Improving the Level and Consistency of Glyphosate-Resistant Canada Fleabane (Erigeron canadensis L.) Control With Bromoxynil-Based Tankmixes Applied Preplant to Soybean

Glyphosate-resistant (GR) Canada fleabane (Erigeron canadensis L.) is a problematic weed in soybean. Bromoxynil-based tankmixes provide control of GR Canada fleabane in monocot crops; however, there is limited research on preplant (PP) applications in soybean. The objective of this study was to determine the best third tankmix partner with glyphosate plus bromoxynil applied PP to improve the level and consistency of GR Canada fleabane control in soybean. Four field trials were conducted over a two-year (2020, 2021) period. Glyphosate plus bromoxynil controlled GR Canada fleabane 52, 58, and 69% at 2, 4, and 8 weeks after application (WAA) respectively. The addition of saflufenacil to glyphosate plus bromoxynil improved GR Canada fleabane control 39% at 2 WAA and the addition of dicamba to glyphosate plus bromoxynil improved GR Canada fleabane control 39 and 29% at 4 and 8 WAA, respectively. The addition of bromoxynil to glyphosate plus 2,4-D ester or dicamba improved GR Canada fleabane control by 36 and 35%, respectively, at 2 WAA. The addition of bromoxynil to glyphosate plus tiafenacil or pyraflufen-ethyl/2,4-D improved GR Canada fleabane control by 19% at 8 WAA. The addition of bromoxynil to glyphosate plus metribuzin improved GR Canada fleabane control 31, 32, and 36% at 2, 4, and 8 WAA, respectively. Antagonism was observed between bromoxynil plus glyphosate plus tiafenacil or halauxifen-methyl 8 WAA, while the interaction between bromoxynil plus glyphosate plus pyraflufen-ethyl/2,4-D was considered synergistic. The addition of metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil or dicamba to glyphosate plus bromoxynil improved the consistency of GR Canada fleabane at 2, 4, and 8 WAA. This study concludes that the addition of saflufenacil or dicamba to glyphosate plus bromoxynil can improve the level and consistency of GR Canada fleabane control applied PP to soybean.


Introduction
Glyphosate-resistant (GR) Canada fleabane was the first documented dicot weed to evolve resistance to glyphosate (Heap, 2020;VanGessel, 2001). GR Canada fleabane can produce thousands of seeds (Bhowmik & Bekech, 1993;Davis et al., 2009) which can move up to 500 km in one dispersal event (Shields et al., 2006). In Ontario, GR Canada fleabane primarily emerges in the spring and in the fall; though, if temperature and soil moisture conditions are favourable, it has been reported to emerge every month of the year except January (Cici & Van Acker, 2009;Main et al., 2006). Seeds within the top 0.3 cm of the soil surface can be stimulated by light and have the greatest likelihood of germinating and emerging (Nandula et al., 2006); though, its ruderal nature allows it to germinate in a variety of landscapes (Weaver, 2001). GR Canada fleabane has been confirmed in 14 countries as of 2020 (Heap, 2020).
GR Canada fleabane interference can reduce soybean yield. GR Canada fleabane interference can result in up to 93% yield loss if no weed management strategies are used (Byker et al., 2013a). In most soybean systems, GR Canada fleabane should be managed before soybean emergence to prevent yield loss (Bruce & Kells, 1990). In conventional-tillage production systems, tillage can be used to manage GR Canada fleabane, but in no-tillage systems, herbicides must be used to manage GR Canada fleabane (Bruce & Kells, 1990). Postemergence (POST) herbicides provide minimal control of GR Canada fleabane in GR and non-GMO soybean (Byker et al., 2013b); therefore, preplant (PP) and preemergence (PRE) herbicides must be used for acceptable control.
Bromoxynil is a photosystem II (PSII)-inhibiting herbicide (Fedtke & Duke, 2005;Shaner, 2014). Bromoxynil is a contact herbicide (Bayer CropScience Inc., 2019) that can outcompete plastiquinone at the Q B -binding site on the D1 protein, which inhibits electron transport, uncouples oxidative phosphorylation, and results in the reactive oxygen species formation and lipid peroxidation (Shaner, 2014;United States Environmental Protection Agency, 1998). Bromoxynil applied POST is not registered for use in soybean since it can cause leaf necrosis and stunting, though the plant may recover with little effect on yield (Andersen et al., 1973;Wax et al., 1974). Bromoxynil is registered to control select annual broadleaf weeds in corn, small grains, alfalfa, carrot, garlic, and onion (Boydston & Al-Khatib, 1993;Bayer CropScience Inc., 2019). Bromoxynil is a contact herbicide that is generally used to control annual broadleaf weeds in monocot crops.
GR Canada fleabane is currently not listed as one of the weeds controlled on the brmoxynil label, but prior research with bromoxynil-based tankmixes in corn and winter wheat has demonstrated GR Canada fleabane control. Mahoney et al. (2017) and Metzger et al. (2019) observed 93 and 94% control, respectively of GR and multiple-herbicide-resistant (MHR) Canada fleabane with bromoxynil (280 g ai ha -1 ) plus atrazine (1500 g ai ha -1 ) applied POST in corn 8 WAA. Quinn et al. (2021) reported 98% GR Canada fleabane control with a premix of pyrasulfotole/bromoxynil (205 g ai ha -1 ) plus AMS (1 L ha -1 ) applied POST in winter wheat 8 WAA. Mahoney et al. (2016) observed bromoxynil (174 g ai ha -1 ) plus pyrasulfotole (31 g ai ha -1 ) applied POST controlled 97% of GR Canada fleabane in winter wheat at 8 WAA. Bromoxynil tankmixes control GR Canada fleabane in monocot crops, which suggests potential for preplant (PP) applications in soybean.

Experimental Methods
Four field experiments were conducted over a two-year period (2020, 2021) at sites with confirmed GR Canada fleabane in southwestern Ontario. A 2 × 8 factorial was organized in a randomized complete block design (RCBD) with 4 replications. Factor One was a control and bromoxynil, and Factor Two was a control, metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil, and dicamba. The plots were 2.25 m [3 glyphosate/dicamba-resistant (GDR) soybean rows at 0.75 m apart] by 8 m. The replicates were divided by a 2.0 m alley.
Herbicide treatments were made PP when majority of the Canada fleabane was 10 cm in height. Glyphosate (900 g ae ha -1 ) was added to each herbicide treatment applied PP. A CO 2 -pressurized backpack sprayer was calibrated to dispense 200 L ha -1 at 240 kPa. The sprayer had four, ULD 120-02 nozzles separated 50 cm apart. The spray width measured 2.0 m. Following PP herbicide applications, a GDR soybean cultivar (DKB12-16) was seeded to about 400,000 seeds ha -1 at a 3.75 cm depth. A glyphosate (450 g ae ha -1 ) cover spray was applied POST to eliminate the effects of other weed species and glyphosate-susceptible Canada fleabane. Site, year, location, soil, and agronomic details are listed in Table 1. Information pertaining to Canada fleabane including height and density at application and the resistance profile of the populations is presented in Table 2. The herbicide treatments that were used in this study are described in Table 3.  Note. a MSO was add to all treatments with halauxifen-methyl (1.0 % v/v) or tiafenacil (0.5 % v/v). b Merge (1 L ha -1 ) was added to treatments with saflufenacil.
Canada fleabane control was assessed as a visual estimate of the reduction in Canada fleabane biomass using a 0 to 100 scale at 2, 4, and 8 weeks after application (WAA). A visual rating of 0 suggested no GR Canada fleabane control and a visual rating of 100 suggested plant death (Canadian Weed Science Society, 2018). Canada fleabane density and biomass were collected 8 WAA. Density was collected by counting Canada fleabane plants in two, 0.25 m 2 quadrats placed randomly in the front and back of each plot. Biomass was collected by cutting the Canada fleabane where the stem meets the soil surface within each quadrat, placing the samples in separate bags for each individual plot, drying the samples to a constant moisture, and then weighing the samples.
A 0 to 100 scale was used to assess visible GDR soybean injury 2 and 4 weeks after soybean emergence (WAE); 0 suggested no soybean injury and 100 suggested soybean death (Canadian Weed Science Society, 2018). At soybean harvest maturity, two GDR soybean rows were harvested per plot using a small plot harvester. Soybean yield and moisture were recorded. Prior to analysis, the grain yield was corrected to a 13.5% moisture content.

Statistical Analysis
SAS 9.4 was the statistical software used in this study. The GLIMMIX procedure was used for the analysis. Herbicide treatment represented the fixed effect and block, block within location, and treatment by location represented the random effects. The Shapiro-Wilk test and the residual plots were utilized when assessing normality. Since no treatment by location interaction was found, the sites were pooled together for the analysis. An arcsine square root back transformation was used to meet normality assumptions for GR Canada fleabane jas.ccsenet.org Journal of Agricultural Science Vol. 14, No. 1; 2022 control at 2, 4, and 8 WAA. Soybean yield data was assessed using the default link identity. A lognormal distribution using the link identity was used to assess GR Canada fleabane density and biomass; the omega procedure was used to back-transform the means from the analysis format (M. Edwards, Ontario Agricultural College Statistician, University of Guelph, personal communication). Mean separations were completed using the Tukey-Kramer's multiple range test (p < 0.10). Significance was indicated by different letter codes within (1) The expected GR Canada fleabane biomass and density was calculated using a modified Colby's equation. The 'W' in the equation represents the non-treated control mean.
A t-test was used to analyze the expected versus the observed values, significance was indicated when p < 0.10. The interactions were considered additive when the observed control value was the same as the expected control value. The interaction was considered synergistic or antagonistic when the observed control values were greater than or less than the expected control values, respectively (Colby, 1967). Conversely, if the observed density or biomass values were greater than or less than the expected density or biomass values then the interaction was considered antagonistic or synergistic, respectively (Colby, 1967). The coefficient of variation was calculated to ascertain the consistency of GR Canada fleabane control for each least square mean estimate.

Soybean Injury
Soybean injury was ≤ 10% at all sites across both years (data not presented).

Glyphosate-Resistant Canada Fleabane Control
The simple effects are presented in Table 4 since the interaction between Factor One (control and bromoxynil) and Factor Two (control, metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil, and dicamba) for GR Canada fleabane control 2, 4, and 8 WAA was statistically significant (p < 0.10).     Note. Abbreviations: SE, standard error. Means accompanied by different letters in a column (a-c) or row (Z, Y) differ significantly based on Tukey Kramer's LSD (P < 0.10). ** Significant at P < 0.01; * significant at P < 0.10 based on a t-test conducted for observed versus expected values. a Values in parentheses represent the expected values from Colby's analysis. b Merge, 1 L ha -1 was added to all treatments with saflufenacil. c MSO, 1 % v/v or 0.5 % v/v was added to all treatments with halauxifen-methyl and tiafenacil, respectively.

Glyphosate-Resistant Canada Fleabane Consistency in Control
The coefficient of variation (CV) is a standardized measure of the standard deviation to the mean and is useful for investigating data variability (Shechtman, 2013). The CV for each least square mean was used as an indicator to demonstrate the consistency of GR Canada fleabane control (Table 5). When comparing CVs, a lower CV would indicate greater consistency in control (Shechtman, 2013).   Note. a Merge, 1 L ha -1 was included in all treatments with saflufenacil. b MSO, 1 % v/v was included in all treatments with halauxifen-methyl and tiafenacil.
At 2, 4, and 8 WAA, the addition of metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil or dicamba to glyphosate plus bromoxynil improved the consistency of GR Canada fleabane control, as indicated by lower CVs compared to the CV of glyphosate plus bromoxynil. When bromoxynil was added to glyphosate plus metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil or dicamba the consistency of GR Canada fleabane control was improved as indicated by lower CVs except for glyphosate plus bromoxynil plus halauxifen-methyl at 4 and 8 WAA, which did not improve the consistency of control. It has been suggested that a third herbicide into the tankmix improves the consistency of GR Canada fleabane control (Mellendorf et al., 2013). Soltani et al. (2020b) reported improved consistency of GR Canada fleabane control when a third herbicide, such as metribuzin (400 g ai ha -1 ), was added to glyphosate (900 g ae ha -1 ) plus 2,4-D ester (500 g ae ha -1 ) or saflufenacil (25 g ai ha -1 ) 8 WAA. Budd et al. (2016) observed an improvement in the consistency of GR Canada fleabane control when metribuzin (400 g ai ha -1 ) was added to glyphosate (900 g ae ha -1 ) plus saflufenacil (25 g ai ha -1 ).
In conclusion, limited research exists on GR Canada fleabane control with bromoxynil-based tankmixtures applied PP to soybean. Glyphosate plus bromoxynil plus metribuzin, tiafenacil, pyraflufen-ethyl/2,4-D, 2,4-D ester, halauxifen-methyl, saflufenacil or dicamba caused ≤ 10% visible soybean injury. At 2 WAA, when saflufenacil was added to glyphosate plus bromoxynil GR Canada fleabane control improved from 52 to 85%. At 4 WAA, when dicamba was added to glyphosate plus bromoxynil GR Canada fleabane control improved from 58 to 95%. At 8 WAA, when dicamba was added to glyphosate plus bromoxynil GR Canada fleabane control improved from 69 to 97%. At 2 WAA, the interaction between glyphosate plus bromoxynil plus tiafenacil, pyraflufen-ethyl/2,4-D or halauxifen-methyl was antagonistic. At 4 WAA, the interaction between glyphosate plus bromoxynil plus pyraflufen-ethyl/2,4-D, 2,4-D ester or halauxifen-methyl was antagonistic. At 8 WAA, the interaction between glyphosate plus bromoxynil plus tiafenacil or halauxifen-methyl was antagonistic; while the interaction between glyphosate plus bromoxynil plus pyraflufen-ethyl/2,4-D was synergistic. The interactions for the other tankmixtures were additive. In conclusion, the addition of saflufenacil to glyphosate plus bromoxynil at 2 WAA and dicamba to glyphosate plus bromoxynil at 4 and 8 WAA improves the level and consistency of GR Canada fleabane control applied PP to soybean.