Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Urban Forestry & Urban Greening 14 (2015) 315–322

Contents lists available at ScienceDirect

Urban Forestry & Urban Greening


journal homepage: www.elsevier.com/locate/ufug

Effect of substrate compost percentage on green roof vegetable


production
Mert Eksi a , D. Bradley Rowe b,∗ , Rafael Fernández-Cañero c , Bert M. Cregg d
a
Istanbul University, Faculty of Forestry, Landscape Architecture Department, 34473 Bahcekoy – Sariyer, Istanbul, Turkey
b
Michigan State University, Department of Horticulture, A212 Plant and Soil Sciences, 1066 Bogue Street, East Lansing, MI 48824, USA
c
University of Seville, Department of Agroforestry Sciences, School of Agricultural Engineering, Carretera de Utrera, Km 1, 41013 Seville, Spain
d
Michigan State University, Department of Horticulture and Department of Forestry, A214 Plant and Soil Sciences, 1066 Bogue Street, East Lansing, MI
48824, USA

a r t i c l e i n f o a b s t r a c t

Keywords: Use of rooftops to produce locally grown vegetables is increasing. However, due to weight restrictions,
Cucumbers shallow substrate depths, and potential harsh environmental conditions, optimizing production can be a
Extensive green roof challenge. Standard industry practices for ornamental extensive green roofs planted with succulents or
Growing media
other herbaceous perennials and grasses dictate that organic matter should be less than 20% of the original
Peppers
substrate mix. In rooftop agriculture, however, maximizing growth and yields are usually a primary
Rooftop agriculture
Urban agriculture objective and the amount of organic matter incorporated into these substrates are a major factor in
this equation. In this study we quantified the optimal percentage of compost in a green roof substrate for
optimizing growth and yields for cucumbers (Cucumis sativus) and peppers (Capsicum annuum). The study
was conducted on raised green roof platforms over a period of 19 weeks and compared six substrates
containing increasing amounts of a commercial compost produced from municipal yard waste (0, 20,
40, 60, 80, and 100%) mixed with a heat-expanded shale and sand base. These treatments were also
compared to a typical garden plot at ground level. Plant performance evaluations such as plant growth,
chlorophyll fluorescence (Fv/Fm) as an indicator of plant stress, and fruit yields were used as an indicator
for optimal substrate mixtures. Generally, the addition of 60 or 80% compost resulted in the greatest plant
growth and fruit yields, although compost influenced growth and yield of peppers to a greater degree
than cucumbers. In addition, the ground garden plots performed poorly which emphasizes the point that
growing vegetables on a rooftop can be advantage because substrates can be engineered to maximize
plant health, although the same could be done with raised beds in a garden plot.
© 2015 Elsevier GmbH. All rights reserved.

Introduction provide the other generally accepted benefits of green roofs such
as stormwater retention, cooling of buildings due to evapotranspi-
One of the key challenges in many cities is providing opportu- ration and shading, and carbon storage provided by sequestration
nities for residents to meet their basic needs of food, water, and in plant biomass, carbon transferred to the substrate via plant lit-
shelter (Lynch et al., 2013). One practice that can help alleviate ter and exudates, and soil organic matter (Oberndorfer et al., 2007;
the food security problem is the adoption of urban agriculture Getter et al., 2009; Magdoff and van Es, 2010).
which involves growing plants and raising animals for food near There are only a few published experimental studies related
population centers (Veenhuizen, 2007). Because land suitable for to green roof agriculture (Elstein et al., 2006; Whittinghill and
growing food is often limited in urban areas, the use of rooftops to Rowe, 2012; Whittinghill et al., 2013) and the effect of compost
grow vegetables can improve nutrition and food security in urban in green roof substrates (Rowe et al., 2006; Nagase and Dunnett,
neighborhoods while reducing dependence on an energy-intensive 2011; Papafotiou et al., 2013; Young et al., 2014). It is well known
global food economy (Gorgolewski et al., 2011; Whittinghill and that organic amendments can improve chemical and physical prop-
Rowe, 2012). In addition to food production, this approach can still erties of soils; increase moisture-holding capacity (Aggelides and
Londra, 2000; Arancon et al., 2004; Beattie and Berghage, 2004;
Evanylo et al., 2008); play a major role in thermal conductivity,
∗ Corresponding author. Tel.: +1 517 355 5191. soil texture, water content, salt concentration, and organic mat-
E-mail address: rowed@msu.edu (D.B. Rowe). ter content (Abu-Hamdeh and Reeder, 2000); enhance soil fertility

http://dx.doi.org/10.1016/j.ufug.2015.03.006
1618-8667/© 2015 Elsevier GmbH. All rights reserved.
316 M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322

(Pinamonti, 1997; Zebarth et al., 1999; Rowe, 2011; Nagase and


Dunnett, 2011); and increase crop yields (Mays et al., 1973; Carter
et al., 2004; Tilston et al., 2005; D’Hose et al., 2014). In addition,
adding organic matter usually improves cation exchange capacity
(CEC) (Beattie and Berghage, 2004), slows nitrification of N, and sta-
bilizes soil chemical properties by increasing soil buffering capacity
and slowly releasing essential nutrients for more sustainable plant
growth (Stamatiadis et al., 1999; Aggelides and Londra, 2000).
Even with all the above mentioned benefits, German green roof
standards state that substrates should not contain more than 20%
compost by volume (FLL, 2008). This is because high levels of com-
post in green roof substrates may lead to leaching of nutrients
(Rowe, 2011), higher levels of N and P in the runoff (Emilsson et al.,
2007; Rowe, 2011), and plant nutrient availability in soils varies due
to type of compost (Mkhabela and Warman, 2005). Also, organic
matter can weigh more than light-weight heat expanded aggre-
gates, especially when it is wet. The structural weight limitations
of any given roof is often the limiting factor on whether a green
roof can be installed. In addition, organic matter will decompose
and may lead to substrate shrinkage (Beattie and Berghage, 2004). Fig. 1. Particle distribution of growing substrates containing 0–100% compost and
the ground plot (GR). Analysis performed in Michigan State University Dept. of Civil
This becomes a problem on a standard green roof as it is not feasible Engineering Lab per ASTM C136-06 Standard Test Method for Sieve Analysis of Fine
or practical to continually replace the substrate on a rooftop. and Coarse Aggregates.
Specifications for typical green roofs as outlined in the FLL stan-
dards (FLL, 2008), were primarily written for ornamental plants
selected for harsh green roof environments. However in rooftop including the ground plots were fertilized with a controlled release
agriculture maximizing growth and yields are usually a primary fertilizer (Osmocote Plus X 15N–9P2 O5 –12K2 O 5-month release
objective and the amount of compost incorporated into these sub- (Everis International BV) at a rate of 70.4 g per plot (11.0 g/m2 ) and
strates are a major factor in this equation. Therefore, the objectives watered to field capacity by hand on the day of planting. No fur-
of this study were: (1) to investigate the feasibility of producing ther fertilization or any pesticides were applied to the study plots.
vegetables (peppers and cucumbers) on shallow extensive green All platforms were exposed to full sun. Supplemental irrigation
roofs and (2) to determine the optimal percentage of substrate was provided to the roof platforms by an automatic irrigation sys-
compost to optimize yields for these crops. tem through a series of overhead sprinklers. Plots were irrigated
three times per week for 15 min during the growing season which
amounted to 2.4 L per section per irrigation event. This irrigation
Materials and methods schedule amounted to 11 mm of water per week in addition to
normal rainfall. The ground plots were not watered at the same
The study was conducted on divided roof platforms at the Michi- rate as the roof platforms as doing so would have resulted in con-
gan State University Horticulture Teaching and Research Center stantly saturated soil due to poor drainage. The ground plots were
(East Lansing, MI). Eighteen sections (plots) were partitioned with watered as needed. Initial planting was performed May 22 and
wood dividers with each section measuring 0.81 m × 0.81 m. Plat- a late frost occurred on May 24 which damaged the cucumber
forms replicated a commercial extensive green roof and included plants so they were replanted on May 30. Pepper plants were not
a waterproofing membrane, drainage mat (XeroFlor XF108, Wolf- affected.
gang Behrens Systementwicklung, GmbH, Groß Ippener, Germany), Measurements of plant height (cm) for peppers and length
moisture retention fabric (Xero Flor XF159) and vegetation carrier (cm) for cucumbers, number of leaves, and chlorophyll fluores-
(Xero Flor XF301). Substrate depth was 12.5 cm in all sections. cence (Fv/Fm) were collected initially at the time of planting,
Six substrate blends with increasing amounts (0, 20, 40, 60, 80, then once a week over the first 5 weeks, and then every other
and 100%) of a commercial compost produced from municipal yard week for the duration of the study. Height of the pepper plants
waste compost (Hammond Farms, East Lansing, MI) were mixed were measured on the main stem from the ground to the top
with heat-expanded shale aggregate (30% Haydite A and 30% Hay- of the plant. Cucumbers generated up to three branches during
dite B, Hydraulic Press Brick Company, Brooklyn, IN), and 40% 2NS the growing period so length was measured on the stem with
sand (Renewed Earth, Kalamazoo, MI) and randomly distributed the longest branch. Once plants began producing fruits, fruit size
among the 18 platform plots (Fig. 1). Each treatment was repli- and fruit weight measurements were recorded instead of leaf
cated three times. In addition, three ground plots representing a counting. The first fruits were observed on peppers in July 3
typical home garden were also established and served as a control and by week 9 (July 26) ripe fruits of peppers and cucumbers
for comparison (Table 1). These ground plots had been maintained were harvested and the number, length or height and diameter
as a garden for the past 17 years and have had compost added to of fruits were recorded weekly. Snapshots of substrate mois-
them on several occasions, but still were primarily composed of ture and temperature were recorded on June 13, July 2, and July
clay. In fact, many chemical properties such as P, K, Mg, CEC; per- 26.
centage of exchangeable bases for K and Mg, and organic matter Chlorophyll fluorescence (Fv/Fm) was measured using a Hansat-
were lower than those properties for the 0 and 20% compost blends ech PEA (Plant Efficiency Analyzer, Hansatech Instruments, Ltd.,
(Table 1). Norfolk, UK) as the maximum quantum efficiency of photosystem
Two types of vegetables were grown in the study: cucumbers II (Fv/Fm) during the process of photosynthesis is an indirect mea-
(Cucumis sativus) and peppers (Capsicum annuum). Plugs of each sure of plant stress. Prior to Fv/Fm measurements, leaf clips were
species (5.1 cm × 5.7 cm × 7.6 cm, 1230 cm3 , 36/flat) were planted attached to randomly selected leaves of each plant for a minimum
on 22 May 2013 in a triangular form (two peppers and a cucum- of 20 min to provide the dark adaptation (Ritchie, 2006; Getter et al.,
ber in each plot) and spaced 25 cm apart from each other. All plots 2009).
M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322 317

Table 1
Chemical properties of substrate mixtures and ground plots.

Component 0% 20% 40% 60% 80% 100% GR Method

Soil pH 7.8 8.4 8.3 8.8 8.2 8.2 7.9 NCR-13 1998
Phosphorus (P) (ppm) 8.0 68.0 129.0 171.0 177.0 194.0 74.0 NCR-13 1998
Potassium (K) (ppm) 18.0 480.0 1254.0 1584.0 2943.0 3420.0 166.0 NCR-13 1998
Magnesium (Mg) (ppm) 39.0 189.0 390.0 654.0 846.0 972.0 252.0 NCR-13 1998
Calcium (Ca) (ppm) 796.0 1299.0 2305.0 3167.0 3761.0 3999.0 2041.0 NCR-13 1998
Cation-exchange capacity (meq/100 g) 4.4 9.3 18.0 25.3 33.4 36.9 12.7 Centrifugation procedure, Warncke, D.D.a
K % of exchangeable bases 1.1 13.2 17.9 16.0 22.6 23.8 3.3 Flame emissionb
Mg % of exchangeable bases 7.5 16.9 18.1 21.5 21.1 22.0 16.5 Colorimetricallyc
Ca % of exchangeable bases 91.5 69.8 64.1 62.5 56.3 54.2 80.2 Flame emissionb
Organic matter % 0.4 2.2 7.9 15.1 21.3 31.0 4.8 Loss-on-ignitiond

Analysis per Michigan State University Soil and Plant Nutrient Laboratory East Lansing, Michigan (0–100% represents compost rates, GR represents Ground Plot).
a
Centrifugation procedure developed by D. D. Warncke, Department of Crop and Soil Sciences, Michigan State University, East Lansing, MI.
b
Flame emission pp. 31–34 Recommended Chemical Soil Test Procedures for the North Central Region; J.R. Brown, North Central Regional Research Publication No. 221;
Revised January, 1998 (Brown, 1998).
c
Colorimetric pp. 31–34 from Recommended Chemical Soil Test Procedures for the North Central Region; J.R. Brown, North Central Regional Research Publication No. 221;
Revised January, 1998 (Brown, 1998).
d
Loss-on-ignition pp. 57–58 from Recommended Chemical Soil Test Procedures for the North Central Region; J.R. Brown, North Central Regional Research Publication No.
221; Revised January, 1998, pp. 57–58 (Brown, 1998).

Substrate volumetric moisture content and temperature were Plant growth


measured on randomly selected days. Substrate moisture levels
were recorded at three points in each plot by inserting a Theta After the initial planting, the number of leaves per cucumber
probe (ML2x; Delta-T Devices, Ltd., Cambridge, UK) with 6.0-cm and pepper plant were counted until the first fruits were produced.
rods into the substrate. The Theta probe instrument has a range Quantitatively, for cucumbers, the 80% compost mix produced the
of 0.0–1.0 m3 m−3 , with accuracy of ±0.01 m3 m−3 for values from greatest number of leaves and the 60 and 80% mixes produced the
0.05 to 0.6 m3 m−3 . However, accuracy was likely lower for values greatest vine length (Table 2). Cucumber plants generated several
below 0.05 m3 m−3 (Delta-T Devices, 1999) and for the substrates stems which were up to 155 cm long and final lengths were 10 times
with higher percentages of organic matter. At the same time, the initial size.
substrate temperatures were measured with a portable thermome- Cucumber plants growing in the ground plots were significantly
ter (Barnant 600–0000 T/C K Type Thermocouple Thermometer, shorter than those in the green roof platforms, however, there were
Barrington, IL) at the same depth as the Theta probe. Weather no significant differences in leaf number among compost treat-
data was continuously recorded at an adjacent weather station ments. Growth rate for the cucumbers plants was relatively low
(Michigan State University Enviro-Weather Project East Lansing until the fifth week of the study (June 20). They eventually showed
(enviroweather.msu.edu) and climate norms between 1981 and a spurt of growth at week 5 and plants growing in the roof platforms
2010 were obtained from the National Oceanic and Atmospheric outperformed those in the ground control plots. The cucumber
Administration (NOAA) National Weather Service Forecast Office plants lost their leaves and were dead by the end of September
Applied Climate Information System (ACIS) project (NOAA National as temperatures decreased, whereas the peppers were still alive
Weather Service, 2013). until week 19 (October 4). The poor performance in the ground
The effect of compost amount on plant growth, chlorophyll flu- plots was likely due to the high clay content which resulted in little
orescence (Fv/Fm), and fruit yields for peppers and cucumbers aeration. Also, drainage conditions were exacerbated due to a wet
were analyzed by One-Way ANOVA tests using Fisher’s LSD com- spring which put these plants behind from the beginning. In con-
parison in Minitab® 16.2.2 (Minitab Inc., State College, PA). All trast, green roof substrates are designed to drain freely when they
data were checked for normality prior to analysis of variance by become saturated. The fact that the roof plots all performed better
using Kolmogorov–Smirnov test (Minitab Inc., State College, PA) than the ground garden plots emphasizes the point that growing
and non-normal data were analyzed after applying a square root vegetables on a rooftop can be an advantage because substrates can
transformation for leaf number data of cucumbers. be engineered to maximize plant health, although the same could
be done with raised beds in a garden plot.
Results and discussion Peppers grown in the 80% compost produced more leaves than
plants grown in the 0, 20, 40, and 100% compost roofing platforms
Weather conditions and the ground plots (Table 2). Peppers showed consistent growth
throughout the growing period. By week 5 (June 20), pepper plants
Weather patterns during the 2013 growing season were very growing in 80% compost experienced the greatest increases in stem
close to NOAA climate norms recorded between 1981 and 2010 length (Table 2). In addition, toward the end of the growing season,
(NOAA National Weather Service, 2013) (Fig. 2). However, the sum- peppers in 100% compost showed a sudden growth spurt (data not
mer of 2013 experienced lower than normal temperatures and had shown).
large temperature swings during July and August. The warmest
month was July and warmest period of the growing season occurred Chlorophyll fluorescence
during week 7 (13–20 July 2013) with an average temperature
of 25.9 ◦ C. Rainfall in June, August and October was higher than Since the efficiency of photosystem II can be quantified by mea-
normal, whereas, July and September had less precipitation than suring the amount of fluorescence emitted during photosynthesis,
climate norms. October was the wettest month with a total rain- chlorophyll fluorescence measurements are an indirect measure of
fall of 118 mm while the wettest week was recorded during week plant stress (Durhman et al., 2006; Maxwell and Johnson, 2000).
13 (August 24–31). Total precipitation during the growing season This is usually reported as the ratio of variable fluorescence to
was 419 mm which was close to the historical total precipitation maximum fluorescence (Fv/Fm). Normal Fv/Fm values for healthy
records. pepper and cucumber plants range from 0.700 to 0.800. Values
318 M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322

Table 2
Plant growth of cucumbers and peppers as influenced by percentage of compost in the substrate and by the ground plots.

0% 20% 40% 60% 80% 100% Ground

Cucumbers
Leaf generationa 5.7 A 5.0 A 7.0 A 5.0 A 8.0 A 5.7 A 5.3 A
Height (cm) 131.7 A 138.3 A 150.0 A 155.0 A 155.0 A 146.7 A 80.0 B

Peppers
Leaf generationa 17.5 CD 21.5 BC 24.3 BC 30.8 AB 33.3 A 21.5 BC 10.6 D
Maximum length (cm)b 41.8 BC 53.2 A 52.3 AB 53.3 A 60.3 A 52.8 AB 39.8 C
a
Final leaf generation at week 5 before fruits were observed on plants. Each mean represents observations from three plots (three cucumbers or six peppers).
b
Refers to mean length of the longest branches generated by plants in three replications.
Means were separated using Fisher’s LSD (alpha = 0.05). Letters in rows indicate differences between substrate types.

Fig. 3. Change in chlorophyll fluorescence (Fv/Fm) over the course of the study as
influenced by percentage of compost in the substrate and in the ground plots (GR)
for cucumbers (A) and peppers (B). Time ranges from week 2 (May 30) to week 20
(October 11).

differences detected between cucumbers in various compost rates


except during weeks 4 and 9 (Table 3). However, during the first
Fig. 2. Environmental conditions during the 2013 growing season. (A) Monthly week, Fv/Fm values of cucumbers ranged from 0.480 to 0.570, val-
average air temperature and solar radiation, (B) monthly total precipitation, and
ues which can be interpreted as plant stress due to transplanting.
(C) weekly ambient air temperature and rainfall during the experiment (May
31–November 1). Local weather data obtained from the Enviro-weather, MSU Michi- After the first week, cucumbers in 60 and 80% compost quickly over-
gan Automated Weather Network’s East Lansing station located at the research site. came the transplant stress and Fv/Fm values reached up to 0.710.
Long-term climate data (1981–2010) derived from NOAA, National Weather Service Two weeks after planting, values for cucumbers growing in 20 and
Forecast Office database. 40% compost also exceeded the 0.700 level while those in 100%
compost struggled to reach the 0.700 level until the third week.
ranging below 0.600–0.700 are an indication that the plant may Cucumbers in 0% compost just passed the 0.700 threshold after the
be experiencing stress (Ritchie, 2006). fourth week while the mean Fv/Fm values in the ground plots never
After planting, Fv/Fm values generally increased until the plants reached 0.700. Fv/Fm values started to decline by week 17 with the
became established by week 4 (Fig. 3) and there were no significant onset of colder temperatures and all plants were dead by week 19.
M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322 319

Table 3
Chlorophyll fluorescence values of cucumbers and peppers as influenced by percentage of compost in the substrate and by the ground plots.

0% 20% 40% 60% 80% 100% Ground

Cucumbers
Week 4 (June 13) 0.659 BC 0.638 C 0.694 A 0.712 A 0.721 A 0.690 ABC 0.552 D
Week 5 (June 20) 0.641 A 0.708 A 0.701 A 0.703 A 0.703 A 0.640 A 0.648 A
Week 9 (July 26) 0.738 AB 0.744 AB 0.767 A 0.730 AB 0.583 C 0.657 BC 0.674 BC
Week 11 (August 8) 0.725 A 0.653 A 0.708 A 0.657 A 0.591 A 0.718 A 0.623 A

Peppers
Week 4 (June 13) 0.664 A 0.735 A 0.734 A 0.727 A 0.718 A 0.677 A 0.366 B
Week 5 (June 20) 0.697 A 0.708 A 0.723 A 0.716 A 0.635 A 0.685 A 0.419 B
Week 9 (July 26) 0.747 A 0.656 B C 0.643 C 0.750 A 0.698 AB 0.698 AB 0.717 AB
Week 11 (August 8) 0.763 A 0.722 AB 0.649 B 0.726 AB 0.670 B 0.766 A 0.713 AB

Each mean represents observations from three plots. Means were separated using Fisher’s LSD (alpha = 0.05). Letters in rows indicate differences between substrate types.

During the first three weeks and the last weeks of the study, Fv/Fm Following establishment, there were no significant differences
values for cucumbers in the roofing platforms were significantly observed among peppers growing on the green roof platforms until
different than those in the ground plots and values in the ground week 9 (July 26 to August 2) or week 11 (August 8–15). Toward the
plots decreased a week earlier than roofing platforms. end of the growing season after week 17 (September 27) all plants
At week 4 (June 12), Fv/Fm values of cucumbers growing in 40, started to deteriorate due to weather conditions.
60 and 80% of compost were higher than all other treatments. A Significant differences among treatments were detected at vari-
heavy rain event began the day before measurements were taken ous times during the study period (Table 3). At week 4 (June 13) and
and continued until the morning of June 13 (46.5 mm). These wet week 5 (June 20) significant differences were detected between all
environmental conditions likely affected the Fv/Fm readings in the green roof platforms and ground plots. At week 9 (July 26) Fv/Fm
ground plots where drainage was poor due to the clay soil. At week values were significantly lower for plants growing in 20 and 40%
9 (July 26), Fv/Fm values were the highest at 40% compost followed compost compared to 60%. Fv/Fm values of plants growing in the
by 0, 20, the 60% plots which had similar values statistically. How- ground began rising during week 5 and by week 11 were as high as
ever, at 80 and 100% compost and in the ground plots, Fv/Fm values those on the roofing platforms, but started to decrease earlier than
were significantly lower. Week 9 was the coolest week of summer those plants in roofing platforms. Fv/Fm values of pepper plants in
with an average air temperature of 17.3 ◦ C. Lower night temper- 100% of compost were the highest, but were only significantly dif-
atures decreased the overall average and may have affected the ferent from 40 and 80% compost (Table 3). By the time the final
plants as lower temperatures would decrease evapotranspiration, chlorophyll fluorescence measurements were taken at week 18
thus increasing substrate moisture, especially in the high compost (September 13), Fv/Fm values had decreased due to senescence of
treatments. A dry substrate should never have been a major factor plant tissues. There was no clear trend among treatments other
in plant stress since the plants were irrigated three times a week. than peppers in the ground plots getting off to a slow start, pre-
Compost influenced substrate temperatures as the 80 and 100% sumably due to the lack of aeration in the clay soil.
treatments consistently exhibited the lowest temperatures which
is likely related to substrate moisture (Fig. 4). Fruit harvest indicators
For peppers, there was limited transplant stress during the first
week of the growing period except for the ground plots. Mean Quantitatively, the highest fresh weight production for cucum-
Fv/Fm values ranged between 0.616 and 0.698 for the roof plat- bers occurred at 60% compost (3403 g) (Table 4). The least
form treatments, but was just above 0.300 for those growing in the production occurred in the ground plots with only eight cucumbers
ground (Fig. 3). In fact, peppers growing in the ground plots expe- produced with a total weight of 1857 g (619 g per plot). All plants
rienced stress up until week 7 when Fv/Fm finally exceeded 0.600. grown on platform produced a greater amount of fresh weight
cucumbers than plants grown in the ground plots (Table 4). Even
though the 60% compost treatment for cucumbers produced 740 g
more per plot of fresh weight than those grown in 80% compost,
the greatest marketable yield was produced in 80% compost where
40% of the cucumbers met USDA standards. Even so, the majority
were still graded below USDA No. 1 standards (60%). Overall, a large
number of cucumber fruits (68%) failed to achieve the expected
standards, fruit size and quality. Only 32% met USDA standards:
2% (USDA No. 1), 16% (USDA No. 2), 14% (USDA No. 3), and the
remainder were evaluated as culls.
Fifty-one peppers were harvested from the 60% compost treat-
ment which also had the highest fresh weight (3903 g). This was
followed by 80% compost with 41 peppers (3201 g). Total fresh
weight of pepper fruit was normally proportional to chlorophyll
fluorescence values as yields increased with higher Fv/Fm values.
As stated, the highest pepper yield occurred at 60% compost which
also corresponded to the least stressed plants according to Fv/Fm
measurements. This was also generally the case with cucumbers
with some exceptions.
Fig. 4. Mean volumetric moisture content and mean substrate temperatures on June
Significant differences existed in yield between the ground plots
13, July 2, and July 26, as influenced by percentage of compost in the substrate and
in the ground plots (GR). Means separated using LSD (alpha = 0.05). Capital letters and the higher percentages of composts on the roofing platforms.
indicate differences in volumetric moisture content among substrate types. On the roofing platforms the lowest pepper yield was harvested
320 M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322

Table 4
Harvest indicators for cucumbers (Cucumis sativus) and peppers (Capsicum annuum) as influenced by percentage of compost in the substrate and by the ground plots.

0% 20% 40% 60% 80% 100% Ground

Cucumbers
Total fresh weight (g) 1835 AB 1995 AB 3013 A 3403 A 2661 A 2430 AB 619 B
Marketable yield (g) 459 AB 507 AB 339 B 760 AB 1022 A 503 AB 103 B
Marketable yield (%) 26 AB 25 AB 13 B 25 AB 40 A 24 AB 14 AB
Number of fruits 5.7 AB 7.0 AB 8.7 A 10.0 A 8.3 A 6.7 AB 2.7 B
Fruit length (cm) 16.9 AB 16.8 AB 17.9 A 17.0 AB 16.3 AB 17.3 AB 15.8 B
Fruit grade 2.6 A 2.2 A 2.4 A 2.0 A 2.4 A 2.4 A 2.5 A

Peppers
Total fresh weight (g) 328 BC 783 AB 877 AB 1301 A 1067 A 979 AB 175 C
Marketable yield (g) 285 BC 740 AB 834 AB 1258 A 1025 A 937 AB 175 C
Marketable yield (%) 83.5 B 93.5 A 94.0 A 96.6 A 95.0 A 95.0 A 100 A
Number of fruits 5.3 B C 10.3 AB 12.6 AB 17.0 A 13.6 A B 14.6 A 3.0 C
Fruit length (cm) 5.9 C 7.0 AB 7.0 AB 7.0 AB 7.4 A 6.8 B 6.3 BC
Fruit grade 3.2 A 2.9 B 2.9 B 3.0 B 3.0 B 3.0 B 3.0 B

Total yield based on the combined fresh weight of all fruit harvested during growing season of 2013. Marketable percent of yield calculated by (marketable yield weight/total
yield weight) × 100 for each plot.
Fruit grades are based on United States Standards for Grades of Cucumbers (USDA, 1958) (1 = 8.9 and 3.2 cm—US No. 1; 2 = 14.0 and 4.8 cm—US No. 2; 3 = 15.2 and 5.7 cm—US
No. 3; 4 = >15.2 and >5.7 cm—Cull. maximum length and diameter—grade).
Each mean represents observations from three plots. Means were separated using Fisher’s LSD (alpha = 0.05). Letters in rows indicate differences between substrate types.

from the plants growing in 0% compost. (Table 4). When com- supported plant growth (Nagase and Dunnett, 2011; Papafotiou
pared to 0% compost and the ground plots, total fresh weight yield et al., 2013; Young et al., 2014). In general, it is well known that
harvested from the plots containing 60% of compost was almost the proper amounts of compost improve soil quality as well as crop
three times higher. production (D’Hose et al., 2014).
A greater percentage of the peppers produced from both the On randomly selected days during the growing season (June
roofing platforms and ground plots met USDA standards for mar- 13, July 2, and July 26) substrate volumetric moisture content and
ketability than cucumbers, but they were at the small end of the substrate temperature were measured. Moisture content of the
standard. This may be the result of a relatively cool summer and substrate mixtures all increased in direct proportion to increasing
cold night temperatures during the 2013 growing season. Polowick amounts of compost (Fig. 4). This is expected, as compost generally
and Sawhney (1985) found that temperature conditions strongly increases water retention in substrates (Giusquiani et al., 1995).
influence development of flowers and fruits of pepper plants and The predominantly clay soil ground plots contained relatively high
lower temperatures (LTR; 18 ◦ C day/15 ◦ C night) had much greater amounts of water since clay particles bind water, and as clay con-
effects on flowers and fruits than intermediate (ITR; 23 ◦ C day/18 ◦ C tent increases, volumetric moisture content also increases (Gong
night) or high (HTR; 28 ◦ C day/23 ◦ C night) temperatures (LTR, ITR et al., 2003).
and HTR refers to low temperature, intermediate temperature, and In previous studies, Papafotiou et al. (2013) found that incor-
high temperature, respectively). The relatively cold night temper- porating compost into green roof substrates resulted in greater
atures during June (minimum temperature 3.9 ◦ C and average low water retention and plant growth for three Mediterranean xero-
temperature 14.2 ◦ C) in East Lansing likely affected the plants. Cli- phytes, Artemisia absinthium, Helichrysum italicum, and H. orientale.
mate conditions reduced the total yield as well as the quality and The shallow (7.5 cm) compost-amended substrate (grape marc
most of the pepper fruits failed to meet USDA Fancy or No. 1 stan- compost:soil:perlite (2:3:5, v/v)) with sparse irrigation resulted in
dards. Most of the peppers fell in the USDA No. 2 grade. USDA similar or greater plant growth of all plant species compared with
No. 1 grade peppers only accounted for 11% of the total harvest a deep (15 cm) peat-amended substrate (peat:soil:perlite (2:3:5,
and 8% were classified as culls. However, most harvested peppers v/v)) with normal irrigation. The source of organic matter is also
could still be classified as “marketable” ranging from 87 to 100% very important. When comparing green waste compost to conifer
depending on the treatment. Quantitatively, the number of fruits bark, Young et al. (2014) reported that the green waste compost
was highest for 60% compost, but was only significantly different increased plant available nutrients, shoot and root growth, shoot
from the 0% compost and the ground plots (Table 4). Fruit grade in N concentration, and chlorophyll content of Lolium perenne (rye-
0% compost was significantly different than remaining treatments grass). However, it also decreased the shoot:root ratio which could
with a higher mean value which can interpreted as lower grade potentially make the plants more subject to drought stress.
peppers. In addition, significant differences were detected in fruit In contrast, substrate temperatures decreased with increasing
length between 80% compost compared to 0% and 100%, and the amounts of compost (Fig. 4). On sunny days during summer, despite
ground plots. its darker color, the 100% compost mixture had the lowest temper-
ature values because it was capable of retaining larger quantities of
Substrate characteristics water which also allows for greater evaporative cooling (Penman,
1948; Monteith, 1981). However, the rate of evapotranspiration
According to German green roof industry guidelines, compost will eventually be limited by moisture availability within the sub-
should not exceed 20% by volume in substrates (FLL, 2008) as the strate (Voyde et al., 2010). The ground plots maintained relatively
organic matter will decompose and result in substrate shrinkage high substrate temperatures and moisture, despite the relatively
(Beattie and Berghage, 2004). However, it is evident in this study low levels of organic matter (Table 1; Fig. 4).
that the addition of greater amounts than 20% compost improved The increase in compost rate significantly enhanced most sub-
substrate moisture, plant growth, and fruit production. However, strate chemical properties (Table 1). This is not surprising as organic
very high amounts such as 100% was found to be detrimental, pre- matter serves as a reservoir for many plant nutrients and also
sumably due to too much substrate moisture (Fig. 4). In other green contributes substantially to substrate CEC, which is important for
roof studies, organic matter also increased substrate moisture and the retention of many nutrients (Powlson et al., 2013). Again, the
M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322 321

properties possessed by a particular compost depend on the source than what is recommended for standard ornamental extensive
and type and could lead to leaching of nutrients into the runoff green roofs planted with succulents or other herbaceous perenni-
(Rowe, 2011). When comparing substrate blends of grape marc als and grasses because of the negative consequences of substrate
compost:soil:perlite (2:3:5, v/v) and peat:soil:perlite (2:3:5, v/v), shrinkage and nutrient runoff. However, in rooftop agriculture
Papafotiou et al. (2013) reported that although both substrates had maximizing growth and yields are usually a primary objective
similar physical properties, higher levels of macronutrients, total and the amount of compost incorporated into these substrates are
P, and K were recorded in the compost-amended substrate. In the a major factor in this equation. Also, management will be more
current study, the substrate containing 0% compost consisted of the intense with the need for higher nutrient levels, irrigation require-
base 60% heat expanded shale and 40% 2NS sand and possessed a ments, and labor for harvesting the produce. In this study, the
low CEC (4.4 meq/100 g) and few nutrients with the exception of drainage provided by the green roof systems with controlled sub-
Ca. Nutrient levels in ground plots were much lower than all of the strate and environment conditions increased production quality
other compost added substrates except for 20% in some cases. The and health of plants on the roofing platforms relative to the ground
relatively low nutrient levels of the ground plots likely influenced plots. The fact that the roof plots all performed better than the
plant growth in addition to the clay soil with poor aeration and actual ground garden plot emphasizes the point that growing veg-
drainage. etables on a rooftop can be advantage because substrates can be
The addition of compost also altered particle size distribution engineered to maximize plant health, although the same could be
(Fig. 1). Plots containing 0% compost can be classified as sandy soils done with raised beds in a garden plot. In this case, the ground
(>80% sand, <10% clay) per the ASTM C136-506 06 Standard Test plots consisted of native clay soil that was lacking in nutrients and
Method for Sieve Analysis of Fine and Coarse Aggregates, which drainage. Rooftop vegetable production can not only improve nutri-
generally have poor physical properties such as low water retention tion and food security in urban neighborhoods, but these roofs
as well as poor chemical fertility (Verhoef and Egea, 2013). The can also supply the ecosystem services normally associated with
addition of compost improved substrate conditions, whereas, in non-food producing green roofs. Additional research is needed to
the ground plots, the high clay content resulted in poor chemical optimize growing substrates, fertilization practices, irrigation, and
and physical properties which influenced plant growth, stress, and environmental conditions for a variety of vegetable species.
yields. With the exception of the 0% compost mix, the ground plots
generally contained lower levels of P, K, and Mg; lower CEC; lower
Acknowledgements
percentages of exchangeable bases for K and Mg, and much less
organic matter (Table 1).
Funding for this study was provided by TUBITAK (The Scien-
In our study, the 0% compost substrate restricted plant growth
tific and Technical Research Council of Turkey) 2219 International
and reduced yields of peppers. Similarly, Rowe et al. (2006) found
Postdoctoral Research Scholarship Programme; Renewed Earth,
that decreasing amounts of compost mixed with heat expanded
Kalamazoo, MI; Ford Motor Company, Dearborn, MI; ChristenDE-
slate reduced survival and growth of several herbaceous peren-
TROIT Roofing Contractors, Detroit, MI; XeroFlor America, Durham,
nials and grasses in green roof substrates. Similarly, Nagase and
NC; the USDA National Institute of Food and Agriculture, and
Dunnett (2011) studied the effects of organic matter in green roof
AgBioResearch at Michigan State University. Mert Eksi and Rafael
substrates on plant growth of Allium schoenoprasum, Limonium
Fernández Cañero were visiting scholars at MSU during 2013–14
latifolium, Melica ciliata, and Nepeta × faassenii. They amended a
and 2013, respectively.
commercial green roof substrate (a crushed brick base containing
less than 4% organic matter) with additional green waste compost
at rates of 0, 10, 25, and 50% by volume. Although there were dif- References
ferences among species, they concluded that the addition of 10%
organic matter was optimal and that those species from nitrogen- Abu-Hamdeh, N.H., Reeder, R.C., 2000. Soil thermal conductivity: effects of den-
sity, moisture, salt concentration, and organic matter. Soil Sci. Soc. Am. J. 64,
rich habitats tended to perform better with higher percentages
1285–1290.
of organic matter. Although, there were only minor differences Aggelides, S.M., Londra, P.A., 2000. Effects of compost produced from town wastes
between cucumbers and peppers, it is likely that there would be and sewage sludge on the physical properties of a loamy and a clay soil. Biore-
sour. Technol. 71, 253–259.
differences among the numerous species of vegetables that are
Arancon, N.Q., Edwards, C.A., Bierman, P., Welch, C., Metzger, J.D., 2004. Influences
typically grown. of vermicomposts on field strawberries: effects on growth and yields. Bioresour.
Whittinghill et al. (2013) tested six vegetables and herbs in three Technol. 93, 145–153.
growing systems—a green roof on a two story building, raised green Beattie, D.J., Berghage, R., 2004. Green roof media characteristics: the basics. In:
Proceedings of 2nd North American Green Roof Conference: Greening Rooftops
roof platforms, and in-ground plantings. Species included tomatoes for Sustainable Communities, Portland, OR, pp. 411–416.
(Solanum lycopersicum), green beans (Phaseolus vulgaris), cucum- Brown, J.R., 1998. Recommended chemical soil test procedures for the north central
bers (Cucumis sativus), peppers (Capsicum annuum), basil (Ocimum region. North Central Regional Research Publication No. 221, pp. 57–58.
Carter, M.R., Sanderson, J.B., MacLeod, J.A., 2004. Influence of compost on the physical
basilicum), and chives (Allium schoenoprasum). Their study resulted properties and organic matter fractions of a fine sandy loam throughout the cycle
with overall higher quality for the in-ground production rather than of a potato rotation. Can. J. Soil Sci. 84, 211–218.
on the green roof platforms. In contrast, our study showed that the D’Hose, T., Cougnon, M., Vliegher, A.D., Vandecasteele, B., Viaene, N., Cornelis, W.,
Bockstaele, E.V., Reheul, D., 2014. The positive relationship between soil quality
roofing platforms outperformed the ground plots. These findings and crop production: a case study on the effect of farm compost application.
point out the importance of chemical and physical properties of Appl. Soil Ecol. 75, 189–198.
growing substrates and environmental conditions. Durhman, A.K., Rowe, D.B., Rugh, C.L., 2006. Effect of watering regimen on chloro-
phyll fluorescence and growth of selected green roof plant taxa. HortScience 41
(7), 1623–1628.
Conclusions Elstein, J., Welbaum, G.E., Stewart, D.A., Borys, D.R., 2006. Evaluating growing media
for a shallow-rooted vegetable crop production system on a green roof. Acta
Horti. 782, 177–183.
This study demonstrates that the addition of proper amount of
Emilsson, T., Berndtsson, J.C., Mattsson, J.E., Rolf, K., 2007. Effect of using conven-
compost to shallow green roof substrates can increase plant growth tional and controlled release fertilizer on nutrient runoff from various vegetated
and yields of cucumbers and peppers under regular irrigation. Gen- roof systems. Ecol. Eng. 29, 260–271.
erally, the addition of 60 or 80% compost resulted in the greatest Evanylo, G., Sherony, C., Spargo, J., Starner, D., Brosius, M., Haering, K., 2008. Soil and
water environmental effects of fertilizer-, manure-, and compost-based fertility
plant growth and fruit yields, although the differences were not practices in an organic vegetable cropping system. Agric. Ecosyst. Environ. 127,
as pronounced for cucumbers. These amounts are much higher 50–58.
322 M. Eksi et al. / Urban Forestry & Urban Greening 14 (2015) 315–322

Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau (FLL) (Society of Pinamonti, F., 1997. Compost mulch effects on soil fertility, nutritional status and
Landscape Development and Landscape Design), 2008. Guidelines for the Plan- performance of grapevine. Nutrient Cycl. Agroecosyst. 51, 239–248.
ning, Construction, and Maintenance of Green-roof Sites. FLL, Bonn, Germany. Polowick, P.L., Sawhney, V.K., 1985. Temperature effects on male fertility and
Getter, K.L., Rowe, D.B., Robertson, G.P., Cregg, B.M., Andresen, J.A., 2009. Carbon flower and fruit development in Capsicum annuum L. Scientia Horticulturae 25,
sequestration potential of extensive green roofs. Environ. Sci. Technol. 43 (19), 117–127.
7564–7570. Powlson, D., Smith, P., Nobili, M.D., 2013. Soil organic matter. In: Gregory, P., Nortcliff,
Giusquiani, P.L., Pagliai, M., Gigliotti, G., Businelli, D., Benetti, A., 1995. Urban waste S. (Eds.), Soil Conditions and Plant Growth. Blackwell Publishing Ltd, UK, pp.
compost: effects on physical, chemical, and biochemical soil properties. J. Envi- 86–131.
ron. Qual. 24, 175–182. Ritchie, G.,2006. Chlorophyll fluorescence: What is it and what do the numbers
Gong, Y., Cao, Q., Sun, Z., 2003. The effects of soil bulk density, clay content and tem- mean? In: USDA Forest Service Proceedings RMRSP-43. U.S. Department of Agri-
perature on soil water content measurement using time-domain reflectometry. culture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO, pp.
Hydrol. Process. 17, 3601–3614. 34–43.
Gorgolewski, M., Komisar, J., Nasr, J., 2011. Carrot City: Creating Places for Urban Rowe, D.B., 2011. Green roofs as a means of pollution abatement. Environ. Pollut.
Agriculture. Monacelli Press, New York, NY. 159, 2100–2110.
Lynch, K., Maconachie, R., Binns, T., Tengbe, P., Bangura, K., 2013. Meeting the urban Rowe, D.B., Monterusso, M.A., Rugh, C.L., 2006. Assessment of heat-expanded slate
challenge? Urban agriculture and food security in post-conflict Freetown, Sierra and fertility requirements in green roof substrates. HortTechnology 16 (3),
Leone. Appl. Geogr. 36, 31–39. 471–477.
Magdoff, F., van Es, H., 2010. Building Soils for Better Crops. Sustainable Soil Man- Stamatiadis, S., Werner, M., Buchanan, M., 1999. Field assessment of soil quality
agement, 3rd ed. SARE Outreach Publications, Waldorf, MD. as affected by compost and fertilizer application in a broccoli field (San Benito
Maxwell, K., Johnson, G.N., 2000. Chlorophyll fluorescence—a practical guide. J. Exp. County, California). Appl. Soil Ecol. 12, 217–225.
Bot. 51 (345), 659–668. Tilston, E.L., Pitt, D., Fuller, M.P., Groenhof, A.C., 2005. Compost increases
Mays, D.A., Terman, G.L., Duggad, J.C., 1973. Municipal compost: effects on crop yield and decreases take-all severity in winter wheat. Field Crop Res. 94,
yields and soil properties. J. Environ. Qual. 2 (1), 89–92. 176–188.
Mkhabela, M.S., Warman, P.R., 2005. The influence of municipal solid waste compost Veenhuizen, R., 2007. Profitability and sustainability of urban and peri-urban agri-
on yield, soil phosphorus availability and uptake by two vegetable crops grown culture. Agricultural Management, Marketing and Finance Occasional Paper
in a Pugwash sandy loam soil in Nova Scotia. Ecosyst. Environ. 106, 57–67. No. 19. Food and Agriculture Organization of the United Nations (FAO),
Monteith, J.L., 1981. Evaporation and surface temperature. Quart. J. R. Meteorol. Soc. Italy.
107 (451), 1–27. Verhoef, A., Egea, G., 2013. Soil water and its management. In: Gregory, P., Nortcliff,
Nagase, A., Dunnett, N., 2011. The relationship between percentage of organic matter S. (Eds.), Soil Conditions and Plant Growth. Blackwell Publishing Ltd., UK, pp.
in substrate and plant growth in extensive green roofs. Landsc. Urban Plann. 103, 269–322.
230–236. Voyde, E., Fassman, E., Simcock, R., 2010. Hydrology of an extensive living roof
NOAA National Weather Service, 2013. National Oceanic and Atmospheric Admin- under sub-tropical climate conditions in Auckland, New Zealand. J. Hydrol. 394,
istration (NOAA) National Weather Service Forecast Office Applied Climate 384–395.
Information System (ACIS) Project (www.nws.noaa.gov/). Whittinghill, L.J., Rowe, D.B., 2012. The role of green roof technology in urban agri-
Oberndorfer, E., Lundholm, J., Bass, B., Connelly, M., Coffman, R., Doshi, H., Dun- culture. Renew. Agric. Food Syst. 27, 314–322.
nett, N., Gaffin, S., Kohler, M., Lui, K., Rowe, D.B., 2007. Green roofs as urban Whittinghill, L.J., Rowe, D.B., Cregg, B.M., 2013. Evaluation of vegetable production
ecosystems: ecological structures, functions, and services. Bioscience 57 (10), on extensive green roofs. Agroecol. Sustain. Food Syst. 37, 465–484.
823–833. Young, T., Cameron, D.D., Sorrill, J., Edwards, T., Phoenix, G.K., 2014. Importance of
Papafotiou, M., Pergialioti, N., Tassoula, L., Massas, J., Kargas, G., 2013. Growth of different components of green roof substrate on plant growth and physiological
native aromatic xerophytes in an extensive Mediterranean green roof as affected performance. Urban For. Urban Green. 13 (3), 507–516.
by substrate type and depth and irrigation frequency. HortScience 48 (10), Zebarth, B.J., Neilsen, G.H., Hogue, E., Neilsen, D., 1999. Influence of organic waste
1327–1333. amendments on selected soil physical and chemical properties. Can. J. Soil Sci.
Penman, H.L., 1948. Natural evaporation from open water, bare soil and grass. Proc. 79 (3), 501–504.
R. Soc. London Series A: Math. Phys. Sci. 193 (1032), 120–145.

You might also like