Weed Technology
www.cambridge.org/wet
Optimizing weed control using dicamba and
glufosinate in eligible crop systems
Grant L. Priess1, Michael P. Popp2 , Jason K. Norsworthy3,
Research Article
Andy Mauromoustakos4, Trenton L. Roberts5 and Thomas R. Butts6
1
Cite this article: Priess GL, Popp MP,
Norsworthy JK, Mauromoustakos A,
Roberts TL, Butts TR (2022) Optimizing weed
control using dicamba and glufosinate in
eligible crop systems. Weed Technol. 36:
468–480. doi: 10.1017/wet.2022.44
Received: 28 February 2022
Revised: 9 May 2022
Accepted: 31 May 2022
First published online: 10 June 2022
Associate Editor:
Amit Jhala, University of Nebraska, Lincoln
Keywords:
agricultural pesticides; best management
practices; economics; production agriculture;
weed management
Nomenclature:
dicamba; glufosinate; Palmer amaranth;
Amaranthus palmeri (S.) Wats.; cotton;
Gossypium hirsutum L.; soybean; Glycine max
(L.) Merr.
Author for Correspondence:
Michael P. Popp, Professor, University of
Arkansas, Department of Agricultural
Economics and Agribusiness, 217 Agriculture
Building, Fayetteville, AR, 72701
Email: mpopp@uark.edu
Doctoral Academy Fellow, University of Arkansas, Crop Soil and Environmental Sciences, Fayetteville, AR, USA;
Professor, University of Arkansas, Agricultural Economics and Agribusiness, Fayetteville, AR, USA;
3
Distinguished Professor, University of Arkansas, Crop Soil and Environmental Sciences, Fayetteville, AR, USA;
4
Professor, University of Arkansas, Agriculture Statistics Lab, Fayetteville, AR, USA; 5Associate Professor,
University of Arkansas, Crop Soil and Environmental Sciences, Fayetteville, AR, USA and 6Assistant Professor,
University of Arkansas Cooperative Extension Service, Lonoke, AR, USA
2
Abstract
A field experiment was conducted in 2019 and 2020 that included six site-years and four
locations in Arkansas to determine the optimal sequence and timing of dicamba and glufosinate
applications when applied alone, sequentially, or in combination to control Palmer amaranth
by size: labeled (<10 cm height) and non-labeled (13 to 25 cm height). Single applications of
dicamba, glufosinate, and dicamba plus glufosinate (not labeled) resulted in less than 80%
Palmer amaranth control, regardless of weed size. The mixture of dicamba plus glufosinate
was antagonistic for Palmer amaranth control and percent mortality. Sequential applications,
averaged over all time intervals and herbicides, improved the percentage of Palmer amaranth
control 11 to 17 percentage points over a single application, regardless of weed size at
application 28 d after final application (DAFA). Palmer amaranth control with glufosinate
followed by (fb) glufosinate and dicamba fb dicamba, pending weed size, were optimized at
intervals of 7 d, and 14 to 21 d, respectively. Because single site of action (SOA) postemergence
herbicide systems increase the likelihood of the development of resistant biotypes and are not a
best management practice (BMP) in that regard; sequential applications involving both
dicamba and glufosinate were more effective. Furthermore, the sequence of application
mattered with a preference for applying dicamba first. Dicamba fb glufosinate at a 14-d interval
was profit-maximizing and the only herbicide treatment that resulted in 100% weed control
when size was <10 cm. For larger weed sizes, economic analysis revealed that dicamba fb
dicamba performed better than dicamba fb glufosinate when no penalty was assigned for using
a single SOA. This resulted in greater yield loss risk and soil weed seed bank in comparison to
timelier weed control with the smaller weed size. Hence, timely weed control and two SOAs to
control Palmer amaranth are recommended as BMPs that reduce producer risk.
Introduction
© The Author(s), 2022. Published by Cambridge
University Press on behalf of the Weed Science
Society of America. This is an Open Access
article, distributed under the terms of the
Creative Commons Attribution licence (http://
creativecommons.org/licenses/by/4.0/), which
permits unrestricted re-use, distribution and
reproduction, provided the original article is
properly cited.
Palmer amaranth has evolved resistance to seven sites of action (SOAs) in the United States
(Heap 2020). The perpetuating evolution of herbicide resistance in Palmer amaranth has
increased pressure on the few remaining preemergence and postemergence herbicide options that
are commonly used by U.S. soybean and cotton producers. The innovation of genetically modified
(GM) herbicide-tolerant soybean and cotton has enabled producers to apply over-the-top postemergence herbicides to combat evolving Palmer amaranth populations. However, a new herbicide
site of action (SOA) has not been developed in almost 30 yr. Therefore, proper management of
the few remaining effective SOA is imperative (Duke 2012; Norsworthy et al. 2012).
Monsanto (which merged with Bayer Crop Science in 2019) commercially launched
glyphosate/dicamba/glufosinate–resistant cotton (XtendFlex®) in 2015, with a subsequent commercialization for use in soybean, in 2021. Specific to this work, the incorporation of multiple
GM traits allows the use of postemergence applications of dicamba, glufosinate, and glyphosate.
Applying two effective SOAs in mixture mitigates the likelihood of target-site herbicide resistance more than applying the herbicides sequentially (Bagavathiannan et al. 2013, 2014; Diggle
et al. 2003); however, dicamba-containing products such as XtendiMax® plus Vaporgrip®
(Monsanto Corporation, St. Louis, MO) and Engenia® (BASF Corporation, Research Triangle
Park, NC) cannot be mixed with glufosinate due to limitations on the Environmental Protection
Agency–approved product labels (Anonymous 2020a, 2020b). Therefore, using the XtendFlex®
technology, dicamba and glufosinate, have to be applied sequentially.
Many factors within sequential herbicide programs can affect their weed control efficacy,
including but not limited to the interval between sequential applications (Meyer et al. 2019),
sequence of herbicides applied (Burke et al. 2005), weed size (Lee and Oliver 1982; Steckel
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
469
Weed Technology
et al. 1997; Wilson 2005), environmental conditions (Ahrens 1994;
Anderson et al. 1993; Coetzer et al. 2001), application technique or
nozzle selection (Etheridge et al. 2001; McKinlay et al. 1974; Meyer
et al. 2015), and cost. To optimize herbicides when using the
XtendFlex® technology, a clear understanding is needed of how
the aforementioned factors can influence the overall weed control
efficacy of sequential applications of dicamba and glufosinate.
Additionally, identifying which weed control program provides
the greatest relative net benefit to the producer from both cost
and weed control efficacy perspectives would be beneficial.
It has been previously shown that the order in which sequential
postemergence herbicides are applied can influence weed control
(Burke et al. 2005). For example, when a contact herbicide such as
glufosinate is applied, a decrease in sequential herbicide absorption
and translocation has been observed. Reductions in sequential
herbicide absorption and translocation were attributed to the rapid
necrosis of plant tissue following the glufosinate application (Burke
et al. 2005). The reduction in absorption and translocation
observed following a glufosinate application may suggest that an
application of glufosinate preceding an application of dicamba
will not optimize postemergence weed control in the XtendFlex®
technology.
Applications of synthetic auxin herbicides such as dicamba can
have adverse effects on sequentially applied herbicides (Priess
et al. 2019). Following an auxin herbicide application, sensitive
plants have been observed to display abnormalities such as epinasty, leaf abscission, and abnormal elongation of aerial structures
(Grossman 2000). The plant symptomology that occurs after an
auxin herbicide application may reduce the overall leaf surface area
of subsequently treated sensitive broadleaf weeds. However,
impacts from the reduction of leaf surface area on the efficacy
of sequentially applied herbicides have not been quantified. In
addition, application of an auxin herbicide causes an upregulation
of detoxifying enzymes (glutathione transferase, cytochrome
P450s), which can affect the metabolism of the applied herbicide
as well as subsequently applied pesticides (Cummins et al. 1999;
Raghavan et al. 2005). Even though auxin herbicides have the
potential to impact efficacy of sequentially applied contact herbicides, an increase in Palmer amaranth efficacy was observed when
2,4-DB was applied 7 d prior to lactofen or acifluorfen, when compared to sequential applications of 2,4-DB (Chahal et al. 2011).
To mitigate the probability of Palmer amaranth evolving resistance to either dicamba or glufosinate when the XtendFlex®
technology is used, timing and order of sequential herbicide applications of the two SOAs need to be optimized to maximize weed
control. Therefore, the objectives of this research were to 1) assess
weed control of single applications of stand-alone dicamba and
glufosinate; 2) determine whether the unlabeled mixture of
dicamba plus glufosinate was synergistic, additive, or antagonistic;
3) evaluate the improvement in weed control over and above single
applications with sequential applications of dicamba followed by
(fb) dicamba, glufosinate fb glufosinate, dicamba fb glufosinate,
and glufosinate fb dicamba, and identify the sequence of herbicide
order that optimized control; 4) investigate whether time intervals
of 0.2, 3, 7, 14, and 21 d between herbicide applications optimized
control of sequential applications of dicamba fb glufosinate and
glufosinate fb dicamba (Table 1); and finally, 5) quantify economic
and production risk by simultaneously comparing not only weed
control effectiveness but also herbicide cost (including application
charges) across weed control options using Monte Carlo simulation to quantify the relative net benefit to producers in dollar terms.
To accomplish this, we estimated cumulative probability density
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Table 1. Experimental treatments.a
Herbicide
Nontreated
Dicamba
Glufosinate
Dicamba þ
glufosinate
Dicamba fb
dicamba
Glufosinate fb
glufosinate
Dicamba fb
glufosinate
Glufosinate fb
dicamba
Rate
–
560 g ae ha−1
656 g ai ha−1
560 g ae ha−1 þ 656
g ai ha−1
560 g ae ha−1 fb
560 g ae ha−1
656 g ai ha−1 fb 656
g ai ha−1
560 g ae ha−1 fb
656 g ai ha−1
656 g ai ha−1 fb 560
g ae ha−1
Time interval
between sequential
applications
Cost of
herbicide
treatmentb
–
–
–
–
US$ ha−1
0
56.03
51.31
85.36
7, 14, and 21 d
112.06
7, 14, and 21 d
102.62
0.2 (6 h), 3, 7, 14,
and 21 d
0.2 (6 h), 3, 7, 14,
and 21 d
107.34
107.34
a
Abbreviation: fb, followed by.
Cost of herbicide treatment includes a custom application fee of US$21.98 ha−1
application−1.
b
functions (CDFs) to visually present the extent to which the range
of outcomes with a particular weed control option compares with
the range of outcomes of alternative weed management options.
Material and Methods
Field Trials
Field experiments were conducted in 2019 and 2020. In 2019,
experiments were conducted in Keiser, AR (35.675128°N,
90.07844°W), near Crawfordsville, AR (35.228428°N, 90.336762°W),
and near Marianna, AR (34.725784°N, 90.735788°W). In 2020,
the experiment was conducted in Fayetteville, AR (36.092002°
N, 94.187002°W), and at the same locations near Keiser and
Marianna. The experiment was designed as a single-factor randomized complete block with four replications (Table 1). Field
location, Palmer amaranth size at the initial application, and soil
information at each site are displayed in Table 2.
Treatments were initiated without an attempt to grow cotton
or soybean on ground with native Palmer amaranth populations
at all locations besides Fayetteville, AR, in 2020, where Palmer
amaranth from Crittenden County, AR, was overseeded. Crop
and weed interactions can influence the ability of herbicide-injured
weeds to recover and produce seed (Evans et al. 2003; Jha and
Norsworthy 2008). As crop density increases, weed biomass and
the ability of weed interference to effect crop yield decreases
(Tollenaar et al. 1994). Because these experiments were conducted
without a crop present, Palmer amaranth had an improved
opportunity to regrow. The presence of a crop would likely affect
the ability of herbicide treatments to control weeds as evaluated
elsewhere (Tollenaar et al. 1994). However, Palmer amaranth
has been observed to partially acclimate to crop shading by increasing leaf area and total leaf chlorophyll concentrations (Jha et al.
2008). To evaluate the effectiveness of the herbicide treatments
without crop competition the experiments were conducted without a crop present.
Plot size at all locations was 1.93 m wide and 6 m long. Prior to
the first herbicide application, two 0.25- to 0.5-m2 quadrants where
the size of the quadrants depended on Palmer amaranth density,
were established in each plot and plants were counted for a
470
density assessment. After initial density assessments were
recorded, either S-metolachlor or dimethenamid-P was applied
over the entire test area at a rate of 1,606 g ai ha−1 or
736 g ai ha−1, respectively, to limit further Palmer amaranth
emergence. Average Palmer amaranth height was also recorded
prior to the initial herbicide application.
Herbicides were applied with hand-held CO2-pressurized
sprayers calibrated to deliver 140 L ha−1 of spray solution at
6.4 kph. Dicamba was applied using TTI 110015-VP (TeeJet,
Springfield, IL) nozzles to ensure an ultra-course droplet using
small-plot spraying equipment (Anonymous 2020a, 2020b).
Glufosinate was applied with AIXR 110015-VP (TeeJet) nozzles.
The mixture of dicamba þ glufosinate was applied with TTI
110015-VP nozzles.
Following herbicide applications, Palmer amaranth control was
visually rated and plants with live tissue were counted in the established 0.25- to 0.5-m2 quadrants 28 d after final application
(DAFA) in each treatment. Estimates of Palmer amaranth control
were rated on a scale of 0 to 100, with 0 being no control and 100
being complete Palmer amaranth death 14 and 28 DAFA. Initial
and final counts were used to calculate a quantitative mortality percentage for each treatment.
Economic Analysis
Pricing for dicamba, the required volatility reducing agents
(VRAs), and glufosinate products labeled for use over-the-top of
XtendFlex® crops were obtained from Helena Agri-Enterprises,
Nutrien Ag Solutions, and Simplot locations in the mid-southern
United States. Cost per liter, as averaged across the different retailers in spring 2021, were converted to cost per hectare using labeled
use rates of each product. Several VRAs were priced including
Sentris (BASF, Lundwigshafen, Germany), VaporGrip Xtra Agent
(Bayer CropScience, St. Louis, MO), and Delta Lock (Loveland
Products, Loveland, CO), and similarly converted to cost per
hectare. Use rates of VRAs were calculated using a spray volume
of 140 L ha−1, which is the minimum amount required by label
(Anonymous 2020a, 2020b). Rebates affecting the cost of the herbicide were not included due to intricacies in the various programs
and difficulty standardizing rebates across products and locations.
Given changing bio-tech trait availability no attempt was made
to calculate longer-term average cost differences across herbicide, VRAs, and deposition aid and drift management adjuvants
(DRAs).
In addition to pesticide and adjuvant input costs, the cost of
application also contributes to the overall expense of herbicide
applications. To standardize treatments a custom application fee
of US$21.98 ha−1 was added to each herbicide application, based
on the average statewide cost of custom ground herbicide applications in Texas (Klose et al. 2019). A total cost of herbicide expense
was calculated for each treatment in Table 1. Other factors that
could affect the cost of these postemergence herbicide applications are the ability to mix residual herbicides to embed these
applications in timely full-season herbicide programs to limit
Palmer amaranth emergence; however, the use of residual herbicides is outside of the scope of this research. The over-the-top
applications of dimethenamid-P and S-metolachlor used to prevent new flushes of weeds in this research were not included in
the cost calculations.
To quantify both weed control effectiveness and treatment cost
implications we calculated relative net monetary benefits across
all treatments and further simulate ranges of likely outcomes
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Priess et al.: Timing Dicamba and Glufosinate
associated with different weed control options using Monte
Carlo simulation with @Risk v7.5 (Palisade Corporation,
Raleigh, NC). Triangular truncated probability density functions
(distributions) were fitted to Palmer amaranth mortality rates from
experimental data for each of the treatments and to the initial
Palmer amaranth density in the field. Because experimental trials
were conducted under high initial Palmer amaranth densities, the
latter distribution was scaled and truncated at 10,764 plants ha−1
based on the very high density found in the Palmer amaranth management software (Lindsay et al. 2017). Using Monte Carlo simulation with 10,000 iterations, Table 1 in the Supplemental Material
lists the parameters describing the distributions as sampled from
the fitted distributions for treatments where herbicide applications were made to <10-cm-tall plants and to plants that were
13 to 25 cm tall. Triangular distributions, truncated between
software-selected minima greater than or equal to 0%, and
maxima of software-selected maxima less than or equal to
100%, exhibited superior fit characteristics in comparison to
beta, normal, exponential, gamma, Weibull, Pareto, Pearson,
Inverse Gauss, Laplace, Levy, logistic, loglogistic, and log normal distributions using the Kolmogorov-Smirnov statistic
reported by @RISK for a majority of the fit distribution comparisons for each treatment alternative.
For each of the 10,000 simulation runs, the initial Palmer amaranth density (PD) was drawn randomly from its fitted distribution
and then also a mortality rate (MRi) for each treatment alternative, i, from respective fitted distributions. This was done to calculate the estimated number of Palmer amaranth plants
remaining (PRi) after spraying individual herbicide treatments
on the two weed sizes tested. Because trial data were collected in
the absence of a crop grown, we estimate yield loss (YLi) on the
basis of PR in percentage terms for soybean as follows (Bensch
et al. 2003):
PRi
YLi ¼ 104:6
1þ
10000
104:6
!
PRi
10000
86:9
!
=100
[1]
The yield loss percentages for each treatment alternative were then
multiplied by a yield-based revenue expectation per hectare using a
soybean price of US$0.37 kg−1 and an irrigated soybean yield of
4,370 kg ha−1 to reflect long-term average dollar loss expectations
from Palmer amaranth for a soybean producer in the study region.
Estimated dollar losses (DLi = YLi multiplied by the revenue
potential of soybean at US$1,606 ha−1) as a result of varying levels
of control of Palmer amaranth across the k treatment alternatives
were compared to obtain an estimate of the relative benefit (RBi) a
producer would obtain by choosing a particular herbicide treatment alternative i over the herbicide treatment with the largest
dollar loss across the k alternatives:
RBi ¼ max DLi
k
DLi
[2]
The weed control option with the least dollar loss is expected to be
preferred by the producer and would lead to the largest RB.
However, the producer spends different dollar amounts on weed
control (WCi) across alternatives. As such, relative cost (RCi) is
the difference between the least-expensive weed control option
across the k alternatives and the chosen alternative i and reflects
added cost for more expensive treatment alternatives in terms of
herbicide cost itself as well as charges for application:
471
Weed Technology
Table 2. Experiment data.
Palmer amaranth
Year
Nearest town
Size at initial
application
Density
Average
(range)
Average (range)
Trial site
2019
Crawfordville,
AR
Production field
2019
Keiser, AR
2019
Marianna, AR
Northeast Research and
Extension Center
Lon Mann Cotton Research
Station
2020
Fayetteville, AR
2020
Keiser, AR
2020
Marianna, AR
University of Arkansas–
Agricultural Research and
Extension Center
Northeast Research and
Extension Center
Lon Mann Cotton Research
Station
RCi ¼ WCi
plants ha−1
2,400,000 (480,000–
5,400,000)
13 (0.5–15.4)
840,000 (120,000–1,400,000)
13 (0.5–13.5)
800,000 (200,000–1,320,000)
20 (1.5–25.4)
760,000 (16,000–2,240,00)
7.6 (0.5–8.0)
1,040,000 (240,000–
1,920,000)
1,280,000 (320,000–
2,880,000)
20 (2–25.4)
min WCi
k
[3]
Finally, the relative net benefit of a particular weed control
method is a function of both RB and RC and summarizes the dollar
impact from a revenue and cost side. It is calculated as follows:
RNBi ¼ RBi
Soil information
cm
7.6 (0.5–8.2)
RCi
[4]
Importantly, no cost is assessed to Palmer amaranth evolving
herbicide resistance to alternatives that use the same SOA as would
be the case for treatments using the same herbicide twice in the
same growing season, nor is a value assigned for weed seed addition to the soil seedbank across treatment alternatives as a function
of PR. Hence, relative net benefit (RNB) values are likely
conservative in the sense that treatments with poor weed control
have further costs.
To summarize, a particular iteration run in the Monte Carlo
simulation would depend on that particular iteration run’s initial
PD, which is the same across all treatment alternatives, and the
randomly chosen mortality rate independently chosen form each
treatment’s fitted mortality rate distribution. Calculating both a
positive or negative RNB is possible as some treatments may be
low cost but also lead to high yield and thereby revenue or dollar
losses given poor control, or they could be effective weed control
options that could be either relatively cheap or expensive. Hence,
the treatment alternative with the highest RNB is superior to the
other treatment alternatives. Such RNBi were iteratively calculated
10,000 times to report both an average RNB and also to estimate
CDFs across 10,000 iterations to reflect differences in riskiness as
well as relative profitability. Treatment alternatives with steeper
curves (e.g., less risk) and those with greater mean values (e.g.,
more RNB) would be preferred by the producer.
Because treatment effectiveness in terms of mortality rates is likely
related across treatments, partial correlation coefficients were used
(Supplemental Material, Tables 2 and 3), to develop CDFs with those
correlations imposed. Treatment comparisons both with and without
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Dundee silt loam (Fine-silty, mixed, active,
thermic Typic Endoaqualfs) with 11% sand, 77%
silt, 12% clay, and 1.95% organic matter, pH 5.5
Sharkey silty clay (Very-fine, smectitic, thermic
Chromic Epiaquerts)
Convent silt loam (Coarse-silty, mixed,
superactive, thermic Fluvaquentic Endoaquepts)
with 9% sand, 80% silt, 11% clay, and 1.8%
organic matter, pH 6.3
Leaf silt loam soil (Fine, mixed, active, thermic
Typic, Albaqualts) with 34% sand, 53% silt, 13%
clay, and 1.5% organic matter, pH 6.2
Sharkey silty clay (Very-fine, smectitic, thermic
Chromic Epiaquerts)
Convent silt loam (Coarse-silty, mixed,
superactive, thermic Fluvaquentic Endoaquepts)
with 9% sand, 80% silt, 11% clay, and 1.8%
organic matter, pH 6.3
those correlations employed were performed to assess whether correlations across treatments made a difference in terms of ranking weed
control management alternatives.
Data Analysis
Data were analyzed by Palmer amaranth size (<10 cm and 13 to 25
cm). A single factor ANOVA was used to assess herbicide treatments in SAS software (version 9.4; SAS Institute Inc., Cary,
NC) using the GLIMMIX procedure. A beta error distribution
was assumed for Palmer amaranth control 14 and 28 DAFA
(McDonald and Xu 1995). Site-years were analyzed by weed size
at the initial application and site locations, and runs were added
as random variables. Experiments conducted at the
Crawfordsville, AR, location in 2019 and Keiser, AR, location in
2020 were considered labeled applications based on the average
Palmer amaranth size at the initial application (Table 2). The other
four experimental runs were pooled because Palmer amaranth
averaged over 10 cm in height at the time of the initial application
(Table 2). Means were separated using Tukey’s highly significant
difference test (α = 0.05). Least significant mean contrasts were
conducted for comparison of single applications versus sequential
applications, dicamba fb dicamba versus glufosinate fb glufosinate,
dicamba fb glufosinate versus glufosinate fb dicamba, dicamba fb
dicamba versus dicamba fb glufosinate, and glufosinate fb glufosinate versus dicamba fb glufosinate (α = 0.05).
To evaluate the interaction of the unlabeled mixtures of
dicamba and glufosinate (Anonymous 2020a, 2020b), Colby’s
method (Colby 1967) was used to assess the type of interaction
occurring when two herbicides are applied in mixture. Colby’s
method requires the calculation of an expected value as follows:
E ¼ ðX þ YÞ
ðXYÞ=100
[5]
where E is the expected value of the herbicide mixture, and X and
Y are values of herbicides when applied alone. A two-sided t-test
was performed comparing the expected value calculated from
472
Priess et al.: Timing Dicamba and Glufosinate
Table 3. Percent control and mortality when <10-cm-tall Palmer amaranth was treated with single and sequential applications of dicamba and glufosinate, averaged
over two site-years.a,b,c
Herbicide
Dicamba
Glufosinate
Dicamba þ glufosinate
Dicamba fb dicamba
Dicamba fb dicamba
Dicamba fb dicamba
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
P-value
Interval between
applications
days
NA
NA
NA
7
14
21
7
14
21
0.2
3
7
14
21
0.2
3
7
14
21
Palmer amaranth control
14 DAFA
28 DAFA
Palmer amaranth mortality
28 DAFA
——————————————————%———————————————
80 EF
74 IJ
92 BCD
76 FGH
65 K
85 E
78 FG
76 HIJ
85 E
82 DEF
86 DEFG
98 ABC
78 FG
97 AB
94 ABC
78 FG
97 AB
98 AB
92 AB
94 ABC
98 AB
83 CDEF
78 GHIJ
92 CD
61 I
72 JK
88 DE
88 BCD
81 FGHI
94 ABCD
95 AB
94 ABC
97 ABC
98 A
94 ABC
95 ABC
96 AB
100 A
100 A
72 GH
95 ABC
98 ABC
89 BCD
90 BCDE
93 BCD
91 ABC
93 ABCD
97 ABC
88 BCDE
83 EFGH
95 ABC
77 FGH
91 ABCD
95 ABC
69 H
87 CDEF
93 BCD
<0.0001
<0.0001
<0.0001
a
Palmer amaranth control and mortality is expressed as a percent of the nontreated.
Abbreviations: DAFA, days after final application; fb, followed by; NA, not applicable.
Means followed by the same letter within a column are not statistically different according to Tukey’s honestly significant difference test (α = 0.05).
b
c
Colby’s equation and the observed values of the mixture (α = 0.05).
If the expected value of the herbicide mixture was statistically
greater than the observed value the mixture was considered
antagonistic. If no difference was found between the observed
and expected value the mixture was considered additive, and if
the observed value was greater than the expected value the mixture
was considered synergistic. The expected value calculated in this
experiment may be considered inflated because glufosinate
alone treatments were applied with an Air Induction
Extended Range (AIXR) nozzle and the mixture of the two herbicides was applied with a Turbo Teejet Induction (TTI) nozzle
to abide by nozzle regulations of dicamba labels, even though
the mixture of dicamba plus glufosinate is not federally labeled
(Anonymous 2020a, 2020b).
Results and Discussion
Site-years included in the analysis for labeled Palmer amaranth size
at the time of the initial application were Crawfordsville, AR, 2019
and Keiser, AR, 2020. For Palmer amaranth control 14 and 28
DAFA and percent mortality the main effect of herbicide treatment
was significant (Table 3). Site-years included in the larger-thanlabeled Palmer amaranth size at the time of the initial application
were Keiser 2019, Marianna 2019, Fayetteville 2020, and Marianna
2020. For Palmer amaranth control 14 DAFA, 28 DAFA, and percent mortality the main effect of herbicide treatment was significant (Table 4). As mentioned previously, all experiments were
over-sprayed with either S-metolachlor or dimethenamid-P prior
to first application of treatments; therefore, control ratings and
mortality percentages reflect emerged plants at the time of initial
application.
To justify the interpretation of the data collected an explanation
of why percent control and mortality data taken at 28 DAFA is
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
used preferentially over the 14 DAFA evaluation is needed. The
percentage of Palmer amaranth weed control at 14 DAFA for
sequential dicamba applications at 7-, 14-, and 21-d intervals
ranged from 4 to 19 and 4 to 11 percentage points lower than
the 28 DAFA evaluation on labeled and larger-than-labeled weed
sizes, respectively (Tables 3 and 4). At 14 DAFA, the systemic
nature of sequential applications of dicamba had not reached
maximum Palmer amaranth control; therefore, comparisons of
sequential applications of dicamba to other sequential applications
at 14 DAFA should not be made. The lack of rapid removal of
Palmer amaranth from crops like cotton or soybean, unlike glufosinate, especially at high densities following application, may have
a negative effect on the crop if competition for resources is still
occurring. Furthermore, the presence of weedy vegetation such
as injured Palmer amaranth and its reflected far-red light perceived
by nearby plants are known to alter crop growth (Afifi and
Swantton 2012; Markham and Stolenberg 2009). However, evaluations at 28 DAFA allowed time for the maximized herbicide efficacy to be reached, and captured any regrowth that occurred from
either dicamba or glufosinate (GLP, personal observation). In the
presence of a crop, some of these sequential treatments may perform slightly different than observed here such as extent of Palmer
amaranth regrowth from a dicamba or glufosinate application if
the crop is approaching canopy formation as noted in previous
research (Meyer and Norsworthy 2020).
Single Applications
Single applications of dicamba or glufosinate applied to <10-cmtall Palmer amaranth provided 76% and 65% control and caused
92% and 85% mortality, respectively, at 28 DAFA (Table 3).
Larger-than-labeled Palmer amaranth plants were not controlled
>90% by a single herbicide application. A single application of
473
Weed Technology
Table 4. Percent control and mortality when 13- to 25-cm-tall Palmer amaranth was treated with single and sequential applications of dicamba and glufosinate,
averaged over four site-years.a,b,c
Herbicide
Interval between
applications
Dicamba
Glufosinate
Dicamba þ glufosinate
Dicamba fb dicamba
Dicamba fb dicamba
Dicamba fb dicamba
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
P-value
days
NA
NA
NA
7
14
21
7
14
21
0.2
3
7
14
21
0.2
3
7
14
21
Palmer amaranth control
14 DAFA
28 DAFA
Palmer amaranth mortality
28 DAFA
——————————————————%———————————————
62 EF
65 GH
57 FG
54 F
59 H
49 G
61 EF
59 H
66 EF
81 BC
85 ABC
88 ABC
79 BC
85 ABC
90 A
73 CD
82 BCD
89 AB
81 BC
77 CDE
77 BCDE
78 BC
76 DEF
75 CDE
63 E
76 DEF
66 EF
67 DE
68 FG
71 DEF
77 BC
76 DEF
72 DE
79 BC
69 FG
84 ABCD
92 A
92 A
89 AB
84 AB
87 AB
89 AB
67 DE
65 GH
65 EF
80 BC
79 BCDE
74 DE
78 BC
75 DEF
80 ABCD
75 CD
81 BCD
83 ABCD
54 F
71 EFG
58 FG
<0.0001
<0.0001
<0.0001
a
Palmer amaranth control and mortality are expressed as a percent of the nontreated.
Abbreviations: DAFA, days after final application; fb, followed by; NA, not applicable.
Means followed by the same letter within a column are not statistically different according to Tukey’s honestly significant difference test (α = 0.05).
b
c
glufosinate or dicamba applied to larger-than-labeled weeds controlled Palmer amaranth 59% and 65% and led to 47% and 59%
mortality, respectively (Table 4). Similarly, Merchant et al.
(2013) and Coetzer et al. (2002) observed that a single application
of glufosinate or dicamba did not provide greater than 76% control
of Palmer amaranth at labeled or above-labeled weed sizes.
Mixtures of Dicamba Plus Glufosinate
Norsworthy et al. (2012) noted that the use of multiple SOAs in
mixture will lessen the risk of herbicide resistance due to an
increase in efficacy and a reduction of selection pressure on a single
herbicide. The mixture of dicamba plus glufosinate applied to
Palmer amaranth <10 cm in height provided 76% control and
did not differ from a single application of dicamba but did provide
an increase of 11 percentage points in control when compared to a
single application of glufosinate 28 DAFA. The mixture of dicamba
plus glufosinate to larger-than-labeled Palmer amaranth did not
result in increased control or mortality when compared to dicamba
or glufosinate alone (Table 4). The mixture of dicamba plus glufosinate was antagonistic when compared to the expected value
reducing Palmer amaranth control by 15 and 28 percentage points
and mortality by 5 and 12 percentage points at the labeled and
above-labeled weed sizes 28 DAFA, respectively (Table 5). These
results are similar to those previously observed (Meyer and
Norsworthy, 2019).
Meyer and Norsworthy (2019) deemed the mixture of dicamba
plus glufosinate to be antagonistic on 30-cm-tall weeds, with
reductions in the Palmer amaranth control of 18 percentage points
when compared with the expected value calculated by Colby’s
equation. When compared to the expected value, the poor efficacy
of the mixture of two SOAs is likely attributed to the use of a TTI
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
(ultra-coarse droplet) nozzle and the inverse nature of the systemic
and contact activity of the two herbicides. Glufosinate efficacy is
optimized at a droplet size of 605 μm (extremely coarse; Butts
et al. 2018). The TTI nozzle used in this experiment for postemergence
applications of dicamba produces an ultra-coarse droplet, thus droplet
size is not optimized for glufosinate efficacy (Anonymous 2020a,
2020b; Butts et al. 2018). Contrarily, Merchant et al. (2013) observed
an increase in efficacy when dicamba plus glufosinate was applied to
Palmer amaranth that was 13 to 25 cm in height with a fine to coarse
droplet nozzle. Meyer et al. (2020) also observed a 46 percentage point
reduction in dicamba translocation when dicamba was mixed with
glufosinate compared to dicamba alone, using radio-labeled herbicides. As mentioned previously, dicamba applications can cause an
upregulation of cytochrome P450 and glutathione S-transferase
enzymes, which can enhance herbicide metabolism (Cummins
et al. 1999; Raghavan et al. 2005). All aforementioned factors including nozzle selection, reductions in systemic translocation of dicamba
caused by rapid necrosis from glufosinate, and/or upregulation of
detoxifying enzymes caused by dicamba could be possible reasons
improved control was not observed when the two herbicides were
mixed (Burke et al. 2005; Cummins et al. 1999; Meyer et al. 2020;
Raghavan et al. 2005).
Palmer amaranth control or mortality percentages did not
reach 100% when a single application of dicamba, glufosinate,
or a mixture of the two herbicides was applied, regardless of weed
size (Tables 3 and 4). To mitigate the selection for resistant biotypes and addition of weed seed to the soil seedbank, a zero-tolerance policy should be implemented (Norsworthy et al. 2012,
2016). Therefore, additional measures will be needed to control
Palmer amaranth plants that survive a single application of
either herbicide or mixture, regardless of weed size at the initial
application.
474
Priess et al.: Timing Dicamba and Glufosinate
Table 5. Effect of mixtures of dicamba and glufosinate on Palmer amaranth control and mortality at 14 and 28 d after final application, separated by labeled and
larger-than-labeled weed sizes.a
Palmer amaranth mortalityb
Palmer amaranth control
14 DAFA
Palmer
amaranth
sizec
cm
<10
13 to 25
Herbicide
dicamba
glufosinate
dicamba þ
glufosinate
dicamba
glufosinate
dicamba þ
glufosinate
Observed
value
Expected
value
————%—————
80
76
78
95
62
54
61
83
28 DAFA
P-valued,e
Observed
value
<0.0001*
—————%————
74
65
76
91
65
59
59
<0.0001*
28 DAFA
Expected
value
86
P-valued,e
Observed
value
<0.0001*
—————%————
92
85
85
99
<0.0001*
Expected
value
57
49
66
78
P-valued,e
0.0025*
0.0042*
a
Abbreviation: DAFA, days after final application.
Palmer amaranth mortality is expressed as a percent of the nontreated.
Labeled Palmer amaranth is <10 cm in height, larger-than-labeled Palmer amaranth is 13- to 25-cm in height.
d
An asterisk (*) denotes significant antagonism based on a two-sided t-test between observed and expected values. Expected values are based on Colby’s equation [E = (X þ Y) – (XY)/100].
e
Significant P-values (≤0.05) are indicated by *.
b
c
Table 6. Least significant means contrast conducted on single applications vs sequential applications and differing sequential applications vs differing sequential
applications analyzed by Palmer amaranth size, evaluation timing, and averaged over site-year.a,b
Palmer amaranth less than 10 cm in heightc
Control 14 DAFA
Control 28 DAFA
d
Contrast
Means
P-value
Single application vs sequential application
Dicamba fb dicamba vs glufosinate fb glufosinate
Dicamba fb glufosinate vs glufosinate fb dicamba
Dicamba fb dicamba vs dicamba fb glufosinate
Glufosinate fb glufosinate vs dicamba fb glufosinate
%
78 vs
79 vs
90 vs
79 vs
79 vs
0.0014*
0.7821
<0.0001*
<0.0001*
<0.0001*
83
79
83
90
90
P-valued
Means
72
93
93
93
81
%
vs 83
vs 81
vs 89
vs 93
vs 93
<0.0001*
<0.0001*
0.0441*
0.7638
<0.0001*
Palmer amaranth 13- to 25-cm in heightc
Control 14 DAFA
Control 28 DAFA
d
Contrast
Means
P-value
Single application vs sequential application
Dicamba fb dicamba vs glufosinate fb glufosinate
Dicamba fb glufosinate vs glufosinate fb dicamba
Dicamba fb dicamba vs dicamba fb glufosinate
Glufosinate fb glufosinate vs dicamba fb glufosinate
%
59 vs
78 vs
80 vs
78 vs
74 vs
<0.0001*
0.1935
<0.0001*
0.1090
0.4902
75
74
71
80
80
Means
61
84
78
84
76
%
vs 78
vs 76
vs 74
vs 78
vs 78
P-valued
<0.0001*
0.0014*
0.0491*
0.0042*
0.4710
a
Sequential applications were averaged over time intervals between sequential applications.
Abbreviations: DAFA, d after final application; fb, followed by; vs, versus.
Average Palmer amaranth height at the time of the initial application.
d
Significant P-values (≤0.05) are indicated by *.
b
c
Sequential Applications
Across all time intervals, an increase of 5 to 11 and 16 to 17 percentage points in control occurred when sequential herbicide
applications were made compared with single herbicide applications at 14 and 28 DAFA, respectively, regardless of weed size
(Table 6). Sequential applications of glufosinate were optimized
at a 7-d interval between applications when initially applied at a
labeled weed size. When applied to labeled-sized Palmer amaranth
10>) cm), glufosinate fb glufosinate at the 7-, 14-, and 21-d intervals provided 94%, 78%, and 72% control and 98%, 92%, and 88%
mortality 28 DAFA, respectively (Table 3). Similarly, Meyer and
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Norsworthy (2020) observed that sequential applications of glufosinate at 7- to 10-d intervals optimized annual weed control. On
larger-than-labeled Palmer amaranth sizes (13 to 25 cm), weed
control and mortality among timing intervals of sequential glufosinate applications did not differ. Control and mortality of largerthan-labeled Palmer amaranth plants following sequential glufosinate applications ranged from 75% to 76% and 66% to 77% at 28
DAFA, respectively (Table 4). Likewise, Meyer and Norsworthy
(2020) observed 84% and 80% Palmer amaranth control when glufosinate at 451 g ai ha−1 was applied sequentially at 7- and 14-d
intervals 3 wk after application.
475
Weed Technology
In terms of visual control ratings of <10-cm-tall Palmer amaranth, sequential applications of dicamba were optimized at the 14and 21-d interval, 28 DAFA (Table 3). A distinctly superior interval between sequential applications of dicamba applied to 13- to
25-cm-tall Palmer amaranth was not observed. Control and mortality of larger-than-labeled Palmer amaranth ranged from 82% to
85% and 88% to 90%, respectively. No differences in Palmer amaranth mortality were observed among sequential applications of
dicamba at 7-, 14-, and 21-d intervals, regardless of weed size
(Table 3 and 4). No sequential application of dicamba or glufosinate resulted in 100% control or 100% mortality of Palmer amaranth (Tables 3 and 4). The risk for selection of resistant
biotypes in the aforementioned single SOA postemergence systems
is high and multiple SOAs should be used to mitigate target-site
based herbicide resistance (Norsworthy et al. 2012). The use of a
single SOA for postemergence control reflects a glufosinate
(LibertyLink™) or Roundup Ready™ Xtend™ system used in an
area where Palmer amaranth has resistance to acetolactate synthase, 5-enolpyruvyl-shikimate-3-phosphate synthase, and protoporphyrinogen oxidase inhibitors. Additional control measures
will have to be taken to mitigate Palmer amaranth seed replenishing the soil seedbank and furthering the selection for resistant
biotypes.
The sequence of sequential herbicide applications influenced
the control level observed in the postemergence two-SOA
XtendFlex® system. Averaged over intervals, dicamba fb glufosinate
provided a 4 percentage point increase in control when compared
with glufosinate fb dicamba sequentially applied to labeled and
larger-than-labeled Palmer amaranth sizes based on a contrast
(Table 6). The increase in control observed when dicamba is
applied prior to glufosinate is likely attributed to adequate absorption and translocation of both herbicides. When a contact herbicide such as glufosinate is applied before a systemic herbicide
such as dicamba a reduction in absorption and translocation of
the systemic herbicide is observed (Sung-Eun et al. 2005).
Future studies should assess the extent to which dicamba absorption and translocation is affected by a prior glufosinate application
at differing time intervals.
When weed sizes were less than 10 cm, >90% Palmer amaranth
control was observed in all sequential herbicide treatments 28
DAFA that included two SOAs, except dicamba fb glufosinate at
the 0.2-d interval, and glufosinate fb dicamba at the 7- and 21-d
intervals (Table 3). Dicamba fb glufosinate at the 0.2-d (6 h) interval was consistently the lowest level of control observed when
dicamba was applied prior to glufosinate 28 DAFA, regardless
of weed size. This observation can likely be attributed to the rapid
reduction in Palmer amaranth groundcover following an auxin
herbicide application (Priess et al. 2019). Following an application
of dicamba with TTI nozzles, a 31 to 36 percentage point reduction
in Palmer amaranth groundcover was observed (Priess et al. 2019).
A dicamba application subsequently reduces Palmer amaranth
groundcover and the surface area of the weed available for intercepting glufosinate. Even though the prior sequence and interval of
the sequential herbicide treatment follows label requirements, an
increase in herbicide cost, application cost, and reductions in
Palmer amaranth weed control does not make dicamba fb glufosinate at a 0.2-d (6 h) interval a sequence likely for adoption by
growers and applicators.
To optimize the use of the two SOAs on labeled weed sizes
dicamba fb glufosinate at the 14-d interval was the only treatment
that provided 100% control and 100% mortality of Palmer
amaranth 28 DAFA (Table 3). On larger-than-labeled Palmer
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
amaranth, dicamba fb glufosinate at the 14-d interval provided
higher control than any other herbicide treatment besides dicamba
fb glufosinate at the 21-d interval at 28 DAFA (Table 6). Findings
from this research lead to the conclusion that dicamba fb glufosinate 14 d later optimizes Palmer amaranth control (Tables 3 and
4). The only time there was zero addition to the Palmer amaranth
soil seedbank was with dicamba fb glufosinate at the 14-d interval
at <10 cm weed size. Importantly, this mitigates further selection
of herbicide resistance by eliminating escapes (Neve et al. 2011).
The optimized use of dicamba fb glufosinate at the 14-d interval
may be explained by a reduction in the interaction between the two
herbicides. Priess et al. (2019) observed that Palmer amaranth
regrowth and an increase in Palmer amaranth groundcover
occurred 14 d after a dicamba application. Therefore, when a
sequential application of dicamba fb glufosinate at a 14-d interval
is made, glufosinate would be applied to actively growing weeds
with increased leaf surface area for herbicide contact when compared with closer time intervals between sequential applications.
In addition, by delaying a subsequent herbicide application by
14 d and targeting actively growing weeds, interactions of herbicide
absorption and translocation may be negligible. Reductions in herbicide absorption and translocation are often attributed to rapid
necrosis of contact herbicides (Meyer et al. 2020). Scarponi et al.
(2005) found that upregulation of herbicide detoxifying enzymes
was maximized 3 d after a metabolic enzyme-inducing seed treatment was applied, and an upregulation of herbicide detoxifying
enzymes was not observed past 7 d. Because auxin herbicides
are known to cause an upregulation of herbicide detoxifying
enzymes, delaying the subsequent herbicide application by 14 d
may alleviate this interaction.
Economic Implications
Dicamba products labeled for use in Xtend® or XtendFlex® crops
averaged US$34.05 ha−1 (including the addition of a necessary
volatility reducing agent) and glufosinate products averaged US
$29.33 ha−1. Excluding technology fees, seed cost, residual herbicides, and herbicide rebate programs, the cost of dicamba and glufosinate are similar but increase with sequential applications given
added application charges regardless of weed size (Tables 7 and 8).
Average mortality rates, as drawn from the fitted distributions
(Supplemental Material Table 1), along with estimated yield loss
and associated relative revenue loss of a hypothetical soybean crop
were calculated using Eq. 1. Average mortality rates (MRs) closely
resemble those reported in Tables 3 and 4 but are slightly different
because they are averages of 10,000 random draws from fitted mortality rate distributions as discussed above. The relative benefit of a
treatment is reported in relation to the most revenue robbing alternative (Eq. 2) using the average Palmer amaranth plant density
(PD) before herbicide application of 5,194 plants ha−1 across all
treatments (Supplemental Material Table 1). Added cost relative
to the most inexpensive treatment reflects RC (Eq. 3) and showcases the single pass with glufosinate to be the cheapest alternative,
whereas sequential applications of dicamba are most expensive
(Tables 7 and 8).
The average relative net benefit (RNB), taking both relative
cost of herbicide treatment and relative benefit from weed control,
calculated at average PD and average MR represents a point estimate on the distribution functions of RNB. Treatment differences
across RNB showcase dicamba fb glufosinate with a 14-d interval
between applications to have the highest RNB of US$54.86 ha−1.
The second-best alternative is a single pass of dicamba at RNB
476
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Table 7. Relative comparisons across treatment alternatives using Monte Carlo simulation and hypothetical soybean revenue loss estimates for simulations of for Palmer amaranth ≤10 cm height.a
Herbicide
a
day
NA
NA
NA
7
14
21
7
14
21
0.2
3
7
14
21
0.2
3
7
14
21
HC
MR
−1
$ ha
$56
$51
$85
$112
$112
$112
$103
$103
$103
$107
$107
$107
$107
$107
$107
$107
$107
$107
$107
%
92.0
85.0
93.6
96.5
89.4
97.2
97.7
88.9
88.0
93.9
95.7
91.7
99.0
95.8
92.0
96.6
95.1
94.9
88.8
RP
−1
plants ha
415
781
333
181
548
147
120
575
622
316
222
430
54
219
417
175
255
263
583
Simulated
average RNB,
no correlationb,d
Simulated
average
RNB,
correlatedc,d
YL
%
4.1
7.5
3.4
1.8
5.4
1.5
1.2
5.6
6.1
3.2
2.3
4.3
0.6
2.2
4.2
1.8
2.6
2.7
5.7
———————————————— $ ha−1 —————————————————————
$66
$53
$5
$48.75
$103.27
$83.99
$120
$0
$0
$0.00
$60.47
$41.13
$54
$66
$34
$32.04
$85.86
$66.41
$30
$90
$61
$29.46
$81.61
$62.69
$86
$34
$61
-$27.22
$29.00
$9.75
$24
$96
$61
$34.97
$86.95
$68.10
$20
$100
$51
$48.70
$100.58
$81.67
$90
$30
$51
-$21.75
$34.37
$15.70
$97
$23
$51
-$28.59
$29.12
$9.47
$51
$69
$56
$12.81
$65.87
$47.33
$36
$84
$56
$27.60
$79.90
$61.14
$69
$51
$56
-$4.85
$49.46
$30.69
$9
$111
$56
$54.86
$106.43
$87.63
$36
$84
$56
$28.05
$80.64
$61.57
$67
$53
$56
-$2.89
$51.54
$32.35
$29
$91
$56
$35.03
$87.08
$68.14
$42
$78
$56
$22.27
$75.02
$56.24
$43
$77
$56
$21.04
$73.87
$54.74
$91
$28
$56
-$27.57
$28.90
$10.03
RB
RC
Abbreviations: fb, followed by; HC, herbicide cost and application charges; MR, Palmer amaranth mortality rate; NA, not applicable; RB, estimated relative benefit; RC, added relative cost; RNB, relative net benefit; RP, expected number of remaining plants
after spray; YL, estimated yield loss.
b
Evaluation was performed with average initial Palmer amaranth plant density and average mortality rates under varying assumptions of correlation among mortality rate probability distributions.
c
See Supplemental Material Table 2 for partial correlation coefficients among weed control options.
d
Bold lettering indicates the top choice (highest RNB = RB – RC) among either single herbicide weed control treatments using different herbicides or their mixture, and again, the most profitable time interval among weed control systems involving two
sequential passes with different combinations of herbicides.
Priess et al.: Timing Dicamba and Glufosinate
Dicamba
Glufosinate
Dicamba þ glufosinate
Dicamba fb dicamba
Dicamba fb dicamba
Dicamba fb dicamba
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Interval
Average
RB – RC at
average
PDd
Estimated
revenue
loss
Weed Technology
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Table 8. Relative comparisons across treatment alternatives using Monte Carlo simulation and hypothetical soybean revenue loss estimates for Palmer amaranth 13 to 25 cm height.a
Herbicide
Dicamba
Glufosinate
Dicamba þ glufosinate
Dicamba fb dicamba
Dicamba fb dicamba
Dicamba fb dicamba
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Glufosinate fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Dicamba fb glufosinate
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Glufosinate fb dicamba
Interval
day
NA
NA
NA
7
14
21
7
14
21
0.2
3
7
14
21
0.2
3
7
14
21
HC
MR
−1
$ ha
$56
$51
$85
$112
$112
$112
$103
$103
$103
$107
$107
$107
$107
$107
$107
$107
$107
$107
$107
%
55.7
39.8
64.7
85.9
89.0
89.7
79.0
64.2
63.8
72.2
71.8
84.8
84.7
80.8
67.6
72.7
81.6
77.1
56.3
RP
YL
−1
plants ha
2,301
3,129
1,832
732
573
536
1,091
1,858
1,880
1,441
1,467
789
796
997
1,682
1,415
954
1,191
2,270
%
18.8
23.8
15.7
7.0
5.6
5.3
10.1
15.9
16.0
12.8
13.0
7.5
7.6
9.3
14.6
12.7
9.0
10.9
18.6
Estimated
revenue
loss
RB
RC
Average
RB – RC at
average
PDd
Simulated
average RNB,
no correlationb,d
Simulated
average
RNB,
correlatedc,d
—————————————————————— $ ha−1 ——————————————————
$303
$79
$5
$74.47
$145.88
$134.78
$382
$0
$0
$0.00
$83.11
$72.61
$252
$130
$34
$95.62
$161.53
$150.10
$113
$269
$61
$208.18
$258.13
$248.15
$90
$292
$61
$231.02
$278.85
$268.29
$85
$297
$61
$236.54
$283.44
$273.47
$162
$220
$51
$168.50
$225.90
$214.70
$255
$127
$51
$75.47
$148.98
$139.44
$258
$124
$51
$73.05
$145.82
$136.79
$206
$176
$56
$119.50
$183.66
$173.78
$209
$172
$56
$116.40
$181.15
$171.36
$121
$261
$56
$204.81
$255.76
$246.06
$122
$260
$56
$203.87
$255.00
$244.50
$150
$232
$56
$176.29
$231.85
$221.36
$235
$147
$56
$90.86
$160.88
$150.15
$203
$179
$56
$122.67
$186.55
$176.36
$144
$238
$56
$182.03
$236.03
$225.97
$175
$207
$56
$150.89
$205.43
$195.11
$300
$82
$56
$26.34
$97.57
$88.07
a
Abbreviations: fb, followed by; HC, herbicide cost and application charges; MR, expected Palmer amaranth mortality rate; NA, not applicable; RB, estimated relative benefit; RC, added cost; RNB, relative net benefit; RP, expected number of remaining plants
after spray; YL, estimated yield loss.
b
Evaluation was performed with average initial Palmer amaranth plant density and average mortality rates under varying assumptions of correlation among mortality rate probability distributions.
c
See Supplemental Material Table 3 for partial correlation coefficients among weed control options.
d
Bold lettering indicates the top choice (highest RNB = RB – RC) among either single herbicide weed control treatments using different herbicides or their mixture and again the most profitable time interval among weed control systems involving two
sequential passes with different combinations of herbicides.
477
478
Priess et al.: Timing Dicamba and Glufosinate
Likelihood of achieving less
than indicated benefit
100%
75%
Glufosinate
Dicamba
D+G
D fb D 21
G fb G 7
D fb G 14
G fb D 3
50%
25%
Palmer amaranth < 10 cm
$ 800
$ 700
$ 600
$ 500
$ 400
$ 300
$ 200
$ 100
$0
-$ 100
0%
Relative net benefit ($ ha-1)
Likelihood of achieving less
than indicated benefit
100%
75%
Glufosinate
Dicamba
D+G
D fb D 21
G fb G 7
D fb G 7
G fb D 7
50%
25%
Palmer amaranth at 13- to 25-cm
$ 800
$ 700
$ 600
$ 500
$ 400
$ 300
$ 200
$ 100
$0
–$ 100
0%
Relative net benefit ($ ha-1)
Figure 1. Comparison of simulated cumulative distribution functions of relative net benefit accounting for relative sales losses and relative weed control cost for single pass vs.
sequential passes of herbicides when applied to large and small weeds using average soybean price and yield expectations as an example. In the legend, abbreviations are D,
dicamba; G, glufosinate; and fb, followed by. The number following the herbicide abbreviation is the interval in days between sequential applications.
of US$48.75 ha−1 or US$6.11 ha−1 less than the best treatment
option for Palmer amaranth plants <10 cm in height (Table 7).
Because MR differed not only in average but also in range, the
simulated average RNB numbers both with and without correlation across treatments were calculated using the 10,000 randomly
drawn observations from treatments with different MR distributions. Averages reported are larger than the point estimate at average PD and average MR as randomly selected observations from
different distributions lead to greater RNB values. Importantly,
however, the ranking of treatment alternatives continues to highlight dicamba fb glufosinate with 14-d gap between application to
showcase the highest average RNB but now at a lesser average difference of US$3.16 ha−1 or the difference in RNB of US$106.43 ha
−1
for dicamba fb glufosinate at 14 d versus the RNB of US$103.27
ha−1 for dicamba alone (Table 7). Imposing correlation across
treatments did not alter the rankings nor the difference among
the top RNB treatments because RNB differences point to the same
optimal choice of using dicamba fb glufosinate with a 14-d interval
and remain on average about US$3 ha−1 apart from the secondbest weed control treatment.
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Economic implications when spraying is delayed to a largerthan-labeled Palmer amaranth size led to a change in the optimal
control program, larger yield losses, and larger RNB differences
across weed control options compared with results shown for small
weed sizes (Tables 7 and 8). The best option following best management practices (BMPs) to differentiate SOAs remains with
dicamba fb glufosinate; however, treatment interval should be
shorter (7 d) than for smaller sized weeds (14 d). The treatment
option, not following BMPs to manage herbicide resistance by
switching SOA, but of highest RNB, is the dicamba fb dicamba
treatment with an interval of 21 d between herbicide application.
The simulated mean difference between dicamba fb dicamba after
21 d and dicamba fb glufosinate after 7 d is US$27.41 ha−1. A similar option of glufosinate fb glufosinate, in terms of not following
herbicide-resistance BMPs, is much less successful in avoiding soil
weed seedbank accumulation given higher PR or lesser mortality
rate. Assuming the producer chooses the dicamba fb dicamba with
a 21-d interval, 1) expected yield loss is nearly five times greater; 2)
does not follow herbicide-resistance BMPs; and 3) leaves on average 9% more Palmer amaranth plants in the field along with
479
Weed Technology
attendant negative soil weed seedbank implications, than if spraying were to occur in a timelier manner.
To portray differences at the mean (50% percentile) and across
the range of observations in RNB by size of Palmer amaranth plant
at time of herbicide application, we plotted estimated CDFs of RNB
across single and sequential pass treatments (Figure 1). To lessen
the number of CDFs to compare, only the best treatments among
single and sequential pass control options are shown and represent
RNB iterations with the correlation among treatments imposed
(Tables 7 and 8 footnote d). Comparing weed control treatment
options by Palmer amaranth size, timely application at smaller
weed size is less risky. The range in weed control efficacy is smaller
leading to steeper CDFs with smaller weed size in comparison to
the larger weed size. Also, the economic implications of choosing a
less-than-optimal weed control option is smaller with the smaller
weed size (RNB CDFs are horizontally closer in proximity). A second observation is that the cheapest control option, involving a single pass of glufosinate alone, has the least downside risk, given its
low cost, but lags behind in terms of upside potential associated
with superior Palmer amaranth control. Over the 10,000 simulations, there is a 43% and 37% chance of glufosinate alone controlling weeds as effectively as the other treatment options while at the
same time being least expensive, when Palmer amaranth plants are
large and <10 cm, respectively (the likelihoods of glufosinate CDF
RNB values remain at US$0). However, the best weed control
options also indicate superior RNB approximately 98% of the time
in comparison to the cheapest option (CDFs cross at 2% cumulative likelihood), using glufosinate once only, in comparison to
dicamba once only (<10 cm) or dicamba fb dicamba after 21 d
when Palmer amaranth plants are 13 to 25 cm in size. Hence,
the cheapest option is inferior 98% of the time in comparison to
better weed control strategies. Finally, while the CDFs are clearly
differentiable in terms of producer preference when weed size is
large, the distinction at smaller weed size between dicamba alone
versus dicamba fb glufosinate after 14 d, for example is less clear. At
the mean (50% cumulative likelihood), the costlier option is preferred as indicated (Figure 1 and Table 7); however, there is more
downside risk with dicamba fb glufosinate after 14 d in comparison
to dicamba alone because that treatment option is costlier. At the
same time, the upside potential is larger with the costlier option. A
risk-averse producer may thus opt for dicamba alone because the
range in relative profitability is smaller. At the same time, however,
reduction in profit risk increases the soil seedbank given the 7 percentage point lower mortality rate with dicamba alone versus
dicamba fb glufosinate after 14 d (Table 7).
Conclusions and Practical Implications
A single application of dicamba, glufosinate, or dicamba plus glufosinate alone did not control Palmer amaranth greater than 80%,
regardless of weed size. Sequential applications of dicamba fb
dicamba and glufosinate fb glufosinate did not result in 100% control of Palmer amaranth at any time interval or regardless of weed
size. In order to mitigate the selection of biotypes with reduced sensitivity to the few remaining effective postemergence herbicides in
XtendFlex® soybean or cotton, producers should adopt sequential
herbicide application regimes and other integrated weed management strategies to completely control Palmer amaranth. Both
dicamba and glufosinate have already experienced a tremendous
amount of selection. The risk for further selection of biotypes with
reduced sensitivity to either herbicide in single SOA sequential herbicide systems is high. To increase the sustainability of herbicides,
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
an optimized sequence and time interval between applications of
dicamba and glufosinate should be used in the XtendFlex® technology. Dicamba fb glufosinate 14 d later was the only sequential postemergence system that provided 100% control and 100% mortality
of Palmer amaranth when herbicides were applied to weeds <10
cm in height. On larger-than-labeled Palmer amaranth sizes, complete control and mortality was not achieved; therefore, weed size
at the time of the initial application is still of the utmost importance. When weed size increases to 13 to 25 cm in height, incomplete control of Palmer amaranth leads to greater variability in
producer returns not only for individual treatment options but also
across treatment options. Economic returns were highest for
dicamba fb dicamba after 21 d, which can create economic pressure
to choose a weed control option that does not follow herbicideresistance BMPs in comparison to using dicamba fb glufosinate
after 7 d, or the next best option in Table 8.
Finally, the competitive impact of weeds that survive contact
and systemic herbicides like glufosinate and dicamba should be
investigated in the future because Palmer amaranth plants
appeared to rapidly regrow 14 d after a glufosinate application
and slower death and limited regrowth was observed from plants
treated with dicamba, in the absence of a crop (GLP, personal
observation). The inability to quickly remove Palmer amaranth
from crops following a dicamba application may result in competition for limited resources for an extended period following application of the herbicide. Conversely, the regrowth of glufosinatetreated Palmer amaranth 14 d after application may also influence
the crop and weed interaction. Changes in the competitiveness of
Palmer amaranth following a herbicide application in the presence
of a crop would likely affect weed seed production.
Supplementary material. To view supplementary material for this article,
please visit https://doi.org/10.1017/wet.2022.44
Acknowledgments. Funding for this research was provided by Bayer
CropScience. The authors declare no conflicts of interest.
References
Afifi M, Swanton CJ (2012) Early physiological mechanisms of weed competition. Weed Sci 60:542–551
Ahrens WH, ed (1994) Pages 149–174 in Herbicide Handbook. 7th ed.
Champaign, IL: Weed Science Society of America
Anderson DM, Swanton CJ, Hall JC, Mersey BG (1993) The influence of temperature and relative humidity on the efficacy of glufosinate-ammonium.
Weed Res 33:139–147
Anonymous (2020a) Engenia® herbicide label. http://www.cdms.net/Labels
Msds/LMDefault.aspx. Accessed: February 29, 2021
Anonymous (2020b) XtendiMax® with VaporGrip® herbicide label. Bayer
Publication 201105.http://www.cdms.net/LabelsMsds/LMDefault.aspx. Accessed:
February 29, 2021
Bagavathiannan MV, Norsworthy JK, Smith KL, Neve P (2013) Modeling the
evolution of glyphosate-resistance in barnyardgrass (Echinochloa crus-galli)
in cotton-based production systems of the midsouthern United States. Weed
Technol 27:475–487
Bagavathiannan MV, Norsworthy JK, Smith KL, Neve P (2014) Modeling the
simultaneous evolution of resistance to ALS- and ACCase-inhibiting herbicides in barnyardgrass (Echinochloa crus-galli) in Clearfield® rice. Weed
Technol 28:89–103
Bensch CN, Horak MJ, Dallas P (2003) Interference of redroot pigweed
(Amaranthus retroflexus), Palmer amaranth (Amaranthus palmeri), and
common waterhemp (A. rudis) in soybean. Weed Sci 51:37–43
Burke IC, Askew SD, Corbett JL, Wilcut JW (2005) Glufosinate antagonizes clethodim control of goosegrass (Eleusine indica). Weed Technol 19:664–668
480
Butts TR, Samples CA, Franca LX, Dodds DM, Reynolds DB, Adams JW,
Zollinger RK, Howatt KA, Fritz BK, Hoffmann CW, Kruger GR (2018)
Spray droplet size and carrier volume effect on dicamba and glufosinate efficacy. Pest Manag Sci 74:2020–2029
Chahal GS, DL Jordan, AC York, EP Prostko (2011) Palmer amaranth control
with combinations of 2,4-DB and diphenylether herbicides. Crop Manag
10:1–7
Coetzer EK, Al-Khatib K, Loughin TM (2001) Glufosinate efficacy absorption,
and translocation in Amaranthus species as affected by relative humidity and
temperature. Weed Sci 49:8–13
Coetzer EK, Al-Khatib K, Peterson DE (2002) Glufosinate efficacy on
Amaranthus species in glufosinate-resistant soybean (Glycine max). Weed
Technol 16:326–331
Colby S (1967) Calculating synergistic and antagonistic responses of herbicide
combinations. Weeds 15:20–22
Cummins I, Cole DJ, Edwards R (1999) A role for glutathione transferases functioning as glutathione peroxidases in resistance to multiple herbicides in
black-grass. Plant J 18:285–292
Diggle AJ, Neve PB, Smith FP (2003) Herbicides used in combination can
reduce the probability of herbicide resistance in finite weed populations.
Weed Res 43:371–382
Duke SO (2012) Why have no new herbicides modes of action appeared in
recent years? Pest Manag Sci 68:505–512
Etheridge RE, Hart WE, Hayes RM, Mueller TC (2001) Effect of Venturi-type
nozzles and application volume on postemergence herbicide efficacy. Weed
Technol 15:75–80
Evans SP, Knezevic SZ, Lindquist JL, Shapiro CA (2003) Influence of nitrogen
and duration of weed interference on corn growth and development. Weed
Sci 51:546–556
Grossmann K (2000) Mode of action of auxin herbicides: a new ending to a long,
drawn out story. Trends Plant Sci 5:506–508
Heap IM (2020) The international survey of herbicide resistant weeds. http://
www.weedscience.com. Accessed: September 2, 2020
Jha P, Norsworthy JK, Riley MB, Bielenberg DG, Bridges W (2008) Acclimation
of Palmer amaranth (Amaranthus palmeri) to shading. Weed Sci 56:729–734
Klose SL, Smith J, Thompson B, Waller M, Zapata S, Keeling W (2019) 2018
Texas agricultural custom rates. Texas AgriLife Extension Publication.
https://agecoext.tamu.edu/wpcontent/uploads/2019/03/TXCustomRates
Survey2018.pdf. Accessed: April 22, 2021
Lee SD, Oliver LR (1982) Efficacy of acifluorfen on broadleaf weeds. Times and
methods for application. Weed Sci 30:520–526
Lindsay K, Popp M, Norsworthy J, Bagavathiannan M, Powles S, Lacoste M
(2017) PAM: Decision support for long-term palmer amaranth (Amaranthus
palmeri) control. Weed Technol 31:915–927
Markham MY, Stolenberg DE (2009) Red:far-red light effects on corn growth
and productivity in field environments. Weed Sci 57:208–215
McDonald JB, Xu YJ (1995) A generalization of the beta distribution with applications. J Econ 66:133–152
https://doi.org/10.1017/wet.2022.44 Published online by Cambridge University Press
Priess et al.: Timing Dicamba and Glufosinate
McKinlay KS, Ashford R, Ford RJ (1974) Effects of drop size, spray volume, and
dosage on paraquat toxicity. Weed Sci 22:31–34
Merchant RM, Sosnoskie LM, Culpepper SA, Steckle LE, York AC, Braxton BL,
Ford JC (2013) Weed response to 2,4-D, 2,4-D-DB, and dicamba applied
alone or with glufosinate. J Cotton Sci 17:212–218
Meyer CJ, Norsworthy JK (2019) Influence of weed size on herbicide interaction
for Enlist™ and Roundup Ready® Xtend technologies. Weed Technol
33:569–577
Meyer CJ, Norsworthy JK (2020). Timing and application rate for sequential
applications of glufosinate are critical for maximizing control of annual
weeds in LibertyLink® soybean. Int J Agron 2020:1–7
Meyer CJ, Norsworthy JK, Kruger GR, Barber T (2015) Influence of droplet size
on efficacy of the formulated products Engenia, Roundup Powermax, and
Liberty. Weed Technol 29:641–652
Meyer CJ, Peter F, Norsworthy JK, Beffa R (2020) Uptake, translocation,
and metabolism of glyphosate, glufosinate, and dicamba mixtures in
Echinochloa crus-galli and Amaranthus palmeri. Pest Manag Sci 76:3078–3087
Neve P, Norsworthy JK, Smith KL, Zelaya IA (2011) Modelling evolution and
management of glyphosate resistance in Amaranthus palmeri. Weed Res
51:99–112
Norsworthy JK, Korres NE, Walsh MJ, Powles SB (2016) Integrating herbicide
programs with harvest weed seed control and other fall management practices for the control of glyphosate-resistant Palmer amaranth (Amaranthus
palmeri). Weed Sci 64:540–550
Norsworthy JK, Ward S, Shaw D, Llewellyn R, Nichols R, Webster T, Bradley K,
Frisvold G, Powles S, Burgos N, Witt W, Barrett M (2012) Reducing the risks
of herbicide resistance: best management practices and recommendations.
Weed Sci 60 (SP1):31–62
Priess GL, Norsworthy JK, Barber LT, Castner MC (2019) Interaction between
sequential applications of dicamba and glufosinate. Page 315 in Proceedings
of the 2019 Beltwide Cotton Conference. New Orleans, LA, January 8–10,
2019
Raghavan C, Ong EK, Dalling MJ (2005) Effect of herbicidal application of
2,4d-dichlorophenoxyacetic acid in Arabidopsis. Funct Integr Genomic
5:4–17
Scarponi LS, Buono DD, Vischetti C (2005) Effect of pretilachlor and fenclorim
on carbohydrate and protein formation in relation to their persistence in rice.
Pest Manag Sci 61:371–376
Steckel GJ, Wax LM, Simmons FW, Phillips WH II (1997) Glufosinate efficacy
on annual weeds is influenced by rate and growth stage. Weed Technol
11:484–488
Sung-Eun K, Yoon-Sung P, Sul-Hwa A, Jae-Chul C (2005) Absorption and
translocation behavior of glyphosate as affected by contact herbicide carfentrazone-ethyl. Kor J Weed Sci 25:62–69
Tollenaar M, Dibo AA, Aguilera A, Weise SF, Swanton SJ (1994) Effect of crop
density on weed interference in maize. Agron J 86:591–595
Wilson RG (2005) Response of dry bean and weeds to fomesafen and fomesafen
tank mixtures. Weed Technol 19:201–206