Response of conventional sunflower cultivars to drift rates of synthetic auxin herbicides

Growth chamber experiments

A bioassay study was conducted in a growth chamber adjusted to 25 ± 1 °C/20 ± 1 °C with a 12/12 h day/night photoperiod to determine the sensitivity level of two commercial sunflower cultivars, Bosfora (Bosfora®; Syngenta Corporation, İzmir, Turkey) and Tunca (Tunca®; Limagrain Corporation, Bursa, Turkey), to sub-lethal rates of florpyrauxifen-benzyl + penoxsulam (BAXIGA® 32.5 OD; Corteva Agriscience Corporation, Istanbul, Turkey) and quinclorac (Facet®; BASF Corporation, Istanbul, Turkey). The growth chamber was set light to HPI-T Plus Metal Halide lamps (400 W). Three to four sunflower seeds were sown in plastic pots (7 × 7 × 8.5 cm) filled with white peat bedding substrate (TS 1, Klassman-Deilmann GmbH consisting of TS 1 fine + 15% perlite). After emergence seedlings were removed so that only two plants remained per pot. A high-performance air conditioning system was used to acclimatize the growth chamber because of adjusting of high temperature released by metal halide lamps. Therefore, the pots were arranged in a randomized complete block design with four replications according to acclimatization.

In the experiment, the seedlings were treated with seven herbicide rates (0, 2.93, 5.86, 11.72, 23.44, 46.88, and 93.75 g ai ha⁻¹ for quinclorac and 0, 0.51, 1.02, 2.03, 4.06, 8.13, and 16.25 g ai ha⁻¹ for florpyrauxifen-benzyl + penoxsulam) at the 2-4 true leaf stage. The quinclorac and florpyrauxifen-benzyl + penoxsulam label rates were 375 and 65 g ai ha⁻¹ respectively (PPPD, 2022). Herbicides that were dissolved in tap water were applied using a motorized backpack sprayer equipped with two flat-fan nozzles mounted on a hand-held boom (Teejet XR11002) and calibrated to deliver 195 L ha⁻¹. Sprayed seedlings were kept in the spraying place for 1 day after treatment. The seedlings were moved to the growth chamber and irrigated using tap water as needed. The seedlings were cut from ground level at 28 DAT and stored in a drying oven at 60 °C for 48 h to determine the above ground dry weight.

Field experiments

Two field experiments were conducted at Bilecik Seyh Edebali University Aşağıköy Agricultural Application and Research Centre (AAARC) in 2021 and 2022 to determine the response of sunflower cultivars to sub-lethal rates of florpyrauxifen-benzyl + penoxsulam and quinclorac under non-irrigated conditions. The soil texture at AARC was silty clay with 2.4% organic matter, pH 8.13. Field studies were arranged within a randomized complete block design with four replications. Conventional sunflower cultivars, Bosfora and Tunca, were sown to 70 cm spacing in plots consisting of four parallel rows on April 11, 2021, and May 02, 2022, respectively. Di-ammonium phosphate was applied at the time of sowing at 80 kg ha⁻¹. In this study, the temperature and rainfall in 2021 and 2022 were close to long-term averages (12.1 and 12.7 °C and 449.2 and 478.9 mm, respectively).

Florochloridone was applied prior to emergence at a 700 g ai ha⁻¹ rate in 2021, while pendimethalin was used two days after sowing at a 1.350 g ai ha⁻¹ rate to control weeds in 2022. Plots consisted of 4 70-cm-spaced rows 10 m long. Two crop rows between the herbicide-applied plots were left as alleys to avoid contamination between the plots.

Quinclorac and florpyrauxifen-benzyl + penoxsulam rates were 11.72, 23.44, and 46.88 g ai ha⁻¹ and 2.03, 4.06, and 8.13 g ai ha⁻¹ respectively, which were equivalent to 3.125, 6.250, and 12.5% of the recommended use rates (PPPD, 2022). The aforementioned rates of herbicides were applied when the sunflower reached 8-10 true leaves. Sunflower injury caused by these herbicide rates was recorded at 7, 14, and 28 DAT based on a scale of 0–100%, where 0% indicated no impact of herbicides, while 100% represented complete plant death. Plants were harvested for yield at the ripening stage. At harvest, five plants from the middle two rows of each plot were randomly selected and harvested by hand on September 05, 2021, and October 04, 2022. The heights of the selected plants, sunflower head diameter (SHD), 1000-seed weight (OTSW), and yield were measured (Serim & Maden, 2014).

Statistical analysis

The data from the dose-response study was evaluated using a nonlinear regression model in R statistical software (RStudio Team, 2023). The DRC package was used to calculate the dose-response curve and parameters with a four-parameter log–logistic model (Equation (1)). (1)${\displaystyle Y=c+\frac{d-c}{1+exp\left(b\left(log\left(x\right)-log\left({\mathrm{GR}}_{50}\right)\right)\right)}}$

where y represents seedling dry matter at herbicide treatment rate x; b, c, d, and GR₅₀ represent slope, lower limit, upper limit, and herbicide rate that reduced seedling dry matter by 50%, respectively.

The data from the field experiments were analyzed with analysis of variance for each herbicide and cultivar, and mean separation was performed with Fisher’s protected least significant difference test at the 5% level of probability. The agricolae package (Mendiburu & Yaseen, 2020) was used in the R statistical software program (RStudio Team, 2023). The Pearson correlation test was used to find a relationship between quinclorac injuries measured at 4 separate times and the yield (and yield components). The test was not performed for FBP because higher rates of FBP killed sunflowers before harvest.

Growth chamber experiment

The sunflower cultivar ‘Tunca’ was severely injured when exposed to low rates of florpyrauxifen-benzyl + penoxsulam (Fig. 1A). However, the lowest rate of herbicide unexpectedly increased shoot length, probably due to the hormetic impact of auxinic herbicides, similar to what was previously reported by Mudge et al. (2021). Overall, shoot lengths decreased as the rates increased. The impact of the highest four rates of herbicide was the most destructive, and the growing points of cultivar ‘Bosfora’ were completely killed at these rates at 28 DAT (Fig. 1B). More than half of the leaf area of seedlings exposed to these rates became necrotic. The response of cultivar Bosfora to FBP was similar to that of cultivar Tunca, except at rates higher than 1.02 g ai ha⁻¹ where it was more injured (Figs. 1A and 1B).

Injury from quinclorac in both cultivars increased over time, with a slight and gradual increase in sunflower injury with rate. By 28 DAT, the highest sunflower injury was caused by the highest rate of quinclorac (Figs. 1C and 1D). The response of these sunflower cultivars to lower rates of quinclorac was very similar; however, cultivar Bosfora was slightly more sensitive to quinclorac than cultivar Tunca at higher rates. As opposed to FBP, sunflower plants of both cultivars treated with quinclorac still had live growing points, and necrotic areas on the seedling leaves were relatively limited even at the highest rates. At the lowest rates, FBP resulted in no damage to these cultivars, while at the lowest dose of quinclorac growth reduction was still observed.

The quantitative response of the cultivar Tunca and cultivar Bosfora to FBP and quinclorac was evaluated via a dose-response assay using a log–logistic model. The GR₅₀ values of FBP for the cultivars Tunca and Bosfora were 1.07 and 0.75 g ai ha⁻¹, respectively (Table 1), nearly 2.9% and 2% of the recommended rate of FBP (65 g ai ha⁻¹). The GR₅₀ values of quinclorac for sunflower cultivars were 14.16 and 7.56 g ai ha⁻¹, 3,8% and 2% of the recommended rate of quinclorac (375 g ai ha⁻¹). Similar to the visual herbicidal impact of FBP and quinclorac, the results show that FBP was slightly more injurious than quinclorac to these cultivars (Fig. 2).

Field experiment

Crop injury

Sunflower crops exposed to FBP and quinclorac responded to low-dose treatment, especially the leaves. FBP resulted in typical auxin symptoms, such as parallel veins, cupping, twisting, chlorosis, and distortion (Fig. 3). Quinclorac also caused these symptoms, except it did not demonstrate distortion (Fig. 4). The severity of symptoms was higher in cultivar Bosfora than in cultivar Tunca and increased as the rates increased. Stunting became apparent at 28 DAT for both cultivars. The highest rate of FBP prevented the establishment of sunflower heads in both cultivars (Figs. 5B and 5D).

Quinclorac and FBP injury was low at 7 DAT, and no difference was found between cultivar Tunca and cultivar Bosfora (Figs. 6A and 6B). The response of sunflower cultivar to FBP increased as the rates rose and reached 75–77.5% at 8.13 g ai ha⁻¹ at 7 DAT. Similar to the growth chamber study, quinclorac injury in cultivar Bosfora (13.75–33.75%) and cultivar Tunca (17.5–30%) was lower than those of FBP. At 14 DAT, the response of the cultivars to the herbicides was generally similar to that at 7 DAT. Sunflower injury due to FBP increased with the increase in herbicide rates from 2.03 to 8.13 g ai ha⁻¹ at 14 DAT. Cultivar Bosfora exposure to quinclorac at 11.72, 23.44, and 46.88 g ai ha⁻¹resulted in 15, 22.5, and 38.75% injury, respectively, while cultivar Tunca was less sensitive to these rates at 14 DAT.

Sunflower injury at the highest rate of FBP was 100% by 28 DAT while the lowest FBP rate resulted in sunflower injury of 76.25–88.75% (Figs. 6A and 6B). Differences between sunflower cultivars became more apparent at 28 DAT. Although the phytotoxicity of quinclorac increased at 28 DAT, crop injury caused by the herbicide was nearly half that of FBP.

Visible injury rates reached the greatest level at harvest. Sunflower injury at 2.03 g ai ha⁻¹ FBP was higher than >90% for cultivar Bosfora and >85% for cultivar Tunca. The highest FBP rate led to complete death of both sunflower cultivars. The injury increased in severity as the quinclorac rate increased in both cultivars. The injury on cultivar Bosfora and cultivar Tunca ranged from 42.5–57.5% and from 36.25–48.75%, respectively. The severity of injury due to quinclorac was limited to at 7 DAT because injury assessment was only performed on leaves, while this influence was more apparent in evaluations of other crop parameters, including plant height, stem structure, size of flower bud and head, were included in the evaluations. Compared to quinclorac, FBP was more phytotoxic to both cultivars, and this destructive impact was observed even from the first assessments at 7 DAT.

The injury rate of auxin herbicides on sensitive row crops has been reported in some studies. For instance, Schroeder, Cole & Dexter (1979) stated that 2,4-D and dicamba resulted in tremendous injury to sunflowers. In another study, Snipes, Street & Mueller (1992) reported that 17 g ha⁻¹ or higher rates of quinclorac may injure cotton when applied at the cotyledon stage, and cotton injury increased with increasing herbicide rate, especially during late-stage applications, such as at the pin-head square stage. Overall, the cotton injury rate reached 59% at 140 g ha⁻¹ in their study, and sunflower injury caused by quinclorac at 46.88 ai g ha⁻¹ was 52-61.25% in our study. A study conducted by Lovelace et al. (2007) revealed that tomato was also among quinclorac-sensitive crops, and crop injury caused by quinclorac applied at 42 g ai ha⁻¹ was 45% at 49 DAT. In our study, the third evaluation was done at 55 DAT, and crop injuries were 48.75–57.5% when herbicide was applied at 46.88 g ai ha⁻¹. These data are consistent with the work of Lovelace et al. (2007). Comparing these results to those of Snipes, Street & Mueller (1992) and Lovelace et al. (2007), sunflower can be classified as a more sensitive crop than cotton.

Florpyrauxifen-benzyl is a relatively new herbicide on the market; therefore, few drift studies are available in the literature. FBP containing 2.03 g ai ha⁻¹ florpyrauxifen-benzyl + penoxsulam caused 76 and 89% injury at 28 DAT on cultivar Tunca and cultivar Bosfora, respectively. In agreement with the results of our study, Miller & Norsworthy (2018b) stated that 3 g ai ha⁻¹ florpyrauxifen-benzyl applied to sunflowers at the three true-leaf stage resulted in 69% injury, at 28 DAT under greenhouse conditions.

Plant height

The heights of the sunflower cultivars constantly declined as the FBP and quinclorac rates increased. All FBP rates resulted in a significant decrease in the plant height of the sunflower cultivars; The lowest rate of FBP resulted in an 8–9% decrease while the highest resulted in a 32–44% decrease (Fig. 7A). Similarly, the height reduction from quinclorac ranged from 13–14% at the lowest rate to 25–32% at the highest. Similar to the findings of this study, Miller & Norsworthy (2018b) also stated that 3 g ai ha⁻¹ florpyrauxifen-benzyl applied to sunflowers at the three true-leaf stage resulted in a 66% plant height reduction at 28 DAT under greenhouse conditions. The lower plant height observed in our study may be due to the application time of herbicide or penoxsulam in the herbicide mixture.

Sunflower head diameter and One thousand seed weight

Increasing FBP and quinclorac rates caused sunflower head diameter (SHD) to decline regardless of the cultivar. The greatest percentage of decline resulting from FBP was recorded at 4.06 g ai ha⁻¹ with 76–77%. At the highest rate of FBP (8.13 g ai ha⁻¹), no flower heads were observed (Figs. 5B and 5D). SHD reduction in cultivar Bosfora and cultivar Tunca from quinclorac ranged from 19–57% and 9–51%, respectively (Fig. 7B). A slight decrease was observed in One thousand seed weight (TSW) compared to plant height and SHD, but this decrease, even at the highest level of herbicide, was no more than 21% (Fig. 7C). Sunflower exposed to higher FBP rates, 4.06 and 8.13 g ai ha⁻¹, could not produce mature seeds. The results showed that TSW was a less susceptible yield component to FBP drift rates than the others.

Yield

Associated with other yield components, significant sunflower yield loss was determined even at lower rates of quinclorac and FBP (Fig. 7D). The highest yield reduction was observed for sunflower cultivars treated with FBP at 4.06 and 8.03 g ai ha⁻¹at 100%. The lowest FBP rate reduced cultivar Bosfora and cultivar Tunca yields by 79.8 and 74.3%, respectively. The yield reduction of cultivar Bosfora from quinclorac rates was 45.1–87.9%, while yield loss of cultivar Tunca caused these rates to range from 16.3–82.3%. Although no study was found in the literature related to the impact of low rates of quinclorac and FBP on sunflower yield and yield components such as SHD and TSW, previous research reported significant reductions in crop growth, yield, and some yield components in other systems. Miller & Norsworthy (2018b) found that soybean yield was reduced 71% when florpyrauxifen-benzyl was applied at 3 g ae ha⁻¹ rate. In another study, Snipes, Street & Mueller (1992) calculated that quinclorac at 50 g ha⁻¹ reduces cotton yield by 25%. On the other side, Lovelace et al. (2007) indicated that quinclorac resulted in yield loss in tomatoes, but the crop may recover itself from the adverse impact of herbicides depending on the herbicide rate and application timing. Our results are consistent with those of the aforementioned studies that lower rates of florpyrauxifen-benzyl were more destructive to crop yields than quinclorac.

Crop cultivars can have various genetic backgrounds depending on the breeding aim; therefore, it is not surprising that they respond differently to abiotic stressors, including drought stress, heat stress, and herbicides. Using herbicide-tolerant crops or less sensitive crop cultivars to herbicide are among the cost-effective and reliable solutions to prevent the injurious impact of herbicide drift on sensitive crops. This practice has been shown in other studies of by France et al. (2022), Zangoueinejad et al. (2021) and Warmund, Ellersieck & Smeda (2022), who showed differences between the tolerance levels of soybean, melon, and tomato cultivars to synthetic auxin herbicides dicamba, 2.4-D, and 2.4-D or dicamba, respectively. The yield data clearly showed that cultivar Bosfora was more sensitive to FBP and quinclorac rates than cultivar Tunca. To reduce the impact of these off-target effects, more cultivars can be screened, and robust cultivars can be selected to reduce the risk of off-target herbicide damage.

Correlation analysis

Remarkably high Pearson correlation coefficients were found between quinclorac injury at 7 or 14 DAT and plant height (Table 2). The negative relations between quinclorac injury at 7 or 14 DAT and SHD, TSW, and/or yield were also high, but their significance was slightly below the relations between quinclorac injury at 7 or 14 DAT and plant height. At 28 DAT, plants treated with quinclorac began to recover from treatment; therefore, quinclorac injury at 28 DAT had a weaker correlation with yield and yield component compared to previous evaluation times. There was a strong positive relationship between plant height and yield, and the importance of the relationship was greater than that of other relationships.

The correlation analysis performed in this research can be a powerful tool to estimate the injurious impact caused by drift rates of synthetic auxin herbicides on yield and yield components long before harvest. In our study, strong relationships between injury rates and yields (or yield components) were similar to those found in previous studies (Lovelace et al., 2007; Marple, Al-Khatib & Peterson, 2008; Daramola et al., 2023). The ability to model injury rates and yield loss resulting from herbicides provide an opportunity for farmers to decide whether to stop or continue current agricultural practices. If herbicide damage reduces farmer’s income below the total expenses, farmers may wish to change their management in order to remain profitable; therefore, correlation analysis between herbicides and yield can be used as a decision-support tool for farmers.