Introduction

Glyphosate has become the most widely used herbicide in grain crops with the adoption of glyphosate-tolerant crops, such as Roundup Ready^® soybean. This intensive use leads to the selection of herbicide-resistant weed biotypes. There are currently 53 weed species with cases of glyphosate-resistant biotypes worldwide (^{Heap, 2021}).

The weeds hairy fleabane (Conyza bonariensis), horseweed (C. canadensis), and Sumatran fleabane (C. sumatrensis) are among the main weeds found worldwide (^{Trainer et al., 2005}). They are annual herbaceous plants with high seed production and are found in various agricultural environments, such as grain crops (^{Lorenzi, 2014}). Altogether, they have 105 reported cases of biotypes that are resistant to herbicides, such as glyphosate, paraquat, and acetolactate synthase (ALS) inhibitors (^{Heap, 2021}). Recent studies showed that there are biotypes of C. sumatrensis with resistance to paraquat (^{Zobiole et al., 2019}) or 2,4-D (^{Queiroz et al., 2020}), and multiple resistance to glyphosate, chlorimuron, and paraquat (^{Albrecht, Pereira, et al., 2020}).

Thus, it is necessary to use other herbicides, such as pre-emergent herbicides, and different modes of action that are effective in controlling hard-to-control weeds, whether herbicide-tolerant or resistant (^{Mueller et al., 2014}). Studies highlight the efficacy of pre-emergent herbicides in controlling Conyza spp. in soybeans. These include flumioxazin (^{Zimmer et al., 2018}), diclosulam (^{Braz et al., 2017}), diuron (^{Moreira et al., 2010}), sulfentrazone (Zimmer et al., 2018), and imazethapyr (^{Hedges et al., 2019}).

In addition to the use of pre-emergent herbicides for the effective management of Conyza spp., burndown herbicides should be used in combination in situations with the presence of these weeds prior to soybean sowing. Among these herbicides, diquat (^{Weaver et al., 2004}), paraquat, and glufosinate (^{Zobiole et al., 2018}) stand out.

The application of pre-emergent herbicides combined with burndown herbicides is believed to be effective in controlling Conyza spp. in the pre-sowing management of soybean. Therefore, the present study aimed at evaluating the efficacy of pre-emergent herbicides sulfentrazone/diuron, imazethapyr/flumioxazin, and diclosulam in combination with burndown herbicides diquat, paraquat, or glufosinate on the control of Conyza spp.

Materials and methods

Two field experiments were carried out in the city of Palotina, state of Parana, Brazil, at coordinates 24°20'44.54" S, 53°51'50.93" W (experiment 1) and 24°20'48.89" S, 53°51'37.58" W (experiment 2) during the 2018-19 growing season.

For both experiments, the soil was classified as clay texture, with 63% clay, 19% silt, and 15% sand. The climate of the region is temperate humid with hot summers (Cfa), according to the Köppen classification (^{Aparecido et al, 2016}), and the weather conditions for the experimental period are illustrated in Figure 1. The area in experiment 1 was previously cultivated with maize and in experiment 2 with wheat. Both areas were infested with Conyza spp. plants up to 15 cm high and with 16 leaves. No flowering plants were observed at the time of application, and plant density was 17 and 20 Conyza plants m^-2 in experiments 1 and 2, respectively.

FIGURE 1 Rainfall and minimum and maximum temperature for the experimental sites. 2018/19 season, Palotina, PR, Brazil. 

The treatments consisted of different herbicide applications in a randomized block design with four replicates (Tab. 1). Application of treatments occurred on October 15, 2018 with sowing of soybean cultivar Monsoy* 5947 IPRO (Monsanto Co. do Brasil, São Paulo, SP, Brazil) on the same day, immediately after application. Herbicides were applied using a C0₂ pressurized backpack sprayer (Pulverizador Pesquisa - Herbicat Ltda, Catanduva, SP, Brazil) equipped with six AIXR 110.015 nozzles at a pressure of 2.5 kgf m^-2 and a speed of 3.6 km/h, providing an application volume of 150 L ha^-1. For experiment 1, the climatic conditions during the application were as follow: temperature of 27.3°C, relative air humidity of 61.5%, and wind speed of 6.4 km/h. For experiment 2, the conditions were as follow: temperature of 29.5°C, relative air humidity of 60.2%, and wind speed of 7.0 km/h.

TABLE 1 Herbicide mixture treatments applied in pre-sowing of soybean to control Conyza spp. 2018/19 season, Palotina, PR, Brazil. 

*Rates at g of acid equivalent (ae) ha^-1 for imazethapyr and at g of active ingredient (ae) ha^-1 for the other herbicides. Addition of 0.5% mineral oil to all treatments.

The experimental units were composed of 3 x 5 m parcels; soybean sowing was carried out after the application of the treatments with 13 seeds m^-1. Six soybean lines were sown in each parcel with a spacing of 0.45 m. The evaluations were carried out in the useful area of the parcel, discarding the external lines and the first and last meters.

The control of Conyza spp. plants was assessed at 7, 14, 21, 28, and 35 d after herbicide application (DAA). Injury symptoms in soybean plants were assessed at 14, 21, 28 and 35 DAA. These assessments were carried out through visual analysis at each experimental unit considering significantly visible symptoms in the plants, according to their development. Scores from 0 to 100% were assigned, where 0 represented the absence of symptoms and 100% the death of the plant (^{Velini et al., 1995}). The treatment without application (without herbicide effect) was used as a reference for evaluations, always with a score of 0, whether for weed control or injuries to soybean plants, as in other studies (^{Braz et al., 2017}; ^{Chahal & Jhala, 2019}; ^{Guerra et al., 2020}).

Data were tested by analysis of variance and F-test (P<0.05) according to ^{Pimentel-Gomes and Garcia (2002}). The means of the treatments were compared by the Tukey test (P<0.05) using the Sisvar 5.6 program (^{Ferreira, 2011}).

Results

In experiment 1, all herbicide treatments were effective in controlling Conyza spp., with values higher than or equal to 83.8% at 7 DAA for all treatments. No differences were detected between herbicide treatments, and all of them were superior to the weedy control treatment (without application) throughout all evaluations. It is worth noting the control for all combinations at 35 DAA with values of at least 94.3% (Tab. 2).

TABLE 2 Control (%) of Conyza spp. from 7 to 35 d after application of herbicide mixtures (DAA). 2018/19 season, Palotina, PR, Brazil (experiment 1). 

*Means followed by the same letter in the line do not differ from each other according to the Tukey test at the 5% probability level. CV - coefficient of variation.

As observed for experiment 1 (area previously cultivated with maize), in experiment 2 (area previously cultivated with wheat) herbicide treatments were effective in controlling Conyza spp. with no differences between combinations from 21 to 35 DAA. Some differences were observed between herbicide treatments at 7 and 14 DAA; however, all exhibited weed control of at least 89.0%. At 35 DAA, a control of Conyza spp. of at least 94.8% was observed for all herbicide treatments (Tab. 3).

TABLE 3 Control (%) of Conyza spp. from 7 to 35 d after application of herbicide mixtures (DAA). 2018/19 season, Palotina, PR, Brazil (experiment 2). 

*Means followed by the same letter in the line do not differ from each other according to the Tukey test at the 5% probability level. CV - coefficient of variation.

For experiment 1, at 14 DAA, no differences were detected between treatments regarding injury symptoms in soybean plants. Treatments with diclosulam caused stronger symptoms, up to 10.3% at 21 DAA; these were superior to almost all other treatments from 21 to 35 DAA. For the treatments composed of the application of sulfentrazone/diuron and imazethapyr/flumioxazin, injury symptoms were 3.8% to 4.3% and 2.3% to 2.5%, respectively at 35 DAA (Tab. 4).

TABLE 4 Crop injury (%) of soybean plants from 14 to 35 d after application of herbicide mixtures (DAA). 2018/19 season, Palotina, PR, Brazil (experiment 1). 

*Means followed by the same letter in the line do not differ from each other according to the Tukey test at the 5% probability level. ns - not significant, means do not differ from each other by the F-test at the 5% probability level. CV - coefficient of variation.

Table 5 shows the injury symptoms in soybean plants due to herbicide application for experiment 2. Differences between treatments were observed in all evaluations. Symptoms were up to 9% (at 28 DAA), while at the last assessment (35 DAA) they ranged from 2.0% to 6.8%, in general with greater potential for injury to the application of diclosulam.

TABLE 5 Crop injury (%) of soybean plants from 14 to 35 d after application of herbicide mixtures (DAA) during the 2018/19 season, Palotina, PR, Brazil (experiment 2). 

*Means followed by the same letter in the line do not differ from each other according to the Tukey test at the 5% probability level. CV - coefficient of variation.

Discussion

In this study, the pre-sowing application of sulfentrazone/ diuron (premix formulation) in combination with diquat, paraquat, or glufosinate was effective in controlling Conyza spp. Similar results were observed by ^{Zimmer et al. (2018)}, where pre-sowing application of sulfentrazone (195 g ae ha^-1) in a mixture with halauxifen (5 g ai ha^-1) + glyphosate (1,120 g ae ha^-1) + cloransulam (25 g ai ha^-1) provided 94% control of C. canadensis at 35 d after application. Sulfentrazone was effective in different management programs to control Amaranthus tuberculatus (^{Schryver et al., 2017}) and other weeds. Sulfentrazone has a variable half-life (34-116 d) according to soil moisture and temperature and is less persistent under conditions of higher humidity and higher temperature (40°C); these aspects interfere with its residual period (^{Brum et al., 2013}). Additionally, sulfentrazone has a spectrum of action on eudicotyledonous weeds and some grasses (^{Rodrigues & Almeida, 2018}). These studies, and the results of the present study, indicate the effectiveness of sulfentrazone in different mixtures on the control of weeds.

The application of diuron with mixtures was effective in controlling Conyza spp. in the present study. In other research, a control of 91% of C. bonariensis at 30 DAA was observed with the application of diuron (200 g ai ha^-1) at mixtures (^{Paula et al., 2011}). Other studies also highlight the effectiveness of diuron in controlling Conyza spp. in different combinations (^{Lamego et al., 2013}; ^{Santos et al., 2015}). Diuron is an herbicide that is especially effective for the control of eudicotyledons and some grasses. In the present study, diuron was effective in controlling Conyza spp. in a premixture with sulfentrazone. The half-life can vary from 40 to 91 d and is generally more persistent in soils with higher levels of clay and organic matter (^{Rocha et al., 2013}). This characteristic helps to explain the results of this study.

The pre-sowing application of imazethapyr/flumioxazin (premix formulation) was also effective for controlling Conyza spp. in this study. Similar results were observed by ^{Albrecht, Albrecht, et al. (2020)}, who found that this premix formulation obtained a control of about 80% of Conyza spp. up to 35 DAA. Imazethapyr has a soil half-life varying from 36 to 98 d (^{Marinho et al., 2019}), and flumioxazin from 13 to 18 d (^{Ferrell & Vencill, 2003}), according to the edaphoclimatic conditions. The persistence of imazethapyr helps to explain the effectiveness of this mixture over time, especially in the emergence of plants.

Pre-sowing application of imazethapyr (100 g ae ha^-1) in a mixture with glyphosate, dicamba, and saflufenacil provided 93% control of C. canadensis 12 weeks after application (^{Hedges et al., 2019}). This herbicide, in mixture with saflufenacil, is also effective in controlling Abutilon theophrasti, Amaranthus retroflexus, and Chenopodium album (^{Underwood et al., 2017}). Pre-sowing application of flumioxazin (76 g ai ha^-1) in combination with halauxifen (5 g ai ha^-1) + glyphosate (1,120 g ae ha^-1) + cloransulam (25 g ai ha^-1), following glyphosate application (1,120 g ae ha^-1) in post-emergence soybean, provided 96% control of C. canadensis at 35 d after the first application (^{Zimmer et al., 2018}). The results of the present study corroborated these findings, indicating the effectiveness of imazethapyr/ flumioxazin in different chemical control programs.

In the present study, diclosulam was also effective in controlling Conyza spp. in combination with diquat, paraquat, or glufosinate. In general, no significant differences were observed between the management adopted for the control of Conyza spp. in initial post-emergence (plants up to 15 cm high) and pre-emergence. This highlights the importance of rotation of mechanisms of action not only in post-emergence, but also in pre-emergence.

Diclosulam is one of the most widely used herbicides in soybean pre-sowing for controlling Conyza spp. and other weeds. ^{Krenchinski et al. (2019)} report control of up to 97.25% of Conyza spp. for diclosulam applied at pre-sowing in soybean, plus halauxifen + glyphosate. Other studies also report the control of Conyza spp. with the application of diclosulam in different combinations (^{Braz et al., 2017}; ^{Zobiole et al., 2018}). This herbicide has a soil half-life varying from 16 to 87 d, according to the edaphoclimatic conditions (^{Lavorenti et al., 2003}) and provides a broad-spectrum control (^{Rodrigues & Almeida, 2018}).

The results of this study indicated that glufosinate may be an alternative in the management of Conyza spp. as a tool for preventing resistance to herbicides. In this sense, it is worth mentioning the recent cases of a C. sumatrensis biotype reported to be resistant to paraquat (^{Zobiole et al., 2019}) or with multiple resistance to glyphosate, chlorimuron, and paraquat (^{Albrecht, Pereira, et al., 2020}) in the state of Parana, Brazil. Paraquat is a photosystem I inhibitor herbicide with the same mechanism of action as that of diquat. These two herbicides are among the main desiccants for pre-sowing application in soybean.

Several studies also highlight the efficacy of glufosinate in controlling Conyza spp. (^{Oliveira Neto et al., 2010}; ^{Tahmasebi et al., 2018}; ^{Zobiole et al., 2018}; ^{Albrecht, Albrecht, et al., 2020}). This herbicide is also effective in controlling other weeds, such as Amaranthus spp. (^{Hay et al., 2019}), Spermacoce latifolia and Richardia brasiliensis (^{Gallon et al., 2019}). Glufosinate can be used in pre-sowing, as in the present study, and in post-emergence of tolerant transgenic cultivars (soybean, maize, and cotton with pat or bar genes).

Regarding symptoms of injury, diclosulam showed higher phytotoxic potential, especially in experiment 1, with symptoms of up to 10.3%. For the other pre-emergent herbicides, minor injuries were observed in the soybean plants. Different studies also report injury symptoms in soybean with the application of diclosulam (^{Osipe et al., 2014}; ^{Braz et al., 2017}). However, plants recover from the symptoms with no effect on the agronomic performance, and this proves the selectivity of the herbicide.

^{Belfry et al. (2016)} found symptoms of 8% and 2% injury for pre-sowing application of flumioxazin and sulfentrazone, respectively. ^{Belfry et al. (2015)} also observed symptoms of 8% and 4% injury for pre-sowing application of flumioxazin and imazethapyr, respectively. ^{Braswell et al. (2015)} observed up to 15% soybean injury for pre-sowing application of diuron. For these herbicides, injury symptoms up to a maximum of 5% were observed in soybean plants at the end of the evaluations. These results confirm the potential selectivity of pre-emergent herbicides for soybeans.

Conclusions

Based on the results of this study we can conclude that it is important to rotate both post-emergent and pre-emergent herbicides since there have been several reports of resistance to glyphosate and ALS inhibitors that are widely used in pre-emergence. We would like to highlight the equivalence of the different herbicide combinations in the control of Conyza spp. in this study, after growing maize (experiment 1) or wheat (experiment 2).

The herbicides sulfentrazone/diuron, imazethapyr/flumioxazin, and diclosulam in mixtures with burndown herbicides (diquat, paraquat or glufosinate) were effective in controlling Conyza spp. at pre-sowing of soybean. The choice of pre-emergent or burndown herbicides must consider the history of use of the area, the weed community, and the cost of management, among other factors. Regardless of the choice, pre-emergent herbicides in combinations with burndown herbicides are important tools for the effective management of weeds.

This study fills a gap in the research since studies with the premixes sulfentrazone/diuron and imazethapyr/flumioxazin are not easily found in the literature in contrast to other products. Glyphosate was not used in this study, which characterizes the management adopted as an alternative to this herbicide.