Evaluating cover crop tolerance to corn residual herbicides using field-treated soil in greenhouse bioassay

Abstract

More growers across the U.S. Midwest are considering interseeding or overseeding cover crops into corn for soil health purposes. One challenge of this practice is the potential injury from soil residual herbicides applied preemergence (PRE) for weed control in corn to the interseeded and overseeded cover crop species. Field-treated soil was collected in 2021 and 2022 at Janesville, WI, and Lancaster, WI, to investigate the impact of PRE residual herbicides on establishment of interseeded and overseeded cover crops via greenhouse bioassay. Soil samples (0 to 5 cm depth) were collected from field experiments at 0, 10, 20, 30, 40, 50, 60, and 70 days after treatment (DAT). Treatments consisted of 14 single and multiple sites of action (SOAs) PRE herbicides plus a nontreated check (NTC). Four bioindicator cover crop species were used in the greenhouse bioassay: annual ryegrass, cereal rye, radish, and red clover. Cover crop biomass was collected 28 d after bioassay seeding. Cover crop species responded differently across herbicide treatments. Annual ryegrass and cereal rye were sensitive to treatments containing herbicide Group 15, whereas Groups 2, 4, 5, 14, and 27 had minimal impact on their establishment when field soil was collected at 30 DAT (interseeding scenario) and 70 DAT (overseeding scenario) compared to the NTC. Radish and red clover were sensitive to herbicide Groups 2, 4, and 27, whereas Groups 5, 14, and 15 had minimal impact on their establishment. Annual ryegrass, radish, and red clover were more sensitive to PRE herbicides containing two and three SOAs than to herbicides with a single SOA. On the basis of these greenhouse bioassay results, cover crop species should be carefully selected depending on the soil residual herbicide when interseeded and overseeded into corn. Field studies will be conducted to validate these results and support recommendations to growers interested in this system.

Introduction

The adoption of cover crops increases the crop diversity in continuous corn and corn–soybean [Glycine max (L.) Merr.] rotations across the U.S. Midwest (Brooker et al. 2020a). Cover crops can provide various benefits, including reduced soil erosion, improved water infiltration, enhanced nutrient cycling, and weed and insect pest suppression (Grint et al. 2022; Schipanski et al. 2014; Wallander et al. 2021). Although only 2% of agricultural hectares in the United States were sown with cover crops in 2017, the increase in cover crop adoption is promising, with a 50% increase in cover cropping from 2012 to 2019 (USDA-NASS 2019; Wallander et al. 2021). In Wisconsin, cover crops were established in 6% of the 3.7 million ha of cropland in 2017 (USDA-NASS 2019). One of the main challenges for successful cover crop establishment in corn cropping systems in the Upper Midwest is the short growing season (lack of degree days) for sowing and establishing cover crops following corn grain harvest (Kladivko et al. 2014; Smith et al. 2019).

In a continuous corn and corn–soybean rotation, cover crop species selection is typically limited to winter cereals, such as cereal rye. Legume species like crimson clover (Trifolium incarnatum L.) and field pea (Pisum sativum L.), as well as brassica species like radish and turnips (Brassica rapa L.), perform poorly when established late after corn grain harvest because of low temperatures (Curran et al. 2018; Noland et al. 2018; Rusch et al. 2020; Singer 2008). Interseeding or overseeding cover crops while the primary crop is still in the field increases growing season length and cover crop biomass potential relative to cover crop planted after harvest, enhancing the ecosystem benefits of cover cropping in corn production systems (Adler and Nelson 2020; Caswell et al. 2019). Herein, interseeding is defined as planting a cover crop early in the growing season when the corn is between the V3 and V8 vegetative growth stages (Smith et al. 2019; Youngerman et al. 2018). In contrast, cover crop overseeding is typically done by aerial seeding just before or at crop physiological maturity (Kladivko et al. 2014). These systems provide winter-sensitive legume cover crop species, such as crimson clover (Peterson et al. 2021; Youngerman et al. 2018) and red clover (Wallace et al. 2017), and brassica species, such as radish and turnips, a wider growing window before the winter (Brooker et al. 2020b).

A concern with interseeding and overseeding is whether soil residual herbicides applied for weed control will injure the cover crops (Adler and Nelson 2020; Brooker et al. 2020b). Researchers have investigated the impact of soil residual herbicides on interseeded cover crops into the V3 to V6 corn growth stages and reported injury depending on the active ingredient and cover crop species. In one interseeding study established at the V5 corn growth stage and conducted in Pennsylvania, annual ryegrass biomass was reduced >80% with pyroxasulfone and S-metolachlor applications, and red clover biomass was reduced >80% with mesotrione, compared to the nontreated control (Wallace et al. 2017). Brooker et al. (2020b) reported that Group 15 herbicides (acetochlor, dimethenamid-P, and pyroxasulfone) reduced annual ryegrass stand >60% at V3 and V6 interseeding timings; Group 2 herbicides (flumetsulam and rimsulfuron) reduced radish stand >70% compared to the nontreated control. The same authors described that cover crops can be interseeded into corn over the V3 and V6 stages but that species selection and herbicide label restrictions should be carefully considered. Thus additional studies are warranted to evaluate response of multiple cover crop species to soil residual herbicides under different soil types and environmental conditions, which are critical components influencing cover crop establishment, herbicide residual activity in the soil, and their interactions (Cornelius and Bradley 2017; Jursík et al. 2020).

Few studies have reported the potential herbicide residual injury to cover crops interseeded at the V3 to V5 corn growth stages and overseeded at the V10 to VT corn growth stages. More research is needed to support herbicide selection that provides effective weed control yet allows establishment of cover crops for growers adopting the interseeding and overseeding systems. Herein the tolerance of four common cover crop species (annual ryegrass, cereal rye, radish, and red clover) to a comprehensive list of labeled corn residual preemergence (PRE) herbicides is evaluated. The main purpose of this study was to investigate potential soil residual herbicide and cover crop combination options for interseeding (∼V3 to V5 corn growth stages) and overseeding (∼V10 to VT corn growth stages) scenarios via greenhouse bioassay.

Materials and Methods

Field Experiment Information

A field experiment was conducted in 2021 and 2022 at the Rock County Farm, Janesville, WI (42.43°N, 89.01°W), and the University of Wisconsin–Madison Lancaster Agricultural Research Station, Lancaster, WI (42.83°N, 90.76°W), to evaluate weed control in corn with multiple soil residual herbicides applied PRE. (For more information about the field experiment and weed control results, see Silva et al. [2023].) Briefly, soil properties, corn hybrids, and seeding rates for each location are summarized in Table 1. The Janesville-2021 and Lancaster-2021 fields had no history of residual herbicide application in the previous season. Authority® First (sulfentrazone [280 g ai ha⁻¹] + cloransulam-methyl [35 g ai ha⁻¹]) (FMC, Philadelphia, PA, USA) was applied PRE in the previous season for the Janesville-2022 field. Sequence® (glyphosate [800 g ai ha⁻¹] + S-metolachlor [1,100 g ai ha⁻¹]) (Sygenta, Greensboro, NC, USA) was applied at the V2 soybean growth stage in the previous season for the Lancaster-2022 field. Monthly average air temperature and accumulated precipitation during the data collection period were obtained from onsite weather stations (WatchDog® 2700, Spectrum Technologies, Aurora, IL, USA) and are summarized in Table 2.

Table 1. Soil properties, corn hybrids, and seeding rates for corn field experiments at Janesville, WI, and Lancaster, WI, 2021 and 2022. ^a,b

^a The experimental areas were managed in a soybean–corn rotation; thus soybean was grown in the previous growing season before the experiment establishment at all experimental sites. Before corn planting, the experimental area was tilled using a field cultivator. Corn was planted 5 cm deep and in 76-cm row spacing at all experimental sites.

^b Abbreviation: OM, organic matter.

^c The Janesville-2021 experimental field was fertilized with 200 kg ha⁻¹ of nitrogen (46-0-0); Lancaster-2021 with 128 kg ha⁻¹ of nitrogen (46-0-0); Janesville-2022 with 112 kg ha⁻¹ of nitrogen (32-0-0) and 32 kg ha⁻¹ of sulfur in the form of ammonium thiosulfate (12-0-0-26S); Lancaster-2022 with 55 kg ha⁻¹ of phosphorus + 112 kg ha⁻¹ of potassium nitrate (4-19-38) applied early spring and 160 kg ha⁻¹ of nitrogen (46-0-0).

Table 2. Monthly average air temperature and total precipitation from April through July at Rock County Farm, Janesville, WI, and Lancaster Agricultural Research Station, Lancaster, WI, in 2021 and 2022 and during the past 30 years. ^a,b

^a Weather data for 2021 and 2022 were obtained from onsite weather stations. The 30-yr average monthly weather data were obtained from the Wisconsin State Climatology Office (∼https://www.aos.wisc.edu/∼sco/clim-history/acis_stn_meta_wi_index.htm).

The field experiment was conducted in a randomized complete-block design with four replications. The experimental units were 3 m wide (4 corn rows) × 9 m long. The treatments consisted of 14 soil residual herbicides applied PRE plus a nontreated check (NTC; Table 3). Herbicides were applied within a day after corn planting (Table 4) using a CO₂ pressurized backpack sprayer equipped with six TeeJet® TTI110015 flat-fan nozzles (TeeJet®, Springfield, IL, USA) spaced 50.8 cm apart at a boom height of 50 cm from the soil surface. The sprayer was calibrated to deliver 140 L ha⁻¹ of spray solution at 241 kPa at a speed of 4.8 km h⁻¹.

Table 3. PRE herbicides, site of action groups, trade names, manufacturers, chemical families, half-lives, and rates used in the corn field experiments. ^a

^a Abbreviation: SOA, site of action.

^b Average field half-life. Data are from Shaner (2014) and the Pesticide Properties Database (http://sitem.herts.ac.uk/aeru/ppdb/en/index.htm).

Table 4. Corn planting and herbicide application dates for each site-year and corn growth stage for each collection date. ^a

^a Abbreviation: DAT, days after treatment.

Soil samples (0 to 5 cm depth) were collected from 14 residual PRE herbicide treatments (including herbicides with single and multiple sites of action [SOAs]) plus the NTC (Table 3) from field experiments conducted at the four site-years. Soil samples were collected at 0, 10, 20, 30, 40, 50, 60, and 70 d after treatment (DAT) to evaluate cover crops’ response to herbicide residuals over time. A handheld 6-cm-diameter soil sampler (Fiskars, Middleton, WI, USA) was used to collect the soil samples. At each sampling time, six soil cores were collected from each plot adjacent to the two central corn rows, combined, and placed in a plastic bag (∼1,000 g). Soil samples were stored in a freezer (−20 C) until the onset of the greenhouse bioassay experiment (approximately 4 mo after the first collection date). The corn growth stage at each soil sampling time was recorded according to Broeske and Lauer (2020; Table 4). No herbicides were applied to the field experiments other than the PRE herbicides evaluated.

Greenhouse Bioassays Using Cover Crops

In the fall of each year, the treated soil samples were used in greenhouse bioassays; that is, in fall 2021, greenhouse bioassays were conducted with the soil samples collected from the Janesville-2021 and Lancaster-2021 field experiments; in fall 2022, greenhouse bioassays were conducted with the soil samples collected from the Janesville-2022 and Lancaster-2022 field experiments. The bioassays were conducted in the Walnut Street Greenhouse at the University of Wisconsin–Madison. Each bioassay experimental unit consisted of a 210-cm³ seed tray cell (6 cm length × 6 cm width × 5.9 cm depth; 804 Series, T.O. Plastics, Clearwater, MN, USA). The four field soil samples from each treatment within a site-year and sampling time were thawed and combined across replications (creating a composite sample). The composite sample was thoroughly mixed by hand to ensure uniform distribution of the herbicide among the eight resulting replicates (treated as four replications and two experimental runs). Experimental units (seed tray cells) were then filled with the respective mixed soil sample.

Annual ryegrass, cereal rye, radish, and red clover (La Crosse Seed, La Crosse, WI, USA) were used as bioindicator species. These species are among the most commonly adopted cover crops across cropping systems in the United States (USDA-SARE 2020) and have been successfully interseeded in Wisconsin corn systems (Smith and Ruark 2022). Germination tests were conducted before setting up the bioassay experiment to investigate seed viability. Seeds were sown in pots filled with soil (four replicates of 25 seeds each), and at 10 d after sowing (DAS), the germinated seedlings were counted. The average percentage of germination for both years was 93%, 85%, 96%, and 94% for annual ryegrass, cereal rye, radish, and red clover, respectively. A preliminary experiment in additive series (Freckleton and Watkinson 2000; Galon et al. 2018) was also conducted in 2021 to determine the cover crop plant density for each species. The cover crop densities evaluated were 1, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, and 22 plants cell⁻¹, which corresponded to 17, 34, 67, 101, 134, 168, 201, 235, 268, 302, 335, and 369 plants m⁻². At 28 DAS, the aboveground biomass of the plants was harvested and dried at 60 C until constant dry biomass was obtained. The constant biomass production was obtained with a density of 8 plants cell⁻¹ for cereal rye and radish (134 plants m⁻²), 10 plants cell⁻¹ for annual ryegrass (168 plants m⁻²), and 18 plants cell⁻¹ for red clover (302 plants m⁻²) (data not shown). For the preliminary study and for the bioassay experiment, each cover crop species was grown in a separate experimental unit (Figure 1).

Figure 1. Cover crop species 14 d after sowing in each experimental unit (left). From bottom to top, the units represent days after treatment in the field from 0 to 70, and from left to right, the units represent the different treatments, starting with the nontreated check. Photos provide a closer view of the cover crops at 14 d (top right) and 28 d (bottom right) after sowing. The experimental units at the top left and bottom left represent radish and red clover, respectively, whereas the experimental units at the top right and bottom right represent cereal rye and annual ryegrass, respectively.

The greenhouse bioassay experiment was conducted as a completely randomized design with four replications. The experiment was repeated in time (two experimental runs) for each PRE herbicide treatment over sampling time and site-year. In 2021, the greenhouses were maintained at 28/25 C day/night temperature and 55% relative humidity. In 2022, the greenhouses were maintained at 24/21 C day/night temperature and 60% relative humidity. The slight differences in day/night temperature and relative humidity in the greenhouses between the two experimental years were because of external fall weather conditions (the greenhouse bioassay experiments were established on September 17, 2021, and September 8, 2022). Greenhouse conditions for both years were set to a 16/8-h day/night photoperiod using high-pressure sodium lightbulbs (400 W) to supplement the natural light. The greenhouse environmental conditions were monitored throughout the experiment using a WatchDog® A150 logger (Spectrum Technologies). Bioassays were watered twice a day and fertigated weekly using 20-10-20 water-soluble fertilizer (Peters® Professional; ICL Fertilizers, Dublin, OH, USA) providing 300 ppm of nitrogen and potassium and 150 ppm of phosphorus. At 28 DAS, bioassay cover crop injury was assessed using a scale of 0% to 100%, where 0% was no visible injury and 100% was complete plant necrosis. The aboveground biomass of indicator cover crop species growing in each tray cell (g pot⁻¹) was harvested, bagged, forced-air dried at 60 C for at least 7 d, and then weighed.

Statistical Analyses

All statistical analyses were performed using R Version 4.2.2 (R Development Core Team 2022). A linear correlation between bioassay injury and aboveground biomass production was performed using Pearson’s analysis (stat_cor function, ggpubr package; Kassambara 2022). Jitter violin plots combined with box plots were generated for annual ryegrass, cereal rye, radish, and red clover data to show the variance of the biomass values combining all treatments over site-years and sampling times.

Analysis of variance (ANOVA) was performed to compare different PRE herbicide treatments within each sampling time (0, 10, 20, 30, 40, 50, 60, and 70 DAT) for each bioindicator cover crop species instead of building multiple response curves over time. ANOVA provided more meaningful results when compared to regression models (data not shown). Bioassay aboveground biomass data for each cover crop species and sampling time were combined over site-years and over the two experimental runs in the greenhouse and analyzed with linear mixed-effect models using the function lmer from the lme4 package (Bates et al. 2015). Square root transformation models were used when fitting the bioassay biomass data to meet the assumptions of normality and homogeneity of variance of residuals for each cover crop species. Back-transformed means are reported in the results. PRE herbicide treatments were included as a fixed effect in the model; greenhouse bioassay experimental run and site-year and experimental run nested within site-year were considered random effects. Models were analyzed using ANOVA (anova function, car package; Fox and Weisberg 2019), and means were separated using Fisher’s least significant difference (LSD) test (emmeans package; Lenth 2022) when a treatment effect was significant (P ≤ 0.05).

The biomass response by PRE herbicide treatment on each cover crop species across soil sampling time and site-year was used to calculate the area under biomass stairs (AUBS). The AUBS was estimated using the audps function of the agricolae package (Mendiburu 2022). The AUBS referred to herein is an adaptation from the area under the disease progress stairs (AUDPS) commonly used in plant pathology to estimate disease progress over time (Simko and Piepho 2012). The AUDPS has also been adopted to estimate crop injury from postemergence (POST) herbicides over distance (Striegel et al. 2021) and herbicide impact on biomass bioindicator species (Ribeiro et al. 2021). The AUDPS (herein after called AUBS) concept applied to our bioassay data resulted in one value to estimate the impact of each residual herbicide applied PRE on the biomass of each cover crop species over sampling time. AUBS corresponds to the area under the step function considering adjusted weight for the first and last DAT. For instance, each biomass value was multiplied by 10 (interval between soil sampling), and the first and last assessment weights were extrapolated in the missing direction using half of the average interval duration between DAT observations (Simko and Piepho 2012). The higher the AUBS value that was obtained, the lower was the PRE herbicide injury on cover crops (Figure 2).

Figure 2. Graphical example of clopyralid + acetochlor + mesotrione herbicide effect on annual ryegrass aboveground biomass production (28 d after sowing the greenhouse bioassay) as a function of days after treatment in the field as calculated by the area under biomass stairs (AUBS). D is cover crop biomass, and n is the interval between days after treatment. The AUBS value was obtained from the simplified equation ${\rm{AUBS}} = \bar y \times \;n$ , where $\bar y$ is the arithmetic mean of all cover crop biomass assessments.

AUBS estimated values for each cover crop species by PRE herbicide treatment were submitted to ANOVA using a linear mixed-effect model following the previously described approaches for biomass data. AUBS values were also estimated for each cover crop species and combined across species by the number of herbicide sites of action of each PRE treatment (one, two, or three SOAs; Table 3). PRE herbicide SOAs were included as a fixed effect in the model. Experimental run and site-year and experimental run nested within site-year were fitted as random effects for each cover crop and all cover crops pooled together. Models were analyzed using ANOVA (anova function, car package; Fox and Weisberg 2019), and means were separated using Fisher’s LSD test (emmeans package; Lenth 2022) when a treatment effect was significant (P ≤ 0.05).

Results and Discussion

Cover Crop Response

The average relative cover crop growth at 28 DAS in the NTC was 0.40 g pot⁻¹ for annual ryegrass, 0.68 g pot⁻¹ for cereal rye, 1.71 g pot⁻¹ for radish, and 0.23 g pot⁻¹ for red clover. Soil residual herbicides applied PRE, measured via greenhouse bioassay 28 DAS, affected cover crop biomass for each field soil sampling time (0, 10, 20, 30, 40, 50, 60, and 70 DAT) (P < 0.01). Pearson’s analysis showed negative correlations between visual injury rating values and biomass production (g pot⁻¹) for annual ryegrass, cereal rye, radish, and red clover at 28 DAS (Figure 3 A). The slope of the regression lines was R = −0.73, R = −0.47, R = −0.72, and R = −0.71 with P < 0.001 for annual ryegrass, cereal rye, radish, and red clover, respectively (Figure 3 A). These results indicate that visual injury rating is associated with cover crop biomass production and suggest that higher visual injury occurred when lower biomass was produced; therefore only the biomass data were considered for fitting the linear mixed-effect models and calculating AUBS values. Jitter violin plots combined with box plots showed the distribution and changes of the biomass values for each cover crop species, including all PRE herbicide treatments and the NTC at all sampling times (Figure 3 B). In general, high biomass values (more jitter points distributed above zero) were observed for cereal rye and radish. A similar shape of violin plots was observed for annual ryegrass and red clover, with a wider base and a high frequency of observation close to zero. This indicates that cereal rye and radish tended to be more tolerant than annual ryegrass and red clover to the residual herbicides applied PRE evaluated herein.

Figure 3. Pearson’s linear correlation between herbicide injury (%) and aboveground biomass (g pot⁻¹) for annual ryegrass, cereal rye, radish, and red clover (A). The black solid lines show the linear trend, and the gray shaded areas represent 95% confidence intervals. Violin plots and box plots representing the aboveground biomass distribution (g pot⁻¹) combined for all treatments and sampling time of each cover crop species (B).

Herein we focus the discussion on the results from the field soil samples collected 30 and 70 DAT, but the complete results are also available in Tables 5 to 8. Using the cover crop greenhouse bioassay data (28 DAS), we assumed a situation of interseeding at 30 DAT (∼V3 to V5 corn growth stages) and overseeding at 70 DAT (∼V10 to VT corn growth stages; Table 4). This decision was taken considering that cover crop interseeding (planted during the vegetative corn growth stage) in Wisconsin can be successful between the V3 and V7 corn growth stages (Smith and Ruark 2022), and the overseeding is adopted just before or at corn maturity (Adler and Nelson 2020; Kladivko et al. 2014). The methodology adopted herein allowed us to evaluate the impact of the soil residual herbicide on the cover crop establishment in the absence of the crop canopy, which can also impact cover crop establishment (Ribeiro et al. 2021; Schmitt et al. 2021).

Table 5. Effect of PRE herbicides on annual ryegrass biomass production at each sampling time and area under biomass stairs estimated for annual ryegrass biomass production by PRE herbicide over time in greenhouse bioassay using field-treated soil from Janesville, WI, and Lancaster, WI, in 2021 and 2022. ^a,b

^a Abbreviations: ACET, acetochlor; ATZ, atrazine; AUBS, area under biomass stairs; BIP, bicyclopyrone; CLOP, clopyralid; DAT, days after treatment; DIM-P, dimethenamid-P; FLUM, flumetsulam; IFT, isoxaflutole; LSD, least significant difference; MES, mesotrione; SAFL, saflufenacil; S-MET, S-metolachlor; SMZ, simazine.

^b Means within a column with the same letter are not significantly different according to Fisher’s LSD test at P ≤ 0.05.

^c Aboveground biomass 28 d after sowing the greenhouse bioassay.

Interseeding and Overseeding Annual Ryegrass Scenario

Annual rye biomass 28 DAS in the interseeding scenario (field soil samples collected at 30 DAT; ∼V3 to V5 corn growth stages; Table 4) was reduced by most soil residual herbicides applied PRE compared to the NTC (Table 5). Atrazine + S-metolachlor, S-metolachlor, S-metolachlor + bicyclopyrone + mesotrione, atrazine + S-metolachlor + bicyclopyrone + mesotrione, atrazine + acetochlor, acetochlor, and saflufenacil + dimethenamid-P had the most detrimental impact on annual ryegrass (0.005 to 0.097 g pot⁻¹) when compared to the NTC (0.419 g pot⁻¹) (Table 5). Clopyralid + acetochlor + mesotrione, acetochlor + mesotrione, and flumetsulam + clopyralid + acetochlor resulted an intermediate negative impact on annual ryegrass (0.172 to 0.342 g pot⁻¹). Only mesotrione, flumetsulam + clopyralid, atrazine, and simazine did not impact annual ryegrass biomass (0.369 to 0.525 g pot⁻¹) compared to the NTC (Table 5).

These results suggest that applying S-metolachlor, acetochlor, and premixes containing Group 15 herbicides is likely to impact annual ryegrass biomass at V3 to V5 corn. Group 15 herbicides are recommended for controlling grass weed species, and S-metolachlor also has an extended half-life, which might result in greater persistence (Shaner 2014) and consequently more risk for annual ryegrass injury. A previous field study reported stand reduction >60% of annual ryegrass interseeded at the V3 to V6 corn growth stages following PRE application of Group 15 herbicides (S-metolachlor, acetochlor, and dimethenamid-P) (Brooker et al. 2020b). Wallace et al. (2017) also observed unacceptable levels of annual ryegrass biomass reduction (>75%) for S-metolachlor (1,790 g ai ha⁻¹) when applied PRE at the V5 corn stage. However, unlike our results, Wallace et al. (2017), in a field study, found that dimethenamid-P (840 g ai ha⁻¹) and acetochlor (1,960 g ai ha⁻¹) applied PRE at standard label rates resulted in less than 20% of annual ryegrass biomass reduction at the V5 stage, which was suggested to be an acceptable level to farmers integrating weed control and soil conservation benefits. Stanton and Haramoto (2019), in a field experiment in Kentucky, reported that saflufenacil (70 g ai ha⁻¹) + dimethenamid-P (560 g ai ha⁻¹) did not reduce initial annual ryegrass density (137 plants m⁻²) 3 wk after interseeding compared to the NTC (196 plants m⁻²); however, the herbicide rate applied was slightly lower compared to the rates applied in our current study (saflufenacil [75 g ai ha⁻¹] + dimethenamid-P [655 g ai ha⁻¹]; Table 3).

For the overseeding scenario (field soil samples collected at 70 DAT; ∼V10 to VT corn growth stages; Table 4), the most injurious PRE herbicides on annual ryegrass 28 DAS were S-metolachlor, S-metolachlor + bicyclopyrone + mesotrione, atrazine + S-metolachlor, and atrazine + S-metolachlor + bicyclopyrone + mesotrione (0.057 to 0.096 g pot⁻¹) compared to the NTC (0.472 g pot⁻¹). Saflufenacil + dimethenamid-P, atrazine + acetochlor, acetochlor, clopyralid + acetochlor + mesotrione, and flumetsulam + clopyralid + acetochlor caused intermediate impacts on annual ryegrass biomass (0.259 to 0.365 g pot⁻¹). Annual rye was not injured by flumetsulam + clopyralid, simazine, atrazine, or mesotrione (0.397 to 0.424 g pot⁻¹) compared to the NTC (Table 5). Validating these results, the AUBS analysis showed that S-metolachlor (2.07), atrazine + S-metolachlor (2.22), S-metolachlor + bicyclopyrone + mesotrione (3.23), and atrazine + S-metolachlor + bicyclopyrone + mesotrione (4.83) provided the lowest AUBS values, which means that these herbicides caused the highest injury to annual ryegrass throughout the soil sampling period. The highest AUBS values were observed for atrazine (32.48), flumetsulam + clopyralid (40.99), and mesotrione (42.08), soil residual herbicides applied PRE that did not injure annual ryegrass compared to the NTC (33.43).

Interseeding and Overseeding Cereal Rye Scenario

High levels of tolerance to soil residual herbicides applied PRE were observed 28 DAS for cereal rye in the interseeding scenario (field soil samples collected at 30 DAT; ∼V3 to V5 corn growth stages; Tables 4 and 6). Cereal rye biomass was not reduced by clopyralid + acetochlor + mesotrione, saflufenacil + dimethenamid-P, atrazine, acetochlor + mesotrione, mesotrione, and flumetsulam + clopyralid + acetochlor (0.652–0.857 g pot^-1) compared to the NTC (0.769 g pot^-1). Simazine, acetochlor, S-metolachlor + bicyclopyrone + mesotrione, and atrazine + S-metolachlor + bicyclopyrone + mesotrione resulted in intermediate injurious (0.559–0.595 g pot^-1) compared to the NTC. S-metolachlor (0.358 g pot^-1), atrazine + S-metolachlor (0.401 g pot^-1), and atrazine + acetochlor (0.476 g pot^-1) were the most injurious PRE herbicides (Table 6).

Table 6. Effect of PRE herbicides on cereal rye biomass production at each sampling time and area under biomass stairs estimated for cereal rye biomass production by PRE herbicide over time in greenhouse bioassays using field-treated soil from Janesville, WI, and Lancaster, WI, in 2021 and 2022. ^a,b

^a Abbreviations: ACET, acetochlor; ATZ, atrazine; AUBS, area under biomass stairs; BIP, bicyclopyrone; CLOP, clopyralid; DAT, days after treatment; DIM-P, dimethenamid-P; FLUM, flumetsulam; IFT, isoxaflutole; LSD, least significant difference; MES, mesotrione; SAFL, saflufenacil; S-MET, S-metolachlor; SMZ, simazine.

^b Means within a column with the same letter are not significantly different according to Fisher’s LSD test at P ≤ 0.05.

^c Aboveground biomass 28 d after sowing the greenhouse bioassay.

For the overseeding scenario (field samples collected at 70 DAT; ∼V10 to VT corn growth stages; Table 4), none of the PRE herbicides tested herein negatively impacted cereal rye biomass (0.495 to 0.666 g pot⁻¹) compared to the NTC (0.530 g pot⁻¹; Table 6). The AUBS findings support the high cereal rye tolerance observed for biomass values (Table 6). Atrazine (51.85), mesotrione (61.71), acetochlor + mesotrione (54.56), saflufenacil + dimethenamid-P (56.62), flumetsulam + clopyralid (63.11), flumetsulam + clopyralid + acetochlor (61.13), and clopyralid + acetochlor + mesotrione (54.75) did not negatively impact cereal rye compared to the NTC (55.38). A previous study also reported that saflufenacil + dimethenamid-P did not injure cereal rye interseeded at 30 DAT compared to the NTC (Smith 2015). Palhano et al. (2018), in field research, described 11% of fall-seeded cereal rye emergence reduction following POST application of mesotrione (Group 27).

Interseeding and Overseeding Radish Scenario

Radish biomass 28 DAS in the interseeding scenario (field soil samples collected at 30 DAT; ∼V3 to V5 corn growth stages; Table 4) was negatively impacted by flumetsulam + clopyralid + acetochlor, flumetsulam + clopyralid, atrazine + S-metolachlor + bicyclopyrone + mesotrione, clopyralid + acetochlor + mesotrione, S-metolachlor + bicyclopyrone + mesotrione, acetochlor + mesotrione, mesotrione, and saflufenacil + dimethenamid-P (0.460 to 1.096 g pot⁻¹) compared to the NTC (1.744 g pot⁻¹; Table 7). The soil residual herbicides applied PRE that contained only Groups 5 and 15 (atrazine, simazine, acetochlor, S-metolachlor, atrazine + S-metolachlor, and atrazine + acetochlor) did not injure radish (1.528 to 1.939 g pot⁻¹) compared to the NTC.

Table 7. Effect of PRE herbicides on radish biomass production at each sampling time and area under biomass stairs estimated for radish biomass production by PRE herbicide over time in greenhouse bioassays using field-treated soil from Janesville, WI, and Lancaster, WI, in 2021 and 2022. ^a,b

^a Abbreviations: ACET, acetochlor; ATZ, atrazine; AUBS, area under biomass stairs; BIP, bicyclopyrone; CLOP, clopyralid; DAT, days after treatment; DIM-P, dimethenamid-P; FLUM, flumetsulam; IFT, isoxaflutole; LSD, least significant difference; MES, mesotrione; SAFL, saflufenacil; S-MET, S-metolachlor; SMZ, simazine.

^b Means within a column with the same letter are not significantly different according to Fisher’s LSD test at P ≤ 0.05.

^c Aboveground biomass 28 d after sowing the greenhouse bioassay.

For the overseeding scenario (field samples collected at 70 DAT; ∼V10 to VT corn growth stages; Table 4), radish biomass was not reduced by atrazine + S-metolachlor, atrazine + acetochlor, S-metolachlor, simazine, acetochlor, atrazine, or saflufenacil + dimethenamid-P (1.225 to 1.621 g pot⁻¹) compared to the NTC (1.344 g pot⁻¹; Table 7). The remaining treatments reduced radish biomass (0.704 to 1.119 g pot⁻¹) compared to the NTC. For the AUBS, only S-metolachlor (155.66), acetochlor (148.29), and atrazine + S-metolachlor (135.95) were not different from the NTC (139.02). Atrazine + acetochlor, atrazine, and saflufenacil + dimethenamid-P presented intermediate AUBS values (88.90 to 127.33), whereas the remaining treatments had the lowest AUBS values (45.00 to 73.80).

On the basis of these results, applications of residual herbicides containing Groups 2, 4, and 27 evaluated in this study are likely to injure radish interseeded into corn at 30 DAT (V3 or V5 growth stage) and 70 DAT (V10 to VT growth stages) because of the short interval between herbicide application and cover crop interseeding or overseeding. Mesotrione (Group 27) and flumetsulam + clopyralid (Groups 2 and 4) herbicide labels list a 26- and 10-mo rotational restriction for canola (Brassica napus L.; Anonymous 2022a, 2022b), respectively, which belongs to the same family as radish and may have similar sensitivity. Brooker et al. (2020b), in a field experiment, also reported that Group 2 herbicides (flumetsulam [56 g ai ha⁻¹] and rimsulfuron [22 g ai ha⁻¹]) caused >70% radish stand reduction into corn at the V3 to V6 stages compared to the NTC. In a greenhouse experiment, these same authors observed that Group 27 (mesotrione [210 g ai ha⁻¹]) resulted in >50% biomass reduction at rates less than field use rates; the authors did not observe radish stand and biomass reduction by a Group 4 herbicide (clopyralid [105 g ai ha⁻¹]). According to our results, delaying radish planting until 70 DAT is likely to reduce injury and biomass reduction if saflufenacil + dimethenamid-P is applied. A previous study reported that fall-seeded radish was not negatively impacted by saflufenacil + dimethenamid-P (735 + 1,470 g ai ha⁻¹; Yu et al. 2015).

Interseeding and Overseeding Red Clover Scenario

Red clover biomass 28 DAS in the interseeding scenario (field soil samples collected at 30 DAT; ∼V3 to V5 corn growth stages; Table 4) was negatively impacted by soil residual herbicides applied PRE that contained Groups 2, 4, and 27 (mesotrione, acetochlor + mesotrione, flumetsulam + clopyralid, S-metolachlor + bicyclopyrone + mesotrione, atrazine + S-metolachlor + bicyclopyrone + mesotrione, flumetsulam + clopyralid + acetochlor, and clopyralid + acetochlor + mesotrione) with a biomass production ranging from 0.000 to 0.027 g pot⁻¹ compared to the NTC (0.253 g pot⁻¹; Table 8). S-metolachlor, acetochlor, saflufenacil + dimethenamid-P, and atrazine did not negatively impact red clover biomass (0.203 to 0.243 g pot⁻¹). The remaining treatments resulted in intermediate injury (0.122 to 0.199 g pot⁻¹).

Table 8. Effect of PRE herbicides on red clover biomass production at each sampling time and area under biomass stairs estimated for red clover biomass production by PRE herbicide over time in greenhouse bioassays using field-treated soil from Janesville, WI, and Lancaster, WI, in 2021 and 2022. ^a,b

^a Abbreviations: ACET, acetochlor; ATZ, atrazine; AUBS, area under biomass stairs; BIP, bicyclopyrone; CLOP, clopyralid; DAT, days after treatment; DIM-P, dimethenamid-P; FLUM, flumetsulam; IFT, isoxaflutole; LSD, least significant difference; MES, mesotrione; SAFL, saflufenacil; S-MET, S-metolachlor; SMZ, simazine.

^b Means within a column with the same letter are not significantly different according to Fisher’s LSD test at P ≤ 0.05.

^c Aboveground biomass 28 d after sowing the greenhouse bioassay.

For the overseeding scenario (field samples collected at 70 DAT; ∼V10 to VT corn growth stages; Table 4), the PRE herbicides that contained Group 27 (mesotrione, acetochlor + mesotrione, S-metolachlor + bicyclopyrone + mesotrione, atrazine + S-metolachlor + bicyclopyrone + mesotrione, and clopyralid + acetochlor + mesotrione) still caused high injury to red clover (0.000 to 0.009 g pot⁻¹) compared to the NTC (0.296 g pot⁻¹; Table 8). PRE herbicides that contained Groups 5, 14, and 15 (atrazine, simazine, acetochlor, S-metolachlor, atrazine + S-metolachlor, and saflufenacil + dimethenamid-P) did not injure red clover (0.0.251 to 0.354 g pot⁻¹) compared to the NTC. The remaining PRE herbicide treatments resulted in intermediate injury (0.231 and 0.153 g pot⁻¹).

The AUBS results (Table 8) support the high red clover sensitivity to soil residual herbicides applied PRE that contain Groups 2, 4, and 27 (flumetsulam + clopyralid [7.07], flumetsulam + clopyralid + acetochlor [4.51], mesotrione [0.16], acetochlor + mesotrione [0.07], S-metolachlor + bicyclopyrone + mesotrione [0.03], atrazine + S-metolachlor + bicyclopyrone + mesotrione [0.25], and clopyralid + acetochlor + mesotrione [0.07]) compared to the NTC (18.94). Although none of the PRE herbicides reached an AUBS equal to the NTC, PRE herbicides containing Groups 5, 14, and 15 caused less injury than the PRE herbicides last mentioned. The PRE herbicides that caused less injury were atrazine (13.98), simazine (11.97), acetochlor (15.89), S-metolachlor (16.63), atrazine + acetochlor (13.04), and saflufenacil + dimethenamid-P (15.88).

The low red clover biomass reduction by atrazine and simazine after 30 DAT may be due to the fast degradation of these herbicides. Mueller et al. (2017) reported enhanced dissipation and a decrease in atrazine persistence in some locations in Wisconsin due to microbial degradation, limiting extended weed control. Our results demonstrate that red clover is highly sensitive to mesotrione (Group 27) applied solo and in the premixes even at 70 DAT. Wallace et al. (2017) reported more than 98% biomass reduction by mesotrione (188 g ai ha⁻¹) and atrazine + S-metolachlor + mesotrione applied PRE at a reduced rate (0.5X) compared to the NTC in silt-loam soil fields at the V3 corn growth stage. Field studies conducted in silt-loam soils have shown that the half-life of mesotrione ranges from 8 to 32 d (Dyson et al. 2002), but mesotrione may persist longer in the soil, depending on the edaphoclimatic conditions (Su et al. 2017), especially pH and organic matter (Dyson et al. 2002; Shaner et al. 2012). For example, as pH decreases, the mesotrione half-life increases (Chaabane et al. 2008; Shaner et al. 2012). Our results at 70 DAT are also supported by other studies that have demonstrated mesotrione carryover injury to rotational crops (Pintar et al. 2020) and fall-seeded cover crops (Palhano et al. 2018). Mesotrione (Group 27) also lists an 18-mo rotational restriction for red clover (Anonymous 2022a), which can explain the high sensibility of red clover up to 70 DAT in our study. These rotational herbicide label restrictions address only potential crop injury and are independent of plant-back interval (PBI) restrictions established by the Environmental Protection Agency (WSSA 2022). If cover crops are planted for soil health purposes, PBI restrictions do not apply. However, if cover crops are planted for livestock feeding, grazing, or human consumption, PBI restrictions must be complied with.

Cover Crops Injury by the Number of Active Ingredients

The estimated cover crop AUBS values analyzed by the number of SOAs showed that the higher the number of SOAs for the products tested is, the higher is the injury, except for cereal rye (Figure 4). For annual ryegrass, the AUBS values followed the order of NTC (33.43) followed by PRE herbicides with one SOA (19.43), two SOAs (13.68), and three SOAs (9.92). The cereal rye AUBS for PRE herbicides with three SOAs (54.2) was not different from the NTC (55.4), whereas PRE herbicides with one SOA (46.7) and two SOAs (49.3) slightly reduced growth compared to the NTC. All SOA numbers negatively impacted radish AUBS (112.8, 94.8, and 59.2 for one, two, and three SOAs, respectively) compared to the NTC (139.0). The same was observed for red clover AUBS values (9.82, 7.5, and 0.59 for one, two, and three SOAs, respectively) compared to the NTC (18.94). The AUBS combined across species followed the same trend, where PRE herbicides with a single SOA (39.1), two SOAs (33.6), and three SOAs (22.5) negatively impacted the cover crops compared to the NTC (53.9).

Figure 4. Area under biomass stairs (AUBS) estimated for annual ryegrass, cereal rye, radish, and red clover, and combined across species by PRE herbicide sites of action (SOAs; one, two, or three SOAs) over time in greenhouse bioassays using field-treated soil from Janesville, WI, and Lancaster, WI, in 2021 and 2022. Jittered points represent replicates, centered solid points denote the means, and error bars represent the upper and lower 95% confidence interval limits. Means were compared using Fisher’s least significant difference, and herbicide treatments with the same letters are not different at P ≤ 0.05.

Although the cover crops tended to be more sensitive to the premixes with multiple SOAs than with a single SOA (Figure 4), premixes with at least two SOAs are necessary to improve weed control success (Silva et al. 2023). In this case, the selection of cover crop species can be more restricted. But acceptable levels of weed control are needed to achieve production goals and enhance the chances of successful establishment of interseeded cover crops (Wallace et al. 2017). Therefore premixes with at least two SOAs should be tested in the field on different weeds and cover crop species to carefully select a herbicide program that can provide effective weed control and successful cover crop establishment.

Practical Implications

All herbicides tested, except atrazine and simazine, resulted in biomass reduction of at least one cover crop 28 DAS in the interseeding scenario (field soil samples collected at 30 DAT; ∼V3 to V5 corn growth stage) and the overseeding scenario (field samples collected at 70 DAT; ∼V10 to VT corn growth stage). Consequently, species selection might be a challenge in the case of using grass–legume cover crop mixtures. Conversely, for each cover crop studied, there were soil residual herbicides applied PRE that did not negatively impact biomass. Cereal rye was the most tolerant cover crop species, followed by radish, red clover, and annual ryegrass. Cereal rye was affected by only 6 out of the 14 total PRE herbicides at 30 DAT and by none of the PRE herbicides at 70 DAT. The chances of injury were higher for annual ryegrass, radish, and red clover when the number of SOAs in a premix was higher. This trend was not observed for cereal rye. These results suggest that certain soil residual herbicides applied PRE are likely to reduce biomass of interseeded (∼V3 corn growth stage) and overseeded (∼VT corn growth stage) cover crops; therefore cover crop species should be carefully selected depending on the residual PRE herbicide applied. This new system can be challenging, but this study shows some potential cover crop options for farmers using the soil residual herbicides applied PRE investigated herein. Moreover, additional field studies are needed to validate these results in different environments and to support recommendations to growers interested in this system.