Introduction
Achieving a sustainable agricultural and food system requires consideration of environmental, economic, and social dimensions across all stages—from food production to its commercialization and consumption. This is particularly relevant for small-scale organic vegetable farmers who, unlike commodity crop growers, sell fresh, unprocessed products directly to consumers, making visual quality a key factor in marketability. However, despite much research (Mohler, Teasdale and DiTommaso, Reference Mohler, Teasdale and DiTommaso2022), weeds remain one of the main threats to high-quality produce for organic vegetable growers (Organic Farming Research Foundation, 2022). Weed competition not only reduces overall yields but also contributes to misshapen and unmarketable produce (McErlich and Boydston, Reference McErlich and Boydston2014).
Without access to synthetic herbicides, organic growers rely on the integration of nonchemical weed management practices like cover cropping, mulching, enhancing crop competitiveness, and mechanical cultivation (Coleman et al., Reference Coleman, Kristiansen, Sindel and Fyfe2024). Yet, on smaller-scale and diversified farms, labor-intensive hand weeding remains a prevalent method of controlling weeds due to the limited mechanization in these systems (Weil et al., Reference Weil, Silva, Hendrickson and Mitchell2017; Coleman et al., Reference Coleman, Kristiansen, Sindel and Fyfe2024). As a result, labor costs on diversified vegetable farms account for upwards of 65%–70% of farm production expenses (Weil et al., Reference Weil, Silva, Hendrickson and Mitchell2017). Rising farm wages and labor shortages make it essential to adopt strategies that optimize weed management labor while also sustaining crop yields and quality (Calvin, Simnitt and Martin, Reference Calvin, Simnitt and Martin2022).
An approach that growers can use to reduce labor demands is to concentrate weed management efforts during the critical period of weed control (CPWC) (Brown, Hoshide and Gallandt, Reference Brown, Hoshide and Gallandt2019), that is, the window of time a crop needs to be weeded to achieve acceptable yields (Nieto, Brondo and Gonzalez, Reference Nieto, Brondo and Gonzalez1968; Knezevic et al., Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002). The CPWC is determined by combining two different metrics: (1) the critical timing of weed removal, or how long the crop can coexist with weeds before there is a significant reduction in yield, and (2) the critical weed-free period, or the minimum amount of time the crop must remain weed free to maintain high yields (Knezevic et al., Reference Knezevic, Evans, Blankenship, Van Acker and Lindquist2002). Despite its potential, reviews of CPWC studies indicate that nearly all studies on the CPWC have been conducted on conventionally grown commodity crops (Van Heemst, Reference Van Heemst1985; Ramesh et al., Reference Ramesh, Vijaya Kumar, Upadhyay and Chauhan2021) In the few vegetable crops where the CPWC has been determined, studies have predominantly been conducted in conventional systems where herbicides and other synthetic inputs had been used (Weaver, Reference Weaver1984; Qasem, Reference Qasem1995; Amador-Ramirez, Reference Amador-Ramirez2002; Swanton, O’Sullivan and Robinson, Reference Swanton, O’Sullivan and Robinson2010; Chaudhari et al., Reference Chaudhari, Jennings, Monks, Jordan, Gunter, McGowen and Louws2016; Fadlallah, Reference Fadlallah2019). Thus, their applicability to organic systems is limited and leaves a knowledge gap for organic vegetable systems where reducing weeding labor costs is particularly important.
Utilizing the CPWC is especially advantageous for crops with poor competitive ability, which require high labor demands to mitigate weed pressure and maximize crop quality. A prime example is carrot ( Daucus carota ), one of the least competitive crops (Van Heemst, Reference Van Heemst1985). Carrots’ lack of competitiveness can be attributed to the crop’s direct seeding, slow emergence, and limited early season growth, delaying canopy closure (Colquhoun, Rittmeyer and Heider, Reference Colquhoun, Rittmeyer and Heider2017). Consequently, organically grown carrots require high labor inputs of hand weeding, with estimates ranging from 100 to 500 hours/ha−1 (Tei et al., Reference Tei, Baumann, Bleeker, Dobrzański, Economu, Fogelberg, Froud-Williams, Hoek, Melander, Rocha, Ruuttunen, Rzozi, Sanseovic, Simonic A, Torma, Ujgur, Van Der Weide, Verschwele, Villeneuve, Zaragoza and Van Laar2002). In addition to decreasing overall yields, weed competition in carrots contributes to reduced crop quality, leading to lower marketable yields due to smaller and deformed roots that are forked, bent, or stunted, ultimately reducing marketability (Dresbøll, Bjorn and Thorup-Kristensen, Reference Dresbøll, Bjorn and Thorup-Kristensen2008; Abdel-dayem et al., Reference Abdel-Dayem, Shady, El-Nady and Ghalwash2022; Kaur et al., Reference Kaur, Morden, Orsat and Singh2022).
In the United States, nearly 40% of produce is discarded before reaching consumers, often due to visual imperfections (Gunders, Reference Gunders2012). Consumers associate imperfect produce with inferior nutritional value, safety, and economic worth (Yuan et al., Reference Yuan, Yi, Williams and Park2019), resulting in substantial quantities of produce being left unharvested or disposed of at the retail level (Porter et al., Reference Porter, Reay, Bomberg and Higgins2018). One strategy to utilize these imperfect vegetables is to incorporate produce with poor marketability into value-added products (Bhatt et al., Reference Bhatt, Lee, Deutsch, Ayaz, Fulton and Suri2018), that is, produce that has undergone a change in its physical state or form to enhance its value (Agricultural Marketing Resource Center, n.d.). This allows producers to reduce food waste while simultaneously opening new revenue sources.
Given the high labor demands and poor competitive ability of carrots, improving weed management and finding alternative markets for imperfect produce is imperative. While CPWC has been applied to optimize carrot yields (Swanton, O’Sullivan and Robinson, Reference Swanton, O’Sullivan and Robinson2010; Fadlallah, Reference Fadlallah2019), it has not been studied in organic vegetable systems. Furthermore, its relevance to crop quality, an essential consideration for vegetable producers marketing fresh produce, remains largely unexplored.
Yield loss and physical deformities caused by weed competition may not be proportional, raising questions about whether CPWC is sufficient to protect both yield and quality. We therefore hypothesized that CPWC defined by visual quality would differ from that defined by yield alone and expected that weed interference could reduce carrot quality even when biomass reductions were limited, thereby lengthening the weed-free period required to maintain marketability.
To address these hypotheses, we assessed the following research questions: (1) What is CPWC for organically grown carrot that allows growers minimize yield losses due to weed competition? and (2) Can the identified CPWC also be used to optimize organic carrot quality? We further evaluated whether imperfect carrots were suitable for value-added product processing by addressing the following question: (3) Can imperfect carrots be used for value-added products by local food processors?
Materials and methods
Experimental design
This study was conducted in 2021 and 2022 at the Towne Harvest Garden, a certified organic research and teaching farm managed by Montana State University in Bozeman, MT, USA. In all, four trials were carried out: 2021 early planting, 2021 late planting, 2022 early planting, and 2022 late planting. Trials were established at separate irrigated field locations each year to reflect typical organic crop rotations. A mustard (Brassica spp.) cover crop preceded carrots in 2021 and winter squash (Cucurbita spp.) in 2022.
Farm-level weed surveys were conducted in carrot production areas at Towne Harvest from 2022 to 2024. Across years, dominant species included warm-season annuals such as Chenopodium album and Amaranthus retroflexus, primarily cool-season annuals including Lamium amplexicaule and Malva neglecta, and perennial species including Cirsium arvense and Taraxacum officinale, with relative abundance varying across years. This assemblage represents a mix of early and mid-season emerging species typical of organic vegetable systems in the region.
During the growing season, the mean annual temperature was 19.9 °C and 19.2 °C in 2021 and 2022, respectively, and total precipitation was 72.1 mm and 90.7 mm in 2021 and 2022, respectively. The soil at the site is a loam with moderate organic matter content (3.5% in 2021, 3.1% in 2022). Soil tests were taken in late May (0–24 cm depth) in both 2021 and 2022. Nitrate-nitrogen concentrations averaged 9.9 ppm in 2021 to 11.5 in 2022 in the 0–6 cm layer and 9.2 ppm in 2021 and 7.4 ppm in 2022 in the 6–24 cm layer. Olsen phosphorus levels ranged from 87 ppm in 2021 and 96 ppm in 2022, and potassium averaged from 704 ppm in 2021 and 773 ppm in 2022.
In all trials, early-season weed management was conducted through multiple tillage passes ahead of planting. To prepare a fine seedbed for carrot seeding, fields were subsequently rototilled and then cultivated with a S-tine, following standard practices for the region. In both years, the cultivar Hercules F1 was selected for its prior success under local conditions, adaptability, cold hardiness, and storage capacity, which make it well suited for fall harvest. Seed was sourced from Johnny’s Selected Seeds (Winslow, ME, USA) and complied with organic production standards. Pelleted carrot seeds were sown using a Jang tractor-mounted seeder to ensure uniform spacing and planting depth. Carrots were planted 5 cm apart and approximately 2.5 cm deep, in four-row beds with 30 cm spacing between rows. Early-season plantings occurred on June 3 (2021) and June 23 (2022), while late-season plantings took place on June 28 (2021) and July 14 (2022). Early and late plantings were established in adjacent beds within the same field.
Each year, the experimental design consisted of a randomized complete block design with four replications per treatment. Each experimental plot was 2.5 m × 4 rows with eight randomly assigned weeding treatments imposed to create a defined duration of intervention: (1) weedy up to the 2-leaf stage, (2) weedy up to the 6-leaf stage, (3) weedy up to the 10-leaf stage, (4) weed free up to the 2-leaf stage, (5) weed free up to the 6-leaf stage, (6) weed free up to the 10-leaf stage, (7) season-long weedy, and (8) season-long weed free. Treatments were maintained by hand weeding once to twice a week, depending on weed pressure.
In 2021, harvests occurred on September 3 (early planting) and September 24 (late planting); in 2022, carrots were harvested on October 5 (early) and October 21 (late). From each plot, 20 carrots were randomly selected for data collection. Root biomass (fresh weight) was measured individually for each of the 20 selected carrots per treatment after removing aboveground foliage. Roots were gently cleaned of soil and weighed immediately after harvest using a digital scale (±0.1 g precision).
Additionally, all 20 carrots from each plot were evaluated for visual quality using a 5-point ordinal scale, where 1 indicated the lowest quality and marketability and 5 represented the highest. Quality ratings were based on overall marketability, including root shape uniformity, incidence of forking or splitting, surface defects, and size relative to fresh-market standards. Carrots receiving a score of 1 were severely misshapen: forked, split, or otherwise damaged and were deemed unmarketable for direct-to-consumer sales. In contrast, carrots rated as 5 had uniform, cylindrical shapes, were undamaged, and of sufficient size for sale. Scores of 2–4 represented intermediate levels of cosmetic deformity, with increasing marketability from 2 (substantially misshapen but partially usable) to 4 (minor imperfections but marketable). Visual quality assessments were conducted by multiple trained evaluators. Prior to scoring and to minimize scoring variability among evaluators, we conducted a calibration session where representative carrots from each quality category were jointly reviewed to standardize the scoring criteria. Of the 20 carrots from each treatment, a randomly selected subset of 5 was used to measure total soluble solids (Brix°), using a digital refractometer (Model HI96801, Hanna Instruments, Woonsocket, RI, USA). Total soluble solids are predominately composed of soluble sugars and are used as a practical indicator of perceived sweetness, and therefore taste, in carrots and other horticultural crops (Magwaza and Opara, Reference Magwaza and Opara2015; Carvalho, Silva and Magalhães, Reference Carvalho, Silva and Magalhães2020).
In addition, we worked with two local food processors which make value-added carrot-based products. Specifically, one food processor evaluated the feasibility of incorporating carrots from each quality category into their curried carrot pickles, while the second assessed their use for spicy fermented carrot chips. Each processor received approximately 2 kg of carrots from each quality category (1–5). Based on the samples provided, processors completed a questionnaire for each quality category, indicating whether the carrots could be used, labor requirements for each quality category of carrots, and how much they would be willing to pay farmers for carrots of varying quality per pound. Despite the small sample size, this collaboration provided preliminary insight into the potential for processing weed-impacted carrots and their marketability within value-added supply chains.
Statistical analysis
To assess treatment and trial effects, a factorial ANOVA was performed for three response variables: carrot root biomass, average marketability/quality scores, and sugar content. Linear models of the response vs. Treatment + Trial were fitted, followed by Tukey’s honest significant difference (HSD) post-hoc tests for pairwise comparisons. Interaction terms (Treatment × Trial) were also tested to evaluate potential variability in treatment effects across trials. To further evaluate how response variables related to each other, Pearson correlation and linear regression analyses were conducted for individual trials for the root biomass versus quality and quality versus sugar associations. All models and boxplot visualizations were implemented in R (R Core Team, 2024) using the multcomp, ggplot2, and dplyr packages.
To generalize results and allow for comparison across years and trials, weeding treatments were correlated with accumulated growing degree days (GDD), that is, the accumulated temperature from crop planting to treatment timing. The GDD was calculated using a base/minimum temperature of 3.33 °C and an upper/maximum temperature of 30 °C for carrots as recommended by Oregon State Extension (Andrews, Reference Andrews2011). Temperature data were collected from an on-site weather station. CPWC was estimated using yields or quality data and GDD following Knezevic and Datta (Reference Knezevic and Datta2015), using the following four-parameter log-logistic regression model:
where Y represents the relative response of the variable of interest (i.e., relative yield or marketability/quality score, relative to the season-long weed-free control), C and D are the lower and upper asymptotes, X is time expressed as GDD, E is the GDD at the inflection point (50% response between C and D), and B is the slope at the inflection point. Separate models were fitted for the weedy and weed-free treatments for each year and planting and for both response metrics (relative yield and quality). Average plot-level quality scores were treated as continuous response variables in dose–response modeling. Analyses were conducted in R (R Core Team, 2024) using the drc package (Ritz et al., Reference Ritz, Baty, Streibig and Gerhard2015), with the l4 function applied to fit the log-logistic response curves.
Model fit was assessed using lack-of-fit tests (modelFit() function), which compared the residual sum of squares of the fitted model to that of an ANOVA model. A nonsignificant p-value (p > 0.05) indicated an adequate fit of the log-logistic model to the data. In addition, model assumptions were evaluated for each site-year analysis by inspecting residual versus fitted value plots to assess homogeneity of variance and by examining normal Q-Q plots to assess residual normality. Across all models, residuals were approximately symmetrically distributed around zero with no consistent patterns of heteroscedasticity. Minor deviations in the distribution tails were observed in some years and trials but were not systemic. Based on these diagnostics, data transformations were not required. For all models, upper (D) and lower (C) limits were fixed using observed values from the experiment: D was set to the maximum observed yield/quality in the season-long weed-free treatment, and C to the minimum yield/quality observed in the season-long weedy treatment for the associated trial ensuring that CPWC estimates reflected the biologically observed yield/quality range.
Based on a 5% acceptable yield/quality loss threshold, the beginning of the CPWC was identified as the point in time at which relative yield or quality fell below 95% of the weed-free maximum, estimated from the weedy treatment model. Similarly, the end of the CPWC was identified as the point at which treatments did not maintain a yield or quality below 95% of the season-long weed-free condition. In most cases, CPWC thresholds were estimated using the ED() function in the drc package in R, which calculates effective dose levels based on the fitted model and provides associated standard errors derived from the model’s parameters. However, in 2022, model weed-free curves extended beyond the observed data range for relative yield, resulting in biologically implausible estimates of the CPWC endpoint (i.e., after harvest date). In these cases, we constrained estimation to the observed GDD range and identified the first point where the fitted curve reached 95% of the weed-free maximum. This approach does not offer standard error estimates, so confidence intervals for these endpoints were not reported. The use of constrained curve based thresholds parallels the methods used in earlier CPWC studies (Weaver, Reference Weaver1984) but uses fitted regression models rather than direct comparison of treatment means. To examine potential trade-offs between yield and quality, the overlap in GDD ranges between CPWC intervals defined by relative yield and quality was calculated. While the relationship between carrot quality and sugar content was assessed, no formal statistical analysis was conducted on the value-added processor survey responses due to the limited sample size (n = 2), and results should be considered cautiously rather than statistically conclusive.
Results
Root biomass, quality, and sugar
Treatment × Trial interaction terms were not significant, indicating that the effects of weeding treatments on response variables were consistent across years and planting timings (quality: p = 0.88; root biomass: p = 0.96; sugar: p = 0.26). However, the main effect of trial was significant for all response variables (p < 0.001), indicating that overall response levels differed among years and planting timings. Weeding treatment affected root biomass (Fig. 1) (p < 0.001, for all trials) and quality (Fig. 2) (p < 0.004, for all trials). To contextualize results, Fig. 3 presents representative carrots from both weedy-until and weed-free-until up to successive carrot leaf stages, highlighting smaller roots with longer weed interference. In contrast, sugar content showed a weak treatment effect in only the early planting trial of 2021 (p = 0.012) and no effect in the others (p = 0.44, for all trials).
Relative carrot yield (%) in response to increasing weed interference duration (weedy, solid black line) and weed-free period (weed-free, dashed red line) across planting dates (early versus late) and years (2021 and 2022). Points represent treatment-level means, and lines represent fitted four-parameter log-logistic models based on accumulated growing degree days (base temperature 3.33 °C) from carrot emergence. The shaded yellow region denotes the relative yield critical period of weed control (CPWC).
Figure 1. Long description
The four panels are arranged in a two-by-two grid. The top row represents Early Planting and the bottom row represents Late Planting. The left column is for the year 2021 and the right column is for 2022.
Each panel shares the same axes: the x-axis is Growing Degree Days ranging from 0 to 1500, and the y-axis is Relative Yield (percent) ranging from 0 to 100.
In every panel, two sigmoidal curves intersect.
1. The Weedy curve (solid black line with open circles) starts at 100 percent yield and follows a sigmoidal decline as growing degree days increase, eventually leveling off at a lower yield floor.
2. The Weed-Free curve (dashed red line with open triangles) starts at a low yield and follows a sigmoidal increase, reaching a plateau near 100 percent yield.
A shaded yellow region, labeled as the Yield C P W C (Critical Period of Weed Control), is bounded by vertical dotted yellow lines. This region highlights the interval where the two curves cross and the yield is most sensitive to weed interference.
Specific panel observations:
- Top-Left (Early Planting, 2021): The C P W C begins around 500 and ends near 1100 growing degree days. The weedy curve drops sharply after 500.
- Top-Right (Early Planting, 2022): The C P W C is wider, starting earlier near 400 and extending to 1100. The weedy curve has a more gradual decline compared to 2021.
- Bottom-Left (Late Planting, 2021): The C P W C is shifted right, starting near 700 and ending near 1200.
- Bottom-Right (Late Planting, 2022): The C P W C starts near 400 and ends near 1100, similar to the early planting of the same year.
Carrot marketabiltiy/quality (qualititive scores 1–5) in response to increasing weed interference duration (weedy, solid black line) and weed-free period (weed-free, dashed red line) across planting dates (early versus late) and years (2021 and 2022). Points represent treatment-level means, and lines represent fitted four-parameter log-logistic models based on accumulated growing degree days (base temperature 3.33 °C). from carrot emergence. The shaded blue region denotes the critical period of weed control (CPWC).
Figure 2. Long description
The four panels are organized in a two-by-two grid. The top row represents 2021 and the bottom row represents 2022. The left column represents Early Planting and the right column represents Late Planting.
* Axes: The horizontal x-axis represents Growing Degree Days ranging from 0 to 1500. The vertical y-axis represents Carrot Quality on a scale of 1 to 5.
* Data Series:
- A solid black line with circular points represents the Weedy treatment. It shows a sigmoidal decline, starting at a high quality score (approximately 3.5 for early planting and 3.0 for late planting) and dropping to a lower score as growing degree days increase.
- A dashed red line with triangular points represents the Weed-Free treatment. It shows a sigmoidal increase, starting at a low quality score (approximately 2.0) and rising to a plateau (approximately 3.5 for early planting and 3.0 for late planting).
* Critical Period: A light blue shaded region, labeled as Quality C P W C, is bounded by vertical dashed blue lines. This region highlights the interval where the two lines intersect and the most significant change in quality occurs.
* Trends: In all panels, the Weedy line and Weed-Free line cross within the shaded C P W C. The Early Planting panels (top) show higher overall peak quality scores compared to the Late Planting panels (bottom). The curves in 2021 and 2022 follow nearly identical trajectories for their respective planting dates.
Representative carrots from weeding treatments demonstrating differences in root size due to weed competiton.
Figure 3. Long description
The carrots are laid out on a gray surface, showing a clear trend where longer weed-free periods result in larger roots. From left to right, the specimens and their labels are:
* Season long weed free: The largest carrot with a thick, long orange root and dense, vibrant green leafy tops.
* Weed free up to 10-leaf stage: Slightly smaller than the first, but still possessing a robust root and full foliage.
* Weed free up to 6-leaf stage: A medium-sized carrot with a noticeably thinner root.
* Weed free up to 2-leaf stage: A small, thin carrot with sparse foliage.
* Season long weedy: Located at the center, this is the smallest specimen, appearing as a tiny, pale, thread-like root with very little green growth.
* Weedy up to 2-leaf stage: A small carrot, slightly larger than the season-long weedy specimen.
* Weedy up to 6-leaf stage: A medium-sized carrot showing recovery in root thickness.
* Weedy up to 10-leaf stage: A large carrot, nearly matching the size of the weed-free specimens, with thick roots and expansive green tops.
Pairwise comparisons using Tukey’s HSD test showed that season-long weedy treatments consistently produced the lowest root biomass and quality scores. For example, in the early planting trial in 2021, carrots from season-long weedy plots weighed 84.8 g ± 6.5 g less, or an 85.8% decrease in root yield, than those from season-long weed-free plots (p < 0.001) and scored 1.3 ± 0.2 points lower in quality (p < 0.001). Although the magnitude of response varied slightly across trials, the direction of treatment effects was consistent.
Correlation and regression analyses were conducted at the plot level within trials (n = 32 plots per trial). A positive association was observed between carrot root biomass and quality, with coefficients of determination ranging from R 2 = 0.45 to 0.61 (all p < 0.001), indicating that higher-biomass carrots tended to receive higher marketability scores. However, these R 2 values indicate that a proportion of variation in quality remained unexplained. Linear regression showed that each 1.0 g increase in root biomass was associated with a 0.01 increase in quality (p < 0.001, R 2 = 0.35). In contrast, the association between quality and sugar content was weak overall (R 2 = 0.01, p = 0.39). Only the early planting trial of 2021 showed a moderate positive association (R 2 = 0.32 p < 0.01), with no associations in other years.
Critical period of weed control
Model fit was adequate across all years and plantings for both response variables, with no evidence of lack of fit (p > 0.05). The CPWC for relative yield varied across years and planting timing, with starting points ranging from 363.1 (± 62.0) GDD (± SE) (2022, early planting) to 656.1 (± 80.0) GDD (± SE) (2021, late planting) and ending points ranging from 1060.7 (± 375.8) GDD (± SE) (2021, early planting) to 1130.6 (± 406.4) GDD (± SE) (2021, late planting) (Fig. 1). The CPWC for quality did not vary as substantially as for relative yield. Still, results varied across years and planting timing, with starting points ranging from 519.33 (± 110.6) GDD (± SE) (2021, early planting) to 590.1 (± 123.2) GDD (± SE) (2022, late planting) and ending points ranging from 1301.2 (± 859.8) GDD (± SE) (2022, early planting) to 1345.7 (± 376.0) GDD (± SE) (2021, late planting) (Fig. 2).
Across years and planting timing, the total length of CPWC for relative yield was generally shorter than the one for quality, but the extent to which they differed varied between years (Fig. 4). In 2021, the CPWC for relative yield was 208.3 GDD shorter than that for quality in the early planting and 284.7 GDD shorter in the late planting. In 2022, the difference was smaller, with the relative yield CPWC 52.4 GDD shorter in the early planting and 56.6 GDD shorter in the late planting, compared to the quality-based CPWC. In both trials in 2022, and the earlier planting in 2021, the relative yield CPWC started earlier than the quality CPWC. In all trials, the quality CPWC ended later than the relative yield CPWC, requiring an average of 233.4 ± 20.0 GDDs to maintain acceptable quality levels.
Comparison of the critical period of weed control (CPWC) based on relative yield (yellow) and crop quality (blue) for organic carrots across years and planting dates. Each horizontal bar represents the CPWC duration in growing degree days (base temperature 3.33 °C). from crop emergence. The shaded gray regions indicate the overlap between yield- and quality-based CPWC estimates.
Figure 4. Long description
The chart uses a horizontal x-axis representing Growing Degree Days G D D ranging from 200 to 1400. The y-axis is divided into four vertical panels: 2021 Early, 2021 Late, 2022 Early, and 2022 Late. A legend at the top right identifies yellow bars as Relative Yield, blue bars as Quality, and gray shaded regions as Overlap.
* 2021 Early: The yellow bar starts at approximately 480 G D D and ends at 1050 G D D. The blue bar starts at 520 G D D and ends at 1300 G D D. The gray overlap region spans from 520 to 1050 G D D.
* 2021 Late: The yellow bar starts at 650 G D D and ends at 1120 G D D. The blue bar starts at 580 G D D and ends at 1350 G D D. The gray overlap region spans from 650 to 1120 G D D.
* 2022 Early: The yellow bar starts at 350 G D D and ends at 1050 G D D. The blue bar starts at 550 G D D and ends at 1300 G D D. The gray overlap region spans from 550 to 1050 G D D.
* 2022 Late: The yellow bar starts at 420 G D D and ends at 1080 G D D. The blue bar starts at 600 G D D and ends at 1300 G D D. The gray overlap region spans from 600 to 1080 G D D.
In all panels, the Quality-based C P W C extends significantly further into the growing season than the Yield-based C P W C.
Value-added products
The spicy fermented carrot chip food processor indicated their willingness to use carrots from all five quality categories though they expressed minor concerns about carrots with a quality score of 1 due to cleaning difficulty. Nevertheless, they considered even the lowest-rated carrots to be at least partially usable. In contrast, the curried carrot pickles company stated they would not use carrots from quality scores of 1, 2, and 3 but could incorporate 4 and 5 quality into their curried carrot pickles.
The price these two food processor companies were willing to pay for carrots increased with quality. The spicy fermented carrot chip food processing company was willing to pay 27% less per kg for the lowest quality carrots (rating 1), and 13% less per kg for low quality carrots (rating 2) compared to middle- to highest-quality carrots (ratings 3, 4, and 5). The curried carrot pickle food processing company was more selective, citing concerns over processing labor and visual appeal. They were not willing to use carrots from the lowest-, low-, and middle-quality categories (ratings 1, 2, and 3) but indicated willingness to purchase high- and highest-quality carrots (ratings 4 and 5), with high-quality carrots valued at 31% less per kg than the highest-quality category.
The reduced value of lower-rated carrots was largely attributed to increased labor demands during processing, that is, cleaning, cutting, and preparing carrots for fermentation. The spicy fermented carrot chip company reported processing rates of ~0.7 kg/h for the lowest-quality carrots (ratings 1 and 2), increasing to ~0.9 kg/h for higher-quality carrots (ratings 3, 4, and 5). In contrast, the curried carrot pickle company related more variable processing time, with rates of ~0.5 kg/h for rating 3, ~0.6 kg/h for rating 4, and ~ 0.9 kg/h for rating 5.
Discussion
To the best of our knowledge, this is the first study to jointly evaluate CPWC in organically grown carrots and to assess the implications of CPWC for both crop yield and quality. Furthermore, we explored how imperfect carrots may be integrated into local value-added processing. Consistent effects of weeding treatment on carrot yields and quality across all trials underscores the influence weed competition has on organic carrot production. Our results agree with various studies that have demonstrated that both the yield and quality of carrots are vulnerable to weed crop competition (Coelho, Bianco and Carvalho, Reference Coelho, Bianco and Carvalho2009; Colquhoun, Rittmeyer and Heider, Reference Colquhoun, Rittmeyer and Heider2020). In all four trials, carrots grown in season-long weed competition had reduced root biomass and lower marketability compared to those in weed-free plots. In contrast, sugar content remained unaffected by weeding treatment, suggesting that weed pressure may have minimal impact on flavor-related traits.
The CPWC for organic carrots was successfully estimated, with the yield-based period beginning between 363.1 and 656.1 GDD (base 3.33 °C) and ending between 1061.0 and 1131.0 GDD (base 3.33 °C). While our findings show some consistency with previous work in conventional systems, they also highlight notable differences. Swanton, O’Sullivan and Robinson (Reference Swanton, O’Sullivan and Robinson2010) determined the minimum weed-free period required for carrots to prevent unacceptable yield loss (i.e., the end of CPWC) to be 414.0 GDD to 930.0 GDD (base 5 °C with a 5% acceptable yield loss). Our results suggest that organic carrot growers must maintain carrots weed free for longer to achieve a similar yield loss threshold as their conventional counterparts. The increased minimum amount of time organic carrots must remain weed free can be attributed to differences in management practices and mechanisms driving weed crop competition between conventional and organic agroecosystems. For example, Swanton, O’Sullivan and Robinson (Reference Swanton, O’Sullivan and Robinson2010), used Linuron (3-(3,4-dichlorophenyl)-1-methoxy-1-methylurea) early in the season to control broadleaf weeds. Linuron is widely used in conventional carrot production due to its pre- and early–post-emergence activity as well as its residual weed suppression (Bellinder, Kirkwyland and Wallace, Reference Bellinder, Kirkwyland and Wallace1997). This early-season herbicide likely extended weed control during the early growth stages of the carrot crop. In contrast, our study relied on mechanical cultivation and hand weeding for early season weed management, which typically provides less consistent and shorter-lived weed control compared to herbicides (Bàrberi, Reference Bàrberi2002; Riemens et al., Reference Riemens, Groeneveld, Lotz and Kropff2007).
Differences in fertilizer application and nutrient dynamics between organic and conventional agroecosystems may have also influenced the length of the CPWC (Evans et al., Reference Evans, Knezevic, Lindquist, Shapiro and Blankenship2003; Tursun et al., Reference Tursun, Datta, Tuncel, Kantarci and Knezevic2015; Little et al., Reference Little, DiTommaso, Westbrook, Ketterings and Mohler2021). Conventional systems typically apply synthetic fertilizers in concentrated, plant-available forms early in the season, promoting rapid crop growth, earlier crop canopy closure, and a flush of fast-growing weeds (Jørnsgård et al., Reference Jørnsgård, Rasmussen, Hill and Christiansen1996; Liebman and Davis, Reference Liebman and Davis2000). In contrast, organic systems rely on mineralization of organic matter and organically approved fertilizers, which release nutrients gradually over time (Liebman and Davis, Reference Liebman and Davis2000). Thus, fertility management strategies may indirectly influence CPWC through their effects on early crop growth and weed competitive pressure.
Contrasting nutrient dynamics coupled with the absence of herbicide use contributes to more weeds typically observed in organic systems compared to conventional systems (Bengtsson, Ahnström and Weibull, Reference Bengtsson, Ahnström and Weibull2005; Edesi et al., Reference Edesi, Järvan, Adamson, Lauringson and Kuht2012). Despite the greater abundance of weeds, organic systems have been found to be more resistant to yield losses from weed competition (Ryan et al., Reference Ryan, Smith, Mortensen, Teasdale, Curran, Seidel and Shumway2009). This resilience has been attributed to increased weed species diversity, more heterogeneous nutrient availability, and shifts in the functional composition of weed (Smith, Mortensen and Ryan, Reference Smith, Mortensen and Ryan2010; Cordeau et al., Reference Cordeau, Wayman, Ketterings, Pelzer, Sadeghpour and Ryan2021). Also, changes in the associated soil-microbial communities (Ishaq et al., Reference Ishaq, Johnson, Miller, Lehnhoff, Olivo, Yeoman and Menalled2017), in turn, modify plant growth and plant–plant interactions with relatively less crop-weed competition in organic cropping systems that in conventional systems (Johnson et al., Reference Johnson, Miller, Lehnhoff, Miller and Menalled2017).
Despite the differences between conventional and organic systems, our results agree with Swanton, O’Sullivan and Robinson (Reference Swanton, O’Sullivan and Robinson2010), as carrots planted later in the growing season required a shorter weed-free period to maintain acceptable yields. Thus, delaying planting can enhance carrot emergence and early growth due to warmer soil temperatures while also reducing the number of weeds that emerge between tillage and planting (Colquhoun, Rittmeyer and Heider, Reference Colquhoun, Rittmeyer and Heider2020).
Although we observed variability in CPWC duration between planting timings, the greatest differences occurred between years. Several factors could explain this pattern. Although all plots were irrigated, 2022 received more rainfall than 2021, which may have supported more flushes of weeds, prolonging weed emergence and interference (Swanton, O’Sullivan and Robinson, Reference Swanton, O’Sullivan and Robinson2010). Additionally, interannual differences in weed communities may have influenced competitive dynamics. Weed species composition, density, and biomass were not systematically quantified within the experimental plots, precluding us to attribute variation in CPWC duration to specific weed community differences.
Differences in prior field management could have also influenced weed pressure and the length of the CPWC. The 2021 field had previously been planted with a mustard cover crop, known for its weed-suppressive properties (Haramoto and Gallandt, Reference Haramoto and Gallandt2005; Björkman et al., Reference Björkman, Lowry, Shail, Brainard, Anderson and Masiunas2015). In contrast, the 2022 field followed a winter squash cash crop that provided less residue-mediated suppression of weed emergence prior to carrot establishment. In agreement, previous research demonstrated that high residue cover crops can delay the critical time for weed removal and, in some cases, shorten the overall CPWC by suppressing early season weed emergence (Korres and Norsworthy, Reference Korres and Norsworthy2015; Price et al., Reference Price, Korres, Norsworthy and Li2018; Kumari et al., Reference Kumari, Price, Korres, Gamble and Li2023a; Kumari et al., Reference Kumari, Price, Korres, Gamble and Li2023b). Overall, our results reinforce the idea that the CPWC should be applied as an adaptive management guideline that accounts for site-specific factors, including prior management history, environmental conditions, and weed community composition (Ramesh et al., Reference Ramesh, Vijaya Kumar, Upadhyay and Chauhan2021).
Our results demonstrate that weed interference reduced not only total carrot yield but also increased the proportion of misshapen or undersized roots. This has vital implications for management decisions when visual quality is closely tied to price premiums (Grewal et al., Reference Grewal, Hmurovic, Lamberton and Reczek2019; Hartmann, Jahnke and Hamm, Reference Hartmann, Jahnke and Hamm2021). Growers must consider trade-offs with economic and labor demands required for more weeding to preserve their desired crop marketability. However, it should be noted that carrot quality is influenced by multiple interacting factors beyond just weed competition, including cultivar characteristics, soil structure, environmental stress, pests, soilborne pathogens, and storage (Suojala, Reference Suojala2000; Johansen et al., Reference Johansen, Thomsen, Løes and Riley2015; Bender, Reference Bender2021; Paparella et al., Reference Paparella, Kongala, Serio, Rossi, Shaltiel-Harpaza, Husaini and Ibdah2024).
We also explored the potential for value-added processing of carrots that do not meet quality standards for fresh market sales. The results show carrot total soluble solids, and thus perceived sweetness, remained unaffected regardless of their quality rating. This suggests that cosmetically imperfect carrots may be suitable for juicing, fermenting, or other processing uses. These findings align with those of Elimelech et al. (Reference Elimelech, Ert, Parag and Hochman2024), who reported that consumers generally did not perceive notable taste differences between visually perfect and imperfect vegetables, and in some cases, rated the imperfect produce as better tasting than its higher visual quality counterpart. However, while total soluble solids serve as a useful indicator of perceived sweetness, a more comprehensive assessment of flavor quality should include additional metrics such as concentrations of 6-methoxymellein (6-MM), soluble phenolics, and organic acids in relation to moisture content, which has been found to be indicative of carrot taste (Talcott, Howard and Brenes, Reference Talcott, Howard and Brenes2001).
Interviews with two local processors suggested that imperfect carrots were acceptable for some value-added products, particularly those focused on flavor rather than form; however, given the limited sample size, these findings should be interpreted as preliminary and context specific. In our study, the processor that used curried carrot sticks that required longer, more uniform carrots compared to the chopped carrot chips was limited in their ability to incorporate misshapen produce. It is possible that other value-added products that use carrot juice or purée like hot sauces and smoothies may be able to more readily incorporate misshapen produce due to lack of limitations of produce size. Specifically, Barone, Donato and Romani (Reference Barone, Donato and Romani2021) demonstrated that transforming misshapen items into value-added products that completely change the shape increases consumer acceptance by making the product a more ‘prototypical’ example of its new category.
In addition to limited incorporation to fit product standards based on shape, food processing companies also reported that lower-quality carrots often required more labor to trim or clean, which reduced processing efficiency and increased costs. These findings suggest that while quality losses due to weed interference may be partially recouped through incorporation into value-added products, labor availability and processing constraints must be considered when designing value-added product recovery strategies.
Conclusion
This study demonstrated that CPWC can serve as a practical decision tool for managing labor demands while maintaining both yield and quality in organic carrot production. Findings demonstrate that CPWC for visual quality extends beyond that for yield, underscoring the importance of quality-based thresholds for growers targeting direct-to-consumer markets. At the same time, the results show that imperfect carrots maintain perceived sweetness levels and may be suitable for select value-added products, especially when the final product form masks initial appearance. By integrating CPWC recommendations with strategies to repurpose lower-quality produce, growers may reduce on-farm food loss, improve labor efficiency, and access new markets. Collectively, these findings establish CPWC as a systems based framework for aligning weed management, market goals, and food loss reduction, addressing key sustainability challenges in diversified organic vegetable systems.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon request.
Acknowledgements
The authors thank Kevin Sheridan, Tyler Kulak, Daniel Engen, Reilly Stack, Wesley Cawood, and Towne Harvest Garden for their assistance with field work and data collection. Additionally, the final version of this manuscript was reviewed by the Nature Research Assistant—Manuscript Adviser (https://researchassistant.nature.com/) for feedback on clarity, content, and grammar on 10/2/2025. AI suggestions were incorporated only in cases that authors judged they would improve the text, and, in these cases, the recommendations were not necessarily accepted verbatim. No changes were made to the research goals, data analysis, or interpretation of the results. All AI-suggested edits are available upon request.
Author contribution
Conceptualization: E.K., A.C., F.M., and M.B.; Drafting: E.K.; Experiments: E.K., F.M., M.B., and A.C.; Formal analysis: E.K.; Funding acquisition: E.K., F.M., and M.B.; Methodology: E.K., A.C., F.M., and M.B.; Resources: M.B. and F.M.; Supervision and project administration: F.M. and M.B.; Writing—review and editing: E.K., F.M., M.B., and A.C.
Funding statement
This material is based upon work that is supported by the National Institute of Food and Agriculture, US Department of Agriculture, under award number 2021-38640-34695 through the Western Sustainable Agriculture Research and Education program under project number [GW22-235]. USDA is an equal opportunity employer and service provider. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the US Department of Agriculture.
Competing interests
The authors report there are no competing interests to declare.
Open access
