Compost trial

Introduction

Compostable organics, such as food waste, accounted for 36% of Metro Vancouver's garbage in 2013. Sending compostables to the landfill uses up limited landfill space and produces methane, which is a powerful greenhouse gas. The region has implemented an organics disposal ban in an effort to divert 70% of its waste from disposal this year. Beginning this month, tipping fees for waste haulers include a 50% surcharge for any loads that contain more than 25% food waste.

The region's green bin programs provide an alternative disposal stream for compostable organics. Food scraps and yard waste collected in household green bins are sent to several municipal composting facilities, including an operation at the Vancouver Landfill, in Delta; Harvest Power, in Richmond; and Net Zero Waste, in Abbotsford. These and other commercial composting facilities in the region help close the nutrient loop by generating natural fertilizer for the region's farms.

I visited the Harvest Power facility in Richmond in February. I was impressed by the scale of the operation, with its capacity to process 240,000 tons of organic waste into compost and biogas annually. I was grateful when the managers offered to donate compost for the student's terrace gardens at KPU's Richmond campus.

Soon after visiting the Harvest Power site, I was contacted by a representative of Net Zero Waste, who also offered to donate compost for the terrace garden project. Since we had two sources of donated municipal compost available, I asked the students to conduct a randomized and replicated experiment to compare the composts, and uncomposted control plots, in terms of crop productivity. We expected the results to be useful to our program, and to regional farmers selecting composts for soil improvement.

Methods

The two terraces east of the Richmond campus entrance were divided into 12 plots, each measuring 1 m by 3 m (Fig. 1). Plots were grouped into two blocks on each terrace, with the three plots nearest the trees in the southeast corner of campus grouped together, to control for possible terrace and shade effects.

Plots in each block were randomly assigned to one of three treatments: 1) Control (no compost); 2) Soil Amender compost from Harvest Power; or 3) Boost compost from Net Zero Waste. Neither compost was blended with sand or topsoil. Composts were applied as a 15 cm mulch layer on top of sandy soil in early April (Fig. 2).

Figure 1. Compost trial location and experimental layout.

Figure 2. Compost trial location before planting. Compost had yet to be applied
to the Net Zero plots in the upper terrace when this image was taken on April 10th.

On April 15th each plot was direct seeded with a row of lettuce (Lactuca sativa cv. 'Drunken Woman'), a row of spinach (Spinachia oleracea cv. 'Savoy'), and a row of Swiss chard (Beta vulgaris cv. 'Pink Flamingo). Rows were spaced 25 cm apart. Plots were irrigated by sprinkler as needed.

Lettuce and spinach were harvested on June 7th (Fig. 3; 53 days after seeding) and Swiss chard was harvested on July 8th (84 days after seeding). All plants were counted and weighed at harvest to calculate total yield, density, and average plant weight. Data analysis was conducted using the R statistical computing environment (Raw data and R coding in Addendum).

Figure 3. Lettuce and spinach harvest, June 7th.

Results

Compost treatment influenced total yield (Fig. 4), yield of individual crops (Figs. 4, 5), and average plant weight (Fig. 6), but not plant density (Fig. 7). Yield and average plant weight were highest in plots amended with compost from Net Zero Waste, and lowest in unamended Control plots. Swiss chard accounted for 72% of the biomass harvested from the study area (Fig. 4), so had a disproportionate influence on the total yield, but similar trends were observed across crops (Fig. 5).

Median values and distributions are compared using 'box and whiskers' charts (interpret) in Figures 4-7.

Figure 4. Total yield of lettuce, spinach, and Swiss chard in unamended
control plots (Ctrl) and plots amended with compost from
Harvest Power (HP) or Net Zero Waste (NZ).

Figure 5. Yield of lettuce, spinach, and Swiss chard in unamended
control plots (Ctrl) and plots amended with compost from
Harvest Power (HP) or Net Zero Waste (NZ).

Figure 6. Average plant weight in unamended control plots (Ctrl) and plots
amended with compost from Harvest Power (HP) or Net Zero Waste (NZ).Weight differed by compost source (left) but not replicate.

Figure 7. Plant count at harvest in unamended control plots (Ctrl)
and plots amended with compost from Harvest Power (HP) or
Net Zero Waste (NZ). Count differed by replicate (right) but not

between compost treatments.

Discussion

Compost improved yield of lettuce, spinach, and Swiss chard. The composts from Harvest Power and Net Zero increased yields by 360% and 700% of the control, respectively. Pronounced yield differences between the treatment and control plots suggest severe nutrient deficiency in the absence of compost, consistent with the low nutrient holding capacity of the high sand, low organic matter soil at the site before treatment. Germination and survival of the direct-seeded crops were similar across treatments, but weight at harvest differed between treatments, suggesting that compost treatment and compost source influenced plant growth.

Compost analysis is needed to determine why crop growth rates differed between the composts tested. Differences in nutrient content, pH, and salinity could all contribute to yield differences such as those observed in this trial.

Acknowledgements

Thanks to Harvest Power and Net Zero Waste for donation of composts used in this trial, and in the remainder of KPU's Richmond campus terrace gardens.

Addendum: Data analysis and presentation in R

A data table called 'Compost' was constructed in R.

Each row of the table contained the following 14 data points for a single plot:

1. Rep: One of four blocks, labeled A, B, C, or D
2. Treat: One of three treatments, labeled Ctrl (Control), HP (Harvest Power), or NZ (Net Zero)
3. letwt: Combined weight of all lettuce, in grams
4. spiwt: Combined weight of all spinach, in grams
5. chawt: Combined weight of all Swiss chard, in grams
6. letcnt: Number of lettuce plants harvested
7. spicnt: Number of spinach plants harvested
8. chacnt: Number of chard plants harvested 
9. letavg: Average weight of a lettuce plant, in grams (letwt/letcnt)
10. spiavg: Average weight of a spinach plant, in grams (spewt/spicnt)
11. chaavg: Average weight of a chard plant, in grams (chawt/chacnt)
12. totwt: Total weight of all plants, in grams (letwt + spiwt + chawt)
13. totcnt: Total number of plants harvested (letcnt + spicnt + chacnt)
14. Plantwt: Average plant weight, in grams (totwt/totcnt)

The null hypothesis that compost source had no effect on total yield was tested by ANOVA. The null hypothesis was rejected, and means were separated by Tukey's Honestly Significant Difference test, which found significant differences between all treatments.

> Compost.aov <- aov(totwt ~ Treat + Rep, data=Compost)

> summary(Compost.aov)

> TukeyHSD(Compost.aov) 

The null hypothesis that compost source had no effect on plant density was tested by ANOVA. The null hypothesis could not be rejected but a significant replicate effect was detected.

> Count.aov <- aov(totcnt ~ Treat + Rep, data=Compost)

> summary(Count.aov)

The null hypothesis that compost source had no effect on plant weight was tested by ANOVA. The null hypothesis was rejected, and means were separated by Tukey's Honestly Significant Difference test, which found significant differences between the two compost treatments, and between Net Zero compost and the Control.

> PlantWt.aov <- aov(PlantWt ~ Treat + Rep, data=Compost)

> summary(PlantWt.aov)

> TukeyHSD(PlantWt.aov)

Figures 4-7 were generated using the following code:

• Figure 4:

> par(mfrow=c(1,2))

> boxplot(totwt~Treat, data=Compost, main="Total yield", xlab="Compost source", ylab="Weight (g)")

> barplot(MeanWt, main="Yield by crop", xlab="Compost source", ylab="Weight (g)", ylim=c(0, 5000), legend.text=c("Lettuce", "Spinach", "Chard"), args.legend = list(x="topleft"))

• Figure 5:

> par(mfrow=c(1,3))

> boxplot(letwt~Treat, data=Compost, main="Lettuce yield", xlab="Compost source", ylab="Weight (g)")

> boxplot(spiwt~Treat, data=Compost, main="Spinach yield", xlab="Compost source", ylab="Weight (g)")

> boxplot(chawt~Treat, data=Compost, main="Chard yield", xlab="Compost source", ylab="Weight (g)")

• Figure 6:

> par(mfrow=c(1,2))

> plot(PlantWt~Treat+Rep, data=Compost, ylab="Weight (g)")

• Figure 7:

> par(mfrow=c(1,2))

> plot(totcnt~Treat+Rep, data=Compost, ylab="Plant count")