Exploring Chemical and Cultural Weed Management for Industrial Hemp Production in Georgia, USA

Abstract

Industrial hemp (Cannabis sativa) production is complex, with strict regulatory constraints and challenges associated with a lack of labeled pesticides due to its status as a novel crop in the US. Four experiments were conducted in 2020 and 2021 to establish herbicide tolerances for hemp production in the coastal plain of Georgia, USA. Objectives included evaluating hemp response to pretransplant or posttransplant herbicides, determining if planting method influenced herbicide injury from residual preplant applied herbicides, and understanding how plastic mulch may influence hemp flower yields. When applied one day prior to transplanting, maximum hemp crop visual injury was less than 12% compared to the untreated control, with acetochlor, flumioxazin, fomesafen, pendimethalin, and norflurazon while dithiopyr, halosulfuron, isoxaben, and isoxaflutole resulted in greater than 50% injury. Posttransplant applications of S-metolachlor, acetochlor, pendimethalin, and clethodim resulted in less than 15% injury while halosulfuron, metribuzin, trifloxysulfuron, imazethapyr, and prometryn applications resulted in greater than 50% injury to plants. Preplant and posttransplant applied herbicides were found to have little effect on total tetrahydrocannabinol (THC), cannabidiol (CBD), or total cannabinoids in the dry flower after harvest. In a separate experiment, injury from halosulfuron and metribuzin was 52% to 56% less when planted with a mechanical transplanter as compared to the practice of using a transplant wheel to depress a hole in the soil followed by hand transplanting. In the final experiment, hemp dry flower yield in a non-plastic mulched (bareground) system was similar to that in a plastic mulched system. However, early season plant above-ground biomass was less in the plastic mulched system, which may have been due to elevated soil temperatures inhibiting early season growth.

1. Introduction

Industrial hemp (Cannabis sativa L.) may be utilized for fiber, grain, or flower production [1]. Interest in growing hemp, particularly for the floral material, has increased following the broad legalization of industrial hemp in the United States with the 2018 Farm Bill [2]. Floral hemp production is focused on harvesting unfertilized female flowers that have high concentrations of cannabinoids and terpenes, which are present in the leaves and flowers of the plant [1]. Interest in floral hemp often focuses on cannabidiol (CBD) production as this compound is reported to have health benefits such as alleviating anxiety or inflammation [3].

Floral hemp production is similar to vegetable production in many ways, suggesting that Georgia, USA vegetable growers could adopt the crop rapidly [4]. Hemp grown for the floral market is widely spaced in fields and often grown using plastic mulch. Many vegetables grown for fresh market produce are transplanted and grown using plastic mulch systems, which provide several benefits, including higher yields, quality, and weed control [5]. A study on tomato yield found that total yield was higher when plants were grown in a plastic mulch system (40 MT·ha⁻¹) compared to bareground production (24.6 MT·ha⁻¹) [6]. Furthermore, that same study reported a larger average fruit weight and a higher percentage of marketable fruit from plants grown using plastic mulch, indicating better fruit quality over bareground production systems. Additionally, Monks et al. [7] found that plastic mulch controlled weeds 70–99% at 8 to 9 weeks after tomato transplanting. In those studies, weed control was only reduced due to weed germination and growth in the plant hole.

Managing weeds is crucial to maximize yield and quality of industrial hemp; however, at this time, there is only one synthetic herbicide, ethalfluralin, currently labeled for hemp in the US, though these labels are restricted to specific states and uses [8]. Previous research conducted in hemp grown for fiber indicates it is sensitive to the application of many herbicides applied prior to (PRE) and after (POST) crop emergence [9,10,11,12,13,14,15,16,17]. Amaducci et al. [18] reported S-metolachlor and acetochlor were effectively used in the production of grain and fiber hemp varieties in China. However, Ortmeier-Clarke et al. [17] reported that several of these herbicides reduced grain hemp plant biomass by at least 50% when applied PRE to two different grain hemp varieties in greenhouse studies. Injury from 2130 g·ha⁻¹ pendimethalin, 563 g·ha⁻¹ metribuzin, 263 g·ha⁻¹ fomesafen, 107 g·ha⁻¹ flumioxazin, and 79 g·ha⁻¹ isoxaflutole in the same study reduced plant biomass 70% to 100%. Clopyralid at 158 g·ha⁻¹ and 76.4 g·ha⁻¹ clethodim applied POST resulted in less than 25% reduction in biomass, while other herbicides such as atrazine (1680 g·ha⁻¹), fomesafen (263 g·ha⁻¹), and imazethapyr (70 g·ha⁻¹) reduced biomass at least 75%. Prior research has reported that hemp response to herbicides was variable. Flessner et al. [10] reported injury of 0% to 10%, 10% to 55%, and 25% to 50% at 30 d after S-metolachlor, fomesafen, and pendimethalin, respectively, were applied immediately after seeding grain and fiber hemp varieties. Variability was also prevalent with POST applied herbicides where injury from sethoxydim and bromoxynil POST ranged from 0% to 15% and 0% to 18%, respectively.

Most available research evaluating hemp responses to herbicides was conducted on direct seeded fiber or grain hemp cultivars. However, the response of transplanted floral hemp to herbicides has not been thoroughly evaluated. Production practices for floral hemp are not the same as for hemp grown for fiber or grain, potentially influencing plant response [4,19]. For example, fiber hemp is typically densely seeded while hemp grown for the floral market is generally transplanted at low populations. It is common for cultivars across cropping systems to respond differently to herbicides [20,21]. Young et al. [22] demonstrated a differential response of rice (Oryza sativa L.) cultivars based on parent origin. It is possible that fiber hemp cultivars vary widely in growth habit and maturity from those selected for floral production [1,23].

In order to provide new insights for potential chemical and cultural weed management options in floral hemp production in the US, three objectives were studied. The first objective was to evaluate floral hemp tolerance to 13 herbicides applied to soils prior to transplanting and floral hemp response to 13 herbicide treatments made 2 to 5 weeks after transplanting. The second objective was to compare how floral hemp responded to preplant residual herbicides when planted with a mechanical transplanter compared to a hand planting technique. The third objective was to compare bareground and plastic mulch production systems for productivity of floral hemp.

2. Materials and Methods

Hemp seeds were planted into potting mix (Miracle-Gro Potting Mix; Scotts Miracle-Gro, Marysville, OH, USA) using 128-cell Styrofoam trays (Speedling, Ruskin, FL, USA). Seedlings were greenhouse-grown where conditions consisted of natural light supplemented with metal halide lamps (400 µmol·m⁻²·s⁻¹) to provide a 24 h photoperiod [24] in Tifton, GA, USA (31°5′ N, 83°5′ W). Greenhouse temperature was set to 27 °C (day/night). From seeding until 7 d after emergence, trays were floated in water, then seedlings were watered overhead twice daily and fertilized using a 67 mg·L⁻¹ nitrogen (N) solution (24N-3.5P-13.3K; Miracle-Gro; Scotts Miracle-Gro) once every 5 d. When plants reached the 3-leaf growth stage (approximately 21 d), they were moved to an outdoor open-air shelter to be hardened off for 4 d prior to transplanting.

Four field experiments were conducted including a pretransplant herbicide screening (2020 and 2021), a posttransplant herbicide screening (2020 and 2021), a planting method comparison (two field locations in 2021), and a plastic mulch vs. non-plastic mulch production system comparison (2020 and 2021) for a total of 8 studies. Each experiment was conducted at the University of Georgia Ponder Research Farm near Ty Ty, GA, USA (31°5′ N, 83°6′ W). Soil was a Tifton loamy sand [25] consisting of 89.6% sand, 8% silt, 2.4% clay, and 0.64% OM with a pH of 6.3.

2.1. Herbicide Screening Experiments

For the two herbicide screening studies (PRE and POST), land was prepared using a disk harrow (International Harvester, Chicago, IL, USA) and fertilized with 560 kg·ha⁻¹ 10N-4.4P-8.3K (Big Bend Agri-Services, Inc., Cairo, GA, USA). Beds were formed while laying drip tape (30-cm emitter spacing, T-Tape; Rivulus, Madera, CA, USA) 8 cm beneath the soil surface, then the bed was shallowly tilled (Maletti, Modena, Italy) to create a level surface for planting. Beds were 1.8 m center to center and 3 cm tall. In-season fertilizer was applied via drip irrigation totaling 196 kg·ha⁻¹ N (7-0-7 Liquid feed; Big Bend Agri-Services, Inc., Cairo, GA, USA). The experimental design was a randomized complete block with 3 replications in 2020 and 4 replications in 2021. Initial plant in-row spacing was 0.6 m, with 10 plants per plot. Biomass was collected by removing every other plant resulting in a final in-row plant spacing of 1.2 m.

Herbicide treatments in the pretransplant experiment consisted of acetochlor, dithiopyr, flumioxazin, fomesafen, halosulfuron-methyl, isoxaben, isoxaflutole, metribuzin, norflurazon, oxyfluorfen, pendimethalin, prometryn, and S-metolachlor (

Table 1. Pretransplant herbicide treatments including product trade name, manufacturer, and herbicide active ingredient (ai) rate.

Herbicide	Product	Manufacturer	Rate ai (g·ha−1)
acetochlor	Warrant	Bayer CropScience, St. Louis, MO, USA	841
dithiopyr	Dimension® EC	Corteva Agriscience, Indianapolis, IN, USA	280
flumioxazin	Valor®SX	Valent U.S.A. Corporation, Walnut Creek, CA, USA	72
fomesafen	Reflex®	Syngenta Crop Protection, Greensboro, NC, USA	210
halosulfuron-methyl	Sandea®	Gowan Company, Yuma, AZ, USA	39
isoxaben	Gallery®	Corteva Agriscience Indianapolis, IN, USA	631
isoxaflutole	Alite™ 27	BASF Corporation, Research Triangle Park, NC, USA	112
metribuzin	Tricor® DF	United Phosphorus, Inc., King of Prussia, PA, USA	210
norflurazon	Solicam® DF	Tessenderlo Kerley, Inc. Phoenix, AZ, USA	138
oxyfluorfen	Goal® 2XL	Corteva Agriscience, Indianapolis, IN, USA	421
pendimethalin	Prowl® H2O	BASF Corporation, Research Triangle Park, NC, USA	1065
prometryn	Caparol® 4L	Syngenta Crop Protection, Greensboro, NC, USA	1121
S-metolachlor	Dual Magnum®	Syngenta Crop Protection, Greensboro, NC, USA	1068

). Herbicide applications were made on 11 June 2020 and 9 June 2021, one day before transplanting hemp seedlings ‘Von’ (Sunbelt Hemp Source, Moultrie, GA, USA). Overhead irrigation of 0.6 cm was implemented after application and prior to planting. Seedlings were transplanted using a mechanical transplanter (Mechanical Transplanter Co., Holland, MI, USA).

For the POST experiment, hemp ’Sunbelt Crush’ (Sunbelt Hemp Source) was planted at the same time as the PRE study, on 12 June 2020 and 10 June 2021 with herbicides broadcast applied when plants reached 13–18 cm in height on 29 June 2020 and 16 July 2021. Herbicide treatments included acetochlor, clethodim, halosulfuron-methyl, imazethapyr, metribuzin, pendimethalin, prometryn, S-metolachlor, and trifloxysulfuron (

Table 2. Posttransplant herbicide treatments including product trade name, manufacturer, and herbicide rate ^a.

Herbicide	Product	Manufacturer	Rate (g·ha−1)
acetochlor	Warrant	Bayer CropScience St. Louis, MO, USA	841
clethodim	Select Max®	Valent U.S.A. Corporation Walnut Creek, CA, USA	136 *
halosulfuron-methyl	Sandea®	Gowan Company Yuma, AZ, USA	26 *
halosulfuron-methyl	Sandea®	Gowan Company	53 *
imazethapyr	Pursuit®	BASF Corporation, Research Triangle Park, NC, USA	35 *
metribuzin	Tricor® DF	United Phosphorus, Inc. King of Prussia, PA, USA	210
metribuzin	Tricor® DF	United Phosphorus, Inc.	420
metribuzin	Tricor® DF	United Phosphorus, Inc.	210 *
metribuzin	Tricor® DF	United Phosphorus, Inc.	420 *
pendimethalin	Prowl® H2O	BASF Corporation	1065
prometryn	Caparol® 4L	Syngenta Crop Protection, Greensboro, NC, USA	1121 *
S-metolachlor	Dual Magnum®	Syngenta Crop Protection	1068
trifloxysulfuron	Envoke®	Syngenta Crop Protection	5 *

^a * indicates NIS at 0.25% v/v was included.

). All herbicide treatments were applied using a carbon dioxide (CO₂)-pressurized backpack sprayer equipped with AIXR 11002 nozzles (TeeJet Technologies, Springfield, IL, USA) calibrated to deliver 140 L·ha⁻¹ at 207 kPa (30.0 psi).

Herbicides in both the PRE and POST experiments were selected due to their efficacy and use in vegetable or row crop production. Documenting crop response was the objective for both herbicide screening experiments as weed control with these herbicides has been well studied. At 2 weeks after transplanting, hemp plants were covered with plastic nursery pots (22 cm diameter by 21 cm tall) and the area between plants was treated with 3.3 kg·ha⁻¹ glyphosate (Roundup PowerMAX II; Bayer CropScience LP, St. Louis, MO, USA), 36 g·ha⁻¹ flumioxazin (Valor SX; Valent U.S.A. Corporation, Walnut Creek, CA, USA), 421 g·ha⁻¹ acetochlor (Warrant; Bayer CropScience LP, St. Louis, MO, USA), and 421 g·ha⁻¹ ethalfluralin (Sonalan; Gowan Company, LLC, Yuma, AZ, USA) to manage weeds in the study. After these herbicides were applied, overhead irrigation of 0.6 cm was implemented prior to removal of pots to facilitate herbicide activation without splashing treated soil on the hemp plants. Weeds were hand removed as needed around the plant. No insecticides or fungicides were applied to plants during growth.

Visual injury on a 0%, no injury, to 100%, plant death, scale was estimated weekly for 6 weeks after treatment (WAT). Above ground biomass was measured 2 to 3 WAT by removing every other plant in a plot by hand, removing roots, and weighing all above ground plant material collected per plot. Plant heights were measured weekly for 4 WAT measuring from the soil surface to the terminal meristem on each plant. Five hemp plants per plot were harvested when it was determined that approximately half of the trichomes on the flower were amber colored, which occurred between 16 and 21 Sep of each year. This harvest period also corresponded to growers in the region who were utilizing the same cultivars. After harvest, plants were inverted and hung in a covered barn under ambient conditions for at least 2 weeks to air dry. After plants were dried, the terminal flower (8–10 cm) of all plants from plots treated with acetochlor, dithiopyr, flumioxazin, fomesafen, norflurazon, pendimethalin, and the nontreated control PRE-treatments and acetochlor, clethodim, pendimethalin, S-metolachlor, and nontreated control POST treatments were collected. Treatments chosen to be tested were those most likely to be implemented in future weed management programs based on crop response during the season. These composite samples from each plot were weighed and included in total floral weight. Floral samples were sent to a commercial laboratory (SJ Labs & Analytics, Macon, GA, USA) for determination of the acidic and neutral (decarboxylated) forms of THC, CBD as well minor cannabinoids, delta-8 tetrahydrocannabinol acid, cannabigerol, cannabinol, and cannabichromene according to Storm et al. [26]. After removal of the apical meristem, the remaining flowers and leaves were hand-removed from all plants and weighed by plot to obtain yield.

Injury, biomass, cannabinoid content, plant height, and yield were analyzed using JMP Pro 15 (SAS Institute, Cary, NC, USA) using the Fit Model procedure. Replication was nested in year, and both were treated as random effects while herbicide treatment was a fixed effect. Means were separated using Tukey’s Honestly Significant Difference (HSD) test at α = 0.05 when the model was significant.

2.2. Transplant Method Comparison

In 2021, an experiment was conducted twice to determine if transplant method affected residual herbicide injury. The experiment was a split-plot two-factor factorial design. The first factor, herbicides, consisted of halosulfuron (Sandea; Gowan Company, Yuma, AZ, USA) at 39 g·ha⁻¹ or metribuzin (Tricor DF; United Phosphorus, Inc., King of Prussia, PA, USA) at 210 g·ha⁻¹ applied preplant as the whole-plot with the split-plot being transplant method including mechanical transplanting or hand-planted into holes punched into the soil. Land preparation and preplant fertilizer was as described previously. Transplanting was accomplished using a mechanical transplanter (Model 5500W; Mechanical Transplanter Co.) that opened a furrow, the plant was then placed in the furrow, and the furrow was closed with two wheels pressing soil around the plant root ball. The hole punch transplanting method was made using a mechanical transplant wheel (Kennco Manufacturing, Ruskin, FL, USA) which compressed the soil making a transplant hole 4 cm deep and wide. Hemp was transplanted by hand after holes were created. Nine plants were planted in each plot at the first planting, and five plants were planted in each plot at the second planting (study). In-row plant spacing was 0.6 m and between-row spacing was 1.8 m. A non-herbicide treated control was included for each planting method. Halosulfuron and metribuzin were selected due to the injury observed in prior herbicide screening studies. Herbicide applications were made using a CO₂-pressurized backpack sprayer equipped with TTI 110015 nozzles (TeeJet Technologies, Springfield, IL, USA) calibrated to deliver 140 L·ha⁻¹ at 276 kPa. Herbicide applications were made on 9 June 2021 (study 1) and 15 June 2021 (study 2) with hemp ’Sunbelt Crush’ transplanted one day after herbicide applications and after irrigation of 0.6 cm. No other pesticides were applied to plants in this study.

Visual injury (0–100%) and plant heights were measured as previously described in the herbicide screening studies. At 4 WAT, all plants in the plot were collected and above ground biomass weighed. Data were analyzed using JMP Pro 15 (SAS Institute) using the Fit Model procedure. Study and replication nested within study were considered random effects and herbicide, planting method, and the interaction between herbicide and planting method were fixed effects. The interaction was significant for all variables and is reported. Means were separated using Tukey’s Honest Significant Difference (HSD) test at α = 0.05 when the model was significant.

2.3. Plastic Mulch and Bareground Production System Comparison

In December 2019 and December 2020, soil was prepared as described previously and raised beds 25-cm tall and 81-cm wide were formed using a bedder shaper (Hendrix and Dail, Inc., Greenville, SC, USA). As beds were formed, chloropicrin and 1,3-dichloropropene (Pic-Chlor 60, TriEst Ag Group Inc., Greenville, NC, USA) at 197 L·ha⁻¹ were injected 20 cm deep in the soil using three fumigant knives spaced evenly across the bed. Immediately following the initial fumigation, a separate bed shaper (Hendrix and Dail) further pressed the bed to a height of 20 cm with the width remaining at 81 cm. During this process metam-sodium (Vapam^® HL; AMVAC, Los Angeles, CA, USA) at 700 L·ha⁻¹ was injected into the bed at a depth of 10 cm with injection knives placed 10 cm apart across the bed. Additionally, drip tape (30-cm emitter spacing, T-Tape; Rivulus, Madera, CA, USA) was placed 3 cm below the bed surface while black low-density polyethylene plastic mulch (Guardian Agro Plastics, Tampa, FL, USA) covered the bed. Bareground plots were established at the same time using the same process except plastic mulch was removed one day prior to planting.

Treatments consisted of production on plastic mulch and production on bare soil with each treatment being replicated four times. On 3 June 2020 and 10 June 2021 holes were pressed through the plastic mulch or into bare soil using a mechanical transplant wheel and 10 hemp plants per plot, hemp seedlings ‘Von’ (Sunbelt Hemp Source), were transplanted by hand. Initial plant spacing was 0.6 m, above ground biomass was collected by removing every other plant resulting in a final plant spacing of 1.2 m. Hemp seedlings were grown as described previously. Soil temperature at planting was measured using a digital pocket thermometer (General Tools & Instruments, Secaucus, NJ, USA) at a depth of 15 cm by inserting the thermometer into the soil and recording the temperature when the reading was stable, after recording the temperature the thermometer was removed.

Above ground biomass was collected 3 WAP and plant heights were measured weekly for 4 WAP as described previously. Hemp trichomes were observed for color changes and when 50% of trichomes were amber colored, hemp was harvested, which occurred on 14 September 2020 and 15 September 2021. Harvest method followed the same procedures as previously described in the herbicide screening experiment. Data for biomass, plant heights, and yield were analyzed using JMP Pro 15 (SAS Institute) using the Fit Model procedure. Replication and replication nested within year were considered random effects [27], while production system was a fixed effect. Means were separated using Tukey’s HSD at α = 0.05 when the model was significant.

3. Results and Discussion

3.1. PRE Herbicide Screening

Hemp response was similar each year, thus data are pooled over site years. Injuries from pretransplant applied herbicides ranged from 6 to 96% (

Table 3. Hemp (Cannabis sativa) ‘Von’ injury, plant height, and fresh weight biomass 2 weeks after planting, and dry flower yield as influenced by pretransplant applied herbicides ^a,b.

Herbicide	Rate (g·ha−1)	Injury		Height		Biomass		Dry Flower
		%
acetochlor	841	6	de	103	a	84	ab	96	a
dithiopyr	280	50	bc	92	ab	98	ab	91	ab
flumioxazin	72	9	de	97	ab	64	abc	103	a
fomesafen	210	10	de	103	a	62	abc	87	ab
halosulfuron	39	55	bc	70	abcd	42	bc	53	bc
isoxaben	631	66	ab	57	cd	55	abc	29	cd
isoxaflutole	112	96	a	14	e	8	c	0	c
metribuzin	210	32	cd	68	bcd	37	bc	74	abc
norflurazon	138	11	de	93	ab	80	ab	104	a
oxyfluorfen	421	25	cde	84	abc	68	ab	87	ab
pendimethalin	1065	10	de	99	ab	87	ab	89	ab
prometryn	1121	33	cd	42	de	35	bc	102	a
S-metolachlor	1068	37	bcd	80	abc	50	abc	94	ab
none	-	0	e	100	a	100	a	100	a
p-value		<0.0001		<0.0001		0.0002		<0.0001

^a Means are separated using Tukey’s HSD at α = 0.05. Means with the same letter within a column are not significantly different. Data are combined over 2020 and 2021. ^b Average plant height, biomass, and dry flower yield per plant for the nontreated control was 14.8 cm, 6.1 g, and 0.69 kg, respectively.

). Maximum injury observed was less than 12% with acetochlor, flumioxazin, fomesafen, pendimethalin, and norflurazon while it ranged from 25 to 37% with oxyfluorfen, S-metolachlor, metribuzin, and prometryn. All other herbicide treatments injured hemp at least 50% compared to the non-treated control. Similar results for S-metolachlor, pendimethalin, flumioxazin and acetochlor were observed in transplanted industrial hemp by Flessner et al. [28]. Previously, visual injury from fomesafen applications to seeded hemp have been reported to be variable with 10 to 11% injury in one site year and greater than 50% in a second site year in each study [9,10,11]. In our study and Flessner et al. [28], the crop was transplanted in a manner that minimized transplant root contact with treated soil, therefore less injury was expected. It is likely that higher injury levels would be observed when using a different planting method, such as a hole punch, where treated soil is pushed into the plant hole and comes into direct contact with roots.

Plant height compared to the non-treated control was reduced 32 to 86% by metribuzin, isoxaflutole, prometryn, and isoxaben, with isoxaflutole resulting in the greatest height reduction (Table 3). Similarly, biomass was reduced 58–92% by metribuzin, isoxaflutole, prometryn, and halosulfuron, and again isoxaflutole resulted in the greatest reduction in biomass. Ortmeier-Clarke et al. [17] reported 100% biomass reduction in seeded industrial hemp treated with isoxaflutole and metribuzin.

Dry flower yield weight was reduced 47 to 100% by isoxaflutole, isoxaben, and halosulfuron. Other herbicide treatments did not significantly reduce yield even though early season injury was severe among several of the treatments. Previous literature has documented that hemp may be able to recover from early season herbicide injury [10]. Additionally, Byrd [9] observed up to 87% injury from PRE herbicides on grain hemp without a difference in yield after harvest. In fresh market vegetable production, injury from herbicides > 10% is considered unacceptable. It stands to reason that excessive injury in floral hemp, which is grown similarly to fresh market produce, would be considered unacceptable as well, even though plants recovered from early season injury with no yield loss in many treatments. Injury acceptability should be kept in mind as further herbicide options for floral hemp production are explored.

Coffman and Gentner [29] hypothesized that cannabinoid content would increase when plants were stressed, so it was of interest to document if herbicide applications influenced cannabinoid content. Plant samples from each plot treated with flumioxazin, fomesafen, acetochlor, pendimethalin, norflurazon, and dithiopyr were selected for cannabinoid testing each year because of their potential for future usage. Laboratory results indicated all treatments including the nontreated control produced total THC levels higher than 0.3% dry weight at harvest, the legal limit for total THC in floral hemp production (

Table 4. Total tetrahydrocannabinol (THC), cannabidiol (CBD), and cannabinoids as a percent dry weight from subsamples collected from hemp (Cannabis sativa) ‘Von’ plants in the pretransplant experiment ^a,b.

Herbicide	THC		CBD	Total Cannabinoids
	%
acetochlor	0.58	abc	12.67	13.05	ab
dithiopyr	0.46	c	11.91	10.36	b
flumioxazin	0.66	ab	11.41	13.11	ab
fomesafen	0.54	bc	11.13	12.13	ab
norflurazon	0.73	a	13.54	14.42	a
pendimethalin	0.56	abc	10.74	12.12	ab
none	0.66	ab	13.51	14.41	a
p-value	0.0009		0.0887	0.0227

^a Means are separated using Tukey’s HSD at α = 0.05. Means with the same letter within a column are not significantly different. Data are combined over 2020 and 2021. ^b Total cannabinoids include neutral and acidic forms of delta-9 and delta-8 tetrahydrocannabinol acid, cannabidiol, cannabigerol, cannabinol, and cannabichromene measured on a dry weight basis.

). Harvest followed recommendations by Darby and Bruce [30] where the crop was monitored and harvested when half of the trichomes observed were amber or milky-colored in the control. Cannabinoids accumulate throughout the season [31], thus, to reduce THC levels and to comply with federal regulations, floral hemp grown for CBD may be harvested earlier than in the present study [2,30]. When compared to the control, no herbicide increased total THC content while dithiopyr showed a reduction of 0.2% (Table 4). CBD content was not influenced by treatment while total cannabinoid content, similar to THC, was lowest for plants treated with dithiopyr (10.36%). Early season injury with dithiopyr of 50% may have delayed crop maturity, resulting in less THC and total cannabinoid accumulation, however, further exploration of early season injury and cannabinoid content is necessary.

In this initial PRE residual applied herbicide screening, isoxaflutole, isoxaben, and halosulfuron were determined not to be viable PRE herbicide options in hemp that is transplanted and grown for flower production due to excessive crop injury. Acetochlor, flumioxazin, fomesafen, norflurazon, and pendimethalin may be candidates for additional research. Other herbicides including dithiopyr, metribuzin, oxyfluorfen, prometryn, and S-metolachlor would need to be evaluated at significantly lower application use rates.

3.2. POST Herbicide Screening

Hemp response to POST herbicide applications was not impacted by year, so data were pooled over each site year. Maximum injury in the POST study was less than 5% with clethodim, pendimethalin, and S-metolachlor while injury was 14% with acetochlor (

Table 5. Hemp (Cannabis sativa) ‘Sunbelt Crush’ injury, plant height, and fresh weight biomass 2 weeks after application, and dry flower yield as influenced by posttransplant applied herbicides ^a,b,c.

Herbicide	Rate (g·ha−1)	Injury		Height		Biomass		Dry Flower
		%
acetochlor	841	14	d	95	a	87	a	93	a
clethodim	136	0	e	115	a	102	a	116	a
halosulfuron	26	83	b	64	b	29	bc	44	bcd
	53	95	a	42	bcd	17	c	20	d
imazethapyr	35	52	c	65	b	42	bc	83	ab
metribuzin	210	91	ab	32	cde	17	c	34	cd
	210 + NIS	94	a	24	def	10	c	35	cd
	420	99	a	4	f	9	c	5	d
	420 + NIS	99	a	3	f	7	c	1	d
pendimethalin	1065	0	e	97	a	35	bc	100	a
prometryn	1121	97	a	8	ef	8	c	13	d
S-metolachlor	1068	4	de	112	a	61	ab	77	abc
trifloxysulfuron	5	92	ab	53	bc	19	c	13	d
none	-	0	e	100	a	100	a	100	a
p-value		<0.0001		<0.0001		<0.0001		<0.0001

^a Means are separated using Tukey’s HSD at α = 0.05. Means with the same letter within a column are not significantly different. Data are combined over 2020 and 2021. ^b Average plant height, biomass, and dry flower yield per plant for the nontreated control were 51.3 cm, 68.5 g, and 0.43 kg, respectively. ^c Nonionic surfactant (NIS) was added at 0.25% v/v.

). These treatments did not influence plant height or dry flower yield when compared to the control and of these herbicides only pendimethalin reduced early season biomass. Injury from both halosulfuron treatments, each metribuzin treatment, imazethapyr, prometryn, and trifloxysulfuron was severe with at least 50% visual injury, a reduction in plant height, and a reduction in biomass when compared to the control. From these herbicide options that caused severe early season damage, hemp dry flower yield was only within 50% of the control for imazethapyr.

Results from the POST herbicide screening study reflect previously reported data. Numerous studies have demonstrated the safety of clethodim or another cyclohexanedione herbicide on grain and fiber hemp [9,17,32]. Sosnoskie and Maloney [33] reported injury greater than 85% following a POST application of prometryn to floral hemp, while Howatt and Mettler [34] documented metribuzin injury of 82% at 14 d to fiber hemp. Grain and fiber hemp biomass reductions of 79 to 81% were reported with field use rates of imazethapyr and trifloxysulfuron [11,17], and halosulfuron injured hemp 60 to 63% during early season [9,10].

Acetochlor, clethodim, pendimethalin, and S-metolachlor were selected for cannabinoid analysis due to their potential for future use in a floral hemp crop (

Table 6. Total tetrahydrocannabinol (THC), cannabidiol (CBD), and cannabinoids as a percentage of dry weight from subsamples collected from hemp (Cannabis sativa) ‘Sunbelt Crush’ plants in the posttransplant study ^a,b.

Herbicide	THC		CBD	Total Cannabinoids
	%
acetochlor	0.52	ab	11.02	11.94
clethodim	0.46	b	10.71	11.41
pendimethalin	0.63	a	12.67	13.8
S-metolachlor	0.53	ab	11.62	12.45
none	0.49	b	11.36	12.4
p-value	0.0108		0.1497	0.071

^a Means are separated using Tukey’s HSD at α = 0.05. Means with the same letter within a column are not significantly different. Data are combined over 2020 and 2021. ^b Total cannabinoids include neutral and acidic forms of delta-9 and delta-8 tetrahydrocannabinol acid, cannabidiol, cannabigerol, cannabinol, and cannabichromene measured on a dry weight basis.

). Neither CBD nor total cannabinoid production was influenced by herbicide treatment. Similar to the PRE study, all samples collected produced THC levels at harvest higher than those currently legally allowed in the USA. This result is expected since both studies were harvested concurrently. Plants treated with pendimethalin had total THC concentrations of 0.63% compared to the 0.49% in the control, while all other treatments had levels similar to the control. Though there was no significant visual injury or reduction in plant height growth detected with pendimethalin, plant biomass was reduced by 65%, indicating plants were affected by this treatment, which may have influenced THC levels. Prior research suggests THC accumulation may be affected by genetics to a greater degree than the environment; however, abiotic stressors (such as herbicide injury) were not evaluated [35]. The interaction between floral hemp variety and herbicide injury on cannabinoid content should be explored in future experiments.

Injury data suggest that halosulfuron, metribuzin, trifloxysulfuron, and prometryn are not viable for broadcast POST applications to transplanted hemp. Imazethapyr is likely not suitable for use due to its effects on crop growth and variable yield response unless application use rates are reduced dramatically [9,10]. S-metolachlor, acetochlor, pendimethalin, and clethodim have potential for use in floral hemp production; however, further research is needed to evaluate these options.

3.3. Transplant Method Comparison

There was a significant interaction between planting method and herbicide for plant injury, height, and biomass in this experiment (

Table 7. Hemp (Cannabis sativa)‘Sunbelt Crush’ injury, plant height, and biomass 4 weeks after planting as influenced by pretransplant herbicide and planting method for hemp (Cannabis sativa) grown in Georgia, USA, in 2021 ^a,b,c.

Transplanting Method	Herbicide	Injury		Height		Biomass
		%
Mechanical	halosulfuron	11	bc	84	ab	57	b
	metribuzin	32	b	65	bc	47	bc
	none	0	c	100	a	100	a
Hole punch	halosulfuron	67	a	51	c	22	cd
	metribuzin	84	a	51	c	18	d
	none	0	c	100	a	100	a
p-value		<0.0001		0.0174		0.0045

^a Means are separated using Tukey’s HSD at α = 0.05. Means with the same letter within a column are not significantly different. ^b Average plant height and biomass for the mechanical transplant method were 25.5 cm and 18.8 g, and 35.6 cm and 67.8 g for the hole punch method. ^c Mechanical transplanting was accomplished using a mechanical transplanter (Mechanical Transplanter Co., Holland, MI, USA) that opened a furrow, the plant was dropped in the furrow, and the furrow was closed with two wheels pressing soil around the plant root. The hole punch transplanting method was made using a mechanical transplant wheel (Kennco Manufacturing, Ruskin, FL, USA) which pushed the soil down, making a transplant hole 4 cm deep and wide. Hemp was transplanted by hand after holes were punched.

). Injury from halosulfuron and metribuzin when using the mechanical transplant method was 52 to 56% less than that observed when punching a hole and transplanting by hand. Differences in crop response to pretransplant herbicide applications have been previously documented. Culpepper et al. [36] observed 2 to 16% more injury from oxyfluorfen applied pretransplant at 140 to 841 g·ha⁻¹ when broccoli (Brassica oleracea var. italica) and collard (Brassica oleracea var. viridis) were planted using the hole punch transplant method compared to using a mechanical transplanter. In the present study and the Culpepper et al. [30] experiment, greater crop safety was observed with the mechanical planting process because transplants are placed below the treated soil surface, thereby lessening contact of the residual herbicide with the transplant root ball. The traditional hole punch approach, by pressing treated soil downward, places the herbicide in the plant hole, thereby surrounding the transplant root ball. This is a critically important observation as it may not only improve the opportunity to obtain more residual herbicides in transplanted hemp but also for specialty crops where herbicide options are limited, and the hand planting approach remains the standard. Observations for both plant height and biomass generally followed those trends observed with hemp injury (Table 7). Results from this experiment indicate that mechanical transplanting can lessen injury from herbicides applied preplant in hemp when compared to the standard hole punch method.

3.4. Plastic Mulch and Bareground Production System Comparison

Plant height at 2 and 4 WAP was not influenced by the use of plastic mulch for production (

Table 8. Hemp (Cannabis sativa) ‘Von’ height, fresh weight biomass, and dry flower yield in a plastic mulch and bareground production system grown in Georgia, USA, in 2020 and 2021 ^a,b.

	Height 2 WAP	Height 4 WAP	Biomass		Dry Flower
	cm		g·plant−1
Bareground	12.7	40.4	41.3	a	596.9
Plastic	11.5	31.4	19.1	b	665.9
p-value	0.3051	0.1472	0.0059		0.6562

^a Means are separated using Tukey’s HSD at α = 0.05. Means followed by different letters are significantly different. ^b Plastic was a black low-density polyethylene plastic mulch covering a raised bed that was 81 cm wide and 20 cm tall.

). Biomass collected at 2 WAP indicated that the bareground system had twice the biomass of plants growing on plastic mulch with a black surface facing upward. Soil temperatures measured using a digital pocket thermometer at a 15-cm depth in the non-plastic mulched system at planting ranged from 31.2 to 33.2 °C compared to 34.9 to 35.2 °C in the plastic mulched system, while the average minimum and maximum air temperatures at the research sites for the first 3 wk after planting were 20.2 to 21.4 °C and 30.8 to 31.1 °C, respectively [37]. Nelson [38] reported above-ground fresh weight of hemp grown in a 30 °C/30 °C air/soil temperature regime was 28% to 40% less than plants grown in a 15 °C/15 °C and 15 °C/30 °C air/soil temperature regime, respectively. Further, as hemp photosynthesis and respiration decrease once temperatures reach 30 °C or greater, and warm air and soil temperatures appear to decrease hemp biomass, it is likely that the reduction in biomass observed for hemp growing in the plastic mulch system was the result of elevated soil temperatures under the plastic.

Even though production system influenced biomass early in the season, hemp plants had similar dry flower yields in plastic mulched and bareground systems (Table 8). Similar results have been reported by Grandon et al. [39], where heights and widths of floral hemp grown on bareground were greater than those of plants grown in plastic mulch systems, but floral yield was similar among systems. These results indicate hemp can be grown in a bareground production system or using plastic mulch without negatively impacting yield with either system. However, future research may address the influence of plastic mulch surface color on the production of hemp. Vegetable growers planting during summer months in Georgia, USA, will only utilize plastic mulch with a white or silver surface to mitigate excessive surface temperatures [40]. Thus, additional efforts are needed to determine the influence of plastic mulch color in relation to soil and air temperatures when growing hemp.

4. Conclusions

These studies have provided the industry with new insights on herbicide use and weed control options for floral hemp production. Herbicide screening studies suggest there are numerous PRE and POST options that may be effective tools to assist hemp producers, especially when using a mechanical transplanting process. Until herbicide options become more widely available, plastic mulch is a viable option to minimize weed interference with hemp, but plastic mulch surface color should be considered before adopting this production system.