Commom ragweed interference in soybean: critical threshold and agronomic mechanismsAbstract
Background: Common ragweed (Ambrosia artemisiifolia) is an aggressive weed with tolerance and resistance to multiple herbicides, hindering chemical control and raising concerns among soybean growers due to its growth and stature. Nevertheless, studies defining a critical interference threshold and a elucidating how this weed impacts soybean yield formation are lacking.
Objective: To determine the critical interference threshold of common ragweed on soybean yield and to characterize the agronomic mechanisms involved.
Methods: Experiments were conducted at four fields with seven weed densities (ranging from 0 to 32 plants m-2), arranged in a completely randomized design with five replicates. Weed growth traits and soybean yield components were analyzed. A regression analysis was performed to identify the critical threshold, and agronomic mechanisms were analyzed through multiple regression and principal component analysis.
Results: Regression analysis, using pooled data from the four fields, identified a breakpoint in soybean yield loss at 2.47 ± 0.52 plants m-2. Below this threshold, each additional common ragweed plant reduced soybean yield by 10.72%, whereas above this density, yield loss was reduced to 1.45% per additional plant, indicating asymptotic competition. At the critical threshold (2.47 plants m-2), soybean yield loss reached ≈26.5%.
Conclusions: Our findings indicate that there is non detrimental infestation density for this species, as the presence of one plant per square meter can result in yield reductions exceeding 10%. These results highlight the need for early preventive weed management strategies in soybean fields.
Keywords:
Soybean; Yield Components; Preventive Management; Asymptotic Competition; Crop Losses
1. Introduction
Ambrosia artemisiifolia L. (common ragweed) is an annual weed species widely distributed in the Northern Hemisphere that has rapidly expanded into subtropical and tropical climates, including agricultural areas of South America (Song et al., 2023). In Brazil, the occurrence of this species in soybean fields has become increasingly frequent, particularly in the southern and central-western regions, where farmers report escalating infestations and growing difficulties in achieving chemical control. Invasive weed species are a major and growing threat to agricultural production (Schuster et al., 2026), undermining yields and driving up management costs.
This scenario is exacerbated by documented multiple herbicide resistance in this species, particularly against ALS, PPO, and EPSPS inhibitors in the United States (D’Amico Jr. et al., 2018). In Brazil, growers are reporting escalating control failures, raising suspicion of herbicide resistance and indicating rapid geographic spread.
Studies conducted in the Northern Hemisphere indicate that common ragweed possesses high competitive potential, significantly reducing soybean yield (Coble et al., 1981; Cowbrough et al., 2003; Hall et al., 2020). Cowbrough et al. (2003), demonstrated that ragweed densities between 0.01 and 0.14 plants m-2 justify control interventions, based on economic threshold estimates adjusted for spatial aggregation of infestations.
These findings highlight the aggressive competitive ability of common ragweed, even at low infestation densities, in the need support for practical weed management decisions. A these estimates are based primarily on economic thresholds, they remain strongly influenced by contextual factors such as grain prices, control costs, and production scale butdo not directly capture the physiological impacts of weed interference on crop yield.
Using its widespread distribution and recognized competitive impact, studies that define functional interference thresholds critical density levels from which measurable physiological and agronomic yield losses begin to occur remain scarce. Moreover, there is a particular lack of studies employing universally applicable agronomic variables; that is, objective indicators that are easily measured in the field and biologically interpretable, such as plant height, biomass, and yield components.
Adopting these types of variables allows for more precise diagnosis of weed interference and enhances comparability among studies conducted across different regions worldwide, ultimately facilitating the formulation of more generalized and practical management recommendations.
In addition to direct competition for resources such as light, water, and nutrients, common ragweed may affect soybean crops through additional physiological pathways, including allelopathy (Formigheiri et al., 2018) and reduced nodulation by Bradyrhizobium japonicum, potentially compromising symbiosis and crop yield (Hall et al., 2020). Such effects may be expressed early in crop development, as suggested by growth analyses conducted by Rockenbach and Rizzardi (2024), which reported rapid biomass accumulation and marked competitive aggressiveness of the Ambrosia artemisiifolia.
Given this context, the objectives of this study were to: (i) quantify the effects of common ragweed interference on soybean yield across different agricultural environments; (ii) identify the soybean interference threshold for common ragweed density through segmented regression; and (iii) characterize the agronomic mechanisms associated with yield losses using multivariate approaches and predictive models. By integrating robust statistical modeling and field validation, we aim to provide biologically interpretable and operationally relevant guidance for preventive management of one of the most problematic emerging weed species.
2. Materials and Methods
2.1 Experimental sites and design
Three experiments were conducted during the 2023/2024 growing season in the municipality of Lapa, Paraná, Brazil. The experimental areas were located on commercial farms with a known history of natural infestation by common ragweed. The geographic coordinates and altitudes of the sites were as follows: Field 1 (25°40′33.55″ S, 49°39′48.55″ W, 930 m), Field 2 (25°39′17.69″ S, 49°57′44.10″ W, 823 m), and Field 3 (25°42′02.60″ S, 49°59′26.94″ W, 944 m).
The regional climate is categorized as Cfb (humid mesothermal with mild summers), with an average annual rainfall between 1,400 and 1,600 mm and average monthly temperatures below 22 °C (Instituto Nacional de Meteorologia, 2025). Weather data recorded during the soybean growing season are shown in Figure 1, with a mean temperature of 20.9 °C and total precipitation of 1,251 mm throughout the crop development period.
Monthly averages of temperature (°C) and precipitation (mm) during the soybean growing season, used to assess common ragweed interference. Lapa (Paraná), 2023/2024 season
At each site, a completely randomized design was adopted with seven treatments representing different densities of common ragweed (0, 1, 2, 4, 6, 8, and 10 plants m–2), with five replicates per treatment, totaling 35 experimental units per location. Each plot was defined as two soybean rows, 1 m long and 0.5 m apart, totaling 1 m2 of usable area. Common ragweed plants were established at the exact target densities, positioned centrally within each unit. This arrangement minimized border effects, ensured uniform interference conditions, and justified the use of a completely randomized design. To complement and strengthen the dataset, additional data from an location independent trial conducted in Vitorino (Paraná, Brazil) were included. This trial evaluated only the effect of common ragweed density on soybean yield, using treatments with 0, 5, 10, 15, 20, 25, and 30 plants m–2.
2.2 Field operations and management
Plots were established based on the natural infestation of the weed. Treatment assignment was performed through prior visual mapping, meaning the target densities for each treatment. This approach ensured that interactions between the crop and the weed developed naturally throughout the growing cycle. No other weed species were present within the delineated plots.
Soybean sowing was carried out between October and November 2023, according to each farm’s operational schedule (Table 1), using the cultivars Brasmax Zeus IPRO® (Fields 1 and 2) and GH 5933 IPRO® (Field 3), with an average planting density of 280,000 to 300,000 plants ha–1 and a row spacing of 0.5 m. For field 4, planting was conducted in November 2022 using the cultivar Vênus CE®, eith a population of 280,000 plants ha–1 and row spacing of 0,5 m (Table 1). Fertilization followed the technical recommendations for each site, with formulated fertilizers (06-21-12 and 10-50-00) applied at sowing and potassium chloride (00-00-60) applied as a topdressing after crop emergence.
Soil physical and chemical properties and soybean cropping information for the experimental sites
The physical and chemical properties of the soils in the three experimental areas. Prior to soybean planting, soil samples were collected from the 0–20 cm layer across all experimental sites.
Weed management prior to soybean emergence involved the application of individual or tank-mixed herbicides, following the standard practices adopted by each grower (Table 2). Management strategies varied across sites and were later considered as interpretative factors for the observed responses. In other words, the yield losses reported here reflect interference that occurred despite attempts at weed control.
Herbicide management practices before and after soybean planting across the experimental sites. Lapa (Paraná), 2023/2024 season
Trade names: Glyphosate (Crucial); Halauxifen + Diclosulam (Paxeo); Saflufenacil (Heat); Glufosinate (Finale); Fomesafen (Flex); 2,4d (Aminol); Flumioxazin + imazethapyr (Zethamaxx).
2.3 Agronomic evaluations
Evaluations were carried out at the R8 soybean growth stage (full maturity). For common ragweed, the following traits were recorded: plant density, height, stem diameter, and shoot dry biomass. Biomass was determined after drying at 60 °C to constant weight in a forced-air circulation oven. Plant height was measured from the soil surface to the highest point of the plant, averaging all individuals within each plot except for plots with more than four plants, in which case the average was based on four randomly selected individuals. The same criterion was applied for measuring stem diameter. Dry biomass was assessed by harvesting all common ragweed plants present in each plot.
For soybean, the following variables were assessed: final plant density (plants m–2), number of pods per plant, plant height (cm), stem diameter (mm), thousand-seed weight (g), height of the first pod (cm), and grain yield (kg ha–1). The height of the first pod was measured from the soil surface. Thousand-seed weight was estimated from the average of four subsamples of 100 seeds each, weighed on an analytical balance and extrapolated to 1,000 seeds. The number of pods, stem diameter, and height of the first pod were determined based on the average of four plants per plot for all treatments.
2.4 Statistical analysis
All statistical analyses were performed in R (version 4.4.3). In this study, we use the term interference threshold to refer to the weed density at which the slope of the soybean yield-loss response changes, as estimated by segmented regression. The term breakpoint is used only for the statistical estimate of that change point. Thus, the interference threshold represents the agronomically relevant density derived from the fitted breakpoint.
The analytical sequence was designed to separate two complementary objectives: (i) identification of the interference threshold for soybean yield loss, and (ii) characterization of the agronomic mechanisms associated with that yield loss. First, soybean yield loss was modeled as a function of common ragweed density using segmented regression, because this approach allows estimation of a change point in the slope of the response curve and is therefore appropriate for identifying the interference threshold. The segmented model was fitted from an initial linear model, and the breakpoint was estimated iteratively using the segmented function. Model support was assessed relative to the simple linear model using the likelihood ratio test (LRT), Akaike information criterion (AIC), and changes in the coefficient of determination (R2), following the weed-interference regression framework proposed by Cousens (1985) and later adapted by Agostinetto et al. (2018).
To support the choice of the predictor used in threshold modeling, we also compared common ragweed density and shoot biomass as alternative explanatory variables for soybean yield loss (Supplementary Material 1). Although both variables were biologically informative, plant density provided the more stable and operationally relevant threshold model and was therefore retained as the primary predictor for estimating the interference threshold.
After identifying the interference threshold, we analyzed the agronomic variables underlying yield loss. Soybean traits (number of pods per plant, final plant density, thousand-seed weight, plant height, and stem diameter) together with weed traits (shoot biomass) were examined using principal component analysis (PCA) based on standardized data. PCA was used as an exploratory tool to describe the multivariate structure of crop–weed interference and to identify which variables were most strongly associated with yield loss across environments. The ordination was visualized with biplots, and field-related multivariate separation was evaluated by PERMANOVA with 999 permutations. K-means clustering was used only as a descriptive complement to visualize similarity among plots in multivariate space; it was not used to define the interference threshold.
Finally, to quantify the direct contribution of the agronomic variables to soybean yield loss, a multiple linear regression model was fitted. Regression coefficients, p-values, and confidence intervals were obtained, and marginal effects were visualized using plot_model from the sjPlot package. In addition, Random Forest (randomForest) and partial least squares regression (PLS; plsr) with cross-validation were used as complementary predictive approaches to test whether the variables highlighted by the regression and PCA remained consistently important across different modeling frameworks. Predictive performance was evaluated using root mean square error (RMSE) and coefficient of determination (R2) in the test set.
3. Results
3.1 Interference threshold and dual-phase yield response to weed density
Segmented regression analysis using the combined dataset from the four experimental fields revealed a nonlinear relationship between common ragweed density and soybean yield reduction (Figure 1). The model estimated an interference threshold at 2.47 plants m–2 (± 0.52 standard error), clearly indicating the presence of two distinct phases in the crop’s response to weed infestation.
In the initial phase, below the breakpoint, each additional common ragweed plant caused an average soybean yield reduction of approximately 10.72% per plant (p < 0.001). Above this interference threshold, the slope decreased to 1.45% per additional plant, suggesting a pattern of asymptotic competition at higher weed densities.
The fixed effects associated with the experimental fields were highly significant. Fields 2 and 3 showed significantly lower yield reductions compared to Field 1 (estimated coefficients of –10.88 and –13.83, respectively; both p < 0.01), while Field 4 did not differ statistically from Field 1 (p = 0.82). No significant effects were observed for the experimental blocks (p > 0.05 in all cases).
The model exhibited good explanatory power, with an adjusted R2 of 0.566, indicating that approximately 57% of the observed variability in yield loss could be attributed to common ragweed density and the fixed effects of the experimental fields.
It is noteworthy that the breakpoint occurred at a relatively low density (2.47 plants m–2), suggesting that soybean is highly sensitive to even low levels of infestation. This finding reinforces the importance of implementing early and effective management strategies to prevent yield losses caused by this weed.
Additionally, a comparative analysis was performed between two potential predictors of soybean yield loss: common ragweed plant density and accumulated shoot biomass. This analysis, detailed in Supplementary Material 1, indicated that plant density showed superior statistical performance and greater operational applicability, and was therefore selected as the primary predictive variable in the segmented regression model.
3.2 Interference mechanisms: multivariate patterns
PCA was conducted to investigate multivariate response patterns of soybean agronomic variables and common ragweed traits, with a focus on their interference effects on crop performance. The PCA produced two main components (PC1 and PC2), which explained 52% and 23% of the total data variance, respectively, accounting for 75% of the variability represented in the first two multivariate dimensions.
The PCA biplot (Figure 2) revealed a distinct distribution of experimental fields in multivariate space, reflecting contrasting agronomic profiles among environments. Field 1 concentrated most of the plots with the greatest yield losses and the highest values of common ragweed biomass, height, and stem diameter. In contrast, Fields 2 and 3 clustered in distinct and opposite regions of the main plane, indicating more stable conditions and less impact from weed interference. This pattern was supported by K-means clustering analysis, which identified three main clusters that partially coincided with the experimental fields, although some overlap was observed among plots in Field 2.
Segmented regression curve fitted to the relationship between common ragweed density and soybean yield reduction, with a breakpoint at 2.47 plants m–2.
The greater yield losses observed in Field 1 may be attributed to differences in sowing date and herbicide management compared to Field 2, despite both using the same soybean cultivar. In Field 2, the pre-emergence application occurred closer to sowing, and different active ingredients were used in mixture, which may have reduced common ragweed emergence and/or development. In Field 3, more pronounced changes were observed, including a different sowing date and cultivar, and no pre-emergence herbicides. However, the later planting date may have allowed the crop to avoid the ideal development window for common ragweed, potentially limiting its growth. The average temperature during the first 30 days of soybean development in Field 3 was 22.3 °C, which may have hindered shoot and root development of common ragweed. According to Knolmajer et al. (2024), optimal development of both shoot and root systems occurs within the temperature range of 29.5 °C to 31.4 °C. Moreover, precipitation levels during this early period were lower compared to Fields 1 and 2, where greater soil water availability favored rapid weed growth.
This separation among fields was statistically confirmed by permutational multivariate analysis of variance (PERMANOVA), which showed a highly significant effect of the “field” factor on the multivariate structure of the analyzed variables (R2 = 0.32; p < 0.001), confirming that environmental and agronomic conditions directly influenced the observed responses.
For the variables that most contributed to plot differentiation, the vector for yield reduction was oriented in the opposite direction of key soybean morphological and yield traits, such as number of pods (numberofpod), final plant density (soybeandensity), thousand-seed weight (TSW), and stem diameter (diameterS). This configuration confirms a negative correlation between weed interference and crop performance, suggesting that common ragweed simultaneously compromises soybean stand, fecundity, and grain filling.
The vector magnitudes for numberofpod, TSW, and biomassC indicate that these were the main structuring variables of PC1, serving as sensitive indicators of interference intensity. Therefore, the biplot not only illustrates the separation among environments but also highlights the agronomic mechanisms directly affected by weed presence (Figure 3).
Biplot from the principal component analysis (PCA) of soybean agronomic variables and common ragweed attributes. Points represent experimental plots distributed across three distinct fields (Field 1, Field 2, and Field 3), while vectors indicate the contribution of each variable to the multivariate structure
3.3 Predictive models: multiple regression, Random Forest, and PLS
The multiple linear regression model fitted to quantify the individual contribution of soybean agronomic variables and weed infestation traits to yield loss, was significant (p < 0.001) and explained 53.4% of the variation in soybean yield reduction (R2 = 0.534; Figure 1).
Among the evaluated predictors, the number of pods per plant remained the strongest negative indicator of yield loss (β = –0.535, p < 0.001), followed by soybean stand density (β = –1.178, p = 0.001) and thousand-seed weight (TSW) (β = –0.497, p = 0.002). These results confirm that weed interference simultaneously compromises crop establishment, fecundity, and grain filling.
In addition, common ragweed biomass showed a positive and significant effect on yield loss (β = +0.023, p = 0.026), reinforcing its role as a direct indicator of competitive intensity in the analyzed environments. Soybean plant height was negatively associated with yield reduction (β = –0.113, p = 0.009), indicating that shorter plants suffered greater interference from the weed. This suggests that competition for resources especially light and nutrients limited vegetative growth, impairing the crop’s ability to maintain vigor and productivity. On the other hand, soybean stem diameter had no effect (p = 0.448) and was not considered a reliable predictor within the model structure.
Complementary predictive models were fitted to confirm the robustness of the results (Supplementary Material 2). Partial least squares regression (PLS) explained 52% of the variance in the response variable using two principal components. The PLS model coefficients confirmed the same key predictors previously identified, reinforcing the statistical and agronomic consistency of the observed effects. Additionally, the Random Forest model indicated that numberofpod, soybendensity, TSW, and biomassC were the variables with the highest relative importance for predicting yield loss, with moderate predictive performance (R2 = 0.39; RMSE ≈ 14.2%).
The correlation matrix between soybean agronomic traits and common ragweed infestation attributes revealed consistent patterns of association between weed interference and crop performance (Supplementary Material 2). The variable yieldReduction showed significant negative correlations with absolute yield (yield, r = –0.93), number of pods per plant (numberofpod, r = –0.43), soybean plant density (soybendensity, r = –0.31), and thousand-seed weight (TSW, r = –0.30), confirming that weed interference compromises multiple yield components.
Conversely, yield reduction was positively correlated with common ragweed biomass (biomassC, r = +0.52) and stem diameter (diameterC, r = +0.39), reinforcing that infestation intensity is directly associated with the magnitude of yield losses. These multivariate patterns corroborate the findings from PCA and regression analyses, indicating that soybean yield is impacted by multiple, simultaneous competition mechanisms.
4. Discussion
The emergence of common ragweed as an aggressive agent, very difficult to control chemically in the field, with suspected resistance in Brazilian soybean fields and rapid geographic expansion, represents a growing challenge for integrated management. Although its impact has already been recognized in international contexts, experimental evidence quantifying its interference under field conditions in Brazil is still scarce. This study makes a significant contribution in this regard by demonstrating that even low initial densities of common ragweed result in substantial yield losses, with early interference affecting multiple reproductive traits of soybean. Jordan et al. (2021) reported that the rapid emergence and vigorous growth of common ragweed make early control essential to avoid losses in soybean production. In line with our segmented regression, multivariate analysis, and complementary data validation, our results reveal that soybean is highly sensitive to the presence of this species even before competition for resources is fully established — as evidenced by significant reductions in final plant density, number of pods, and thousand-seed weight, even at low infestation levels. These findings suggest that interference begins early, possibly due to allelopathy or competition during the crop establishment phase, underscoring the importance of preventive actions and conservative control thresholds for this emerging weed.
Table 3 Coefficients of the multiple linear regression model for predicting soybean yield loss| Variable | Estimate | p-value | Interpretation |
|---|---|---|---|
| Intercept | 155.94 | < 0.001 | Estimated maximum yield (in the absence of interference, with all predictors at zero) |
| Number of pods | –0.535 | < 0.001 | Each additional pod per plant lost → +0.54% yield reduction |
| Soybean plant density | –1.18 | 0.0012 | Each plant per m2 lost → +1.18% yield reduction |
| Thousand-seed weight | –0.497 | 0.0024 | A 1 g reduction in TSW → +0.5% yield reduction |
| Common ragweed biomass | +0.023 | 0.0260 | Each additional gram of common ragweed biomass → +0.023% yield reduction |
| Soybean height | –0.113 | 0.0088 | Each 1 cm decrease in soybean height → +0.11% yield reduction |
| Soybean stem diameter | +1.02 | 0.448 | Not significant (p = 0.45) |
The analyses also showed that the three experimental fields exhibited distinct agronomic profiles, largely as a result of differences in management practices. The most notable contrast, particularly between Fields 1 and 2, was related to the use of pre-emergence herbicides. While Field 1 received a single active ingredient, Field 2 was treated with a mixture of two pre-emergence herbicides, which may have delayed weed emergence and, consequently, postponed the critical period for weed control (CPWC). These results support the findings of Roncatto et al. (2023), who demonstrated that mixtures of pre-emergence herbicides are effective strategies for broadening the control spectrum, extending residual activity, and mitigating the effects of weed interference in soybean.
4 .1 Critical interference threshold for common ragweed
Under the evaluated conditions, we found no yield-neutral infestation level. Segmented regression identified a yield-loss breakpoint at 2.47 ± 0.52 plants m–2 that marks a change in slope—not the onset of interference: losses accrue steeply below the breakpoint (≈10.72% per additional plant) and continue to accumulate above it at a lower marginal rate (≈1.45% per plant). Consequently, even minimal, early-season infestations can compromise yield. This risk is compounded by the species’ high dispersal potential, especially when plants reach reproductive maturity by harvest, facilitating spread during harvest operations. In Nebraska, USA, Barnes et al. (2018) reported soybean yield losses of 25–61% from a single common ragweed plant per meter of row, with light competition identified as the primary interference mechanism— evidence of the species’ exceptional competitive ability under certain environments.
The steep slope observed before the inflection point indicates that yield loss increases more sharply at low infestation levels, while above the threshold the curve begins to level off, suggesting an asymptotic competition effect. This behavior characterizes early interference as the main cause of yield reduction and reinforces the need for timely weed control.
Evidence from Canada (Cowbrough et al., 2003) and from our Brazilian study shows that soybean is highly sensitive to interference by common ragweed even at very low densities, highlighting the species’ ecological plasticity across contrasting edaphoclimatic conditions while maintaining strong competitive effects. Cowbrough et al. estimated economic thresholds as low as 0.01–0.14 plants m–2, depending on the acceptable level of economic loss. The convergence on very low thresholds, despite the different methodological bases (economic vs. functional), underscores the pronounced competitive potential of common ragweed.
The strong competitive ability of common ragweed likely stems from its high absolute and relative growth rates in both the shoot and root systems. Rapid canopy development enables efficient light interception and early shading, while vigorous root growth enhances capture of soil resources; together these traits confer a marked advantage over soybean (Rockenbach, Rizzardi, 2024). Furtheremore, in a study conducted in Austria, researchers found that a single common ragweed plant per square meter was capable of reducing the average weight and number of Bradyrhizobium japonicum nodules by 56%, resulting in an 18% reduction in final soybean yield.
In our dataset, ragweed shoot biomass emerged as one of the most influential variables across both multivariate and predictive analyses (Supplementary Material 2), underscoring dry-mass accumulation as a key mechanism of agronomic interference. For example, at a density of 1 plant m–2, common ragweed reduced soybean yield by 10.72%, whereas Amaranthus hybridus at the same density caused only 4.32–5.09% losses (Sossmeier et al., 2025). Losses increased steeply up to the critical density (~2.47 plants m–2), at which the predicted cumulative reduction reached ≈26.5%, and continued to accrue beyond that point at a lower marginal rate (as shown above). The rapid early development of both shoots and roots documented by Rockenbach and Rizzardi (2024) is consistent with these patterns and supports a practical recommendation for very early intervention—ideally immediately post-emergence— before rapid biomass accumulation and canopy closure intensify competition.
4.2 Competition-mediated impacts on soybean stand, growth, and yield formation
Multivariate and predictive analyses allowed us to infer the most probable agronomic mechanisms associated with common ragweed interference in soybean. The observed patterns indicated consistent effects on crop establishment, fecundity, and grain filling, suggesting that interference occurs through multiple morphophysiological pathways. PCA showed that common ragweed shoot biomass was one of the main vectors of variation across plots, standing out as a functional indicator of competitive intensity.
The shoot growth of common ragweed is considered rapid: at 25 days after emergence (DAE), plants accumulated 0.83 g of dry mass, which increased to 6.64 g at 35 DAE—an increment of 0.58 g per day. However, daily accumulation up to 25 DAE was only 0.03 g, indicating that growth intensifies substantially after this period, highlighting the species’ fast and aggressive development (Rockenbach, Rizzardi, 2024). Furthermore, the field separation was statistically confirmed by PERMANOVA, demonstrating that, despite environmental variation, the interference mechanisms were consistent.
D’Amico Jr. et al. (2018) reported that populations of common ragweed resistant to ALS, PPO, and EPSPS inhibitors in the eastern United States were capable of reducing soybean yield by more than 40% with just two plants per meter of row. In our study, conducted in Brazil, we identified a functional interference threshold as low as 2.47 plants m–2, associated with substantial yield losses. We estimated that two plants m–2 caused an approximate 17% reduction in soybean yield. Although this value is lower than that reported in the U.S., it still represents a substantial and concerning agronomic impact—especially considering the extremely low infestation level. Notably, the biotype evaluated in this study also showed poor control by glyphosate and has been described as highly aggressive in commercial fields.
Despite differences in climate, management practices, and soybean genotypes, the data converge on a common pattern: when common ragweed is present from the early stages of the crop, it severely compromises soybean performance—even at low densities. This consistency across biotypes and geographic regions reinforces the need for preventive management strategies tailored to local contexts. More efficient control alternatives include the use of pre-emergence herbicide mixtures, as applied in Field 2, and delayed sowing dates that avoid peak periods of weed development. These strategies were associated with lower yield losses in our study. Chudzik et al. (2025) demonstrated that soil management practices and the timing of soybean planting are critical in systems where weed biotypes exhibit an extended emergence window.
Multiple regression, Random Forest, and PLS models consistently identified the number of pods per plant, soybean stand density, thousand-seed weight (TSW), and common ragweed biomass as the variables most strongly associated with yield loss. The convergence of these approaches demonstrates that interference compromises crop establishment, fecundity, and grain filling—i.e., multiple physiological stages of soybean development.
Our results contrast with those of Coble et al. (1981), who reported that the presence of common ragweed during the first six weeks after soybean emergence did not significantly reduce crop yield. Although our experiment did not isolate specific periods of weed–crop coexistence, analysis of agronomic variables revealed a significant reduction in soybean stand density with increasing weed density. This effect indicates that interference began in the early stages of the cycle, compromising crop establishment a critical yield component. Thus, even if common ragweed does not immediately reduce yield in the short term, as suggested by Coble et al., our data show that it can indirectly affect yield by limiting early crop stand, with downstream impacts on reproductive components. This difference in interpretation may be related to environmental conditions, soybean genotypes, or the simultaneous emergence of weed and crop in our study.
In addition to the competitive effects obser ved in the field, laboratory studies conducted by Formigheiri et al. (2018) demonstrated that aqueous extracts of common ragweed exhibit allelopathic effects on soybean seed germination and early seedling development, significantly reducing germination speed, radicle length, and seed vigor. Although our study was conducted under field conditions and did not isolate allelopathic effects, the reduction in soybean stand density observed in infested areas suggests that, beyond resource competition, common ragweed may exert direct physiological effects on crop establishment establishment, increasing its aggressive potential. These findings support the hypothesis that interference begins early and through multiple pathways, ultimately compromising soybean stand and, consequently, yield components.
4.3 Implications for management and research
The identification of a steep yield loss per additional plant up to the threshold of 2.47 plants m–2 provides a strong empirical basis for adopting stricter management strategies against common ragweed. The absence of a “safe tolerance level” suggests that even the mere presence of this species should be considered an agronomic risk, particularly in regions where early emergence coincides with sensitive crop stages. Recent syntheses for common ragweed also recommend pairing pre-plant/pre-emergence residual herbicides with early detection and spatially targeted control to curb early-season interference and to limit resistance selection pressure through diversified modes of action (Knolmajer et al., 2025).
Our findings further confirm the aggressiveness of even a single common ragweed plant and highlight the need for zero-tolerance approaches in infested soybean fields (Hall et al., 2020). Moreover, the species’ long seed persistence and high dispersal potential mean that even small escapes can sustain infestations; harvest-time sanitation and roguing of late cohorts are therefore critical complements to chemical control (Knolmajer et al., 2025). Taken together, these observations underscore the need to anticipate emergence patterns and to implement integrated programs as early as possible ideally before crop establishment combining residual herbicides with cultural practices that accelerate canopy closure and with hygiene measures that prevent seed set and spread (Knolmajer et al., 2025).
The robustness of our results demonstrated by the consistency of explanatory variables across multiple analytical methods supports the recommendation of using weed biomass as an indicator of competitive intensity in multivariate studies, while plant density remains the most practical and operational metric for defining agronomic thresholds in the field.
Ultimately, the findings reinforce the necessity of early monitoring and preventive control, prioritizing integrated approaches that combine chemical management, soil coverage, and cultural practices that promote rapid soybean canopy closure. This study also opens avenues for future research focused on additional physiological variables such as light interception, photosynthetic efficiency, and stress responses under extreme conditions (e.g., drought or heat) to deepen the understanding of weed interference mechanisms. Such variables are particularly relevant under ongoing climate change scenarios, as altered temperature and precipitation patterns may shift competitive dynamics and affect the efficacy of both cultural and chemical control strategies.
5. Conclusions
This study demonstrated that common ragweed significantly interferes with soybean productivity, with yield losses occurring even at an infestation density as low as one plant per square meter. Yield reduction increased rapidly with rising weed density up to an interference threshold of 2.47 plants m–2, beyond which losses continued to accumulate, albeit at a slower rate, indicating there is no safe level of infestation for this species, as even minimal presence can substantially affect crop performance.
Multivariate and predictive analyses converged to show that interference by common ragweed compromises multiple soybean yield components—especially pods per plant, final plant density, and thousand-seed weight— and is tightly associated with accumulated weed biomass. The agreement across analytical approaches reinforces the robustness of these findings.
Interference is established early in the crop cycle and acts through multiple morphophysiological pathways, underscoring the need for proactive, integrated management. Early scouting, prevention of biomass buildup, strategic use of pre-emergence residual herbicides, maintenance of soil cover, and cultural practices that hasten soybean canopy closure are essential to keep common ragweed below the interference threshold. From a practical management perspective, infestations approaching 2 plants m–2 should already be considered a warning level for intervention. This means that soybean growers should act before common ragweed biomass accumulates and crop reproductive traits are irreversibly affected.
Supplementary Materials
Supplementary material 1
Acknowledgments
We thank the farmers who generously provided their land, making this research possible.
-
Funding
This research did not receive external funding.Additional support was provided by the Coordination for the Improvement of Higher Education Personnel (CAPES) through master’s fellowships granted to R.S.T and R.A.W, and a postdoctoral fellowship granted to M.Z.S.A.A.M.B. is a research productivity fellow of the National Council for Scientific and Technological Development (CNPq).
Data Availability
The datasets generated and analyzed during this study are not publicly available due to confidentiality.
References
-
Agostinetto D, Silva DRO, Vargas L. Soybean yield loss and economic thresholds due to glyphosate resistant hairy fleabane interference. Arq Inst Biol. 2018;84(1):1-8. Available from: https://doi.org/10.1590/1808-1657000022017
» https://doi.org/10.1590/1808-1657000022017 -
Barnes ER, Jhala AJ, Knezevic SZ, Sikkema PH, Lindquist JL. Common ragweed (Ambrosia artemisiifolia L.) interference with soybean in Nebraska. Agron J. 2018;110(2):646-53. Available from: https://doi.org/10.2134/agronj2017.09.0554
» https://doi.org/10.2134/agronj2017.09.0554 -
Chudzik G, Nunes JJ, Arneson NJ, DeWerff RP, Ferreira VS, Proctor C et al. Postemergence giant ragweed management as affected by soil and cover crop management, soybean planting time, and preemergence herbicide application. Weed Technol. 2025;39:1-8. Available from: https://doi:10.1017/wet.2024.110
» https://doi.org/10.1017/wet.2024.110 -
Coble HD, Williams FM, Ritter RL. Common ragweed (Ambrosia artemisiifolia) interference in soybeans (Glycine max). Weed Sci. 1981;29(3):339-42. Available from: https://doi.org/10.1017/S0043174500062081
» https://doi.org/10.1017/S0043174500062081 -
Cousens R. An empirical model relating crop yield to weed and crop density and a statistical comparison with other models. J Agric Sci. 1985;105(3):513-21. Available from: https://doi.org/10.1017/S0021859600059396
» https://doi.org/10.1017/S0021859600059396 -
Cowbrough MJ, Brown RB, Tardif FJ. Impact of common ragweed (Ambrosia artemisiifolia) aggregation on economic thresholds in soybean. Weed Sci. 2003;51(6):947–54. Available from: https://doi.org/10.1614/02-036
» https://doi.org/10.1614/02-036 -
D’Amico F, Besanҫon T, Koehler A, Shergill L, Ziegler M, VanGessel M. Common ragweed (Ambrosia artemisiifolia L.) accessions in the Mid Atlantic region resistant to ALS, PPO, and EPSPS inhibiting herbicides. Weed Technol. 2024;38(1):1-10. Available from: https://doi.org/10.1017/wet.2024.11
» https://doi.org/10.1017/wet.2024.11 -
Formigheiri FB, Bonome LTS, Bittencourt HH, Leite K, Reginatto M, Giovanetti, LK. [Allelopathy of Ambrosia artemisiifolia on germination and growth of maize and soybean seedlings]. Rev Cienc Agrar. 2019;41(3):151-60. Portuguese. Available from: https://doi.org/10.19084/RCA18074
» https://doi.org/10.19084/RCA18074 -
Hall, RM, Wagentristl H, Karrer G, Winter A, Czerny R, Kaul HP. Common ragweed (Ambrosia artemisiifolia L.) causes severe yield losses in different soybean varieties by reducing the infection potential of Bradyrhizobium japonicum Julius-Kühn-Arch. 2020;464:1-8. Available from: https://doi.org/10.5073/jka.2020.464.009
» https://doi.org/10.5073/jka.2020.464.009 -
Instituto Nacional de Meteorologia – INMET. [Automatic station A807]. Curitiba; 2000. Portuguese. Available from: https://tempo.inmet.gov.br/TabelaEstacoes/A001
» https://tempo.inmet.gov.br/TabelaEstacoes/A001 - Jordan T, Nice G, Smeda R, Sprague C, Loux M, Johnson B. Biology and management of common ragweed. Glyphosate, weeds, and crops series (GWC 14). West Lafayette: Purdue University Extension; 2021.
-
Knolmajer B, Jócsák I, Taller J, Keszthelyi S, Kazinczi G. Common ragweed: Ambrosia artemisiifolia L.: a review with special regards to the latest results in protection methods, herbicide resistance, new tools and methods. Agronomy. 2025;15(8):1-35. Available from: https://doi.org/10.3390/agronomy15081765
» https://doi.org/10.3390/agronomy15081765 -
Roncatto E, Barroso AA, Albrecht AJ, Novello BD, Silva RG, Backes CB. Shortening critical period of weed control at soybean by residual herbicide mixtures. Adv Weed Sci. 2023;41:1-7. Available from: https://doi.org/10.51694/AdvWeedSci/2023;41:00009
» https://doi.org/10.51694/AdvWeedSci/2023;41:00009 -
Rockenbach AP, Rizzardi MA. Growth analysis of Ambrosia artemisiifolia. Rev Ceres. 2024;71:1-12. Available from: https://doi.org/10.1590/0034-737X2024710001
» https://doi.org/10.1590/0034-737X2024710001 -
Schuster MZ, Gastal F, Doisy D, Charrier X, Moraes A, Médiène S, Barbu CM. Weed regulation by crop and grassland competition: critical biomass level and persistence rate. Eur J Agron. 2020;113:125963.j.eja. Available from: https://doi.org/10.1016/j.eja.2019.125963
» https://doi.org/10.1016/j.eja.2019.125963 -
Schuster MZ, Barroso AAM, Gastal F. Grassland phases as ecological buffers: Reducing invasive weeds even under drought-induced stress in cropping systems. Agric Ecosyst Environ. 2026;395:109939. Available from: https://doi.org/10.1016/j.agee.2025.109939
» https://doi.org/10.1016/j.agee.2025.109939 -
Song XJ, Liu G, Qian ZQ, Zhu ZH. Niche filling dynamics of ragweed (Ambrosia artemisiifolia L.) during global invasion. Plants. 2023;12(6):1-16. Available from: https://doi.org/10.3390/plants12061313
» https://doi.org/10.3390/plants12061313 -
Sossmeier SG, Bortoluzzi MP, Rizzardi MA. Interference and economic damage threshold of smooth pigweed in soybean. Pesq Agropec Bras. 2025;60(1):1-9. Available from: https://doi.org/10.1590/S1678-3921.pab2025.v60.03768
» https://doi.org/10.1590/S1678-3921.pab2025.v60.03768
Edited by
-
Editor in Chief:
Carol Ann Mallory-Smith
-
Associate Editor:
Paul Te-Ming Tseng
Source: 
The equation describes yield loss as a function of weed density, accounting for fixed effects of field and block: Yield Reduction (x) ={ 5.91 + 10.72·x + β_field + β_block se x ≤ 2.47; 5.91 + 10.72·2.47 + 1.45·(x – 2.47) + β_field + β_block se x > 2.47} Where x represents common ragweed density (plants m–2); the slope 10.72 corresponds to yield loss per plant below the breakpoint; the post-breakpoint slope (1.45) results from 10.72 – 9.27; and β_field + β_block are the fixed effects of field and block, respectively
BiomassC (common ragweed biomass); DiameterS (soybean diameter); Numberfood (number of pods); HeightS (soybean height); Soybendensity (population density); YeldReduction (reduced productivity).