Comparative Analysis of Mechanical In-Field Corn Residue Shredding Methods: Evaluating Particle Size Distribution and Rating of Structural Integrity of Corn Stalk Segments

Abstract

The European corn borer is a major pest of corn that overwinters in corn stubble and stalks. Shredding these residues disrupts the larvae’s habitat or directly harms them. A corn header has been engineered with a new type of cutting tool on its horizontal choppers, featuring sharp edges and dulled flails, to shred corn stubble near the soil surface. This study investigated the effect of the dulled flails on the shredding intensity of corn stover. Field trials compared flail knives with standard knives for particle size distribution of corn stover and structural integrity of corn stalk segments. Additionally, a common two-step method, which involved a standard knives-equipped corn header followed by tractor-driven flail mowers, was tested. The flail knives reduced the mean particle size by 3.6 mm compared to the standard knives. Subsequent processing with tractor-driven flail mowers, following the corn header using standard knives, led to a reduction in mean particle size by 11.8 mm. It also further reduced the number of incompletely destroyed stalk segments. However, completely intact internodes were scarce in all methods. Given that flail knives enhance shredding intensity without a second processing step, this concept is concluded to be effective for corn stover shredding.

1. Introduction

The European Corn Borer (ECB), Ostrinia nubilalis (Hübner, 1796), is recognized as one of the most important insect pests affecting corn. The ECB poses a threat to large agricultural areas worldwide and climate change is expected to lead to increased pest pressure [1,2,3,4,5,6,7]. The primary damage is caused by its larvae feeding inside the corn stalks and ears, which affects the transportation of water, nutrients, and assimilates. Additionally, the mining activity undertaken within the stalks reduces the mechanical integrity of the corn plant and can lead to stalk breakage, consequently leading to challenging harvesting conditions and heightened losses of corn ears [5,8,9]. If heavily infested, yield losses in grain corn can range from 5–40% [6,10,11]. In addition to direct damage, injuries to the corn plant caused by the ECB promote infection by fungal pests such as Fusarium spp. [9,12,13,14,15,16].

Due to the significance of the ECB, various approaches for direct pest control have been pursued. Nonetheless, the management of ECB over the past 25 years, particularly in North America, has primarily relied on the cultivation of transgenic Bt-corn varieties, which effectively suppress ECB [5,17,18]. Bt-corn can reach efficiencies up to 97–100% [10,19]. In most EU countries, the cultivation of Bt-corn is prohibited, leading to reliance on biological control agents (primarily Trichogramma spp.) and chemical and biological insecticides for the direct control of the ECB, but the efficiencies are dependent on optimal timing of the application and influenced by year-specific effects [19,20,21,22]. Therefore, agronomic and cultural measures, such as crop rotation and tillage practices, have been implemented as additional preventive control measures. Furthermore, corn residue shredding has long been recommended [7,17,23,24].

The corn residue encompasses all plant organs except for the corn kernels. Unless otherwise specified, the following distinction is made between corn stubble and corn stover. Corn stubble refers to the portion of corn stalks that remains rooted in the soil after harvest. Corn stover includes all plant components that remain loosely on the soil surface after harvest. Corn stalks comprise the largest proportion of corn stover, ranging from 42% to 56% of the total dry mass, followed by leaves (18–27%), cobs (15–21%), and husks (8–13%) [25,26,27,28,29].

During the winter, the ECB persists in its mature larval stage, primarily within its tunnels inside the corn stubble and stalks, where it uses the enclosed space for protection. By the time of the corn harvest, the majority of the larvae are located in the corn stubble, generally within 30 cm above the ground surface, or from the second node downward [23,30]. By shredding (downsizing and/or splitting) the corn stubble, the overwintering habitat of the ECB can be disrupted, thereby reducing the population of larvae that survive into the subsequent growing season. Even though the majority of larvae are found in the stubble, it is also essential to shred the stalk segments within the corn stover, as they too house a portion of the larvae. Moreover, the larvae are mobile, meaning that if they are not directly killed by shredding the stubble, they can colonize new stalk segments [23].

The effectiveness of shredding corn residue in managing ECB has been shown in studies through assessments that count the number of larvae before and after shredding [7,23,31]. However, the direct correlation between the degree of destruction of corn stubble and corn stalk segments and the mortality of ECB larvae has not been extensively studied. It is assumed that slightly crushing the corn stalks, such as by driving over them, is not sufficient to deprive the ECB of its winter habitat [32]. Sieve analyses, along with visual assessments and ratings, are often conducted for the comparative evaluation of corn residue shredding methods, without drawing direct conclusions on ECB mortality. It is assumed that the corn stalks and stubble must be shredded to a degree where they do not provide sufficient space for the approximately 25 mm long larvae [33,34] to overwinter, or are split open, allowing water to penetrate and inadequately protect the larvae. Even though there are suggested procedures to compare different shredding methods for corn stover [35,36] and cereal straw [37], the literature presents a variety of approaches, complicating the comparability of different studies.

The initial stage of corn stover shredding occurs during harvesting using combine harvesters and corn headers. The level of shredding intensity largely depends on the configuration of the corn header. This includes the design of the snapping rolls equipped with cutting edges or blades, as well as optional attachments such as horizontal choppers [38]. The research conducted by Handler et al [39] underlines these differences. Three distinct models of six-row corn headers designed for 70 cm row spacing (manufactured by Carl Geringhoff GmbH & Co. KG, Ahlen, Germany) were investigated. The evaluated models included the Rota Disc (featuring three-roller row units), the Horizon Star (equipped with three-roller row units and horizontal choppers), and the Mais Star (featuring two-roller row units and horizontal choppers). The two corn header models with three-roller row units were found to shred 90% of the stalk mass shorter than 20 cm. Only 1–2% of the stalk mass was longer than 30 cm. In the case of the Mais Star with two-roller row units, however, only 62% of the stalk mass was shorter than 20 cm, and 18% was longer than 30 cm. The Mais Star series only split about 68% of the stalk parts longitudinally. In contrast, the Rota Disc and Horizon Star achieved over 90%. In this context, it should be noted that only the portion of the plant material processed by the corn header was considered. Plant material that passed through the threshing mechanism of the combine harvester was not taken into account.

Due to the insufficient processing capability of current corn headers towards corn stubble, an additional pass is required to effectively break down the corn stubble. Several devices, including disc harrows and knife rollers, are used for this task. However, tractor-driven flail mowers have demonstrated the highest shredding intensity towards corn stubble and stover and are often preferred for this task [40]. There exists a variety of flail mower types, which differ in terms of their working width, the design and arrangement of the shredding tools, and shredding intensity [41,42]. Field tests and studies have investigated how various flail mower variants affect the shredding intensity of corn stubble and stover. These studies demonstrated that flail mowers can effectively process standing corn stubble, resulting in stubble heights ranging from 4 cm to 10 cm [31,43]. However, processing flattened or bent corn stubble poses challenges. If the stubble is lying on the ground or if it no longer provides enough resistance, it may not be adequately captured by the tools of the tractor-driven flail mowers, leading to longer residual stubble and a reduced overall processing quality [1,32,35,44].

One way to address this issue is to shred the corn stubble before it is run over by the combine harvester or following machinery, by having the corn header perform this task. Recently, two corn headers capable of shredding corn stubble close to the ground surface were introduced.

The primary objective of our joint project was to develop and test such a solution. The Horizon Star* III (HS3) corn header (Carl Geringhoff GmbH & Co. KG, Ahlen, Germany) was introduced in 2019 and is specifically designed to operate close to the soil surface. It incorporates a new toolset (flail knives) for horizontal choppers to effectively shred the corn stubble near ground level. These cutting tools feature dulled flails at their ends to hit the corn stubble. For a comprehensive examination of the technical details and factors influencing its power consumption, refer to Ramm et al. [45].

The Stubble Cracker System (Claas Selbstfahrende Erntemaschinen GmbH, Harsewinkel, Germany), introduced in 2022, combines two rotating skids into one unit. Each skid follows one row of corn stubble and is equipped with two hammers to shred the corn stubble. These units are mounted at the rear of the corn header. Herter and Schwaer [46] conducted a comparison between a standard corn header, a corn header equipped with the Stubble Cracker System, and a two-step method which involved harvesting with the standard corn header followed by a pass with a tractor-driven flail mower. The shredding intensity of corn stubble was assessed by visual inspection and rating of the corn stubble using a method inspired by Brunotte and Voßhenrich [35]. The standard corn header achieved a quality score of 14%, the corn header with the Stubble Cracker System achieved 87%, and the two-step method achieved 100%.

With the ability of corn headers to sufficiently shred corn stubble, the question arises whether an additional pass with a tractor-driven flail mower is still necessary. However, tractor-driven flail mowers not only shred the corn stubble but also the corn stover, leading to a reduction in particle size. Sieve analyses of corn stover by Grosa et al. [31] demonstrated that, depending on the flail mower model, the coarse plant material mass (>63 mm) decreased by approximately 15% to 40%, thereby benefiting the middle (<63 mm) and fine (<30 mm) fractions. Specifically, the front-mounted flail mower equipped with hammer flails was the most effective, reducing the coarse fraction larger than 63 mm by 40%, with corresponding increases in the medium (>30 mm; +15%) and fine fractions (<30 mm; +25%). Kirchmeier and Demmel [47] were among the first to attempt combining harvesting and corn residue shredding by mounting flail mowers on the combine harvester and testing it in comparison to a tractor-driven flail mower. Results of the sieve analyses from two experimental sites over three years showed a wide range. Based on the total weight of the samples, the proportion of plant material larger than 45 mm decreased by 11 to 55 percentage points after flail mowing compared to before flail mowing. Similarly, the proportion of plant material larger than 200 mm was reduced by 18 to 37 percentage points due to flail mowing. Uppenkamp et al. [43] also found that flail mowers with hammer flails achieved the highest shredding intensity of corn stover, with 81.9% to 88.4% of the Corn-Cob-Mix residue being smaller than 45 mm, followed by the plate blades with 84.2% to 85.0%. The tested rotary mower achieved 79.2% to 81.7%, depending on the configuration, while the flail mower with Y-blades only managed to cut 59.6% of the corn stover shorter than 45 mm in this study.

This raises the question of whether the new tool design of the HS3 corn header, with flail knives attached to its horizontal choppers, not only achieves its desired effect on the corn stubble but also enhances the shredding intensity of the corn stover due to the dulled flails, compared to standard knives. To address this question, field trials were conducted over a period of three years at six different farm fields across Germany. The main focus was on investigating two configurations of the HS3: the first equipped with standard knives and the second with the new flail knives. Additionally, a third method, commonly used in Germany, which involved harvesting with the corn header equipped with standard knives followed by a pass with a tractor-driven flail mower, was tested as a benchmark at the same time. The first objective was to compare these methods based on the particle size distribution of the processed corn stover. Therefore, a total of 315 corn stover samples were separated by size into six fractions using a cascade sieve system with sieve opening diameters of 67, 30, 16, 8, 4, and 2 mm. The retained mass per sieve deck was determined by weighing. To describe the cumulative particle size distribution of the corn stover samples, Rosin-Rammler-Sperling-Bennet (RRSB) distributions were applied. Based on these distributions, the mean particle size, median particle size, 10th percentile, 90th percentile, and the interpercentile range were calculated for each corn stover sample. To compare the treatments, two-sided simultaneous 95% confidence intervals were calculated using statistical mixed models. The second objective was to compare the corn stover shredding methods based on the weight, number, and structural integrity of the corn stalk segments within the processed corn stover. For this purpose, stalk segments were manually selected from the sieve residues of the 67 and 30 mm sieves, rated using a five-level assessment scale, and weighed and counted for each level separately.

2. Materials and Methods

2.1. Experimental Approach

The field trials involved the utilization of the Horizon Star* III (HS3) corn header which has already been described by Ramm et al. [45]. This corn header encompasses eight rows with a row spacing of 75 cm. The HS3 is based on the three-roller Rota Disc^® technology (Carl Geringhoff GmbH & Co. KG, Ahlen, Germany). This process involves the corn stalks being pulled through counter-rotating disc rotors equipped with 15 cutting discs. These cutting discs engage with the two slotted snapping rolls, shredding the corn plant in a nearly diagonal cutting pattern. Each row unit of the HS3 is equipped with a horizontal chopper positioned in the front area on the right-hand side of the row unit. The horizontal choppers are designed to hold two cutting tools. The cutting tools feature sharp cutting edges and are additionally equipped with dulled flails at their ends. As the horizontal blade executes the cutting action, these flails strike the corn stubble at an angle, causing it to be shredded down to the root (Figure 1 and Figure A1). The nominal rotational speed is 2630 rpm and the outer diameter of the horizontal choppers is 575 mm.

The HS3 has been tested in two different configurations regarding the tools of the horizontal choppers. In addition to the flail knives, the corn header was tested using standard knives (simple straight knives; Figure A2) to imitate the previous model, Horizon Star* II (HS2). In the configuration with flail knives, the corn header was operated at the lowest possible cutting height setting to achieve the desired effect of stubble shredding. In the configuration with standard knives, a common cutting height of approximately 15–25 cm was set. During the harvest, a constant speed of 6 km/h was maintained.

As a third additional experimental variant, the corn stover was processed using a commonly practiced method. Following the harvest of the corn using the HS3 equipped with standard knives on its horizontal choppers, tractor-driven flail mowers (provided by the test farms) were additionally employed to break down the corn residue. This method, particularly common in Germany, was also tested at the same time as a benchmark. Thus, three different methods of corn stover processing were tested: flail knives, standard knives, and the practice method (

Table 1. Overview of corn stover shredding methods evaluated in this research.

Corn Stover Shredding Method	First Pass: Harvest (Configuration of the HS3 Corn Header)		Second Pass: Post-Harvest Residue Management
	Cutting Tools	Cutting Height Setting
Flail Knives	Flail Knives	Lowest possible for stubble shredding	None
Standard Knives	Standard Knives	15–25 cm corn stubble height	None
Practice Method	Standard Knives	15–25 cm corn stubble height	Tractor-driven Flail Mowers

).

To ensure the tested technology operated effectively under the typical conditions of Germany’s various grain corn cultivation regions (weather conditions, soil types, harvest dates, etc.), on-farm experiments were carried out at six different experimental sites nationwide from 2018 to 2020. The trials extended from Baden-Württemberg in the south through North Rhine-Westphalia and Lower Saxony to Schleswig-Holstein in the north. The field trials were set up as randomized complete block designs. The width of the plots corresponded to the working width of the corn header (6 m). Due to site-specific conditions such as field size, spacing of tramlines, and topography, it was necessary to adjust the number of blocks, replications, and the length of the plots for the individual test sites (

Table 2. Test sites.

Year	Site	Flail Mower	Plot Length [m]	Blocks (Replications)	Samples per Plot	Total Samples
2018	Zeutern (49.1786, 8.6587)	Sauerburger (Y-Blades)	125	3	6	54
2018	Steinheim (51.8460, 9.1316)	Müthing (Hammer flails)	125	3	6	54
2019	Stettfeld (49.1858, 8.6343)	Maschio (Hammer flails)	75	6	3	54
2019	Bückeburg (52.2676, 9.0851)	Müthing (Hammer flails)	75	6	3	54
2020	Kraichtal (49.1442, 8.7332)	Sauerburger (Y-Blades)	75	6	3	54
2020	Timmaspe (54.1348, 9.8916)	Sauerburger (Hammer flails)	75	5	3	45

). The experimental plans are displayed in the Appendix A (Figure A3, Figure A4, Figure A5, Figure A6, Figure A7 and Figure A8). In total, 315 corn stover samples were collected.

The plot boundaries were marked using paint spray and surveyed prior to harvesting using the SST FieldRover II 10.4 (SST Development Group Inc., Stillwater, OK, USA) surveying software, which stores GPS positioning data transmitted by the RTK-GPS rover (AgGPS 442; Trimble Inc., Sunnyvale, CA, USA) to the laptop.

2.2. Sample Collection and Preparation

The corn residue left in the field after harvest consists of stalks, cobs, husks, leaves, and tassels (referred to collectively as corn stover) along with corn stubble, which is the portion of the stalks that remains rooted in the soil after harvest.

For sample collection, an aluminum frame measuring 100 cm × 75 cm was placed longitudinally to the sowing direction on top of the corn residue (Figure 2). This allowed for the capture of an area corresponding to one row of corn over a length of 1 m (with 75 cm row spacing). The sample collection was conducted outside the paths of both the tire tracks of the combine harvester and the tractor equipped with a flail mower. The corn stover within the aluminum frame was carefully collected using a small rake and packed into perforated bags. The corn stubble was left behind. To establish storage stability, the corn stover samples were dried for a period of 12 to 14 days using a hot air blower in a drying chamber.

2.3. Sieve Analysis

The sieve analysis of corn stover was conducted based on the methodology developed by the Thünen Institute of Agricultural Technology (Braunschweig, Germany) [36]. The corn stover was divided into six fractions (sieve opening diameters: 67/30/16/8/4/2 mm). For this purpose, a six-stage sieve system was utilized (Figure 3). The 2 mm sieve deck was used to separate possible adhesions of soil particles; thus, the passage through the 2 mm sieve was discarded. The sieve decks were equipped with round-hole perforated metal sheets. The sieve surface area per sieve deck was approximately 2200 cm² with an inclination of approximately 5°. The sieves were horizontally moved at a rate of 250 rpm by an eccentric drive with a 50 mm stroke. The sieving process was continued until the sieve system cleaned out. The retained mass per sieve deck was determined by weighing.

The cumulative relative mass retained ($R_m$) on the sieve $i$ with a sieve opening diameter $x_i$ (mm) was calculated using Equation (1),

$$R_m\left(x_i\right)=\frac{{{\displaystyle \sum }}_{z=1}^i{M}_z}{M}$$

(1)

where $M_z$ (g) is the mass retained on the $z$ sieve and $M$ (g) is the total mass of the corn stover sample. To describe the cumulative particle size distribution of the corn stover samples, Rosin-Rammler-Sperling-Bennet (RRSB) distributions were applied,

$$R\left(x\right)=exp\left(-{\left(\frac{x}{x\prime }\right)}^n\right)$$

(2)

where $R\left(x\right)$ is the cumulative relative weight of particles larger than $x$ (mm), $x^{\prime }$ is the position coefficient and $n$ the width coefficient. Taking the logarithm twice results in a linear relationship between $ln\left(ln\left(\frac{1}{R\left(x\right)}\right)\right)$ and $ln\left(x\right)$.

$$ln\left(ln\left(\frac{1}{R\left(x\right)}\right)\right)=n ln\left(x\right)-n ln\left(x^{\prime }\right)$$

(3)

Based on Equation (3) the coefficients of the RRSB distributions for each corn stover sample were estimated performing linear regression. Based on the particle size distributions, the following parameters were calculated for each corn stover sample, where Γ is the gamma function.

$$\mathrm{Mean} \mathrm{particle} \mathrm{size} \overline{x}=x^{\prime }\mathsf{\Gamma }\left(1+\frac{1}{n}\right)$$

(4)

$$\mathrm{Median} \mathrm{particle} \mathrm{size} x_{50}=x^{\prime } ln{\left(2\right)}^{\frac{1}{n}}$$

(5)

$$10\mathrm{th} \mathrm{percentile} x_{10}=exp\left(\frac{ln\left(-ln\left(0.9\right)\right)+n ln\left(x^{\prime }\right)}{n}\right)$$

(6)

$$90\mathrm{th} \mathrm{percentile} x_{90}=exp\left(\frac{ln\left(-ln\left(0.1\right)\right)+n ln\left(x^{\prime }\right)}{n}\right)$$

(7)

$$\mathrm{Interpercentile} \mathrm{range} IPR=x_{90}-x_{10}$$

(8)

2.4. Scoring of the Structural Integrity of Corn Stalk Segments

Following the sieve analysis, the structural integrity of corn stalk segments contained within the corn stover was assessed. The evaluation is based on the scoring system presented in

Table 3. Scoring system for evaluating the structural integrity of corn stalk segments.

Scoring Level	Definition
0	stalk segment is completely frayed or sidewall at least 50% opened
1	between 30% and 50% of the stalk segments’ sidewall is opened
2	less than 30% of the stalk segments’ sidewall is opened, yet clearly damaged
3	frontal section of the stalk opened, nodes missing, sidewall mostly undamaged
4	stalk segment is intact, both the frontal section and sidewall are undamaged

, which was inspired by the methodology introduced by Brunotte and Voßhenrich [35].

Plant material that has passed through the 30 mm sieve is assumed to be fragmented to an extent where it cannot offer sufficient protection for ECB’s overwintering. This assumption is based on the random inspection of plant material from individual samples that passed through the 30 mm sieve. During this inspection, it was found that no stalk segments could potentially provide a closed space for the overwintering of ECB larvae, due to the degree of fragmentation (Figure 4, bottom row). As such, it did not require further assessment and was directly assigned to scoring level 0. Therefore, the evaluation of structural integrity focused specifically on the stalk segments found in the sieve residues of the 30 mm and 67 mm sieves (Figure 4, top row). Moreover, the assessment explicitly pertained to stalk segments; other plant components, like leaves, were also assigned to scoring level 0. The same applied to stalk segments shorter than 30 mm that may still have been present in these sieve residues.

From the sieve residues of both the 30 mm and 67 mm sieves, all stalk segments not corresponding to scoring level 0 and with a minimum length of 30 mm and a diameter greater than 5 mm were manually selected and allocated to the respective scoring levels. The mass of these stalk segments per sieve fraction was determined by weighing. These threshold values were chosen based on three factors. The first was the average length of the larvae, which is approximately 25 mm [33,34]. Secondly, Losey et al. [48] found that larvae demonstrate a notably low survival rate and a high tendency to abandon stems with diameters smaller than 5 mm. Thirdly, it has been observed that the majority of the larvae are located in the thicker, lower sections of the stalks by the time of harvest [23,30].

In the next assessment step, incompletely destroyed stalk segments (scoring level > 0) from each sieve fraction and scoring level were further divided into two categories based on their diameter, either larger or smaller than 10 mm, and their count was noted. Given that ECB larvae consume the plant moving towards the root base over the growing season, the risk of larvae inhabiting the upper stalk sections with small diameters at harvest time is reduced. This subdivision also accounts for the spatial requirements of the larvae.

Figure 5 shows the flow chart for the assessment of the structural integrity of corn stalk segments within the sieve residues.

2.5. Statistical Analysis

The statistical software R 4.3.2 [49] was used to analyze and evaluate the measures of central tendency and dispersion of the particle size distributions (Equations (4)–(8)), the total weight of the corn stover samples and the percentage mass of incompletely destroyed stalk segments and the number of incompletely destroyed stalk segments by diameter. The data evaluation started with the definition of appropriate statistical mixed models for the target variables:

• Mean particle size
• Median
• 10th percentile
• 90th percentile
• Interpercentile range
• Total sample weight
• Percentage mass of incompletely destroyed stalk segments
• Number of incompletely destroyed stalk segments (diameter > 5–10 mm)
• Number of incompletely destroyed stalk segments (diameter > 10 mm) [50,51].

The models included treatment (flail knives, standard knives, practice method) as a fixed factor. The experimental design (randomized complete block design) was taken into account by corresponding nested random effects (site, block, plot). The residuals were assumed to be (approximately) normally distributed and to be homoscedastic (or heteroscedastic where necessary). These assumptions are based on a graphical residual analysis. In order to compare the multiple treatments, two-sided simultaneous 95% confidence intervals were calculated.

3. Results

3.1. Sieve Analysis

The corn stover samples collected during the field trials exhibited no statistically significant differences in sample mass among the different treatments (

Table A1. Two-sided 95% confidence intervals for comparison of the different levels of treatment.

Parameter	Contrast	Estimate	Lower Bound	Upper Bound
Total sample weight (g)	standard knives − flail knives	7.54	−37.53	52.62
	practice method − flail knives	−30.36	−75.44	14.72
	practice method − standard knives	−37.91	−82.98	7.17
Percentage mass of incompletely destroyed stalk segments	standard knives − flail knives	1.00	0.17	1.84
	practice method − flail knives	−2.65	−3.37	−1.93
	practice method − standard knives	−3.65	−4.38	−2.92
Number of incompletely destroyed stalk segments (diameter > 5–10 mm)	standard knives − flail knives	0.31	−0.73	1.36
	practice method − flail knives	−1.63	−2.54	−0.72
	practice method − standard knives	−1.94	−2.89	−0.99
Number of incompletely destroyed stalk segments (diameter > 10 mm)	standard knives − flail knives	2.09	0.53	3.65
	practice method − flail knives	−4.27	−5.61	−2.93
	practice method − standard knives	−6.36	−7.76	−4.97

). The average weight of the dried samples was 666.5 g for the corn header configuration with flail knives, 665.7 g for the configuration with standard knives, and 640.9 g for the practice method (Figure 6).

The box plots in Figure 7 depict the relative sample mass retained per sieve for the different treatments. It is evident that the use of flail knives resulted in more intense shredding of corn stover compared to standard knives. On average, there was a 5.4% reduction in the proportion of coarse plant material, defined as material >30 mm (retained by both the 67 mm and 30 mm sieves combined). As expected, the use of tractor-driven flail mowers following prior harvesting with the corn header in the standard knives configuration (practice method) led to a greater reduction in coarse plant material >30 mm than changing from standard knives to flail knives. On average, the additional processing of the corn stover using a flail mower resulted in a reduction in the proportion of coarse plant material >30 mm by 23.4%. Accordingly, the difference between the flail knives configuration and the practice method amounted to 18.0%.

The described differences between the treatments were accordingly reflected in the measures of central tendency and dispersion of the RRSB distributions (

Table 4. Model estimated means for the measures of central tendency and dispersion of the RRSB distributions by treatment (mm).

	Flail Knives	Standard Knives	Practice Method
$\overline{x}$	31.90	35.50	23.74
$x_{50}$	29.98	33.60	22.33
$x_{10}$	11.82	13.73	8.90
$x_{90}$	54.47	59.67	40.44
$IPR$	42.65	45.91	31.58

). The mean and median particle sizes, as well as the 10th and 90th percentiles, exhibited the same trend. The dulled flails at the ends of the flail knives resulted in a slight decrease in particle size compared to the standard knives. However, as anticipated, the impact of flail mowing, represented by the disparity between the standard knife configuration and the practice method, was significantly more pronounced. The greater intensity in shredding corn stover shifted the particle size distribution towards smaller particles, consequently reducing the interpercentile ranges and indicating a more homogeneous plant material.

Table 5. Two-sided 95% confidence intervals for comparison of the different levels of treatment for the measures of central tendency and dispersion of the RRSB distributions (mm).

Parameter	Contrast	Estimate	Lower Bound	Upper Bound
$\overline{x}$	standard knives − flail knives	3.60	2.07	5.12
	practice method − flail knives	−8.17	−9.69	−6.64
	practice method − standard knives	−11.76	−13.29	−10.24
$x_{50}$	standard knives − flail knives	3.62	2.26	4.99
	practice method − flail knives	−7.65	−9.02	−6.28
	practice method − standard knives	−11.28	−12.64	−9.91
$x_{10}$	standard knives − flail knives	1.91	1.33	2.49
	practice method − flail knives	−2.92	−3.50	−2.34
	practice method − standard knives	−4.82	−5.41	−4.24
$x_{90}$	standard knives − flail knives	5.20	2.24	8.16
	practice method − flail knives	−14.03	−17.00	−11.07
	practice method − standard knives	−19.24	−22.20	−16.27
$IPR$	standard knives − flail knives	3.27	0.55	5.98
	practice method − flail knives	−11.07	−13.79	−8.35
	practice method − standard knives	−14.33	−17.05	−11.61

presents the two-sided 95% confidence intervals for the contrasts between different treatment levels (all pairwise comparisons). Furthermore, it provides the model-based estimated mean differences for the measures of central tendency and dispersion of the RRSB distributions. Importantly, none of the confidence intervals contain zero, indicating that all the mean differences are statistically significant.

The difference of the mean particle size was 3.6 mm between the standard knives configuration and the flail knives configuration. At least a 2.07 mm difference is to be expected, as indicated by the upper bound of the confidence interval. The difference between the flail knives configuration and the practice method was −8.17 mm. If flail mowing were omitted and the flail knives were used instead of the standard knives, in the worst-case scenario, an increase in the mean particle size of 9.69 mm (lower bound) can be expected, and in the best-case scenario, an increase of 6.64 mm (upper bound).

The difference in the 90th percentile of −19.24 mm between the standard knives configuration and the practice method, along with a −14.03 mm difference between the flail knives configuration and the practice method, suggests that flail mowing may significantly reduce the proportion of large, intact stalk segments potentially suitable for the ECB.

3.2. Rating of the Structural Integrity of Corn Stalk Segments

Figure 8 shows the percentage of incompletely destroyed stalk segments longer than 30 mm (Scoring Level > 0) in the sieve residues, relative to the total sample weight by treatment. With an average proportion of the total sample mass of 5.23%, 6.35%, and 2.50% for the flail knives, standard knives, and practice method treatments, respectively, the incompletely destroyed stalk segments constituted only a very small fraction of the overall sample mass. Nevertheless, the differences are statistically significant (Table A1), and flail mowing more than halved the mass of these stalk segments.

Despite the low percentage of incompletely destroyed stalk segments in the total sample mass, in the single-digit range, their number is considerable. As shown in Figure 9, the average sample contained 5.36, 5.70, and 3.73 incompletely destroyed stalk segments with diameters between >5 mm and 10 mm for the flail knives, standard knives, and practice method treatments, respectively. The analysis revealed that there was no significant difference in the number of incompletely destroyed stalk segments with diameters ranging from >5 mm to 10 mm when comparing the flail knives treatment to the standard knives treatment. However, statistically significant differences were observed when comparing both the standard knives treatment and the flail knives treatment to the practice method. Additionally, there were, on average, 10.37, 12.42, and 6.10 incompletely destroyed stalk segments with a diameter > 10 mm. In this diameter range, all differences between the treatments were found to be statistically significant (Table A1). Considering the sample area of 0.75 m² per sample, this equates to approximately 7.15, 7.60, and 4.97 incompletely destroyed stalk segments per square meter (i.e., 71,466, 76,000, and 49,733 per hectare) with a diameter between >5 mm and 10 mm, and 13.83, 16.56, and 8.13 incompletely destroyed stalk segments per square meter (i.e., 138,266, 165,600, and 81,333 per hectare) with a diameter > 10 mm, respectively.

Further categorization of these segments by scoring levels reveals that none of the treatments left a significant number of undamaged stalk segments corresponding to scoring level 4 behind, as evidenced by a median of 0 for all treatments and both stalk diameter ranges (Figure 10).

However, most intact stalk segments were found in the samples treated with standard knives. At least one intact stalk segment with a diameter > 10 mm was found in 15.5% of these samples. Including stalk segments with diameters between > 5 mm and 10 mm, this percentage increases to 21.4%. In contrast, samples processed by the newly developed flail knives contained intact stalk segments in only 16.2% of samples, with 9.5% of samples having at least one intact stalk segment with a diameter > 10 mm. In the practice method, these proportions were further halved to 8.6% and 4.8% of the 105 total samples.

When considering the number of intact stalk segments relative to the sampling area, it was observed that the corn header with standard knives configuration left an average of 0.32 intact stalk segments per square meter, equivalent to scoring level 4 with a diameter > 10 mm, which is less than one stalk segment per three square meters. In the flail knives configuration, the corn header left an average of 0.13 segments per square meter. With the practice method, this average was further reduced to just 0.06 intact stalk segments (scoring level 4; diameter > 10 mm) per square meter.

Regarding the other scoring levels, it was similarly observed that the practice method left behind the fewest number of stalk segments, followed by the flail knives corn header configuration, and finally the standard knives configuration.

4. Discussion

4.1. Discussion of the Experimental Approach

The experimental approach chosen for this study was feasible for implementation as on-farm experiments. The sampling of corn stover using an aluminum frame was time-consuming but practical. As described in Section 2, the corn stover samples were exclusively taken from areas that were not previously passed by the tires or tracks of the combine harvester or subsequently by the tractor with a flail mower. However, it is anticipated that the flail mowers will exhibit reduced shredding intensity on the corn stover within the wheel tracks compared to areas outside the tracks. This is particularly evident under moist soil conditions, which are commonly encountered during the harvest period of grain corn in Germany. The deeper wheel tracks and the cohesion between the soil and corn stover make it more challenging for the flail mower to pick up and process the crop residues. Consequently, obtaining representative samples within the wheel tracks becomes challenging as well. Depending on the working width of the corn header and the design (front or rear attachment) and working width of the flail mower, significant portions of the field may have been traversed by the vehicle’s tires or tracks before being processed by the flail mower. This fact must be considered when interpreting the results of this study. The effect of the flail mowers is likely to have been overestimated regarding the intensity of corn stover shredding for the whole plot or field.

A solid database was successfully generated regarding the experimental questions, allowing even relatively small differences in the particle size distribution of the different methods of corn stover shredding to be demonstrated as significant. However, for future investigations, consideration should be given to adjusting the gradations of the sieve hole diameters in order to further fractionate the coarse material (>30 mm), which regularly accounted for more than 50% of the plant material. Testing an equidistant gradation of sieve hole diameters (e.g., 90/70/50/30/10/2 mm) may be worthwhile. In this case, a fine sieve deck (2 mm) should be retained for the separation of mineral impurities. Nigon et al. [37] achieved good results by adding a pre-screening section, which sorts out long, tangled, bent, or otherwise irregularly shaped pieces of wheat residue that can lodge in the sieves and decrease the screening efficiency.

Drying the corn stover samples using a commercially available hot air blower proved to be a suitable approach to ensure storage stability until sieving. Considering the defined sample area of 0.75 m², determining the moisture content of the corn stover samples at harvest and subsequent drying in a lab-based forced air oven could enable the assessment of fresh and dry mass yield. These parameters could serve as additional influencing factors in the analysis, providing insights into how different corn stover shredding methods react to varying yields and moisture content of corn stover. To achieve this, moisture content determination at harvest would need to be performed immediately by weighing the samples in the field, and the drying of samples would have to be performed just in time for sieve analysis to prevent reabsorption of moisture by the plant material. However, due to the large volume of samples, this would require significant drying capacity. Determining these parameters based on subsamples, with the aim of reducing the volume to be dried, is problematic as it may be challenging to obtain small representative sub-samples given the broad particle size distribution. Another approach could involve sieving and weighing the corn stover directly after harvest without prior drying, followed by drying the individual sieve fractions to determine their dry mass. Considering the time required for sieving, this would likely be practical only if fewer samples per experimental year were to be investigated compared to the number of samples in this study (96–105 samples per year). The significant drying capacity would still be necessary in this case. Thus, the alternative would be to determine them at the field, block, or, at best, plot level, which would significantly limit subsequent data analysis.

The additional scoring for the structural integrity of corn stalk segments further enhanced the informative value of the studies. It demonstrated that even with a significant difference in the degree of size reduction, as seen between the standard knives and practice method treatments, there does not necessarily have to be a higher number of intact stalk segments in the grain corn residue. However, it is important to consider the substantial time and effort involved.

4.2. Methods of Corn Stover Shredding

The HS3 was designed with the primary aim of shredding corn stubble during the harvesting pass to minimize processing quality issues caused by the stubble being run over by the combine harvester or other vehicles. Sieving analyses and assessments of the structural integrity of corn stalk segments in this study have, however, revealed that the newly developed flail knives of the HS3 also contribute to further shredding of the corn stover. This is evidenced by a mean particle size 3.6 mm smaller than that achieved with standard knives, as well as a slight reduction in the mass and quantity of incompletely destroyed corn stalk segments with scoring levels greater than zero. Nevertheless, in terms of particle size, the performance of the HS3 with flail knives has not reached the level of tractor-driven flail mowers following the harvest, as indicated by a greater mean particle size difference of 11.76 mm between the practiced method and the standard knives configuration. The mass of incompletely destroyed stalk segments was also only half of that measured for the corn header configuration with flail knives, and the quantity of incompletely destroyed stalk segments was also lower than for the flail knives treatment.

However, when categorizing the incompletely destroyed stalk segments by scoring levels, it became clear that the most concerning level 4—intact corn stalk segments with nodes still attached at the ends and intact sidewalls—were only found in a relatively small amount. This suggests that the HS3 and its three-roller row units, even when equipped with standard knives, effectively slice open the stalks. Moreover, a significant portion of these stalk segments had a diameter of less than 10 mm. These segments are unlikely to host ECB larvae due to the larvae’s space requirements, considering that these smaller diameter stalks belong to the upper portion of the corn plant, while the larvae tend to be located in the lower parts of the stalks at harvest time [48,52].

Regarding the remaining scoring levels 1–3, it was also observed that the HS3 equipped with flail knives can reduce the number of corresponding corn stalks compared to the standard knives configuration, but not to the level of the tested tractor-driven flail mowers. However, these stalk segments are at least damaged enough to allow water ingress, making them less likely to provide a suitable overwintering habitat for ECB larvae. To date, no distinct research has been conducted that analyzes the correlation between the level of destruction of corn stalk segments and the winter mortality rate of ECB larvae in detail. Additionally, it is important to note that the mass of completely destroyed stalk segments represented only a single-digit percentage of the total mass of the corn stover samples, although stalk segments constitute 42% to 56% of the total dry mass of corn stover [25,26,27,28,29].

When considering the differences between the two-step practice method—harvesting with the corn header equipped with standard knives followed by a pass with a tractor-driven flail mower—and the single-step methods, that is, simply processing the residue with the HS3 equipped with either standard or flail knives during the harvesting pass, three additional factors must be considered.

Firstly, prior research exploring the application of tractor-driven flail mowers to grain corn and Corn-Cob-Mix crop residues has shown a broad spectrum in shredding intensity. This variation is largely dependent on factors such as the flail mower model, the design of the cutting tools, and differences across experimental years and sites [31,43,47]. In general, the results derived from the practice method treatment in this study align with these previously observed ranges. However, it is important to note that in our study, a variety of tractor-driven flail mower models equipped with diverse types of cutting tools were tested. This approach served as a benchmark, encompassing a substantial portion of the available models and tool sets (Table 2). Nevertheless, due to the experimental approach employed, the specific effects attributed to the different types of tool sets for tractor-driven flail mowers could not be isolated and analyzed separately. Despite this limitation, a consistent pattern was evident across all experimental sites. Independent of the tractor-driven flail mower type or the type of tool set used, the corn headers (whether equipped with flail knives or standard knives) were consistently unable to achieve a higher shredding intensity compared to the tractor-driven flail mowers, as indicated by the particle size distributions. Notably, the variability in the shredding intensity of the tractor-driven flail mowers tested in this study was not particularly large, as indicated by the boxplots of ‘Percentage sample mass retained per sieve’ (Figure 7). This consistency was observed despite the utilization of four distinct flail mowers across three years at six different experimental sites. On the other hand, the corn header variants, both standard and flail knives, seemed to exhibit a greater tendency towards variability.

Secondly, as illustrated in Section 4.1, the shredding intensity by flail mowers may be affected by the compaction of residues from being run over by the combine harvester, probably causing an overestimation of the overall shredding intensity. A practical method is required to collect representative samples of shredded corn stover from tire tracks, despite their compaction into the soil, to be able to investigate this effect.

Thirdly, the HS3 is outfitted with three-roller row units, which, according to studies by Handler et al. [39], already achieve very intensive shredding of the corn stover in terms of both particle size and structural integrity of the stalks. This confirms the observations from this study that the three-roller row units intensively shred the corn stover, posing a question about the comparative performance of flail mowers versus the HS3 equipped with flail knives, especially when used on residues initially processed by corn headers with conventional two-roller row units. These two-roller units, much more prevalent than three-roller systems and available in various designs regarding the cutting edges of the snapping rollers, establish a different baseline for the subsequent processing by flail mowers. Given the current state of technology, the practice method tested in this study, which combines harvesting with a corn header equipped with three-roller row units followed by processing the crop residues with a tractor-driven flail mower, likely represents the most intensive method in terms of shredding intensity available today.

5. Conclusions

This study focused on comparing the newly developed flail knives for the horizontal choppers of the Horizon Star III (HS3) against standard knives and also compared them against a two-step practice method. Key findings include:

• Shredding intensity improvement: The HS3 with flail knives reduced the coarse plant material (>30 mm sieve residues) by 5.4% and achieved a 3.6 mm smaller mean particle size compared to standard knives.
• Benchmark comparison: Against the two-step practice method, the HS3’s flail knives showed lower shredding intensity. The tractor-driven flail mowers further reduced the mean particle size of corn stover by 11.28 mm and 7.65 mm compared to the standard and flail knives treatments, respectively. In terms of coarse material reduction, it achieved 23.4% and 18.0% less coarse plant material compared to the standard and flail knives treatments, respectively.
• Assessment of incompletely destroyed stalk segments: The flail knives, standard knives, and practical method treatments left an average of 5.36, 5.70, and 3.73 stalk segments (diameter > 5 mm and ≤10 mm), and 10.37, 12.42, and 6.10 stalk segments (diameter > 10 mm) per corn stover sample, respectively. Notably, no treatment resulted in a significant number of undamaged stalk segments (scoring level 4), with a median of 0 across all treatments and diameter ranges.

The study underscores the HS3’s improved effectiveness with flail knives over standard knives, marking a notable step forward in corn stover shredding during the harvesting pass. The additional impacts of the dulled flails at the knife tips led to further shredding of the corn stover and were the cause for the enhancement, yet the study also indicates the need for further research. Future studies should focus on the interactions between corn header and tractor-driven flail mower models, the impact of pre-mowing residue compression into the soil, and factors like moisture content on shredding intensity. Additionally, the effectiveness of the HS3 in shredding corn stubble requires further investigation.