This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
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
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
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)
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
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
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
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, 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
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
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
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.
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
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.