Materials and MethodsTwo experiments were designed to compare the effect of the spray nozzle and carrier volume on hazelnut sucker control. A season-long study was designed to compare treatments after four applications following available recommendations (Olsen and Peachey, 2013). A short-term study followed a similar treatment structure but was concluded 28 d after treatment.Season-long study.The season-long study was conducted in 2017 and 2018 in a mature hazelnut orchard located in Amity, OR. The site is a well-drained Woodburn silt loam soil (Soil Survey Staff, 2020). The orchard was a 10-year-old Jefferson cultivar planted at 6 × 6 m. Irrigation was applied using a single drip line per tree row. Suckers were located up to 0.5 m around the main trunk, and all trees had 15 suckers on average at the initiation of the experiments. The first application was made at an average sucker height of 14 ± 5 cm. Treatments were applied using a CO2 pressurized backpack sprayer equipped with a three-nozzle boom. An application consisted of one pass to each side of the tree row. The sprayer was calibrated before each application to ensure the carrier volume was delivered within ± 5% of the targeted volume. Variable travel speed and different spray nozzles were used to achieve desired volumes. Four consecutive applications were performed 4 weeks apart between May and August of each year. In 2018, the experiment was repeated in a different section of the same orchard.The treatments compared were the systemic herbicide 2,4-D (Saber; Loveland Products, Loveland CO) at 1060 g·ha−1 acid equivalent (a.e.), and the contact herbicide glufosinate (Rely 280; BASF, Durham, NC) at 1150 g·ha−1 a.i. Treatments included ammonium sulfate at a rate equivalent to 10 g·L−1. A non-ionic surfactant was included in the 2,4-D treatment at 0.25% v/v (Rainier EA, Wilbur Ellis, Aurora, CO). Herbicides were applied at 187 or 374 L·ha−1 at 275 kPa of pressure. We tested the extended range (XR) flat-fan nozzle (TeeJet Technologies, Glendale Heights, IL) XR11002 at 187 L·ha−1 and XR11004 for 374 L·ha−1; these nozzles generate fine and medium droplets (based on manufacturer’s information) (TeeJet Technologies, 2020). We also tested drift-reduction nozzles Turbo TeeJet Induction (TTI) TTI11002 and TTI11004 at 187 L·ha−1 and XR11004 for 374 L·ha−1, respectively. According to the manufacturer, these nozzles have a pre-orifice chamber and air-induction to generate ultra-coarse droplets. Droplet nomenclature is based on the American Society of Agricultural and Biological Engineers (ASABE) standard S-572.1 (ASABE, 2009). Manual pruning was included as a reference, with suckers removed at the same time the herbicide treatments were applied. An experimental unit consisted of eight trees, and each individual tree was treated as a sub-sample. An average assessment was given to each plot.Assessments.Sucker control was estimated on a scale of 0 to 100, where 0 is no control, and 100 is complete control. The average height of 10 suckers per experimental unit was assessed at 28 d after treatment (DAT) for each application. Sucker dry weight and cross-sectional area were recorded 28 d after the final treatment. The cross-sectional area was measured by recording the diameter of 20 suckers per experimental unit, using a caliper (Fisherbrand Traceable Digital Calipers; Thermo Fischer Scientific, Waltham, WA). Diameter data were transformed into a cross-sectional area. Crop damage was not observed during the experiment.Short-term studies.The short-term field studies followed the same experimental design and evaluation as previously described, except that treatments were applied only once, and the experiment was concluded after 28 d. Hazelnut growers treat suckers about every 4 weeks or less. A total of four studies were conducted in 2018. Two studies were in Canby, OR (lat. 45°17′N, long. 122°39′W), on a Latourell loam soil (Soil Survey Staff, 2020). Trees were spaced at 6 × 6 m and were rainfed. The orchard consisted of 12-year-old ‘McDonald’ trees. The first trial was initiated in May when suckers were 15 cm ± 4 cm in height, and the second study was initiated in June about 4 weeks after the first application of a burndown herbicide by the grower. The suckers were 14 ± 4 cm in height. Two different sections of the same orchard were used for each study.Two other trials were conducted in Corvallis, OR (lat. 44°29′N, long. 123°13′W) on ‘Jefferson’ hazelnut trees growing in a Chehalis silt loam soil (Soil Survey Staff, 2020). The first trial began in May when suckers had reached 16 ± 5 cm in height. The second study was initiated in June in a different area of the same orchard when suckers were 14 ± 5 cm in height. The site of the second study was treated 4 weeks prior with a burndown herbicide by the grower. Assessments were made as described in the season-long study. Crop damage was not observed during the experiment.Statistical analysis.The long-term experiment was a four-factor factorial study organized as a randomized complete block design with four replicates. The factors were application timing, herbicide, nozzle type, and carrier volume. Data were submitted to a generalized linear mixed model using RStudio (RStudio Team, 2020). Random factors included experimental year, block, and their interactions, because the focus of the study was to infer across years. The data were analyzed using the package glmmTMB version 1.01 (Brooks et al., 2017). Control data were analyzed with beta distribution, a suitable distribution for proportions, and percentages of continuous data (Douma and Weedon, 2019). Height, weight, and cross-sectional area were analyzed using a Gaussian distribution. Means were compared using Tukey’s honestly significant difference test. Orthogonal tests were designed for specific comparisons of the effects of carrier volume and nozzle type for each herbicide. We observed a significant effect of application timing and treatment. Data were analyzed by the application event.The short-term studies were analyzed as a three-factor factorial study organized as a randomized complete block design with four replicates. The factors were herbicide, nozzle type, and carrier volume. The experiment was repeated in four locations. Random factors included experimental location, block, and their interactions. Data analysis was like that described for the long-term study.