Evaluating the role of post-harvest glyphosate application in enhancing weed control in winter wheat

Impact of herbicide programs on weed flora and wheat yield

The study was designed as a split-plot experiment with four replications. The main plots were assigned to the herbicide regime post-harvest glyphosate potassium salt (PHG) treatment and No PHG, and the subplots were allocated to the weed control treatments. Each plot was 3 m × 3.5 m. A two m alley was left between blocks, and a one m alley was left between parcels (Fig. 2A). The experiment was conducted in a field with winter wheat grown for 3 consecutive years. Four of the eight blocks were allocated for PHG application, whereas the remaining four were used for No PHG (conventional method) application. These blocks were marked, and the same treatments were repeated for 3 years. The same plot was used for each treatment when the experiment was repeated in the second and third years in the winter wheat.

The treatments used were as follows: T1: PHG + Post-emergence tribenuron-methyl, T2: PHG + Weedy check, T3: No PHG + Post-emergence tribenuron-methyl, T4: No PHG + Weedy check (Control), T5: PHG + Hand weeding, and T6: No PHG + Hand weeding. Hand weeding was performed 3 times at 10 day intervals after tillering began. Although this treatment is a time-consuming and labor-intensive method, it is advised by local advisers and is practiced by Turkish growers whose fields are covered by volunteers and weeds that have not been controlled by herbicides, such as feral rye or herbicide-resistant weeds.

Glyphosate potassium salt (ROUNDUP STAR^®; Bayer Crop Science, İstanbul, Türkiye) and tribenuron-methyl (GRANSTAR^®; FMC Turkey, İstanbul, Türkiye) were applied using a field sprayer and motorized backpack sprayer adjusted to deliver 200 l ha⁻¹ at 200 kPa pressure, respectively (Figs. 2B and 2C). Glyphosate potassium salt was applied at 2.646 kg ai ha⁻¹ one week after wheat harvest as a post-harvest treatment to control grass and broadleaf weeds in the stubble. Tribenuron-methyl was applied at a rate of 7.5 g ai ha⁻¹ at the 3–4 leaf stage of broadleaf weeds. Glyphosate potassium salt has no residual activity, whereas tribenuron-methyl has a short residual activity in the soil (Shaner, 2014).

The weed flora was determined at the end of April 3 weeks after tribenuron-methyl treatment. Weeds in each plot were recorded from one m² and identified according to the Flora of Turkey and the North Aegean Islands (Davis, 1965–1985). During the harvest time, spikes in a one m² area were harvested from each plot. The harvested spikes were placed in bags, brought to the laboratory, and threshed. The grain yield was calculated on a per-hectare basis by converting the results from the harvest area. Weed density was calculated by using the Formula (1) (Nkoa, Owen & Swanton, 2015). (1)${\displaystyle \mathrm{D}=\frac{\sum \mathrm{Y}}{\mathrm{Sa}}}$ where D is the density of the weed species; ∑Y is the count of individual plants of weed species placed in a quadrat; and Sa is the surface area of the quadrat (one m²).

The collected data on wheat grain yield were subjected to variance analysis using the Agricolae package (Mendiburu & Yaseen, 2020) in R statistical software (R Studio Team, 2024). The means were compared using the Least Significant Difference test at the 5% probability level. The wheat yield data were subjected to ANOVA separately for each year because the post-harvest herbicide treatment by weed control treatment interactions were significantly different (P < 0.05).

Impact of herbicide programs on the weed seed bank

The weed seed bank was determined for three consecutive years by using the method described by Mayor & Dessaint (1998), with slight modifications. Soil samples were collected from each plot using a soil borer (five cm diameter) at a 10 cm soil depth, and each soil sample was nearly 500 g. The soil samples were first sieved (four mm × four mm) to remove unwanted materials, such as debris and soil particles, and then sieved using a precise sieve (0.25 mm × 0.25 mm) to separate the soil aggregates. The soil samples were placed in aluminum trays filled with tap water and left for 24 h (Fig. 2D). The slurry was gently mixed and washed using a sieve set under tap water. The seeds were dried using towel papers, put in paper bags, and stored in a cooler at +4 °C until classification. Weed species in the seed bank were identified using the Flora of Turkey and the North Aegean Islands (Davis, 1965–1985).

The impacts of the PHG and weed control treatments on Average Weed Seed Density (AWSD) data were subjected to variance analysis using the Agricolae package (Mendiburu & Yaseen, 2020) in R statistical software (R Studio Team, 2024). The means were compared using the Least Significant Difference test at the 5% probability level to determine the statistical significance of the differences among the treatments. AWSD data were subjected to ANOVA separately for each year because year, PHG treatment by weed control treatment interactions were significantly different (P < 0.05).

Impact of the PHG on the critical time for weed control (CTWR)

The CTWR in wheat was calculated by assessing the relationship between wheat grain yield and the duration of weed presence during the growing season. In the weedy control treatment, weeds were allowed to compete with the crop for the entire season, whereas in the weed-free control, weeds were removed manually at 10-day intervals to prevent competition with the crop. Weed removal times were adjusted at 10-day intervals, from 0 to 110 days after wheat emergence (DAE). The grain yield obtained from the weed-free control treatment was considered the maximum yield, and the percentage yield loss was calculated by comparing the grain yield from the weedy control treatment to that of the weed-free control.

The experiment was carried out for three consecutive years in a field grown continuous winter wheat. The experimental design was a split-block design with three replications, where the main plots were assigned to the PHG and No PHG, and the subplots were allocated to the length of the weed-free period, weed-free, and weedy season-long controls. Three of the six blocks were allocated for PHG application, whereas the remaining three were used for No PHG (conventional method) application. These blocks were marked, and the same application was repeated for 3 years. Each plot was two m × three m. A non-linear regression analysis was performed on the yield loss data to model the relationship between weed presence duration and wheat grain yield (Knezevic & Datta, 2015). The log–logistic model was then used to estimate the CTWR (Formula 2). (2)${\displaystyle Y=c+\frac{d-c}{1+exp\left(b\left(log\left(x\right)-log\left(E\right)\right)\right)}}$ where Y is the grain yield; c is the lower limit; d is the upper limit; x is the weed removal timing expressed in growing degree days (GDD) after wheat emergence; E is the GDD at the inflection point (I₅₀); and b is the slope around the I₅₀. The GDD was calculated according to Formula (3) (Mcmaster & Wilhelm, 1997): (3)${\displaystyle GDD=\sum \left[\frac{T_{\max }+T_{\min }}{2}\right]-T_{\mathrm{base}}}$ where GDD, T_max, T_min, and T_base were growing degree days, daily maximum temperature, daily minimum temperature, and base temperature (≥0 °C) for wheat growth, respectively. The data are presented separately for each year because the year and herbicide treatment interaction were significant (P < 0.05).

Wheat yield loss

Wheat yield losses were affected by the timing of weed removal, PHG, and year (Fig. 4). The wheat yield decreased with increasing duration of weed interference during 2021–2023. The grain yields in season-long weed-free plots with PHG were 5.127 t ha⁻¹ in 2021, 7.668 t ha⁻¹ in 2022, and 7.120 t ha⁻¹ in 2023, respectively. The wheat yield in the season-long weedy plots with PHG was 46.1, 75.5, and 71.2% lower than that in the season-long weed-free plots. Similarly, the yields in the season-long weed-free plots without PHG were 5.211 t ha⁻¹, 6.510 t ha⁻¹, and 6.635 t ha⁻¹, respectively, during the same period. The wheat yield in the season-long weedy plots without PHG was 51.3, 47.5, and 76.6% lower than that in the season-long weed-free plots. Notably, the smallest yield decrease was observed in 2022 compared with 2021 and 2023, which was due mainly to weed presence and reduced rainfall (38.93% lower than the average).

Wheat yield loss due to weeds is reported to be 3.0–34.4% in the USA, 2.9% in Canada, and 20–32% in Pakistan (Flessner et al., 2021; Chhokar, Sharma & Sharma, 2012). These rates were lower than our results, which was caused mainly by the specific weed society in our cropping system and rainfall. Under soil moisture deficiency such as in our study, competition between weeds and wheat is greater meaning reduced wheat yields when compared to favorable conditions (i.e., high moisture) (Ihsan, El-Nakhlawy & Ismail, 2015). Webster, Cardina & Loux (1998) investigated the impact of post-harvest herbicide application (glyphosate+2,4-D) at three application times on corn yield and weed biomass in the corn field grown in the winter wheat-corn cropping system. They found that weed biomass was significantly reduced by post-harvest herbicide application without herbicide treatment, as the corn yield was 2.75 times higher than that of non-treated control, similar to our findings.

Critical time for weed removal

The CTWR was calculated using the log–logistic model described based on a 5% acceptable yield loss by Ritz et al. (2015). In the first year of the study, the CTWR began at nearly the same point in both the No PHG and the PHG treatments (Table 4, Fig. 4). Specifically, the CTWR was initiated at 711.9 GDD in the conventional system and 718.5 GDD in the PHG treatment, showing minimal differences between the two treatments in terms of the onset of critical weed pressure. In contrast, the apparent effect of the PHG treatment on the CTWR was observed in the second year. In 2022, the CTWR started at 416 GDD, which corresponds to 16.6 DAE in the No PHG treatment; meanwhile, the CTWR was delayed until 516.5 GDD (22.3 DAE) in the PHG treatment. The PHG treatment delayed the CTWR from 465.6 GDD (14.5 DAE) to 661.2 GDD, equivalent to 31.6 DAE in 2023. This means a substantial delay of 17.1 days compared with the No PHG treatment.

Previous studies have shown that the CTWR varies according to numerous factors, including the growing season, crop species, agronomic practices, and environmental factors (Contreras et al., 2022). However, few studies about CTWR in wheat have been reported in the literature. The CTWR started 2-3 weeks after seed emergence or 28 or 30 days after sowing (Agostinetto et al., 2008; Morsy et al., 2020; Chaudhary et al., 2008). These times agree with our results, which varied from 16.6–31.47 DAE in No PHG treatment.

The relationship between weed abundance and CTWR has been reported in many crops (Charles et al., 2019; Charles et al., 2020; Williams, Ransom & Thompson, 2007). Nearly all the results are more or less the same as weed density decreases, and CTWR is delayed. To reduce weed density in non-herbicide resistant crops, total herbicides might be used at various times, such as pre-sowing, pre-emergence, or post-harvest. These treatments improve weed suppression throughout the stubble or early growth stages and alter the CTWR. Pre-emergence herbicide treatments delay the CTWR by 15–31 d in soybean (Roncatto et al., 2023), 3–21 d in corn (Ulusoy et al., 2021), 6–12 d in sunflower (Knezevic et al., 2013), 31 d in dry bean (Beiermann et al., 2022), and 25–32 d in popcorn (Barnes et al., 2019). The impact of PHG treatment on the CTWR in our study, 5.7–17.1 d, was in line with these studies.