Dairy farmers in New England are looking for ways to (1) reduce the need for imported grain and (2) minimize the impacts that their operations have on the environment. They have also identified weed control as one of their most challenging management issues. Intercropping annual feed grains with perennial pasture could reduce the need for grain imports and intensive tillage for weed management, and may lessen ecosystem disservices from dairy production and the need for purchased off-farm inputs. This experiment investigated different approaches to establishing annual feed grains into a perennial pasture community to identify optimal combinations of tillage and perennial pasture community suppression that effectively balance weed pressure and grain production. A six-treatment experiment was piloted in 2012 at the UNH Organic Dairy Research Farm in Lee, NH. In 2013, the experiment was expanded and established as a corn-soybean rotation at the UNH Kingman Research Farm in Madbury, NH. At the Kingman Farm, we established nine cropping system treatments, each replicated four times in a randomized complete block design. Our cropping system treatments spanned two disturbance intensity gradients allowing us to explore ecosystem service tradeoffs associated with low external input annual grain crop production. In 2013, corn grain yields averaged 4.8 and 5.6 Mg ha-1 in the best performing treatments (a conventionally managed strip-tillage system and an organically managed full-tillage system, respectively). Grain yield was affected more by the method of perennial pasture community suppression than by the method of seedbed preparation and was highest where the perennial community was completely eliminated either via herbicide or full tillage. Weed abundance appears to be an emerging tradeoff associated with intense soil disturbance. These results illustrate the important link between soil disturbance and weed abundance in agroecosystems. These results also suggest that pasture recovery following intercropping with feed grains can be improved by implementing minimum tillage approaches to feed grain establishment.
According to the 2009 USDA Economic Research Service Report Summary, small organic dairy farms (fewer than 70 cows) in the Northeast and upper Midwest make up over 80% of the organic dairy operations in the U.S., and rely more on pasture feed than the predominantly grain-fed organic dairy operations of the west. While a pasture-based system is less expensive than purchasing feed quality grains, the output (cwt milk per cow) is significantly diminished, and labor requirements are more intense when compared to con?ning cows and feeding harvested forages, making small, pasture-based dairy operations less profitable (McBride and Greene, 2009). This model incentivizes increasing the scale of organic dairy operations at the cost of the ecosystem services typically generated in closed-loop farming operations.
Due to the short growing season, dairy farmers in the Northeast must purchase between 25 and 50% of their winter feed (Maltby, 2012). The cost of feed grains has more than doubled since 2005 and is projected to continue the climb. This escalation in the cost of feed grain only decreases profitability for farmers, threatening the survival of small dairy farming operations (Maltby, 2012). By incorporating annual feed grains into pastureland, farmers may be able to produce a percentage of their winter dairy grain needs, thereby increasing productivity of their pastureland and decreasing the overall costs of inputs.
Organic feed grain production often requires frequent tillage to prepare seed beds and suppress weed growth (Smith et al., 2010). Conventional tillage can damage soil structure and can result in a loss of soil organic matter, reducing water holding capacity, and increasing the risk of soil erosion (Lal, 2004). Intercropping could reduce the need for inter-row cultivation for weed control and if grains are intercropped with legumes could reduce the need for supplemental nitrogen (Hauggaard-Nielsen et al. 2008).
Minimizing off-farm inputs such as fertilizer not only reduces environmental contaminates (such as nitrate and trace gas emissions) but also capital costs to farmers. In 2010 U.S. farmers spent 21 million dollars on fertilizers (including lime and soil conditioners) (USDA NASS 2012). One potential approach to increasing the efficiency of nitrogen cycling and reducing trace gas emissions in agricultural systems is to reduce tillage in conjunction with the maintenance of perennial ground cover (Robertson and Vitousek, 2009). Reduced tillage, however, particularly in organic cropping systems, can result in increased weed populations which can ultimately reduce crop yields (Smith et al., 2010). Therefore, identifying intercropping systems that effectively balance soil disturbance with the maintenance of perennial cover so as to effectively cycle nutrients, build soil organic matter, and reduce erosion potential, is a priority for organic research.
We sought to answer the following questions: What combination of practices maximizes provisioning ecosystem services (e.g., grain yield)? What are the tradeoffs associated with annual-perennial intercropping, and can these be minimized? How do these tradeoffs manifest over the longer-term? By incorporating annual feed grains into perennial pasture systems, we expect to expand production operations within organic dairy systems. In addition to enhancing overall system productivity, reducing costs and increasing net farm income, an integrated dairy forage operation may also improve soil quality relative to a conventionally-tilled system, which would contribute to improved soil aggregate structure and water holding capacity, off-site water quality, as well as building belowground habitat for soil communities (Kladivko, 2001). The first two years of this project served as the initial stages of a long-term agricultural research experiment. During year one and two, we sought to set the stage for integrating annual feed grains into perennial legume pasture, as well as establish a baseline of inquiry for a variety of future research questions.
- Objective 1: Integrate a rotation of annual feed grains (corn-soybean-sunflower) into a perennial legume pasture, and determine the optimal tillage system based upon crop yield, weed abundance, soil quality, and capital cost. We successfully integrated corn (2012 & 2013) and soybean (2014) into a perennial pasture of alfalfa and orchard grass. Due to a shift in locations (moving from the Organic Dairy in Lee, NH to the Kingman Research Farm in Madbury, NH), we have yet to integrate sunflower into the three-year rotation, as we started at year one again in 2013 to develop a baseline of the full three-year rotation at Kingman Research Farm. Sunflower will be planted in spring 2015.
- Objective 2: Assess soil quality in response to the treatments, by using standard soil quality indicators, to determine how intercropping intensity affects soil properties—nitrogen availability, soil carbon, moisture, and organic matter—and microbial activity along the gradient of management intensity. Soil samples were collected at the beginning of the season in both 2013 and 2014, and sent to the Cornell Soil Test lab, in Ithaca, NY. Additionally, soil moisture and Chlorophyll readings (using a SPAD meter) were taken regularly throughout the season. At this point, we only have one year to compare to the baseline, related to one crop in the three-year rotation (corn). While we cannot yet determine how the full range of treatments in the three-year rotation will affect soil quality, we have been able to make comparisons across the gradient of management intensity for year one. Soil samples collected in the spring of 2015 will allow us to analyze treatment affects from year two of the rotation (soybean).
- Objective 3: Calculate overall crop yield and weed abundance, and examine how community assembly (pasture recovery) could affect weed suppression and interspecific competition. Community assembly was determined for the pasture prior to planting each year, and weed community assembly and abundance were determined at peek biomass (mid to late August) each year. We were able to examine how weed abundance affected corn yield in both 2012 and 2013. As we have yet to harvest the soybean, we cannot determine the effect of weed abundance on overall crop yield.
- Objective 4: Perform a cost-benefit analysis for each cropping system treatment, and develop a grain production management plan for best practices within pasture-based dairy systems. A grain production management plan cannot be finished until the three-year rotation has been completed.
In 2012, a pilot study with six different management treatments, each replicated four times in a randomized complete block design, were established at the Organic Dairy Research Farm in Lee, NH (Fig. 1). Corn was planted in all treatments except in the Pasture treatment (T1), in mid-June. Aboveground weed biomass was measured at peak biomass (mid-August, 2012), and again in May, 2013. Weeds were sorted to species, dried to constant biomass and weighed.
Treatments examined in 2012 (UNH Organic Dairy Research Farm, Lee, NH)
T1: Pasture—3 yr old stablished mixture of alfalfa and orchardgrass (control)
T2: Minimal Till—Corn planted into living pasture following mowing and subsoil tillage with a Yeoman’s Plow
T3: Minimal Till w/ Undercut—Corn planted into living pasture following mowing and subsoil tillage with undercutting (Yeoman’s Plow with undercut knives)
T4: Strip Till—Corn planted into living pasture following mowing and strip-tillage (Unverferth Ripper Stripper)
T5: Conventional Till—Corn planted following full-tillage (moldboard plowed and disked, not intercropped)
T6: Conventional Till w/ Interseeding—Corn planted following full-tillage with crimson clover inter-seeded at final cultivation
The pilot experiment conducted at the UNH Organic Dairy Research Farm provided useful data on the weed and pasture community response to the treatments; however, due to several unanticipated personnel and technical challenges, we decided to move the study to a nearby UNH research farm for the 2013-2014 iteration of the study (Fig. 2). The field experiment is now located at the University of New Hampshire Kingman Research Farm in Madbury, NH. By moving the study to the Kingman Farm we were also able to include several conventional herbicide-based treatments in addition to the six original organic-based treatments, for a total of nine treatments.
The treatments were established into a two-year old alfalfa stand. The treatments included two controls, a pasture control (an established mixture of alfalfa and grass) (T9), and a full-tillage control with inter-row cultivation (not intercropped) (T1). These were compared to seven intercropping treatments differing in disturbance intensity and glyphosate usage: Full tillage, with inter-row cultivation and inter-seeding of crimson clover following harvest (T2); strip tillage, with glyphosate burn-down and inter-seeding of crimson clover following harvest (T3); strip tillage following a glyphosate burn-down (T4); strip tillage and legume seeded following harvest (T5); no till following a glyphosate burn-down (T6); no till following a glyphosate burn-down and inter-seeding of crimson clover at V6 (T7); and no tillage and inter-seeding of crimson clover following harvest (T8). Corn was planted in all treatments except T9. Manure was applied to the entire site prior to treatment establishment. No additional inputs of fertilizer or pesticides were applied after corn planting.
Aboveground biomass (alfalfa pasture and weed community composition) was collected in June 2013 to determine baseline conditions prior to the establishment of the treatments at the Kingman Research Farm. Soil samples were collected on June 18, 2013 and analyzed at the Cornell Soil Lab for an initial comprehensive soil health assessment. To evaluate differences in corn establishment and early season growth across the treatments, we measured corn populations and height and used a SPAD meter to measure leaf chlorophyll content. Moisture readings were taken regularly throughout the growing season and additional soil samples were collected in spring 2014 and submitted to the Cornell Soil Health Test prior to soybean planting. The weed and pasture community was measured in mid-August 2013 and 2014. Corn grain was harvested with a plot combine in fall 2013.
Treatments 2013-2014 (UNH Kingman Research Farm, Madbury, NH)
T1: Full-tillage + inter-row cultivation
T2: Full tillage + inter-row cultivation + legume inter-seeded at final cultivation
*T3: Strip-tillage + glyphosate burn-down + legume inter-seeded at final cultivation in T2
T4: Strip-tillage + glyphosate burn-down
*T5: Strip-tillage + legume inter-seeded at final cultivation in T2
T6: No-till + glyphosate burn-down
T7: No-till + glyphosate burn-down + legume inter-seeded at final cultivation in T2
*T8: No-till + legume inter-seeded at final cultivation in T2
T9: Alfalfa (control)
* Treatment modified in 2014 due to loss of alfalfa in all cropped treatments and expected high weed pressure.
2012 – 2013 Corn yield, weed abundance, and pasture recovery (UNH Organic Dairy Research Farm)
From the establishment of the pilot experiment at the Organic Dairy Research Farm we were able to assess late summer weed abundance and community composition in response to the tillage and intercropping treatments, and determine the extent of pasture recovery in each treatment in the following spring. We found that pasture recovery was highest in the minimally-tilled treatments compared to the conventionally-tilled treatments, which had a larger proportion of weed species, but lower total crop and weed biomass. Plant communities in the minimum-tillage treatments were more similar to the pasture control and were less temporally variable than were plant communities in the conventionally-tilled treatments.
Plant community composition differed across treatments, with Amaranthus spp., Chenopodium album, Digitaria sanguinalis, and Setaria viridis dominating the conventionally-tilled treatments and perennial pasture species dominating the plant communities in the minimally-tilled treatments (Fig. 3). Total non-crop plant biomass increased with tillage intensity, with higher weed abundance in conventionally tilled treatments compared to the minimally-tilled intercrop treatments and pasture control (Fig. 4). Corn grew poorly across all treatments and was not harvested for grain; however we did measure corn biomass. Corn biomass was highest in the conventionally tilled treatments and lowest in the minimally tilled treatments (Fig. 5). By spring 2013, pasture species comprised the largest proportion of the non-crop biomass in minimally-tilled treatments (T1-T4) compared to conventionally-tilled treatments (T5 and T6), which had a larger proportion of weed species, but lower total community biomass (Figs. 6 & 7). Plant communities in the minimum-tillage treatments were more similar to the pasture control and were less temporally variable from summer 2012 to spring 2013 than were plant communities in the conventionally-tilled treatments (Fig. 8).
2013 – 2014 Soil Quality, Pasture Response, and Grain Yield (UNH Kingman Research Farm)
Our cropping system treatments spanned two disturbance intensity gradients allowing us to explore ecosystem service tradeoffs associated with low external input annual grain crop production. Corn grain yields averaged 4.8 and 5.6 Mg ha-1 in the best performing treatments (T4 and T2, respectively). Grain yield was affected more by the method of perennial forage community suppression than by the method of seedbed preparation and was highest where the perennial community was completely eliminated either via herbicide or full tillage. Relative to the control (T9, alfalfa) most treatments resulted in slightly lower scores for non-provisioning ecosystem service indicators. Weed abundance appears to be an emerging tradeoff associated with intense soil disturbance (Fig. 9). We expect to observe additional tradeoffs to emerge as this long-term study continues.
Based upon the Cornell Soil Health Test (CSHT), several soil quality indicators were used as proxies for ecosystem services including soil available water content, soil organic matter, organic N, fraction of SOM that is organically bound N, soil respiration, and total soil quality (a composite index based on the full suite of biological, chemical, and physical soil indicators measured by the CSHT). Relative to the alfalfa control (T9) most treatments resulted in slightly lower scores for non-provisioning ecosystem service indicators (Fig. 10).
While we initially established this experiment at the Organic Dairy Research Farm, the shift to the Kingman Research Farm required us to restart the experiment in year two. This setback did not allow us to complete specific objectives within the original time-frame outlined in the proposal. To develop a comprehensive grain management plan for farmers, we need data from each of the three annual feed grain rotations. Restarting the experiment has also delayed video production. However, our new timeline is to produce both videos this winter and make them available summer 2015. Additionally, the new location of the experiment is not conducive to farmer the field day activities that we had originally envisioned (the site is not within walking distance of the main research area at Kingman Farm). Thus, we were not able to include the site in the farmer field day that was held at Kingman Farm in October this year.
The cost of purchasing feed grain and managing weeds are two of the biggest challenges facing small-scale dairy producers today. If a low-input, low-management grain production option can be made feasible, farmers could increase not only the economic viability but also the long-term sustainability of small-scale dairy production in New England. This experiment is one effort to determine the most effective management practices that could help meet these goals. Because it is a long-term cropping systems experiment, we expect additional impacts and outcomes to emerge over time.
Education & Outreach Activities and Participation Summary
- Wilhelm, J.A. 2015. Intercropping for on-farm grain production and environmental quality in perennial forage systems. UNH Graduate Student Research Conference, Durham, NH.
- Warren, N.D., Wilhelm, J.A., Smith, R.G. 2014. Intercropping for on-farm grain production and environmental quality in perennial forage systems. Ecological Society of America Annual Meeting, Sacramento, CA. August 10-15.
- Wilhelm, J.A., Smith, R.G. 2013. Disturbance and Recovery in a Pasture Plant Community along a Tillage and Intercropping Intensity Gradient. Ecological Society of America Annual Meeting, Minneapolis, MN. August 4-9.
- Wilhelm, J.A. 2013. Shifts in weed community composition and abundance along an intercropping and soil disturbance intensity gradient. UNH Graduate Student Research Conference, Durham, NH
- Wilhelm, J.A., Smith, R.G. 2013. Shifts in weed community composition and abundance along an intercropping and soil disturbance intensity gradient. Annual Meeting of the Weed Science Society of America. Baltimore, MD. Feb. 4-7.
We collected extensive video and photo documentation of the treatments being established at the Kingman Research Farm. With data from two full seasons, we have now begun to combine these data with the video footage and photos to develop an educational video for farmers. This video is intended to assist farmers in making informed decisions about pasture management strategies and reduced tillage implements, and will be targeted to dairy farmers in the Northeast. By sharing our research findings with other researchers and extension personnel, we hope to contribute valuable knowledge of how each management system functions. It was also our intention to conduct this experiment so that farmers do not have to take the financial risk experimenting with different tillage methods themselves; they can learn from our efforts. We are currently working with iVideo to integrate simple graphs and statistical analyses with video, audio, and photo images. The result will be a concise, easy to understand video that will be posted to YouTube, and will be available to Cooperative Extension staff as well as other farmer-oriented organizations.
Local farmers from Farm Hack (http://farmhack.net/home/) visited our experimental plots at Kingman Farm to learn about our work, and to test equipment they developed for taking aerial imagery of crops to aid decision-making related to agricultural inputs and adaptive management (Fig. 11).
The need for reducing the cost of feed grain and competition with weeds on dairy farms in New England has been clearly articulated by the dairy farming community. Farmers have also expressed a willingness to implement management practices that have larger overall ecosystem services benefits to their farming operation, provided yields are shown to be consistent (even if the average yield is lower than conventional practices). If annual crop feed grain production can be successfully integrated into perennial pasture on dairy farms, farmers could potentially reduce the amount of off-farm inputs required for their operations. The results would be lower costs to farmers and increased soil quality, ultimately making the overall operation more ecological.
Areas needing additional study
Our results show the importance of both soil disturbance and intercropping intensity on annual grain yield production. We were able to determine what combination of practices maximizes provisioning ecosystem services, as well as some of the tradeoffs associated with annual-perennial intercropping. However, as this experiment involves a multi-year crop rotation, the longer-term effects of the different treatments will likely emerge over time. Additionally, the full three-year rotation must be completed before conclusions can be drawn about the overall functionality of the management systems. Therefore, the immediate next steps are to harvest soybean, continue baseline measurements, and plant the third year of the rotation (sunflower) in spring 2015.