Short Term Cotton Lint Yield Improvement with Cover Crop and No-Tillage Implementation

No-tillage has been used for mitigating wind erosion on the Southern High Plains US for decades. This study investigated the effects of tillage and nitrogen (N) fertilizer timing on cotton lint yield, fiber quality, and seed N content during a three-year transition from conventional tillage (CT) to a no-tillage system both with a wheat (Triticum aestivum) cover crop (NTW) and without a cover crop (NT). Lint yield was different between tillage systems within each year with the NTW system producing greater lint yield than the CT system in the second and third year of the transition period. The concentration of cotton seed N was not different within years, although it was decreased in the no N added control in the third year. Cotton fiber strength was increased in the NTW system compared to the CT system in the second year of the study. However, the CT system produced increased fiber strength compared to the other two systems in 2018 and is likely the result of late-season weather conditions. It was determined that implementing a NTW system may increase lint yield within the first few years and has no effect on most fiber quality parameters, especially in environmentally challenging conditions.


Introduction
The first tillage implements brought to the Southern High Plains (SHP; MLRA 77C) [1] to manage soil included the moldboard plow, which allowed crop production on the mixed grass prairies of the SHP and turned them into highly productive soils [2]. Intensive tillage of soil on the SHP combined with a period of intense drought, led to extreme cases of wind erosion in the period now known as the Dust Bowl [3]. Wind erosion has reportedly caused up to a 40% reduction in cotton (Gossypium hirsutum L.) lint yield [4], and under certain conditions, such as extremely high wind speeds and blowing soil, it can cause total crop failure. Government agencies have been established across the USA since the Dust Bowl to promote the use of less intensive forms of soil management and the use of cover crops to reduce the effects of wind erosion [5]. However, tillage continues to be used on the SHP and throughout Texas agriculture, with an estimated 60% of acres being managed using conventional tillage (CT) [6].
Cotton is the predominant crop grown on the SHP, with production reaching over three million bales in 2017 [7], with a standard US cotton bale weight of 217.7 kg, this amounts to over 650 million kg cotton lint produced on the SHP in 2017. The cotton produced on the SHP in 2017 accounted for about 15% of cotton production in the USA [8]. Many studies across the USA have examined the effects of soil and nutrient management on cotton yield and fiber quality. However, research involving soil and nutrient management practices and their impacts on cotton production is lacking on the SHP.

Site Descriptions
The study was conducted at the Texas A&M AgriLife Research and Extension Center in Lubbock, Lubbock County Texas (33.687 • , −101.827 • ). The 30-year average rainfall and temperature  for the study site were 486 mm and 15.9 • C, respectively. Maximum monthly temperature, average monthly temperature, and monthly cumulative precipitation is for the study years is presented in Figure 1. Inorganic N (NO 3 − and NH 4 + ) wet deposition was calculated to be 1.73 kg ha −1 in 2018 at this research site, based on wet deposition rates for that year collected at the Muleshoe National Wildlife Refuge Bailey [28]. The soil was an Acuff loam described as fine-loamy, mixed, superactive, thermic Aridic Paleustolls [29]. Soil characterization was conducted prior to the beginning of this study in 2016 [30] with an average pH of 7.4 and nitrate (NO 3 − -N) concentration determined to be less than

Experimental Design and Treatments
The study used a split-plot design with tillage system as the main factor and N fertilizer application timing (N treatment) as the split factor. Tillage systems included: no-till with a winter wheat (Triticum aestivum) cover crop (NTW), no-till winter fallow (NT) and conventional tillage winter fallow (CT) and were replicated 3 times. Nitrogen fertilizer application timings were arranged in a randomized complete block within each tillage. Nitrogen treatment included the following three levels: (1) no-added N (control); (2) 100% of N applied as a pre-plant application (PP); (3) 100% of N side-dressed applied (SD) at the cotton growth stage of pinhead square; (4) 40% of N applied PP and 60% SD applied (SPLIT); and, (5) 100% of N applied PP with a N stabilizer product (STB). Limus ® Nitrogen Management (N-butyl-thiophosphoric triamide and N-Propyl-thiophosphoric triamide, BASF Corporation, Florham Park, NJ, USA), a dual-action urease inhibitor, was used as the stabilizer product. Within each of the three replicates, tillage as main plots were randomly assigned to 4 rows (1 m spacing) and N fertilizer treatments were randomly assigned within each main plot. There was a total of 45, 4-row plots 15 m in length.

Field Management
All field management practices are summarized in Supplementary Table S1

Soil Characterization
Soil characterization was conducted prior to beginning the study in 2016 with composite samples collected from each replication of the tillage system. Samples were composited because no All N fertilizer was applied at a total rate of 168 kg ha −1 as urea ammonium nitrate (UAN, 32-0-0) via knife injection using a coulter fertilizer applicator. On 10 May 2016, 11 May 2017, and 19 May 2018 PP N treatments were applied and on 13 July 2016, 20 July 2017, 16 July 2018 SD N treatments were applied to respective plots. Fields were prepared by shredding stalks with a four-row John Deere shredder (Moline, IL, USA) in all tillage systems for both 2016 and 2017, and the CT plots were then disked to a depth of 5-8 cm with a four-row John Deere offset disk. Trifluralin [α,α,α-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine] was applied at 0.84 kg a.i. ha −1 and incorporated using a four-row spring tooth harrow to a depth of 5-8 cm in the CT system on 22 February 2016, 22 March 2017, and 24 April 2018. A bed lister was then used to re-form planting beds in the CT system. Tillage was conducted using a sweep cultivator in the CT system one time in each growing season at a depth of 5-8 cm. A rotary hoe was used to scratch all tillage systems about a week after planting in 2017, to a depth of about 1 cm, to encourage seed emergence and prevent soil crusting. Glyphosate was applied once in 2016 and 2017 for weed control, in addition to cover crop termination, at a rate of 2.7 kg a.i. ha

Soil Characterization
Soil characterization was conducted prior to beginning the study in 2016 with composite samples collected from each replication of the tillage system. Samples were composited because no N treatments had been implemented at the time of sampling. Samples were collected to a depth of 15 cm. The average soil pH across all tillage systems was 7.4 and did not differ with tillage system (Table 1). Organic carbon (OC) for this soil was 5.2 g kg −1 when averaged across tillage systems. Total N (TN) averaged 0.7 g kg −1 across tillage systems, and neither OC nor TN differed due to the tillage system. It was determined that the wheat cover crop in the NTW system reduced NO 3 − -N concentrations compared to the NT and CT systems without a cover crop, which supports previous reports of N use by cover crops (Table 1, [20]). No differences were determined for P, K, Ca, Mg, S, and Na among tillage systems.

Plant Nitrogen Analysis
Cotton seed was collected after ginning and acid delinted using sulfuric acid (66 deg. Baumé). In 2016, seeds were frozen at 20 • C before grinding. In 2017, seed was frozen with liquid N before grinding. Delinted seed was ground to pass a 2-mm sieve using an Analysenmühle mill (Analytical mill, IKA Works INC, Wilmington, NC, USA). Nitrogen concentrations were determined for seed via combustion analysis using an Elementar Vario Max CN (Elementar, Ronkonkoma, NY, USA) at the Texas A&M AgriLife Research and Extension Center in Vernon, TX [31][32][33] in 2016 and 2017. In 2018, cotton seed was collected from ginned cotton samples and analyzed via combustion analysis using an LECO elemental analyzer (LECO Corporation, St. Joseph, MI, USA) by Waters Agricultural Laboratories, Inc. (Camilla, GA, USA) for N concentration.

Nitrogen Use Efficiency
Agronomicefficiency (AE) of N was calculated as the increase in yield per unit of N applied to determine the best management strategy for timing of N fertilizer application. Agronomic N use efficiency was calculated for all years of the study (2016-2018) and was defined as: (Y-Y 0 )/F, where Y was lint yield with N fertilizer applied, Y 0 was the control yield with no N applied, and F was the rate of fertilizer application in kg ha −1 [34]. Agronomic use efficiency was calculated within year due to a significant year effect on cotton lint yield.

Statistical Approach
Data were analyzed using Proc GLIMMIX at a significance level of a < 0.10 using SAS software version 9.4 [35]. The GLIMMIX procedure is a generalized linear mixed model that can incorporate random effects, and it can also be used to fit statistical models to data with nonconstant variability, as well as where the response is not normally distributed [36]. Year effects were determined as the interaction of year with tillage system, N treatment, and both tillage system and N treatment with year treated as a random effect. Main-plot treatments (NTW, NT, CT) and split-plot treatments (control, PP, SD, SPLIT, STB) were treated as fixed effects. Replication was treated as a random effect for cotton lint yield, fiber quality, and stand count due to the lack of N treatment effects, and thus the statistical testing of only tillage differences. In addition, cover crop biomass, was analyzed with replication as the random effect due to only one tillage system being tested. Finally, seed N concentration and AE were tested using replication and replication by tillage as random effects for within year analysis. Fischer's protected LSD (a < 0.05) was used to separate means of significant impacts.

Cover Crop Biomass
On the SHP, wheat cover crop growth is heavily dependent upon winter precipitation, which was scarce in 2017 and 2018. Due to the lack of precipitation, and the late establishment of the cover crop in 2018, cover crop biomass was not collected in 2018. Wheat biomass data were previously published [30]. The aboveground biomass of the wheat cover crop in 2016 and 2017 was affected by the interaction of year and N treatment (p < 0.001) and was analyzed within each year of the study. Nitrogen treatment did not affect cover crop biomass in the first year of the study (p = 0.279). No treatments had been implemented prior to planting this cover crop, so no difference was expected. Mean cover crop biomass in 2017 was 3180 kg ha −1 . Nitrogen treatment did affect cover crop biomass in the second year of the study (p = 0.027), with the PP (1547 kg ha −1 ) SD (1795 kg ha −1 ) and SPLIT (1655 kg ha −1 ) treatments producing greater biomass than the control (822 kg ha −1 ). In addition, the SPLIT treatment produced greater biomass than the STB (1092 kg ha −1 ) treatment. When analyzed within year, there is no relationship between cover crop biomass and cotton lint yield (2016: p = 0.913; 2017: p = 0.53).

Fiber Quality
Fiber strength was the only fiber quality parameter affected by the interaction of year and tillage system (p < 0.001) and was analyzed within year. Tillage system affected fiber strength in 2017 (p = 0.011) and 2018 (p < 0.001) with the NTW system having greater fiber strength compared to the NT and CT system in 2017 (Figure 3). In 2018, the CT system produced cotton lint with greater fiber strength than the NTW and NT systems ( No split plot (N treatment), or interaction effects between the tillage system and N treatment were determined within any year of the study (Table 2). In 2016, the CT system produced approximately 300 kg ha −1 and 415 kg ha −1 more cotton lint than the NTW and NT systems, respectively (Figure 1). The 2016 growing season was the first season following NTW and NT implementation and cover crops, and it was determined that tillage system affected lint yield (Table 2), with the CT system producing greater cotton lint than the NTW and NT systems (Figure 2). Tillage also affected lint yield in both 2017 and 2018 (Table 2), with the NTW system producing greater lint yield compared to the CT system in both years (Figure 2). There was not a difference between the NT system compared to the NTW and CT systems in 2017 and 2018 ( Figure 2). Nitrogen treatment did not affect lint yield in any of the three study years (Table 2).

Fiber Quality
Fiber strength was the only fiber quality parameter affected by the interaction of year and tillage system (p < 0.001) and was analyzed within year. Tillage system affected fiber strength in 2017 (p = 0.011) and 2018 (p < 0.001) with the NTW system having greater fiber strength compared to the NT and CT system in 2017 (Figure 3). In 2018, the CT system produced cotton lint with greater fiber strength than the NTW and NT systems (  Due to the determination of a year effect on fiber strength additional fiber quality parameters were analyzed within the year including: micronaire, uniformity, and length. No differences between tillage systems were determined for any of the additional fiber quality parameters, with the exception of micronaire in 2017 (p = 0.002) where the NTW (3.2) and NT (3.1) systems produced cotton lint with more micronaire than the CT system (2.9).

Agronomic Use Efficiency
The agronomic use efficiency of N was analyzed within each year due to its dependence on yield as a significant factor which was affected by study year. No effect on AE was determined for tillage system, N treatment, or their interaction in the three years of the study (Table 3).   Due to the determination of a year effect on fiber strength additional fiber quality parameters were analyzed within the year including: micronaire, uniformity, and length. No differences between tillage systems were determined for any of the additional fiber quality parameters, with the exception of micronaire in 2017 (p = 0.002) where the NTW (3.2) and NT (3.1) systems produced cotton lint with more micronaire than the CT system (2.9).

Agronomic Use Efficiency
The agronomic use efficiency of N was analyzed within each year due to its dependence on yield as a significant factor which was affected by study year. No effect on AE was determined for tillage system, N treatment, or their interaction in the three years of the study (Table 3).

Plant Seed N
Seed N concentration was analyzed within year due to the significant effect of year on cotton yield. There were no seed N concentration differences in 2016 or (Table 4). In 2018, reduced N concentrations in the control compared to the rest of the N treatments were determined (Table 4, Figure 4).

Cover Crop Biomass
An increase in cover crop biomass was expected for the SD and SPLIT treatments compared to the PP and STB treatments, as they represent the mid-season application of N for this study, and thus are likely to have greater residual N for the cover crop to use after the cotton growing season. In addition, an increase in cover crop biomass was expected between added N treatments and the control, and was determined for the SD and SPLIT treatments. However, there was no difference in biomass between mid-season applied treatments, SD and SPLIT, and the PP treatment which indicates potentially similar NUE for the cotton crop regardless of N fertilizer application timing leaving low levels of N for the cover crop to use. The lack of a relationship between cover crop biomass and cotton lint yield in 2016 was expected, as the wheat cover crop was grown prior to any N treatments being implemented, and there was likely consistent plant available N across the entire study area. After one year of treatment implementation, there was still no correlation between wheat aboveground biomass and yield, which may point to a longer time period for the presence of wheat

Cover Crop Biomass
An increase in cover crop biomass was expected for the SD and SPLIT treatments compared to the PP and STB treatments, as they represent the mid-season application of N for this study, and thus are likely to have greater residual N for the cover crop to use after the cotton growing season. In addition, an increase in cover crop biomass was expected between added N treatments and the control, and was determined for the SD and SPLIT treatments. However, there was no difference in biomass between mid-season applied treatments, SD and SPLIT, and the PP treatment which indicates potentially similar NUE for the cotton crop regardless of N fertilizer application timing leaving low levels of N for the cover crop to use. The lack of a relationship between cover crop biomass and cotton lint yield in 2016 was expected, as the wheat cover crop was grown prior to any N treatments being implemented, and there was likely consistent plant available N across the entire study area. After one year of treatment implementation, there was still no correlation between wheat aboveground biomass and yield, which may point to a longer time period for the presence of wheat cover to impact lint yield. Any cover crop growth on the SHP should provide intrinsic benefits to the system as a major purpose for cover crop use is wind erosion mitigation through increasing the turbidity of wind at the soil surface, reducing the potential for wind damage.

Cotton Lint Yield
The reduced lint yield in 2017 and 2018 compared to 2016 was likely due to residual N use in 2016 which increased cotton lint yields compared to 2017 and 2018 and may have also been affected by environmental conditions (Figure 2). Environmental conditions in 2017 and 2018 included reduced precipitation and higher temperatures early in the growing season, which may have decreased cotton seedling emergence, thus reducing yield. When evaluated in 2018, it was determined that the no-till systems (NTW and NT) had more abundant plant stands compared to the CT system, likely indicating the inherent benefits of NT on the SHP as it relates to reducing wind erosion. However, NT without a cover crop is often not a viable system in semi-arid, high-wind areas such as the SHP, where producers often need to run implements, sand fighters, to form soil clods after a rain [5]. These clods increase the turbidity of the air at the soil surface and can reduce wind erosion. When using a strictly NT system without a cover crop, there may be the potential for more significant wind erosion and crop damage due to the erodibility of sandy soils without crop residue [5] and the lack of tillage or sand fighting in the NT system to increase turbidity.
The maintenance and improvement between NT systems with and without a cover crop compared to CT systems ( Figure 1) agrees with recent research on the semi-arid Texas Rolling Plains, where cotton lint yield was not affected by tillage practice during the transition from CT to conservation tillage system and increases in lint yield were determined once treatments had established [14,15]. Cotton lint yield in all three tillage systems in 2017 and 2018 were reduced compared to the tillage systems in 2016.
The increase for the NTW system compared to the CT system in the latter years may be due to early season protection of the cotton seedlings from harsh climatic conditions including high temperatures, decreased precipitation and wind erosion (Figure 1a-c). Temperatures reaching 39 • C in May and 44 • C in June and high winds throughout the early growing season in 2017 likely reduced cotton seedling emergence and vigor in both the CT and the NT plots (data not collected). The residue from the winter cover in the NTW system likely protected from blowing sand, allowing cotton seedlings to establish better compared to the CT and NT system. It is prudent to note the lack of difference in cotton lint yield between the NTW and NT systems in all three years of the study, although numerical increases were present for the NTW system ( Figure 1). The increase for the NTW and NT system compared to the CT system also indicates the inherent benefit of wind erosion mitigation of NT on the SHP. Wind erosion can cause severe crop damage and even crop failure on the SHP [5], while the mitigation of wind erosion by using a cover crop and NT maintains or slightly improves cotton lint yield soon after implementation, as determined in this study. Although not significant, there was a numerical increase in plant abundance for the NTW system compared to the NT system with an average increase of over 8000 plants ha −1 . This agrees with the numerical, but not significant, increase in cotton lint yield for the NTW system compared to the NT system in 2018 (Figure 1). Often, it is assumed that there will be a yield deficit in the first few years of cover crop and NT implementation particularly in dry climates, although the increase in lint yield for the NT and NTW systems compared to the CT system agrees with previous research with NT systems where no difference in cotton yield was determined across studies, especially in dry climates [37].
Often, yield benefits are not seen for several years or at all following implementation of conservational tillage systems. Lewis et al. [19] reported in a long term study in the SHP, that after 17 years a CT system without a cover crop had greater cotton lint yields than a NT system with a rye cover crop [19], which may point to variable impacts across the ecoregion due to soil type and other unmeasured factors. When compared to average irrigated lint yields on the SHP (1087 kg ha −1 ) [38], lint yield in the NTW system (approximately 1032 kg ha −1 ) in 2017 was just below the average yield, while in 2018 lint yield in the NTW and NT systems (1142 kg ha −1 and 965 kg ha −1 , respectively) were numerically greater than the average lint yield across the SHP (808 kg ha −1 , NASS 2018).
The lack of a N treatment effect on cotton lint yield supports previous research that determined the timing of N fertilizer application does not affect yield when the rate of application is unchanged [25,26]. The lack of differences in lint yield between N treatments and the control may be due to several factors including: residual N deeper in the soil profile than was measured, which did not result in N deficiency in the control with no added N; environmental conditions; and, NO 3 − -N in the irrigation water. Agronomic efficiency was not affected by altering the timing of N fertilizer application for any year of this study (Table 3). Related to the lack of N treatment differences on cotton lint yield, it was expected that AE would not vary between the timings of N application. The lack of AE difference was also likely due to the application of N through irrigation water, which would apply a small amount of N at various points throughout the year, potentially leading to comparable N use efficiencies across the tillage systems and N treatments of this three-year study. The concentration of NO 3 − -N in the irrigation water was determined to be 8.44 mg kg −1 in 2018, and NO 3 − -N additions through irrigation were determined based on this concentration (Table 4). This concentration of NO 3 − -N is not uncommon in the area, and falls within the general average of 3 to 10 mg kg −1 in Texas [39]. Average N addition due to irrigation was about 39 kg NO 3 − -N ha −1 per year [39]. With this addition of N in the irrigation water, the timing of N application becomes more complicated as NO 3 − -N is added with each irrigation. This reduces the ability to accurately predict which timing of N fertilizer would be most effective for these systems. Unplanned N addition through the irrigation water may mask any effects of one fertilizer application timing compared to the others and should be accounted for when examining these types of systems in the future.

Fiber Quality
Fiber strength was affected by tillage system in 2017 and 2018 ( Figure 3). Late season rains have long been known to decrease fiber strength [40]. The shift in precipitation from July/August in 2017 to September/October in 2018 with the potential for reduced evaporation in the NTW and NT systems compared to the CT system was likely the cause of the difference between years. Previous research indicates little to no difference in fiber quality between conservation tillage implementation, with or without cover crops, compared to CT systems [9,10,13]. The addition of N fertilizer has been reported to increase fiber strength [22,41], although this was not determined in this study, likely due to the addition of NO 3 − -N through irrigation water and the use of a single N rate across the study with the exception of the control.
When other fiber quality parameters were analyzed within year, it was determined that micronaire was affected by the tillage system in 2017. Micronaire is used as an indication of fiber maturity, with values between 3.5-3.6 and 4.3-4.9 being considered as the base level. However, all micronaire values determined for the cotton lint produced in 2017 would be classified as discount grades (3.4 and under, 5.0 and above), while the average micronaire value across the SHP in 2017 was 3.2 [42]. Micronaire increases in the NTW and NT systems compared to the CT system in 2017 may be due to better establishment due to the benefits of no-tillage systems and thus earlier boll development, resulting in greater heat unit accumulation and a more mature cotton fiber [43]. Due to the large effect of breeding and environment on fiber quality, it is important to continue to evaluate these parameters as cotton varieties and the environment changes on the SHP.

Plant Seed N
Cotton seed N concentration is a good measure of plant N use as it has been shown to constitute up to 70% of cotton N uptake [44]. A reduction in seed N concentration was determined for the control compared to the applied N treatments in 2018 (Table 4, Figure 4). The reduction in seed N concentration in the control would be expected as there was no application of N fertilizer. The delayed reduction in seed N concentration between N treatments and the control until the third year (2018) potentially indicates the use of residual N to satisfy cotton N demands in the control during the first two years. Previous research has indicated a critical seed N concentration threshold for yield reduction of about 35 g kg −1 [45] which all N treatments surpassed during this study with the exception of the control in 2018 (33.6 g kg −1 , Figure 4). Egelkraut, Kissel, Cabrera, Gascho and Adkins [45] determined that above this critical threshold no lint yield increase would be observed, which was supported by the lack of a N treatment effect on cotton lint yield over this three-year period ( Table 2). Previous studies have indicated a correlation between N fertilizer rate and cotton seed N concentration, with seed N in excess of 35 g kg −1 indicating over-fertilization [46]. Due to an average seed N concentration of 38.7 g kg −1 , and the lack of correlation between seed N concentrations and cotton yield, it is possible that fertilizer N rates are above optimal based on this metric. In addition, with all seed N concentrations being greater than the critical threshold with the exception of the control in 2018, it is clear that the addition of N through irrigation water has a significant effect on cotton N uptake. With these seed N concentrations providing support for irrigation N addition effects on cotton lint yield as well, and further indicating irrigation N addition effects on cotton lint yield on the SHP.

Conclusions
The use of a combination of a wheat cover crop and NT can improve yield on the SHP in the first three years of implementation compared to CT without a cover crop. When considering the integration of cover crops and NT in a semi-arid region which experiences extremes in climatic conditions, it is important to consider the early season protection of the soil and cotton seedlings. In addition, in-season moisture stress mitigation afforded by the cover crop residue and reduced tillage is a potentially important benefit of a NTW system. The ability to mitigate extreme heat and wind is important on the SHP where these conditions are common. When making recommendations for best N management practices, timing of fertilizer application is usually included, although in newly implemented conservation tillage systems on the SHP it is critical to also consider irrigation N addition in any nutrient calculations. This is important to consider, as the rate of N application may be a greater determinant for AE in this eco-region and should be evaluated further.