A greenhouse experiment was conducted in the Plant Sciences Center at Auburn University, Auburn, AL. The experiment was conducted two times from June to August 2021. Five replications of each treatment were placed in a randomized block design in each run. The dimensions of soil flats were 55.88 × 29.21 × 3.81 cm. Weed seeds were planted at 70.69, 12.25, 12.25, and 64.32 million seeds ha⁻¹ for Palmer amaranth, ivyleaf morningglory, sicklepod, and large crabgrass, respectively. Weed seeds were thoroughly mixed with organic soil and placed over the top of the soil (sandy loam)-filled flats. Cereal rye biomass was harvested by hand from a field in May 2021, oven-dried for 3 days at 60°C, and then trimmed to match the length of the soil flat. Then, soil flats were covered uniformly by four different biomass levels of rye straw (2,800 kg ha⁻¹ as low, 5,600 kg ha⁻¹ as medium, 8,400 kg ha⁻¹ as high, and 11,200 kg ha⁻¹ as highest) and a check with no residue added.

A field experiment was conducted at E.V. Smith (EVS) Auburn University Research and Extension Center (Field Crops Unit; 32.4417°N, 85.8974°W) near Shorter, Alabama, and Wiregrass Research and Experimental Station (WREC) Headland, AL (31°30′ N, 85°17′ W) from late October 2022 through August 2023. The soil characteristics at the EVS site were Compass sandy loam (coarse-loamy, siliceous, subactive, thermic Paleudults) with pH 6.2 and 0.8% organic matter. At the WREC site, the soil was a Dothan fine sandy loam (fine-loamy, siliceous, thermic Plinthic Paleudult) with pH 6.0 and 1.1% organic matter. The soil sampling depth of pH and organic matter was up to 10 cm.

The experiment was conducted in a split-plot design with six replications; the main plots were four seeding rates (45, 90, 135, and 180 kg ha⁻¹) of cereal rye and a winter fallow check. Plot size measured 20 by 12 feet. In the subplot, flumioxazin herbicide and a non-treated (NT) check were considered. The cereal rye ‘Wrens Abruzzi’ was planted utilizing JD 7730 and a Great PlainsR no-till drill (Great Plains Salina, KS 67401) with GreenStar GPS in the third week of November 2022 at both locations. The cover crop was mechanically rolled via a three-section bar roller-crimper to level the biomass residue on the soil surface in the third week of April 2023. After the mechanical rolling of cover crops, burndown was performed with an application of glyphosate (Roundup Powermax^®; Monsanto Company, St. Louis, MO, USA) 0.91 kg ae ha⁻¹ and glufosinate (Liberty; Bayer Crop Science, Research Triangle Park, NC, USA) 0.57 kg ai ha⁻¹. Flumioxazin (Valor^®, Valent U.S.A. Corporation, Walnut Creek, CA, USA) at 71.5 g ai ha⁻¹ was applied approximately 3–4 weeks after cover crop termination. The herbicide was applied using a CO₂-pressurized backpack sprayer equipped with TTI 11004 nozzles (TeeJet^® Technologies, Glendale Heights, IL, USA) at 276-kPa calibration to deliver 280 L ha⁻¹. At the EVS location, precipitation was received during the week of herbicide application ( Figure 1 ); however, at the WREC location, approximately 12 mm of irrigation was provided within 24 hours of herbicide application.

In the greenhouse experiment, data collection such as weed counts and weed biomass was performed every 10 days for 30 days. Mostly, the emergence of ivyleaf morningglory and sicklepod occurred within 10 days after planting. However, emergence for Palmer amaranth and large crabgrass was also observed at 20 and 30 days after planting, with weed emergence decreasing over time. Subsequently, all weed counts were summed to show the total counts for each weed species in the graphs.

In the field study, visual weed control ratings on a scale of 0–100 (0 = no control and 100 = complete control), weed counts for each species, and weed biomass collection between 3 and 4 weeks after herbicide application based on randomly selected 0.25-m² quadrats per plot at each location was performed. In both greenhouse and field experiments, weed biomass collected samples were placed into a drier at 65°C for 72 h, and then the dry weight of weed biomass was recorded.

Data were analyzed using the PROC GLIMMIX model, and means were separated using Tukey’s honestly significant difference (HSD) post-hoc test at α < 0.05 in the SAS statistical software version 9.4 (SAS Institute, Cary, NC, USA). Normality was checked using the qq plot and Shapiro–Wilk test. There was no statistically significant difference between the two runs in the greenhouse; hence, data were pooled. The fixed effects were cereal rye biomass and herbicide treatment, while replication was considered a random effect. The SigmaPlot software (version 13.0; Systat Software, San Jose, CA, USA) was used for curve-fitting regressions and to estimate coefficient values and coefficient of determination (R²), which were utilized to assess the fitness of each regression curve.

A three-parameter logistic model was used for fitting weed counts and weed biomass for each species individually against cereal rye biomass.

where y is the weed counts and weed biomass, x ⁰ represents the inflection point, b is the slope of the curve or growth rate, α is the asymptote, and x depicts the amount of biomass residue.

All coefficients were found to be statistically significant (p < 0.05), indicating a robust relationship of weed counts and weed biomass with cereal rye residue in the case of Palmer amaranth with R² = 0.99 for both regressions ( Figures 3A, B ; Tables 1 , 2 ). Based on the logistic curve, reducing the Palmer amaranth counts by 50%, 75%, and 90% compared to the fallow treatment required 3,112, 6,520, and 10,490 kg ha⁻¹, respectively, of cereal rye biomass. It was observed that weed counts and weed biomass of Palmer amaranth decreased with increasing biomass of cereal rye residue. A previous study by Ryan et al. (2011) also claimed that increasing cereal rye biomass residue suppressed weeds and decreased weed biomass, including pigweeds (Amaranthus spp.).

However, the slope was non-significant (p > 0.05) for weed counts and weed biomass of ivyleaf morningglory with R² = 0.39 for counts and R² = 0.90 for weed biomass. The slope was non-significant for sicklepod (p > 0.05) with R² = 0.96 for weed counts and R² = 0.92 for weed biomass. Hence, there was no relationship between increasing cereal rye biomass with the reduction of weed counts and weed biomass for ivyleaf morningglory and sicklepod. A decreasing trend in crabgrass counts and biomass was observed with increasing cereal rye biomass, showing R² values of 0.98 for weed counts and 0.87 for biomass.

There was a significant effect of seeding rate on cover crop biomass production; however, the effect of location and their interaction was not significant. The seeding rate of 90 kg ha⁻¹ of cereal rye provided significantly greater cover crop biomass as compared to the 180 kg ha⁻¹ seeding rate ( Figure 4 ).

Palmer amaranth and ivyleaf morningglory were the dominant weeds throughout the field studies at both locations. A logistic three-parameter curve was fitted on weed counts of Palmer amaranth and ivyleaf morningglory and biomass of Palmer amaranth against cereal rye biomass residue.

where y represents counts and biomass of Palmer amaranth and counts of ivyleaf morningglory, x ⁰ represents the inflection point, b represents the slope of the curve or growth rate, α represents the asymptote, and x depicts the amount of biomass residue.

For the morningglory biomass, the Gompertz equation was the best fit and was fitted against cereal rye biomass residue.

where y represents weed biomass for ivyleaf morningglory, x ⁰ represents the inflection point, b represents the slope of the curve or growth rate, α represents the asymptote, and x depicts the amount of biomass residue.

The correlation between cereal rye biomass, visual control rating, weed biomass, and weed counts of morningglory and Palmer amaranth was estimated and represented in a correlation graph ( Figure 6 ). In the graph, the size of the circle and the color intensity in the circle were used to depict the correlation strength, with bigger circles and darker colors indicating stronger correlations between variables. The blue color indicates a positive correlation, while the red color shows a negative correlation between variables. Furthermore, the correlation is represented from 1 to −1, in which 1 means a positive correlation, −1 means a negative correlation, and 0 means no correlation between variables.

We examined the interaction of seeding rate with herbicide treatments regarding weed control. The results showed that across all seeding rates of cover crops, the effect of weed control was similar for the herbicide treatment. Furthermore, biomass residue was grouped into categories (<2,000 kg ha⁻¹ as A, 2,000–4,000 kg ha⁻¹ as B, 4,000–6,000 kg ha⁻¹ as C, and >6,000 kg ha⁻¹ as D) to assess the interaction of herbicide with levels of biomass residue for weed control; however, no significant effect was found. Herbicide wash-off from cover crop residue is expected due to precipitation/irrigation after herbicide application, leading to its release into the soil. The overall effect of herbicide was found significant (p < 0.01). It has been observed that Palmer amaranth control was 95% when treated with flumioxazin herbicide, while the biomass residue of NT checks only provided approximately 50% control of Palmer amaranth at 7 weeks after cover crop termination ( Table 4 ). However, 4 weeks after cereal rye termination, it was observed that cereal rye provided up to 70% control of Palmer amaranth (data not shown). Significantly lesser counts of Palmer amaranth were found in herbicide-treated plots compared to the NT check. Therefore, herbicide is still required to achieve excellent control of this troublesome weed. In the case of the flumioxazin-treated check, we found that it worked well regardless of cover crop biomass. According to Reeves et al. (2005), at 60 days after planting, weed control achieved using cereal rye with preemergence herbicide was comparable to that achieved in winter fallow checks using both preemergence and postemergence herbicides.

Ivyleaf morningglory control was 90% when treated with flumioxazin while 30% under NT check. Similarly, fewer counts of ivyleaf morningglory were observed in plots that received flumioxazin herbicide compared to the NT check. Previous research studies have also found that preemergence-applied flumioxazin controls many broadleaf weed species such as morningglories and pigweeds (Cranmer et al., 2000; Wilcut et al., 2000). Specifically, flumioxazin as preemergence showed high efficacy and provided 94% control of Palmer amaranth at 4 weeks after application (Whitaker et al., 2010).

The incorporation of cereal rye into the cropping system would not only provide weed suppression but also offer significant soil health benefits, given the importance of cover crops. Furthermore, the integration of cereal rye with preemergence-applied flumioxazin herbicide could be an effective strategy to control glyphosate-resistant Palmer amaranth, which is a major challenge for growers in the Southern USA. According to Palhano et al. (2018), the incorporation of a highly productive cover crop system in combination with preemergence herbicides not only offers early-season weed control but also provides flexibility in scheduling postemergence herbicide application. Additionally, a previous research study at the same site has also demonstrated that cereal rye, in combination with herbicides, provided greater weed control compared to winter fallow with herbicide treatment (Kumari et al., 2024b). Hence, adopting a cereal rye cover crop would be a better tool for weed management.