The management of field treatments, tillage methods, type of cash crop, and cover crop species used varied slightly across five sampled locations in MLRA 134, four of which were in eastern Arkansas, and one was in western Tennessee (Table 1). Each location was divided between a no-cover crop and the cover crop treatment. The primary, summer cash crops grown were cotton, soybean,

^†Cover crops were first established in the Fall of the listed year and were re-established every consecutive fall through 2019.

and corn. Planting and establishment of a cover crop occurred the fall of each year after commercial crops were harvested.

Site one was in St. Francis County near Haynes, Arkansas. This site was initially sampled in December 2019. Cotton was grown on beds each of the three summers on both the cover crop and no-cover crop treatments. Cereal rye was the cover crop for both Fall 2018 and 2019. The producer at Site one chose to practice a NT system in both the cover crop and no-cover crop treatment areas.

Site two was in Clay County near Piggott, Arkansas. This site was initially sampled in October 2019. Cash crops included corn and cotton, with corn being grown once in the no-cover crop treatment only with cotton grown on the cover crop treatment in Summer 2018. Cotton was grown on beds as the primary crop in each of the other growing seasons. Cereal rye was the cover crop for both Fall 2018 and 2019. Re-shaped beds were used on the cover crop and no-cover crop treated fields but, samples from the no-cover crop treatments were collected from a stale-seed bed prior to re-shaping.

Site three was at the Shelby County Agricultural Extension Center, near Germantown, Tennessee. This site was initially sampled in May 2020. Cotton was grown on beds each production season in both the treated and untreated fields. A cereal rye cover crop was used in 2020, but no cover crop was used in the treated field in 2018. Conventional tillage was used for the no-cover treatment.

Site four was in Cross County near Cherry Valley, Arkansas. This site was initially sampled in December 2019. Alternating production seasons and starting with corn on both fields, corn and soybeans were grown on beds. Cereal rye, black oats (Avena strigose), crimson clover (Trifolium incarnatum), and Austrian winter pea (Pisum sativum subsp. Arvense) were used as the cover crops both winter seasons. The producer at Site four chose to practice a NT system in both the cover crop and no-cover crop treatment areas.

Site five was in Green County near Paragould, Arkansas. This site was initially sampled in November 2019. Non-bedded soybeans were grown as the primary cash crop during the three growing seasons in both treatments. A combination of cover crops was used at this site and included cereal rye, black oats, and crimson clover. The producer at Site five chose to practice a conventional tillage system in the no-cover crop treatment area and a no-tillage system in the cover crop treatment.

The experimental design for this study was a completely random design consisting of a single field treatment, with or without cover crops, imposed in five fields at five locations within MLRA 134. The cover crop/no-cover crop treatments were established in either two halves of the same field or in two adjacent fields within the same soil map unit. For certain soil properties, only treatment was formally assessed, while, for other soil properties, soil depth and/or aggregate size class were also formally assessed.

Bulk density samples were collected from five random locations, on top of the bed in bedded fields, in each treatment at each field location using a 4.7-cm-diameter core chamber and slide hammer from soil depths of 0 to 10 and 10 to 20 cm. Samples were oven-dried at 70˚C for 48 hours and weighed for bulk density determination [16].

For texture analysis, Mehlich-3 extractable nutrients, SOM, and soil pH, 25 to 30 individual soil cores were collected throughout the treatment areas to a depth of 15 cm with a 2.5-cm-diameter push probe (from the tops of the beds in bedded fields) and combined into a single composite sample for each treatment area at each location. When field treatment size was larger than 8.1 ha (20 acres), the area was divided into two separate composite samples to ensure that no composite sample encompassed more than 8.1 ha. Soil samples were oven-dried at 70˚C for 48 hours and then crushed to pass through a 2-mm sieve to remove coarse fragments and/or coarse roots. Soil particle-size analyses were conducted using a modified 12-hr hydrometer method to determine sand, silt, and clay concentrations [17]. Using the oven-dried, sieved soil, soil pH and Mehlich-3 extractable Ca²⁺, Mg²⁺, and Na⁺ and SOM concentrations were also measured. Soil pH was potentiometrically measured using an electrode in a 1:2 (m/v) soil-to-water mixture. Plant-available soil nutrient concentrations (i.e., Ca, Mg. and Na) were determined after extraction using the Mehlich-3 extractant in a 1:10 (m/v) soil-to-solution mixture [18] and measured by inductively coupled plasma-atomic emission spectrometry (ICP-AES; CIROS CCD model; Spectro Analytical Instruments, MA) [19]. Soil organic matter concentrations were measured by weight-loss-on-ignition, where a muffle furnace was used for 2 hours at 360˚C [20]. The soil did not effervesce when exposed to dilute hydrochloric acid, thus all soil C in the SOM was assumed to be organic C. Measured SOM and Ca, Mg, and Na concentrations were converted to contents using the measured bulk densities and 10-cm sample depth and were reported as either kg or Mg ha⁻¹. Calculations were performed assuming that bulk density was constant from the 0 - 10 and 0 - 15 cm soil depths.

Three random samples were also collected from the top of the bed (in bedded fields) with a 7.4-cm-diameter core chamber and slide hammer from the 0- to 10-cm depth. The core was removed from the chamber and split into 0- to 5- and 5- to 10-cm sections for water-stable aggregate determinations [10].

Following previous procedures [10] [21] [22] [23], individual soil samples for aggregate stability were manually broken apart into smaller pieces, sieved moist through a 6-mm mesh screen and left to air dry at ~21.1˚C for 7 days. After air-drying, 150 g of soil from a sample were weighed and placed on top of the nest of sieves in the wet-sieve apparatus. The nest of sieves contained the following sieve sizes in decreasing order: 4-, 2-, 1-, 0.5-, and 0.25-mm. The nest of sieves was attached to an arm that mechanically oscillated the nest of sieves with the soil sample in a 40-cm-diameter by ~ 120-cm-tall column of tap water at 30 cycles per minute for five minutes. After the mechanically imposed disturbance, the nest of sieves was removed and separated. The soil aggregates that had been retained on each sieve were manually transferred into a pre-weighed, aluminum container with a wash bottle. Samples were left to settle for approximately 10 minutes, excess water was slowly decanted, making sure no soil aggregates or sediment were discarded. Samples were then placed into a forced-draft oven to oven-dry at 70˚C for 24 hours. After oven-drying, soil samples were weighed to determine the water-stable aggregate fraction by aggregate size class (i.e., > 4-, 2- to 4-, 1- to 2-, 0.5- to 1-, and 0.25- to 0.5-mm sizes. Visibly obvious coarse fragments were picked out manually from the largest two size classes, weighed, and the coarse fragment mass was subtracted from the oven-dry soil mass. In addition, total water-stabile aggregates were calculated by summing the mass of soil aggregates retained on all five sieves and dividing by the original 150 g of air-dried soil. The three replications of each soil treatment sample were conducted one after another before the non-aggregated soil that passed through the 0.25-mm sieve was removed from the bottom of the wet-sieve apparatus. The wet-sieve apparatus was filled with fresh water to process the three replications of the next treatment sample.

A one-factor analysis of variance (ANOVA) was conducted using the PROC GLIMMIX procedure in SAS (version 9.4, SAS Institute, Inc., Cary, NC) to evaluate the effect of treatment (cover crop and no-cover crop) on sand, silt, and clay, pH, and extractable soil Ca, Mg, and Na and SOM concentrations and contents. Soil pH and extractable soil Ca, Mg, and Na, and SOM concentrations and contents were analyzed using a gamma distribution, while sand, silt, and clay were analyzed using a beta distribution. A two-factor ANOVA was conducted to assess the effect of treatment, soil depth (0 to 5 and 5 to 10 cm), and their interaction on the total water-stable aggregate concentration using a beta distribution. Soil bulk density was analyzed using a two-factor ANOVA to evaluate the effect of soil depth (0- to 10- and 10- to 20-cm depths), treatment, and their interaction using a gamma distribution. A three-factor ANOVA was conducted to evaluate the effects of treatment, aggregate size class (>4, 2 to 4, 1 to 2, 0.5 to 1, and 0.25 to 0.5 mm), soil depth, and their interactions on water-stable aggregate concentrations using a beta distribution. Significance was judged at P < 0.05. When appropriate, means were separated by least significant differences (P < 0.05). All statistical analyses were conducted using SAS version 9.4.

Though all five sites evaluated in this study were from within MLRA 134, initial soil properties in the top 15 cm varied among sites. Sand ranged from 0.09 to 0.26 g·g⁻¹, silt ranged from 0.62 to 0.82 g·g⁻¹, and clay ranged from 0.07 to 0.16 g·g⁻¹ among all individual sample replicated across all five sites (Table 2). Even with a range of sand, silt, and clay, the texture of all five sites was a silt loam. Soil pH ranged from 5.9 to 7.2 (Table 2). This range of pH is ideal for most commercial crops and eliminates concern regarding pH-limiting conditions for plants within each of the five sites [24]. Extractable soil Ca ranged from 688 to 1479 mg·kg⁻¹, extractable soil Mg ranged from 86.0 to 309 mg·kg⁻¹, and extractable soil Na ranged from 7.0 to 29.0 mg·kg⁻¹ (Table 2). The nutrient concentration ranges among the studied sites are indicative of fertile soil that is in good condition for growing a variety of crops, including the cash crops and cover crops grown in this study [25].

Soil organic matter concentration ranged from 16 to 27 g·kg⁻¹ (Table 2). Characterizing initial soil properties among sites included in this study showed that most agronomically relevant properties were within a range that is adequate for proper crop growth and production and establishes a baseline condition to which future assessments could be compared to directly quantify and evaluate change-over-time results [25].

The soil property dataset generated and evaluated in this study represented a survey of agricultural sites using cover crops for various durations in MLRA 134 within the LMRV, where locations were intentionally selected to result in large variability. Therefore, any resulting significant differences between cover-crop and non-cover-crop treatments will be substantive despite large, inherent, initial, soil property variability.

All soil properties measured in the top 15 cm were unaffected (P > 0.05) by cover crop treatment (i.e., cover and no-cover; Table 3). Sand, silt, clay, pH, and SOM concentration averaged 0.16, 0.74, and 0.10 g·g⁻¹, 6.49, and 20.8 g·kg⁻¹, respectively (Table 3). Extractable soil Ca, Mg, and Na concentrations averaged 1043,

^†Means in a row with different letters are different at P < 0.05. *Measured soil concentrations were converted to contents using measured bulk densities.

184.5, and 14.2 mg·kg⁻¹, respectively (Table 3). Sand, silt, and clay are inherent soil properties that were not expected to change due to imposing a cover crop treatment.

The insignificant differences in pH and SOM concentrations between cover crop treatments (i.e., cover crop and no-cover crop) can be attributed to the warm, humid climate of the area and the short duration between treatment establishment and soil sample collection, which was < 24 months across all sites, except the Cross County site, which had treatments in place for approximately four years at the time of sampling. Over a longer period of consistent management, it is expected that the effects of a cover crop treatment would create significant differences compared to no-cover crop for near-surface soil pH and SOM concentration, as the benefits of cover cropping generally increase over time [26] [27].

Eastern Arkansas has a warm, humid climate with mild winters [14] that encourages a rapid rate of SOM decomposition [28], especially when the soil is tilled [16]. The climatic factor, in combination with the short duration of the presence of the cover crop treatment, are likely responsible for the lack of SOM concentration differences between cover crop treatments (Table 3). However, the lack of significant differences among initial soil properties across the study sites aids in the evaluation of other dynamic soil properties. Study sites with similar textures, pH, SOM, and extractable soil nutrient concentrations allow for more accurate assessments of vegetative treatments, such as a cover crop or no-cover crop, without concern for how different relatively static soil properties could affect other dynamic soil properties, such as soil bulk density and water-stable aggregate concentration.

Soil bulk density differed between cover crop treatments (P = 0.03) and differed between soil depths (P < 0.01) (Table 4). Averaged across soil depths, bulk density was greater with no-cover crop (1.27 g·cm⁻³) than with cover crops (1.24 g·cm⁻³) (Table 5). The decrease in bulk density in cover crop-treated fields may have been due to the addition of organic matter in the form of belowground root biomass, despite similar SOM contents across treatments, to increase porosity and reduce compaction and due to the aboveground biomass protecting the soil surface from rainfall impacts [29]. Averaged across cover crop treatment, bulk density was 1.1 times greater in the 10 - 20 cm depth (1.31 g·cm⁻³) than in the 0 - 10 cm depth (1.20 g·cm⁻³) (Table 5). As plant roots grow between soil aggregates, the roots function as a partitioner and separate the aggregates that have been compacted by farm equipment [30]. Soil bulk density naturally increases with depth due to the weight of overlying soil, but, in the case of this study, the change in bulk density was likely due to the influence of the CC treatment. The difference in depth affected is most likely due to the design of the root systems developed by the cover crops used. The wide, fibrous root systems of cereal rye,

^†Means within a treatment group with different letters are different at P < 0.05

oats, clover, and winter pea expand to a shallower depth compared to cover crops with a taproot system and, as a result, the upper 10 cm of the profile had a lower bulk density [31].

Using the measured soil bulk density data with measured concentration data on a sample-by-sample basis, extractable soil Ca, Mg, Na, and SOM contents were calculated. Like concentrations alone, extractable soil Ca, Mg, Na, and SOM contents were unaffected (P > 0.05) by cover crop treatment and averaged 1873, 332.3, and 25.4 kg ha⁻¹ and 37.3 Mg ha⁻¹, respectively (Table 3).

Water-stable aggregate (WSA) concentrations were unaffected (P > 0.05) by cover crop treatment or soil depth (0 - 5 and 5 - 10 cm) but differed (P < 0.01) among aggregate size classes. Water-stable aggregate concentration in the 0- 0.25-mm size class (0.138 g·g⁻¹) was largest and differed among that in the other four size classes (Figure 1). Water-stable aggregate concentration in the 0.50-1.0-mm size class (0.101 g·g⁻¹) was greater than that in the 1-2- (0.080 g·g⁻¹) and >4-mm size class (0.053 g·g⁻¹) but was like that in the 2 - 4-mm size class (Figure 1). Water-stable aggregate concentration in the 1 - 2-mm size class was like that in the 2 - 4-mm size class (0.085 g·g⁻¹) but was also greater than that in the >4 mm size class, which had the lowest WSA concentration among the five size classes (Figure 1). Water-stable aggregates being unaffected across both cover crop treatment and soil depth is likely explained by the short duration of this study (<24 months for Sites 1, 2, 3, and 5 and ~48 months for Site 4). An increasing proportion of small aggregates, as resulted in this study, may increase

aggregate transportability during runoff events to potentially increase soil erosion losses.

Like the results of the current study, Smith et al. [10] reported a general decrease in WSA concentrations with increasing size class in a silt-loam soil in eastern Arkansas after 15 years of consistent management in a wheat-soybean, double-crop system. It was predicted that, if the duration of cover crop treatment was longer, then the effects of cover crop treatment on WSA would be greater. Long-term residue and water management practices, like cover crop treatments, have been shown to aid in the increased soil aggregate stability and result in increased amounts of water-stable aggregates in larger size classes [10]. Allowing for continuous management with cover crop treatments for 10 years has shown to improve soil properties, including the percent WSA in the 2- to 4- and 1- to 2-mm size classes.

Summing the WSA concentrations across all size classes, and in contrast to soil bulk density, total (TWSA) concentration was unaffected (P > 0.05) by cover crop treatment and soil depth (Table 4). Total WSA concentration averaged 0.457 g·g⁻¹ across all cover crop treatments and soil depths (Table 5). Total WSA concentrations were most likely unaffected by cover crop treatment and soil depth because of the short duration of between establishment and soil sampling for this study. It was expected that within the 24 months that the two seasons of cover crops were implemented few significant changes would have occurred. Like WSA, TWSA would be expected to increase as cover crop treatments continue over time due to the increase in soil aggregate stability and the formation of new and larger soil aggregates. Cover crops aid in the formation of improved soil structure over time and it follows that, as structure improves, an increase in TWSA would occur as well [32] [33].

It is reasonable to expect that, as the duration of cover crop use increases, many of the listed benefits of cover crops that did not differ between cover-crop treatments in this study, such as SOM content and TWSA, would manifest themselves more compared to similar soils without a cover crop [13] [29]. This implies that the use of cover crop treatments as a means of improving soil properties is a longer-term commitment, and many of the benefits will not be realized in the short-term, as was characterized by this study. The climate in which cover crops are being used should also be taken into consideration, as the time for certain properties to be affected, such as SOM, may vary across climatic zones [16] [28].

Soils with a silt-loam surface texture, such as the ones specifically included in this study, are most vulnerable to soil erosion, particularly water erosion, and exist in much of the most productive agricultural areas in the US. Improving soil structure (i.e., decreasing soil bulk density), root matrix entanglement, and additions of organic matter to the soil from the use of cover crops can potentially contribute to reduced soil erosion and improved overall soil health. Not only should cover crops be used on vulnerable soils, but it is likely necessary to commit to long-term implementation of cover crops on vulnerable soils if the full benefits of cover crops are to occur.