Introduction
Weeds are the most damaging of pests in agronomic production systems, with competition from weeds causing greater yield losses than insects or pathogen pathogens (Oerke Reference Oerke2006). Herbicides are the primary source of control in corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] in the United States (Gianessi and Reigner Reference Gianessi and Reigner2007). Selective herbicides, such as dicamba, a Group 4 herbicide, have been useful for several decades for targeting specific weeds in monocot field crops (Canode and Robocker Reference Canode and Robocker1970). Dicamba controls broadleaf weeds within grass crops, as grasses are able to effectively metabolize the herbicide to prevent injury (Chang and Born Reference Chang and Born1971). With the recent commercialization of dicamba-resistant soybean varieties, the use of dicamba has increased substantially. Approximately 60% of the soybean acres in the United States were dicamba resistant in 2019 (Unglesbee Reference Unglesbee2019).
Weed management practices impose selection pressures that drive shifts in weed communities. Documented cases of weed shifts due to changes in weed management have occurred as a result of tillage, irrigation systems, herbicide use, and crop rotation (Brim-DeForest et al. Reference Brim-DeForest, Al-Khatib, Linquist and Fischer2017; Davis et al. Reference Davis, Gibson, Bauman, Weller and Johnson2009; Johnson and Coble Reference Johnson and Coble1986; Johnson et al. Reference Johnson, Davis, Kruger and Weller2009; Menalled et al. Reference Menalled, Gross and Hammond2001; Tuesca et al. Reference Tuesca, Puricelli and Papa2001). Shifts in weed species have also been observed when comparing glyphosate-resistant cropping systems with conventional herbicide systems. Late-emerging weed species have been shown to be more prevalent in cropping systems that use POST herbicide applications (Swanton et al. Reference Swanton, Gulden, Sikkema, Tardif and Hamill2010). Johnson et al. (Reference Johnson, Davis, Kruger and Weller2009) discussed the concern of weed populations shifting to more problematic and herbicide-resistant weed biotypes that will reduce the usefulness of technologies such as glyphosate-resistant crops.
One of the most prevalent weed shifts in modern history occurred as a result of the wide-scale adoption of glyphosate-resistant corn, soybean, and cotton (Gossypium hirsutum L.) (Dill et al. Reference Dill, Cajacob and Padgette2008). This led to more than one application per season of glyphosate being applied to many acres with little diversity in herbicide strategies. This process selected for glyphosate-resistant biotypes, which Young (Reference Young2006) argued would be a negative consequence of this technology. In the absence of an integrated weed management approach, weed shifts from tolerant to resistant species will occur as a result of widespread adoption of dicamba-resistant soybean. Both species richness and evenness were greater when crop rotations were not implemented or were continually planted to glyphosate-resistant traits (Young et al. Reference Young, Gibson, Gage, Matthews, Jordan, Owen, Shaw, Weller and Wilson2013). A survey of Nebraska farmers in 2017 showed that 20% had planted soybean resistant to dicamba and glyphosate, and that 60% of those used dicamba, glyphosate, or the combination of the two as their only source of POST weed control (Werle et al. Reference Werle, Oliveira, Jhala, Proctor, Rees and Klein2018). Although research has shown that dicamba can be a valuable tool for controlling problematic weeds (Chahal and Johnson Reference Chahal and Johnson2012), it is important to use multiple effective SOAs to reduce the selection pressure for dicamba-resistant biotypes (Shergill et al. Reference Shergill, Bish, Biggs and Bradley2017).
Dicamba-resistant soybean varieties were introduced to be commercially grown before the 2017 growing season and were developed to aid producers in controlling problematic herbicide-resistant broadleaf weeds. There are several herbicide-resistant weed species that present challenges to Indiana growers, including horseweed [Conyza canadensis (L.) Cronquist], giant ragweed (Ambrosia trifida L.), waterhemp [Amaranthus tuberculatus (Moq.) Sauer], and Palmer amaranth (Amaranthus palmeri S. Watson) (Heap Reference Heap2020). Glyphosate-resistant C. canadensis was detected in anywhere from 15% to 78% of populations across all regions of Indiana more than 12 yr ago (Davis et al. Reference Davis, Gibson and Johnson2008). These herbicide-resistant weeds pose a threat to corn and soybean yield. Ambrosia trifida can reduce yields of soybean as much as 52% with densities of only 2 plants 3 m−2, and can reduce corn yields by up to 90% under high A. trifida densities (Baysinger and Sims Reference Baysinger and Sims1991; Harrison et al. Reference Harrison, Regnier and Webb2001). The addition of dicamba as an active ingredient in soybean will provide in-season control options for several glyphosate-resistant broadleaf weed species, as the addition of dicamba to glyphosate increased the control of glyphosate-resistant A. palmeri, A. tuberculatus, and C. canadensis to at least 95% (Johnson et al. Reference Johnson, Bradley, Young, Matthews, Marquardt, Slack, York, Culpepper, Hager, Al-Khatib, Steckel, Moeching, Loux, Bernards and Smeda2010). Byker et al. (Reference Byker, Soltani, Robinson, Tardif, Lawton and Sikkema2013) found that 900 g ae ha−1 glyphosate + 600 g ae ha−1 dicamba applied preplant, followed by a POST application of 900 g ha−1 of glyphosate + 300 g ha−1 of dicamba, resulted in at least 95% C. canadensis control in dicamba-resistant soybeans across three locations in Ontario, Canada. However, Spaunhorst and Johnson (Reference Spaunhorst and Johnson2016) reported that utilizing dicamba as a PRE herbicide alone could result in less than 50% control of glyphosate-resistant A. palmeri, but when used with metribuzin, control increased more than 30%. It will be important to utilize dicamba-resistant soybeans with multiple other SOAs and residual herbicides in years that both corn and soybean are grown.
Currently there are only two reported species that have evolved resistance to dicamba in the United States (Heap Reference Heap2020). The two species with reported resistance are kochia [Bassia scoparia (L.) A.J. Scott] and prickly lettuce (Lactuca serriola L.), both common in small grains production, where crop rotation is minimal and dicamba is applied year after year and auxins are heavily relied upon (Fernandez-Cornejo et al. Reference Fernandez-Cornejo, Nehring, Osteen, Wechsler, Martin and Vialou2011). Therefore, there is a reasonable concern that dicamba resistance will evolve as auxin-resistant soybean varieties are used on more acreage because of overuse of this SOA for controlling other herbicide-resistant broadleaf weeds. Although dicamba resistance could evolve, a shift toward more tolerant monocot species is likely to occur without application of residual herbicides and would likely be a more immediate concern.
The objective of this research was to identify shifts in the weed community, in terms of both weed density and species richness, in a corn–soybean rotation with varying levels of dicamba selection pressure over 7 yr.
Materials and Methods
Field Sites
A field trial in a corn–soybean rotation was initiated in 2013 and continued through the 2019 growing season. Experiments were conducted at the Throckmorton Purdue Agricultural Center, Lafayette, IN (40.30°N, 86.90°W). Global positioning system coordinates were taken to mark the corners of the trial areas due to the long-term nature of this project to ensure the trials remained in the same location throughout the 7-yr period. Corners of the trial area were additionally marked to ensure trial remained in the same location from year to year. The site was a conventional-till site on a Toronto-Millbrook (fine-silty, mixed, superactive, mesic Udollic Epiaqualfs) complex that was chisel plowed in the fall and disked and field cultivated in the spring. The soil has an organic matter of 2.6%, a pH of 6.1, and a CEC of 10.6 meq 100 g−1. Fertility programs were adjusted based on soil test values. Corn was planted in 2013, 2015, 2017, and 2019 at a rate of 80,000 seeds ha−1, while soybean was planted in 2014, 2016, and 2018 at a rate of 350,000 seeds ha−1. Corn hybrids used had traits that conferred resistance to both glufosinate and glyphosate. Soybean varieties used were resistant to dicamba and glyphosate.
Experimental Design and Herbicide Strategies
The experimental design was a random complete block with six replications. Plot size was 6-m wide and 18-m long. Herbicide strategies were developed to evaluate weed community shifts as SOAs are implemented into a 2-yr corn–soybean cropping system. The four treatments were labeled as follows: (1) dicamba reliant, (2) diversified glyphosate, (3) diversified dicamba, and (4) fully diversified consisting of three, six, seven, and eight SOAs (Table 1). Herbicides were chosen in order to control the problematic weeds. Herbicide applications were made with a 3-m CO2 -propelled backpack sprayer calibrated to deliver 140 L ha−1 at 4.8 km h−1. Flat-fan AIXR11002 nozzles (TeeJet Technologies, 1801 Business Park DR, Springfield, IL 62703) were used to apply treatments that did not contain dicamba, while TTI11003 nozzles (TeeJet Technologies) were used for dicamba applications. Herbicide application dates for each year can be found in Table 2. During corn years, two applications were made to each plot, including a PRE, followed by a POST when weeds were 10- to 15-cm tall or when corn reached 76 cm in height. POST applications were made in mid-June and are hereafter referred to as “early-summer” evaluations. During soybean years a PRE was applied followed by a POST application when weeds were 10 to 15 cm. An early-POST followed by a late-POST application was used in the dicamba-reliant strategy due to the lack of a soil residual herbicide in soybean years.
Herbicide strategies used in both corn and soybean years.a
a Table adapted from Legleiter (Reference Legleiter2017: 156–157).
b SOA, site of action.
c Before 2018 Clarity® was used in place of XtendiMax®.
Herbicide application dates from 2013 to 2019.
Field Data Collection
To monitor changes over time in weed density and species richness within plots, two 1-m2 quadrats were placed 4.5 m from the back and front of each plot and 1 m from plot edges. These quadrats were placed in the same location every year. Weed densities and species richness were recorded before POST herbicide applications. Trials were harvested once crops reached physiological maturity, and grain weight and moisture for reach plot were recorded. Weed density and richness were partitioned into total weed measurements, as well as separated into dicot and monocot categories. Data were subjected to ANOVA using the PROC GLIMMIX procedure in SAS v. 9.4 (SAS, 100 SAS Campus Drive, Cary, NC 27513-2414), with year treated as a repeated measure. Means separation was conducted using Tukey’s honestly significant difference (HSD) test (α = 0.05).
Soil Seedbank Data Collection
Before spring tillage, 16 soil cores were randomly sampled from each plot to assess weed seedbank composition. Cores measured 5.7 cm in diameter and were collected to a depth of 7.6 cm, resulting in approximately 3,000 cm3 of soil from each plot. Pareja et al. (Reference Pareja, Staniforth and Pareja1985) determined that 85% of all seeds in a reduced-tillage system, and 28% in a conventional-tillage system were in the top 5 cm of soil. Cores were homogenized and placed into 25 by 50 cm soil flats in a Purdue University greenhouse in West Lafayette, IN, where the seeds were allowed to germinate for 8 wk. Greenhouse conditions were established as a 16-h photoperiod with 600-W high-pressure sodium lights, with a temperature of approximately 26 C.
Weed density and species were recorded every 2 wk, and weeds were removed by hand after each recording date. Following data collection at the 4th week, the soil was mixed thoroughly to promote germination of additional seeds remaining in flats. After the 8-wk period was complete, the soil was discarded. Weed densities and species richness are presented per 3,000 cm3. Other studies have presented soil seedbank densities in terms of area (e.g., m−2; Carter and Ivany Reference Carter and Ivany2006; Conn et al. Reference Conn, Cochrane and Delapp1984; Menalled et al. Reference Menalled, Gross and Hammond2001; Moonen and Barberi Reference Moonen and Barberi2004). Seed density on a volume basis (cm3) has also previously been used to compare spatial analysis methods within soil seedbanks (Bigwood and Inouye Reference Bigwood and Inouye1988). The data were analyzed all together, referred to as total and separated into monocots and dicots. All data were subjected to ANOVA using the PROC GLIMMIX procedure in SAS v. 9.4, with year serving as a repeated measure, and means were separated using Tukey’s HSD test (α = 0.05).
Results and Discussion
Early-Summer POST Application Weed Densities and Species Richness
Total, monocot, and dicot weed densities and species richness were all influenced by an interaction between herbicide strategy and year (Table 3). Total weed density was highest in the dicamba-reliant treatments (Figure 1), and generally higher in the soybean years than in the corn years. Monocot densities increased over time in the dicamba-reliant treatment, but remained constant in the other three treatments. Dicot densities in the dicamba-reliant treatment declined over time, but like monocot densities, tended to be higher in soybean years than in corn years. By 2019, monocots accounted for more than 90% of the total weed density.
ANOVA table for the influence of herbicide strategy, year, and the interaction of the two on in-field, mid-June, and soil seedbank total, dicot, and monocot density, and species richness from 2013 to 2019.
(A) Total, (B) monocot, and (C) dicot in-field densities at the Throckmorton Purdue Agricultural Center (8343 US-231, Lafayette, IN) in mid-June. Standard error bars shown. Asterisks represent differences in mean separation according to Tukey’s honest significant difference (HSD) test (P ≤ 0.05) within year as influenced by an interaction between year and herbicide strategy at early-summer evaluations. Corn was grown in 2013, 2015, 2017, and 2019, and soybeans were grown in 2014, 2016, and 2018.
The dicamba-reliant treatment had a higher dicot species richness compared with all other treatments in years that soybeans were grown (Figure 2). This is possibly due to the lack of one dominant weed species, which allowed other dicot species to find a niche that is usually inhabited by a dominant weed species such as A. trifida or Amaranthus spp. Total, dicot, and monocot species richness were always higher in soybean years in the dicamba-reliant strategy (Figure 2). The increased species richness in soybean years is likely either due to the lack of atrazine used, or soybeans being a less competitive crop compared with corn for some weed species (Knake and Slife Reference Knake and Slife1962; Moolani et al. Reference Moolani, Knake and Slife1963).
(A) Total, (B) monocot, and (C) dicot in-field species richness at the Throckmorton Purdue Agricultural Center (8343 US-231, Lafayette, IN) in mid-June. Standard error bars shown. Asterisks represent differences in mean separation according to Tukey’s honest significant difference (HSD) test (P ≤ 0.05) within year as influenced by an interaction between year and herbicide strategy at early-summer evaluations. Corn was grown in 2013, 2015, 2017, and 2019, and soybeans were grown in 2014, 2016, and 2018.
The research presented in this study is the first to date to evaluate weed community shifts in dicamba-resistant soybeans rotated with corn and showed that species richness will increase if dicamba is used with only glyphosate, and atrazine in corn. Shergill et al. (Reference Shergill, Bish, Biggs and Bradley2017) evaluated weed shifts in dicamba-resistant continuous soybean, but did not evaluate shifts in dicamba-resistant soybeans rotated with corn. Species richness was highest in the dicamba-reliant strategy, resulting in three more species at the end of the trial compared with all other herbicide strategies (data not shown).
Using six or more SOAs with residual herbicides in the fully diversified strategy in both corn and soybean years resulted in a 98% decrease in density compared with using three SOAs with a residual only in corn years. Using six or more SOAs reduced species richness and reduced weed densities compared with a herbicide strategy that only implemented dicamba and glyphosate in soybean years with the addition of atrazine in corn years. In other published research, weed populations were reduced by 50% in continuous corn when more mechanisms of action were included (Wilson et al. Reference Wilson, Young, Matthews, Weller, Johnson, Jordan, Owen, Dixon and Shaw2011). The addition of corn in rotation with dicamba-resistant soybeans will have large implications on weed communities due to the broad spectrum of herbicides used in corn that are not used in soybeans, such as atrazine and 4-hydroxyphenylpyruvate dioxygenase herbicides (e.g., mesotrione, tembotrione, and topramezone).
Soil Seedbank Weed Densities and Species Richness
Total and monocot weed densities within the soil seedbank were influenced by a year by herbicide strategy interaction, while dicot weed densities were only influenced by year (Table 2). Total and monocot densities showed a gradual decline from 2013 to 2015, but were highest in the dicamba-reliant strategy in 2018 and 2019 (Figure 3). The dicamba-reliant strategy had greater than 43% higher total weed densities than all other treatments due to having a higher monocot density that was at least 71% higher than in the other treatments. All herbicide strategies had similar dicot densities. A less diversified herbicide strategy with only three SOAs and no residual herbicides in soybean years over the course of 7 yr has resulted in those plots having higher total weed densities, composed primarily of monocots.
(A) Total and (B) monocot soil seedbank densities from the Throckmorton Purdue Agricultural Center (8343 US-231, Lafayette, IN). Standard error bars shown. Asterisks represent differences in mean separation according to Tukey’s honest significant difference (HSD) test (P ≤ 0.05) within year as influenced by an interaction between year and herbicide strategy. Corn was grown in 2013, 2015, 2017, and 2019, and soybeans were grown 2014, 2016, and 2018.
Only monocot weed species richness was influenced by a herbicide strategy by year interaction, while total and dicot weed species richness were influenced by each factor individually (Table 3). Species richness decreased as more SOAs were implemented into herbicide programs, and the fully diversified strategy resulted in the lowest dicot weed species richness, supporting the in-field observations (Table 4). The soil seedbank assessments reflects what occurred within early-summer evaluations, as the dicamba-reliant treatment had higher total weed density compared with other treatments and was primarily composed of monocots.
Influence of herbicide strategy on total and dicot weed species richness in the soil seedbank averaged over 2013 to 2019.
a Means followed by the same letter with columns are not different according to Tukey’s honest significant difference (HSD) test (P ≤ 0.05).
Crop Yield
Differences in crop yield were not observed in this research. The results from this experiment suggest that over a 7-yr period negative impacts from weed shifts on yield are not likely.
This research is the first published report to evaluate weed community response to varying levels of dicamba selection pressure in dicamba-resistant soybean rotated with corn. Soil seedbank analysis supported the observations from early-summer evaluations. Both the total and monocot weed densities in the dicamba-reliant treatment had at least 43% more weeds than the next highest treatment. The total weed densities taken in early summer were 84% monocots, while the densities within the soil seedbank were only 56% monocots. Wilson et al. (Reference Wilson, Miller, Westra, Kniss, Stahlman, Wicks and Kachman2007) also reported that in-field densities of B. scoparia in a 6-yr study were supported by soil seedbank analysis, as both decreased due to varying herbicides strategies. We demonstrated that utilizing six or more SOAs and residual herbicides in both corn and soybean years reduced weed densities compared with three SOAs and using a residual herbicide only in corn years. Shergill et al. (Reference Shergill, Bish, Biggs and Bradley2017) found similar results, as residual herbicides applied before POST glyphosate applications resulted in a greater than 50% reduction in weed densities and a 250% increase in yield compared with glyphosate alone. However, reduction in weed densities and increase in yield were not observed in the first year, but after the 4th year.
Species richness within the dicamba-reliant treatment was higher for total, dicot, and monocots by 3.5, 2, and 1.3 species, respectively, compared with all other herbicide strategies at early-summer evaluations. The other three treatments did not differ from one another. A similar result occurred within the soil seedbank, as the dicamba-reliant treatment had 0.7 more species than the next highest treatment. A shift into more diverse weed species occurred due to fewer SOAs being implemented in both corn and soybean years, but more importantly due to the lack of a residual herbicide in years when soybean was grown. Shifts in weed seedbanks were expected due to the changes in herbicide management practices, as differences in arable weed seedbanks have been observed due to differences in production practices in both organic and conventional cropping systems (Rotchés-Ribalta et al. Reference Rotchés-Ribalta, Armengot, Mäder, Mayer and Sans2017). Ovejero et al. (Reference Ovejero, Soares, Oliveira, Fonseca, Berger, Soteres and Christoffoleti2013) and Tharp and Kells (Reference Tharp and Kells2002) have shown that the use of residual herbicides increases overall weed control. Jhala et al. (Reference Jhala, Sandell, Sarangi, Kruger and Knezevic2017) reported that soil-applied residual herbicides followed by residual POST applications provided 82% control of common waterhemp [Amaranthus rudis (Moq.) Sauer], while strategies without a PRE herbicide reported 45% control at harvest. Using sequential applications of soil-residual herbicides also reduced weed densities and weed species richness in the present study.
We demonstrated that farm operators need to utilize both multiple herbicide SOAs and sequential applications of residual herbicides to decrease the densities and species richness of weed communities in corn rotated with dicamba-resistant soybeans, which has not been previously evaluated.
Acknowledgments
This research received no specific grant from any funding agency or the commercial or not-for-profit sectors. No conflicts of interest have been declared.
