Evaluating Scale-appropriate Technology for Organic No-till Vegetable Production

Cover Crop Termination

Cover crop termination was successful (>90%) across all equipment types and termination methods 3 weeks after the late termination date, although termination success was somewhat higher in late-crimped treatments compared to early-crimped treatments (99% vs. 95%; p<0.001). The result that termination success was uniformly >90% across treatments indicates that - at least in this location and year - larger-scale (heavier) equipment was not necessary to achieve successful rye termination, and that delaying termination (eg. from late flower to late milk) is not required for small-scale equipment, thus opening up the cropping season by approximately two weeks. Earlier cash crop planting may or may not result in a yield benefit for all crops, but is generally thought of as a way to increase marketable yields across the season by enabling the use of longer-season varieties (often higher-yielding) or by increasing the harvest window before pest pressure limits yield. 

Weed Pressure

Weed pressure at Lomax Farm, measured by both peak weed biomass (sampled in July) and time spent weeding, was not affected by equipment type (tractor, walk-behind tractor or manual) because weeding was performed manually across treatments. Interestingly, weed pressure was not affected by termination timing (early vs. late-crimped), despite somewhat higher cover crop biomass in the late-terminated treatment (7408 lb/acre vs. 6710 lb/acre, p=0.0099). However, weed pressure was strongly affected by cover crop & weed management method (crimp, mow, landscape fabric or till), where weed biomass was, unsurprisingly, lower in plots treated with landscape fabric (143 lbs/acre) compared to other methods (983-1168 lbs/acre; p=0.008). It was somewhat surprising that weed biomass in the tilled treatments averaged 1047 lbs/acre, where weeds grew only along plot edges where soil was tilled but not covered in plastic mulch. Time spent weeding at Lomax followed a similar trend: weeding time was lowest in the landscape fabric treatment (14 hours/acre), highest in crimped and mowed treatments (179 and 281 hours/acre, respectively), and intermediate in tilled treatments (35 hours/acre; p<0.0001). Weed biomass was high relative to weeding time in tilled treatments because weeds grew only along plot edges, making their control quicker than in crimped and mowed treatments, where weeds were evenly distributed. 

Weed pressure at Red Scout Farm followed a similar pattern to Lomax Farm, although weeding time was substantially higher on a per-acre basis, owing to a zero weed tolerance policy at Red Scout Farm in combination with a much smaller study area that was more manageable with limited staff. Weed biomass was not quantified at Red Scout Farm as a result of this policy. Time spent weeding was lowest in plots with landscape fabric (15 hours/acre) and much higher in other treatments, especially mowed treatments (494 hours/acre). In contrast to Lomax Farm, weed pressure at Red Scout Farm appeared to be influenced by termination timing, with early-crimped treatments requiring more time to maintain weed free (357 hours/acre) compared to late-crimped treatments (241 hours/acre). While treatments at Red Scout Farm were not replicated (although crimping timing was effectively duplicated and consistent across walk-behind & manual methods), and results should be interpreted cautiously, the difference in weed pressure between early- and late-crimped treatments at Red Scout Farm suggests that in some years or locations - particularly where weed pressure is high -  growers may benefit from delaying crimping to maximize cover crop biomass, possibly reducing weed pressure over the course of the season. 

Finally, regarding time spent weeding across treatments: weeding was performed manually across all treatments, including treatments based on tractor or walk-behind tractor use, and it’s likely that weeding time could be reduced in those treatments if a suitable high-residue cultivator was used for weed control. Because weeding time drove total labor hours (see below), the ability to take advantage of existing tractor-scale equipment (or custom-built equipment for a walk-behind tractor) could dramatically improve the profitability and spatial thresholds for tractor and walk-behind tractor production systems. 

Yields

Butternut squash yields at Lomax Farm are not reported due to a near crop failure that resulted from a suite of production challenges, primarily pressure from Lolium, Bromus and Digitaria weed species and from squash bugs (Anasa tristis). Yields at Red Scout Farm, however, were much higher, and ranged from 3.5 - 8.8 lbs/plant (15,125 - 38,100 lbs/acre; average of 27, 280 lbs/acre) across treatments. Yield was highest in the landscape fabric treatment (38,100 lbs/acre), and lowest in late-crimped treatments (17,105 lbs/acre averaged across manual and walk-behind). Yield in mowed and early-crimped treatments were intermediate (25,385 and 33,000 lbs/acre, respectively).  It is not clear why yields were low in both late-crimped treatments (walk-behind tractor and manual), where weed pressure was low. Because treatments were not replicated, yield data from Red Scout Farm were averaged across treatments in the subsequent economic analysis, and no yield data from Lomax Farm was used. 

Production Costs

The cost of butternut squash production (for one acre, per year) for each system at Lomax Farm varied widely from $4,317/1-acre - $8,145/1-acre (Figure 1A). These production costs were driven, first, by labor costs ($1544/1-acre - $4365/1-acre), which in turn were driven by time spent weeding (Figure 2), and therefore were strongly influenced by treatment. Labor costs were lowest in the tilled treatments ($1,688/1-acre, averaged across tractor and walk-behind tractor), where beds could be set up efficiently and where weed pressure was low on account of plastic mulch. While time spent weeding was lowest in the landscape fabric treatment, labor costs were somewhat higher ($2,611/1-acre) than in tilled treatments because the application and breakdown of landscape fabric is not mechanized and must be performed manually, unlike plastic mulch in the tilled treatments. Labor costs were highest in the mowed treatments ($4,290/1-acre, averaged across tractor and walk-behind tractor) due to time spent weeding, and intermediate in the crimped treatments ($3,147/1-acre, averaged across all six crimped treatments). Equipment costs were the secondary driver of total costs for one acre per year, where the annual cost to own a tractor and associated equipment was the most expensive ($2,912/1-acre averaged across tractor-based systems), especially equipment for the tilled system ($3,375/1-acre). The annual cost to own a walk-behind tractor and associated equipment was approximately half that of a tractor ($1,412/1-acre averaged across walk-behind tractor-based systems), and the cost of equipment for the tilled system was highest in this category as well ($1,725/1-acre). Among manual treatments, the cost of the manual crimper was negligible ($2/1-acre) compared to tractor- and walk-behind tractor-based systems. The cost of materials, fuel and maintenance was consistent across all crimped and mowed treatments (average of $1,068/1-acre), but higher for tilled ($1,500/1-acre) and landscape fabric ($1,760/1-acre) treatments, owing to the cost of plastic mulch and landscape fabric. Total cost of production at Red Scout Farm was higher overall due to more time spent weeding, but followed a similar pattern across treatments, with overall costs similar among manual treatments (crimped and landscape fabric), highest in mowed, and intermediate in roll-crimped treatments (Figure 1B)

Breakeven Yields and Profitability

Breakeven yields, based on butternut squash value of $2/lb, ranged from 2,159 - 4,073 lbs/acre at Lomax Farm, were lowest in manual treatments (due to low production costs), and highest in the tractor-based treatments (due to high production costs). Breakeven yields in mowed treatments were higher than all other treatments (4,073 and 3,365 lbs/acre for tractor and walk-behind tractor treatments, respectively). Breakeven yields were higher at Red Scout Farm due to the higher cost of production (from weeding time), ranging from 3,270 - 5,415 lbs/acre, and followed a similar pattern across treatments. Because achieving $2/lb squash value may be limited to direct-to-consumer markets, where 1+ acre of a single crop is not often grown, two other $/lb scenarios were used to determine breakeven yields: $1/lb and $0.50/lb. In these scenarios, breakeven yields ranged from 4,317 - 8,145 lbs/acre ($1/lb) and 8,635 - 16,290 lbs/acre ($0.50/lb) at Lomax Farm. While yields were only reported for Red Scout Farm, butternut squash yields were substantially higher than breakeven yields, averaging 27, 280 lbs/acre across treatments. Because of this, butternut squash could be valued much lower to breakeven at 27,280 lbs/acre - as low as $0.40/lb in the most costly-to-produce treatments, mowed. 

Profitability (net $/acre), determined only for Red Scout Farm due to crop failure at Lomax Farm, ranged from $43,730/acre - $48,019/acre, using average squash yield (27,280 lbs/acre) and squash valued at $2/lb. Similar to analysis of breakeven yield, we explored two other valuation scenarios to determine profitability when squash is valued lower: $1/lb and $0.50/lb. In the $1/lb scenario, profitability ranged from $16,450/acre - $20,739/acre, and under the $0.50/lb scenario, profitability ranged from $7,810/acre - $7,100/acre (based on average yield value of 27,280 lbs/acre). 

Spatial Thresholds of Production Costs and Time

The spatial analysis indicated that, overall, when scaled from 0.1 - 10 acres, the total cost of production increased linearly with acreage for all treatments, but increased the most per acre in mowed treatments, and the least per acre in tilled treatments, owing primarily to differences in time spent weeding. At larger acreages (greater than approximately 3 acres), cover crop and weed management methods (crimp, mow, landscape fabric or till) were more important for total production costs than the equipment used (tractor, walk-behind tractor or manual), with mowed treatments being the most expensive, tilled treatments the least expensive, and crimped and landscape fabric treatments intermediate (Figure 3A). At smaller acreages (less than than approximately 3 acres), equipment type was more important for total production costs than cover crop and weed management methods, with tractor treatments the most expensive, followed by walk-behind tractor, followed by manual (Figure 3B). This transition in the importance of equipment at smaller acreages to method at larger acreages occurs between 0.8 and 2.8 acres, depending on method or equipment (Figure 3B). Thus, the critical farm acreage for selecting equipment and method (based on total production costs) ranges from approximately 1-3 acres, depending on preferred method or equipment. 

When cost data are expressed on a per-acre basis (distinct from total cost, presented above), per-acre costs decreased with increasing acreage for tractor- and walk-behind tractor-based treatments, as economies of scale are achieved and the cost of equipment is spread across increasing acreage. In contrast, per-acre costs of manual treatments remained essentially the same with increasing acreage, because economies of scale can not be achieved without equipment (Figure 4). Similar to total production cost results, per-acre costs were highest in mowed treatments, lowest in the tilled treatments (above 2 - 2.5 acres), and crimped and landscape fabric treatments were similar (above 3 acres) and intermediate between these. These results indicate that per-acre costs are more sensitive to cover crop and weed management methods (crimp, mow, landscape fabric and till) than to the equipment used (tractor, walk-behind tractor, and manual) above 2 acres (although per-acre costs of mowed treatments were highest regardless). Results also indicate that manual methods are surprisingly efficient ($/acre spent) and competitive with tractor- and walk-behind tractor-based technologies: they’re cheaper per acre than roller-crimper (mechanized) treatments up to at least 5 acres, and consistently cheaper per acre than mowed treatments, regardless of acreage. Similarly, walk-behind tractor treatments were cheaper ($/acre) than their tractor-based counterparts to at least 5 acres. These results indicate that per-acre costs of tractor-based treatments are more expensive than manual or walk-behind treatments up to at least 5 acres. It should be reemphasized, however, that production costs were driven by time spent weeding, and weed control in tractor and walk-behind tractor treatments was performed manually; thus, it’s likely that production cost could be decreased substantially in these treatments if equipment-based weed control (using high-residue cultivation) could be implemented.

The final component of the spatial analysis is the effect of each treatment on maximum farmable acreage as a function of available labor hours (40 hours/week for the 5-month production period is approximately 850 hours). Maximum acreage achievable in this 5-month full-time equivalent scenario was, among no-till treatments, highest for the landscape fabric treatment (3.8 acres), followed by crimped treatments (3.2∓0.2 acres), followed by mowing treatments (2.4∓0.1 acres; Figure 5). Maximum farmable acreage in the tilling treatments was highest (5.6 acres, walk-behind tractor; 6.6 acres, tractor). Similar to total cost and per-acre cost results, if suitable weed control equipment were used in tractor and walk-behind tractor treatments, then the maximum farmable acreage could increase substantially.

Conclusion

In conclusion, among the no-till methods compared in this study, it is clear that crimping was preferable to mowing, given the intense weed pressure and resulting drag on both profits and yields. In this particular year at Lomax Farm, where weed pressure was high and weeding was performed manually across all treatments, the cost of weeding drove economic and spatial results, making manual methods efficient ($/acre and hours/acre invested) and profitable, and competitive with their mechanized counterparts (especially below 3 acres). Additionally, the relatively low hours/acre invested when using landscape fabric enables more acreage to be covered (3.8 acres in 850 hours) compared to crimping and mowing (2.4 - 3.2 acres in 850 hours), and if higher yields on landscape fabric (as demonstrated at Red Scout Farm) are consistent across locations and years, it may prove to be the most profitable no-till method - limited perhaps by available labor hours and not upfront costs. 

Finally, the fact that tilled treatments outperformed no-till treatments by most metrics (weed pressure, costs per acre, scalability) - owing mostly to decreased weed pressure, resulting in less time invested and higher yields per acre - highlights the need for improvement in organic no-till systems.

We see two options for improving organic no-till vegetable production:

Option 1) To mechanize weed control in crimped or mowed systems - with a strong preference for crimping to mowing - to drive down overall production costs by decreasing time spent weeding, and to potentially improve yields. Cover crops should also be managed to ensure high biomass and consistent termination for good weed suppression, and soil nutrients should be managed to account for nitrogen tie-up and the slow release of nitrogen by cover crops upon decomposition. While fertigation technology can address in-crop nitrogen limitation imposed by slow-to-decompose cover crop mulches, and hand weeding can be used in high-value crops, it is our opinion that crimped and mowed systems for most vegetable crops will always be hamstrung by limitations in in-crop weed control (especially by in-row weeds), regardless of whether high-residue cultivation can be implemented. This is because in crimped and mowed systems there is a high risk of weeds setting seed in any given year, leading to higher weed pressure and ever-more-challenging weed control in subsequent years.

For this reason we are in favor of Option 2) development of a new technology to mechanize the application of landscape fabric or similar material to provide season-long weed suppression and improve nutrient supply to crops from more rapid decomposition of cover crop mulches. A technology such as this, if affordable like many plastic mulch layers, could bring successful organic no-till vegetable production within reach for many more farmers, and have similar yield and economic benefits that plastic mulch had several decades ago, as well as significant benefits to soil health and on-farm conservation.

Thus, we advocate that agricultural engineers and the agricultural equipment industry work to develop and deploy such a technology, as we see it having multiple benefits to soil health, natural resource conservation, and bottom lines compared to other methods and equipment tested in this project. However, in the meantime, we advocate for the use of better in-crop weed control (high residue cultivation) and skillful weed and nutrient management outside of the crop production window covered by this project.

We are looking forward to a second year of this project, to be implemented during the 2021 growing season (funded separately).