Developing Soil Fertility Systems for Urban Agriculture in the Upper Midwest

Final Report for FNC12-874

Project Type: Farmer/Rancher
Funds awarded in 2012: $7,239.00
Projected End Date: 12/31/2013
Region: North Central
State: Minnesota
Project Coordinator:
Alexander Liebman
Stone's Throw Urban Farm
Expand All

Project Information


We tested the impact of different cover crops and the addition of compost on small-scale urban production systems. I specifically want to learn how cover crops can be best integrated into intensive vegetable production systems and how their impacts can be measured to ensure best practices. Rye/vetch, oats/peas, and clovers, with or without the addition of compost, were established following cash crops in the fall of 2012. I measured metrics of soil health such as organic matter, soil bulk density, and pH, in summer 2012 and spring 2013 to determine the before/after influence of cover crops. I also measured cover crop biomass to determine the extent to which different cover cropping methods contributed to soil organic matter.

Overall, soil organic matter and nitrate concentration increased across treatments with no clear association with cover cropping method. The pH decreased across all treatments, also with no connection to cover cropping or compost application. Bulk density levels varied largely yet overall increased across all treatments, with higher rates of increase where compost was applied with or without cover cropping.


Stone’s Throw Urban Farm is a 3-acre vegetable operation located on previously vacant lots throughout the cities of Saint Paul and Minneapolis in Minnesota. The farm produces food for a 100-person Community Supported Agriculture (CSA) program, the Mill City Farmer’s Market, and several restaurant and wholesale accounts. The farm is committed to growing high-quality produce and a sustainable livelihood while improving the ecological health of its land and engaging the surrounding community in on-farm education and equitable food access. Growing food in an urban setting has great potential for increasing local foods availability, decreasing transportation costs, and educating an urban populace on the pressing agricultural issues of our time. However, production systems are under-researched. We are intent on developing scientific, agro-ecological approaches to improving the capacity of urban agriculture.

We developed this experiment to learn how the integration of different cover cropping strategies could help our urban farm to improve soil health while not conflicting with production demands. As an urban farm, implementing agricultural production on previously marginal land, soil building and conservation strategies are imperative to the long-term success of our operation. However, due to our limited acreage and insecure land tenure (two major issues facing the development of urban agriculture), long-term cover cropping strategies are impractical. In addition, the short growing season of northern climates, such as Minnesota, creates a small window for the growth of both cover and cash crops. Therefore, we are experimenting with ways in which we can build soil structure and organic matter without sacrificing cash crop space.

This experiment was an initial step at determining 1) how cover-cropping methods fit into our production scheme and 2) how different cover cropping methods differed in their ability to improve soil quality. We experimented with the establishment of rye/vetch, oats/peas, and clover in the fall, after the harvest of a main crop or intercropped with an existing tomato crop.

Project Objectives:

In 2012, we had several main objectives. These were to:
– Establish cover cropping practice within existing farm system
– Measure soil bulk density for 2013 comparison
– Measure cover crop biomass for 2013 comparison
– Measure soil nitrogen, organic matter and pH for 2013 comparison
– Photo journal to track cover-crop growth
– Track costs/labor
– Host several outreach events to surrounding agricultural and community garden community

As summarized in the 2012 midterm report, these tasks were completed. During the 2013 season, our objectives were:
– Sample all plots for follow-up soil tests, bulk density measurements, and biomass measurement for comparison to 2012 collected data
– Data analysis: compile data and synthesize
– Continue outreach
– Qualitative descriptions of residue decomposition and success of succeeding crops These activities are summarized in the Accomplishments/Milestones section below.


Click linked name(s) to expand


Materials and methods:

I planted three types of cover crops: rye/vetch, oats/peas, and clover. Treatments were split between compost applied + cover crop, no compost + cover crop, and solely compost. Due to spatial constraints, clover was planted without a compost + clover comparison. Preceding crops varied across lots.

Seeding rates were determined by studying MOFGA (Maine Organic Farmers and Gardeners Association) recommended seeding rates. I increased the rates ~ 30% for direct seeding due to chronic problems with poor germination in the urban setting due to low soil moisture and poor soil fertility.


CROP, MOFGA (lbs/acre), STUF (lbs/acre), STUF (lbs/500 sq. ft.)
Rye, 15-30, 45, 0.50
Vetch, 50-60, 75, 0.86
Oats, 15-30, 45, 0.50
Peas, 50-75, 85, 1.00
Clover, 9-20, 30, 0.33

I modified the lbs./500 sq. ft. seeding rates to match bed dimensions at all sites.

All direct seeding used an Earthway Seeder. Oats/peas and rye/vetch were mixed, placed into hopper, and seeded with seed plate #1002-22 (beets/swiss chard). Clover was planted with seed plate #1002-5 (radish-med/leek/asparagus/spinach). Existing tomatoes were intercropped, with placement of seed as close as possible to tomato plant and spreading away from plant throughout the bed.

At one site (Oliver) and two beds at Aldrich, it was necessary to broadcast seed due to crop type/density. Existing eggplants, peppers, melons, and summer squash did not allow for use of Earthway seeder due to potential damage of harvestable crops. Broadcast seeding rates were seeded at double the MOFGA seeding rate.

All cover-cropping sites were planted during the first weeks of August. Sites where the cash crop had been removed were tilled. Sites with existing tomatoes, eggplants, or pepper stands were broadcast. Despite planting on schedule with a high seed density, cover crop germination and growth was extremely spotty. We believe this is due to two main factors. Lack of precipitation (no significant rainfall events in August and September 2012) impeded cover crop germination. In addition, in the intercropped cover-crop stands, soil disturbance due to harvest (cover crop was planted mid-August but we continued to harvest nightshades until late September thereby dragging bins across soil surface) may have disrupted growth and germination.

Soil samples were taken the first week of July and submitted to the University of Minnesota Research Analytical Laboratory for testing. We also sampled soil for bulk density measurements at that time. Bulk density soil samples were stored, dried for two days at 120 degrees Fahrenheit and measured on 10/21/12. Volumes were determined using a 100 ml graduated cylinder and density determined by dividing soil mass by soil volume.

Cover crop biomass was sampled 6 weeks after planting. We constructed a 1 foot x 1 foot square and cut all vegetation within square to measure cover crop growth in comparison to weed growth. We randomly sampled vegetation from two places per bed. Biomass was dried at 120 degrees Fahrenheit for two days and stored. Biomass was weighed on 10/25/12 at the University of Minnesota, Weed Ecology Lab of Professor Nick Jordan.

In the spring of 2013, sites were re-sampled for follow-up data. Biomass and soil samples were dried, stored, and weighed in fall 2013 at the University of Minnesota, Weed Ecology lab of Professor Nick Jordan. Data analysis was completed in December 2013.

Research results and discussion:

Organic Matter
We observed a small increase in percent organic matter across plots in the compost (C), clover-no compost (CNC), oats/peas – compost (OPC), oats/peas – no compost (OPNC) and rye/vetch – compost (RVC) treatments. In the rye/vetch – no compost (RVNC) treatment, we observed a small decrease in percent organic matter (-.5%). However, this decrease is not statistically significant. The largest increase occurred in the C treatment (+1.7%).

Due to our intensive use of compost to rehabilitate our soils, all soils had high initial levels of organic matter (average 8.5%). This is likely not representative of the entire soil profile, as compost is exclusively incorporated within the top six inches of the soil horizon. However, at the time of sampling, compost was well incorporated. This complicates my understanding of soil quality as I often equate health with levels of organic matter. Soil texture and the variability in organic matter sources are also very likely important factors in healthy soil.

Nitrate levels
We observed a large increase across all treatments from fall to spring. Initial readings appeared largely adequate (20-50 ppm) yet jumped to over 100 ppm in almost all treatments. As soil was sampled in May after a long, snowy spring, I would not have expected the large quantity of available nitrogen. We did not observe a discernable link to compost application or specific cover crop treatment. The highest change occurred in the OPC trial (+ 109 ppm) yet CNC (+101.9 ppm) and RVNC (+106.4 ppm) were higher than RVC (+64.2 ppm)

We observed a .3 – .6 decrease in pH levels across all treatments. The compost-only treatment had the largest decrease (.6). I had expected some decrease in pH in the compost treatments, as additions of nitrogen would increase acidification. I did not expect to observe such decreases of pH in the non-compost trials.

Soil bulk density
Density measurements varied widely, ranging from .728 – 1.25 g/ml in 2012 to .627 – 1.74 in 2013. Across all treatments there was an increase in bulk density. CNC had the smallest percentage increase (1.47%) while compost-only had the largest percentage increase (13.36%). Both RVC (11.67%) and OPC (9.82%) had higher percent increases than their non-compost counterparts, RVNC (10.95%) and OPNC (8.03%). I had approached this experiment from the viewpoint that compaction would be a serious issue facing our farm and measuring changes in density would be a good way to determine if root growth from cover cropping was increasing porosity. Compost and cover crops might increase soil bulk density in specific areas by binding soil particles but I am not convinced of this. As soil samples were dried, moisture is unlikely to confound this result.

Cover-crop biomass
OPNC and RVC treatments produced the highest amount of cover crop biomass at 1647.29 lbs./acre and 1607.27 lbs./acre, measured May 2013, respectively. Compared to October 2012 biomass measurements (six weeks after planting), OPNC had the greatest growth increase, from 135.4 lbs./acre to 1647.29 lbs./acre (change of 1511.89 lbs./acre). Compared to RVNC, this is quite significant (601.81 change, 233.86 lbs./acre to 835.17 lbs./acre). RVC performed similarly to OPNC, increasing from 196.87 lbs./acre to 1607.27 lbs./acre (change of 1410.4 lbs./acre). Unsurprisingly, estimated lbs. N per acre was linked to biomass growth, with the largest amounts for OPNC and RVC at 49.42 lbs./acre and 48.22 lbs./acre, respectively.

Based on this biomass data, it does not seem that there is a strict correlation between compost application and cover crop growth in our experiment. It is possible that rye/vetch is more responsive to compost application (RVC = 1607.27 lbs./acre versus RVNC = 835.17) than oats/peas (OPC = 714.81 lbs./acre versus OPNC = 1647.29 lbs./acre).

Surprisingly, clover decreased in sampled biomass from fall to spring, from 68.18 lbs./acre to 20.64 lbs./acre. I am quite intrigued by the potential of clover as a spring-planted intercrop beneath brassicas and nightshades. However, fall planting and subsequent spring incorporation does not seem to be the best use of clover.

For all of the above section please see attached EXCEL sheets for graphs and raw data.

Overall, I found no conclusive evidence across compost applications and cover crop trials that signal a specific best treatment for improving soil health. Longer-term studies would help to elucidate specific patterns across treatments.

Qualitative observations:
– Urban unpredictability
o Despite our best intentions, farmland in the city is tampered with by the abundance of neighboring residents. This existence in the quasi private-public domain led to a specific management difficulty. When measuring spring cover crop biomass, one of our sites, “Oliver” had been mowed to the ground. My suspicion is that a neighboring resident in this upper-class neighborhood had looked disapprovingly upon our tall, mature cover crop stands. I was not able to collect accurate biomass data from this plot for 2013 comparisons.

– Importance of fall moisture o As described in depth in the midterm report, the lack of fall moisture created difficult conditions for cover-crop germination and growth. In one plot, “Aldrich”, cover crops failed to establish. An unpredicted outcome was the variable success of intercropped cover crops based upon their proximity to drip-tape lines. Small leaks in drip-tape lines created highly successful growth pockets where, two feet away from the leakage, cover crop was significantly smaller and less dense. A few instances of overhead sprinklings would have greatly improved germination.

– Residue Decomposition
o Residue incorporation and decomposition varied greatly across treatments. The rye/vetch mixtures were the most difficult to incorporate. Successfully killing the plantings required two disc passes and two rototilling passes. On the contrary, oats and peas, having winterkilled were easy to incorporate. We performed all tillage and cover crop incorporation using a walk-behind, BCS tractor with disc and rototiller attachments. This equipment set-up was inadequate for easy cover crop incorporation. This seems like a unique challenge as many of our sites are too large for hand-powered tools while also being too small for four-wheeled tractor jobs.

– Cutworm
o Cutworm (family Noctuidae) was an issue in the 2013 spring. While we did not observe specific correlations across treatments, nightshade transplants suffered in cover cropping areas in one of our test sites, “Maryland”. We hypothesize that the combination of cool spring temperatures and dense residue provided ideal conditions for cutworms.

Impact of Results/Outcomes

We accomplished almost of all our initial objectives. As seen in the above section, we completed the data collection and analysis of our experiment. We also linked to the greater sustainable agriculture community in several ways, speaking at events, leading tours, and dialoguing with academics with a great wealth of knowledge in soil management systems. This has refined our young farm’s thinking and led to the creation of several additional research questions.

I was not able to accurately analyze hours spent implementing cover crops to discern their labor costs. I made a record-keeping error and did not separate normal cover-cropping implementation tasks from experiment-specific tasks. This is an important consideration, especially as we experiment with mulching, composting, and continued cover-crop trials. Small-farm operations must make decisions according to a wide variety of variables and labor is a major factor.

Cover crop impact on subsequent yield was omitted from the experimental design. High variability in spring planting schedules and crop choice made it impossible to implement this part of the experiment.

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

In 2012, outreach regarding the experiment included:

– Anne Pfeiffer, Community and Regional Food Systems, University of Wisconsin-Madison, toured the farm for an afternoon. She is also researching how cover crop systems may improve the fertility of urban farm/market garden systems. She is researching if spring-planted cover crops followed by late summer cash crops may be an effective means of improving soil fertility. She is also measuring yield differences in intercropped systems. We are planning to share insights as we gain more clarity regarding the outcomes of our experiments.

– University of Minnesota class, Ecology of Managed Ecosystems, taught by Professor Nick Jordan, visited the farm twice. About 60 students total walked through our farm plots in two separate groups. This was an excellent forum to display our intention in designing a cover-cropping system and showing the growth differences across treatments.

– I spoke at a University of Minnesota, What’s Up in Sustainable Agriculture (WUSA) event in mid-September. This informal 1-hour lecture period takes place over the lunch hour. About 10 students attended this session. The people attending my talk ranged from first-year agronomy students to student-run Cornercopia Organic Farm participants to graduate students studying corn/soy-cropping systems.

In 2013, we completed additional outreach work.

– University of Minnesota Extension hosted an urban agriculture expo in April 2013. I led a lecture on soil health and included findings from the experiment to highlight the potential impacts of cover cropping on urban production systems.

– I continued dialogue with Anne Pfeiffer, Community and Regional Food Systems, University of Wisconsin-Madison, who was completing a research report, titled, “Innovation in Urban Agricultural Practices Responding to Diverse Production Environments. This article is forthcoming in the journal of Renewable Agriculture and Food Systems.

– During the Ecological Society of America conference in August 2013, the New World Agriculture and Ecology Group visited our farm sites during two organized farm tours. Kirsten Nelson, professor at the University of Minnesota in the Department of Forest Resources, coordinated these tours. About 50 people, mainly professors and graduate students, visited the farm. During these tours, I highlighted our soil management practices and cover-cropping trials.

– University of Minnesota, Ecology of Managed Resources class visited again for two farm tours during October 2013.

Project Outcomes


Potential Contributions

On a farm-level, this experiment contributed greatly to our observational focus. By trialing several parts of our farm for this experiment, we closely observed changes in these areas. The record keeping elements of the experiment carried over to other management areas and we tracked farm data such as total yield to a much greater extent. It is our goal to create a research agenda and culture for our farm and this foray into on-farm experimentation helped us to learn what is necessary to achieve that goal. As we learn from the mistakes and shortcomings of this initial experiment, we feel better equipped to approach future research questions.

Throughout the experimental period, we maintained active dialogues with urban farmers throughout the upper Midwest, especially in the Twin Cities, Madison, and Chicago. It is our hope that in the coming years these urban farmers adopt and modify our cover-cropping strategies.

Future Recommendations

As the experiment progressed, many improvements to the experimental design became evident. It became clear that a variety of additional follow-up experiments would be necessary to refine our thinking. A few ideas for future work, summarized below are:

– Comparison of fall-planted cover crops to spring-planted cover crops o While fall-planted cover crops leave a short window for successful establishment, spring-planted cover crops can take advantage of a long growing season. Intercropped clovers or under-seeded oats/peas could be an important improvement for our operation.

– Evaluation of intercropping of cover-crops on nitrogen availability and yield of cash crop o How can intercropping of cover crops best be accomplished to limit negative impact on cash crops? o How does the timing of establishment impact the trade-off between cover crop establishment/growth and cash crop success?

– Comparison of the soil organic matter, pH, and microbial impacts of cover cropping versus other soil improvement strategies (i.e. mulching, composting)
o In nutrient-rich urban areas, available materials for composting and mulching are abundant. Despite the extremely necessary and valid focus on cover-cropping systems in rural areas on farms of a larger scale, cover crops may not be the optimal soil-building method for urban, intensive vegetable operations.
o What is the impact of composting without cover crops on soil biology and structure?
o How are the energy costs associated with composting vs. cover cropping related to increases in soil carbon? What methods create a net increase in soil organic matter and carbon?

– How can irrigation improve cover crop establishment and growth in a drought-year? How can we quantify the energy and economic costs associated with irrigation to determine the appropriateness of watering cover crops?

Our farm remains committed to asking these questions and developing methods to test them on our operation.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.