospectra 2011).
The use of HPS in greenhouses was shown to significantly increase production in
plants in the early 1970’s, and was shown to have beneficial effects in tomato
production (McAvoy et al. 1984). The increase in plant growth was due to
increased lighting duration in greenhouses and was extended to 18 hours daily (as
opposed to the sole use of sunlight, ~10hr) (McAvoy et al. 1984). Adverse
effects, such as arrested flower development, foliar discolouration or slow growth
of the plant’s apex occurred from the use of HPS lighting on tomato plants
(McAvoy et al. 1984). These adverse effects were explained by the lack of blue
light in the HPS spectrum, which is essential for proper plant growth (Brazaitytė
et al. 2009). The spectral range of HPS lights focuses heavily in the yellow
wavelengths (Figure 4), while lacking red and blue wavelengths, which are
essential for photosynthesis (Figure 1).
Figure 4: High Pressure Sodium Spectrum (Agriculture Solutions 2012)
Page 23 of 115
3.4. LEDs in Plant Research:
The action of light spectra has been a focus of study since the late 19th century,
but it wasn’t until 1919, with research from Garner and Allard on the
photoperiodic response of flowering in plants that the field of photobiology
developed (Garner et al. 1920). Research has focused on the effect of different
lighting apparatus on plants (Cosgrove et al. 1981; Barro et al. 1989; Brown et al.
1985; Lefsrud et al. 2008; Maruo et al. 2012), and now, LEDs are a major focus
for research in this field. The response of plant parts to LEDs has been researched
since the early 1990’s (Bula et al. 1991; Hoenecke et al. 1992). The rapid
increase in LED technology has been driving this research (Brazaityė et al. 2009).
LED performance has become more than 20 times more efficient from the mid-
1970s to the mid-1990s (Crawford 1992), creating more and more lumens out of
less watts (0.2 lumens per watt in 1970, 10 lumens per watt in 1990). New
colours have been added, such as orange and green lights, but also non-visible
colours such as far-red and ultra-violet (XiaoYing et al. 2011, 2012). Hoenecke et
al. (1992) described the high-output LEDs used for their research as having a
bandwidth of ± 30 nm with a peak at 660 nm, and Xiaoying et al. (2012) reported
new LEDs with a bandwidth of ±15 nm, permitting for much better focus on the
most efficient wavelengths for photosynthesis. This advancement in wavelength
control is not the only benefit LEDs have over other commercially used lighting
sources for plant development: LEDs are characterized by a relative small mass (<
1 g each), small volume, relatively cool emitting temperature, longevity of over
100 000 hours (Folta 2005), linear photon output, wavelength specificity and the
range of possible wavelengths which can be created (Goins et al. 1997; Brazaityė
et al. 2009; Massa et al. 2008) are all characteristics which make them better
suited to crop production than earlier greenhouse lighting systems. Bula et al.
(1991) proposed that the advancement of LEDs would be the solution for growing
plants in space, and that they could potentially be used as a sole light source for
crop growth, as opposed to supplemental lighting. NASA has focused on the
necessary ratio of red light to blue light in order to grow plants in outer space with
Page 24 of 115
LEDs as their sole source of light, such as for a life-support system on Mars
(Massa et al. 2008). Due to the low availability of blue LEDs, red LEDs were
supplemented with blue fluorescent lights, and the research focused on the
possibility of sole-source or supplemental lighting for outer-space missions, and
controlled environment research (Massa et al. 2008). With the increased
availability of the full range of colors of LEDs and the decreasing costs of
fabrication, LEDs are becoming viable for extensive commercial applications,
such as tomato greenhouses in northern latitude countries; and are being
researched extensively (Brazaitytė et al. 2009). Tennessen et al. (1994) explained
that research on the effect of different wavelengths on plants will greatly benefit
from LEDs since they work as well (and sometimes better, with a quantum yield
of photosynthesis of 0.0027 with red LEDs versus 0.0022 with white light from a
metal halide lamp) as other lighting systems, are more reliable, easily repeatable,
and much more portable.
3.5. Effects of Different Wavelengths
Different wavelengths of the light spectrum have been found to have specific
effects on plant morphology, physiology, photosynthesis efficacy and flowering
capabilities (Menard et al. 2006). Most wavelengths have been shown to have
positive and negative aspects, thus research is focusing more on proper
combinations of light (Massa et al. 2008). Deregibus et al. (1983) showed that red
light (600-700 nm) is beneficial to grasses from the Lolium spp., by increasing the
tiller rate by 20% and increasing the leaf area by 15% without showing any
morphological drawbacks. Red light (650 nm) is shown to have beneficial effects
on tomato and cucumber (Menard et al. 2006). Menard (2006) showed that red
light (650 nm) slightly reduces internode length in cucumber, benefited fruit color
and post-harvest conservation, but can also increase the total amount of starch
molecules inside the chloroplasts in tomatoes by up to 10%.
Page 25 of 115
Okamoto et al. (1996) reported that both red and blue light can be used by
chlorophyll during photosynthesis, and explained that blue light is beneficial to
plant morphology and overall health: blue light (450 nm) has been shown to
heavily suppress stem elongation in multiple plant species. The effect has been
shown to have effects which can last for many hours into the night (after
termination of the supplemental lighting) in cucumber (Cucumis sativus), pea
(Pisum sativum) and mung bean (Vigna radiata), while plants such as sunflower
(Helianthus annuus), azuki bean (Vigna angularis), and zucchini (Cucurbita
pepo.) undergo dark recovery within the first half hour of night (Cosgrove 1981).
Blue light has shown to decrease dry mass and increase stem length in marigold
seedlings (three times longer stems than under the control or the red supplement),
while it has been shown that the same light would increase dry mass for Salvia
(Salvia spp.) seedlings (Heo et al. 2002). Poudel et al. (2008) showed that blue
light promoted stem growth suppression in grape transplants, but that it also
increased the concentration of chlorophyll in the shoots by 40 to 67% dependant
on the cultivar when compared to red light, making the smaller plants (40%
smaller than when compared to red light) and more effective at photosynthesis.
Overall, blue light is not as effective for photosynthesis as red light, since it
inhibits leaf growth by reducing cell expansion and reduces the total amount of
chlorophyll in the leaves (Goto 2003). Due to this lower efficiency, researchers
tend to undervalue the use of blue light and not consider it in high proportions for
plant growth (Goto 2003). A lack of blue light has been shown to have very
adverse effects on plant morphology: low number of chloroplasts, lower thickness
of cell walls, and low spongy mesophyll tissues (Goto 2003). Blue light was
shown to stimulate stomata opening, and increasing the rate of photosynthesis by
up to 30% in some species (Menard et al. 2006). Growth responses (enhanced
total dry matter, more flowering…) to blue light have been shown to be more
significant and rapid than similar growth responses to red lights (Goto 2003).
Goto (2003) explains that some of the blue light responses are not dependent on
the ratio of blue to red, but just on the total irradiance of blue light. The ratio of
blue to red light is shown to be the most important factor when using LEDs, since
Page 26 of 115
having both blue light and red light increases plant biomass growth and fruit
production by over 20% when compared to plants grown under only one of the
wavelengths (Goto 2003; Lefsrud et al. 2008; Brazaitytė et al. 2009; XiaoYing et
al. 2011, 2012).
Far red light has been shown to have a strong biological significance, but is not as
readily available on the market as blue and red LEDs. The low availability of far
red LEDs is due to the limited value for human applications or the industrial
sectors (Kubota 2012). Far red has been shown to induce and increase flowering
in many plants such as Gypsophila paniculata and Arabidopsis (Hori et al. 2012),
potato (Miyashita et al. 1995), increase leaf area in lettuce (Kubota et al. 2012),
and stem growth in pepper (Brown et al. 1995). Far red light does not impact
photosynthesis directly, and doesn’t promote dry matter production (Goto 2003).
Supplementing far-red light to red light increases plant growth and health
significantly, by increasing internode length, increasing plant biomass and
increased photosynthesis (Miyashita et al. 1995; Goto 2003). Brown et al. (1995)
suggested that the addition of blue light to red light is much more crucial than the
addition of far-red light to red light.
Plant response to light from the red and the blue spectra has been documented
extensively (Menard et al. 2006). The results show that plant response to LEDs
(and different wavelengths) depends on the plant species and in some cases even
the cultivar. Cosgrove (1981) showed that three different cultivars of cucumber
were not impacted in the same way by the blue light stem development
suppression; Burpee’s Pickler type cucumber suffered from rapid inhibition but
underwent dark recovery (return to normal growth overnight), Levo type pickle
suffered from rapid inhibition and did not undergo dark recovery and finally
Lemon type pickle suffered from the strongest rapid inhibition and remained
stable during the night (Cosgrove 1981).
Other wavelengths have been studied, such as orange and green, showing that
they significantly reduced photosynthesis when compared to other wavelengths in
tomato plants (XiaoYing et al. 2011, 2012). A 200% reduction in photosynthetic
Page 27 of 115
response was observed for both orange and green wavelengths when compared to
a blue and red light composite, a 120% reduction was observed when compared to
blue light and a 100% reduction was observed when compared to red light
(XiaoYing et al. 2011, 2012). When used in conjunction with blue and red
wavelengths however, green light may have a beneficial effect on overall plant
growth. No statistical differences for production or photosynthetic response were
found between the treatments with only red and blue wavelengths and the
treatment with green light added to the red and blue, but the one with added green
had slightly higher stomatal growth and plant size (XiaoYing et al. 2011). Yellow
and green lights were shown to greatly increase stem length (by over 40% when
compared to white light) and decrease leaf area (by over 50% when compared to
white light) and resulted in the leaves being more brittle and lighter in colour (less
pigments) and lower overall net mass of the plants (XiaoYing et al. 2011).
3.6. Effect of Intensity
Lefsrud et al. (2006) reported that changes in irradiance levels result in significant
changes in the measured accumulation of carotenoids and chlorophylls in kale.
Dorais et al. (1990) showed that an increase in photosynthetic photon flux density
from 100 μmol m-2 s-1 to 150 μmol m-2 s-1 (supplied from 400-W HPS bulbs) will
increase production by 10%, 16% and 14% for tomato plants (Lycopersicon
esculentum Mill. cv. Caruso) grown in low density (2.3 plants m-2), variable
density (between 2.3 plants m-2 and 3.5 plants m-2) and high plant density (3.5
plants m-2) respectively (Dorais et al. 1990).
McAvoy et al. (1984) showed that increasing light irradiance has beneficial
effects on fruit production. Five experiments were run on tomato plants during
the first six months of 1982, using 400-W HPS bulbs to produce 100 μmol m-2 s-1,
125 μmol m-2 s-1 and 150 μmol m-2 s-1. Fruit yield, mass, cluster size and percent
fruit set all increased significantly: average number of fruit increased from 4.4 per
plant (for 100 μmol m-2 s-1) to 5 (for 150 μmol m-2 s-1), mass increased from 2 kg
Page 28 of 115
per m2 (for 100 μmol m-2 s-1) to 2.9 kg per m2 (for 150 μmol m-2 s-1) (McAvoy et
al. 1984).
A study was done by Tennessen et al. (1994), submitting tomato plants to a range
of intensities from 0 μmol m-2 s-1 to 1400 μmol m-2 s-1, using red LEDs (660 nm)
or water filtered light from a xenon arc. It was show that the photosynthetic
response was similar for the xenon white light and for the red LEDs;
photosynthesis increased until close to 800 μmol m-2 s-1 (with each increase in
intensity causing diminishing increases in photosynthetic rates), and then
remained constant up to 1400 μmol m-2 s-1 (a slight loss occurred with the red
LEDs after 1000 μmol m-2 s-1). Stomatal conductance behaved similarly to the
rate of photosynthesis up to 800 μmol m-2 s-1, but higher intensities decreased the
conductance. It was also shown that the red LEDs promoted higher
photosynthetic rates than the white light until 250 μmol m-2 s-1, where the white
light becomes more efficiently used (Tennessen et al. 1994).
Page 29 of 115
4. MATERIALS AND METHODS
4.1. Plant Care
All the culturing methods and the tomato care were based on the methods used by
Savoura (Portneuf, QC), a large tomato production greenhouse. The greenhouse
set points were 21oC during the day and 16oC during the night. The watering
system was on for 3 minutes every 20 minutes, starting at 06:00 and ending at
22:00. The supplemental lighting was left on for 16 hours a day, from 06:00 to
22:00. The nutrient solution supplied to the plants via the drip irrigation system
was based on a full strength Hoagland’s solution (Hoagland et al. 1950) modified
by Savoura (proprietary information), of which 40L (nutrient solution) were used
per week (20L of the nitrogen solution and 20L of the micronutrient solution).
The water flow was 40 ml per minute, with total irrigation of approximately 6L
per plant per day (based on 1 ml per 1 J of irradiance, instructed by Savoura). The
nutrient solution was monitored three times a week to maintain proper electro-
conductivity (EC) and pH level. The nutrient solution was mixed automatically,
as needed using one hundred times concentrated stock solutions. Plants were
pollinated by hand, shaking the flowers with a Q-tip.
A central computer in the greenhouse was used as the control system. It
controlled the lighting (on at 06:00 and off at 22:00), the irrigation (on for 3
minutes every 20 minutes during the same hours as the lights), and the ventilation,
coupled with a mister, in order to keep the temperature close to the set points.
The internal greenhouse temperature was monitored and controlled; relative
humidity was monitored, but not controlled.
The LED arrays were positioned at an inter-canopy height, no more than 10 cm
below the top of the plants. The lights height was increased weekly at the start of
the experiment, and every two weeks towards the end, in order to keep the LED
arrays at a level of 75% the plant growth. Once the plants reached the maximum
height of 2.1 m (7 feet) for the greenhouse bays, the array height could not be
increased any more.
Page 30 of 115
At harvest, fresh mass was determined by separating and weighing both aerial
plant biomass (the rooting system was discarded) and fruit biomass. All fruit
greater than 2 g were counted and weighed (fruit under 2 g were included as plant
biomass). Fruit and flower numbers were counted at two weeks intervals during
the first run, every month for the second run and at final harvest. Fruit was
harvested throughout the experiment at the first observed red pigmentation
(considered ripe), counted and individually weighed. Plants were pruned
according to Savoura’s methods (approximately every 2 weeks), with fresh and
dry biomass measured for each plant (Savoura unpublished data). The Savoura
pruning method was necessary to reduce the leaf nodes to between 10 and 14, to
remove tertiary branching of the stem (a primary and secondary stem were
maintained), but no fruit clusters were removed from the primary or secondary
stems. All fruit were counted (above 2 g); however, only healthy looking flowers
were included. Weekly pruning removed some suckers with flowers: these
flowers were included in the earlier counts, but were not considered in later
counts.
The fresh biomass harvested (aerial and fruit) was dried according to the ASABE
standard (2007), with a temperature of 65 ºC for no less than 72 hours and
subsequently weighed. One representative ripe fruit was collected from each
plant during the final harvests, freeze dried, and stored at -80oC for future fruit
quality measurements. The remaining fruit were made available to Macdonald
Campus students for consumption. No fruit were subject to sensory evaluation.
The plant measurements taken during the experiment included fresh and dry mass,
total and marketable fruit yield, fruit and flower counts, and ratio of fruit to
biomass. Fruit were also separated into two categories depending on size: fruit
from 2 g to 90 g and fruit that weighed more than 90 g. Savoura uses 90 grams as
an internal standard for the minimum mass at which fruit are acceptable to be sold
on the market (Savoura unpublished data).
Page 31 of 115
4.2. Experimental Setup
4.2.1. Greenhouse Setup
A 7.6 m x 12 m (24’ 10” x 39’ 5”) greenhouse room (set in a north-south
orientation, in the southern west corner of the greenhouse) was used for the
experiment. Long wire-mesh tables (1.2 m high, or 4’) were used as a base for the
tomato plants, in order to ensure more even airflow (with unobstructed areas
underneath the plants. The tables were also set in the north south orientation, with
three tables (4.4 m x 1.6 m, or 14’ 5” x 5’ 4”) in the northern end and three tables
(6.1 m x 1.6 m, or 20’1”x 5’4”) in the southern end of the pod. These tables were
separated into 2 m x 1.6 m (6” 7” x 5’ 4”) sections, which were used for testing.
The southern, longer tables could each accommodate 3 sections, and the northern,
shorter tables could accommodate 2 sections, giving a total of 15 sections for this
experiment. Figure 5 (below) is a map of the greenhouse setup, with the specific
location of each light treatment during the two runs (Table 1, below).
Table 1: Greenhouse Section Placement (with regards to the section numbers in Figure 5, below)
Section Run 1 Run 2 1 5:1 Low 50% LED 2 5:1 Med Red Top 3 5:1 High 10:1 High 4 10:1 Low Red Bot 5 10:1 Med 19:1 High 6 10:1 High 5:1 Med 7 19:1 Low HPS 8 19:1 Med 5:1 Low 9 19:1 High 19:1 Low 10 HPS 19:1 Med 11 Control 10:1 Med 12 Red Top 100% LED 13 Red Bot Control 14 100% LED 10:1 Low 15 50% LED 5:1 High
Page 32 of 115
Figure 5: Diagram of the greenhouse used in the experiment
Greenhouse Bay with 7 benches (top set) 240 by 64 inch and (bottom set) 173 by 64 inches. Each treatment zone is at least 80 by 64 inches each. Plants were grown in two rows of 4 in a north-south orientation, as shown in zone 11 during the first experiment, and in 2 rows of 3 in a north-south orientation, as shown in zone 13 during the second experiment.
This bay is in the south west corner of the greenhouse with the south and west walls open to the outside environment and the other two walls open to another bay (north wall) or walkway (east wall), the entry door is in the east wall.
Table 1 above lists the placement of each treatment for the first and second experimental run (with respect to the section numbering in this figure)
3
2
1
6
5
4
9
8
7
11
10
Space
not
used
13
12
15
14
N
D
o
o
r
Page 33 of 115
Each section was surrounded by a double layer of 2.44 m (8 feet) 80% shade cloth
(8MK808, Harnois, St-Thomas, QC) which prevented 96% of light from passing
through, which greatly reduced the cross-contamination of light between
treatments, permitting analysis of all sections independently. A layer of 2.44 m (8
feet) 60% shade cloth (8MK608, Harnois, St-Thomas, QC) was also installed on
top of all the sections, under the roof of the greenhouse. This layer was used in
order to reduce the amount of sunlight attained during the summer (July to
September 2011) to under 20 mol m-2 day-1, in order to simulate winter lighting
conditions. During the second experimental run (January through April 2012),
this layer of 60% shade cloth was removed, in order to get the full winter sunlight.
The amount of natural light received by the plants throughout both experimental
runs was comparable.
A pressure driven drip irrigation system was installed in order to ensure even
distribution of water to the plants. A separate tubing system was built for the
north and the south tables. A 1/2” (5/8” Exterior) black flexible hose (14350
Biofloral, Montreal, QC) was run along the center of each table, and 8 pressure
compensating drip emitters were installed in each section (24 emitters on the
hoses on the tables from the south side, and 16 emitters on the hoses on the tables
from the south side). The emitters used were 19 L h--1 (5 GPH) Antelco agri-drip
emitters (14257 Biofloral, Montreal, QC). Each drip emitter was connected to a
30 cm (1’) 1/4” dripping hose (14314 Biofloral, Montreal, QC) with an 8”
Antelco dripper for 1/4” hose (14259 Biofloral, Montreal, QC), which supplied
water to one plant (set into the rockwool near the stem). The three table length
hoses from each side (North and South) were connected to the greenhouse water
supply. A double tank system was connected to the greenhouse water supply, in
which were stored the two nutrient solutions needed to grow the plants (refer to
section 4.2.)
Tomato plants cv. Trust (esculentum) (grafted onto cv. Maxifort (hirsutum)) were
obtained from Ontario Plant Propagation (St-Thomas, ON), 55 days after seeding.
Page 34 of 115
Average aerial mass at 55 days was 68.1 g (± s.e. 9.1 g) fresh mass and 4.8 g (±
s.e. 0.7 g) dry mass (10 plants). For the first run, the plants were received as two
single stemmed plants per rockwool cube (5.5 ” x 2.5” x 3.5”) and 4 plants were
placed per rockwool slab (Pargro QuickDrain slab 40”x6”x3” Biofloral, Montreal,
QC). Eight plants (2 rockwool slabs) were placed in each zone, fitted between the
light arrays. For the second run, the plants were received as one double stemmed
plant per rockwool cube (5.5 ” x 2.5” x 3.5”) and 3 plants were placed per
rockwool slab (Pargro QuickDrain slab 40”x6”x3” Biofloral, Montreal, QC). Six
plants (2 rockwool slabs) were placed in each zone. Therefore, there were 8
plants in the first experimental run and 6 plants in the second experimental run.
There were two runs of the experiment: once in summer (July-October 2011) and
once in winter (January-April 2012). A two month period was allotted between
the experimental runs in order to minimize the risk of potential pathogens
carrying over from one experimental run to the other. Each experimental run had
two harvest times: 70 days after being placed in the greenhouse and 120 days after
being placed in the greenhouse. The specific location of each light treatment into
the greenhouse sections was randomly allotted at the beginning of each
experimental run. Half the plants in each section (4 during the first experimental
run and 3 for the second) were randomly selected at the 70 day mark and
harvested, while the remaining plants were grown for the full 120 days.
4.2.1. LED Setup
A full factorial design with two treatment levels was implemented in order to
properly test the different LED ratios and intensities. Three ratios of red to blue
light (5:1, 10:1 and 19:1) and three intensities (Low: ~100 µmol m-2 s-1, Med:
~115 µmol m-2 s-1 and High: ~135 µmol m-2 s-1) were tested. Other experimental
treatments were included to compare the LED treatments to current greenhouse
standard lighting procedures: HPS and lighting from below the plant. The final
fifteen treatments were:
Page 35 of 115
Table 2: Treatment list and description
5:1 Low 100 µmol m-2 s-1
5:1 Med 115 µmol m-2 s-1
5:1 High 135 µmol m-2 s-1
10:1 Low 100 µmol m-2 s-1
10:1 Med 115 µmol m-2 s-1
10:1 High 135 µmol m-2 s-1
19:1 Low 100 µmol m-2 s-1
19:1 Med 115 µmol m-2 s-1
19:1 High 135 µmol m-2 s-1
100% LED Replicate of the 10:1 High (135 µmol m-2 s-1)
50%:50%
LED:HPS
10:1 Ratio LED coupled with HPS (total: 115 µmol m-2 s-1)
Red Top 100% Red light (130 µmol m-2 s-1)
Red Bot 100% Red light (110 µmol m-2 s-1), LEDs placed at the
bottom of the section, shining upwards into the plant canopy
100% HPS Full HPS light, (105 µmol m-2 s-1)
Control No supplemental lighting
The LED arrays were prototypes from General Electric Lighting Solutions
(Lachine, QC). These consisted of 1.78 m x 8 cm x 2 cm (70 in x 3 in x 0.8 in)
linear fixtures, on which were placed an array of 16 LEDs. A reflective coating
around each LED permitted for better dispersion of the light into the plant canopy.
Each section was fitted with either 3 light fixtures for the Low levels (with the
lights set along the length of the section (Figure 6) or 6 lights (same setup as for
the Low, but having 2 lights side by side in order to increase the intensity). The
LED arrays were designed to provide light at a 45 degree angle in both directions.
They consisted of sixteen modules of three LEDs and a refractor (designed to
project the light in the correct direction), alternating left and right. This way, 8
Page 36 of 115
groups of LED plus a refractor provide the light to each direction.
Photosynthetically active radiation (PAR) and irradiance levels (W m-2) were
measured to determine light maps of each section, at the beginning of the first
experimental run, the end of the second experimental run and once in between
(see instrumentation sub-section 4.3. for more details).
Figure 6: Light and plant setup for each section
4.3. Instrumentation
Each treatment had a temperature sensor (S-TMB-002; Hobo, Bourne, MA),
located 30 cm below the middle lamp array (except for the 100% HPS and the
Control, where the sensor was placed in the plant canopy at the same height as for
the other sections). Two treatments were randomly selected with
temperature/relative humidity sensors (S-THB-008; Hobo, Bourne, MA), one on
the north side and one on the south side. Light sensors were used in eight selected
treatments using pyranometers (S-LIB-M003; Hobo, Bourne, MA) and quantum
sensors (S-LIA-M003; Hobo, Bourne, MA), recording data for every minute
during the entire experiment. For these eight treatments, a light sensor was
Page 37 of 115
installed underneath the lamps in the middle of the experimental treatment (to
measure the light from the LED’s) and one was installed above the middle LED
array to measure the sunlight irradiance. For the 100% HPS and the Control, the
light sensor was installed in the plant canopy at the same height as the top sensor
for the other treatment sections. During the first experimental run, light sensors
were placed in the following treatments: 5:1 Low, Med and High, 10:1 High, 19:1
High, HPS, Control, and Red Bottom. During the second experimental run, the
chosen treatments for the light sensors were: 10:1 Low, Med and High, 5:1 Low,
19:1 Low, Red Top, 100% HPS and Control. The environmental measurements
were collected with three data loggers (U30; Hobo, Bourne, MA), and
downloaded to a computer.
A light irradiance map was created by taking nine point measurements beneath
each bulb or array, at a distance of 30 cm (12 inches), 1 m (39 inches) and 1.5 m
(59 inches) from each bulb or array in three locations of each LED array (10 cm
from the ends of the array and the middle location). Two measurement types
were taken for each sensor point: one at a vertical orientation directly below the
bulb and one at a 45º angle relative to the bulb always pointed in a western
direction (at 30 cm toward the plants, first reading measuring 3 arrays, then 2
arrays and one array). This difference in angle was required since the
spectroradiometer measures light in a perpendicular orientation relative to the
sensor and does not include light coming from any other direction; while the
LED’s were designed to direct their light at the plant canopy (to either side of the
lamp at a 45 degree angle). The HPS light data was measured at four levels, top
of the rockwool, starting height of the transplants (1 m, 1.5 m and 30 cm below
the bulb). The second and final light maps were performed with no plants present
in the greenhouse, in order to minimize shade. All readings were taken at least
half an hour after sunset with a spectroradiometer (PS-100, Apogee Instruments,
Logan, UT) with integration of the irradiance to obtain the PAR irradiance values.
A Li-COR spherical quantum sensor was used in the second and the final light
measurements (LI-193; Lincoln, Nebraska), for comparison. Light data are
provided in Table 3 (mid experiment run) and Tables D3-D6.
Page 38 of 115
Table 3: End of first run and start of second run light map data (January 5th 2012). Compiled data is from average of before (after calibration correction) and end (Li-COR and Spectroradiometer) data
Zone Spectroradiometer Li-COR Average of before and
after measurements
Daily Light Integral from artificial light (mol day-1)
% of Artificial
Light Average
Vertical Average
45º Average Average
5:1 LOW 103.4 135.5 119.4 113.9 98.3 8.50 37.3 5:1 MED 110.2 128.7 119.4 145.8 113.1 9.77 40.6 5:1 HIGH 107.0 136.4 121.7 188.9 132.9 11.48 44.5 10:1 LOW 88.8 121.8 105.3 123.4 104.7 9.05 38.8 10:1 MED 100.1 119.9 110.0 145.6 114.5 9.89 40.9
10:1 HIGH 115.6 140.1 127.8 233.3 151.9
13.13 47.9
19:1 LOW 104.6 139.5 122.0 112.1 104.5 9.03 38.7 19:1 MED 96.8 121.4 109.1 170.6 120.7 10.42 42.2
19:1 HIGH 131.4 154.0 142.7 177.6 139.4
12.04 45.7
LED 131.0 147.4 139.2 204.1 142.3 12.29 46.2 50% 95.2 135.0 115.1 131.3 105.1 9.08 38.8
Red TOP 119.2 149.4 134.3 175.2 133.2 11.50 44.6 Red BOT 82.9 105.6 94.3 98.1 91.2 7.88 35.5 Control 0.0 0.0 0.0 0.2 0.1 0.01 0.1
HPS Light Maximum Irradiance
Middle (1 m) 49.1 49.1 64.7 50.4 4.36 23.4 Height (1.5 m) 91.8 91.8 153.9 106.9 9.24 39.2 Top (30 cm from lamp) 158.7 158.7 427.9 253.9 21.93 60.5
Light map was for the LED bulbs in a vertical and 45 degree angle at a distance of 30 cm from the bulb in 9 locations per section (full darkness, with other bulbs on). The HPS was measured at three heights (of the chamber (approximate height of the seedlings), top of the chamber and 30 cm from the bulb (in a vertical orientation). All measurements were taken with a spectroradiometer and Li-COR spherical quantum sensor.
Page 39 of 115
4.4. Statistical Analysis
Data was analyzed according to the GLM (generalized linear model; ANOVA)
procedure of SAS (Cary, N.C.), as well as the MIXED procedure. The data was
separated for 70 day and 120 day harvests. The models used involved the Run
factor (experimental run 1 or 2), the Treatment factor (1 through 15, signifying
5:1 Low, 5:1 Med, 5:1 High, 10:1 Low, 10:1 Med, 10:1 High, 19:1 Low, 19:1
Med, 19:1 High, 100% LED, 50% LED, 100% HPS, Red top, Red bottom,
Control, respectively), and the Section Placement in the greenhouse factor
(numbered 1 through 15). The Run and Section Placement factors were treated as
random. A second model was run to make observations about the factorial data
only. The same Run and Section Placement factors were used, but the Treatments
factor was separated into two factors:
- Light ratio (1, 2, 3 for 5:1, 10:1 and 19:1 ratios of red light to blue light)
- Light intensity (100 μmol m-2 s-1, 115 μmol m-2 s-1, 140 μmol m-2 s-1, the
first statistical models were run with 1, 2 and 3, but then switched to the actual
μmol m-2 s-1 values in order to run the regression curves)
The different quantities tested in this experiment were:
- Total marketable fruit number (fruit weighing more than 90 grams)
- Total marketable fruit mass (fruit weighing more than 90 grams)
- Total number of fruit (fruit weighing more than 2 grams)
- Total fruit mass (fruit weighing more than 2 grams)
- Total plant biomass, fresh mass (at harvest)
- Total plant biomass, dry mass (after drying for a minimum of 72 hours,
following the standard ASABE protocol)
- Ratio of fruit mass to fresh plant biomass
- Ratio of marketable fruit to fresh plant biomass (fruit weighing more than
90 grams)
Page 40 of 115
Based on the statistical method used, plant number was not an independent
variable (since there was only one measurement per plant), thus having a
heterogeneous number of plants for each experimental run (8 plants versus 6
plants) does not impact the statistical analysis. The statistical models used can
handle this heterogeneity.
A standard Tukey-Kramer test was applied to determine significant differences
between treatments. Significance was determined with α ≤ 0.05. The relationship
between experimental dependent variables and treatments was determined by
regression analysis. Orthogonal polynomials were used to study changes
associated with treatments by partitioning the sums of squares into components
associated with linear and quadratic terms. Due to the strong constraints of the
experiment (large plant variability, limited space, limited time, large number of
different factors needing testing), the power of the statistical analysis was
calculated post hoc between 10% and 38% depending on the quantity being
tested.
Page 41 of 115
5. RESULTS
5.1. Fruit
Fruit data is presented for different categories below. The first category is the
number of marketable fruit, followed by the mass for the marketable fruit. Total
number of fruit and total fruit mass are the subsequent categories, and finally,
average fruit mass is presented. For the factorial model equations below,
quadratic relationships were attempted for all equations, but did not improve the
R2, thus only linear regressions are presented.
5.1.1. Marketable Fruit Number
The top 5 producers of total number of marketable fruit were 50%:50% LED:HPS
(with a total of 118 fruit over 90 grams), Red Top (115 fruit), 5:1 Med (113 fruit),
19:1 High (109 fruit) and 5:1 High (104 fruit). This total data is presented in
Figure 7. For more detailed data see Tables A1 (second run data), A2 (first run
data) and B1 (statistical box plot) in the appendix
Please note that only the factorial portion of the experiment could have regression
equations created, and are explained below.
From the factorial data only, a regression curve with R2 = 0.68 and 4.7 fruit
standard error was fitted. The regression equation was (in number of fruit):
# of Fruit over 90 g = 4.5 + Intensity*0.1 + Light Ratio*(-1.5) (Eqn. 1)
Where Intensity is the amount of light in the treatment (100 μmol m-2 s-1 for Low,
115 μmol m-2 s-1 for Med, 140 μmol m-2 s-1 for High) and Light Ratio is the ratio
of red light to blue light (1 for 5:1, 2 for 10:1 and 3 for 19:1). From this
regression curve, an increase in light intensity promotes a small increase in the
average number of marketable fruit: 1.5 extra marketable fruit when going from
Low intensity to Med intensity, and 2.5 extra fruit from Med intensity to High
Page 44 of 115
5.1.3. Total Fruit Number
For the total number of all fruit, the highest five producing treatments were 5:1
High (385 fruit). 5:1 Med (358 fruit), 5:1 Low (341 fruit), 19:1 High (315 fruit)
and 100% LED (310 fruit) (Figure A3). Some statistically significant differences
were reported, but the Control was statistically different from most other
treatments, specifically: 5:1 Low, 5:1 Med, 5:1 High, 10:1 High, 19:1 Med, 19:1
High, 100% LED, 50%:50% LED:HPS, Red Top and Red Bottom. This data is
presented in Figure 9. For more detailed data see Tables A5 (second run data),
A6 (first run data) and B3 (statistical box plot) in the appendix.
When analysing the factorial data only, a regression curve with R2 = 0.67 and a
standard error of 6.8 fruit was fitted. The regression equation was (in number of
fruit):
# of Fruit = 23.9 + Intensity*0.12 + Light Ratio*(-4.5) (Eqn. 3)
Where Intensity is the amount of light in the treatment (100 μmol m-2 s-1 for Low,
115 μmol m-2 s-1 for Med, 140 μmol m-2 s-1 for High) and Light Ratio is the ratio
of Red Light to Blue light (1 for 5:1, 2 for 10:1 and 3 for 19:1). From this
regression curve, an increase in Light intensity promotes a small increase in the
average number of total fruit: 1.8 extra marketable fruit when going from Low
intensity to Med intensity, and 3 extra fruit from Med intensity to High intensity
(for an average increase of 4.8 fruit when going from Low to High). An increase
in the percentage of red light decreases the amount of total fruit produced in the
treatment: 4.5 less fruit when going from 5:1 to 10:1 and 4.5 less fruit when going
from 10:1 to 19:1 (for an average decrease of 9 fruit when going from 5:1 to
19:1).
Page 49 of 115
5.2. Vegetative Biomass
5.2.1. Fresh Mass
Total fresh mass includes the biomass weighed during harvest, as well as the
biomass harvested from the weekly pruning of extra suckers and leaves. The
largest fresh plant biomass was attributed to 19:1 High with 37.8 kg, followed by
Red Bottom with 37.4 kg, 100% LED with 35.3 kg, 50%:50% LED:HPS with
34.5 kg, and 10:1 Med with 33.9 kg. Most light treatments gave very similar
fresh biomass to the plants, with only 3 treatments getting less than 30 kg of
production: Control with 14.7 kg, 5:1 Low with 25.1 kg and HPS with 26.9 kg.
This resulted in an average fresh plant mass between 2700 g for the 19:1 High and
1050 g for the Control. This data is presented in Figure 13. For more detailed
data see Tables A9 (second run data), A10 (first run data) and B5 (statistical box
plot) in the appendix.
When considering the factorial data only, a regression curve with R2 = 0.64 and
standard error of 831.5 g was fitted.
The regression equation was (in grams):
Vegetative Fresh Biomass = 1982.2 + Intensity*5.1+ Light Ratio*426.3 (Eqn. 5)
Where Intensity is the amount of light in the treatment, in μmol m-2 s-1 (100 for
Low, 115 for Med, 140 for High) and Light Ratio is the ratio of Red Light to Blue
light (1 for 5:1, 2 for 10:1 and 3 for 19:1). From this regression curve, an increase
in light intensity promotes a small increase in the plant fresh mass: 76.5 g increase
when going from Low to High, 127.5 g increase when going from Med to High
(for an average increase of 204 g from Low to High). An increase in the
percentage of red light significantly increases the amount of plant fresh mass
produced in the treatment: an extra 426.3 g is produced on average when going
from 5:1 to 10:1, and an extra 426.3 g is produced on average when going from
10:1 to 19:1 (average increase of 852.6 g when going from 5:1 to 19:1).
Page 52 of 115
5.2.3. Dry to Fresh Biomass Ratio
A dry mass to fresh mass ratio was tested, for the full data set. No statistical
differences were measured, with Control being the lowest (6.0%) and 5:1 Low
being the highest (7.8%). This ratio is used to determine if the plants become
more woody or retain more water (succulent) focusing on the generated biomass
(Clifford 1987). According to the literature, the average value for tomato plants is
between 6 to 8% (Heuvelink 2005). The top 5 treatments for dry to fresh biomass
ratio were 5:1 Low (7.8%), 5:1 High (7.6%), 10:1 High (7.6%), 19:1 High (7.4%)
and 19:1 Med (7.4%). Data can be found in Figure 15, with a supplemental
statistical bar graph in Figure B10.
Figure 15: Dry to fresh biomass ratio No statistically significant differences were observed.
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
rati
o o
f d
ry t
o f
resh
bio
mas
s
Page 53 of 115
5.2.4. Fruit and Flower Counts
The fruit and flower counts occurred on August 2nd, 9th, 12th, 16th, 25th, Sept
1st and 28th for the first run of the experiment, and on Feb. 8th, 28th, and April
27th for the second run (Table 4 with additional data in Table C5, appendix). The
5:1 ratio treatment had the most flowering and fruiting, followed by the 10:1, then
19:1 treatments, similarly to the first run. The 5:1 and 10:1 ratios, with middle or
high intensity had the best fruiting results. The 5:1 High had the highest fruit
count (191 fruit in the first run and 194 fruit in the second run), followed by 5:1
Med (192 fruit in the first run and 166 fruit in the second run), 5:1 Low (173 fruit
in the first run and 168 in the second run) and 10:1 High (161 fruit for the first run
and 143 for the second run). Control was the lowest with 13 fruit in the first run
and 10 fruit in the second, followed by HPS with 53 fruit in the first run and 96 in
the second.
Page 54 of 115
Table 4: Summary fruit and flower counts for the tomato plants (8 per zone for the first replication until September 1st and 4 after, 6 per zone for the second replication)
Date 5:1 TOTAL 10:1 TOTAL 19:1 TOTAL Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 78 61 106 43 102 42 9-Aug 161 85 153 61 118 55 12-Aug 163 116 113 106 85 75 16-Aug 114 169 77 104 63 90 25-Aug 114 277 79 214 71 148 1-Sep 70 411 53 286 41 206 28-Sep 54 205 54 137 54 95 Run 2 8-Feb 58 296 56 294 44 221 28-Feb 300 285 227 248 214 229 27-Apr 316 394 664 280 521 318
LOW TOTAL MED TOTAL HIGH TOTAL
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 106 15 99 88 81 43 9-Aug 107 27 148 103 177 71 12-Aug 119 40 112 145 130 112 16-Aug 73 59 86 146 95 158 25-Aug 96 152 76 218 92 269 1-Sep 71 192 33 323 60 388 28-Sep 42 145 56 126 64 166 Run 2 8-Feb 33 261 37 235 88 315 28-Feb 206 219 211 238 324 305 27-Apr 452 311 617 315 432 366
Page 57 of 115
5.4. Environmental Data
5.4.1. Temperature
The greenhouse temperature from December 31st to April 27th ranged from 40.2
ºC to 0.9 ºC. These extreme values occurred due to the greenhouse climate
control system shutting down twice during the month of February, and once in the
months of March and April. The extremes only occurred for a few minutes for
the heat, and a couple of hours for the cold. The average temperature was 16.9 ºC
with a standard deviation of 3.4 ºC, and was very consistent throughout the
greenhouse. The values broken down per month and per section can be found in
Table D2 for the first run and Table D1 for the second run. For the first run, the
average temperature was 21.8ºC with a standard deviation of 2.4ºC. These
conditions are similar to the conditions in a commercial tomato production
greenhouse with short extremes (38.7ºC and 9.9ºC for the highs and lows).
5.4.2. Relative Humidity
The relative humidity was measured in three locations with less than 2%
difference between the sensors. The relative humidity profile was very similar
during both runs of the experiment. Relative humidity had an average of 63%,
with a range from 94.5% to 13.5%. Typically, the average daily highest humidity
recorded was 82% +/- 5% with an average low of 30% +/- 10%.
Page 58 of 115
5.4.3. Light Sensor Map
The light sensor data is presented in summary and as mapped averages (Table 3,
section 4.3 and tables D3-D6, appendix). The spectroradiometer was recalibrated
for the data in Table D3, with the correction factor used in Table D5. An LI-193
Spherical Quantum Sensor collected light data and is presented in Tables D3-D4.
A spectroradiometer was used from GE and this data is presented in Table D6, but
due to the very large values obtained from this device it was not used for any
analysis (peak values of over 1600 μmol m-2 s-1 were obtained, making it close to
10 times higher than the expected values, or the other sensor recordings). The
large fluctuations in intensity observed with this instrument made it unreliable for
the purpose of the experiment. The data used in the calculations is the combined
data from the spectroradiometer and spherical light sensor taken between both
experimental runs, since those two sensors were found to be the most reliable and
stable for LED readings in this experiment. The first light data (before the
experiment) is reported but due to variation between the different sensors was not
accepted or used in the calculations. The remaining irradiance levels (end of first
run and end of second run) were very similar, with Low intensities having ~ 100
μmol m-2 s-1 (average of 102.5 μmol m-2 s-1 and 101 μmol m-2 s-1 for the second
and last light maps, respectively), while the Med had a level of 115 μmol m-2 s-1
(average of 116.1 μmol m-2 s-1 and 115.9 μmol m-2 s-1 for the second and last light
maps, respectively) and the High had a level of 135 μmol m-2 s-1 (average of 141.4
μmol m-2 s-1 and 130 μmol m-2 s-1 for the second and last light maps, respectively).
The HPS irradiance level average in the chamber was 110 μmol m-2 s-1 (average
of 106.9 μmol m-2 s-1 and 113.2 μmol m-2 s-1 for the second and last light maps,
respectively).
The experiment monitored the natural light irradiance and artificial light provided
to the plant through a number of sensors within the plant canopy, above the plant
canopy below the lights, and above the lights below the shade cloth. The
maximum natural irradiance measured in the test area was at 513 μmol m-2 s-1 at
Page 59 of 115
noon, which is very similar to the maximum of 566 μmol m-2 s-1 measured during
the summer under the shade cloth. During the summer run, all treatments
received almost equal amounts of natural light (small differences in shading using
the different lighting arrays), but during the winter run, it is possible that the angle
of incidence of the sun had an impact on the distribution of the sunlight. The total
light provided by natural light was an average of 14.3 mol day-1 with a standard
deviation of 6.2 mol day-1. Artificial light provided to the treatments varied from
8.5 mol day-1 for 5:1 low to a high of 13.1 mol day-1 for 10:1 high, the HPS
treatment was at 9.2 mol day-1 (Table D3). The inclusion of the natural light
increased the average daily light integral to an average from 26.3 mol day-1 for
19:1 High, a low of 22.8 mol day-1 for 5:1 Low, and 23.5 mol day-1 for HPS.
These results show that the percentage of total irradiance (including sunlight and
LED light) provided to the plant as artificial light sit in between 37% for the 5:1
Low treatment and 47.9% for the 10:1 High (Table 3, section 4.3). The HPS
treatment was at 39.2%, and the Control was around 0.1%. Irradiance levels
measured within the plant canopy at 1.5 m from the rockwool cubes (bottom of
the leaves on the plants), averaged less than 2 mol day-1.
Table 3 (section 4.3) shows the light map taken during the time in between both
experimental runs. Table 3 data was used for all calculations of light levels for
the experiment, such as the regression equations. Data from the final light map is
presented in Table D3. Table D4 shows the first light map, which was done only
with the spectroradiometer. Table D5 takes the data from D4 and applies the
correction factor sent to us by the manufacturer for the calibration before the final
light map. Little differences were found between Table 3 and tables D3 and D4,
making the use of Table 3 data for calculations viable.
Page 60 of 115
6. DISCUSSION
6.1. Ratios chosen
The primary focus of this experiment was to test the effect of the LED ratio and
light intensity (the 3x3 factorial). The 5:1, 10:1 and 19:1 ratios were chosen from
previous research performed at McGill University (unpublished data, Biomass
Production Laboratory), which showed that these ratios resulted in improved
photosynthetic responses with plants at the seedling (3-5 true leaves) plant age.
The three intensity levels were chosen following the supplementary light levels
used by Savoura in their greenhouses.
A range of factors or comparison tests were included:
HPS (a standard greenhouse light in order to compare the LEDs to
currently used technology)
50%:50% LED:HPS (a mix of HPS and 10:1 LEDs)
Control (no supplemental lighting, only sunlight)
100% Red Top (can be used to compare with the main tests, to illustrate
the need for blue light, but was installed as a side experiment, to compare
to the 100% bottom)
100% Red Bottom (used as a side experiment, in order to see the effects of
supplying the light from below the leaves of the plants, had an effect on
production)
100% LED (10:1 ratio, meant to be at a higher intensity than the High, at
160 µmol m-2 s-1). Due to the high heat emanating from the fixtures at this
intensity, the biomass which came in contact with the leaves would
experience heat blanching and stop growing. Therefore, the intensity was
lowered to 135 µmol m-2 s-1, making it a replicate of the 10:1 High
treatment.
Page 61 of 115
Most of the earlier work examining the impact of wavelength occurred under
ultra-low irradiance levels (<13 μmol m-2 s-1) (McCree 1971, 1972a; Inada 1976;
Sager et al. 1982; and Ogawa et al. 1973). Providing plants with elevated
intensities of light at narrow wavelengths will provide a better understanding of
how they utilize these wavelengths. Given that pigments absorb light at specific
wavelengths, conditions should exist whereby the wavelength intensity reaches a
level that result in the optimum light absorbance or possibly, the degradation of
plant pigments. Knowledge concerning the potential range of wavelength
intensities will promote more accurate design of plant lighting systems and
procedures, maximizing plant production
6.2. Fruit
6.2.1 Marketable Fruit
It is important to note that the majority of the fruit harvested were not fully
mature, at the 70 day and the 120 day mark. Thus, the large majority of the fruit
considered non-marketable (< 90 g) would continue growing in a normal
industrial setting, and reach the cut-off (since they are not harvested on a specific
date but after they have matured). Note that the ripe fruit were almost exclusively
considered marketable fruit, since they were harvested when mature. Therefore,
in order to get a proper overview of the effectiveness of a treatment, the results
cannot be based solely on the marketable fruit quantities; total fruit number and
mass also play a large role in the effectiveness of a treatment within this study.
Coupled with the shade cloth used to prevent light contamination from one
treatment to another which blocked out a portion of the sunlight, it results in a
lower overall production than what would be expected in a commercial
greenhouse using the LED lighting system.
The highest producing treatments of marketable fruit were 50%:50% LED:HPS,
5:1 High and 19:1 High . The 100% red treatments (both the top and the bottom)
were ranked fourth and fifth. The 19:1 High, and both 100% red treatments were
Page 62 of 115
much stronger with the production of total marketable fruit during the second run
when compared with the first run. This contradicts what the literature says about
red light, which is that it promotes biomass production and doesn`t promote
fruiting as much as blue light (Brown et al. 1995, Hoenecke et al. 1992). It is
possible that due to the increased biomass production, these treatments grew
faster during the winter (where the sun was weaker and at more of an angle), and
reached the top of the shade cloth (where more sunlight was available) quicker,
resulting in more production than the treatments which grew slightly slower. It
was also theorized that it could be due to how these treatments were placed in the
south-eastern side of the experiment, closest to the windows (Figure 5), which
benefited from more direct sunlight (with the angle of incidence and lower
irradiance due to the winter, could make a difference). 50%:50% LED:HPS had
the highest marketable fruit production during the second run, and was located in
the south-eastern corner, which started the idea that the location might be a factor
of influence in the study; however 50%:50% LED:HPS was rated second best
(behind 5:1 Med) during the first run, where it was located at the exact opposite
corner of the greenhouse. Angle of incidence of the sun was shown to be not
significant in the statistical analysis. Being ranked in the top 2 for total
marketable fruit production in both runs shows that 50%:50% LED:HPS is one of
the top treatments for marketable fruit. This agrees with research from Menard et
al. (2006) which shows that supplementing blue and red light to HPS creates
much higher production than with HPS alone. The 5:1 ratio had the highest
overall ratio, with all three intensities ranking in the top 50% of the experiment
for both experimental runs. This result is supported by the literature, where blue
light promotes fruiting and flowering (Menard et al. 2006, Goto 2003), but
contradicts Miyashita et al. (1995) who reported that red light, and not blue light,
would increase flowering.
Page 63 of 115
6.2.2. Total Fruit
The lower limit for considering a fruit was 2 g. This was selected because fruit
mass under this level are hard to distinguish from aborted flowers or aborted fruit;
during the first experiment especially, many small fruit and flowers were aborted
due to the extreme heat and a cut off was needed in order to remove them as
viable fruit. Unfortunately, due to the size cut off, a certain number of viable fruit
have not been considered (especially during the second run, where less fruit were
aborted since less high temperature extremes were felt). These small viable fruit
would have grown into full fruit if the experiment had been allowed to run for
longer, or in a production greenhouse.
The highest producing treatments were 5:1 High, 5:1 Med, 100% Red top,
50%:50% LED:HPS and 5:1 Low. As expected from the literature, higher
intensities bring forth more production (McAvoy et al.1984, Tennessen et al.
1994), with all ratios producing more in the High than the Med or the Low
irradiance levels. The 5:1 ratio was the top ratio from this experiment, with all
three intensities ranking in the top 5 treatments. This agrees with research from
Menard et al. (2006) and Goto (2003), who found that blue light promoted
flowering and fruiting.
The 19:1 High treatment resulted in much higher fruit production for the second
run than for the first run. This may be because 19:1 High had an increased
biomass production from the increased red light, as shown by Brown et al. (1995)
and Hoenecke et al. (1992). This higher biomass could have allowed the plants to
reach the natural sunlight faster than the other plants (due to the side wall
screening lowering the total light incoming from the sun). This extra production
was also observed in both the 100% Red treatments (top and bottom), but not to
the same extent. An increase in biomass from red LED light has been reported
(Brown et al. 1995; Hoenecke et al. 1992). This increase in the earlier stage
allowed for more sunlight towards the end of the experiment (since the plants
could grow taller, they received unobstructed sunlight quicker, increasing the
overall fruit production). This would not have occurred in an industrial
Page 64 of 115
greenhouse setting, since there would be no shade cloth impeding the sunlight on
the plants, and all treatments would have benefited from maximum sunlight from
the start. Second, due to the placement of the 19:1 High during the second run (in
the middle), the plants were less impacted by the cold (Adams et al. 2001), due to
a larger distance from the windows (data shows that the minima in this section is
over one degree higher than in sections closer to the windows). From the
temperature data however, the 100% LED (10:1 ratio) had a very similar
temperature to 19:1 High (and was also in the more sheltered spots), but did not
perform as well. Thus it is assumed that the red light was the driving factor of the
increased production (Barro et al. 1989), not temperature, which caused early
stem elongation leading to taller plants and more overall light (height was not
measured in the plants during this experiment).
Fruit quality, taste and color were not monitored or tested during the experiment.
No observable differences were noticed in these traits between the treatments
during or after the experiment upon consumption of the fruit.
6.3. Vegetative Biomass
The plants require a minimum amount of leaves to maximize fruit production but
extra vegetative production was pruned in order to maximize yield, since extra
suckers and leaves drain the plant of energy which could be focused into the fruit.
Pruning was performed every couple of weeks, and the biomass from the prunings
was weighed fresh and dry, and added into the final biomass data. The fresh and
dry biomass had very similar patterns to each other; the 5:1 ratio did have the
highest dry to fresh ratio, which coincides with data found in the literature,
showing that blue light promotes an increase in dry mass in tomato plants
(Menard et al. 2006).
The top treatments for both fresh biomass and dry biomass were 19:1 High and
Red Bottom, with the other 19:1 treatments and the 100% Red top in the top 50%
of treatments for biomass. This follows the literature which shows that red light
Page 65 of 115
produces more biomass (Brown et al. 1995; Hoenecke et al. 1992), since all
treatments with high red light were found consistently towards the top.
100% LED, 50%:50% LED:HPS and 10:1 Med were in the top 5 treatments,
which contradicts what was seen by McAvoy et al. (1984) and Dorais et al.
(1990), where an increase in intensity promotes an increase in biomass. All of
those treatments use the 10:1 ratio, yet the 10:1 High was behind both the
50%:50% LED:HPS and the 10:1 Med, which had 115 μmol m-2 s-1 as opposed to
the 140 μmol m-2 s-1 for the high. 100% LED is a replicate of the 10:1 High and
had increased biomass when compared to the two mid-level treatments which
could suggest that the 10:1 High treatment produced lower than expected due to
plant variability, which would then confirm the research which showed that
intensity had a positive effect on plant biomass (Dorias et al. 1990; McAvoy et al.
1984). 100% Red bottom produced more than 100% Red top, goes against the
research (since the red top received 130 μmol m-2 s-1, as opposed to the 110 μmol
m-2 s-1 received by the red bottom), resulting in less intensity creating higher
biomass. This would suggest that illuminating plants from the bottom could
increase overall plant growth. Moss (1964) showed that illumination from below
was at least as efficient as illumination from above; this research suggests that
illumination from below could increase biomass production significantly.
Ideally, all flowers will become fruit, but some flower abortion, as well as fruit
abortion was noted. The elevated temperature at the end of July, along with the
cold temperatures in February both could have limited the fruit set. Fruit and
flower counts are typically not used to make decisions in selecting significant and
non-significant treatment effects but are used to provide support for the decision
from the yield data (marketable and total).
Page 66 of 115
6.4. HPS to LED comparison
One major difference between the HPS lighting and the LED lighting is that the
HPS were fixed at the top, to mimic conventional greenhouse setup, while the
LEDs were set at an inter-canopy height. Thus, the LED illuminated plants
benefited from a higher light intensity at the start of the run, while the HPS plants
had less light to begin the experiment. Once the plants grew past a height of ~1.5
m, the HPS illuminated plants received more light than the LED treatments. This
increased light could explain that the LED treatments consistently outperformed
the HPS in every category; this is a benefit inherent to the design of the LED
systems, which can be placed closer to the plants with much less risk of damaging
or burning them (Brazaitytė 2009). Inter-canopy lighting is impossible with HPS
lighting due to the heat emitted by the lamps (Hogewoning et al. 2007).
6.5. Top versus Bottom Lighting Systems
The Red Top and the Red Bottom treatments were directly compared to each
other. The major difference between the two treatments was the set-up with the
bottom treatment lights placed at the bottom of the section and shone upwards,
while the red top LEDs were placed at the top, similarly to all other treatments.
The Red Bottom had slightly less light levels than the Red Top. Overall, the Red
Bottom section slightly underperformed when compared to the Red Top in
fruiting categories, but created more biomass. No statistical differences were
measured between these two treatments. The differences between the two
treatments are more pronounced towards the end of the run. This could be
explained by the lower light levels or because the plants in the Red Bottom
section were growing more side stems and suckers very close to the LED arrays,
and did not grow as tall (observed but not measured). This slight difference in
Page 67 of 115
yield between the two treatments observed towards the end of the run could be
explained by the Red Top plants growing taller more quickly, reaching the sun
(above the shade cloth) and benefiting from the added natural sunlight quicker.
While, the Red Bottom plants were observed as more bushy and shorter. No
conclusions can be drawn from the statistics, but the production results were
similar for both treatments, and the biomass production was higher in the Red
Bottom despite the light intensity difference. This could suggest that the light
coming from below was at least as beneficial as the light coming from above
(Moss 1964), and possibly even more beneficial than traditional lighting from
above.
6.6. Light Measurement Techniques
The point source nature of the LED arrays makes it very difficult to measure the
light intensity with conventional light measuring devices. Typical
spectroradiometers and light measuring devices are built to measure light levels
from the sun or other light sources which disperse light in every direction (such as
HPS lighting), and therefore are not as compatible with the LED arrays. A series
of irradiance tests were performed with a range of different light measuring
devices, in order to find the most appropriate solution for measuring the LED
point source light and arrays. The first test was performed using a pyranometer
(total solar radiation) (MP-100, Apogee Instruments, Logan UT) and a quantum
meter (measures only the PAR) (MQ-100, Apogee Instruments, Logan UT). It
was found that the light measurement from both devices was very variable (30%
differences in readings), under weighing blue light and over weighing red light
(Apogee 2012a, 2012b), and therefore they could not be used for the experiment.
The light fixtures were created to supply most of their light irradiance at a 45o
angle, as shown below.
Page 68 of 115
The light measurements were taken at horizontal, vertical and 45o, at three
different heights, at 3 points along the length of each light fixture, in order to get a
better overview of the light received by the plants. A base was created to place
the light measuring device at the exact same distance and angle from the lamp in
each position, but it was found that the measurements were still very variable
(close to 25% variability in the readings). The way the light fixtures are created,
LEDs are facing alternatively to the right or to the left (in order to get the 45o
angle of incidence, the LEDs were not designed to face down), with a reflecting
cone around each LED for better dispersion. Therefore, on the other side of the
LED, the light levels decrease dramatically. The same occurs for light
measurements from the bottom. Due to the reflecting cones, the light level
difference from being exactly under one of the LEDs or a couple centimeters off
makes a significant difference (60% of total radiation) in the light reading. Since
the light measuring devices only measure light coming directly from above, it
does not take into account the light originating at an angle from the LEDs
adjacent to the one sitting directly above the light measuring device (Figure 18
depicts the light direction from the LED arrays, showing that most of the light
does not go downwards).
Figure 18: Light direction from the LED arrays. Longer distance shows higher
light levels.
Page 69 of 115
Due to the relative simplicity of the quantum meter and the pyranometer, a more
complex spectroradiometer (BLACK-Comet Concave Grating Spectrometer,
StellarNet Inc., Tampa, FL) was tested, with the same measurement positions
used. Previously, the Apogee Instruments light sensor provided a value of 30
μmol m-2 s-1 for the 10:1 Low (average), while the spectroradiometer provided a
value of 68.4 μmol m-2 s-1; showing that the hand held sensor recorded half of the
expected value. Since the hand-held sensors (Apogee Instruments, MP-100 and
MQ-100, Logan, UT) were calibrated for solar irradiation (Apogee 2012a, 2012b)
they may not properly report wavelengths outside the 460-660 nm range. The
spectroradiometer was much more reliable than the quantum meter and
pyranometer, but the same problem with variability due to spatial position was
found. The spectroradiometer was designed for conventional overhead light
sources, and only records light incoming from directly above the sensor. The
field of view of a typical spectroradiometer was shown to be 10o (Mac Arthur et
al. 2007). With adjustments to the positioning (directly facing the LED for the
measurements), a light map could be created using the spectroradiometer, but it
was still not adequate for proper measurements of the light received by the plants.
Please note in Table D6 a spectroradiometer provided by General Electric (GE)
was used; the values have been provided but were significantly higher than any of
the other measurements taken with other instruments and were excluded from the
data analysis. The instrument was much more unreliable, with peaks at 1600
1600 μmol m-2 s-1, while some readings were lower than the expected values. Due
to this large variability, this instrument was not used for calculations. Note that
having a light level above 1600 μmol m-2 s-1 is considered brighter than the sun
and over 10 times higher than what was expected from the greenhouse lighting
system..
An underwater spherical quantum sensor (LI-193, Li-COR, Lincoln, NE) was
tested. The spherical quantum sensor was developed to better understand the light
dispersion in underwater biological experiments, by measuring the photon flux
coming from all directions. This sensor was tested due to its ability to record the
photon flux coming from all directions, which was better suited for measuring the
Page 70 of 115
LED lighting in this experiment (since the LED arrays do not give out uniform
light). The typical angular response to light is shown below (Figure 19). It is
seen that some light loss occurs, mainly due to the sensor at the base of the
sphere, but that the reading of the light coming from the upper 180o comes with
no significant loss. The same sensor positions were used to take the
measurements (3 heights, 3 positions along the length of each lamp, horizontal,
vertical and 45o). This sensor was well adapted to the LED fixtures, and gave us
much more accurate measurements of the lighting. This was the first sensor
which was found to accurately measure the light received by the plants, since it
took into account the light coming from multiple LEDs at each spot. Some losses
were found: when the bulb is placed sideways, the light measurement was 20%
lower than when placed right side up, which can be explained by the loss of
efficiency in the angular response graph. The Li-COR underwater quantum
sensor was chosen as the most reliable sensor for this research. It was also shown
to be the most adapted sensor for measuring light from LED light arrays available
on the market.
Figure 19: Typical angular response of the LI-193 (Li-COR 2012)
Page 71 of 115
Both sensors used for the light testing have some limitations: the
spectroradiometer’s angle of incidence is extremely narrow, giving unstable
readings from the point source LED’s, and the spherical Li-COR may have
limitations in the spectral response curves. Provided in this report is an average of
all of this data recorded in order to try and reduce the error that might occur from
this variability. The spectroradiometer had been calibrated before taking these
readings.
6.7. Production Issues
A certain number of production issues occurred during both experiments, possibly
having a slight effect on the overall performance of the plants in the experiment,
and have been outlined below. The production issues were separated per
experimental run.
Almost of all these production issues occurred in a fashion where they impacted
all treatments equally, thus they did not impact the statistical comparison between
treatments. However, some impacts may have resulted in a difference in absolute
values between the results from this experiment and what would be expected from
a commercial greenhouse. The exceptions which could have resulted in statistical
differences since they didn’t impact all treatments were: the higher temperature of
the bulbs for the Med ratio test treatment, the lamps failing (Red bottom, 19:1
Low and 10:1 Low), the lower irradiance level for the HPS treatment, the
increased powdery mildew on 5:1 Med (only in the first run of the experiment),
and edge effects from solar loads and the lamp falling and crushing a plant in the
19:1 Med. The lamp failures did not impact the results statistically. It is unsure
how high of an impact the higher bulb temperature (Med level) or the lower
irradiance in the early stages of the HPS has had on the final results but based on
the statistical models, no major statistical impact was noticed. The results of the
plants showed a larger difference than this lower light level would explain.
Page 72 of 115
6.7.1. First Run:
The plants were received as two plants per rockwool cube in the first run, which
was not as expected. Changes in irrigation and original planned spacing had to be
put in place, which meant that the plants had to be irrigated manually. The
computer control system was not automated at first, resulting in this manual
irrigation taking a week longer than expected. No impact was experienced in the
statistics, since all the plants experienced this similarly. Once the automatic
system was running, it took four days in order to get the correct amount of
irrigation (Savoura internal standard). Overwatering occurred during this period,
but due to the quick draining nature of the rockwool, excess water was flushed.
No impact was experienced, since all plants experienced this in the same fashion.
The nutrient solution supplied to the plants was in the wrong ratio (much higher in
iron and magnesium and lower in nitrates) for the first 4 weeks of the experiment.
No impact was noticed in the statistical analysis, since all plants were impacted in
the same manner. This could however partly explain why the total fruit mass
produced during this run was slightly lower than during the second run (no
statistical differences). A nutrient solution stronger in nitrogen during the early
stages of plant growth could increase plant mass by close to 50% (Ma et al. 2006).
The first pruning was delayed during the first two weeks, allowing for increased
biomass, flower and fruit counts. A secondary stem was kept in order to limit the
loss of fruiting, but the extra energy used for the second stem could lower the total
fruit production. All plants experienced the same, hence no impact was measured
in the statistics.
The red bottom LEDs stopped working twice due to water vapour condensing in
the housing (corrected after a week the first time and five days the next). Since
only one of the three lamps stopped working and the problem was fixed relatively
quickly, no impact was noticed. Some lamps became unplugged (the weight of
the lamps pulled down on the plugs at the beginning of the experiment) for a
Page 73 of 115
maximum of 2 days, but were fixed within a few hours, with no measurable
differences. 19:1 Low and 10:1 Low bulbs suffered from failures, and one light in
each section was off for 3 days and 7 days, respectively.
The LED systems originally used in the Medium intensity level (115 μmol m-2 s-1)
were shown to heat up to 88oC (due to the heat generated from the electric box
powering the LEDs as well as the heat accumulated from the sunlight on the black
casing). Since the lamps were set in an intercanopy setting, the leaves and stems
closest to the lamps experienced burning and blanching. As described by
Hogewoning et al. (2007), heat over 40oC close to leaves and stems will burn
them, lowering productivity. This would have lowered the total production from
the plants, since the burnt organs (stems or leaves) lost part of their functionality
(resulting in stunted growth or poor photosynthesis). This was corrected after 3
weeks, by using the same type of system as for the High intensity treatments (2
lamps side by side) which does not generate as much heat. No stems were
impacted and no statistical effect was measured. Heat waves during the summer,
where maximum temperatures were over 30ºC (for a period of 2 weeks, max
recorded was 38ºC) could be responsible for the flower abortion observed. The
heat of the summer has been shown to cause abortion of the flowers and fruit in
tomato plants (Sato et al. 2001). This abortion was observed for all plants in the
greenhouse, and could have an (<10%) impact on the total production from the
first run when compared to the second run. During the peak of the summer, the
fog system failed for 2 days, preventing the main cooling mechanism from
working. One of the fans also failed, and had to be replaced.
Powdery mildew was noticed on some plants, mostly in the 5:1 Med intensity
section. The powdery mildew originated in the 5:1 Med treatment, and slowly
contaminated all treatments in the greenhouse. The occurrence of powdery
mildew in a resistant strain of tomato plants such as the one used for this
experiment is rare, but can occur from exposure (Jones et al. 2001). Up to 25%
decrease in yield potential is possible if the plants are not treated, and the fruit
mass and volume can be decreased by up to 30% (Jones et al. 2001). This could
Page 74 of 115
explain why the 5:1 Med treatment had smaller fruit than the other top treatments
during the first run.
The measure of irradiance from the LEDs changed over time of output.
Corrections were made every two weeks, in order to make sure the voltage stayed
in the correct range. It was observed that one lamp in the 10:1 High lost voltage
(decreased light intensity) much more rapidly than others, and had to be checked
more frequently.
The HPS intensity was lower than expected at first. After 6 weeks this was
corrected, by adding a second lamp. A decrease in yield from the HPS plants is
expected from this lower intensity, but a conjecture was made for correction
(accounting for a 50% loss in production, which is higher than the loss
calculated). The solar light irradiance at different location in the greenhouse was
different. This has a low impact during the summer (when the sunlight was
directly overhead) but may have an impact during the late fall (when the sun was
lower). The southern treatments (5:1 low, 10:1 low and 19:1 low) could have
received more solar irradiance, as well as the eastern treatments (all 5:1
treatments, control and HPS). The difference in production is not expected to be
significant, and since the top treatments continued producing better during the
second run despite less favourable locations, the difference of solar light
irradiance was assumed to be not significant.
6.7.2. Second Run:
Plants were brought in as one double stemmed plant per rockwool (as expected
before the first run), and with six plants per section. This differs from the
previous run, but the statistics hold despite this inconsistency, since the number of
plants in a treatment is part of the error term for the statistical model. The plants
were brought in on December 31st in sub-zero weather, and some plants (primary
leaves) were severely damaged by the cold (Control plants, 19:1 Med and 10:1
Med sections were the most affected). This freezing damage could explain some
Page 75 of 115
of the reduction observed between these three treatments from the first and second
runs.
The heating system failed two days in February (explaining the extreme low
temperatures felt in the greenhouse), leaving extremely low temperatures in the
greenhouse. This cold impacted all plants in the greenhouse similarly, thus did
not have an impact on the statistics. The cooling system also broke down during
6 hours on the hottest day in February, as well as on the hottest days in March and
April (explaining the extreme high temperatures seen in the greenhouse). Both
the cold and the heat caused a slight amount of flower abortion, as seen during the
first run (albeit much less). The fan nearest the 10:1 low section would not close
when it was off, so cold air was allowed in. The same problem was observed with
the fan nearest the HPS section, during the coldest day in February. Both were
fixed within a few days. No impact was observed, except for the HPS section,
where 2 plants were observed to have been slightly damaged by the cold air.
They had stunted growth and less fruit than the others in the section. These 2
plants were randomly chosen to be harvested for the 70 day harvest, which
occurred less than a week after the symptoms were noticed, reducing the HPS
production values relative to the other 70 day harvests.
Similar problems from the first run were encountered, such as the incidence of the
sun (which is much lower during the winter). This might have a small effect on
production. The shade cloth has been partially removed in order to compensate
(no shade cloth above the plants). This could have resulted in the increased
results from 19:1 High, and 100% Red, since the added biomass growth due to
more red light permitted them to grow taller quicker. The lamps continued to
have voltage drops as experienced in the first run. This was corrected every two
weeks, by readjusting the voltage to the correct values for each lamp. The impact
was not statistically significant.
The 5:1 Med bulbs were unplugged three times for a total of 4 days. 19:1 Low
was broken for 15 days and was replaced. No impact on the statistics was shown.
The HPS lamp in the 50% section would shut off after a few hours during the first
two weeks of the run and was changed.
Page 76 of 115
Water lines would disconnect for some plants, due to the lines being older and
stiffer. Reconnecting the line and adding water fixed the problem, since each
plant had a water line, water could move between plants through the larger
rockwool cubes (no plant was ever completely dry).
One lamp in the 19:1 MED section fell due to a faulty wire holding it up. It
slightly damaged a plant on its way down and completely severed another. This
plant had to be harvested and was added in the 70 day report, making for four
plants in the 70 day harvest and two plants in the 120 day harvest.
6.8. Observations for future research
A number of observations were made during the experiment which would require
further research to understand what the driving influence of the results was.
A large number of clusters were seen turning into secondary stems or leaves. This
phenomenon is typically associated with a lack of light (Savoura unpublished
data), undermining tissue dissociation. This lack of light intensity should only
happen on aborted or small fruit clusters. However, it was noticed under LED
treatments with full fruit clusters, where the fruit continued to grow normally.
These results (Figure 20 below) do not agree with what Savoura expected, and
since it was only noticed on LED treatments, further research would need to be
performed to explain the phenomenon.
The lower and older leaves under the LED treatments had numerous spotting and
were more brittle, called intumescences (Rud 2009). These spots are typical of
Solanaceous crops, are cultivar dependent, and are characterized by individual
epidermal cells swelling and bursting. The process is non-reversible, and has
shown to appear on tomato plants from the Maxifort variety used as the base for
the plants (Rud 2009), but not on the Trust variety. Minor occurrences of
intumescences do not impact production or growth, but more severe cases lead to
necrosis of the leaves and can dampen the overall production by up to 50% in
extreme cases (Rud 2009).
Page 77 of 115
Figure 20: Two examples of viable clusters turning into secondary stems.
(harvested during the second run)
Intumescences were observed in all LED treatments, but not in the HPS or control
treatments. Higher intensity treatments were more severely impacted, but the
intumescences were relatively mild compared to what is shown in the literature
(Rud 2009). The worst cases are depicted in Figure 21, and should not have
impacted fruit production too heavily, since they were only observed on older
leaves and did not take up over 50% of the leaves. From these results, these spots
could be an interaction between the nutrient solution and the LED lighting. An
improved nutrient mix might have to be developed in order to compensate for this
interaction. Intumescences have been shown to occur when the moisture uptake
by the plant is higher than what the plant can evacuate (due to low vapor pressure
deficit or over watering of the leaves) (Rud 2009), which could suggest that the
Page 78 of 115
LED light promotes water uptake by the rooting system, since all plants were
submitted to the same watering conditions and only the LED treatments were
affected.
Figure 21: Leaf Burn. Observed on older leaves under the LED light fixtures
The regression curves resulted in strong fits for linear curves when applied to the
factorial models. From these results, higher ratios (i.e. 1:1 red to blue) and higher
intensities should be tested to determine the potential maximums.
Page 79 of 115
Powdery mildew was not observed during the second run but was observed in the
first run. To reduce contamination issues, a delay of over a month was allowed
between experimental runs to reduce the risk of disease carry-over. During the
first run, the 5:1 Med ratio had the most severe case of powdery mildew
potentially due to an over-exertion of the plant’s fruiting capabilities. Powdery
Mildew is rare in tomato strains which are resistant (such as the one used for this
experiment), but can occur due to exposure (Jones et al. 2001). Since no exposure
was expected, further research could determine if the overexertion theory is
correct.
Figure 22: Powdery Mildew. 5:1 Med section, during the first run.
For future research, it would be interesting to test a factorial comparison between
the 50%:50% LED:HPS and the 5:1 ratio to obtain regression curves across
increasing intensities of light.
Page 80 of 115
Future research would need to include quality (color, size, shape, taste, Brix, etc.)
with the yield data to provide a more complete evaluation of marketable fruit
quality beyond size that was reported in this document. An investigation into fruit
quality should be planned (the concentration of carotenoids in the fruit dependent
on the treatments), as well as a nutrient analysis of leaf samples, in order to
understand if a specific nutrient was also absorbed more readily in the plants
under LEDs.
A major issue from this experiment that needs to be addressed by all players
working with LED lights for horticultural purposes is the uniformity and accuracy
of light measurement techniques, equipment, and calculations for LEDs. Methods
and equipment used to measure other lighting sources (solar, HPS, incandescent,
etc.) are not consistent when measuring LED light arrays.
It can be theorized that a better sensor for LED arrays could be achieved by
placing multiple spectroradiometer type sensors in a bulb formation, in order to
get an accurate measurement of light coming from every direction, instead of a
bulb which refracts all light incoming into a single point.
Page 81 of 115
7. CONCLUSIONS
The most obvious conclusions that can be made from this experiment are that the
top five fruit producing light treatments were 5:1 High, 5:1 Med, 19:1 High, Red
Top and 50% LED:50%HPS. However, statistically no difference was measured
between any of the treatments, with the exception of the control which was
statistically different from all of the high producing treatments. The high
irradiance level was the top producer for each ratio, with the largest fresh and dry
vegetative biomass occurring within the 19:1 High LED. The top three treatments
for the largest mass of total fruit were 5:1 High, 5:1 Med, and 19:1 High. The top
three treatments of total number of fruit were 5:1 High, 5:1 Med, and 5:1 Low.
The top three treatments for marketable number of fruit were 50%:50%, Red Top
and 5:1 Med. The top three for total marketable fruit mass was 50%:50%, 5:1
High, and 19:1 High. The differences between these top three treatments (and
sometimes the top 6) were not statistically different. From this data, the top LED
treatment recommended is the 5:1 High, but the 50%:50% surpassed the 5:1 High
treatment for marketable fruit, meaning it can also be considered. From the
regression analysis of the factorial experiment, it can be reported that higher
levels of light increased production, while increased levels of red light (relative to
blue) resulted in less fruit. The regression also reported that increased levels of
light resulted in increased biomass, and increased red light (relative to blue)
resulted in more biomass. The treatment recommended for a newly built
greenhouse is the 5:1 High LED treatment (consistently found to be ahead in total
yields). For a pre-existing greenhouse however, the recommended treatment
should be the 50%:50% LED:HPS. This is due to this treatment not performing
significantly different than the 5:1 treatments, but not requiring complete removal
of the HPS lights, lowering renovation costs while still increasing overall
production significantly. The conclusions reached in this report were not based
on any financial consideration but only on the results from the test and time of
installation of the different systems within the greenhouse. Overall, it was shown
that the top LED treatments (5:1 High, 5:1 Med and 19:1 High) as well as the
Page 82 of 115
50%:50% treatment consistently outperformed the HPS treatment, and thus these
treatments can be considered an improvement over traditional HPS lighting for
greenhouses.
Page 83 of 115
8. REFERENCES
Adams, S.R., K.E. Cockshull, C.R.J. Cave. 2001. Effect of Temperature on the Growth and Development of Tomato Fruit. Ann. Bot. 88 (5):869-877
Agriculture Solutions Inc. 2011. Hortilux Ultra Ace Halide to HPS Conversion Bulb Catalogue.
Anstis, P.J.P., D.H. Northcot. 1974. Chlorophyll Accumulation by Callus Tissues of Glycine-Max. Planta 116:105-108
Apogee Instruments Inc. 2012. Apogee Pyranometer Technical Information. www.apogeeinstruments.co.uk/apogee-pyranometer-technical-information/
Apogee Instruments Inc. 2012. Apogee Quantum Sensor Technical Information. www.apogeeinstruments.co.uk/apogee-instruments-quantum-sensor-technical-information/
Argus Controls. 2010. Lighting and Lighting Controls in Greenhouses. Argus Control Systems Ltd pamphlet
Barro, F., P. De la Haba, J.M. Maldonado, A.G. Fontes. 1989. Effect of Light Quality on Growth, Contents of Carbohydrates, Protein and Pigments, and Nitrate Reductase Activity in Soybean Plants. J. Plant Physiol. 134:586-591
Barta, D.J., T.W. Tibbitts, R.J. Bula, R.C. Morrow. 1991. Evaluation of Light Emitting Diode Characteristics for a Space-Based Plant Irradiation Source. Adv. Space Res. 12(5): 141-149
Björn, L.O. 1967. Light Requirement for Different Steps in Development of Chloroplasts in Excised Wheat Roots. Physiologia Plantarum 20:483-499
Brazaitytė, A., P. Duchovskis, A. Urbonavičiūtė, S. Samuoliene, J. Jankauskienė, V. Kazenas, A. Kasiulevičiūtė-Bonakėrė, Z. Bliznikas, A. Novièkovas, K. Breivė, A. Žukauskas. 2009. After-Effect of Light Emitting Diodes Lighting on Tomato Growth and Yield in Greenhouse. Sodininkystė ir Daržininkystė 28:115-126
Brown, C s., A.C. Schuerger, J.C. Sager. 1995. Growth and Photomorphogenesis of Pepper Plants Under Red Light-Emitting Diodes with Supplemental Blue or Far-Red Lighting. J. Am. Soc. Hort. Sci. 120:808-813
Bula, R.J., R.C. Morrow, T.W. Tibbitts, D.J. Barta. 1991. Light-Emitting Diodes as a Radiation Source for Plants. Hortscience 26(2):203-205
Christie, J.M., W.R. Briggs. 2001. Blue Light Sensing in Higher Plants. J. Biol. Chem. 276 (15): 11457-11460
Clifford Evans, C. 1987. The Quantitative Analysis of Plant Growth, Blackwell Scientific Publications, University of California Press, Berkeley and Los Angeles, CA
Page 84 of 115
Cook, R., L. Calvin. 2005. Greenhouse Tomatoes Change the Dynamics of the North American Fresh Tomato Industry. USDA ERR-2:1-81
Cosgrove, D.J. 1981. Rapid Suppression of Growth by Blue Light: Occurrence, Time Course and General Characteristics. Plant Physiol. 67:584-590
Craford M.G. 1992. LEDs challenge the Incandescents. Circuits and Devices Mag., IEEE 8(5):24-29
Demmig-Adams, B., A.M. Gilmore, W.W. Adams, III. 1996. In Vivo Functions of Carotenoids in Higher Plants. FASEB J. 10:403-412
Deregibus, V.A., R.A. Sanchez, J.J. Casal. 1983. Effects of Light Quality on Tiller Production in Lollium spp. Plant Physiol. 72:900-902
Dorais, M., A. Gosselin, M.J. Trudel. 1990. Annual Greenhouse Tomato Production Under a Sequential Intercropping System Using Supplemental Light. Scientia Horticulturae 45:225-234
Frank, H.A., R.J. Cogdell. 1996. Carotenoids in Photosynthesis. Photochem. Photobiol. 63:257-264
French, C s. 1991. Action Spectra of Photosystems 1 and 2 Scaled by Comparison of Their Sums With Absorption Spectra of Photosynthetic Plants. Photosynthetica 25:67-74
Folta, K.M., L.L. Koss, R. McMorrow, H.H. Kim, J.D. Kenitz, R. Wheeler, J.C. Sager. 2005. Design and Fabrication of Adjustable Red-Green-Blue LED Light Arrays for Plant Research. BMC Plant Biol. 5 (17)
Fu, M., L. Yang, J. Zhang. 2011. Study of Light Emitting Diodes for the Application of Plant Growth in Green House.Intl. Conf. on Elec. Packing Tech. & High Density Packaging 1106-1110
Garner, W.W., H.A. Allard. 1920. Effect of the Relative Length of Day and Night and other Factors of the Environment on Growth and Reproduction in Plants. , J. Agric. Res. 18:553
Giacomelli, G.A., K.C. Ting, D.R. Mears. 1994. Design of a Single Truss Tomato Production System. Acta Horticulturae 361:77-84
Goins, G.D. N.C. Yorio, M.M. Sanwo, C s. Brown. 1997. Photomorphogenesis Photosynthesis, and Seed Yield of Wheat Plants Grown Under Red Light-Emitting Diodes (LEDs) With and Without Supplemental Lighting. J. Exptl. Bot. 48 (302):1407-1413
Goto, E. 2003. Effects of Light Quality on Growth of Crop Plants under Artificial Lighting. Environ. Control in Biol. 41(2):121-132
Heber, U., V.A. Shuvalov. 2005. Photochemical Reactions of Chlorophyll in Dehydrated Photosystem II: Two Chlorophyll Forms (680 and 700nm). Photosynthesis Research 84:85-91
Heliospectra. 2011. Revolutionary Light Systems for Plants. Presentation made to the Cleantech Forum in Amsterdam for 2011
Page 85 of 115
Heraut-Bron, V., C. Robin, C. Varlet-Grancher, A. Guckert. 2001. Phytochrome Mediated Effects on Leaves of White Clover: Consequences for Light Interception by the Plant Under Competition for Light. Annals of Botany 88:737-743
Heo, J., C. Lee, D. Chakrabarty, K. Paek. 2002. Growth Response of Marigold and Salvia Bedding Plants as Affected by Monochromic or Mixture Radiation Provided by a Light-Emitting Diode (LED). Plant Growth Regulaltion 38:225-230
Heuvelink, Ep. 2005. Tomatoes. Volume 13 of Crop Production Science in Horticulture Series. Cambridge, MA. CABI, 2005.
Hoagland, D.R., D.I. Arnon.1950. The water-Culture Method for Growing Plants Without Soil. Calif. Agr. Expt. Sta Circ. 347. Berkeley, Calif
Hoenecke, M.E., R.J. Bula, T.W. Tibbitts. 1992. Importance of ‘Blue’ Photon Levels for Lettuce Seedlings Grown Under Red-Light-Emitting Diodes. HortScience 27:427-430
Hogewoning, S.W., G. Trouwborst, G.J. Engbers, J. Harbinson, W. van Ieperen, J. Ruijsch, O. van Kooten. 2007. Plant Physiological Acclimation to Irradiation by Light-Emitting Diodes (LEDs).Acta Hort. 761:183-192
Holonyac, N., S.F. Bevacqua. 1962. Coherent (Visible) Light Emission from Ga(As1-xPx) Junctions. Appl. Phys. Lett. 1:82-83
Hopkins, W.G., N.P.A. Huner. 2004. Introduction to Plant Physiology, 3rd ed. John Wiley and Sons, Hoboken, NJ
Hori, Y., K. Nishidate, M. Nishiyama, K. Kanahama, Y. Kanayama. 2011. Flowering and Expression of Flowering-Related Genes Under Long Day Conditions with Light-Emitting Diodes. Planta 234:321-330
Inada, K. 1976. Action Spectra for Photosynthesis in Higher Plants. Plant and Cell Physiology 17:355-365
Jones, H., J.M. Whipps, S.J. Gurr. 2001. The Tomato Powdery Mildew Fungus Oidium neolycopersici. Molec. Plt. Patho. 2(6):303-309
Jones, R.W., R.W. Sheard. 1977. Effects of Blue and Red Light on Nitrate Reductase Level in Leaves of Maize and Pea Seedlings. Plt. Sci. Lett. 8:305-311
Koski, V.M., C s. French, J.H.C. Smith. 1951. The Action Spectrum for the Transformation of Protochlorophyll to Chlorophyll a in Normal and Albino Corn Seedlings. Archives of Biochemistry and Biophysics 31:1-17
Köst, H.P. 1988. CRC Handbook of Chromatography Plant Pigments. CRC Press, Inc.
Kubota, C. P. Chia. Z. Yang, Q. Li. 2012. Applications of Far-Red Light Emitting Diodes in Plant Production Under Controlled Environments. Acta Hort. 952:59-66
Page 86 of 115
Lefsrud, M.G., D.A. Kopsell, D.E. Kopsell, J. Curran-Celentano. 2005. Air Temperature Affects Biomass and Carotenoid Pigment Accumulation in Kale and Spinach Grown in a Controlled Environment. HortScience 40: 2026-2030
Lefsrud, M.G., D.A. Kopsell, D.E. Kopsell, J. Curran-Celentano. 2006. Irradiance Affects Biomass, Elemental Concentrations and Carotenoid Pigments in Kale and Spinach Grown in a Controlled Environment. Physiologia Planatarum 127:624-631
Lefsrud, M.G., D.A. Kopsell, C.E. Sams. 2008. Irradiance from Distinct Wavelength Light-Emitting Diodes Affect Secondary Metabolites in Kale. HortScience 43:2243–2244
Li-COR, Inc. 2012. Underwater PAR Measurement Catalogue, Li-COR Environment Division. Lincoln, NE
Lian, M.L., H.N. Murthy, K.Y. Paek. 2002. Effects of Light Emitting Diodes (LEDs) on the In-Vitro Induction and Growth of Bulblets of Lilium Oriental Hybrid ‘Pesaro’. Scientia Hort. 94:365-370
Lu, N., T. Maruo, M. Johkan, M. Hohjo, S. Tsukagoshi, Y. Ito, T. Ichimura, Y. Shinohara. 2012. Effects of Supplemental Lighting with Light-Emitting Diodes (LEDs) on Tomato Yield and Quality of Single-Truss Tomato Plants Grown at High Plant Density. Environ. Control. Biol. 50:63-74
Ma, C.H., T. Talb. 2006. Development of Starter Solution Technology as a Balanced Fertilization Practice in Vegetable Production. Acta Hort. 700:167-172
Mac Arthur, A.A., C. MacLellan, T.J. Malthus. Determining the FOV and directional response of field spectroradiometers. The 5th EARSeL Workshop on Imaging Spectroscopy. 2007
Massa, G.D., H.H. Kim, R.M. Wheeler, C.A. Mitchell. 2008. Plant Productivity in Response to LED Lighting. HortScience 43:1951-1956
Marcelis, L.F.M. ; Broekhuijsen, A.G.M.; Nijs, E.M.F.M.; Raaphorst, M.G.M. 2006. Quantification of the growth response of light quantity of greenhouse grown crops. Acta Horticulturae 711:97 – 103
Matsuda, R., K. Ohashi-Kaneko, K. Fujiwara, E. Goto, K. Kurata. 2004. Photosynthetic Characteristics of Rice Leaves Grown Under Red Light With or Without Supplemental Blue Light. Plant and Cell Physiology 45(12):1870-1874
Ménard, C., M. Dorais, T. Hovi, A. Gosselin. 2005. Developmental and Physiological Responses of Tomato and Cucumber to additional Blue Light. Acta Horticulturae 711:291-296
McAvoy, R. 1984. The Use of High Pressure Sodium Lights in Greenhouse Tomato Crop Production. Acta Horticulturae 148:877-888
McCree, K.J. 1971. Significance of Enhancement For Calculations Based on the Action Spectrum for Photosynthesis. Plant Physiology 49:704-706
Page 87 of 115
McCree, K.J. 1972a. The Action Spectrum, Absorption and Quantum Yield of Photosynthesis in Crop Plants. Agricultural Meteorology 9:191-216
McCree, K.J. 1972b. Test of Current Definitions of Photosynthetically Active Radiation against Leaf Photosynthesis Data. Agric. Meteorol., 10:443-453
Menard, C.,T. Hovi, A. Gosselin, M. Dorais. 2006. Developmental and Phsyiological Responses of Tomato and Cucumber to Additional Blue Light. Acta Hort. 711:291-296
Miyashita, Y., Y. Kitaya, T. Kozai. 1995. Effects of Red and Far-Red Light on the Growth and Morphology of Potato Plantlets In-Vitro: Using Light Emitting Diode as a Light Source for Micropropagation. Acta Hort. 393:189-194
Mortensen, L.M., E. Stromme. 1987. Effects of Light Quality on Some Greenhouse Crops. Scientia Horticulturae 33:27-36
Moss, D.N. 1964. Optimum Lighting of Leaves. Crop Science 4(2):131-136
Mukai, T., M. Yamada, S. Nakamura. 1999. Characteristics of InGaN-Based UV/Blue/Green/Amber/Red Light-Emitting Diodes. Jpn. J. Appl. Phys.38: 3976–3981
Ogawa, T., Y. Inoue, M. Kitajima, K. Shibata. 1973. Action Spectra for Biosynthesis of Chlorophylls a and b and β-carotene. Photochemistry and Photobiology 18:229-235
Ohashi-Kaneko, K., M. Takase, N. Kon, K. Fujiwara, K. Kurata. 2007. Effect of Light Quality on Growth and Vegetable Quality in Leaf Lettuce, Spinach and Komatsuna. Environ. Control Biol. 45 (3):189-198
Okamoto, K. T. Yanagi, S. Takita. 1996. Development of Plant Growth Apparatus Using Blue and Red LED as Artificial Light Source. Acta Hort. 440:111-116
Poudel, P.R., I. Kataoka, R. Mochioka. 2008. Effect of Red- and Blue- Light-Emitting Diodes on Growth and Morphogenesis of Grapes. Plant Cell Tiss. Organ. Cult. 92: 147-153
Rud, N.A. 2009 Environmental Factors Influencing the Physiological Disorders of Edema on Ivy Geranium (Pelargonium Peltatum) and intumescences on tomato (Solanum Lycopersicum). Thesis Kansas St. U., Manhattan, KS. 2009. K-State Electronic Theses, Dissertations, and Reports: 2004 -. Web, 01 Dec 2012
Sager, J.C., J.L. Edwards, W.H. Klein. 1982. Light Energy Utilization Efficiency for Photosynthesis. Transactions ASAE 25:1737-1746
Sandmann, G. 2001. Carotenoid Biosynthesis and Biotechnological Application. Arch. Biochem. Biophys. 385:4-12
Schuerger, A.C., C s. Brown, E.C. Stryjewski. Anatomical Features of Pepper Plants (Capsicum Annuum L.) Grown Under Red Light-Emitting Diodes Supplemented with Blue and Far-Red Light. Annals of Botany 79:273-282
Statistics Canada. 2005. Greenhouse, Sod and Nursery Industries Catalogue # 22-202-x18
Page 88 of 115
Statistics Canada. 2011. Greenhouse, Sod and Nursery Industries Catalogue # 22-202-x
Sato, S., M.M. Peet, R.G. Gardner. 2001. Formation of parthenocarpic fruit, undeveloped flowers and aborted flowers in tomato under moderately elevated temperatures. Scientiae Holticulturae 90 (3-4):243-254
Taiz L., E. Zeiger. Plant Physiology, 2nd ed. Sinauer Associates, Inc. Sunderland, MA. 1998.
Tamulaitis, G., P. Duchovskis, Z. Bliznikas, K. Brievė, R. Ulinskaitė, A. Brazaitytė, A. Novičkovas, A. Žukauskas. 2005. High Powered Light Emitting Diode Based Facility for Plant Cultivation. J. Phys. D: Appl. Phys. 38:3182-3187
Tennessen, D.J., E.L. Singsaas, T.D. Sharkley. 1994. Light-Emitting Diodes as a Light Source for Photosynthesis Research. Photosynthesis Research 39:85-92
Thayer, S s., O. Bjorkman. 1992. Carotenoid Distribution and Deepoxidation in Thylakoid Pigment-Protein Complexes from Cotton Leaves and Bundle-Sheath Cells of Maize. Photosynthesis Research 33:213-235
Virgin, H.I. 1993. Effectiveness of Light of Different Wavelengths to Induce Chlorophyll Biosynthesis in Rapidly and Slowly Greening Tissue. Physiologia Plantarum 89:761-766
Walters, R.G., P. Horton. 1995. Acclimation of Arabidopsis Thaliana to the Light Environment: Regulation of Chloroplast Composition. Planta 197(3):475-481
Wu, M.C., C.Y. Hou, C.M. Jiang, Y.T. Wang, C.Y. Wang, H.H. Chen, H.H. Chang. 2007. A Novel Aapproach of LED Light Radiation Improves the Antioxidant Activity of Pea Seedlings. Food Chem. 101:1753-1758
XiaoYing, L., G. ShiRong, X. ZhiGang, J. XueLei. 2011. Regulation of Chloroplast Ultrastructure, Cross-Section Anatomy of Leaves, and Morphology of Stomata of Cherry Tomato by Different Light Irradiations of Light-Emitting Diodes. HortScience 46(2):217-221
XiaoYing, L., G. ShiRong, C. TaoTao, X. ZhiGang, J. XueLei. 2012. Regulation of the Growth and Photosynthesis of Cherry Tomato Seedlings by Different Light Irradiations of Light Emitting Diodes (LED). African J. of Biotech. 11(22):6169-6177
Xie, W. 1992. Room Temperature Blue Light Emitting p‐n Diodes from Zn(S,Se)‐Based Multiple Quantum Well Structures. Applied Physics Letters 60 (16):1999-2001
Yorio, N.C., G.D. Goins, H.R. Kagie,R.M. Wheeler, J.C. Sager. 2001. Improving Spinach, Radish, and Lettuce Growth Under Red Light-Emitting Diodes (LEDs) with Blue Light Supplementation. HortScience 36(2): 380-383
Page 89 of 115
APPENDIX A: Formatted Data
Figure A1: Total marketable fruit for the second run for the tomato plants.
Figure A2: Total marketable fruit for the first run for the tomato plants.
0
10
20
30
40
50
60
70
80
90
5:1
LO
W
5:1
MED
5:1
HIG
H
10
:1 L
OW
10
:1 M
ED
10
:1 H
IGH
19
:1 L
OW
19
:1 M
ED
19
:1 H
IGH
LED
RED
TO
P
RED
BO
T
50
/50
HP
S
CO
NTR
OL
Seco
nd
ru
n t
ota
l nu
mb
er
of
mar
keta
ble
fru
it
70 day
120 day
Total
0
5
10
15
20
25
30
35
40
45
50
5:1
LO
W
5:1
MED
5:1
HIG
H
10
:1 L
OW
10
:1 M
ED
10
:1 H
IGH
19
:1 L
OW
19
:1 M
ED
19
:1 H
IGH
LED
RED
TO
P
RED
BO
T
50
/50
HP
S
CO
NTR
OL
Firs
t ru
n t
ota
l nu
mb
er
of
mar
keta
ble
fru
it
70 day
120 day
Total
Page 90 of 115
Figure A3: Total mass of marketable fruit for the second run for the tomato plants.
Figure A4: Total mass of marketable fruit for the first run for the tomato plants.
0
2000
4000
6000
8000
10000
12000
14000
Seco
nd
ru
n t
ota
l mas
s o
f m
arke
tab
le f
ruit
(g)
70 day
120 day
Total
0
1000
2000
3000
4000
5000
6000
Firs
t ru
n t
ota
l mas
s o
f m
arke
tab
le f
ruit
(g)
70 day
120 day
Total
Page 91 of 115
Figure A5: Total fruit count during the second run
Figure A6: Total fruit count during the first run. * The HPS conjecture is overestimated; since it keeps into account the 20% increase during the 70 day run while for the 120 day run it would be closer to 5% increase.
0
20
40
60
80
100
120
140
160
180
200
Seco
nd
ru
n t
ota
l fru
it c
ou
nt
70 day
120 day
Total
0
20
40
60
80
100
120
140
160
180
200
5:1
LO
W
5:1
MED
5:1
HIG
H
10
:1 L
OW
10
:1 M
ED
10
:1 H
IGH
19
:1 L
OW
19
:1 M
ED
19
:1 H
IGH
LED
RED
TO
P
RED
BO
T
50
/50
HP
S
CO
NTR
OL
Fruit Count
Firs
t ru
n t
ota
l fru
it c
ou
nt
70 day
120 day
Total
HPS Conjecture
Page 92 of 115
Figure A7: Total fruit mass during the second run
Figure A8: Total fruit mass during the first run
* The HPS conjecture is overestimated; since it keeps into account the 20% increase during the 70 day run while for the 120 day run it would be closer to 5% increase.
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
Seco
nd
ru
n t
ota
l fru
it m
ass
(g)
70 day
120 day
Total
0
2000
4000
6000
8000
10000
12000
5:1
LO
W
5:1
MED
5:1
HIG
H
10
:1 L
OW
10
:1 M
ED
10
:1 H
IGH
19
:1 L
OW
19
:1 M
ED
19
:1 H
IGH
LED
RED
TO
P
RED
BO
T
50
/50
HP
S
CO
NTR
OL
Fruit Weight
Firs
t ru
n t
ota
l fru
it m
ass
(g)
70 day
120 day
Total
HPS Conjecture
Page 93 of 115
Figure A9: Total plant fresh biomass (excluding fruit) for the second run
Figure A10: Total plant fresh biomass (excluding fruit) for the first run
0
5000
10000
15000
20000
25000
30000
Seco
nd
ru
n t
ota
l fre
sh b
iom
ass
(g)
70 day
120 day
Total
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
Firs
t ru
n t
ott
al f
resh
bio
mas
s (g
)
70 day
120 day
Total
Page 94 of 115
Figure A11: Total plant dried biomass (excluding fruit) for the second run
Figure A12: Total plant dried biomass (excluding fruit) for the first run.
APPENDIX B: Statistical Data (Box Plots)
0
200
400
600
800
1000
1200
1400
1600
1800
Seco
nd
ru
n t
ota
l dry
bio
mas
s (g
)
70 day
120 day
Total
0
200
400
600
800
1000
1200
1400
5:1
LO
W
5:1
MED
5:1
HIG
H
10
:1 L
OW
10
:1 M
ED
10
:1 H
IGH
19
:1 L
OW
19
:1 M
ED
19
:1 H
IGH
LED
RED
TO
P
RED
BO
T
50
/50
HP
S
CO
NTR
OL
Dry Biomass
Firs
t ru
n t
ota
l dry
bio
mas
s (g
)
70 day
120 day
Total
Page 95 of 115
For the following Figures, the mean is represented by a diamond shape, the median as a horizontal line, the lower quartile and the upper quartile as the ends of the boxes, the whiskers as the lowest and highest value observed, and circles for outliers.
Figure B1: Total number of marketable fruit per plant for the 120 day harvests
Page 96 of 115
Figure B2: Total marketable fruit mass per plant, for the 120 day harvests
Figure B3: Total number of fruit per plant, for the 120 day harvests
Page 97 of 115
Figure B4: Total mass of fruit per plant, for the 120 day harvests
Figure B5: Total fresh biomass per plant for the 120 day harvests
Page 98 of 115
Figure B6: Total dry biomass per plant for the 120 day harvests
Figure B7: Ratio of dry to fresh biomass production
Page 99 of 115
Figure B8: Ratio of fruit mass to fresh biomass for the 120 day data
Figure B9: Ratio of marketable fruit mass to fresh biomass for the 120 day data
Page 100 of 115
Figure B10: Total number of red fruit per plant, for the 120 day experiment
Page 101 of 115
APPENDIX C: Raw Data Table C1: Total fruit data (green and red fruit for the 2 replications (4 plants harvested after 70 and 120 days for the first one, and 3 plants harvested at 70 and 120 days for the second).
Fruit Count
Total Fruit Mass
Average Mass
Fruit Count
Total Fruit Mass
Average Mass
Fruit Count
Total Fruit Mass
Average Mass
Run 1 5:1 LOW 5:1 MED 5:1 HIGH 70 day 58 1549.2 26.71 92 3355.6 36.47 84 2541.8 30.26 120 day 115 7998.8 69.55 100 7896.5 78.97 107 7866.9 73.52
Total 173 9548 55.19 192 11252.1 58.6 191 10408.7 54.5 Run 2 70 day 46 1327.6 28.86 33 1016.1 30.79 55 1909.3 34.71 120 day 122 11727 96.12 133 13557.3 101.93 139 14908.5 107.26
Total 168 13054.6 77.71 166 14573.4 87.79 194 16817.8 86.69 Run1 10:1 LOW 10:1 MED 10:1 HIGH 70 day 13 342 26.31 46 2682.6 58.32 54 2625.3 48.62 120 day 75 4277.6 57.03 60 4131.1 68.85 107 7033.3 65.73
Total 88 4619.6 52.5 106 6813.7 64.28 161 9658.6 59.99 Run 2 70 day 34 849.7 24.99 52 1927.2 37.06 44 1612.2 36.64 120 day 84 7456.2 88.76 97 9696.1 99.96 99 10293.2 103.97
Total 118 8305.9 70.39 149 11623.3 78.01 143 11905.4
83.25
Page 102 of 115
Run 1 19:1 LOW 19:1 MED 19:1 HIGH 70 day 12 526.2 43.85 28 1055.8 37.71 54 2026.2 37.52 120 day 41 2470.1 60.25 71 4286.6 60.37 64 4560.5 71.26
Total 53 2996.3 56.53 99 5342.4 53.96 118 6586.7 55.82 Run 2 70 day 39 1138 29.18 21 582.8 27.75 69 2483.8 36
16 489.5 30.59 120 day 105 9432.4 89.83 85 8895.7 104.66 128 14885 116.29
Total 105 9432.4 89.83 101 9385.2 92.92 128 14885 116.29 Run 1 LED RED TOP RED BOT 70 day 46 1465.8 31.87 34 1465.7 43.11 42 1427.1 33.98 120 day 114 7225.5 63.38 78 6605.9 84.69 74 4600.8 62.17
Total 160 8691.3 54.32 112 8071.6 72.07 116 6027.9 51.96 Run 2 70 day 56 1967 35.13 26 1072.1 41.23 32 759 23.72 120 day 94 10493.8 111.64 130 14170.4 109.00 120 13941.8 116.18
Total 150 12460.8 83.07 156 15242.5 97.71 152 14700.8 96.72 50/50 HPS CONTROL
70 day 42 1608.3 38.29 16 378.3 23.64 0 0 --- 120 day 64 5452.7 85.2 37 3018 81.57 13 1190.8 91.6
Total 106 7061 66.61 53 3396.3 64.08 13 1190.8 91.6 Run 2 70 day 54 1796.6 33.27 4 88 22 0 0 --- 120 day 109 14302.5 131.22 92 9069.3 98.58 10 624.5 62.45
Total 163 16099.1 98.77 96 9157.3 95.39 10 624.5 62.45
Page 103 of 115
A fourth plant had to be weighed in the 19:1 MED section, since a lamp fell and killed it. The values in red are the sum of all four plants, while the values underneath were found using the best, the worst and the average of the two middle plants. (The 120 data for 19:1 Med therefore only includes 2 plants) Table C2: Red fruit harvested data during the 120 day experiment (4 plants harvested at 70 days, 4 at 120 days). Red fruit were harvested from all plants when the first red pigmentation was noticed on the green fruit. No red fruit was harvested at the 70 day mark during the run 2.
Run 1 5:1 LOW
5:1 MED 5:1 HIGH
70 day 3 9 2 120 day 78 72 73
Total 81 81 75 Run 2
120 day 13 20 39
Run 1 10:1 LOW
10:1 MED 10:1 HIGH
70 day 0 7 6 120 day 36 45 72
Total 36 52 78 Run 2
120 day 2 16 18
Run1 19:1 LOW
19:1 MED 19:1 HIGH
70 day 5 0 1 120 day 17 44 46
Total 22 44 47 Run 2
120 day 10 15 21
Run 1 LED RED TOP RED BOT
70 day 7 5 4 120 day 73 59 35
Total 80 64 39 Run 2
120 day 10 28 7
Run 1 50/50 HPS CONTROL 70 day 2 0 0 120 day 48 19 7
Total 50 19 7 Run 2
120 day 23 2 0
Page 104 of 115
Table C3: Greater than 90 g fruit data during the first (4 plants harvested at 70 days and 4 at 120 days) and the second (3 plants harvested at 70 days) replications. The total number of fruit (count) from each treatment and mass of these fruit is recorded.
Run 1 5:1 LOW
5:1 MED 5:1 HIGH
Count 37 39 33 Mass 4057.4 4399.2 4049.4 Run 2 Count 55 70 70 Mass 8431.5 10745.4 12167.4
Run 1 10:1 LOW
10:1 MED 10:1 HIGH
Count 17 19 21 Mass 1900.2 2105.2 2315.7 Run 2 Count 36 54 63 Mass 5246.7 8138.6 8576
Run 1 19:1 LOW
19:1 MED 19:1 HIGH
Count 13 13 27 Mass 1404.7 1349.2 2905.1 Run 2 Count 47 44 81 Mass 7127.4 7231.1 13162.8
Run 1 LED RED TOP RED BOT
Count 26 32 20 Mass 2803.2 3868.4 2433.9 Run 2 Count 59 79 62 Mass 9539 11535.6 11856.5 Run 1 50/50 HPS CONTROL Count 37 14 5 Mass 4166.6 1673.5 735.4 Run 2 Count 77 48 10 Mass 12794.8 7442.4 624.5
Page 105 of 115
Table C4: Biomass data (fresh and dry) during the first (4 plants harvested at 70 days and 4 at 120 days) and the second (3 plants harvested at 70 days) replications.
Fresh Mass
Average Plant
Dry Mass
Average Plant
Fresh Mass
Average Plant
Dry Mass
Average Plant
Fresh Mass
Average Plant
Dry Mass
Average Plant
Run 1 5:1 LOW 5:1
MED 5:1 HIGH
70 day 3857.2 964.3 279.4 69.9 4961.4 1240.3 354.7 88.7 5060.3 1265.1 402.3 100.6 120 day 7001.1 1750.3 526.1 131.5 7838.7 1959.7 569.3 142.3 10016.9 2504.2 713 178.3
Total 10858.3 1357.3 805.5 100.7 12800.1 1600 924 115.5 15077.2 1884.6 1115.3 139.4 Run 2 70 day 6129.3 2043.1 423.1 141 5501.6 1833.867 358.9 119.6 6414.5 2138.167 414.1 138 120 day 8983.2 2994.4 755.7 251.9 14587.8 4862.6 1141.3 380.4333 11827.3 3942.433 1043.8 347.9333
Total 15112.5 2518.75 1178.8 196.47 20089.4 3348.23 1500.2 250.03 18241.8 3040.30 1457.9 242.98
Run 1 10:1 LOW 10:1
MED 10:1 HIGH
70 day 4212.2 1053.1 285.2 71.3 5142.9 1285.7 350.8 87.7 4585 1146.3 371.8 93 120 day 10000 2500 709.1 177.3 8794.1 2198.5 583.8 146 10677.5 2669.4 825.9 206.5
Total 14212.2 1776.5 994.3 124.3 13937 1742.1 934.6 116.8 12832.8 1604.1 992.9 124.1 Run 2 70 day 5341.7 1780.567 407 135.7 6365.9 2121.967 348 116 6152.9 2050.967 368.9 123 120 day 14826.3 4942.1 1105.9 368.6333 14357.9 4785.967 1221.6 407.2 11208.6 3736.2 829.8 276.6
Total 20168 3361.33 1512.9 252.15 20723.8 3453.97 1569.6 261.60 17361.5 2893.58 1198.7 199.78
Run 1 19:1 LOW 19:1
MED 19:1 HIGH
70 day 4117.2 1029.3 268.2 67.1 4163 1040.8 282.8 70.7 5995.5 1498.9 449.9 112.5 120 day 8371.2 2092.8 580.9 145.2 11868.2 2967.1 850.4 212.6 10939.4 2734.9 836.3 209.1
Page 106 of 115
Total 12488.4 1561.1 849.1 106.1 16031.2 2003.9 1133.2 141.7 16934.9 2116.9 1286.2 160.8 Run 2 70 day 5653.1 1884.367 355.7 118.6 7653.8 1913.45 544.3 136.1 6404.7 2134.9 449.1 149.7
4602.5 1534.167 334.9 111.6 120 day 15782.4 5260.8 1284.2 428.0667 10759.7 3586.567 906.3 302.1 15553.3 5184.433 1092 364
Total 21435.5 3572.58 1639.9 273.32 18413.5 3068.92 1450.6 241.77 21958 3659.67 1541.1 256.85
Run 1 100% LED RED
TOP RED BOT
70 day 5216.1 1304 365.8 91.5 4783.2 1195.8 302.9 75.7 3915.7 978.9 309.4 77.4 120 day 10395.8 2599 763.9 191 9753.3 2438.3 671.9 168 9951.6 2487.9 640.5 160.1
Total 15611.9 1951.5 1129.7 141.2 14536.5 1817.1 974.8 121.9 13867.3 1733.4 949.9 118.7 Run 2 70 day 8058 2686 546 182 4477.8 1492.6 295 98.3 5827.7 1942.567 426.1 142 120 day 13233 4411 1088.6 362.8667 12250.5 4083.5 961.4 320.4667 18947.2 6315.733 1108.9 369.6333
Total 21291 3548.50 1634.6 272.43 16728.3 2788.05 1256.4 209.40 24774.9 4129.15 1535 255.83 Run 1 50/50 HPS CONT 70 day 4333 1083.3 311.6 77.9 3978.3 994.6 251.9 63 1590.1 397.5 89.8 22.5 120 day 9577.1 2394.3 667.7 166.9 6783 1695.8 441.8 110.5 3879.7 969.9 231.3 57.8
Total 13910.1 1738.8 979.3 122.4 10761.3 1345.2 693.7 86.7 5469.8 683.7 321.1 40.1 Run 2 70 day 5688.9 1896.3 365 121.7 4118.4 1372.8 306.5 102.2 1245.9 415.3 68.9 23 120 day 15649 5216.333 1037 345.6667 12774.6 4258.2 1414.1 471.3667 8118 2706 604.5 201.5
Total 21337.9 3556.32 1402 233.67 16893 2815.50 1720.6 286.77 9363.9 1560.65 673.4 112.23
Page 107 of 115
A fourth plant had to be weighed in the 19:1 MED section at the 70 day mark of the second run, since a lamp fell and killed it. The values in red are the sum of all four plants, while the values underneath were found using the best, the worst and the average of the two middle plants.
Page 108 of 115
Table C5: Fruit and flower counts for both replications. The total counts are provided
Date 5:1 LOW
5:1 MED
5:1 HIGH
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 27 10 25 36 26 15 9-Aug 38 14 61 42 62 29
12-Aug 58 25 49 47 56 44
16-Aug 46 39 38 68 30 62
25-Aug 44 76 34 102 36 99
1-Sep 24 115 14 148 32 148 28-Sep 13 74 22 72 19 59 Run 2 8-Feb 103 20 82 10 111 28 28-Feb 80 78 86 92 119 130 27-Apr 52 122 138 133 126 139
10:1 LOW
10:1 MED
10:1 HIGH
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 43 2 40 29 23 12 9-Aug 41 4 46 32 66 25
12-Aug 38 5 37 37 38 34
16-Aug 19 8 24 42 34 54
25-Aug 34 44 15 69 30 101
1-Sep 31 44 10 95 12 148 28-Sep 19 44 5 29 30 64 Run 2 8-Feb 77 5 106 19 111 32 28-Feb 66 56 84 81 98 90 27-Apr 235 84 290 97 139 99
19:1 LOW
19:1 MED
19:1 HIGH
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 36 3 34 23 32 16 9-Aug 28 9 41 29 49 17
12-Aug 23 10 26 31 36 34
16- 8 12 24 36 31 42
Page 109 of 115
Aug 25-Aug 18 32 27 47 26 69
1-Sep 16 34 9 80 16 92 28-Sep 10 27 29 25 15 43 Run 2 8-Feb 81 8 47 8 93 28 28-Feb 73 72 68 38 88 104 27-Apr 165 105 189 85 167 128
LED
RED TOP
RED BOT
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 72 17 29 11 23 13 9-Aug 33 31 23 18 29 11
12-Aug 27 34 25 18 26 15
16-Aug 20 46 21 17 23 40
25-Aug 32 95 30 66 19 56
1-Sep 33 114 20 82 16 72 28-Sep 30 65 12 70 23 40 Run 2 8-Feb 101 13 80 31 81 4 28-Feb 101 90 63 91 111 64 27-Apr 331 94 203 130 231 120
50/50 HPS
CONTROL
Run 1 Flower Fruit Flower Fruit Flower Fruit 2-Aug 22 12 2 0 2 0 9-Aug 27 16 27 0 13 0
12-Aug 34 20 35 3 13 3
16-Aug 27 27 17 4 6 2
25-Aug 42 64 34 30 19 10
1-Sep 24 85 38 33 12 9 28-Sep 24 40 27 49 18 11 Run 2 8-Feb 90 25 38 2 0 0 28-Feb 87 112 57 15 4 0 27-Apr 266 109 138 92 86 10
Page 110 of 115
APPENDIX D: Weather and Lighting Data
Table D1: Temperature data during the second run
5:1 LOW 5:1 MED 5:1 HIGH Temp SD MAX MIN Temp SD MAX MIN Temp SD MAX MIN January 16.8 2.7 24.4 10.4 16.5 2.3 23.4 5.3 16.2 2.4 23.6 10.5 February 15.8 4.1 28.6 1.7 16 4.2 28.2 2 15.1 3.7 27.1 1.8 March 17.4 3.7 34.5 9.7 17.3 3.5 29.7 9.9 17.3 3.4 28.9 8.7 April 17.8 4 34.3 10.6 17.4 3.6 32.6 10.9 18.2 3.9 39.2 10.9 10:1 LOW 10:1 MED 10:1 HIGH Temp SD MAX MIN Temp SD MAX MIN Temp SD MAX MIN January 16.2 2.4 23.6 10.5 16.6 2.5 23.6 8.7 16 2.3 22.5 5.9 February 15.1 3.7 27.1 1.8 15.7 4.2 28.9 1.8 15.7 4 26.9 2.5 March 17.9 3.9 33.5 9.2 17.6 3.6 29.5 10 17.2 3.7 30.9 10.1 April 18.4 4.2 40.2 9.1 18.5 4 39.3 10.7 17.4 3.6 32.5 10.8 19:1 LOW 19:1 MED 19:1 HIGH Temp SD MAX MIN Temp SD MAX MIN Temp SD MAX MIN January 16.7 2.6 24.3 10.4 16.9 2.6 25.3 10.6 16.1 2.3 22.1 5 February 16.1 4.2 28.8 1.6 17.2 2.3 20.4 2.5 15.8 4 27 2.2 March 17.6 3.9 34.5 9.6 18.2 4 31.2 10.3 17 3.5 29.2 9.5 April 18.3 4.3 36 10.5 18.6 4.1 38.4 10.7 17.3 3.6 32.6 10.6
LED RED TOP RED BOT Temp SD MAX MIN Temp SD MAX MIN Temp SD MAX MIN January 16.4 2.4 23.6 10.1 16.3 2.3 22.8 5.8 16.7 2.6 22.9 10.3 February 16.6 4 29.3 2.4 16.2 4.1 28.1 2.2 15.6 3.8 27.2 2 March 17.7 3.9 31.1 9.6 17.4 3.7 30.7 9.9 16.8 3.4 29.3 9.3 April 17.4 4.3 39.5 10.8 17.7 3.9 33.9 11 16.7 3.3 31.6 10.1 50/50 HPS CONTROL Temp SD MAX MIN Temp SD MAX MIN Temp SD MAX MIN January 16.2 2.4 23.2 5 16.7 2.8 26.3 7.3 16.7 2.5 24.9 11 February 15.9 4 28.1 2 15.9 4.3 31.5 0.9 15.9 3.9 29.4 2 March 17.2 3.6 31.7 9.4 16.9 3.6 31.1 2.5 16.8 3.1 29.2 9.7 April 17.6 3.8 33.7 10.6 16.3 3.2 31.6 10.4 17.3 3.4 35.6 10
Page 111 of 115
Table D2: Temperature data during the first run 5:1 LOW 5:1 MED 5:1 HIGH
Temperature St Dev Max Min Temperature St
Dev Max Min Temperature St Dev Max Min
21.4 2.1 33.6 11.2 23.1 2.0 33.1 16.1 21.3 2.1 32.9 11.1 10:1 LOW 10:1 MED 10:1 HIGH
Temperature St Dev Max Min Temperature St
Dev Max Min Temperature St Dev Max Min
22.1 1.8 33.3 9.9 23.2 2.2 32.6 16.4 20.7 2.8 32.2 10.5 19:1 LOW 19:1 MED 19:1 HIGH
Temperature St Dev Max Min Temperature St
Dev Max Min Temperature St Dev Max Min
21.9 1.7 35.2 10.8 23.0 2.1 31.7 16.3 21.4 2.1 35.2 11.1 LED RED TOP RED BOT
Temperature St Dev Max Min Temperature St
Dev Max Min Temperature St Dev Max Min
21.2 2.2 31.2 10.5 20.9 2.5 35.8 10.4 21.5 2.6 38.7 10.8 50/50 HPS CONTROL
Temperature St Dev Max Min Temperature St
Dev Max Min Temperature St Dev Max Min
21.2 2.2 33.5 11.1 21.3 3.0 38.5 9.9 21.0 3.1 36.9 10.7
Page 112 of 115
Table D3: End of second run light map data (May 1st 2012)
Zone Spectroradiometer Li-COR Daily Light Integral (mol day-1)
% of Artificial
Light Average Vertical
Average 45º angle Average Average
5:1 LOW 98.2 152.9 125.6 118.1 8.5 37.3 5:1 MED 119.1 154.2 136.7 152.4 9.8 40.6 5:1 HIGH 130.3 154.2 142.3 190.9 11.5 44.5 10:1 LOW 98.6 126.6 112.6 113.7 9.1 38.8 10:1 MED 113.2 148.7 130.9 134.3 9.9 40.9 10:1 HIGH 125.2 151.4 138.3 160.7 13.1 47.9 19:1 LOW 98.4 143.9 121.2 112.8 9 38.7 19:1 MED 114.4 141.7 128.1 139.7 10.4 42.2 19:1 HIGH 125.3 144 134.7 177.6 12 45.7
LED 131 163 147 206.6 12.3 46.2 0.5 114.7 142 128.3 110.9 9.1 38.8
Red TOP 119.3 150.2 134.8 147.6 11.5 44.6 Red BOT 104 137.1 120.6 115.1 7.9 35.5 Control 0.4 0.1 0.3 0.6 0 0.1
HPS Light Maximum Irradiance Middle (1 m) 38 38 37.3 4.4 23.4 Height (1.5 m) 96 96 127.8 9.2 39.2 Top (30 cm from lamp) 162.6 162.6 447.7 21.9 60.5
Page 113 of 115
Table D4: Initial light map data (August 11th 2011), with correction from updated calibration (next page)
LED Light Maps Average Vertical
Average 45º
angle Horizontal Average
Daily Light Integral (mol
day-1)
% of Artificial
Light
5:1 LOW 51.625 71.875 13.75 61.75 5.34 27.2 5:1 MED 61 87 14.25 74 6.39 30.9 5:1 HIGH 70.125 106.25 29.875 88.125 7.61 34.7 10:1 LOW 73.875 97.25 20.625 85.5 7.39 34.1 10:1 MED 78.875 96.75 33.875 87.875 7.59 34.7 10:1 HIGH 83.875 105.25 23 94.625 8.18 36.4 19:1 LOW 67.375 91.625 19.5 79.5 6.87 32.4 19:1 MED 71.125 93.375 25.75 82.25 7.11 33.2 19:1 HIGH 91.25 104.375 22.375 97.875 8.46 37.2
LED 69.75 97.25 116.75 83.5 7.21 33.5 RED TOP 78 102.125 73.75 90 7.78 35.2 RED BOT 74 88.25 38 81.125 7.01 32.9
50/50 58.5 79.375 39 69 5.96 29.4 HPS Light Starting Irradiance Maximum Irradiance
Floor 13.5 10.875 Middle (1 m) 42.625 37.5
Plant height(1.5 m) 30 75 6.48 31.2 Top (30 cm from
lamp) 260 175
Light map was for the LED bulbs in a vertical and 45 degree angle at a distance of 30 cm from the bulb in the center of the chamber at three locations for each vertical and angled measurement (full darkness, with other bulbs on). The HPS was measured at three heights (top of rock wool cube, middle of the chamber (approximate height of the seedlings), and top of the chamber (30 cm from the bulb in a vertical orientation). All measurements were taken with a spectroradiometer, corrected from calibration on Dec 15, 2011.
Page 114 of 115
Table D5: Initial summary of light map data (August 11th 2011), without calibration correction
5:1 LOW 5:1 MED 5:1 HIGH PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) AVERAGE 49.4 AVERAGE 59.2 AVERAGE 70.5 10:1 LOW 10:1 MED 10:1 HIGH PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) AVERAGE 68.4 AVERAGE 70.2 AVERAGE 75.7 19:1 LOW 19:1 MED 19:1 HIGH PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) AVERAGE 63.6 AVERAGE 65.8 AVERAGE 78.3 LED RED TOP RED BOT PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) AVERAGE 66.9 AVERAGE 72 AVERAGE 65.9 50/50 HPS CONTROL PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1) PAR VALUES (μmol m-2 s-1)
AVERAGE 55.2+(24 from HPS) AVERAGE 24 AVERAGE 0
Average after change 60
Page 115 of 115
Table D6: GE Spectroradiometer data. End of run 1 and start of run 2 light map data (January 5th 2012) Zone GE Spectroradiometer
Average Vertical
Average 45º angle
Average
5:1 LOW 692.5 302.0 497.3 5:1 MED 654.3 861.4 757.9 5:1 HIGH 963.6 1056.1 1009.9 10:1 LOW 673.9 402.9 538.4 10:1 MED 692.3 663.4 677.9 10:1 HIGH 966.2 743.5 854.9 19:1 LOW 614.3 546.0 580.1 19:1 MED 757.1 694.0 725.5 19:1 HIGH 846.9 836.1 841.5
LED 829.0 583.0 706.0 50% 912.5 509.2 710.9
Red TOP 905.0 465.2 685.1 Red BOT 738.3 541.8 640.0 Control 1.0 1.0
HPS Light Middle (1 m) 918.6 918.6
Height (1.5 m) 1045.0 1045.0 Top (30 cm from
lamp) 2106.9 2106.9
Light map was for the LED bulbs in a vertical and 45 degree angle at a distance of 30 cm from the bulb in 9 locations per section (full darkness, with other bulbs on). The HPS was measured at three heights (of the chamber (approximate height of the seedlings), top of the chamber and 30 cm from the bulb (in a vertical orientation). All measurements were taken with a spectroradiometer.
