Introduction
Sweetpotato is a valuable commodity in the United States; approximately 58,500 ha were harvested in 2018 with an estimated value of $654 million (USDA 2019a, 2019b). The majority of sweetpotato production areas are concentrated in North Carolina, California, Mississippi, and Louisiana (USDA 2019a). Sweetpotato is fourth most valuable crop for North Carolina, with a production value of more than $236 million (USDA 2019a). ‘Covington,’ an orange-fleshed, table-stock cultivar (Yencho et al. Reference Yencho, Pecota, Schultheis, VanEsbroeck, Holmes, Little, Thornton and Truong2008), constitutes 88% of the state’s hectarage (Schultheis Reference Schultheis2016). Weeds are severe pests in sweetpotato each year; they are capable of interfering with sweetpotato for sunlight, moisture, and nutrients, and can negatively affect yield and quality (Barkley et al. Reference Barkley, Chaudhari, Jennings, Schultheis, Meyers and Monks2016; Coleman et al. Reference Coleman, Chaudhari, Jennings, Schultheis, Meyers and Monks2016; Meyers et al. Reference Meyers, Jennings, Schultheis and Monks2010; Seem et al. Reference Seem, Creamer and Monks2003). Palmer amaranth is the most common and troublesome weed in North Carolina sweetpotato (Webster Reference Webster2010).
Palmer amaranth is an obligate, dioecious, annual weed (Franssen et al. Reference Franssen, Skinner, Al-Khatib, Horak and Kulakow2001) originating from northern Mexico and southwestern United States (Ward et al. Reference Ward, Webster and Steckel2013). It is a C4 plant (Wang et al. Reference Wang, Klessig and Berry1992), with peak photosynthetic rates occurring from 36 to 46 C (Ehleringer Reference Ehleringer1983). Palmer amaranth grows at a rate of 0.18 to 0.21 cm per growing degree day (Horak and Loughin Reference Horak and Loughin2000). Palmer amaranth grows faster and accumulates more dry matter than other Amaranthus spp., including redroot pigweed (A. retroflexus) (Sellers et al. Reference Sellers, Smeda, Johnson, Kendig and Ellersieck2003). Each female plant can produce 200,000 to 600,000 seeds plant−1, depending on time of emergence (Keeley et al. Reference Keeley, Carter and Thullen1987). Palmer amaranth is highly opportunistic, with 100% of viable seed germinating at maturity when optimal temperature and moisture are present (Steckel et al. Reference Steckel, Sprague, Stoller and Wax2004). The tall growth habit of Palmer amaranth allows it to intercept light that would otherwise reach the sweetpotato canopy, and an inverse relationship with Palmer amaranth light interception and sweetpotato storage root yield was reported by Meyers et al. (Reference Meyers, Jennings, Schultheis and Monks2010). In the same study, the authors reported 36% to 68% reduction in marketable yield of ‘Beauregard’ and Covington sweetpotato from season-long interference of 0.5 to 6.5 Palmer amaranth m−1 of crop row. Similarly, Basinger et al. (Reference Basinger, Jennings, Monks, Jordan, Everman, Hestir, Bertucci and Brownie2019) reported from 50% to 79% marketable yield loss in sweetpotato from Palmer amaranth at density of 1 to 8 plants m−1 of crop row. However, to our knowledge, no research has been conducted to determine the critical time of Palmer amaranth control in sweetpotato.
The critical period for weed control (CPWC) is the period in which weeds must be controlled in the crop to avoid loss in yield and quality (Knezevic et al. Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002; Weaver and Tan Reference Weaver and Tan1983). The CPWC represents the overlap between two separately measured competition components: (1) the critical timing of weed removal (CTWR), defined as the maximum amount of time after crop transplanting that weeds can be tolerated without significant yield reduction; and (2) the critical weed-free period, defined as the minimum length of time that weed emergence must be prevented to ensure weed growth does not diminish crop yield (Knezevic et al. Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002). The CPWC can vary depending on the aggressiveness of the crop or weed, crop cultivar, row spacing, planting density, weed species, weed density, environmental conditions, and crop management (Agostinho et al. Reference Agostinho, Gravena, Alves, Salgado and Mattos2006; Knezevic et al. Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002; Norsworthy and Oliveira Reference Norsworthy and Oliveira2004). A CPWC from 2 to 6 wk after transplanting (WAP) has been reported in Beauregard sweetpotato for a mixture of naturally occurring weeds (Seem et al. Reference Seem, Creamer and Monks2003). Because of the fecundity and control challenges of Palmer amaranth, a threshold of 0 plants m−1 is recommended (Norsworthy et al. Reference Norsworthy, Griffith, Griffin, Bagavathiannan and Gbur2014), which makes the CTWR more useful than the critical weed-free period for Palmer amaranth in sweetpotato. A complete understanding of Palmer amaranth CTWR can be helpful to growers to develop effective management strategies. No studies, however, to our knowledge, have been conducted to evaluate the effect of duration of Palmer amaranth interference on sweetpotato. Thus, our objective for this study was to determine the critical timing of Palmer amaranth removal in sweetpotato to avoid reduction in sweetpotato storage root yield and quality.
Materials and Methods
Studies were conducted at the Horticultural Crops Research Station near Clinton, North Carolina (35.02°N, 78.28°W), in 2016 and 2017 in fields with a history of severe Palmer amaranth infestation. Soils were Orangeburg (fine-loamy, kaolinitic, thermic Typic Kandiudults) and Goldsboro (fine-loamy, siliceous, subactive, thermic Aquic Paleudults) loamy sands in 2016 and 2017, respectively. For both soils, pH was 6.4 and humic matter content was less than 1%. At sweetpotato transplanting, all plots were free of emerged Palmer amaranth, due to recent tillage and bedded row formation. Covington sweetpotato plants in field propagation beds were cut by hand and then the nonrooted cuttings transplanted to raised bed rows on July 20, 2016, and May 31, 2017, with a commercial mechanical transplanter. Plots were two rows, each 1-m wide by 6.1-m long. The first row was a weed-free sweetpotato border row and the second row was the treatment row. The experimental design was a randomized complete block with five and four replications in 2016 and 2017, respectively.
Treatments included hand removal of Palmer amaranth at 2, 3, 4, 5, 6, 7, 8, and 9 WAP and plots were kept weed-free thereafter to simulate weed control beginning at various times during the season. A season-long weedy check was included for comparison. Because weeds cannot be hand removed until after emergence, a true weed-free control was not possible. The first removal timing (2 WAP) represented the first feasible opportunity to remove weeds. Palmer amaranth growing in the center of the row (approximately a 30-cm wide band) of each treatment was not controlled from sweetpotato transplanting until each treatment was implemented. Palmer amaranth removal consisted of cutting the base of the stem at the soil surface to avoid disturbance of sweetpotato roots. After Palmer amaranth removal, treatment plots were maintained weed free for the remainder of the season by weekly hand weeding. Border rows were maintained weed free season long by hand removal and cultivation. Across all studies, large crabgrass [Digitaria sanguinalis (L.) Scop.] and goosegrass [Eleusine indica (L.) Gaertn.] were controlled with clethodim (Select Max®, 0.05 kg ai ha−1;Valent USA Corp., Walnut Creek, CA). Broadleaf weeds other than Palmer amaranth were removed weekly by hand.
At each removal timing, height, density counts, and shoot dry biomass of Palmer amaranth were determined for that treatment. Palmer amaranth height was measured from the soil surface to the apical meristem from 10 randomly selected plants plot−1. For shoot dry biomass determination at each removal timing, all Palmer amaranth plants in the treatment plot were harvested, then placed immediately in paper bags, dried at 70 C for at least 1 wk, and then weighed. Palmer amaranth growth measurements were not obtained in the season-long weedy plots because plants had senesced by harvest. Sweetpotato storage roots were harvested at 84 and 119 d after planting in 2016 and 2017, respectively, using a tractor-mounted commercial chain digger to place roots on the soil surface. The relatively early sweetpotato harvest in 2016 was enacted because of an impending hurricane. Storage roots were sorted by hand into jumbo (>8.9 cm in diameter), no. 1 (>4.4 cm but <8.9 cm), and canner (>2.5 cm but <4.4 cm) grades (USDA 2005). Marketable yield was calculated as the sum of jumbo and no. 1 grades. Due to an early harvest and lack of jumbo-grade storage roots, marketable (no. 1 only) and no. 1 yields were the same for 2016. Relative marketable or no. 1 yields were calculated as a percentage of the weed-free yield.
Data were checked for variance of homogeneity and normality before statistical analysis by plotting residuals, and data were transformed when necessary. Yield classes were subjected to square root transformation, whereas Palmer amaranth height and density were subjected to log transformation, and shoot dry biomass data were subjected to log cube root transformation. Transformed data were subjected to a combined ANOVA using PROC MIXED (SAS, version 9.4; SAS Institute, Inc., Cary, NC). Year, timing, and the interaction were considered fixed effects and replications within years was a random effect. The relationship between relative marketable yield and Palmer amaranth height was examined by regressing the least square means at each removal timing for relative marketable yield on the corresponding means for Palmer amaranth height, separately by year, using R, version 1.1.383 (https://support.rstudio.com/hc/en-us). Nonlinear regression (SAS PROC NLIN) was carried out on least square means, separately by year, to describe the effect of duration of Palmer amaranth interference on relative marketable and no. 1 yields and Palmer amaranth height and dry shoot biomass. Because a true weed-free control was not feasible, weed-free marketable and no. 1 yield for each year were estimated using the resulting fitted regressions (Cousens Reference Cousens1985; Streibig et al. Reference Streibig, Combellack, Pritchard and Richardson1989) and percent weed-free yield was then calculated for each yield component in each year. The regression model was then applied to yield expressed as a percentage of estimated weed-free yield.
The three-parameter logistic model (Knezivich et al. Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002) was used to describe the relationship between duration of Palmer amaranth interference and relative marketable or no. 1 yield (Equation 1):
$${\rm{Y = 100\, \times \,[(1\,/\,\{ exp[c\, \times \!(T - d)]\! + \,f\} )\; + \;[(f - 1)\,/\,f]}}]$$
where Y is the yield (percentage of the estimated weed-free yield), T is time expressed as WAP, d is the inflection point, and c and f are constants.
The three-parameter logistic model was used to describe the relationship between duration of Palmer amaranth interference and Palmer amaranth height or shoot dry biomass (Equation 2):
$${\rm{Y = a\;/\;[(1\; + \;c)\; \times }}\,{\rm{exp}}\left( {{\rm{\!- b\; \times \;T}}} \right){\rm{]}}$$
where Y is the height or Palmer amaranth shoot dry biomass, a is the upper asymptote, T is weeks after planting, and c and b are constants.
Results and Discussion
Because of the significant year by treatment interaction for all the measured variables (P < 0.05), data are presented by year.
Palmer Amaranth Density, Height, and Shoot Dry Biomass
The removal timing had no effect on Palmer amaranth density, which ranged from 31 to 62 and 64 to 124 plants m−2 during 2016 and 2017, respectively (data not presented). This suggests that Palmer amaranth emergence occurred immediately after sweetpotato was transplanted, and intense weed pressure was present throughout the sweetpotato growing season.
The removal timing had a significant effect on Palmer amaranth height and shoot dry biomass. The relationship between removal timing and Palmer amaranth height or shoot dry biomass was described by the three-parameter logistic model (Figure 1). Palmer amaranth height increased rapidly from 3 to 6 WAP and plateaued around 7 and 8 WAP for 2016 and 2017, respectively (Figure 1A). The maximum Palmer amaranth height was 155 cm at 7 WAP in 2016 and 190 cm at 8 WAP in 2017. Sellers et al. (Reference Sellers, Smeda, Johnson, Kendig and Ellersieck2003) reported a maximum change in Palmer amaranth height from 4 to 6 WAP with maximum height up to 160 cm tall by 8 WAP. Greater Palmer amaranth height during 2017 could be contributed to higher density and subsequent intraspecific competition as compared with 2016. Likewise, Massinga et al. (Reference Massinga, Currie and Trooien2003) observed Palmer amaranth plants at higher densities were taller with fewer lateral branches than plants growing at lower densities.
The influence of Palmer amaranth removal timing on (A) height (HT) and (B) shoot dry biomass (SDB) in 2016 and 2017 at Clinton, NC. Points represent observed means ± SE. Lines represent predicted values. HT2016 = 158.04 / [(1 + 200.13) × exp(−1.32x)]; R 2 = 0.99; HT2017 = 193.54 / [(1 + 183.97) × exp(−1.04x)]; R 2 = 0.99. SDB2016 = 2.96 / [(1 + 112.66) × exp(−0.92x)]; R 2 = 0.97. SDB2017 = 82.80 / [(1 + 300.94) × exp(−0.42x)]; R 2 = 0.93.
Palmer amaranth shoot dry biomass increased as removal timing was delayed (Figure 1B). In 2016, an increase in shoot dry biomass occurred until 6 WAP and then plateaued by 7 WAP. However, in 2017, the response differed, and a continuous increase in shoot biomass was observed as removal timing was delayed. The difference in Palmer amaranth growth between years could be due to an earlier sweetpotato planting in 2017 compared with 2016. Similarly, Keeley et al. (Reference Keeley, Carter and Thullen1987) reported that early-emerged (between March to June) Palmer amaranth plants produced greater biomass and more seeds compared with late-emerged (between July and October) plants.
Sweetpotato Yield
The relationships between duration of Palmer amaranth interference and relative marketable or no. 1 sweetpotato yields were described by a three-parameter logistic model in 2016 and 2017 (Figure 2). Compared with the estimated weed-free yield, season-long Palmer amaranth interference reduced marketable (85% and 95%) and no. 1 (85% and 97%) yield in 2016 and 2017, respectively (Figure 2 A and B). Sweetpotato yield response to season-long Palmer amaranth interference is similar to previous research in sweetpotato (Basinger et al. Reference Basinger, Jennings, Monks, Jordan, Everman, Hestir, Bertucci and Brownie2019; Meyers et al. Reference Meyers, Jennings, Schultheis and Monks2010) and was greater than reported in other crops (e.g., 54% in cotton) (Morgan et al. Reference Morgan, Baumann and Chandler2001), 79% in soybean [Glycine max (L.) Merr.] (Bensch et al. Reference Bensch, Horak and Peterson2003), 63% in grain sorghum [Sorghum bicolor (L.) Moench ssp. bicolor] (Moore et al. Reference Moore, Murray and Westerman2004), 45% in grafted watermelon [Citrullus lanatus (Thunb.) Matsum. & Lakai] (Bertucci et al. Reference Bertucci, Jennings, Monks, Schultheis, Louws and Jordan2019a), and 68% in peanut (Arachis hypogaea L.) (Burke et al. Reference Burke, Schroeder, Thomas and Wilcut2007).
The influence of Palmer amaranth removal timing on (A) marketable (MKT), and (B) no. 1 sweetpotato storage root yields in 2016 and 2017 at Clinton, NC. MKT sweetpotato yield is a combination of jumbo and no. 1 grade roots. Points represent observed means ± SE. Lines represent predicted values. No. 12016 and MKT2016 = 100 × [(1 /{exp[1.87 (x − 4.31)] + 1.16}) + [(1.16 − 1) / 1.16]], R 2 = 0.99; No.12017 = 100 × [(1 /{exp[0.70 (x − 5.97)] + 1.07}) + [(1.07 − 1) / 1.07]], R 2 = 0.96; MKT2017 = 100 × [(1 /{exp[0.90 (x − 5.34)] + 1.05}) + [(1.05 − 1) / 1.05]], R 2 = 0.99.
Sweetpotato yield loss displayed a strong inverse linear relationship with Palmer amaranth height in both years (Figure 3). There was a 0.6% and 0.4% decrease in yield for every cm of Palmer amaranth height in 2016 and 2017, respectively. A 5% reduction in yield was observed by the time Palmer amaranth reached a height of 27 cm (2.8 WAP) and 10.8 cm (2.3 WAP) in 2016 and 2017, respectively (Figure 3). Sweetpotato has a vining growth habit and develops a canopy less than 50 cm tall (Huaman Reference Huaman1987; Meyers et al. Reference Meyers, Jennings, Schultheis and Monks2010). Palmer amaranth was taller than sweetpotato during the majority of the growing season and was much taller (> 1 m) after 5 WAP both years (Figures 1A and 4). The interception of light by Palmer amaranth foliage had a negative impact on crop growth and yield (Keeley et al. Reference Keeley, Carter and Thullen1987; Meyers et al. Reference Meyers, Jennings, Schultheis and Monks2010; Rowland et al. Reference Rowland, Murray and Verhalen1999). Meyers et al. (Reference Meyers, Jennings, Schultheis and Monks2010) reported Palmer amaranth caused intense shading of the sweetpotato canopy, which only allowed 68% to 82% light (6 to 10 WAP) to reach the sweetpotato canopy at densities of 6.5 plant m−1, which consequentially reduced sweetpotato yield and quality. Authors also reported a linear relationship between Palmer amaranth light interception and sweetpotato yield loss with yield loss increasing as shading increased.
Relationship between marketable storage root yield and Palmer amaranth height in 2016 and 2017 at Clinton, NC. Marketable sweetpotato yield is a combination of jumbo and no. 1 grade roots. Points represent observed means, and lines represent predicted values. Y 2016 = −0.61x + 111.49; R 2 = 0.97. Y 2017 = −0.44x + 99.81; R 2 = 0.98.
Palmer amaranth interference (A) 3, (B) 5, and (C) 7 wk after planting in sweetpotato at Clinton, NC, in 2017.
A 5% acceptable yield-loss threshold has been used in other crops to measure CTWC (Bertucci et al. Reference Bertucci, Jennings, Monks, Schultheis, Louws, Jordan and Brownie2019b; Chaudhari et al. Reference Chaudhari, Jennings, Monks, Jordan, Gunter, McGowen and Louws2016; Everman et al. Reference Everman, Clewis, Thomas, Burke and Wilcut2008). However, the choice of 5% is an arbitrary measurement of yield loss and can be altered on the basis of the economics of weed management or the risk that the farmer is willing to take (Knezevic et al. Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002). The model illustrated that sweetpotato yield loss due to Palmer amaranth interference prevailed throughout the growing season and increased with the delay of removal timing (Figure 2A and 2B). The CTWR for Palmer amaranth in sweetpotato was 2 WAP in 2016 and 2017. The requirement of an early CTWR to prevent yield loss emphasizes the importance of early-season scouting and Palmer amaranth removal in sweetpotato fields. Any delay in removal can result in substantial yield reductions and fewer premium quality roots (no. 1).
These results show that Palmer amaranth interference caused significant yield reduction in sweetpotato when management was not initiated immediately after transplanting or early in the season. Palmer amaranth interference for 2 wk can cause up to 5% marketable sweetpotato yield loss, indicating that this weed can be detrimental to yield even if allowed to grow for a short time after emergence. The CTWR for Palmer amaranth can be achieved through use of PRE herbicides and other weed management tools (e.g., cultivation, hand removal) in sweetpotato. The high level of Palmer amaranth interference with sweetpotato results in the need for effective control as soon as sweetpotato is transplanted.
Acknowledgments
The authors thank the North Carolina Agricultural Foundation; the Department of Horticultural Science, North Carolina State University (NCSU); the North Carolina Agricultural Research Service (NCSU), and North Carolina Department of Agriculture and Consumer Services for support of this research. The authors also thank Nicholas Basinger, Matthew Bertucci, and Matthew Waldshmidt for assistance with study implementation and data collection. The authors thank the staff at the Horticultural Crops Research Station, Clinton, North Carolina. No conflicts of interest have been declared.
