Advanced sustainable cropping systems for dairy farms in the Northeast

5. Hypotheses

The injected manure (IM) and the reduced herbicide (RH) cropping system treatments will produce similar crop yields to the broadcast manure (BM) and standard herbicide (SH) treatments in the manure and pest rotations. These enhanced best management cropping systems (IM and RH) will perform similarly in milk production and dairy farm profitability to the more typical best management cropping system practices (BM and SH).
2019 New Weed Management Rotation Comparison

In 2019, we began testing a new combination of Integrated Weed Management (IWM) practices to test the hypothesis that increased crop diversity and lengthened crop growing season will enable reduced herbicides relative to standard herbicide management. The IWM rotation compares the utilization of both cultural and mechanical weed control to a more standard herbicide control program utilized in the SH rotation. Specifically, the IWM split terminates perennial alfalfa and orchardgrass in spring to extend weed competition in spring prior to corn planting while the SH treatment terminates the alfalfa and orchardgrass in the fall prior to corn planting. In the IWM split, the rye planted after the corn silage entry is harvested as ryelage allowing a longer growth and for weed suppression as well as removing weed biomass from the ryelage harvest itself. Sorghum sudangrass follows the ryelage which again allows for weed suppression via two to three cuttings for forage. The soy after rye cover is planted green, again allowing for increased growth for weed suppression and increased habitat for beneficial predatory insects. Finally, oats are added to the first-year alfalfa and orchardgrass to help suppress weeds as the slower growing perennial crops establish. We hypothesized that these weed suppression methods would allow for four less herbicide applications compared with the SH rotation.

2019 Weed Rotation Planting Green Experiment Hypotheses

Planting green was integrated into the IWM of the Weed rotation in fall 2018 as an integrated weed management tactic. Planting green is a practice where cover crop termination is delayed until at or after cash crop planting. Treatments include a standard cover crop termination (early killed) and planting green practice.
Soybean yield will not differ when comparing planting green practices to standard cover crop termination.
Total late season weed biomass will not differ when comparing planting green to standard cover crop termination.

The higher cover crop biomass in planting green practices will reduce summer annual weed densities and heights at post-emergent timing when compared to standard cover crop termination.

Dairy Farm Comparison

In 2018, we added an additional hypothesis that a dairy farm that employs the diversified conservation practices we have been simulating will perform similarly in milk production and profitability to a typical “Business as Usual” Pennsylvania dairy farm. This modification enabled us to merge with the USDA ARS-LTAR (Agricultural Research Service Long Term Agriculture Research Network) to sustain the project beyond the NESARE USDA funding that ended in December 2019. We substituted a “Business as Usual” dairy cropping system for the PEST SH management plots to simulate a typical Northeastern dairy farm that grows only the forages for the same-sized dairy herd. The Weed Management rotation was added in 2019 and replaced the PEST RH plot management.

Long-Term Weeds Analysis in the Pest Rotation

We hypothesized that compared to relying on herbicides for weed control in no-till, reducing herbicide inputs and relying on cultural and mechanical weed control would: increase weed dry matter (biomass) result in greater weed biomass over time, but not affect cash crop yield iv. or returns to investment.

Herbicide Resistant Weed Management: Integrated Weed Management Experiment and Marestail Experiment

In fall of 2017, Agronomy MS student Haleigh Summers initiated two satellite experiments to evaluate integrated weed management practices that a dairy farm could adopt to control glyphosate herbicide resistant weed species. Seeds of marestail and pigweed to simulate Palmer amaranth or waterhemp were supplemented in microplots in the experiment in fall 2017. When marestail seeds did not germinate in spring, she supplemented them again in spring 2018. Although populations were lower than expected, pigweed seeds did germinate. Supplemented marestail never germinated; therefore, in spring 2018 we established a separate experiment in a field with a resident marestail population to evaluate spring glyphosate-resistant marestail control strategies.
Herbicide Resistant Weed Control with Integrated Weed Management Experiment

Hypotheses:

• Rotating to a perennial forage crop with regular cuttings will reduce the amount and size of pigweed plants by means of mechanical control by the end of the first year.
• A spring-planted oats nurse crop in alfalfa will perform similarly to a post-emergence herbicide application.
• A fall-planted cover crop will reduce the size and density of weeds in the cash crop the following year, whether it is fall-planted oats with alfalfa or rye cover crop with soybean.
• Termination of a fall-planted rye cover crop at soybean planting (“planting green”) will allow the rye to grow larger than if termination occurs two weeks before planting, and therefore be more effective at shading out small weed seedlings.

Marestail Experiment

Hypotheses:

• As expected, based on herbicide product labels, plants shorter than 10-cm are more susceptible to herbicide than taller plants.
• Herbicides that include high rates of active ingredients that specifically control broadleaves will more effectively control taller plants than glyphosate and other herbicides.
• Diverse management practices such as tillage and frequent mowing will be effective strategies to reduce weed biomass in a field
• Diverse control strategies, whether herbicide or mechanical, will be most effective at reducing total weed biomass when applied early in the season.

Integrated Pest Management and Pesticide Management

Similar to previous years, we tested the central hypothesis that compared to crop rotations with little crop diversity and pre-emptive pesticides, crop rotations with greater crop diversity and integrated pest management (IPM) that conserves natural enemies and habitat for beneficial organisms will interrupt pest populations, increase predator populations, and will maintain pest populations below the economic threshold.

Incorporating Cover Crop and Soil Organic Matter Nitrogen Availability into a Fertilizer Recommendation Tool

The increasing use of the winter cover crop (CC) as a nitrogen (N) management tool has necessitated a practical understanding of its effects on succeeding cash crops. In not realizing the need to increase fertilizer application to maintain yield in the face of N immobilization, or the opportunity to decrease fertilizer application to credit CC-N supply, growers may fail to realize the full economic and environmental benefits of CCs. Traditionally, N fertilizer recommendations have been based on yield targets, where a fixed N recommendation is calculated from the desired yield. Some regional N recommendation systems, such as the Penn State Agronomy Guide (Agro. Guide), contain credits for historical manure applications and previous legume crops. However, the difficulty of estimating N supply from CCs and soil organic matter (SOM) has prevented their incorporation into current N fertilizer recommendation systems. The pre-sidedress nitrate test (PSNT) uses early season nitrate levels to tailor N fertilizer recommendations to N mineralization rates. However, considering the change in N cycling dynamics resulting from CC and SOM interactions, and the fact that CCs were not included in the development of the PSNT, the PSNT might not effectively account for N mineralization dynamics following CCs. Charles White led previous work to develop a system of linear equations, and a resultant graphical tool, that can predict CC- and SOM-N supply to a cash crop.

The objective of this research is to advance and test the N supply tool’s capability to predict an N fertilizer requirement for corn based on an accounting of N supplied by CCs and SOM. We hypothesize that the N supply tool’s fertilizer recommendation will be sufficient to sustain N-supply to corn, such that yields will be equal to or better than corn fertilized using the locally-adapted yield-based N recommendation or the PSNT-derived N fertilizer recommendation. To provide a robust test of the N supply tool, Master’s student Anthony Colin finished his second year of research within the rye silage corn grain cropping sequence of the DCS project’s manure rotation.

Soil Health Indicator Hypotheses

We hypothesized that compared to no-till, annual conservation tillage (chisel-disk) in a corn-soybean crop rotation without cover crops or perennials would reduce indicators of soil health such as labile carbon, percent organic matter, total carbon sequestration and water stable aggregates in the top 0-5 cm of soil immediately after tillage.

Occasional full tillage within a no-till crop rotation that integrates perennials and cover crops and primarily employs no-till will not permanently reduce soil health indicators. Specifically, we hypothesized that in the PEST rotation implementing full tillage (primary and secondary tillage with a moldboard plow and chisel disk) once in the 6 year rotation will initially reduce labile carbon, percent organic matter, total carbon sequestration and water stable aggregates in the top 0-5 cm of soil immediately after tillage, but that these indicators will recover to similar levels in later years in the rotation that in other years employs no-till and integrates perennial alfalfa and orchardgrass forage, double and cover crops .

The effects of termination timing on microbial nitrogen cycling dynamics

Research has primarily focused on identifying how changes to the termination time of the forage, alfalfa/orchard grass (AO), impacts microbial nitrogen cycling dynamics. In previous years of the experiment the AO was chemically terminated in the spring of the 3rd year. However, with wetter springs, it is becoming increasingly difficult to terminate the AO in the spring and still plant corn in May, as some orchardgrass recovers. Further, in one year armyworm infestation in the orchardgrass reached the economic threshold. Therefore, the project has been evaluating terminating the AO in the late fall of the 2nd year which some farmers do to be able to plant earlier in the following spring. Currently, the SDCS plots include a fall terminated AO and a spring terminated AO in the manure rotation. Our main objectives and hypotheses are as following:

O1: Determine if there are differences in nitrous oxide (N2O) emissions in the fall terminated AO (2nd year) compared to the spring terminated AO (3rd year).

H1: In the fall, after fall termination of AO, N2O emissions will be highest in the fall-terminated, but overall N2O emissions will be highest in the spring terminated AO.

O2: Evaluate whether termination time of AO impacts microbial nitrogen cycling gene abundances

H2: Spring termination will have the greatest impact on microbial nitrogen cycling gene abundances compared to fall termination. Microbes will be more active in the spring due to the higher temperatures expected and changes in the community is expected as carbon and nitrogen become available from terminated AO.

6. Materials and Methods

We initiated two diverse, 6-year dairy crop rotations using farm-scale equipment at 1/20th scale of 240 acres on 12 acres of Penn State’s Agronomy Research Farm near University Park, PA. Cropping system strategies include no till, manure injection, cover crops, a cover crop roller, perennial legumes, green manure crops, and winter canola. Within each crop rotation we have evaluated innovative management strategies to address no-till management challenges: i. no-till manure management to conserve manure nutrients and reduce losses to the environment with shallow-disk manure injection; and ii. Reducing herbicide and insecticide use with cultural and mechanical integrated weed management practices, and biological and integrated pest management practices for insect and slug management. To examine the effect of diverse rotations managed with IPM practices on insect pests and beneficial insect populations, a corn-soybean grain rotation was also included for comparison purposes at the experimental site on an additional 2 acres. In the corn-soybean rotation, we compare the manure management strategies (injected or broadcast manure), inorganic fertilizer, and in 2016 we included a minimum chisel-disk tillage (Chisel) treatment to evaluate the effects of tillage on soil health parameters. In 2010, our USDA ARS team also established twelve lysimeter plots (50 x 90 ft.) on a slope to compare total nutrient losses and crop uptake between manure and crop management strategies on an additional 1.3 acres.

Until 2018, the NESARE experiment was a nested split-split plot, full crop entry design. All crop phases of each rotation are planted each year on 0.25 acre plots that are replicated four times (12 crop entries for the 2-year rotations, 2 crop entries for the corn-soybean rotation X 4 replications = 56 plots). Each crop entry is the main plot (120 ft. x 90 ft.) that is divided in half (60 ft by 90 ft) to compare the management strategies over time (ex. injected manure or broadcast manure). In 2014 and 2015 we compared green manure crops (red clover vs. crimson clover); the manure rotation has nested split-split plots (30 x 90 ft.) within both manure management strategies. According to the results of our dairy nutrition model, which develops rations for the virtual dairy herd, the farm was already producing sufficient high-quality forage in the manure rotation and there was a need for more low-quality forage production for dry cows and heifers. Therefore, we replaced red clover with rye silage while still having the crimson clover green manure crop at the split-split plots. Crimson clover and rye silage are planted after two or three cuts of sorghum sudangrass, respectively. Typically, we were not able to harvest sorghum sudangrass a third time in early fall, because crimson clover needed to be planted by the end of August to ensure successful establishment. Therefore, in 2018, the crimson clover green manure was removed from the rotation and replaced by rye silage which is harvested for forage the following spring prior to corn planting.

In the PEST rotation there were nested split-split plots (30 x 90 ft.) within the reduced herbicide treatment in the soybean and corn crops that compared two post-planting weed control strategies (high-residue cultivator or post-emergent herbicide). Since the cultivator required wider row spacing (30 inch, 76 cm) than is typical for no-till soybeans, we had also included split-split plots in the soybean standard herbicide treatment to compare the soybean row spacing of 30 inch (76 cm) to a more typical 15 inch (38 cm) row spacing.

In 2018, to compare the performance of the diversified conservation dairy cropping system to a typical Northeastern dairy farm rotation and to secure new and continued future funding from the USDA Long-Term Agricultural Research (LTAR) network, we added a business as usual (BAU) rotation. The BAU farm represents the common Northeast dairy farm that has increased its herd size but has not proportionally increased land area. Therefore, the farm grows all of the herd’s forages and imports feed crops such as corn grain and soybean. To produce all of the forages for the same size 65 milking cow herd as our other experimental dairy farm rotations, the BAU dairy farms half as much land (49 ha.), and therefore applies manure more frequently to corn fields in spring and fall. To do this, we modified the SH pest control rotation to include the 6 yr-dairy rotation (3 yrs alfalfa+grass/3 yr corn silage with cereal rye cover). After consultation with the Advisory panel, we decided it is most representative to apply the dairy manure in the BAU rotation to the three corn entries in both spring and the following fall as well as the fall before planting alfalfa and orchardgrass (Fig. 1). The transition to the BAU rotation in spring 2018, included extending the growing period of the third year of the Alfalfa + Orchardgrass crop, replacing both canola and soybeans with additional corn silage, and relying solely on surface manure application (see 2020 Cropping Schematic, Fig. 1). Similar to the Control rotation, the BAU corn had Bt traits and the seed was coated with neonicotinoid insecticides. Corn and alfalfa were also sprayed with preventative post-plant insecticide applications after planting. The susceptible-BAU first year alfalfa was also sprayed with an insecticidal spray about 10-14 days after each cutting. In 2018, all soybeans in the RH treatment were planted at a row spacing of 30 inches and the SH treatment was replaced with BAU corn. In 2018, we compared the Pest/BAU rotation alfalfa and orchardgrass forage and corn silage crop entries that were paired in side-by side 60 ft x 90 ft main management-plots.

New Weed Management Rotation Comparison

In 2019, we began testing a new combination of new Integrated Weed Management (IWM) practices to test the hypothesis that increased crop diversity and lengthened crop growing season will allow for reduced herbicides relative to the standard or BAU Weed rotation. To this end we converted the Reduced Herbicide (RH) PEST 2018 rotation into the new Weed management comparison by splitting the original 60’ wide plots into two 30’ wide plots: the Integrated Weed Management (IWM) rotation and the Standard Weed (SH) rotation (see 2020 schematic, Figure 1). The IWM rotation employs both cultural and mechanical weed control tactics to more standard herbicide control program in the SH rotation. Specifically, the IWM split terminates perennial alfalfa and orchardgrass in spring with herbicide to extend weed competition until spring while the SH treatment terminates the alfalfa and orchardgrass in the fall prior to corn planting. In the IWM split, the rye planted after the corn silage entry is harvested as ryelage allowing longer spring growth for weed suppression as well as removing weed biomass during the ryelage harvest. Sorghum sudangrass follows the ryelage which via two to three forage harvests provides multiple mechanical weed control events. Soybean after rye cover is planted green, allowing for increased growth for weed suppression and increased habitat for beneficial predatory weed seed insects. Finally, oats are added to the first-year alfalfa and orchardgrass to help suppress weeds as the slower growing perennial crops establish. These weed suppression methods should allow for four less herbicide applications compared with the SH rotation. We sampled weed biomass prior to harvesting two crop entries in the: i. rye prior to corn silage followed by rye or rye silage to sorghum sudangrass and ii. prior to the soybean harvest.

Figure 1. 2019-2020 NE SARE Dairy Cropping Systems Schematic

2019 New Weed Rotation Planting Green Experiment

The treatments were compared in a split-plot design with four replicates. The cover crop in the standard cover crop termination treatment (early killed) was terminated 3 weeks prior to soybean planting. The cover crop in the planting green treatment was terminated on the same day as soybean planting. Cereal rye was drilled at 135 kg ha-1 in both treatments on October 17, 2018. Artificial weed seedbanks of approximately 400 common lambsquarters (Chenopodium album) seeds, 800 redroot pigweed (Amaranthus retroflexus) seeds, and 200 large crabgrass (Digitaria sanguinalis) seeds were established in a 1 m² microplot within each plot. Early killed plots were sprayed with 32 fl oz/ac of glyphosate (Roundup PowerMax) and 16 fl oz/ac of 2-4,D ester on April 30, 2019. Rye biomass samples were collected just prior to termination and soybeans were planted using a no-till planter equipped with a ZRX integrated roller system on May 22, 2019. One day after planting, both treatments were sprayed with 22 fl oz/ac of Roundup PowerMax and early killed plots were sprayed with 3.75 fl oz/ac of Fierce on May 22, 2019. The artificial weed seedbanks were excluded from the Fierce pre-emergence herbicide application. Summer annual weed density and heights from the artificial weed seedbanks were recorded on July 2, 2019. Late season weed biomass samples were collected randomly within each plot on August 15, 2019. All biomass was dried and weighed. Soybeans were harvested on October 11, 2019. Data were analyzed in R (v. 3.6.1) with treatment as a fixed effect and block as a random effect. Analysis of variance was used to test rye biomass, weed densities, weed biomass, and soybean yield responses to treatments. Treatment LSMEANS were considered significantly different at P ≤ 0.05.

We are evaluating multiple performance indicators: crop yield and quality, soil health, nutrient conservation, greenhouse gas emissions, weed and insect populations; and farm profitability. Crop yields were collected for each crop entry point in the BAU, MANURE, PEST, and Weed rotation in each year. We continue to collect three subsamples for forage or feed for quality analysis from each of our main management treatments in the MANURE rotation, as well as the BAU and Weed management rotation at every crop harvest in the NESARE Dairy Cropping Systems Trial. We simulate making corn, alfalfa-grass, sorghum sudangrass, and rye silage and soybean roasting in the laboratory prior to submitting them to Dairy One for forage quality analysis. Data were analyzed with a mixed ANOVA model using SAS JMP Pro 14 with main management and crop phase within a rotation as fixed effects and blocks as random. Treatment LSMEANS were considered significantly different at P ≤ 0.05.

Dairy Farm Comparison

The dairy scientist, Virginia Ishler uses the quality and yield data to develop rations for the virtual dairy herd, and a dairy nutrition model to simulate a dairy herd’s milk production under the three different cropping system scenarios: i. the BAU farm with imported feed crops, ii. the enhanced conservation cropping systems (the inject manure and reduced herbicide managements until 2019) compared to the, iii. more typical conservation practices (broadcast manure and standard herbicide managements until 2018 or the broadcast manure and reduced herbicide management in 2018). Starting in 2019, we began comparing the BAU farm performance to the conservation manure and new weed management rotations: i. the enhanced conservation cropping systems (IM and IWM management) compared to the, ii. more typical conservation practices (BM and SH).

The cropping year for the Virtual Dairy Operation began in 2010 with the assumption that a crop farmer was transitioning into dairy. The operation had the storage structures for ensiled feeds and basic facilities to house animals. The hay-crop season started in May 2010 and youngstock were started on feed in August 2010. The lactating herd was brought in on October 2010 after corn silage was harvested.

Long-Term Weeds Analysis in the Pest Rotation

This experiment began in 2010 and was completed in 2019. Treatments included a standard herbicide treatment and a reduced herbicide treatment compared in a split-plot within a diversified six-year dairy crop rotation (Fig. 2). Herbicides were reduced by: banding herbicide over the corn and soybean crop rows, using inter-row cultivation, reducing the number of herbicide applications, planting an annual companion crop with the perennial forage rather than using herbicide, and plowing once in six years to terminate the perennial forage.
There were three 3-year phases of weed management with changes to improve the cropping system performance or integrate new research questions (Fig. 2). The experiment is a full-entry crop experiment with each phase of the six-year rotation planted each year in a randomized complete block design with four replications. Beginning in 2013, plots were split again to create split-splits in row crops for both treatments. The SH treatment added a split of 38-cm row soybeans as opposed to 76-cm rows in soybeans. The RH treatment added a split where POST herbicide was broadcast rather than cultivation in both corn and soybean crops. Weed biomass data were collected at 10-15 weeks after planting in soybean and corn crops, at each forage harvest in the establishment year, and the first harvest of second year forage. Over the nine years, herbicides were reduced: i. in soybean by 4% in the RH-POST and 32% in the RH-Cult, ii. in corn by 30% in RH-POST and 44% in RH-Cult, and iii. in forage establishment by 37% in the RH treatment compared to the SH treatment. we also rotated modes of action to reduce the selection pressure for herbicide resistant weeds.

Figure 2: Six-year crop rotation with Reduced Herbicide (RH) and Standard Herbicide (SH) treatments

All data were analyzed with PROC MIXED of SAS® 9.4. Weed biomass was transformed by taking the log of one plus the biomass to normalize the distribution. Data were analyzed with repeated measures of each plot (9-m x 27-m) by year with an autoregressive structure and the Kenward-Rogers approximation. Treatment, year, crop, and the interactions of these terms were fixed effects, while block and the interaction of block and treatment were random effects. The SLICE test was used to test our hypotheses and when there were significant interactions to conduct a partitioned F-test analysis of the LSMEANS for an analysis of the simple or main effects. Treatments were considered significantly different at p<0.05. When the split-treatments of RH-POST and SH-narrow were included, we only compared split-treatments to RH-Cult and SH-wide respectively and did not include Phase 1 in this analysis.

We compared the weed biomass among the three phases by replacing year with phase in the same statistical PROC MIXED model of the log-transformed weed biomass to test the hypothesis that weed biomass would increase more in the RH treatment over time. We used the SLICE test when there were significant interactions to conduct a partitioned F-test analysis of the LSMEANS for an analysis of the simple or main effects, to test if weed biomass was greater in the RH treatment in later phases of the experiment.

Due to the different scales of crop yield, yield data for each crop were analyzed separately with year, treatment, and the interactions of these terms were fixed effects, while block and the interaction of block and treatment were random effects and the Satterthwaite approximation was used. Total annual forage harvest was summed and analyzed for perennial forage crop yields and annual weed biomass. A statistical test revealed 2010 to be an outlier year for the forage establishment year, most likely due to the high annual seeding rate, so we removed 2010 from the statistical analysis of the forage crops.

Herbicide Resistant Weed Management: Integrated Weed Management Experiment and Marestail Experiment

Herbicide Resistant Weed Control with Integrated Weed Management Experiment

This experiment is a nested split-plot design with four blocks in 2018 and five times in 2019. In the soybean treatment, there are three nested treatments of: i. no cover crop, ii. a rye cover crop terminated 10-14 days before soybean planting, and iii. a rye cover crop terminated at soybean planting. All soybean treatments received a burndown of 1.3 kg/ha of glyphosate. Emerged pigweed was covered at burndown application timing to simulate glyphosate-resistance. In the alfalfa treatment there are four randomized nested treatments of with or without an oat cover crop and with or without an oat nurse or companion crop. All alfalfa treatments were planted with a no-till drill on May 1, 2018 and April 24, 2019 and all soybean treatments were planted on May 24, 2018 and May 16, 2019.

One square meter in each plot was supplemented with pigweed seeds at a rate of 1,500 seeds/m2 to simulate a Palmer amaranth or waterhemp population. At each alfalfa harvest, this weed-supplemented microplot was harvested and then sorted into alfalfa, oats, pigweed, and other weeds. The harvested biomass was then dried and weighed. All soybean plots were treated with 0.6 kg ha-1 glufosinate (for post-emergent weed control on July 3, 2018 and June 28, 2019. In soybeans, density counts were taken in each supplemented microplot before and after herbicide applications. Ten weeks after planting, the pigweed biomass was harvested and dried from each supplemented microplot. Data were were log-transformed and analyzed with SAS 9.4 PROC MIXED with crop and treatment as fixed effects and block as a random effect and we used the Satterthwaite approximation. Two-sided F-tests for homogeneity of variance of pigweed biomass in soybean and total annual pigweed biomass in alfalfa, other weed biomass, cover crop biomass, alfalfa yield, and soybean yield and weed density counts, proved years could be combined due to similar variances (Mclntosh 1983). Year was included as a random variable with block nested in year. Weed biomass measured in multiple alfalfa harvests and density counts in soybean were analyzed with repeated measures and the Kenward-Roger approximation. Least square means were determined, and Tukey’s least significant difference was used to compare treatments at p<0.05. The slice test was used for pre-planned contrasts to test hypotheses: i. planting a fall cover crop prior to spring crop establishment compared to no cover crop and ii. spring planting an oat nurse crop with alfalfa compared to not including an oats nurse crop with or without a cover crop.

Marestail Experiment

This experiment is a randomized complete block design with five blocks. Treatments are listed in Table 1. Burndown herbicides were applied on May 17, 2018 and May 8, 2019. Soybean plots were no-till planted in 76-cm rows at 494,000 seeds ha-1 with a no-till planter on May 24, 2018 and May 23, 2019, except the 2018 Chisel disc plots which were planted on May 25, 2018. Burndown application before soybean occurred early to accommodate the label requirements for applying 2,4-D prior to planting Xtend soybean. Oats were planted on May 29, 2018 and April 23, 2019. After the oat harvest and a burndown application, sorghum sudangrass was planted on July 11, 2018 and July 9, 2019 and POST applications on soybean were made on July 2, 2018 and June 25, 2019.

Oats were harvested on July 10, 2018 and July 9, 2019 and sorghum sudangrass was harvested on August 24 and October 3, 2018 and August 10, 2019. We only harvested sorghum sudangrass once in 2019 due to a cold spring and less growth. Soybean plots were harvested on November 12, 2018 and November 20, 2019 and adjusted to 13% moisture. Due to regulations, Enlist soybean plots were destroyed before flowering in 2018; so, we were unable to harvest weed biomass or soybean yield in those plots.

Marestail plants were classified into one of four size categories i. small (not bolting, less than 5-cm wide), ii. medium (not bolting, greater than 5-cm wide), iii. large (bolting, shorter than 10-cm), and iv. extra-large (bolting, taller than 10-cm) and marked. These marked plants were reassessed before and after each management strategy such as an herbicide application, mowing, or tillage. At ten weeks after planting soybeans, all weeds were harvested for their biomass and sorted into marked marestail, unmarked marestail, and other weeds. This biomass was then dried and weighed. Missing plants were not included in this calculation and data were analyzed with SAS 9.4 PROC MIXED with weed control program and size as fixed effects and block as a random effect. Biomass data was transformed by adding 1 and taking the log prior to analysis. Tukey’s test was used to compare treatment differences among LSMEANS and were considered significantly different at p < 0.05. When interactions were significant, the slice test was used and treatments were considered significantly different when p<=0.05.

In 2019, when soybean competition with weeds was reduced by deer damage, soybean biomass was included as a covariate to explain weed biomass. Because there was not a soybean biomass covariate in 2018, years could not be combined and were analyzed separately. Deer damage was not uniform across control programs in 2019; so, we only compared the strategies when analyzing total weed biomass. Reported values are back transformed. Control efficacy of marked horseweed plants was analyzed using the lme4 program in R (R Core Team 2018) with an empirical logit link function and a binomial response distribution (Collett 2000; Bolker et al. 2009; Warton and Hui 2011; Stroup 2015). Size was treated as a split-plot in program or strategy and both were modeled as fixed effects, and block and the interaction of program or strategy x block as random effects. Estimates of control efficacy of programs or strategies were compared with Tukey HSD and considered different at p< 0.05. Reported values were back-transformed to report predicted probabilities

Table 1: Weed control program treatment differences of soybean variety, burndown, pre-emergence and post-emergence herbicides

Integrated Pest Management and Pesticide Management

During the 2019 growing season, we continued to monitor pest and predator populations.
Both the BAU and Control rotations were planted with transgenic Bt seeds coated with neonicotinoid insecticides. In these two rotations corn and alfalfa were also sprayed with preventative post-plant insecticide applications. In the BAU rotation, we sampled three corn entries, two following rye and one following alfalfa. Within the Corn-soybean Control rotation, we also compared the pest populations in corn among the three no-till nutrient treatments (IM, BM, and inorganic fertilizer: Fert) to the BM treatment that was chisel-disked in both the corn and soybean production years. The Weed and Manure rotations were planted with non-Bt/untreated seeds and were not sprayed with a preventative insecticide post-plant. These rotations relied on IPM for insect pest control.

Early Damage in Corn

To assess early damage in corn (slug, cutworm, and army worm), two 10-ft-row sections were randomly selected in each split-split plot of corn. Within that 10-ft section, each individual plant was assessed for damage and given a rating between 0 and 4 depending on proportion of leaf area removed by early season pests (0: no damage, 1: <25%, 2: 25-50%, 3: 50-75%, 4: >75%). This assessment was done at the V2 growth stage.

European Corn Borer (ECB) Damage

European corn borer (ECB) damage assessments were completed close to corn harvest to quantify damage from both first- and second-generation European corn borers. We assessed damage in rows four and eight of each corn split-split. In each of these rows, we sampled two plants every 20 feet and counted the number of entry holes. Through this process, we sampled 6 plants per row, or 12 plants per split-split. In addition to counting entry holes, we also tallied the number of lodged plants while walking through the rows.
To test the hypothesis that crop diversity, IPM, pest scouting and pest control practices that conserve natural enemies could control pests similarly to pre-emptive pest control in crop rotations with less diversity, we compared multiple measures of slug, early season caterpillars, predation of pests, slug predators, corn crop damage and plant establishment in the first three years (Phase 1) and the second three years (Phase II) of the experiment. We conducted analysis of variance with GLM in R, and used the LSMEANS package to perform planned contrasts of the diverse IPM (HiDiv-IPM) cropping systems compared to the low diversity pre-emptive (LoDiv-PP) pesticide control corn-soy cropping system, and considered treatments statistically significant when P<0.05.

For the last four years of the experiment (2016-2019), we compared corn pest damage ratings (due to slugs and early season insect pests) and the proportion of plants that established at the V2 stage with ANOVA for a mixed model using SAS JMP Pro 14 with treatments as fixed effects and year and blocks as random. Analyses were conducted for treatments and corn entries within a rotation as well as across the HiDiv-IPM and LoDiv-PP systems. At the V5 stage we only have complete data from 2017 and 2018, and we conducted the ANOVA with a mixed model with SAS JMP Pro 14 analysis, but year was a fixed effect. Treatment LSMEANS were considered significantly different at P ≤ 0.05 and we used Tukey’s separation of means test for LSMEANS that were considered significantly different at P ≤ 0.05. We also compared European Corn Borer damage between HiDiv-IPM and LoDiv-PP the cropping systems with ANOVA for a mixed model with SAS JMP Pro 14 analysis. Treatment LSMEANS were considered significantly different at P ≤ 0.05 and we used Tukey’s separation of means test for LSMEANS that were considered significantly different at P ≤ 0.05

Incorporating Cover Crop and Soil Organic Matter Nitrogen Availability into a Fertilizer Recommendation Tool

The CC C:N ratio is a determinant factor in the mineralization or immobilization of CC- and SOM-N. The tool’s equations first predict C:N ratio using SOM levels, nitrate-N available to the CC at planting, growing degree days between planting and termination as a measure of CC maturity, and the proportional seeding rate of non-legume CC species in the CC. The fall and spring CC C:N ratio, N content of the CC, and SOM are then used to predict SOM- and CC-N supply, which is expressed as relative yield (RY). This relative yield represents the predicted yield of a corn crop receiving only N supplied by CCs and SOM divided by the yield potential (YP) of a fully N-sufficient corn crop. The next step is to translate the potential management advantages of this pre-season prediction of N supply into an actionable N fertilizer recommendation (NFR) using the physiological N requirement (Nr) of corn. We propose Eq. 1. as a means of doing this.

Eq. 1 FR = Nr [YP – ( RY*YP)]

Treatments in the previous year’s rye silage to corn grain sequence were constrained to a comparison of the PSNT and N supply tool (CC Tool). In 2018, the fertilizer recommendations were calculated using a 10.04 Mg ha-1 yield goal. In 2019, we reduced the yield goal to 8.6 Mg ha-1 based on a linear regression of previous yields by planting date. Additional plot space allowed us to include a treatment from the PSU Agronomy Guide (Agro Guide) and an additional PSNT treatment calculated using the same 10.04 Mg ha-1 yield goal used previously. The lower-yield-goal PSNT (PSNT 137), higher-yield-goal PSNT (PSNT 160), N supply tool (CC Tool), and Penn State Agronomy Guide (Agro Guide) recommendations were each applied to 15 X 90 ft (6-row) plots.
To calculate the N fertilizer recommendations from the CC Tool and PSNT recommendation systems, samples were taken between rye planting in October 2018 and the V6 growth stage of the corn grain in June 2019. We took 15 soil samples (0-20 cm depth) per plot in November of 2018. SOM (%) was determined by the Penn State Agricultural Analytical Services Lab via loss on ignition for input into the CC Tool. The rye silage was topdressed with 89.9 kg ha-1 N in April and harvested mid-May of 2019. To adjust for the effect of fertilization on CC C:N ratio, 20 mg kg-1 nitrate-N was used in the CC Tool calculations. We based this adjustment on previous experience for typical soil nitrate values under fertilized conditions. GDD accumulation of 1763 (deg F; base 32) was calculated using the Network for Environment and Weather Applications (NEWA) Degree Day Calculator. Manure was broadcast on May 24th, 2019, 12 days before corn grain was planted. Using availability factors from the Agro. Guide, we subtracted 37.6 kg ha-1 N from the CC Tool recommendation. To account for the harvest of the rye silage, we used a CC biomass N content of 0 Kg ha-1 N. All treatments received 3.8 kg ha-1 N at planting, but this was only subtracted from the Agro Guide and CC Tool rates. PSNT soil samples (0-30 cm depth) were collected in June, subsampled and extracted using 2M KCl, and analyzed for nitrate-N colorimetrically on a microplate spectrophotometer using the Griess reaction with vanadium(III) as a reducing agent. Lastly, to calculate the Agro. Guide recommendation, we multiplied our yield goal, 8.6 Mg ha-1, by 17.84 kg N Mg-1 grain. In addition to subtracting the same spring manure-N availability credit, we calculated and subtracted 14.6 kg ha-1 N of plant-available N from the Agro. Guide recommendation to account for the historical frequency of manure application within this rotation. Treatment rates were side dressed as UAN 30% N solution in July.

To assess sufficiency of recommendation methods, yield was estimated by harvesting two representative rows from each plot in late November of 2019. Yields were analyzed both by year and across years with SAS 9.4 PROC MIXED. Yield was a fixed effect using stand counts taken in October as a covariate, and block was a random effect. Differences determined by Fisher’s LSD were considered significant when p<0.05.

Soil Health Indicator Research

Routine soil data has been collected every three years during the dairy cropping systems project, beginning in 2010 (collection years were 2010, 2013, 2016, and 2019). In the Control or Corn-Soy rotation in 2016 we initiated yearly chisel-disking treatment in one of the 30 ft x 90 ft split-split plots in the corn-soy rotation that had been receiving broadcast manure prior to corn planting. In 2016 we began chisel-disking these split-split plots prior to planting soybean and in the corn phase after broadcasting manure and prior to planting corn. One of the other 30 ft x 90 ft split-split-plots that had been receiving in-organic fertilizer prior to corn since 2012 was converted to a broadcast manure prior to corn and no-till treatment and the other two 30 ft x 90 ft split-split-plots corn year nutrient treatments were continued: no-till with synthetic fertilizer application and no-till with manure injection.

In the BAU/Weed management rotations, no-till (BAU) plots that were SH until 2018 were converted to the BAU and compared to the Weed rotation plots that from 2010 to 2019 in August had received full tillage (moldboard plow followed by chisel-disk) to terminate alfalfa and orchardgrass once every six years.

In early spring 2019 prior to field operations, five soil samples were collected at 0-5 cm and 5-15 cm per split-split plot in the Corn-soybean, Control rotation and 10 samples in the BAU and WEED split plot (main management level) were collected and combined for soil labile carbon and percent organic matter analysis. For water stable aggregate (WSA) stability, ten soil samples were collected to 15 cm that were stored in sealed plastic boxes until they could wet sieved to obtain 1-2 mm macroaggregates which were then air-dried at room temperature for 24 hours. We measured % WSA with a modified standard wet-sieving technique. We used slaking instead of misting for rewetting the air-dried aggregates, because slaking is more rigorous and would detect the differences among the crop treatments better than misting and we increased the duration of wet-sieving to five minutes. Four grams of the air-dried 1-2 mm aggregates were transferred to the 0.26 mm size sieve of the standard sieving machine. Samples were then submerged in distilled water for 5 minutes before the sieving started. The sieves were raised and lowered for 5 minutes @ 36 cycles minute-1 with a stroke length of 1.0 cm. Material that passed through the sieve after 5 minutes was collected in the metal can, oven dried, and weighed to estimate unstable aggregates. The soil that remained on the sieve was then subjected to in 0.2 M sodium-hexametaphosphate solution to disperse sand from the stable aggregate soil fraction. Disrupted aggregates were gently pushed through the smallest screen with water and a rubber policeman. Sand or coarse material was collected on the screen, washed into pre-weighed weigh boats, dried, weighed, and subtracted from the amount of aggregates collected after wet sieving to calculate the percentage of sand-corrected aggregate stability.

Labile carbon or active carbon was estimated according to the POXC (Permanganate Oxidizable Carbon) method developed by Weil, et al. 2003 but was modified for 2.5 g soil sample. In a 50 mL graduated polypropylene conical centrifuge tube 2.0 mL of a 0.2M KMnO4 with distilled water to the 20 mL mark were swirled into a solution with 2.5 g of dry soil. The sample was shaken by hand for 2 minutes (about 120 strokes/minute) and then allowed to stand for 10 minutes in a rack that was protected from sunlight. A standard curve was created using 0.01M and 0.02M KMnO4 standard solutions. Then 0.05 ml of the liquid from the settled soil - KMnO4 sample was transferred to a tube of distilled water that was brought up to the 50 mL mark and shaken and the absorbance of the sample was read. To estimate the amount of C oxidized, we assumed that 1 mol MnO4 is consumed (reduced from Mn+7 to Mn+4) in the oxidation of 0.75 mol (9000 mg) of C:
POXC (mg kg–1) = (0.02 Mol/L – (a+b*absorbance))*(9000 mg C/Mol)*(0.02 L soln./0.0025 kg soil)
where, 0.02 Mol/L is the initial solution concentration, a is the intercept and b is the slope of the standard curve, 9000 is mg C (0.75 Mol) oxidized by 1 Mol of MnO4 changing from Mn+7 to Mn+4, 0.02 L is the volume of KMnO4 solution reacted, and 0.0025 is the kg of soil used.

Percent soil carbon was analyzed at the USDA-ARS Pasture lab with an Elementar Vario Max elemental analyzer with combustion analysis of carbon to CO2 and measured with a gas chromatograph. The percent carbon was converted to percent organic matter using the conversion factor from Ranney, 1969 (% organic matter = 0.35 + 1.80 x % soil carbon). Total soil carbon was calculated by correcting soil percent carbon for soil bulk density that was sampled in the treatments in fall 2019.

The effects of termination timing on microbial nitrogen cycling dynamics

Nitrous oxide emissions were measured in fall of 2019 before fall termination (FT) into the beginning of December. A Gasmet portable gas analyzer was used to take the N2O field measurements from both the fall terminated (FT) and spring terminated (ST) plots. Measurements were converted to grams of N2O-N per hectare per day (g N2O-N/ha/day) for analyses. Mixed linear models were performed to determine if termination time (i.e. fall versus spring), day of measurement (DOY) or termination*DOY had a significant effect on N2O emissions using SAS (SAS Institute, USA).

Rhizosphere soil was sampled in the FT and ST plots before FT, and before ST to assess changes to the microbial N cycling gene abundances in the spring in the FT and ST plots. Rhizosphere samples were extracted by taking approximately 10 cm * 10 cm samples from the field, removing roots from the cores, and then washing the roots to remove the closely adhered soil particles. Samples were centrifuged to pellet the soil, DNA was extracted, and quantitative PCR (qPCR) was performed to quantify three nitrogen cycling genes, nirK, nosZI, and nrfA.