Final report for ONE22-411
Project Information
What if we could grow our cash crops in a perennial living mulch? What if the living mulch could cover the soil, suppress weeds and fix nitrogen? With technical assistance from UMass Extension, American Farmland Trust and others, Sawyer Farm started growing cash crops in a perennial Clover Living Mulch System (CLMS) in 2020. Our early results have been very promising, but yield drag is an issue for most crops. Figuring out which crops and varieties work best is important, but documenting the labor savings in CLMS is equally important: ultimately, we are concerned with profit per acre, not yield per acre.
Crops growing in the perennial Clover Living Mulch System at Sawyer Farm in 2022. These crops were all transplanted (or direct-seeded, in the case of dry beans) into a clover sod established in 2020. Left to right: green pole beans and cherry tomatoes (trellised), dry bush beans (direct seeded), delicata squash, peppers. Our study looked at two different methods to reduce clover-cash crop competition. First, we hypothesized that larger plug sizes would reduce competition, and second, that more frequent mowing events would reduce competition (but also increase labor). At Sawyer Farm, we grew three crops, and used three different mowing regimens across three different plug sizes, with a bare soil control. At Four Corners Farm, we grew two crops, with three different mowing regimens, with a bare soil control.
2023 was a historically wet year. All of our findings and all of the data comes with that caveat. For example, we lost almost all crops -- clover and bare soil -- at Sawyer Farm, due to its high water table.
Here are the conclusions we feel are justified based on the data and observation:
- Plug size resulted in more robust plant growth in clover and in bare soil. In other words, larger plugs probably lead to faster growth in most cases -- not just in CLMS. However, our no-till transplanter (Checci & Magli Dual Gold with custom no-till modifications) will not accommodate plug sizes over 50s, and treats 50s pretty rough, so 72s are our max size by default, even though larger plugs would probably perform better. Bottom line: Use as large a plug as you can, given equipment constraints, and greenhouse space constraints.
- Mowing regimen had no clear effect on yield. We were surprised. We think mowing may keep the clover in an aggressively growing vegetative state rather than letting it produce seed and senesce. Mowing doesn't seem to justify the labor, at least in terms of yield. Bottom line: Mow the clover right before transplanting, and mow it again 7-10 days afterward to ensure the clover doesn't overtop your transplant. Once your transplant is taller than the clover, there is probably no yield benefit to additional mowing, though you may want to mow again to set back any weeds that emerge through the clover canopy.
- At Sawyer Farm, we lost nearly all yield as plants slowly drowned. We got 0% yield from bare soil, whereas clover plots did yield something. There was really no data worth collecting; it was an all-around disaster. At Four Corners Farm, bare soil plots outyielded clover plots significantly for both cabbage and squash. This may be due to the fact that we didn't have enough downpressure on the transplanter packing wheels, so the transplants didn't have good soil contact. Or it may be due to less biological activity in that field compared to Sawyer. Or it may be due to insufficient sulfur and/or boron in that field. Or it may be due to the clover shading the soil so the saturated soil dried out more slowly, which was a bad thing in such a wet year. Bottom line: We don't know why CLMS plantings underperformed at Four Corners Farm.
- The soil health benefits of CLMS were super clear after just one season, especially in a wet year like this one. Bottom line: SOM, SOC, aggregation, respiration and nutrient levels were all higher in CLMS than in bare soil at the end of the season. We observed one erosion event after another in the bare soil, while we only saw clear water running off of the CLMS plots.
Four Corners Farm in late June, 2023, after a heavy rainfall. Cabbage transplants, ~2 weeks old in clover (on the left) and bare soil (on the right).
We disseminated our results through a field walk sponsored by CISA, and we presented our findings at NOFA-MA, Iowa Organic Conference, MOFGA, and the MA State Soil Health Technical Subcommittee. This SARE-funded project inspired us to launch Momentum Ag, which now funds on-farm CLMS research at 20 farms across the Northeast. For more information, please visit www.momentumag.org.
The project seeks to identify key factors that affect soil health, vegetable yields, and labor in a perennial clover living mulch system (CLMS).
- For each crop (cabbage, winter squash, and paste tomatoes), we will identify the smallest plug size growers can use without experiencing yield drag.
- For each crop, we will identify whether mowing clover according to variable high frequency versus variable low frequency affects yields, and track labor across both mowing regimens. Our data will provide information to guide growers in determining the appropriate mowing system for their operation.
- Each crop will have an organic, tilled control that will allow us to measure labor and yield differences between CLMS and the control. This data will empower farmers to make informed economic decisions about CLMS.
- Soil health indicators will be measured at the end of the 2023 season in CLMS versus control plots.
Organic no-till is in its infancy. Despite significant research efforts, organic no-till systems have yet to be adopted in any meaningful way. At Sawyer Farm, increasingly intense rainfall events led us to search for less tillage-intensive alternatives to the cover crop-based systems we’d been using since 2010. Despite rigorous cover-cropping regimes and various tillage-reduction techniques, we had unacceptable levels of erosion on our sloping vegetable fields. In the last few years, I have become very involved in a number of no-till farmers’ groups, and have reached out to dozens of academics and researchers to help find reduced tillage solutions that might work for Sawyer Farm. The results of this research have confirmed that we are not alone in our search, and that many, or even most, vegetable growers struggle to find no-till/low-till systems that fit their farm.
We first tried transplanting into rolled rye, a system developed at the Rodale Institute, in Kutztown, PA. It was difficult to establish rye early enough, because most of our crops are still in the ground in September, and because rye does not reach anthesis in Western MA (zone 5b) until mid-June, which is after most of our crops need to be established. It generally requires an annual tillage event, to establish rye after crop harvest, which means that we had bare soil right at the time of year when increasingly intense hurricane-driven rain events are most likely.
We then tried tarping – widely used at other small farms in our area – but it requires an amount of labor that makes it unprofitable for most of our crops (we mostly grow low-value storage crops like cabbage, onions, and carrots). It also trades tillage for plastic; the solution we’re hoping for won’t require this trade-off in terms of externalities.
Finally, we had success with a modified transferred mulch system, but the amount of hay needed puts a natural limit on the scalability of that system (it takes a minimum of five acres of hay to adequately mulch one acre of vegetables). To scale the system would have meant huge inputs of hay grown off-farm. That system also has mechanization problems – there is no transplanter that will reliably perform in 6-8” of hay mulch.
My initial interest in transplanting into perennial clover came out of a decade of experience with undersowing Dutch White clover into established crops. We wondered if it would be possible, instead of plowing down the clover before spring planting, to transplant directly into it. I found an abundance of academic research regarding clover living mulches (detailed in the Previous Work section), but none that described an actual production system. I couldn’t find any farmers using a perennial clover living mulch system (CLMS).
In 2020 and 2021, we began implementing CLMS. Though not every crop and every experiment has been a success, CLMS has worked extremely well in many cases and is well worth pursuing. It vastly reduces the negative soil health impacts of tillage; is well-suited to the Northeast’s growing conditions; and, while unlikely to lend itself to certain crops (e.g. direct-seeded carrots), we feel confident that, with modifications and improvements, it can work for nearly all transplanted crops, and possibly large, direct-seeded crops as well. It requires little in the way of equipment, and offers a seasonal flexibility that is appealing to growers. It is a very scalable system – a one-acre grower can use a $15 bulb auger and a lawnmower, while a 100-acre grower could use a no-till transplanter and a row-mow unit. Most importantly, CLMS requires only one tillage event every five years, to renew the clover stand.
However, competition between the clover and the cash crop can lead to yield drag. This competition and the resultant yield depression is the recurring theme of academic papers on perennial clover living mulches. Some of that loss can be minimized by system modifications, such as plug size and mowing frequency (the subject of the proposed study), and some can be compensated for by reduced labor costs (which we propose to track across CLMS versus tilled controls). Plug size is critical – the larger the plug, the more aggressively the transplant can compete with clover during the crop's establishment phase. On the other hand, larger plugs require more greenhouse space, more potting soil, and reduce transplanting efficiency. Correlating plug size and yield would allow us to identify the minimum plug size needed for success. Mowing also reduces the competitiveness of the clover and thus increases yields. On the other hand, each mowing event adds significant labor.
To my knowledge, transplant size has not been studied in a perennial clover living mulch system, though in our experience this factor is critical to success. Mowing has often been studied (see Previous Work), but the frequency seems to be chosen arbitrarily -- once versus twice; twice versus four times, etc. As a farmer, I know that every management technique is highly specific to the crop and the season. Arthur and I have developed two data-driven mowing regimens that would be useful in guiding farmers' attempts to implement CLMS (please see Materials and Methods for details).
Our goal is to gather data to help growers make their own decision about whether CLMS could work for their crop mix, land base, and labor conditions.
Most producers operate on extremely tight margins and are loath to try new systems while juggling loans, payrolls, and overhead. They immediately recognize the potential benefits of CLMS, but need hard data on yield drag, baseline suggestions for plug size and mowing regimens, and labor assessments in order to justify their own trials.
Cooperators
- - Technical Advisor (Educator)
- - Producer
- - Technical Advisor (Researcher)
Research
Field Sites (Objectives 1, 2, 3, and 4)
The fields at Four Corners Farm and Sawyer Farm are sandy loams. Four Corners Farm field management history: Until 2017, the field was in conventional silage corn-on-corn. From 2018-2021 it was fallow. In the fall of 2021, the field was plowed and seeded to rye. In 2022, the field grew a variety of vegetables, nearly all of which were undersown with clover in July. In 2023, the control plot was in a block that had not been seeded to clover, and the experimental plots were in the clover. Sawyer Farm field management history: The main 4 acre block at Sawyer Farm has rotated between a diversity of vegetables since 2010, with extensive cover cropping (primarily undersown clover, but also oats/peas and rye). The block chosen for the experimental plots was in three-year-old Dutch White clover. It was immediately adjacent to a block that had been in an annual cover crop fallow year. In August 2023, the control block was seeded down to oats and peas.
Experimental Design (Objectives 1 and 2)
Both experiments explored vegetable production in a perennial clover living mulch system (CLMS). In spring, vegetables were planted into the clover to explore the relationship of transplant plug size (Objective 1) and clover suppression (Objective 2) to vegetable yield. Labor time for clover mowing and weeding for CLMS vegetable production was compared with organic bare-soil production in control plots (Objective 3).
At Sawyer Farm, for each of three vegetable crops (cabbage, winter squash, and paste tomatoes) a 500-foot bed was planted in Dutch White clover. The clover was established in July, 2021. Another 500-foot bed of each crop was planted in bare soil. These crops were selected for their diversity in terms of families and growth habits. The CLMS and bare soil plots were adjacent in order to minimize in-field variability.
Each vegetable crop was planted in a split-plot randomized complete block design with three blocks. Suppression level was the main plot and plug sizes were the subplots.
There were two suppression treatments, high and low. The first maximized suppression by mowing clover before its regrowth/recovery period has been achieved. Recovery periods vary by temperature, moisture, and nutrients. On-farm observations indicate that full recovery can occur within as few as ten days during early spring growth, whereas mid-late summer recovery times can take a month or more. In the high suppression treatment, clover was mown at a height of 4", well before the Dutch White clover reaches its 6"+ maximum. (This treatment is referred to as high suppression in the "Layout" upload in supporting materials.) Mowing frequency thus varied according to clover recovery times. The second suppression treatment was also variable frequency, but was less intensive overall (This is referred to as low suppression in the "Layout" upload.) Mowing occurred according to the ratio between clover height and crop height. The clover was mown when its height was greater than 1/3 of average crop height (or width, in the case of squash). This treatment required more frequent mowing during crop establishment, but mowing ceased when crops reach ~18" in height or width (Dutch White clover growth typically peaks at 6"). These mowing treatments are in contrast to fixed-frequency mowing treatments found in LM research, and aim to quantify Sawyer Farm's on-farm PCS management system.
At Four Corners Farm, two crops were trialed -- Ruby Perfection storage cabbage, and Delicata winter squash. All plugs were 128s. Mowing treatments were the same as those utilized at Sawyer Farm. Beds were 150' long. Control beds were adjacent to clover beds.
Agronomic Management (Objectives 1, 2, 3, and 4)
Tomatoes and winter squash were planted one row per bed at 20-inch in-row spacing. Cabbage was planted in two rows per bed at 20-inch in-row spacing. At Sawyer Farm, each subplot of tomatoes and squash contained 40 plants, and cabbage 80 plants. At Four Corners Farm, there was no plug-size treatment, so each 150' row was its own plot with different mowing suppressions. Each plot contained 90 squash plants or 180 cabbage plants.
The plants will be started in the greenhouse between 2 and 8 weeks before transplanting in late May, as follows: Squash 2 weeks (72 tray), 3 weeks (48 tray), and 4 weeks (2.5" pot) prior to transplanting. Cabbage 5 weeks (72 tray), 6 weeks (48 tray), and 7 weeks (2.5" pot). Tomato 6 weeks (72 tray), 7 weeks (48 tray), and 8 weeks (2.5" pot).
Dutch white clover was mowed before transplanting and bare-soil plots were prepped with a stone burier for transplanting. To plant into clover, a modified Checci and Magli Dual Gold no-till transplanter was used. The same modified planter worked well on the bare soil control plots.
Following transplanting, clover suppression treatments was implemented by mowing the clover according to low/high treatments shown in the treatment layout. A push lawnmower and string trimmer were used to do the mowing. Bare-soil control plots were cultivated with tractors or by hand, as needed. In the bare-soil plots, the number of cultivations for each crop and the average cultivation time during the growing season was recorded so that weeding labor time in bare-soil plots could be compared to both clover suppression treatments.
The vegetables were be fertilized with organic fertilizers following soil test recommendations.
All plots were observed weekly to note any differences in pest pressure or growth habit and qualitative notes were collected to document farm-scale issues, plot, or treatment differences.
The vegetable crops at Four Corners Farm were harvested in early fall. At Sawyer Farm, we recorded yield data where possible, but we essentially lost the entire harvest in both clover and on bare soil due to heavy rains, a high water table, and waterlogged soils. Though clover plots outyielded bare soil plots, all yields were so dramatically low that we don't think there are any conclusions worth drawing from Sawyer's harvest data.
Data Analysis (Objectives 1 and 2)
At Sawyer Farm, data analysis was severely compromised by the weather. Early observations, measurements and photographs indicated that the larger plug sizes outgrew smaller plug sizes in both clover and bare soil, but those crops died before we were able to collect harvest data.
At Four Corners Farm, data analysis was possibly compromised by transplanter setup issues and lack of good root ball to soil contact at transplanting. Still, there was no relationship between mowing frequency and yield. This is important information for growers going forward.
Comparison of PCS and bare-soil production (Objective 3)
The best performing PCS treatments were compared with standard bare-soil production methods in terms of yield variables, labor time, logistical complexity, soil health assessment and any other differences. These comparisons were not statistically analyzed since it is was not possible to randomize bare-soil and PCS plots at either farm given basic management constraints at these two commercial farms. The differences in crop yields were considered with differences in labor time and soil health.
Soil Health Comparison (Objective 4)
Following harvest, composite soil samples were collected in collaboration with American Farmland Trust from both living mulch plots and tilled plots and sent to Cornell University for Comprehensive Assessment of Soil Health in PCS and bare-soil plots. These composite soil samples were analyzed for soil pH, organic matter, wet aggregate stability, active carbon, total carbon, total nitrogen, and total protein.
In the end, due to rain and water-table related crop loss at Sawyer Farm, most of our results were derived from Four Corners Farm data.
At Four Corners, we suspect that there was some issue with the newly modified no-till transplanter. We likely should have had more weight on the rear to pack the plugs into the clover. Transplanters are designed with bare soil in mind, and even "no-till" transplanters generally use a conventional design with some small modifications, so getting some loose soil firmly packed around the rootball without significantly damaging the clover sod and opening up a bare soil strip for weeds is a challenge. This may not have been an issue in a normal year, but we had a drought from mid May through late June and there is no irrigation at Four Corners, so poorly packed plants likely suffered a good deal. (There is a water tank on the transplanter, but that only goes so far...).
We think this led to the loss of many transplants in the clover plots at Four Corners, and may have stunted others.
However, something else was going on at Four Corners, too. Those plugs that survived the drought kept up with or even outperformed the bare soil plots throughout July. By mid-August, however, the crops in clover kind of 'stalled out' and failed to keep pace with the bare soil crops. We sent in sap samples for analysis but saw no stark differences between clover and bare soil crops. (All crops were 'washed out' though and below optimum range for most nutrients, probably due to heavy rains and leaching).
We have two theories for this stalled growth. First, the soil dried out more slowly under the clover than on bare soil, probably due to shading, but possibly also due to better infiltration. In a normal year, moisture retention in July is a good thing, but this year it wasn't. So it may be that the bare soil crops had an extra week or two during a critical stage in their growth when there was some oxygen in the rhizosphere. This was our pet theory during the growing season, but end-of-season soil tests showed much higher soil respiration rates in the clover than in bare soil -- unsurprising, but perhaps undermining the idea that the soil was oxygen deficient in the clover through the season.
Our second theory comes from looking over the soil health test results at the end of the season. Every nutrient was higher in the clover except for sulfur (and, to a lesser extent, boron). The clover 'holds' about 15 ppm/acre of sulfur in its living tissue. Only 5 ppms of sulfur were left over for the cash crop. Perhaps sulfur levels were adequate in the earlier part of the season, but were not enough to meet our crops' needs as they matured.
Overall, we were disappointed in the clover yields at Four Corners. At Sawyer Farm, in previous years, cabbage and tomatoes had consistently performed as well or very nearly as their bare soil counterparts. Squash had been a bit more variable. So we had expected closer yields between clover and bare soil at Four Corners.
Still, labor was -- as anticipated -- significantly lower in the clover than in bare soil. The crew enjoyed working in the clover way more than in the bare soil. And in a wet year like this one, it was depressing to watch soil erode out of the bare soil sections. The labor savings weren't enough to compensate for the dramatic yield losses we saw in clover-grown crops at Four Corners this year, but they do offset >15% yield loss. This coming season, we will be keeping a very close eye on (1) good rootball to soil contact at transplanting and (2) adequate sulfur going into the season.
* We think that the timing of the soil test is important when calculating sulfur amendments. If the test is taken before the clover is well established, we will aim for 15 ppms above recommended levels, assuming that 15 ppms will be picked up by the clover. If the test is taken while the clover is well established, we will aim for levels on the high end of optimum.
We sought to reduce clover/cash crop competition through larger transplant plug sizes and different mowing regiments, and to track labor and yield to assess profitability. 2023 was incredibly wet and we lost important data points (especially at the Sawyer Farm site).
- Plug size resulted in more robust plant growth in clover and in bare soil. In other words, larger plugs probably lead to faster growth in most cases -- not just in CLMS. However, our no-till transplanter (Checci & Magli Dual Gold with custom no-till modifications) will not accommodate plug sizes over 50s, and treats 50s pretty rough, so 72s are our max size by default, even though larger plugs would probably perform better. Bottom line: Use as large a plug as you can, given equipment constraints, and greenhouse space constraints.
- Mowing regimen had no clear effect on yield. We were surprised. We think mowing may keep the clover in an aggressively growing vegetative state rather than letting it produce seed and senesce. Mowing doesn't seem to justify the labor, at least in terms of yield. Bottom line: Mow the clover right before transplanting, and mow it again 7-10 days afterward to ensure the clover doesn't overtop your transplant. Once your transplant is taller than the clover, there is probably no yield benefit to additional mowing, though you may want to mow again anyway, to set back any weeds that emerge through the clover canopy.
- At Sawyer Farm, we lost nearly all yield as plants slowly drowned. We got 0% yield from bare soil, whereas clover plots did yield something. There was really no data worth collecting; it was an all-around disaster. At Four Corners Farm, bare soil plots outyielded clover plots significantly for both cabbage and squash. Bottom line: While reduced labor hours in the clover plots somewhat compensated for the yield loss, the bare soil plots were still more profitable. Further work needs to be done to optimize management and increase yields in CLMS. One intriguing piece of data that emerged from the soil tests indicates that sulfur may be particularly important in managing CLMS. All nutrient levels were higher in clover than in bare soil at the end of the season (likely due to less leaching), with the glaring exception of sulfur (5 ppms in clover; 20 ppm in bare soil). We hypothesize that clover aggressively forages for S, and that S should be at the high end of optimal ranges in CLMS in order to accommodate the needs of clover and cash crops. There may be a similar dynamic with boron.
- The soil health benefits of CLMS were super clear after just one season, especially in a wet year like this one. Bottom line: SOM, SOC, aggregation, respiration and nutrient levels were all higher in CLMS than in bare soil at the end of the season. We observed one erosion event after another in the bare soil, while we only saw clear water running off of the CLMS plots.
Four Corners Farm, Worthington, MA. Cabbage ~2 weeks after transplanting in clover (on the left) and in bare soil (on the right).
Education & Outreach Activities and Participation Summary
Participation Summary:
CISA, AFT, ATTRA, and Berkshire and Hampshire/Hampden County Conservation Districts hosted a very successful field day at Four Corners Farm on July 20, 2023. There were 60 attendees, including farmers and ag professionals who got to see CLMS in action.
I had the opportunity to present our SARE findings at two different sessions of the Maine Organic Farmers and Gardeners Association annual meeting. One event had 40 participants, and the other had 15.
I also presented our results in two sessions at the Iowa Organic Conference to a total of 90 farmers and ag professionals.
Piyush Labhsetwar volunteered on the SARE project this season and helped conceptualize the results. He presented to an audience of 60 at NOFA-MA's winter conference.
I presented to a group of 15 NRCS employees and other members of the Massachusetts State Soil Health Technical Subcommittee.
ATTRA staff visited Four Corners Farm and Sawyer Farm to learn more about CLMS. They published a blog on the practice, and released the first of a four-part video series on CLMS.
I will present our findings at a CISA-sponsored webinar this March.
Learning Outcomes
Many farmers expressed an interest in CLMS. They primarily cited (1) CLMS ability to hold soil in place in a wet year like 2023. (2) The aesthetic appeal of CLMS. (3) The potential for CLMS to reduce tillage at scale.
Project Outcomes
This project generated a lot of farmer interest and we were able to launch a non-profit -- Momentum Ag -- to help fund CLMS trials. 17 farmers established clover in 2023, and will be running CLMS trials in 2024. Momentum's first annual winter meeting took place in December. Farmers discussed BMPs for clover establishment and began designing trials.
Though this SARE study left many questions unanswered in terms of BMPs for CLMS, the initial results, and the aesthetics of the clover living mulch plots themselves were enough to encourage other farmers to trial the practice on their own farm.
If you are interested in joining our community of farmers implementing and problem-solving CLMS, you can apply at www.momentumag.org/farm-enrollment. Participants' receive trials funding, free technical assistance, and free soil testing.
The weather fundamentally prohibited good data collection in 2023. We lost most crops at Sawyer Farm. All yield data from both farms and across experimental and control plots was flattened, making conclusions difficult.
However, much valuable data and many valuable insights were gained. See Summary.
We would not make major changes to methodology if we were to do it again, but we would cross our fingers and hope for a more 'normal' year.
There is so much we still need to learn about CLMS! N and S dynamics are important and interesting. Soil moisture dynamics and infiltrations rates are fascinating. Continuing to document SOC and SOM. Crop species and varieties trials. Temperature reduction and stability. These questions all bear on optimizing agronomic management for CLMS' success.
We feel that CLMS has promise in the humid temperate zone from the Northeast and MidAtlantic across to Eastern KS and up to MN. Row croppers, small grain growers and vegetable producers would all benefit from considering how CLMS might build soil health and increase profitability.