1. Introduction
Improving the economic resiliency on dairy farms is crucial to sustaining dairy operations and regional food security. Feed has long been the largest expense on dairy farms and other livestock operations in the Northeastern United States [
1,
2]. Increasing the on-farm forage inventory with dual-purpose cover crops (DPCCs) can reduce feed costs, increase land-use efficiency, and improve nutrient management [
1,
3]. Dual-purpose winter grain crops have been successfully utilized in major grain-producing regions of the United States [
4,
5]. The concept of DPCCs also utilizes small winter grains grown after summer cash crops, as reviewed by Sulc and Franzluebbers [
6]; the dual-purpose nature refers to the crop serving as both a cover crop and a forage crop [
7,
8,
9].
Traditional cover crops contribute significant ecosystem services [
10,
11]. In addition to providing many of the same benefits as traditional cover crops, DPCCs offer a tangible economic benefit by reducing the need to purchase off-farm feed, which may incentivize growers to use cover crops when they are not strongly motivated by an interest to improve the long-term production capacity of their soils, or increase environmental sustainability [
10,
12,
13]. Traditional cover crops are also used to increase nutrient availability in the soil, and to supply a nutrient credit to cash crops [
13]. DPCCs are not grown for this purpose. Rather, DPCCs remove nutrients from the soil when harvested [
1,
14].
Manure application rates have traditionally been determined based on the nitrogen (N) requirements of silage corn (
Zea mays L.). Consequently, phosphorus (P) is often overapplied, and soils on dairy farms frequently accumulate high levels of plant-available P. These high P levels can be implicated in the nonpoint source pollution of lakes, streams, and rivers [
15,
16]. The prevention and management of a high soil P is important to mitigate the risk of P contamination in the surrounding environment [
17,
18]. In dairy systems, traditional cover-crop management returns captured P to the soil. By contrast, DPCCs have the potential to remediate the levels of soil P [
8,
13]. Part of the P captured in feed can be exported off-farm in the form of milk or meat. The production of additional forage reduces the need to purchase off-farm feed, thus reducing P inflows from off-farm. Extra forage production may create the opportunity to sell feed, resulting in a P export. Capturing P from fields with excessive levels also creates the opportunity to redistribute P to fields with lower P contents.
Furthermore, as both N and P can lead to the eutrophication of water systems, the management of both nutrients must be prioritized when using cover crops to provide environmental protection for waterways and bodies [
19,
20]. In this manner, DPCCs may be an important component of on-farm nutrient cycling, and sustainable manure application.
The DPCC residue that remains in the field following harvest returns some N, P, and carbon (C) to the system. However, little information is available on the total residue left after DPCC harvest, and the associated N, P, and C that are returned to the soil. Quantifying N and P returned by DPCC residues is necessary to predict nutrient cycling and nutrient mass balances. As cover-crop residues help build soil organic matter, quantifying the C of the returned residues after a DPCC harvest is critical to understand how this practice may affect soil health [
21,
22].
The general decomposition trends of cover crops have been well documented [
23] and the 25:1 rule, which states that decomposition without nitrogen immobilization will occur at C to N ratios below 25:1, was established nearly 100 years ago [
24]. Reports indicate that when soil microbes compete with plants for available soil N, including synthetic fertilizer amendments, N becomes unavailable for plant uptake, as it is incorporated into the microbial biomass [
25,
26]. In accordance with this existing knowledge, we hypothesized that because DPCC residue is primarily composed of dense stems that are likely to have a high-C and low-N composition, the residues will decompose slowly and release little N, as demonstrated by Quemada and Cabrerra [
27]. Moreover, because the degradation of residues with high C:N ratios is associated with microbial assimilation of inorganic N and thereby N-immobilization [
28,
29,
30], we hypothesized that the high C:N residue left after removing DPCC biomass above 7.6 cm would encourage a community of decomposers that would compete for N and induce N-immobilization in corn production systems.
Reports on the impact of small-grain cover-crop residues on subsequent corn silage yield include both negative and neutral effects [
31,
32,
33]. The effect of DPCCs on subsequent corn silage yield is different from a traditional cover crop, due to reduced biomass, removed nutrients, and high C content of residues. Furthermore, research is ongoing to determine the most effective small cereal grain in the DPCC system, with great interest in the rye (
Secale cereale L.), wheat (
Triticum aestivum), and triticale (xTriticosecale) as candidate crops.
Despite the potential negative effects on corn silage, we hypothesized that DPCCs would be a suitable source of forage and increase total forage production. We also hypothesized that DPCCs would remove substantial N and P from the manured field and ameliorate P-accumulation. This three-year field study evaluated rye, wheat, and triticale as DPCCs in a manured field, and their potential impact on subsequent corn silage production.
2. Materials and Methods
2.1. Experimental Site and Field Management
Field experiments were conducted from September 2015 to September 2018 at the University of Massachusetts Crop and Animal Research and Education Farm, Deerfield, Massachusetts (42°28′37″ N, 72°36′2″ W). The soil of the experiment site is a Hadley fine sandy loam, which is characterized as coarse-silty, mixed, superactive, nonacid, mesic Typic Udifluvents (Natural Resource Conservation Service, 2013). The weather conditions for the duration of this experiment are presented in
Figure 1.
Soil samples were taken, to a depth of 0.2 m, sent to the University of Massachusetts Soil and Plant Tissue Testing Laboratory, and analyzed using the Modified Morgan solution.
The pH ranged from 6.2 to 6.7 throughout the duration of the study. Lime was applied as recommended, 1500 kg ha−1 year on average. In two of the three years of this experiment, phosphorus was in the “above optimum” range (>14 ppm) in the experimental field. In each year of the experiment, liquid dairy manure obtained from a local dairy farm was applied at a rate of 75,000 L ha−1 on 8/29/16, 9/4/17, and 8/30/18, and incorporated into the soil using a disk harrow within 24 h after application, to preserve the ammonia N. The rate of application is consistent with local practices. Manure analysis indicated that per 1000 L there were 2 kg total Kjeldahl nitrogen, 0.7 kg P2O5, 1.8 kg K2O, 1.1 kg Ca, and 0.6 kg Mg, which was more than sufficient to address nutrient requirements indicated by soil tests.
Cover crops were planted on 8/30/16, 9/5/17, and 9/1/18. In each replication, one plot was left fallow. DPCCs were harvested in late May, followed immediately by termination with herbicide. Corn was planted within one week following DPCC termination. Start-up N was applied at the time of corn planting, and sidedress N was applied in the first week of July, approximately six weeks after planting. Corn was harvested in the last week of August/first week of September.
2.2. Experimental Design and Management
The experiment was laid out in a completely randomized block design, with four replications per year in each of the three years of the DPCC portion of the experiment, including decomposition. The DPCC treatment had four levels: wheat (var. Arapahoe), triticale (var. NE426GT), and rye (var. Prima) planted on 1 September each year, as well as a no-DPCC plot left fallow to serve as a negative control. All three winter grains were planted at the seeding rate of 123 kg ha−1 using a seven-row grain drill modified for research plots. Experimental plots were 5 m wide and 6 m long.
The DPCCs were harvested at Feekes stage 10, boot stage, on 5/4/16, 5/16/17, and 5/15/18. In each plot, two one-meter-long samples were hand-harvested and portioned into biomass above 7.6 cm, to mimic the mowing height of a flail chopper, and biomass below 7.6 cm, i.e., DPCC residue (
Figure 2). These samples were used for biomass and laboratory analysis.
A second, larger stubble sample from each plot was collected, mixed, and placed into decomposition bags. Decomposition bags were made from a 0.425 mm mesh screen, and were 35 × 35 cm. Seventy-five grams of fresh tissue was placed into each litter bag, and immediately returned to the field, in the original plots from which the residues were collected. Bags were placed on the soil surface and secured in place with landscaping staples. The mesh size allowed for bacterial and fungal decomposers, but did not allow for earthworms. The smaller mesh size prevented the overestimation of decomposition, and the research site historically had a very small earthworm population (observationally, ≤1 earthworm per 1000 cm3) so the impact was considered to be negligible.
Decomposition samples were collected eight times, at 1, 2, 3, 4, 5, 7, 9, and 11 weeks after being placed in the field. Samples were removed during sidedress N application. The dry weights of the separately collected residue samples were used to determine the species-specific dry weights associated with the fresh residues placed in each decomposition bag.
After collecting subsamples, the DPCC plots were mowed with a flail chopper, which removed biomass above 7.6 cm from the experimental plots. Pre-emergent herbicide (S-metolachlor) and post-emergent herbicide (glyphosate) were applied to terminate the cover crops and to control spring weeds. Herbicide was also applied to the no-DPCC plots. No further herbicide was applied, and weed control was effective throughout the summer.
Corn was planted on 5/20/16, 5/22/17, and 5/18/18, with a no-till planter, at a rate of 78 thousand plants ha
−1, and with no start-up nitrogen fertilizer. Synthetic N in the form of calcium ammonium nitrate (27%) was added when corn plants reached approximately 28 cm, based on the pre-sidedress nitrate test (PSNT) average for the field [
34], which is approximately 50 kg ha
−1 each year. The original corn hybrid selected for this experiment, planted in 2016, performed poorly due to hybrid-specific issues not associated with the experiment itself. Thus, corn data from 2016 were not included in the silage data analysis. In 2017 and 2018, corn hybrid Dyna-Gro D32RR56, 92 RM, was used for this experiment.
Corn was harvested by hand when it reached 50% milk-line (9/2/17 and 8/31/18). Samples were taken from a 3 m length of the row, and ears and stover fresh weight were measured separately. The number of plants and ears was recorded. Two stover samples and three ear samples were randomly selected and dried in a forced-air oven (Gruenberg Oven Company, Williamsport, PA, USA) at 80 °C to constant weight. The yield of corn silage was adjusted and reported at 70% moisture.
2.3. Laboratory Analysis
The oven-dried samples were weighed for biomass. Samples for laboratory analysis were ground with a Foss Mill (Foss Cyclotec 1093, Hilleroed, Denmark) to pass through a 0.42 mm screen.
For the DPCC and decomposition samples, a 0.2 g subsample was used for nitrogen analysis according to the Kjeldahl method (Standard Method 4500-N (Org) C. Semi-Micro-Kjeldahl). Samples were then analyzed with a Lachat8500 flow injection analysis spectrophotometer, Lachat Total Kjeldahl Nitrogen Method (TKN) Number 13-107-06-2-D (Zellweger Analytical, Milwaukee, WI, USA). For orthophosphate analysis, 0.2 g of tissue was weighed into porcelain crucibles, and placed in a combustion oven for 24 h at 500 °C. After the crucibles cooled, phosphorus was brought into solution with 10% hydrochloric acid. The ash and acid mixture was filtered and then analyzed using Lachat8500, Lachat Orthophosphate Method Number 10-115-01-1-V. With the exception of crude protein, which was calculated from the TKN value, feed-quality assessments, including the parameters used to determine the milk value of DPCCs, were evaluated with near-infrared reflectance (NIR) spectroscopy (Unity Scientific, Milford, MA, USA). Milk 2006 was used to estimate the milk value of DPCCs. The DPCC residue samples were sent to the Soil and Tissue Testing Laboratory at the University of Massachusetts Amherst for C:N analysis on an elemental analyzer.
2.4. Statistical Analysis
The DPCC data used in this study represent four replications per year, for a total of 12 replications. Total values represent both the harvested portion of DPCCs, and the residue remaining in the field (
Table 1). Statistical analyses were performed using the GLM procedure in SAS, Version 9.4 (SAS Institute, 2016). The main effects were DPCC species plus a “no-DPCC” treatment (4 levels), replications (12 levels), and decomposition collection dates (8 levels). The year was not included in the model for this long-term study, because we were not interested in year-specific details, did not measure climactic or environmental conditions in the field to justify the treatment of the year as a fixed variable, and were interested in consistently detectable trends over a triennium. Statistical significance was declared at
p ≤ 0.05. For statistically significant effects of discrete variables, mean separation was determined by Tukey’s honestly significant difference test. For statistically significant effects of continuous variables, regressions were conducted using the appropriate polynomial determined in Proc MIXED.
4. Discussion
4.1. Total and Harvested DPCC Biomass/Yield and Captured N and P
Feed is by far the largest annual expense on dairy farms in the Northeastern United States (Annual Northeast Dairy Farm Summary, 2010–2019). The results of this study suggest that the incorporation of DPCCs into dairy systems can successfully offset feed expenses by producing additional, high-quality forage, without significantly reducing corn silage yield.
DPCCs provided an average of 4.1 Mg ha
−1 of additional forage. The yields of the harvested DPCCs in the current study were consistent with the earlier reports [
35,
36]. The total DPCC biomass production (stubble plus harvest) of 7.6 Mg ha
−1 is also consistent with the total cover-crop biomass reported in the literature [
37,
38].
Although there were slight variations in the feed values of the DPCCs studied, all three species were of fair value (
Table 1). The lower RFV of rye can be attributed to a slightly more mature developmental stage, and to the growth pattern of rye, which produces a higher stem to leaf ratio [
39]; stems have a higher indigestible fiber content than leaves [
40]. The high RFV of wheat and triticale is reflected in the associated milk production potential of these crops. Per Mg, wheat and rye are associated with the greatest milk production potential. Although yield differences among DPCCs were not statistically significant, the lower yields of wheat and triticale as compared to rye, combined with higher RFV values, resulted in equivalent predictions of milk production on a per hectare basis. Milk production per Mg was estimated using a combination of field, lab, and book values, and the estimates are consistent with the reported wet chemistry analysis [
41]. Plant breeding may tip the scales to create rye varieties with a higher milk value, or wheat or triticale varieties that maintain their milk value and can consistently produce high yields. In either event, milk value per hectare would increase.
Total DPCC biomass captured 102 kg N ha
−1 and 35 kg P ha
−1, equivalent to 68% of both the N (150 kg ha
−1) and P (52 kg ha
−1) applied in the form of manure in the fall. The captured nitrogen results are in alignment with the findings of Hashemi et al. [
19] that on average, a winter rye cover crop recovered 119 kg N ha
–1 from manured soil when planted in early September.
An earlier report indicated that 40% of the N applied in manure is available within the first year of application [
16]. If only 40% of manure N is available in the first year, that would be equal to 60 kg ha
−1 of manure-available N for DPCCs to uptake. As the DPCCs captured an average of 102 kg N ha
−1, 70% more than predicted based on Eghball and Power’s report [
16], this suggests that the DPCCs scavenged available N from sources additional to manure N applied prior to planting.
In the absence of a cover crop, these nutrients would be subject to loss to the environment during the fall-to-spring months [
19]. Ultimately, approximately 60% of the captured nutrients were removed from the field in the harvested DPCCs. In manured systems where N and P can accumulate in amounts exceeding the corn silage removal rates, the introduction of DPCC harvest reduces the field nutrient load and temporarily immobilizes a portion of total soil nutrients. This combined effect may potentially be more efficient in reducing the risk of nutrient loss than the use of traditional cover-cropping practices alone.
4.2. Characteristics of DPCC Residue
It was interesting to find that DPCC residue contained 46% of the total crop biomass. Corresponding to the substantial portion of total biomass in the residue, approximately 40% of the captured N and P remained in the field within the residue. The high-C input associated with the DPCC residue (1.6 Mg ha−1, on average) indicates that DPCC residue alone has the potential to build soil carbon. These findings are important because some researchers are concerned that the DPCC concept jeopardizes the agroecological benefits associated with traditional cover-crop practices.
The small-winter-grain DPCCs should be considered a valuable cover-cropping strategy, even though the total amount of returned biomass is less than traditional cover-cropping with small winter grains that are not harvested. Just as different cover-crop species provide different results so, too, do different cover-crop management schemes provide different results. Accordingly, both the species and the management scheme should be selected relative to the intended outcome.
It is important to note that our study did not measure root biomass. A report indicated that belowground cover-crop residues, i.e., roots and their associated products, were three times more likely to persist in soil C pools as compared to the aboveground inputs [
42]. The combination of substantial aboveground inputs in the DPCC system, and the associated ’albeit unmeasured’ belowground residue, demonstrates a strong potential to provide similar benefits to many popular cover crops, in building soil organic matter, and contributing to soil health without sacrificing soil C building. Surely, future DPCC research must place a strong emphasis on root biomass quantification and soil-C cycling, to fully evaluate the system.
The distribution of DPCC biomass in the harvested and returned portions can likely be attributed to proportions of stem and leaf. The harvested portions of DPCCs are leafier and less dense than the unharvested residue, characteristics that simultaneously make the DPCCs acceptable for use as forage. The thick base of DPCCs is primarily composed of the fibrous, high-carbon stems of many tillers, as well as the crown tissue. Thus, the dense residues left behind after harvest contain almost half of the total aerial biomass.
4.3. DPCC Residue Decomposition Trends
It was not unexpected that there was no significant impact from the DPCC species on the decomposition trend over time. Although there were significant differences in the C percentage of DPCC species (
p ≤ 0.05), these findings are likely of little biological significance, as the C:N ratios of the DPCC species were not significantly different. C:N, and not C percentage, is the primary influencer on crop decomposition trends [
43].
Interestingly, rye and triticale were composed of the greatest percentages of water, and returned the greatest amounts of water to the field. The 4 Mg ha
−1 of rye residue left in the field returned 5500 L of water. Triticale and wheat left 33% and 47% less total fresh material in the field, respectively, and returned an associated 4300 and 3600 L of water. It is well known that moisture affects the decomposition rates of organic material [
44]. Despite the differences in moisture associated with species, as aforementioned, the species differences did not contribute to significantly different (
p ≥ 0.05) decomposition trends as may be expected. Presumably, environmental moisture conditions outweighed any effect of the initial moisture content of the residue itself.
The initial amount of decomposition in week one (20% of the original amount) likely came from the small amounts of leafy residue remaining in the residue. The retarded decomposition for the remainder of the season is characteristic of high-C stems. Earlier reports [
25] indicated that the differences in stem and leaf C:N ratios resulted in differential decomposition of these tissues, which supports our interpretation of this decomposition trend. Despite the continuous decomposition of biomass, N was only released from the residue in the first week after DPCC termination, at which time 20% of the initial N content was released. It is likely that leaf matter, which has a lower C:N than stems, and thus decomposes faster, is the primary contributor to N release in this period.
The consistent biomass decomposition of the DPCC residues paired with the lack of N release indicates that the C portion of the residue is being decomposed while the N portion is not, which can result in N immobilization dynamics. These conclusions are further supported by the observation that the N concentration of the DPCC residues steadily increases throughout the season. Our findings are consistent with an earlier study [
45] in which increased C mineralization from wheat straw residues throughout the growing season was coupled with decreasing soil N levels attributable to N immobilization by microbial decomposers. In the same study, elevated levels of fungal and bacterial enzymes associated with the degradation of C-, H-, and O-based plant matter, i.e., cellulases and hemicellulases, were observed.
The degradation of plant cell walls results in biomass decomposition without N release, thus increasing N concentrations within the residues that remain intact, and failing to provide supplemental soil N. As a result, soil microbes decomposing high-C tissues must obtain N to support their metabolism from alternative sources in the environment, which can create microbial competition with crops for the available soil N. In a corn system, managing fertility by applying fertilizer based on recommendations may not reverse N immobilization dynamics, as the recommendations generated by the pre-sidedress nitrate test account for only the needs of the corn, and not decomposers in the soil. Therefore, systems with high-C residues may therefore require additional N inputs to avoid yield penalties.
Despite potential N immobilization dynamics, there were no yield penalties to total corn silage yields compared to the no-cover-crop treatment. When adding in-season N, the amount of N that was mineralized from soil organic matter remained unaccounted for. The soils at the research site had 1.8–2.1% organic matter in each year of this study. The N mineralized from this organic matter may have provided the additional fertility needed, and may have prevented a yield penalty in the succeeding corn silage.
Depending on the corn planting date, sidedress N application would occur between week seven and nine after DPCC termination, at which time only 40% of the DPCC residue had decomposed relative to the starting amount, and the N release had stagnated. This creates the risk of accidental under-fertilization in the face of N tie-up, as the sidedress N plans for only the needs of the corn, and not the needs of microbial decomposers. Such an effect may be exacerbated because only the high-C portions of the cover crop are being returned, and the lower-C leaf matter that could supply compensatory N is removed.
On dairy farms, the organic matter added to soil in the form of manure may supply additional fertility to compensate for N immobilization. The soil organic matter built up over time from cover-crop inputs continues to be mineralized in subsequent seasons. Future research should focus on the role of high-C inputs in maintaining regular and adequate N mineralization from soil organic matter, and overcoming possible immobilization dynamics. It is possible that additional fertility may be needed for a short time to compensate for the high C inputs, and that the DPCC systems would eventually stabilize.
4.4. Corn Silage Yield and Quality
DPCCs did not show significant negative effects on corn silage yields, and the ear contribution indicated good silage quality following all treatments [
46]. There was a significantly lower proportion of ear in corn planted after triticale than after wheat and rye; however, the proportion of ear in corn planted after all DPCC treatments was not significantly different from that of corn following the no-cover-crop treatment. These effects of cover crops on corn yield, even with the DPCC modification, are in agreement with the earlier findings of Snapp and Surapur [
34] that suggested that there is no effect of a rye cover crop on subsequent corn production. However, we hypothesized that discrepancies in the literature surrounding the debate over the effect of cereal cover crops on corn production may be attributed to two items: firstly, differences in corn genotypes, as some hybrids may be more suitable to this system than others, which is an avenue for future research; secondly, differences in field management, including the cover-crop termination strategy, and the corn-planting implements.
5. Conclusions
This study investigated the reduction in feed cost, and alleviation of high levels of P, which are considered two major issues in dairy farms. Averaged across three years and three species, DPCCs produced 7.6 Mg ha−1 dry matter, with the potential of producing 3987 kg of milk per hectare of land. Additionally, excluding the roots, DPCCs captured 35 kg ha−1 P and 100 kg ha−1 N, which demonstrates efficient on-farm nutrient cycling and keeps the orthophosphate active, and prevents it from becoming unavailable to plants. There was 3.5 Mg ha−1 high-C stubble left in the field after the DPCC harvest, which can significantly contribute to the soil C cycling. The N release from decomposing biomass was well behind biomass decomposition, which can result in N immobilization in succeeding corn. However, in the current manured-field study, the DPCCs did not cause yield or quality penalty in succeeding corn silage.