1. Introduction
In conservation systems, cover crops are utilized to improve soil properties and to enhance cash crop growth. The benefits associated with cover crops include reduced soil erosion, reduced runoff, and increased infiltration and water holding capacity. According to Reeves [
1], cover crops also provide weed suppression due to mulching and allelopathic effects, increased soil organic carbon, and reduced soil compaction.
Cover crops must be terminated at the appropriate growth stage to optimize biomass production and to create a thick mulch layer on the soil surface through which cash crops are direct seeded [
2,
3]. The planting of cash crops is performed by placing seeds into soil with desiccated cover crop residue on the soil surface. Thus, the proper management of cover crop residues is key to the effective no-till planting of cash crops into residue cover without interfering with the planting operation.
Common practice in accelerating the termination process is to utilize herbicides, such as glyphosate (Roundup
TM), as a supplement to rolling and crimping. This practice, however, is not permitted in organic systems, where the use of commercial herbicides is prohibited. Because of this restriction, mechanical termination must be as effective as chemical application. There are different cover crop management methods. One is mechanical termination utilizing rolling and crimping technology [
4,
5]. This requires injuring the plant with the crimping bars without severing stems. A roller-crimper flattens the plants and damages the stems with the crimping bars (equally spaced on the roller’s drum), crushing them to enhance desiccation. The desiccated residue forms a thick layer of mulch over the topsoil to protect the soil surface from harmful rainfall energy, thus reducing soil erosion and runoff. It also reduces weed germination and growth, and conserves water for the subsequent cash crop [
6].
Questions are often brought up about soil compaction when discussing reoccurring rolling and crimper over the same area. In response to these concerns, a field experiment was conducted by Kornecki et al. (2013) [
7] to evaluate soil compaction during 2007–2009 at the Cullman, Alabama location after rolling cereal rye and mixture (rye, crimson clover, and hairy vetch) once, twice, or three times. Researchers [
7] concluded that multiple rolling operations did not cause soil compaction on the non-wheel traffic area. During drought conditions in 2007, soil strength (a soil compaction indicator) for the rolled residue was 42% lower compared to the higher soil strength for cereal rye and mixture controls (7.2 MPa), implying that rolled down cover crops help reduced soil strength by covering the soil surface with a mulch layer and conserving moisture. Furthermore, rolling two or three times over the same non-wheel traffic area did not increase soil compaction. Higher gravimetric soil moisture in 2008 and 2009 lowered soil strength below 2 MPa, a critical value for root penetration resistance (Taylor and Gardner, 1963) [
8].
Another field experiment conducted by Kornecki (2020) [
9] with different roller-crimpers evaluated soil strength in both the plot area (area where only roller-crimper bars contacted) and in the tractor’s wheel track, before and after rolling cover crops. This study indicated that rolling two or three times over the same plot area did not cause soil compaction either in the plot area or wheel track. Data also showed that for the top 15 cm of soil, the soil strength value did not surpass 2 MPa, a level of restriction for root penetration [
8], and was solely related to changes in gravimetric soil moisture content (GMC).
A typical cover crop grown in the southern US is cereal rye (
Secale cereale L.). It can produce between 3 to 11 metric tons ha
−1 of biomass, providing benefits that include allelopathic weed suppression and a mulch effect due to improved soil surface cover [
10]. Cereal rye has the potential to generate high amounts of biomass both above and below the soil. Rye is known for taking in unused nitrogen from previous crops (scavenging), as well as bringing up potassium near the soil surface via the root system. However, the benefit from the nitrogen uptake is often not noticed in the succeeding cash crop due to the slow mineralization after termination or maturity [
11]. According to Ashford and Reeves [
2], rye termination rates above 90% were sufficient to plant a cash crop into desiccated residue without competing for resources. Legume cover crops, such as crimson clover and hairy vetch, are also used in the southern US to produce biomass and fix nitrogen from the air, which can then serve as a nitrogen source for subsequent crops in organic production in which artificial fertilizers are prohibited [
11]. Crimson clover can produce biomass in the range of 3923 to 6165 kg ha
−1 of dry matter, whereas hairy vetch produces slightly less at a range of 2578 to 5604 kg ha
−1 dry matter, with vines that can reach up to 3.7 m in length. Nitrogen fixation from crimson clover can provide from 78 to 168 kg ha
−1 of nitrogen and even more for hairy vetch, which provides a range of 101 to 224 kg ha
−1 [
11].
In the southern US, the mechanical termination of cover crops should be performed at least 3 weeks before planting a cash crop into flattened residue to minimize competition for soil moisture and nutrients [
12]. As with any method that leaves residue on the soil surface, mowing can interfere with planting operations. There is also a possibility for cover crop re-growth after mowing, depending on the species and timing of termination [
13]. Nevertheless, mowing is an effective method for rapidly terminating a cover crop at its appropriate growth stage before seed development. Mowed residues are left on the soil surface, resulting in cooler, wetter soils compared to cover crops that have been terminated with tillage and then incorporated. Cash crops can be no-till planted into a mowed cover crop. Flail mowers contain many small, double-edged knives that uniformly distribute finely cut residue on the soil surface. They generally require more horsepower than rotary mowers and tend to leave residue more finely chopped and evenly distributed on the soil surface, which accelerates decomposition compared to sickle bar and rotary mowers (Barbercheck and Borrelli, 2020) [
14].
Depending on the farm scale, fruit and vegetable producers use different power source sizes, including larger four-wheel tractors to small two-wheel walk-behind tractors. Therefore, producers need no-till equipment compatible with both large and small-scale power sources that are used on farms, especially since the number of urban farms producing fruit and vegetables for local farmers’ markets in the USA has been increasing. This increase is driven by consumer demand for healthy and fresh locally grown produce. Local farms usually own light four-wheel tractors, such as Kubota (Kubota Tractor Corporation, Osaka, Japan), or small walk-behind tractors, such as BCS (BCS S.p.A, Abbiategrasso (MI)–Italy), which has interchangeable attachments (Kornecki, 2015) [
15]. The majority of these small farms realize the benefits of utilizing cover crops that are widely promoted by USDA agencies, including the ARS and NRCS, to increase the sustainability of agriculture production. A factor limiting the adoption of cover crops is the lack of commercially available equipment needed to manage cover crops, such as roller-crimpers, that are compatible with different farm scales. In addition, the tradition of tilling soil for vegetable production is still strong in this region. Conventional tillage, however, causes increased soil erosion and nutrient loss, increases soil strength, and depletes soil organic carbon content [
16,
17].
Therefore, the objective of this study was to evaluate the effects of rolling and crimping and flail mowing crimson clover, cereal rye, and hairy vetch cover crops on cantaloupe yield in a no-till system.
2. Materials and Methods
A replicated field experiment was initiated in 2009 at the North Alabama Horticulture Research Center, in Cullman, Alabama (34.18° N; 86.85° W; 244 m above the sea level). The experiment was conducted on Hartsells fine sandy loam soil (fine-loamy, siliceous, sub-active, thermic Typic Hapludults) that had a bulk density of 1.49 g cm−3 and an organic matter content of 1.3%. Prior to this no-till cantaloupe experiment, a tomato test performed on plasticulture was conducted in the experimental area. Before planting cover crops for this test, the field was disked in the fall of 2009. No tillage was conducted between each growing season for this experiment.
The Hartsells soil consists of moderately deep, well-drained, moderately permeable, loamy residual soil, formed from weathered acidic sandstone, containing a thin band of shale or siltstone and comprising 11% clay, 26% silt, and 63% sand. At the top layer (depth from 0 to 127 mm), the soil is dark-grayish brown with a weak, fine granular structure. Here, it is very powdery, with many fine roots (10%) and sharp fragments of sandstone. In preparation for this test, a general analysis soil sample was collected that indicated a pH of 6.4 and in which liming was not recommended prior to initiating this test. The soil’s CEC was 8.37, with extractable nutrients from the test results that include: phosphorus (
p = 166 kg ha
−1), potassium (K = 290 kg ha
−1), magnesium (Mg = 229 kg ha
−1), and calcium (Ca = 2042 kg ha
−1), which were all at the optimum level (high or very high) for cantaloupe production. The main field activities at each growing season are shown in
Table 1.
The experiment was a randomized complete block design (RCBD) in which three cover crops were used: crimson clover (
Trifolium incarnatum, L.), cereal rye, and hairy vetch (
Vicia villosa, L.). The selected cash crop for this experiment was cantaloupe (
Cucumis melo, L.). The cover crop residue management treatments were assigned as depicted in
Figure 1, with selected cover crops and their termination treatments (equipment).
Crimson clover, cereal rye, and hairy vetch were planted each fall (2009–2011) at a rate of 22 kg ha
−1, 112 kg ha
−1, and 22 kg ha
−1, respectively, utilizing a Tye no-till drill from AGCO Corporation, Duluth, GA, USA (
Figure 2).
In each spring (for the 2010–2012 growing seasons for cantaloupe) before rolling and crimping or flail mowing the cover crops, biomass samples from each plot were collected (0.25 m
2 sample area). Additionally, the plant heights (of 10 randomly selected plants per plot) for each cover crop were recorded. Biomass samples were oven dried using a programmable electric shelf oven, Model No. SC-400 [
18], with forced convection air flow for 72 h at a temperature set to 55 °C (Grieve Corporation, Round Lake, IL, USA) to estimate the amount of dry biomass produced in each plot for each cover crop. The cereal rye cover crop was rolled and crimped at the early milk growth stage (Zadoks scale: #73; [
19]), which is a desirable growth stage for termination that typically produces an optimum level of biomass [
20]. Crimson clover was terminated at the early flowering stage [
21] and hairy vetch was terminated at the 80% bloom stage [
22], both of which are recommended growth stages for mechanical termination.
To terminate cover crops, two experimental roller-crimpers developed at NSDL were tested: a two-stage roller-crimper (
Figure 3), and a powered roller-crimper (
Figure 4). For comparison, a commercially available flail mower was utilized to manage cover crops, as the majority of farmers already own flail mowers on their farms (
Figure 5).
The effectiveness of mechanical termination using roller-crimpers and flail mowers was compared with the untreated (non-rolled) cover crop. However, the control treatment with respect to cantaloupe yield was the flail mowing method. Transplanting cantaloupe into flail-mowed residue is a more practical operation for a no-till system. In contrast, in Alabama, planting the cantaloupe seedlings into an untreated, non-rolled cover crop is impractical.
Cover crop termination rate data using the previously mentioned equipment were collected and evaluated 7, 14, and 21 days after treatment application, denoted as days after treatment (DAT), and compared to the control (non-rolled cover crops). During this period, volumetric soil moisture content (VMC) was also measured. The results were compared to the untreated cover crops. Termination rates were estimated using a handheld light sensor-based chlorophyll meter SPAD 502 (Konica-Minolta, Ramsey, NJ, USA). This portable sensor is capable of instantly measuring the chlorophyll content or “greenness” of plants. The SPAD 502 quantifies slight changes or trends in plant health long before they are visible to the human eye and provides a means of non-invasive measurement. The meter is clamped over leafy tissue and an indexed chlorophyll content reading (usually from 0 to 50.00) is recorded in less than 2 s. The data logging version of the SPAD 502 (item 2900DL) allowed for readings to be compiled more easily for statistical analysis. Since the state of the plant’s greenness is related to its chlorophyll activity (e.g., 50 for healthy plants, and 0 for a dead plant with no chlorophyll activity), this concept was used to detect different stages of cover crop termination due to plant senescence from injury caused by mechanical termination using the roller-crimpers. To obtain an adequate assessment for each plot, five readings of plant tissue were collected in each plot by manually clamping the chlorophyll meter on randomly selected plants and storing the readings in the data logger. These five readings were averaged for each plot. Cover crop termination rates on a scale of 0% (no injury symptoms) to 100% (complete death of all plants) were based on a procedure described in Kornecki et al. [
25]. The percentages of cover crop termination rates were transformed using an arcsine square root transformation method [
26], but this transformation did not result in a change in the analysis of variance (ANOVA); thus, non-transformed means are presented. Similarly, volumetric water content (VWC in %) was measured (five readings per each plot and averaged) using a portable TDR 300 moisture meter with 0.12 m long rods from Spectrum Technologies (Aurora, Illinois). The selection of 0.12 m long stainless-steel rods was based on the necessity of determining the volumetric moisture content (VMC) and water availability in the root zone (~0.06 m depth) of the transplanted cantaloupe seedling. Monthly weather information for all growing seasons is presented in
Figure 6 and includes rainfall, minimum temperature, and maximum temperature [
27].
Three weeks after cover crop termination, the cantaloupe seedlings (cultivar Athena) were transplanted directly into cover crop residue utilizing a modified RJ one-row no-till vegetable transplanter (RJ Equipment, Blenheim, ON, Canada;
Figure 7).
The commercially available transplanter was modified at the NSDL by outfitting it with a custom-built subframe and vertical subsoiling shank to alleviate soil compaction while transplanting the cantaloupe seedlings in a single operation. Prior to transplanting the cantaloupe seedlings, 43 kg ha−1 of nitrogen was applied to all plots using calcium nitrate.
Post-transplanting, 11.2 kg ha−1 of nitrogen was applied weekly, alternating between calcium nitrate and general purpose 20-20-20 water-soluble fertilizer. Weekly fertilization for the entire test was performed using drip irrigation.
Data were analyzed using the general linear model procedures in SAS PROC GLM (SAS Institute, 2016) [
28]. All dependent variables were first analyzed to measure the effect of years as fixed effects, which was used to measure the magnitude of year interaction with cover crops and treatments. The treatment means were separated using the Fisher’s protected least significant differences (LSD) test at the 10% probability level with respective probabilities (
p-values), meaning there is a 10% chance that weather and nature could have contributed to the differences observed, which is common with agricultural field experiments. Cover and termination treatments were considered to be fixed effects and blocks were considered to be random effects [
29]. Where interactions between treatments and weeks or years occurred, data were analyzed separately. In contrast, where no interactions were present, data were combined.