Introduction
Annual bluegrass (Poa annua L.) is a globally distributed winter grass species commonly found in turfgrass, roadsides, orchards, vineyards, and numerous agroecosystems (Mitich Reference Mitich1998). Its ecological plasticity, high fecundity, short life cycle, and allotetraploid genome (2n = 4x = 28) contribute to its success as a weed and enable rapid adaptation to selection pressures (Benson et al. Reference Benson, Sheltra, Maughan, Jellen, Robbins, Bushman, Patterson, Hall and Huff2023; Robbins et al. Reference Robbins, Bushman, Huff, Benson, Warnke, Maughan, Jellen, Johnson and Maughan2023). Globally, P. annua is one of only two weed species confirmed to have evolved resistance to 12 herbicide modes of action (Heap Reference Heap2025), posing a significant problem in production systems that rely heavily on chemical weed control (Ghanizadeh et al. Reference Ghanizadeh, Mesarich and Harrington2020; Rutland et al. Reference Rutland, Bowling, Russell, Hall, Patel, Askew, Bagavathiannan, Brosnan, Gannon, Gonçalves, Hathcoat, McCarty, McCullough, McCurdy and Patton2023).
Hazelnuts (Corylus avellana L.) are a high-value perennial crop in Oregon, generating more than US$170 million in annual farm gate value (USDA-NASS 2023) and sustaining many family-operated farms in the Willamette Valley. Clean orchard floors facilitate mechanical harvesting and reduce pest pressure; however, weed management in hazelnut orchards is challenging due to limited effective control options (Miranda and Moretti Reference Miranda and Moretti2024). Nonchemical weed control strategies are rarely adopted or are logistically challenging to implement, leaving growers dependent on a narrow set of herbicide modes of action (Moretti Reference Moretti2021). Standard management programs typically include applications of glyphosate and glufosinate during the dormant season, late fall or early spring applications of clethodim, glufosinate, rimsulfuron, and paraquat, along with pendimethalin, indaziflam, or other preemergence herbicides to inhibit weed emergence (Miranda and Moretti Reference Miranda and Moretti2025; Moretti Reference Moretti and Becerra-Alvarez2025; Pedroso and Moretti Reference Pedroso and Moretti2022).
In recent years, hazelnut growers, crop consultants, and field managers across Oregon’s Willamette Valley have reported escapes of P. annua following applications of glyphosate, paraquat, clethodim, and pendimethalin, even when applied at the recommended maximum allowed rates and optimal application timings. These field observations raise concerns about the evolution of herbicide resistance, especially given the crop’s history of repeated herbicide use. Herbicide resistance in weeds can be classified based on the number and mode of action of the herbicides involved. Single resistance refers to resistance to one herbicide or mode of action (Délye et al. Reference Délye, Jasieniuk and Le Corre2013). Multiple resistance occurs when a weed is resistant to two or more herbicides with different modes of action (Bobadilla and Tranel Reference Bobadilla and Tranel2024; Délye et al. Reference Délye, Jasieniuk and Le Corre2013). In contrast, cross-resistance refers to resistance to two or more herbicides that share the same mode of action, even if the weed was only exposed to one of them (Bobadilla and Tranel Reference Bobadilla and Tranel2024; Délye et al. Reference Délye, Jasieniuk and Le Corre2013).
Previous studies have demonstrated that P. annua expresses all major forms of herbicide resistance, including single, multiple, and cross-resistance. Early cases reported triazine-resistant populations due to the Ser-264 mutation in the psbA gene (Kelly et al. Reference Kelly, Coats and Luthe1999), and this mutation also conferred cross-resistance to amicarbazone (Perry et al. Reference Perry, McElroy, Dane, van Santen and Walker2012). Multiple resistance has since been documented, with populations resistant to acetyl-CoA carboxylase (ACCase), acetolactate synthase, and photosystem II inhibitors across turfgrass and cropping systems (Bowling et al. Reference Bowling, McCurdy, De Castro, Patton, Brosnan, Askew, Breeden, Elmore, Gannon, Gonçalves, Kaminski, Kowalewski, Liu, Mattox and McCarty2024; Rutland et al. Reference Rutland, Bowling, Russell, Hall, Patel, Askew, Bagavathiannan, Brosnan, Gannon, Gonçalves, Hathcoat, McCarty, McCullough, McCurdy and Patton2023; Singh et al. Reference Singh, dos Reis, Reynolds, Elmore and Bagavathiannan2021). Poa annua can evolve resistance through both target-site mutations and non–target site mechanisms, making it one of the most adaptable and problematic weeds globally (Laforest et al. Reference Laforest, Soufiane, Patterson, Vargas, Boggess, Houston, Trigiano and Brosnan2021). These examples illustrate the adaptability of P. annua and emphasize the need for proactive resistance management.
Given the frequency of suspected resistance cases, there is an urgent need to characterize the extent and patterns of herbicide resistance in P. annua within Oregon’s hazelnut production systems. Because of its high adaptability and rapid evolutionary potential, P. annua can also serve as an indicator species for emerging resistance problems, providing early warning for shifts in susceptibility in other troublesome weeds. The objectives of this study were to: (1) determine the resistance status of P. annua accessions collected from hazelnut orchards in the Willamette Valley to clethodim, pendimethalin, paraquat, and glyphosate; (2) quantify resistance levels using seed and whole-plant dose–response assays; and (3) identify novel cases of resistance, including potential cross-resistance among herbicides with similar modes of action. This research establishes the first documented cases of several herbicide resistances in P. annua in Oregon and the United States, complementing earlier reports of resistance in other production systems. It also provides a new rapid-detection seed assay and offers a baseline for integrated management and stewardship programs aimed at delaying the evolution of resistance in perennial cropping systems.
Materials and Methods
Site Selection and Sample Collection
Between 2022 and 2023, mature P. annua plants were collected from hazelnut orchards across Oregon’s Willamette Valley (Figure 1). Sampling focused on sites where growers reported control failures following applications of glyphosate (5-enolpyruvylshikimate-3-phosphate synthase [EPSPS] inhibitor; Group 9), clethodim (ACCase inhibitor; Group 1), paraquat (photosystem I [PSI] electron diverter; Group 22), or pendimethalin (microtubule assembly inhibitor; Group 3). In total, 10 suspected resistant accessions were collected: 1 for glyphosate (Gly.R), 4 for clethodim (CLR.OAK, CLR.ST, CLR.16, and CLR.AA), 3 for paraquat (PQR.10, PQR.11, and PQR.12), and 2 for pendimethalin (PER.14 and PER.20). Plants were transplanted to the greenhouse and grown under controlled conditions (22/18 C day/night; 12-h photoperiod). Individual plants were bagged at the boot stage using pollen-proof pollination bags (Mini Bag 3D M15.30.15, PBS International, Scarborough, UK) to prevent cross-pollination.
Figure 1. Distribution of Poa annua accessions collected from hazelnut orchards across Oregon’s Willamette Valley for herbicide-resistance screening. The photo on the left shows a hazelnut orchard with P. annua escapes following herbicide application, illustrating poor control in field conditions. The map on the right indicates the geographic origin and herbicide-resistance profile of each accession tested. Symbols represent confirmed resistance to clethodim (blue square), glyphosate (pink triangle), paraquat (green hexagon), and pendimethalin (orange inverted triangle). Black squares denote susceptible accessions (S-LB and S-ORG).
Accessions were advanced one to two generations using single-seed descent to stabilize resistance phenotypes and reduce within-accession variability (Lei et al. Reference Lei, Gordon, Liu, Sade, Lovell, Rubio Wilhelmi, Singan, Sreedasyam, Hestrin, Phillips, Hernandez, Barry, Shu, Jenkins and Schmutz2023). In each generation, seeds/plants were grown in the greenhouse and treated with the maximum allowed herbicide rate to which resistance was suspected. Survivors were selected and advanced to produce seed under greenhouse conditions. This process was repeated to develop homogeneous resistant accessions, which were subsequently used for all dose–response experiments. In addition, two susceptible accessions were obtained from areas with no known history of herbicide use: one from an organically managed orchard at Oregon State University’s Research Farm (S-ORG) and another from a turfgrass area on the Oregon State University’s main campus (S-LB).
Seed-based Dose–Response Assays
Seed-based bioassays were conducted as a rapid, cost-effective method to detect resistance and quantify sensitivity to individual herbicides. Protocols were adapted from Perez et al. (Reference Perez, Beckie, Cawthray, Goggin and Busi2021). For each accession, 10 seeds were placed in 90-mm petri dishes lined with two layers of Whatman No. 1 filter paper (Maidstone, UK) and moistened with 4 ml of aqueous herbicide solution. Seven to eight herbicide concentrations were used per herbicide (Table 1). The herbicides and tested ranges were as follows: glyphosate, 0 to 300 µM; clethodim, 0 to 5 µM; paraquat, 0 to 25 µM; and pendimethalin, 0 to 48 µM. Petri dishes were sealed with Parafilm® to prevent moisture loss and incubated at 20 C in a growth chamber with a 12-h light/dark photoperiod. Treatments were replicated three times, and the entire experiment was repeated. Each herbicide was treated as an independent experiment.
Table 1. Herbicides, application rates, and adjuvants used in seed-based and whole-plant experiments evaluating Poa annua resistance to select herbicides in Oregon hazelnut production.
a Adjuvants: Ammonium sulfate (Amsol, WinField United, Arden Hills, MN 55126) at 10 g L⁻¹ plus methylated seed oil (HASTEN-EA, Wilbur-Ellis, Aurora, CO 80237) at 8.9 g L⁻¹ were included in the whole-plant assays.
b Adjuvants: Ammonium sulfate (Amsol, WinField United) at 10 g L⁻¹ plus nonionic surfactant (Rainier®, Wilbur-Ellis) at 2.5 g L⁻¹ were included in the whole-plant assays.
Seed viability was assessed after 14 d. Seeds were classified as viable if the radicle and/or coleoptile extended ≥5 mm (Ngo et al. Reference Ngo, Boutsalis, Preston and Gill2017). Nongerminated seeds or those exhibiting necrosis or arrested growth were classified as nonviable.
Whole-Plant Dose–Response Assays
Greenhouse experiments were conducted to validate seed-based bioassay results and assess resistance at the whole-plant level using preemergence or postemergence application protocols (Table 1). Experiments were arranged in randomized complete block designs with four replicates per treatment and were conducted twice. Plants were grown in a greenhouse at Oregon State University, Corvallis, OR (44.57°N, 123.29°W), regulated to 22/18 C day/night temperatures and a 12-h photoperiod with supplemented artificial lighting. All herbicide applications were made using a research sprayer (DeVries Manufacturing, Generation III, Hollandale, MN) equipped with a single TP8003E flat-fan nozzle (TeeJet® Technologies, Glendale Heights, IL), placed 45 cm above the canopy and calibrated to deliver 187 L ha−1 of spray solution at 207 kPa.
For the preemergence experiments involving pendimethalin, 10 seeds were sown per 0.5-L pot filled with sterilized Willamette silt loam soil (fine-silty, mixed, superactive, mesic Pachic Ultic Argixerolls). Pendimethalin was applied at rates ranging from 0 to 8,800 g ai ha⁻¹ (labeled rate: 2,240 g ai ha⁻¹) immediately after seeding and covering the seeds. Seedling emergence was recorded at 35 d after treatment (DAT). Plants were rated as “alive” if they had emerged and grown to ≥1-cm tall.
For the postemergence experiments involving glyphosate, clethodim, and paraquat, seeds from each P. annua accession were germinated in acrylic boxes (22 × 11 × 4.5 cm; Hoffman Manufacturing, Corvallis, OR), containing two blotter papers and 50 ml of distilled water, and incubated at 22 C under a 12:12-h light/dark cycle and a light intensity of 300 μmol m−2 s−1. After 10 d, single seedlings were transferred to individual 0.5-L pots filled with commercial potting mix (SS#4 PC RSi, Sun Gro® Horticulture, Agawam, MA). Poa annua plants were treated at the 2- to 3-tiller growth stage (BBCH-23) with seven to eight rates of the following herbicides: glyphosate 0 to 1,680 g ae ha⁻¹ (labeled rate: 840 g ae ha⁻¹); clethodim 0 to 540 g ai ha⁻¹ (labeled rate: 135 g ai ha⁻¹); and because of safety compliance concerns associated with paraquat’s high toxicity, it was replaced with diquat, another bipyridinium herbicide and PSI inhibitor, with treatment rates ranging from 0 to 4,500 g ai ha⁻¹ (labeled rate: 560 g ai ha⁻¹). Cross-resistance between paraquat and diquat has been consistently reported in resistant populations (Hawkes Reference Hawkes2014). Aboveground biomass was harvested at 35 DAT, dried at 60 C to constant weight, and recorded. Plant regrowth was assessed visually at 28 d following biomass harvest; plants showing any regrowth were scored as “alive” (1), while those with no regrowth were scored as “dead” (0).
Cross-resistance
To evaluate potential cross-resistance, selected P. annua accessions with confirmed resistance were exposed to herbicides with a similar mode of action but from different chemical families. Accession CLR.16, identified as clethodim resistant, was treated with fluazifop (another ACCase inhibitor) at rates ranging from 0 to 3,200 g ai ha⁻¹ (labeled rate: 210 g ai ha⁻¹). PER.14 and PER.20, both resistant to pendimethalin, were treated with pronamide, a microtubule inhibitor, at rates from 0 to 11,200 g ai ha⁻¹ (labeled rate: 2,240 g ai ha⁻¹). Fluazifop and pronamide were applied as a postemergence treatment to plants at the 2- to 3-tiller stage, using the same greenhouse conditions, spray setup, and evaluation timelines as the postemergence whole-plant assays. Biomass was harvested at 35 DAT, followed by regrowth assessments at 28 d postharvest.
Statistical Analysis
All data were analyzed using the drc package (Ritz et al. Reference Ritz, Baty, Streibig and Gerhard2015) in R (R Core Team 2025). For fluazifop, aboveground biomass data collected at 35 DAT were used to evaluate herbicide response, as survival data did not fit a log-logistic model due to its limited effectiveness on P. annua and high survival across all accessions. A three-parameter log-logistic model (LL.3) was fit to the biomass data to estimate the dose required to reduce dry weight by 50% (GR50). For all other herbicides, survival data were analyzed using a two-parameter log-logistic model (LL.2) with a binomial error distribution to estimate the dose required to reduce plant survival by 50% (LD50). Pairwise comparisons among accessions were performed using the compParm() function, and significant differences were determined at P < 0.05. Accessions were classified as resistant if the LD50 was statistically higher (P < 0.05) than the LD50 of susceptible accessions (S-LB or S-ORG) and the resistance index (RI), calculated as the ratio of the resistant to susceptible LD50, was ≥2. Data from experimental runs were pooled for analysis, as no significant treatment by run interactions (P > 0.05) were detected.
Results and Discussion
ACCase-inhibiting Herbicides (Clethodim and Fluazifop)
Clethodim resistance was confirmed in all four P. annua accessions (CLR.OAK, CLR.AA, CLR.16, and CLR.ST) with suspected resistance (Figure 2). In seed-based bioassays, LD50 values ranged from 1.6 to 2.7 µM compared with 0.4 to 0.5 µM for susceptible accessions (Figure 2A), corresponding to RIs of 3.2 to 6.8. Whole-plant assays confirmed this trend, with LD50 values of 41 to 43 g ha⁻¹ for CLR.16, CLR.AA, and CLR.OAK, and 217 g ha⁻¹ for CLR.ST, compared with 14 to 19 g ha⁻¹ for susceptible accessions (RI = 2 to 10). Accession CLR.16 was also evaluated for cross-resistance to fluazifop. Due to fluazifop’s poor efficacy on P. annua, biomass reduction (GR50) was used rather than survival. GR50 values were 1,670 g ha⁻¹ for CLR.16, 2,140 g ha⁻¹ for S-LB, and 470 g ha⁻¹ for S-ORG. These results indicate that CLR.16 did not show a clear shift in sensitivity relative to the susceptible populations, and therefore no consistent evidence of cross-resistance to fluazifop was observed.
Figure 2. Clethodim resistance in Poa annua accessions from Oregon hazelnut orchards based on seed-based and whole-plant dose–response assays. (A) Estimated clethodim LD50 values (µM) from seed assays for susceptible and three putative resistant accessions. (B) Estimated clethodim LD50 values (g ha⁻¹) from whole-plant assays for the five P. annua accessions. (C) Cross-resistance to fluazifop, estimated LD50 values (g ha⁻¹) from whole-plant assays for CLR.16 and two susceptible accessions. (D) Seed assay response to clethodim across concentrations. (E) Whole-plant clethodim response showing differential survival across a gradient of doses. (F) Whole-plant response to fluazifop, with CLR.16 surviving rates that controlled susceptible accessions. Error bars represent standard error; asterisks indicate significant differences (P<0.05) relative to susceptible accessions.
To our knowledge, this is the first global confirmed report of clethodim resistance in P. annua, and it is of particular concern, because clethodim is one of the few postemergence herbicide options available for selective grass control in hazelnut orchards and other perennial broadleaf systems. ACCase resistance in other grasses such as rigid ryegrass (Lolium rigidum Gaudin) (De Prado et al. Reference De Prado, Osuna, Heredia and De Prado2005; Vila-Aiub et al. Reference Vila-Aiub, Yu, Han and Powles2015), blackgrass (Alopecurus myosuroides Huds.) (Délye et al. Reference Délye, Matéjicek and Michel2008; Moss et al. Reference Moss, Cocker, Brown, Hall and Field2003), and barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] (Iwakami et al. Reference Iwakami, Ishizawa, Sugiura, Kashiwagi, Oga, Niwayama and Uchino2024) has been linked to target-site mutations (e.g., Ile-1781-Leu, Trp-2027-Cys, Asp-2078-Gly) or enhanced metabolic detoxification (Pan et al. Reference Pan, Guo, Wang, Shi, Yang, Zhou, Yu and Bai2022). Given that P. annua is an allotetraploid species with a high degree of genetic variability, it is plausible that both target-site and non–target site mechanisms could contribute to the resistance observed here.
Microtubule-inhibiting Herbicides (Pendimethalin and Pronamide)
Resistance to pendimethalin was confirmed in both suspected resistant populations PER.14 and PER.20 evaluated through seed and whole-plant dose–response assays (Figure 3A–D). In seed-based assays, LD50 values for resistant accessions ranged from 22 to 42 µM, compared with 0.9 and 1.3 µM for susceptible accessions, corresponding to RIs of 17 to 47. Whole-plant assays corroborated these results, with LD50 values of 1,800 g ha⁻¹ for PER.14 and 3,550 g ha⁻¹ for PER.20, compared with 360 (S-ORG) and 590 (S-LB) g ha⁻¹ in susceptible accessions, RI = 3 to 10. Both accessions, PER.14 and PER.20, were also cross-resistant to pronamide (Figure 3E), with LD50 values of 4,500 g ha⁻¹ for PER.14 and 8,800 g ha⁻¹ for PER.20, compared with 550 g ha⁻¹ for S-ORG and 600 g ha⁻¹ for S-LB, RI = 7 to 16. Although pendimethalin (a dinitroaniline) and pronamide (a benzamide) belong to different chemical classes, both inhibit microtubule polymerization, thereby disrupting spindle fiber formation during cell division. Cross-resistance between these herbicides has been reported in other grass species such as green foxtail [Setaria viridis (L.) P. Beauv.] (Délye et al. Reference Délye, Menchari and Michel2005) and goosegrass [Eleusine indica (L.) Gaertn.] (Russell et al. Reference Russell, Peppers and Rutland2022) and is often associated with alterations in the α-tubulin target site (e.g., point mutations such as Thr-239-Ile or Met-268-Thr) or, in some cases, with enhanced herbicide metabolism. The confirmation of pendimethalin resistance in P. annua is particularly concerning for hazelnut production systems, where this herbicide is essential for early-season weed management. Loss of pendimethalin efficacy could accelerate early-season weed establishment, increase in-crop competition, and contribute to rapid replenishment of the weed seedbank. The observed cross-resistance to pronamide further restricts available options within the microtubule inhibitor mode of action.
Figure 3. Resistance to microtubule-inhibiting herbicides in Poa annua accessions from Oregon hazelnut orchards. (A) Estimated LD50 values for pendimethalin (g ha⁻¹) from whole-plant assays for two susceptible (S-LB, S-ORG) and two resistant accessions (PER.14, PER.20). (B) Whole-plant response to increasing pendimethalin rates showing survival in PER accessions, even at double the maximum labeled rate, 8.8 kg ha⁻¹. (C) Seed assay response to pendimethalin indicating root elongation inhibition in S-LB and reduced sensitivity in PER accessions. (D) Estimated LD50 values for pendimethalin (µM) from seed-based assays for susceptible and resistant accessions. (E) Estimated pronamide LD50 values (g ha⁻¹) from whole-plant assays showing cross-resistance in PER.14 and PER.20. Error bars represent standard error; asterisks indicate significant differences (P<0.05) relative to susceptible accessions.
PSI Electron Diverter Herbicides (Paraquat and Diquat)
Paraquat resistance was confirmed in all three populations with suspected resistance, PQR.10, PQR.11, and PQR.12, based on both seed and whole-plant responses (Figure 4A–C). In seed-based bioassays, LD50 values for resistant accessions ranged from 1.8 to 51 µM, compared with 0.6 to 1 µM for susceptible accessions. Whole-plant assays showed a similar trend, with LD50 values of 180 to 640 g ha⁻¹ of diquat in resistant accessions versus 30 to 50 g ha⁻¹ in susceptible accessions. These differences correspond to RIs ranging from 2 to 85, depending on the assay. This is the first confirmed case of paraquat resistance in P. annua in the United States. Paraquat resistance has been reported in several weed species globally, including P. annua (Clay Reference Clay1989), often arising from diverse physiological mechanisms. Documented resistance mechanisms include vacuolar sequestration of the herbicide, which prevents paraquat from reaching its site of action in the chloroplasts (Brunharo and Hanson Reference Brunharo and Hanson2017), altered chloroplast membrane transport that reduces paraquat uptake into the chloroplast (Li et al. Reference Li, Mu, Bai, Fu, Zou, An, Zhang, Jing, Wang, Li and Yang2013), and enhanced antioxidant or reactive oxygen species–scavenging activity that mitigates the oxidative damage caused by paraquat (Hawkes Reference Hawkes2014; Ye and Gressel Reference Ye2000). Non–target site resistance mechanisms, such as enhanced detoxification and metabolic processes, have also been suggested in other grass weeds (Jugulam and Shyam Reference Jugulam and Shyam2019) and may contribute to the moderate levels of paraquat resistance observed here. Cross-resistance to other PSI electron diverter (bipyridinium) herbicides such as diquat was observed here and has been documented in other species (Moretti et al. Reference Moretti, Bobadilla and Hanson2021). The confirmation of paraquat resistance in P. annua is a major concern for hazelnut production systems in Oregon, where paraquat is a key postemergence contact herbicide used for rapid burndown of emerged weeds.
Figure 4. Paraquat, diquat, and glyphosate resistance in Poa annua accessions from Oregon hazelnut orchards based on seed and whole-plant assays. (A) Seed assay responses to paraquat, showing reduced sensitivity in accessions PQR.10, PQR.11, and PQR.12 compared with susceptible accessions. (B) Estimated paraquat LD50 values (µM) from seed assays. (C) Estimated diquat LD50 values (g ha⁻¹) from whole-plant assays. (D) Estimated glyphosate LD50 values (µM) from seed assays, indicating resistance in Gly.R. (E) Seed assay representative image showing survival of Gly.R at higher glyphosate concentrations. (F) Estimated glyphosate LD50 values (g ae ha⁻¹) from whole-plant assays, confirming resistance in Gly.R. Error bars represent standard error; asterisks indicate significant differences (P<0.05) relative to susceptible accessions.
EPSPS-inhibiting Herbicide (Glyphosate)
Glyphosate resistance was confirmed in the Gly.R accession using both seed and whole-plant assays (Figure 4D–F). The LD50 in seed assays for Gly.R was 2,100 µM, three to six times that of susceptible accessions (340 to 490 µM). Whole-plant LD50 values for Gly.R were 190 g ae ha⁻¹, compared with 60 to 95 g ae ha⁻¹ in S-LB and S-ORG, corresponding to RIs of 3 to 4. This is the first documented glyphosate resistance in P. annua from Oregon. Possible mechanisms include EPSPS target-site mutations (Pro-106-Ala, Pro-106-Leu), copy number variation, or non–target site mechanisms such as reduced translocation or enhanced metabolism (Barua et al. Reference Barua, Malone, Boutsalis, Gill and Preston2022; Brunharo et al. Reference Brunharo, Morran, Martin, Moretti and Hanson2019; Chen et al. Reference Chen, Huang, Zhang, Wei, Huang, Chen and Wang2015; Gaines et al. Reference Gaines, Zhang, Wang, Bukun, Chisholm, Shaner, Nissen, Patzoldt, Tranel, Culpepper, Grey, Webster, Vencill, Sammons and Jiang2010; Wakelin and Preston Reference Wakelin and Preston2006). The moderate resistance level observed in Gly.R is consistent with findings by Vukovic et al. (Reference Vukovic, Mattox, Kowalewski, McNally, McElroy and Patton2024), who noted that even low increases in LD50 can significantly reduce control efficacy under field conditions. Such moderate resistance may indicate an early stage of resistance evolution, potentially driven by metabolic processes rather than target-site mutations (Chen et al. Reference Chen, Huang, Zhang, Wei, Huang, Chen and Wang2015).
Significance and Management Implications
This study documents four significant and unprecedented cases of herbicide resistance in P. annua: the first global case of clethodim resistance, the first reports of pendimethalin and glyphosate resistance from Oregon, and the first case of paraquat resistance reported in the United States. Collectively, these results underscore the rapid evolutionary potential of P. annua and highlight the urgent need for integrated weed management strategies in perennial crops. Although P. annua is not generally considered competitive with hazelnut trees and is sometimes used as a ground cover, its presence poses significant risks through cross-contamination with other crops. Many Oregon hazelnut growers also produce grass seed, where P. annua contamination can jeopardize seed purity standards and export markets (Mengistu et al. Reference Mengistu, Mueller-Warrant and Barker2000). The introduction of resistant P. annua biotypes into grass seed production systems would therefore have serious economic consequences, as these crops demand strict varietal purity and are highly sensitive to weed seed contamination (Mengistu et al. Reference Mengistu, Mueller-Warrant and Barker2000; Vukovic et al. Reference Vukovic, Mattox, Kowalewski, McNally, Bigelow, Meyers, Brosnan and Patton2023). In this context, rapid diagnostics such as seed-based bioassays could help identify resistant populations early, allowing growers to adjust management strategies to prevent spread across production systems. This broader landscape-level risk highlights the importance of integrating diagnostics and stewardship practices, not only to sustain hazelnut production but also to protect Oregon’s grass seed industry.
The results also align with global patterns of resistance in turfgrass systems, illustrating the rapid adaptability of P. annua under selection pressure (Rutland et al. Reference Rutland, Bowling, Russell, Hall, Patel, Askew, Bagavathiannan, Brosnan, Gannon, Gonçalves, Hathcoat, McCarty, McCullough, McCurdy and Patton2023). Notably, our results contrast with those of Bowling et al. (Reference Bowling, McCurdy, De Castro, Patton, Brosnan, Askew, Breeden, Elmore, Gannon, Gonçalves, Kaminski, Kowalewski, Liu, Mattox and McCarty2024), who conducted a national herbicide resistance survey and reported no resistance in P. annua accessions collected from Oregon turfgrass systems. This work also provides the first resistance baseline for P. annua in Oregon hazelnuts, enabling future monitoring, trend analysis, and proactive management planning. Our results also align with the patterns of ethofumesate resistance in P. annua from grass seed production reported by Vukovic et al. (Reference Vukovic, Mattox, Kowalewski, McNally, Bigelow, Meyers, Brosnan and Patton2023), in which diverse sources of seed exhibited variable but persistent resistance traits. In this study, we also presented the first application of seed-based bioassays for detecting resistance to clethodim and paraquat in P. annua, and the first seed assay–based confirmation of paraquat resistance in any weed species. These assays proved to be rapid (<2 wk), cost-effective, and scalable, making them well suited for on-farm resistance diagnostics. As demonstrated by Cutulle et al. (Reference Cutulle, McElroy, Millwood, Sorochan and Stewart2009) and Perez et al. (Reference Perez, Beckie, Cawthray, Goggin and Busi2021), seed-based screening methods can effectively detect resistance to mitotic and ACCase inhibitors, holding promise for practical implementation. Such early detection capabilities can help growers and advisors adjust timely management decisions before resistance becomes widespread. In the long term, seed-based assays could complement molecular diagnostics, particularly in cases where resistance mechanisms are unknown.
For Oregon hazelnut production, resistance to clethodim, pendimethalin, glyphosate, and paraquat eliminates several key chemical options in a system already constrained by limited registered herbicides. With few registered alternatives and minimal adoption of mechanical or cultural control, resistance management must shift from reactive measures taken after control failure to proactive, diversified strategies (Islam and Monjardino 2025; Simard and Laforest Reference Simard and Laforest2024). Short-term strategies should prioritize rotating herbicide modes of action; tank mixing; and integrating non-chemical methods such as cultivation, mowing, and cover cropping (Beckie and Reboud Reference Beckie and Reboud2009; Busi et al. Reference Busi, Powles, Beckie and Renton2020; Norsworthy et al. Reference Norsworthy, Ward, Shaw, Llewellyn, Nichols, Webster, Bradley, Frisvold, Powles, Burgos, Witt and Barrett2012). Long-term strategies should include the development and adoption of diagnostic tools, both seed-based and molecular, alongside the adoption of non-chemical weed management technologies.
Despite these findings, this study had several limitations. Mechanisms of resistance were not investigated; thus, the molecular or physiological basis, whether target-site mutations, gene amplification, or enhanced metabolism, remains unknown. No evaluation was conducted on the potential fitness costs of resistance, which may influence the long-term persistence of resistant biotypes. Additionally, only a subset of Oregon orchards was surveyed, potentially underestimating the geographic extent of resistance. Furthermore, the study did not evaluate the performance of herbicide mixtures, sequential applications, or integrated programs under commercial orchard conditions. Future research should prioritize (1) mechanism characterization via sequencing and gene expression analysis, (2) fitness cost studies, (3) expanded statewide surveys, and (4) field-scale evaluation of non-chemical strategies. Such efforts are crucial for achieving sustainable weed management in perennial systems, which are facing growing challenges from herbicide resistance.
Acknowledgments
The authors thank the crop consultants, who facilitated our connections with farmers. We are grateful to the farmers for their collaboration and for generously granting access to their orchards.
Funding statement
This research was supported by the Oregon Hazelnut Commission and the Ferrero Hazelnut Company, which made this work possible.
Competing interests
The authors declare no conflicts of interest.
Open access
