1. Introduction
Palmer amaranth [
Amaranthus palmeri (L.) S. Wats] is one of the several pigweed species that are problematic in row crops in the southeastern United States. Compared to other pigweed species, such as common waterhemp [
Amaranthus rudis (L.) Sauer], redroot pigweed (
Amaranthus retroflexus L.) and tumble pigweed (
Amaranthus albus L.), Palmer amaranth produced the highest dry weight, leaf area and height [
1]. Palmer amaranth grows relatively quickly and can attain a height of 2 m or more [
1]. It is a dioecious plant with tremendous seed production potential and rapid seed germination [
1,
2,
3]. A single female plant can produce more than 600,000 seeds, depending upon density, which have an average diameter of 1.0 mm [
2]. It has exceptional drought tolerance [
4,
5,
6,
7]. Additionally, Palmer amaranth can grow under low light conditions, such as dense crop canopies [
8]. Palmer amaranth interference and subsequent yield losses have been documented in several crops, such as cotton, corn, cucurbits, grain sorghum, peanut, potato, soybean, sweet potato and several vegetable crops [
9,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23].
Until recently, glyphosate-resistant cotton production systems were very effective for managing a broad spectrum of weeds, including Palmer amaranth [
24,
25]. However, the evolution of glyphosate resistant Palmer amaranth has forced cotton producers to explore other management options and integrated approaches. These included inversion tillage and adoption of glufosinate-resistant varieties [
26]. Additionally, resistance to dinitroaniline herbicides has also been reported in some Palmer amaranth populations [
26,
27].
The role of tillage in altering the distribution, abundance, composition of species, as well as seedling emergence patterns, has been well documented [
28,
29,
30,
31,
32,
33,
34,
35,
36]. In conservation tillage systems, where soil incorporation is minimized, seeds accumulate near soil surface. Contrarily, the soil disturbance resulting from various tillage practices places weed seeds at different depths that vary in availability of moisture, diurnal temperature fluctuation, light exposure and activity of predators [
37]. Moldboard plowing buries weed seeds deeply in the soil; however, deeper burial may lead to long-term weed problems, because of increased seedbank longevity [
38].
The type of tillage implement used to till the soil greatly influences the vertical distribution and density of seeds within the soil profile [
28,
30,
36,
39,
40,
41,
42]. Inversion tillage implements bury a large proportion of the weed seed, while non-inversion tillage implements leave more of the seed near the soil surface [
28]. Previous research demonstrated that more than 60% of weed seedbank was concentrated in top 5 cm following either a no-till or chisel plow [
41]. Therefore, considering the inability of small Palmer amaranth seed to emerge from depths greater than 7.5 cm and a light requirement for germination, moldboard plowing followed by a conservation system may reduce the Palmer amaranth populations to manageable levels.
Currently, an integrated weed control system utilizing high-residue cover crops as a weed management tool is gaining popularity [
43,
44,
45,
46,
47,
48,
49,
50,
51,
52,
53,
54]. Cover crops provide early season weed control by reducing light transmission and quality, altering soil temperature, physically suppressing weed emergence and allelopathy [
55,
56,
57,
58]. Cereal rye (
Secale cereale L.) has been documented as having high biomass potential, early season weed suppression and allelopathic properties by several researchers [
44,
46,
47,
48,
59,
60,
61,
62]. It has also been observed that cereal rye residue alone was effective at reducing glyphosate-resistant Palmer amaranth emergence by 94% in the row middle and 50% within the cotton row [
63]. Others reported that the use of high residue cover crops in conjunction with chemical and cultural weed control tactics provided effective control of established glyphosate-resistant Palmer amaranth, while helping to prevent the development of resistance in glyphosate-susceptible populations [
64].
Considering the magnitude of current herbicide resistance problems, inversion tillage can likely improve control of glyphosate-resistant Palmer amaranth. However, increased input costs and potential soil erosion are significant challenges for growers. Integration of cover crops and glufosinate-resistant cotton technology may be viable alternatives in the light of these economic and environmental considerations. Therefore, with the current Palmer amaranth management challenges, a field study was conducted to evaluate the role of primary inversion tillage, cover crops and secondary tillage methods for Palmer amaranth management in glufosinate-resistant cotton.
2. Materials and Methods
2.1. Experimental Design and Establishment
A three year field experiment was conducted from fall 2008 through harvest 2011 at the E.V. Smith Research Center, Field Crops Unit near Shorter, AL, on a Compass sandy loam soil (coarse-loamy, siliceous, subactive, thermic Plinthic Paleudults) with 1.9 to 2.1% organic matter and pH 6.2 to 6.4. The experiment occupied a site that had been in continuous strip-tillage for previous six years. The entire experimental area was infested with glyphosate-susceptible Palmer amaranth prior to experiment establishment, and the subsequent treatments remained in the same location for three years without re-randomization of treatments. Treatments consisted of a factorial arrangement of two levels of soil inversion—fall inversion tillage (IT) and non-inversion tillage (NIT)—three levels of winter cover crops—cereal rye, crimson clover (
Trifolium incarnatum L.) and none (
i.e., winter fallow)—and four different spring tillage methods, resulting in a 24-treatments test. The four spring tillage methods were disk followed by (fb) chisel plow (DCH), disk fb field cultivator (DCU), disk twice (DD) and a no-tillage control (NT). The experimental design consisted of a split-split plot treatment restriction in a randomized complete block design with three replicates. Soil inversion, cover crop and spring tillage were assigned to the main plots, sub plots and sub-sub plots, respectively. The size of the main, sub and sub-sub plots were 43.9 m by 9.1 m, 14.6 by 9.1 m and 3.6 by 9.1 m, respectively. A schedule of operations performed each year is given in
Table 1.
Table 1.
Schedule of operations performed during the experiment.
Table 1.
Schedule of operations performed during the experiment.
| Experimental Years |
---|
Operations | 2008–2009 | 2009–2010 | 2010–2011 |
Broadcasting of Palmer amaranth seed | 19 Nov 2008 | – | – |
Fall inversion tillage | 19 Nov 2008 | – | – |
Planting of cover crops | 20 Nov 2008 | 6 Jan 2010 | 2 Dec 2010 |
Rolling and termination of cover crops | 22 Apr 2009 | 18 May 2010 | 19 Apr 2011 |
Subsoiling | 23 Apr 2009 | 24 May 2010 | 26 Apr 2011 |
Planting of cotton | 1 Jun 2009 | 27 May 2010 | 5 May 2011 |
Fertilization (16-16-16) | 1 Jun 2009 | 27 May 2010 | 5 May 2011 |
POST application | 16 Jun 2009 | 16 Jun 2010 | 24 May 2011 |
Graminicide application (Sethoxydim + COC) | 13 Jul 2009 | 8 Jul 2010 | 6 Jul 2011 |
LAYBY application | 14 Aug 2009 | 16 Aug 2010 | 19 Jul 2011 |
Cotton defoliation | 26 Oct 2009 | 14 Oct 2010 | 13 Sep 2011 |
Cotton harvesting | 9 Nov 2009 | 20 Oct 2010 | 30 Sep 2011 |
In the fall 2008, approximately 28 million native glyphosate-susceptible Palmer amaranth seeds were broadcast per hectare to ensure a sizeable seedbank of this weed. Prior to broadcasting, Palmer amaranth seed germination was tested by placing 25 seeds on commercial germination paper in four petri dishes at 35 °C. Seeds were kept moist with tap water inside closed petri dishes. Seeds were considered germinated when the radicle emerged 1 mm. Germination percentage was calculated as the number of germinated seeds divided by the total number of seeds multiplied by 100. Two weeks after initiation, 87% of the seeds germinated. One half of each block was subject to fall inversion tillage (IT) by moldboard plowing (30 cm) immediately fb one pass each of a disk and field cultivator; the other half was under non-inversion tillage (NIT) using a within-row subsoiler equipped with pneumatic tires to close the subsoiling slot. During the fall of each year, cereal rye (cv. ‘Elbon’ in 2009 and 2010 and ‘Wrens Abruzzi’ in 2011) and crimson clover (
Trifolium incarnatum L. cv. ‘Dixie’) cover crops were seeded at rates of 101 and 28 kg seed ha
−1, respectively, in both IT and NIT. Different cereal rye cultivars had to be used due to seed availability; Wrens Abruzzi has been shown to be more allelopathic [
65]. In 2009 and 2010, frequent rains delayed both the harvesting of cotton and subsequent planting of cover crops. Cereal rye cover was fertilized using 34 kg ha
−1 of a 33-0-0 fertilizer. A winter fallow control was also included as check.
2.2. Cover Crop Management
Cover crops were rolled with a three section straight bar roller (Bigham Brothers Inc., Lubbock, TX, USA; Mention of a trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the USDA or Auburn University and does not imply endorsement of a product to the exclusion of others that may be suitable) in late April or early May using a JD 7730 equipped with an AutoSteer GPS. Cover crop rolling was immediately followed by an application of glyphosate (Roundup Weathermax®, Monsanto Company, St. Louis, MO, USA) at 0.84 kg ae ha−1 plus glufosinate (Ignite®, Bayer Crop Science, Research Triangle Park, NC, USA) at 0.49 kg ai ha−1; the mixture was needed to enhance crimson clover termination. Cover crop biomass samples were taken prior to desiccation, and dry biomass was recorded. The entire experimental area was sub-soiled in May to 45 cm depth to break hardpans. A within-row subsoiler equipped with pneumatic tires only, to close the subsoiling slot, was used. Sub-soiling was followed by planting of glufosinate-resistant cotton (cvs. FM 1845 LLB2 in 2009 and FM 1735 LL in 2010 and 2011, Bayer Crops Science, Research Triangle Park, NC, USA). Each year, cotton was fertilized using 211 kg ha−1 of 16-16-16 fertilizer at the time of planting.
2.3. Secondary Tillage and Weed Management
The DCH tillage consisted of a single pass of 3 m disk fb a single pass of 1.8 m chisel plow, DD consisted of double pass of 3 m disk and DCU was a single pass of 3 m disk fb a single pass of 4.1 m field cultivator. A single postemergence (POST) application of a tank mixture of glufosinate at 0.60 kg ai ha−1 (Ignite®, Bayer Crops Science, Research Triangle Park, NC, USA) plus S-metolachlor (Dual II Magnum®, Syngenta Crop Protection, Inc., Greensboro, NC, USA) at 0.54 kg ai ha−1 tank mixture was made to 5 to 7.5 cm Palmer amaranth, between 15 and 20 days after planting cotton, with an ATV-mounted sprayer delivering 145 L ha−1 with flat-fan spray tips. In 2011, an additional POST application of glufosinate at 0.60 kg ai ha−1 plus S-metolachlor at 0.54 kg ai ha−1 was carried out three weeks after the first POST application. A last directed POST application (LAYBY) of prometryn (Caporal®, Syngenta Crop Protection, Inc., Greensboro, NC, USA) at 0.84 kg ai ha−1 plus MSMA (Drexel Chemical Company, Memphis, TN, USA) at 1.4 kg ai ha−1 was carried out approximately 2 months after the first POST application. Sethoxydim (Poast Plus®, Bayer Ag. Products, Research Triangle Park, NC, USA) was applied at 0.28 kg ai ha−1 as needed to maintain grass control.
2.4. Palmer Amaranth Sampling and Visual Control Ratings
Palmer amaranth density was determined both between (BR) and within (WR) the cotton rows before the POST application and again before the LAYBY application. BR Palmer amaranth density was recorded as number of plants in a quadrat (0.25 m−2) randomly placed at four different positions between the second and third row of a four-row cotton plot. Similarly, WR Palmer amaranth density was recorded from a quadrat (0.25 m−2) randomly placed at four different positions within the second and third rows. Palmer amaranth control was also assessed visually, for the entire plot, at weekly intervals on a 0 to 100 scale, where 0 and 100 indicate no control and complete control, respectively. Palmer amaranth was hand-pulled from all the plots before the LAYBY application, but following density counts and control ratings to facilitate cotton harvest. Cotton yields were determined by mechanically harvesting the two central rows within each four-row plot with a spindle picker.
2.5. Statistical Analysis
Data were analyzed using generalized linear mixed models or linear mixed models methodology as implemented in SAS® PROC GLIMMIX based on the underlying design, which was a randomized complete block design (r = 3) with a split-split-split plot in time restriction on randomization. Soil inversion, cover crop, spring tillage method, year and all their interactions were treated as fixed effects. Block and Block × treatment factors were treated as random effects. The split plot nature of the experiment requires five different residual terms: (1) block × soil inversion as the appropriate error term for soil inversion; (2) block × soil inversion × cover crop as the appropriate error term for cover crop and its interaction with soil inversion; (3) block × soil inversion × cover crop × spring tillage method as the appropriate error term for spring tillage and its interaction with soil inversion and cover crop; (4) block × year as the appropriate error term for year; and (5) the residual error term as the appropriate error term for all interactions effects of year with the remaining factors. The factor year is of a repeated measures nature that induces a covariance relationship because of the lack of re-randomization. All the standard covariance models were evaluated, but none improved the AICC fit statistic, which is a penalized -2log likelihood. However, grouping the residual variance by year using the “random _residual/group = year” option in SAS gave a slightly improved fit. Fit was improved by creating variance groups, even though the maximum F-test of residuals among the three years did not detect heterogeneous variances. Palmer amaranth density data were analyzed using a lognormal distribution function, and back transformed means along with 95% confidence intervals are reported. Palmer amaranth control rating data at three and six weeks after application were arcsine-transformed, and back transformed means along with 95% confidence intervals are reported. No transformation was required for cover crop biomass and cotton yield data. Multiple means’ comparisons of significant effects were made using the “Adj = simulate” option in SAS PROC GLIMMIX at the 5% significance level.