Knowledge of weed seeds present in the soil seedbank is important for understanding population dynamics and forecasting future weed infestations. Quantification of the weed seedbank has historically been laborious, and few studies have attempted to quantify seedbanks on the scale required to make management decisions. An accurate, efficient, and ideally automated method to identify weed seeds in field samples is needed. To achieve sufficient precision, we leveraged YOLOv8, a machine learning object detection to accurately identify and count weed seeds obtained from the soil seedbank and weed seed collection. The YOLOv8 model, trained and evaluated using high-quality images captured with a digital microscope, achieved an overall accuracy and precision exceeding 80% confidence in distinguishing various weed seed species in both images and real-time videos. Despite the challenges associated with species having similar seed morphology, the application of YOLOv8 will facilitate rapid and accurate identification of weed seeds for the assessment of future weed management strategies.

Recently, farmers in Brazil have observed a decline in efficacy of glyphosate, chlorimuron, and imazethapyr control of smooth pigweed (Amaranthus hybridus L.). The objectives of this study were to quantify the resistance of Amaranthus in Brazil to glyphosate and acetolactate synthase (ALS)-inhibiting herbicides, elucidate the mechanism of resistance, and assess the frequency of shifts in sensitivity to glyphosate and chlorimuron in Brazil. Dose–response assays were conducted in a greenhouse with glyphosate, chlorimuron, and imazethapyr. This was followed by sequencing of the EPSPS and ALS genes. Additionally, 740 Amaranthus populations across several Brazilian states were monitored over 4 yr, subjected to a single discriminatory dose of glyphosate and chlorimuron. The populations BR18Asp051 and BR21Asp205 were resistant to glyphosate, chlorimuron, and imazethapyr. The elevated resistance level to glyphosate in these populations is attributed to multiple amino acid substitutions (TAP-IVS) in the EPSPS gene; and cross-resistance to sulfonylureas and imidazolinones is conferred by the Trp-574-Leu substitution in the ALS gene in both populations. Overall, resistance distribution indicated that 88% of the sampled populations were considered sensitive to glyphosate, while 66% were sensitive to chlorimuron. Furthermore, 10% of the samples demonstrated multiple resistance to both active ingredients. A shift in glyphosate sensitivity was observed in four states in Brazil; however, sensitivity shifts to chlorimuron were more widely dispersed in Brazilian agricultural regions.

Information regarding the prevalence and distribution of herbicide-resistant waterhemp [Amaranthus tuberculatus (Moq.) Sauer] in Minnesota is limited. Whole-plant bioassays were conducted in the greenhouse on 90 A. tuberculatus populations collected from 47 counties in Minnesota. Eight postemergence herbicides, 2,4-D, atrazine, dicamba, fomesafen, glufosinate, glyphosate, imazamox, and mesotrione, were applied at 1× and 3× the labeled doses. Based on their responses, populations were classified into highly resistant (≥40 % survival at 3× the labeled dose), moderately resistant (<40% survival at 3× the labeled dose but ≥40% survival at 1× the labeled dose), less sensitive (10% to 39% survival at 1× the labeled dose), and susceptible (<10% survival at 1× the labeled dose) categories. All 90 populations were resistant to imazamox, while 89% were resistant to glyphosate. Atrazine, fomesafen, and mesotrione resistance was observed in 47%, 31%, and 22% of all populations, respectively. Ten percent of the populations were resistant to 2,4-D, and 2 of 90 populations exhibited >40% survival following dicamba application at the labeled dose. No population was confirmed to be resistant to glufosinate. However, 22% of all populations were classified as less sensitive to glufosinate. Eighty-two populations were found to be multiple-herbicide resistant. Among these, 15 populations exhibited resistance to four different herbicide sites of action (SOAs); 7 and 4 populations were resistant to five and six SOAs, respectively. All six-way-resistant populations were from southwest Minnesota. Two populations, one from Lincoln County and the other from Lyon County, were resistant to 2,4-D, atrazine, dicamba, fomesafen, glyphosate, imazamox, and mesotrione, leaving only glufosinate as a postemergence control option for these populations in corn (Zea mays L.) and soybean [Glycine max (L.) Merr.]. Diversified management tactics, including nonchemical control measures along with herbicide applications from effective SOAs, should be implemented to slow down the evolution and spread of herbicide-resistant A. tuberculatus populations.

Barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.] is increasingly infesting imidazolinone-tolerant (IMI-T) rice (Oryza sativa L.) fields in China, and E. crus-galli imazamox resistance has become the major concern for weed management in IMI-T rice fields. In this study, the susceptible population JLGY-3 (S) and the suspected resistant population JHXY-2 (R) collected from IMI-T rice fields were used as research subjects. When treated with imazamox, the JHXY-2 (R) population showed a high level of herbicide resistance with a resistance index of 31.2. JHXY-2 (R) was cross-resistant to all five acetolactate synthase (ALS) inhibitors from different chemical families, but sensitive to herbicides inhibiting acetyl-CoA carboxylase. To understand the reason why JHXY-2 (R) was resistant to imazamox, we performed experiments to characterize potential target-site resistance (TSR) and non–target site resistance (NTSR) mechanisms. A Trp-574-Leu amino acid mutation in ALS and low imazamox ALS sensitivity were the main mechanisms underlying imazamox resistance in this JHXY-2 (R) population. There was no significant difference in ALS gene expression and ALS protein abundance between R and S populations. High-performance liquid chromatography–tandem mass spectrometry analysis showed enhanced metabolism of imazamox in JHXY-2 (R), which was in contrast to the results of pretreatment with a metabolic enzyme inhibitor. Treatments with cytochrome P450 monooxygenase/glutathione S-transferase (P450/GST) inhibitors did not alter the resistance level of JHXY-2 (R) against imazamox. Transcriptome sequencing showed that there was almost no significant difference in the expression of P450 and GST metabolic enzyme genes between R and S populations, and only GSTU1 showed a significant upregulation in the R population, further clarifying the NTSR mechanism of JHXY-2 (R). In conclusion, amino acid mutation and higher enzyme activity of ALS are the main causes of imazamox resistance in JHXY-2 (R). However, given the differences in imazamox residues in the leaves of E. crus-galli, there may still be undetectable NTSR mechanisms that are causing imazamox resistance in the R population.

The resistance to fenoxaprop-P-ethyl, a herbicide that inhibits acetyl-CoA carboxylase (ACCase), has emerged in shortawn foxtail (Alopecurus aequalis Sobol.) since the 1990s, presenting a considerable challenge to wheat (Triticum aestivum L.) production in China. One of the primary mechanisms responsible for this high-level resistance is the presence of mutations at codons 1781, 2041, and 2078 in the ACCase gene. However, the conventional methods used to detect these mutations, such as polymerase chain reaction (PCR) and gene sequencing, are time-consuming and labor-intensive. To address this issue and enable the prompt and effective detection of these common ACCase mutations in A. aequalis, a loop-mediated isothermal amplification (LAMP) strategy was developed. The LAMP assay specifically targets the Ile-1781-Leu and Asp-2078-Gly mutations within the ACCase gene of A. aequalis. Through the optimization of primers, systems, and conditions, the LAMP assay enables rapid differentiation between wild-type individuals and mutants of A. aequalis carrying either of these two mutations. Including SYBR Green I dye in the final reaction mixtures enables detection of the target mutation through a noticeable color change that can be observed with the naked eye. It is noteworthy that the sensitivity of the LAMP assay was approximately 104-fold greater than that of conventional PCR methods. Additionally, a derived cleaved amplified polymorphic sequence (dCAPS) assay was established for each mutation to distinguish between homozygous and heterozygous mutants. Overall, the developed LAMP assay could efficiently detect the Ile-1781-Leu and Asp-2078-Gly mutations in the ACCase gene of A. aequalis, offering significant advantages for the monitoring and management of fenoxaprop-P-ethyl resistance.

Pollinators risk exposure to insecticide residue when visiting weedy flowers in urban landscapes. Previous research shows that pollinators are routinely exposed to a variety of pesticides, but herbicides have exhibited minimal toxicity and did not contribute to the modeled risk quotients. Herbicides from different modes of action may deter pollinators from visiting turfgrass weeds, but their temporal influence on floral quality and pollinator foraging is unaddressed. Research experiments were conducted at Blacksburg, VA, in 2023 to assess the effect of four herbicides on floral morphology and ultraviolet (UV) reflectance of three different UV floral classes of weeds and associated pollinator foraging visits. Among 1,080 assessments per weed species, honeybees (Apis mellifera), bumble bees (Bombus spp.), solitary bees (Chelostoma florisomne), and flies (Diptera spp.) accounted for 94%, 2%, 3%, and 1%, respectively, of the total pollinator visitations on white clover (Trifolium repens L.) inflorescences; 71%, 2%, 0%, and 27%, respectively, on dandelion (Taraxacum officinale F.H. Wigg.) flowers; and 0%, 0%, 78%, and 22%, respectively, on bulbous buttercup (Ranunculus bulbosus L.) flowers. Pollinator visitation and floral quality were temporarily affected by herbicide application, with some herbicides eliminating food resources, while others transiently impacted floral quality and density. The combination of 2,4-D + dicamba + MCPP and topramezone eliminated pollinator foraging visits, but on differing temporal scales of 3 d for auxins and 14 d for topramezone. Halosulfuron and sulfentrazone transiently suppressed floral quality and density, with varying degrees of deterrence on pollinators depending on the weed species. All evaluated herbicides reduced radiometric UV reflectance of T. officinale petal apices, but only synthetic auxin and topramezone reduced digitally assessed floral UV-reflecting area. Petal UV reflectance appears to contribute but not solely influence pollinator foraging behavior. UV-absorbing and UV-reflecting flowers differed in UV-reflectance response to herbicides, but pollinators were similarly deterred. Results suggest that herbicides may offer a variety of management solutions to pollinator deterrence in areas slated for insecticide treatment, including long-term or transient deterrence with potential food-resource preservation.

Genera of the Orobanchaceae family are holoparasites that parasitize various hosts. Several members of this family cause severe damage to diverse crop plants. While the biological and life cycles of these parasites have been studied, their metabolism has received little attention, most of which focused on Egyptian broomrape [Orobanche aegyptiaca Pers.; syn.: Phelipanche aegyptiaca (Pers.) Pomel]. This study aimed at obtaining more knowledge about the primary metabolic profiling of four parasite species belonging to the Orobanchaceae family—sunflower broomrape (Orobanche cumana Wallr.), Orobanche cernua, P. aegyptiaca, and branched broomrape [Orobanche ramosa L.; syn.: Phelipanche ramosa (L.) Pomel.]—that developed on tomato (Solanum lycopersicum L.) as a host. Gas chromatography–mass spectrometry analysis demonstrated that significant differences in metabolite content occur between species belonging to Orobanche compared with those belonging to Phelipanche. This finding adds another layer to the separation of these two genera in addition to morphological separation. Moreover, each of these four species exhibits different metabolic profiles, indicating that the parasites can absorb the host’s metabolites but also have the ability to self-regulate their metabolites in order to grow and develop.

Buffalobur (Solanum rostratum Dunal) is an invasive weed in China, and identifying its pathogens is crucial for developing effective biological control measures. In this study, leaf samples from S. rostratum showing typical disease symptoms were collected in Liaoning and Jilin provinces, China. The isolated fungal pathogens were identified based on their morphological characteristics and by using molecular biology techniques. Pathogenicity was assessed by artificially inoculating spore suspensions from the fungal pathogen onto the seeds, isolated leaves, and plants of S. rostratum. The safety of the fungal pathogens for eight other plant species was also evaluated. We then identified the following five fungal pathogens causing disease in S. rostratum in Liaoning and Jilin provinces: Alternaria alternata, Epicoccum sorghinum, Fusarium equiseti, Curvularia hawaiiensis, and Nigrospora oryzae. These fungal pathogens exhibited pathogenicity, with N. oryzae exhibiting the strongest pathogenicity and highest safety. Nigrospora oryzae demonstrated the highest inhibition rate against the radicle germination length of S. rostratum and showed robust pathogenicity toward both isolated leaves and plants. Notably, despite inducing mild reactions in corn (Zea mays L.), grain sorghum [Sorghum bicolor (L.) Moench], rice (Oryza sativa L)., and tomato (Solanum lycopersicum L.), N. oryzae did not have any detrimental effect on the growth of these plants.

The Amaranthus genus contains numerous agronomic weedy species that are widely distributed across the United States. The seeds of many Amaranthus species are morphologically indistinguishable. The Minnesota Department of Agriculture declared Palmer amaranth (Amaranthus palmeri S. Watson) a prohibited noxious weed seed in 2016. Any Amaranthus spp. seeds that are identified as contaminants during routine seed testing require genetic testing to determine whether A. palmeri is present. This research aimed to validate and optimize the molecular detection of A. palmeri in seed. We refined the DNA extraction from pools of Amaranthus spp. seeds ranging in size from 1 to 100 seeds to improve sample testing throughput. Real-time polymerase chain reaction (qPCR) using primers developed by Murphy and colleagues correctly identified the presence of A. palmeri genetics in 84 samples containing 0, 1, 2, 25, or 50 A. palmeri seeds in samples containing up to 100 seeds. The method specificity was examined using 17 Amaranthus species and 4 hybrids of unknown genetics. The Murphy regular qPCR cycle amplified Watson’s amaranth (Amaranthus watsonii Standl.), spleen amaranth (Amaranthus dubius Mart. ex Thell.), and spiny amaranth (Amaranthus spinosus L.), which would result in false-positive calls; however, the fast cycle only identified A. watsonii as a potential false positive. Examination of the Murphy primer binding site revealed an identical sequence for A. palmeri and A. watsonii. Additional markers were evaluated and optimized for use in qPCR to eliminate the risk of a false positive. The additional markers did eliminate the amplification of A. dubius and A. spinosus but did not eliminate the amplification for A. watsonii. Currently, A. watsonii is not known to be distributed outside of its limited native range and is not expected to be encountered in samples.

Carolina geranium (Geranium carolinianum L.) growth in planting holes in commercial strawberry [Fragaria × ananassa (Weston) Duchesne ex Rozier (pro sp.) [chiloensis × virginiana]] fields is a serious problem in Florida. This study aimed to evaluate the effects of different G. carolinianum densities on strawberry growth and yield in plasticulture production systems. Geranium carolinianum densities were 0, 0.4, 0.8, 1.9, 2.7, and 3.8 plants m−2 equally distributed on the plastic-mulched bed top within the planting holes. Geranium carolinianum density did not affect plant height; however, seed production and season-end biomass were negatively correlated with density in Season I. There was a negative, linear correlation between weed density and berry yields. With each increase in G. carolinianum per square meter, the total annual yield was lowered by 554 and 935 kg ha−1, in Seasons I and II, respectively. Our data clearly indicate that G. carolinianum emerging in the transplant holes of strawberry and competing throughout the season has a significantly negative effect on total berry yield.

Goosegrass [Eleusine indica (L.) Gaertn.] is one of the most problematic weeds in plasticulture strawberry [Fragaria × ananassa (Weston) Duchesne ex Rozier (pro sp.) [chiloenis × virginiana]] production systems in Florida. A 2-yr trial was implemented to evaluate the effects of different E. indica densities on strawberry growth and yield. Eleusine indica densities evaluated were 0, 0.4, 0.8, 1.9, 2.7, and 3.8 plants m−2 equally distributed on the plastic-mulched bed top within the strawberry transplant holes. Eleusine indica density did not affect E. indica height or biomass. However, E. indica seed production was positively correlated with E. indica density in Season 1 and negatively correlated with E. indica density in Season 2. A negative linear regression was observed between E. indica density and strawberry yield in both seasons. For each increase in E. indica plants per square meter, strawberry yield was reduced by 316 and 2,356 kg ha−1 for Seasons 1 and 2, respectively. Our results highlight the importance of achieving adequate E. indica management to minimize yield losses.

Weedy rice (Oryza sativa f. spontanea Auct. ex Backer) is a troublesome annual weed from the Gramineae family and infests rice (Oryza sativa L.) fields globally, with a notable presence in China. However, limited information is available regarding the effects of diverse environmental factors on its germination and emergence. A better understanding of the seed biology and ecology of weedy rice is crucial for developing effective weed management strategies. Experiments were conducted to evaluate the effects of temperature, light, soil burial depth, wheat (Triticum aestivum L.) crop residue amount, salt stress, osmotic stress, and radiant heat on the germination and seedling emergence of weedy rice. Weedy rice exhibited robust germination (>98%) when exposed to varying day/night temperatures (20/15 to 35/30 C) and remained unaffected by light conditions. Seedling emergence was not influenced within the top 5-cm soil layer, where 100% of the seedlings emerged. However, emergence decreased as the soil burial depth increased, eventually resulting in no emergence from a burial depth of 11 cm. The soil burial depth required for 50% of the maximum emergence was 8.3 cm. Seedling emergence ranged from 97% to 100% across different amounts of the wheat straw residue cover (0 to 10,000 kg ha−1). The sodium chloride concentration and osmotic potential required for 50% were 230.8 mM and −0.5 MPa, respectively. No germination was observed when weedy rice seeds were exposed to >110 C pretreatment (radiant heat for 5 min), indicating that residue burning could reduce infestation of weedy rice. The insights gained from this study contribute valuable knowledge to enhance the integrated management of weedy rice in China.

Purple nutsedge (Cyperus rotundus L.) is one of the world’s resilient upland weeds, primarily spreading through its tubers. Its emergence in rice (Oryza sativa L.) fields has been increasing, likely due to changing paddy-farming practices. This study aimed to investigate how C. rotundus, an upland weed, can withstand soil flooding and become a problematic weed in rice fields. The first comparative analysis focused on the survival and recovery characteristics of growing and mature tubers of C. rotundus exposed to soil-flooding conditions. Notably, mature tubers exhibited significant survival and recovery abilities in these environments. Based on this observation, further investigation was carried out to explore the morphological structure, nonstructural carbohydrates, and respiratory mechanisms of mature tubers in response to prolonged soil flooding. Over time, the mature tubers did not form aerenchyma but instead gradually accumulated lignified sclerenchymal fibers, with lignin content also increasing. After 90 d, the lignified sclerenchymal fibers and lignin contents were 4.0 and 1.1 times higher than those in the no soil-flooding treatment. Concurrently, soluble sugar content decreased while starch content increased, providing energy storage, and alcohol dehydrogenase activity rose to support anaerobic respiration via alcohol fermentation. These results indicated that mature tubers survived in soil-flooding conditions by adopting a low-oxygen quiescence strategy, which involves morphological adaptations through the development of lignified sclerenchymal fibers, increased starch reserves for energy storage, and enhanced anaerobic respiration. This mechanism likely underpins the flooding tolerance of mature C. rotundus tubers, allowing them to endure unfavorable conditions and subsequently germinate and grow once flooding subsides. This study provides a preliminary explanation of the mechanism by which mature tubers of C. rotundus from the upland areas confer flooding tolerance, shedding light on the reasons behind this weed’s increasing presence in rice fields.

The Northern Great Plains of the United States is a major production region for organic pulse crops that are prone to yield losses due to weeds. Weed management in organic systems relies on the integration of several tactics to stack additive effects and for redundancy to deal with variable efficacy of individual weed management practices. To address the need for effective, integrated weed management, we conducted a 2-yr trial that evaluated the effects of planting date, seeding rate, and preemergent weed control practices (shallow tillage and flame weeding) on weed biomass, crop density, and yield in organically managed chickpea (Cicer arietinum L.) in Montana. Stacking weed management practices increased yields. Early planting had the largest effects on yields, increasing them by 1.8- to 3.6-fold compared with planting 10 to 14 d later. Increasing the seeding rate from the standard rate (43 viable seeds m−2) by 50% increased yields by 47% from 889 to 1,304 kg ha−1. Both preemergent weed control practices increased yields by 40% to 50 % relative to the non-weeded controls. By integrating all three practices, yields of organic chickpea increased greater than 6-fold from 318 kg ha−1 in the controls to 2,006 kg ha−1. The effects of weed control treatments on midseason weed biomass were complex and variable. Although efficacy of the cultural (seeding rate and planting date) and physical (preemergent) treatments on weed biomass varied between years and when combined with other treatments, their full integration, that is, early planted, at higher seeding rate, and preemergent weed control, produced consistently lower weed biomass (84% reduction on average) compared with the standard grower practice (later planting, standard seeding rate, no preemergent weed control). The results lend support to the concept of integrating multiple weed management practices to achieve weed control and high yields in organically managed crops.

Opportunistic use of limited resources is often attributed to invasive species, and as a mature vine, old man’s beard (Clematis vitalba L.) is known to have devastating negative impacts on the trees it colonizes. No previous experimental studies have been published on how easily C. vitalba seedlings can colonize ground covered by other established vegetation. This species has had an increasing presence in forestry blocks and riparian zones in New Zealand, both of which usually maintain some grass cover. To determine the importance of vegetative ground cover for preventing ingress of new C. vitalba plants, this study looked at seedling emergence through the soil and establishment of C. vitalba within four different levels of grassy cover at three sites: (1) ground kept bare after vegetation removal; (2) ground bare at C. vitalba seed sowing, but thereafter allowed to recover; (3) vegetative cover trimmed to 4 cm high at C. vitalba sowing, and then allowed to recover; and (4) unmanaged vegetation. At the highest level of vegetation density (unmanaged vegetation), no C. vitalba seedlings were ever detected throughout a 1-yr monitoring period. At lower ground cover densities, poor seedling emergence was observed, with a maximum of 36% of seeds sown in bare plots producing a seedling. Also, seedlings did not survive past 1 yr, except in bare plots or in plots where vegetation grew sparsely. However, seedlings that did survive began producing multiple stems within 6 mo of emergence. These results indicate that obstacles to seedling emergence and poor development at the young seedling stage when vegetative cover is dense severely limit C. vitalba’s chances to invade new sites via seed. Yet some successful seedling recruitment does occur due to the magnitude of the propagule pressure on the landscape and the difficulty of maintaining high-density ground cover across large areas throughout the year.

Mounting cases of herbicide-resistant waterhemp [Amaranthus tuberculatus (Moq.) Sauer] in the U.S. Midwest have renewed the interest in nonchemical weed management strategies. Field experiments were conducted in 2021 and 2022 to quantify the effectiveness of a commercial combine equipped with a seed impact mill in preventing A. tuberculatus seed return to the soil seedbank in soybean [Glycine max (L.) Merr.]. Amaranthus tuberculatus seed shattering before crop harvest was quantified. Amaranthus tuberculatus started shattering seeds during the last week of August in both years. Overall, 51% of A. tuberculatus seeds were retained on the plant at harvest on October 23, 2021, compared with 61% at harvest on October 7, 2022. Viability of shattered A. tuberculatus seeds ranged from 84% to 94%. Additional seed shattering occurred when plants were disturbed by the combine header during soybean harvest, which caused 15% and 9% shattering in 2021 and 2022, respectively. Amaranthus tuberculatus seeds passed through the impact mill were grouped in three categories: no damage, moderate damage, and severe damage. In 2021, A. tuberculatus seeds with moderate damage had 26% lower germination and viability than seeds with no visible damage. In 2022, seed germination and viability of no-damage seeds did not differ from seeds with a moderate level of damage. No severely damaged seed germinated or tested viable in either year. Altogether, impact mill treatment reduced the number of germinable seeds by 87% compared with the no–impact mill treatment. These results indicate that seed impact mills can be a useful tool in Iowa soybean production to help manage multiple herbicide–resistant A. tuberculatus populations. However, A. tuberculatus seed shattering before crop harvest reduces the overall effectiveness of seed impact mills in preventing seedbank replenishments.

Herbicide drift to sensitive crops can result in significant injury, yield loss, and even crop destruction. When pesticide drift is reported to the Georgia Department of Agriculture (GDA), tissue samples are collected and analyzed for residues. Seven field studies were conducted in 2020 and 2021 in cooperation with the GDA to evaluate the effect of (1) time interval between simulated drift event and sampling, (2) low-dose herbicide rates, and (3) the sample collection methods on detecting herbicide residues in cotton (Gossypium hirsutum L.) foliage. Simulated drift rates of 2,4-D, dicamba, and imazapyr were applied to non-tolerant cotton in the 8- to 9-leaf stage with plant samples collected at 7 or 21 d after treatment (DAT). During collection, plant sampling consisted of removing entire plants or removing new growth occurring after the 7-leaf stage. Visual cotton injury from 2,4-D reached 43% to 75% at 0.001 and 0.004 kg ae ha−1, respectively; for dicamba, it was 9% to 41% at 0.003 or 0.014 kg ae ha−1, respectively; and for imazapyr, it was 1% to 74% with 0.004 and 0.03 kg ae ha−1 rates, respectively. Yield loss was observed with both rates of 2,4-D (11% to 51%) and with the high rate of imazapyr (52%); dicamba did not influence yield. Herbicide residues were detected in 88%, 88%, and 69% of samples collected from plants treated with 2,4-D, dicamba, and imazapyr, respectively, at 7 DAT compared with 25%, 16%, and 22% when samples were collected at 21 DAT, highlighting the importance of sampling quickly after a drift event. Although the interval between drift event and sampling, drift rate, and sampling method can all influence residue detection for 2,4-D, dicamba, and imazapyr, the factor with the greatest influence is the amount of time between drift and sample collection.

Machine vision–based herbicide applications relying on object detection or image classification deep convolutional neural networks (DCNNs) demand high memory and computational resources, resulting in lengthy inference times. To tackle these challenges, this study assessed the effectiveness of three teacher models, each trained on datasets of varying sizes, including D-20k (comprising 10,000 true-positive and true-negative images) and D-10k (comprising 5,000 true-positive and true-negative images). Additionally, knowledge distillation was performed on their corresponding student models across a range of temperature settings. After the process of student–teacher learning, the parameters of all student models were reduced. ResNet18 not only achieved higher accuracy (ACC ≥ 0.989) but also maintained higher frames per second (FPS ≥ 742.9) under its optimal temperature condition (T = 1). Overall, the results suggest that employing knowledge distillation in the machine vision models enabled accurate and reliable weed detection in turf while reducing the need for extensive computational resources, thereby facilitating real-time weed detection and contributing to the development of smart, machine vision–based sprayers.

Effective pesticide application is dependent on precise and sufficient delivery of active ingredients to targeted pests. Water-sensitive papers (WSPs) have been used to estimate the stain coverage, droplet density, droplet size, total spray volume, and other spray-quality metrics by analyzing deposit stains using image analysis software. However, because WSPs are expensive, they are typically distributed along unidimensional transects at intervals of 0.5 m or more, which comprises 0.5% or less of the total treated area. This might limit the ability to accurately represent the deposition of agricultural sprayers with irregular patterns, such as agricultural drone sprayers in the early developmental stage. This study introduces a novel approach utilizing white Kraft paper and a blue colorant proxy for assessing spray deposition. A custom Python-based image analysis tool, SprayDAT (Spray Droplet Analysis Tool), was developed and compared with traditional image analysis software, DepositScan. Both models showed increased accuracy in detecting larger objects, with SprayDAT generally performing better for smaller droplets. DepositScan underestimated the total deposited spray volume by up to 2.7 times less compared with the colorant extraction assessed via spectrophotometry and the predicted output based on flow rate, coverage, and speed. Accuracy of software-estimated spray volume declined with increasing total stain coverage, likely due to overlapping stain objects. Droplet density exhibited a Gaussian trend, with peak density at approximately 22% stain cover, offering evidence for overlapped stains for both DepositScan and SprayDAT as stain cover increased. Both models showed exponential growth in volumetric median diameter (VMD) with increasing stain cover. SprayDAT is freely accessible through an online repository. It features a user-friendly interface for batch processing large sets of scanned images and offers versatility for customization to meet individual needs, such as adjusting spread factor, updating the standard curve for spray volume estimation, or modifying the stain detection threshold.

Agricultural spray drone (ASD) use in managed turfgrass has been given limited attention in the scientific literature. Further, deposition patterns of ASD spray have been obscured in previous research by ambient wind, crop canopy interference, and limited sampling resolution. Using a continuous sampling method involving blue colorant and water sprayed over white Kraft paper that was assessed via digital image analysis of stain objects and referenced spectrophotometric analysis of extractant, deposition metrics were estimated across a 29.3-m transect perpendicular to an ASD or ground-sprayer spray swath. The ASD applies very fine droplets that are highly concentrated with herbicide, similar to ultra–low volume treatments, that improved smooth crabgrass [Digitaria ischaemum (Schreb.) Schreb. ex Muhl.] control compared with a ground sprayer when the ASD was operated 2 m above the turf. Unfortunately, these very fine droplets also drift, leading to four times greater droplet density at distance of almost 12 m away from the targeted spray swath following an operational height of 10 m compared with 2 m. As ASD operational height increases, drift and effective swath width at 30% coefficient of variation uniformity increases, while effective application rate, total deposition, and D. ischaemum control by quinclorac herbicide decreased. Total deposition decreased 6% for each meter increase in ASD operational height, likely due to evaporation. The potential losses due to evaporation are a serious consideration for ASD use that has received little attention in the scientific literature. Our data suggest that ASD operational height should be as low as possible, but modification of spray systems may be needed to improve homogeneity of spray pattern.