Hostname: page-component-7c8c6479df-27gpq Total loading time: 0 Render date: 2024-03-28T09:40:00.043Z Has data issue: false hasContentIssue false

Kochia (Kochia scoparia) Emergence Profiles and Seed Persistence across the Central Great Plains

Published online by Cambridge University Press:  15 June 2017

J. Anita Dille*
Affiliation:
Professor, Associate Professor, and Graduate Research Assistant, Kansas State University, Manhattan, KS 66506
Phillip W. Stahlman
Affiliation:
Professor, Kansas State University, Hays, KS 67601
Juan Du
Affiliation:
Professor, Associate Professor, and Graduate Research Assistant, Kansas State University, Manhattan, KS 66506
Patrick W. Geier
Affiliation:
Assistant Scientist and Professor, Kansas State University, Garden City, KS 67846
Jarrett D. Riffel
Affiliation:
Professor, Associate Professor, and Graduate Research Assistant, Kansas State University, Manhattan, KS 66506
Randall S. Currie
Affiliation:
Assistant Scientist and Professor, Kansas State University, Garden City, KS 67846
Robert G. Wilson
Affiliation:
Professor (retired) and Post-doctoral Research Associate, University of Nebraska–Lincoln, Scottsbluff, NE 69631
Gustavo M. Sbatella
Affiliation:
Professor (retired) and Post-doctoral Research Associate, University of Nebraska–Lincoln, Scottsbluff, NE 69631
Philip Westra
Affiliation:
Professor, Colorado State University, Fort Collins, CO 80523
Andrew R. Kniss
Affiliation:
Associate Professor, University of Wyoming, Laramie, WY 82071
Michael J. Moechnig
Affiliation:
Assistant Professor, South Dakota State University, Brookings, SD 57007
Richard M. Cole
Affiliation:
Weed Management Technical Lead (retired), Monsanto Company, St. Louis, MO 63167
*
*Corresponding author’s E-mail: dieleman@ksu.edu
Rights & Permissions [Opens in a new window]

Abstract

Timing of weed emergence and seed persistence in the soil influence the ability to implement timely and effective control practices. Emergence patterns and seed persistence of kochia populations were monitored in 2010 and 2011 at sites in Kansas, Colorado, Wyoming, Nebraska, and South Dakota. Weekly observations of emergence were initiated in March and continued until no new emergence occurred. Seed was harvested from each site, placed into 100-seed mesh packets, and buried at depths of 0, 2.5, and 10 cm in fall of 2010 and 2011. Packets were exhumed at 6-mo intervals over 2 yr. Viability of exhumed seeds was evaluated. Nonlinear mixed-effects Weibull models were fit to cumulative emergence (%) across growing degree days (GDD) and to viable seed (%) across burial time to describe their fixed and random effects across site-years. Final emergence densities varied among site-years and ranged from as few as 4 to almost 380,000 seedlings m−2. Across 11 site-years in Kansas, cumulative GDD needed for 10% emergence were 168, while across 6 site-years in Wyoming and Nebraska, only 90 GDD were needed; on the calendar, this date shifted from early to late March. The majority (>95%) of kochia seed did not persist for more than 2 yr. Remaining seed viability was generally >80% when seeds were exhumed within 6 mo after burial in March, and declined to <5% by October of the first year after burial. Burial did not appear to increase or decrease seed viability over time but placed seed in a position from which seedling emergence would not be possible. High seedling emergence that occurs very early in the spring emphasizes the need for fall or early spring PRE weed control such as tillage, herbicides, and cover crops, while continued emergence into midsummer emphasizes the need for extended periods of kochia management.

Type
Weed Biology and Ecology
Copyright
© Weed Science Society of America, 2017 

Kochia is a major weed in dryland and irrigated crops of the western three-quarters of the United States as well as in pastures, roadsides, and waste areas. Kochia is often observed as the first summer annual weed species to emerge in fields exploiting the limited spring soil moisture in the arid to semi-arid regions of the Great Plains (Friesen et al. Reference Friesen, Beckie, Warwick and Van Acker2009). Kochia was introduced from Europe in the mid- to late-1800s. By 1930, it was observed widely across the central United States (Eberlein and Fore Reference Eberlein and Fore1984). Historically, kochia abundance increased during and following hot, dry years (Blackshaw et al. Reference Blackshaw, Larney, Lindwall, Watson and Derksen2001; Eberlein and Fore Reference Eberlein and Fore1984; Forcella Reference Forcella1985). Kochia thrives in droughty, nutrient-deficient saline soils where few other plants will grow, allowing it to exploit ecological niches wherever they exist. When kochia invades high-quality agricultural lands, it can grow to 2 m in height, develop a root system to 4 m in depth (Phillips and Launchbaugh Reference Phillips and Launchbaugh1958), and produce a woody stem more than 2 cm in diameter. Kochia exhibits a tumbling mechanism of seed dispersal whereby the stem dehisces at the base, and the rounded, stiff, mature plants are driven across the landscape by strong winds, dropping seeds as they bounce along at speeds up to 65 km h−1 (Baker et al. Reference Baker, Beck, Bienkiewicz and Bjostad2008; Becker Reference Becker1978; Stallings et al. Reference Stallings, Thill, Mallory-Smith and Lass1995). This very efficient method of seed dispersal provides a rapid means for kochia to spread to new ecological niches, helping generate its reputation as one of the fastest-spreading weeds in the United States and Canada (Blackshaw et al. Reference Blackshaw, Larney, Lindwall, Watson and Derksen2001; Forcella Reference Forcella1985). Kochia seed is known to have very low or no dormancy and thus does not produce a persistent seedbank, unlike many other annual weed species (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008; Zorner et al. Reference Zorner, Zimdahl and Schweizer1984).

POST weed control had shifted to predominantly nonresidual glyphosate-based programs in several crops and during fallow periods throughout the Great Plains in the mid-2000s, resulting in a dramatic increase in glyphosate-resistant (GR) kochia populations in no-tillage cropping systems (Godar and Stahlman Reference Godar and Stahlman2015). In Kansas, a GR biotype of kochia was first identified in 2007 with subsequent confirmation of GR biotypes across the Great Plains of the United States and Canada (Godar Reference Godar2014; Heap Reference Heap2017; Waite et al. Reference Waite, Thompson, Peterson, Currie, Olson, Stahlman and Al-Khatib2013). Additionally, numerous kochia biotypes have developed cross- and multiple resistance to acetolactate synthase, photosystem II, synthetic auxin, and glycine inhibitors that have been traditionally used for weed control in winter wheat and in fallow (Bell et al. Reference Bell, Nalewaja and Schooler1972; Derksen et al. Reference Derksen, Anderson, Blackshaw and Maxwell2002; Mengistu et al. Reference Mengistu, Christoffers and Lym2005; Primiani et al. Reference Primiani, Cotterman and Saari1990; Thompson et al. Reference Thompson, Thill and Shafii1994; Varanasi et al. Reference Varanasi, Godar, Currie, Dille, Thompson, Stahlman and Jugulam2015; Waite et al. Reference Waite, Thompson, Peterson, Currie, Olson, Stahlman and Al-Khatib2013).

The confirmation of GR kochia from geographically distant sites throughout the Great Plains is yet another example of how this ubiquitous dicot weed has successfully used genetic adaptation to thrive and spread. Biological information is needed on the populations found in these areas to aid management of kochia. It is known that kochia emerges very early in the spring and seed persistence in the soil is very short (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008; Zorner et al. Reference Zorner, Zimdahl and Schweizer1984), but there is no information on geographic influence on emergence patterns, nor is there any evidence that burial depth influences seed viability. Thus, the specific research objectives were to measure and describe the emergence profile of naturally occurring kochia populations to determine how early kochia emergence occurs, the pattern of emergence over time, and the persistence of seed in the soil at field sites across the Great Plains in order to aid producers in managing this weed species.

Materials and Methods

Seedling Emergence Experiment

In February of 2010 and again in 2011, field sites were identified by coauthors across five states—Kansas, Colorado, Wyoming, Nebraska, and South Dakota (Table 1)—where naturally occurring kochia populations were occurring in cropland or non-cropland sites. Crop sites were located within and near the edges of actively cropped fields; non-cropland sites were not being actively cropped (outside field, near buildings), were undisturbed, and had kochia present. Three to eight permanent quadrats were marked at each site in which kochia seedlings were counted and removed as they emerged on a weekly basis. Count data (plants m−2) were converted to cumulative percent emergence and related to cumulative growing degree days (GDD) using a Tbase air temperature of 0 C for kochia (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008). Accumulation of GDD began January 1 in each year using air temperature data from a weather station located near each site.

Table 1 Mean kochia density and observed range (low and high) for total emergence (plantsm−2) observed across sites and field types identified as cropland or non-cropland in 2010 and 2011.

a Abbreviations: Crop, cropland; NC, non-cropland; NT, no-tillage system; T, tilled system.

Seed Burial Experiment

In the fall of 2010 and 2011, seeds were collected from multiple kochia plants located near the emergence quadrats from different emergence cohorts. Up to three emergence cohorts were created by spraying out all emerged kochia and allowing later emerging seedlings to grow and survive until fall. Each cohort was separated in time by 2 wk between spray applications. Bulk seed samples were sent to the Weed Ecology Lab, Department of Agronomy, Manhattan, KS, to be cleaned using sieves and an air-column separator (Seedburo Equipment, 2293 S. Mt. Prospect Road, Des Plaines, IL). Cleaned kochia seeds from each site and cohort were counted using a seed counter (Seedburo 801 Count-A-Pak, Seedburo Equipment, 2293 S. Mt. Prospect Road, Des Plaines, IL) to create 100-seed lot packets. Protective mesh packets each containing 100 kochia seeds were made from 5 by 5 cm squares of 30-mesh wire screen (TWP, 2831 Tenth Street, Berkeley, CA). Packets were sent back to each respective site to be buried. At each site, 15-cm-deep holes were dug in the soil, a wire cage of 6.4-mm mesh (0.25-inch hail screen) was inserted, soil was replaced inside to be able to bury packets at all three different depths within each cage (10 cm, 2.5 cm, and on the surface at 0 cm) or one depth per cage separately (only at Lingle, WY), and a wire cap was placed on top of the cage to keep packets in place and deter rodent predation. Sixteen cages (three burial depths per cage) were established at all sites, except at Lingle, WY, where 48 cages (each burial depth separately) were established in each year. The cages represented four replications with four planned exhumation times for each cohort per site. Seeds collected in fall 2010 were buried by December with four exhumations planned for March and October of 2011 and 2012. Another set of packets was buried with seeds collected in fall 2011 and exhumations in March and October of 2012 and 2013.

After packets were exhumed, they were sent to the Weed Ecology Lab, Department of Agronomy, Manhattan, KS, where seeds were removed from each packet and placed in individual petri dishes with filter paper and tap water in a growth chamber set at 20/10 C day/night temperatures and a 12-h photoperiod. Germination counts were taken every 3 to 4 d for a total of 21 d; a seed was considered germinated when its radicle was greater than 2 mm in length. Petri dishes were allowed to dry for 30 d and then rewetted; subsequently germinating seeds were counted or remaining seeds were tested with a “crush” test to identify hard, viable seed (Borza et al. Reference Borza, Westerman and Liebman2007; Sawma and Mohler Reference Sawma and Mohler2002). The percentage of viable seed from each exhumation was determined based on n=100 total seed packet−1 for each burial year, site, cohort, and depth of burial (Ullrich et al. Reference Ullrich, Buyer, Cavigelli, Seidel and Teasdale2011).

Statistical Modeling Approach

A nonlinear mixed-effects modeling (NLME) method was used to analyze both the seedling emergence data and seed persistence data. To be specific, for the seedling emergence data, it was the relationship between cumulative percent kochia emergence and cumulative GDD, while for the seed burial data it was the relationship between percent viable seed and length of seed burial, based on the following flexible mixed-effects Weibull model (Clay et al. Reference Clay, Davis, Dille, Lindquist, Ramirez, Sprague, Reicks and Forcella2014; Davis et al. Reference Davis, Clay, Cardina, Dille, Forcella, Lindquist and Sprague2013; Ratkowski Reference Ratkowski1983):

(1a) $$y_{{ij}} \,{\equals}\,\O _{{1i}} {\minus}\O _{{2i}} {\rm exp}\left[ {{\minus}{\rm exp}\left( {\O _{{3i}} } \right)x_{{ij}}^{{\O _{{4i}} }} } \right]{\plus}\in_{{ij}} $$

where y ij is the jth response (either cumulative percent emergence or percent viable seed) at ith site-year, x ij is the jth value of the cumulative GDD for seedling emergence data or the length of seed burial (fraction of year) for the viable seed data at ith site-year; Ø1i (Asym) represents the asymptote as x→∞ and Ø2i (Drop) represents the difference between asymptote Ø1i and asymptote as x →0; the logarithm of the rate constant (lrc) is Ø3i ; and the shape parameter (pwr) is Ø4i .

(1b) $$\eqalignno{ &#x0026; \O _{i} \,{\equals}\left[ {\matrix{ {\matrix{ {\O _{{1i}} } \cr {\O _{{2i}} } \cr {\O _{{3i}} } \cr } } \cr {\O _{{4i}} } \cr } } \right]{\equals}\left[ {\matrix{ {\matrix{ {\beta _{1} } \cr {\beta _{2} } \cr {\beta _{3} } \cr } } \cr {\beta _{4} } \cr } } \right]{\plus}\left[ {\matrix{ {\matrix{ {b_{{1i}} } \cr {b_{{2i}} } \cr {b_{{3i}} } \cr } } \cr {b_{{4i}} } \cr } } \right]{\equals}\,{\bf{\beta }} {\plus}{\mib{b}} _{i} \cr &#x0026; {\bf b}_{{\bf i}} \,\sim\,N\left( {0,\Psi } \right), \in_{{ij}} \,\sim\,N\left( {0,\sigma ^{2} } \right) $$

where the site-year-specific parameter vector, Ø i , consists of a vector of fixed effects ) and a vector of random effects ( b i ) associated with the ith group, with a variance–covariance matrix (ψ) that is assumed to be a diagonal matrix. The random errors ij ) are assumed to be independent and identically normally distributed. Note that when Drop2i ) is positive, then the expected regression function has an increasing trend and it is used to model the cumulative percent emergence, and when Drop2i ) is negative, then the expected regression function has a decreasing relationship and it is used to model percent viable seed data over time.

The model fit was based on maximum likelihood estimation method using the library ‘nlme’ of R v. 3.3.1 (R Core Team 2016). To determine whether a reduced model with fewer parameters should be used, or which random effects should be eliminated, or how useful some of the covariates were to explain between-group variability, we used common statistical model selection criteria such as the Akaike information criterion (AIC), Bayesian information criterion (BIC), or likelihood ratio test.

Seedling emergence data were modeled separately by two groups: 11 site-years for Kansas and 6 site-years for Wyoming and Nebraska. Seedbank persistence data were modeled separately by two groups: 10 site-years that had multiple-cohort data and 6 site-years that had single-cohort data. For single-cohort data, significant covariates were identified that included site, field type (cropland or non-cropland), and year of seed burial and used Equation 2:

(2) $$\O _{{3i}} \,{\equals}\,\beta _{3} {\plus}\gamma _{1} site_{{1i}} {\plus}\gamma _{2} site_{{2i}} {\plus}\gamma _{3} \,field_{{3i}} {\plus}\gamma _{4} \,year_{i} {\plus}b_{{3i}} $$

Cumulative GDD and calendar date for the start (10% emergence) and GDD for the midpoint (50%) and end (90%) of kochia emergence were calculated based on best NLME model fit for each site-year.

Results and Discussion

Over the course of the study, observations of kochia seedling emergence and seedbank persistence were made at sites ranging from Garden City, KS, in the south to Brookings, SD, in the north, and from Fort Collins, CO, in the west to Manhattan, KS, in the east. A total of 17 site-years had useful observations for seedling emergence (Table 1), and 16 site-years had useful observations for seedbank persistence.

Seedling Emergence

The maximum number of kochia that emerged varied widely across the sites and field types in 2010 and 2011 but are not atypical of kochia densities (Table 1). For example, lowest observed density was 4 plants m−2 in one plot in 2010 at Garden City, KS, in a no-tillage cropland environment with several years of aggressive herbicide management (RS Currie, personal observation), and up to 379,100 seedlings m−2 in a plot in a non-cropland environment in 2010 at Hays, KS (Table 1). Average kochia densities in cropland environments were 2,137 and 4,755 seedlings m−2 in 2010 and 2011, respectively, while in non-cropland environments, populations were 70,440 and 18,530 seedlings m−2 in 2010 and 2011, respectively. The increased number of seedlings in non-cropland environments was likely due to high plant density and seed rain from the previous year when no active weed control was applied in those environments. Average kochia density observed in fields in Manitoba was 2,292 seedlings m−2 and was characterized as a much greater density level than typically observed in cultivated fields in Manitoba (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008). Typical end-of-season populations observed in western Nebraska in untreated check plots were 16 to 24 kochia plants m−2 at soybean harvest, but in the following winter wheat crop, kochia densities ranged from 39 to 1,950 plants m−2 depending on year (Wicks et al. Reference Wicks, Martin and Hanson1997). Kochia population size can increase quickly if left uncontrolled from one year to the next.

Kochia seedlings were initially observed in 2010 by the middle of March (day of year [DOY] 70), and in 2011 by the end of February (DOY 60) and early March depending on the site (Figure 1). Across most site-years, seedling emergence quickly achieved 100% cumulative emergence, often within 2 to 3 wk after initial observations. However, the rate of seedling emergence varied widely among site-years when described using DOY (Figure 1). This is not unusual when examining emergence timing across a wide geography, as similar observations were seen with giant ragweed (Ambrosia trifida L.) (Davis et al. Reference Davis, Clay, Cardina, Dille, Forcella, Lindquist and Sprague2013) or common sunflower (Helianthus annuus L.) (Clay et al. Reference Clay, Davis, Dille, Lindquist, Ramirez, Sprague, Reicks and Forcella2014) across the midwestern United States. Kochia seedlings are known to be the first summer annuals to emerge (Werle et al. Reference Werle, Sandell, Buhler, Hartzler and Lindquist2014), but DOY will vary across the Great Plains region.

Figure 1 Cumulative kochia seedling emergence, by day of year, for 17 site-years across the Great High Plains region of the United States. NC, non-cropland; C, cropland; NT, no-tillage system; T, tilled system

Seedling emergence patterns for kochia populations across 11 site-years in Kansas were analyzed separately from the 6 site-years in Nebraska and Wyoming, using NLME. After comparison among a pool of candidate models, the most parsimonious NLME Weibull model for seedling emergence in Kansas had a fixed shape parameter value for Ø4i (pwr)=1, and only Ø1i (Asym) and Ø2i (Drop) had random effects using Equation 1 (Table 2). The type of field (cropland or non-cropland) was not significant in explaining the variation in seedling emergence patterns, thus this model predicted kochia seedling emergence among the 11 site-years across Kansas (Tables 2 and 3). The quality of the fitted NLME model can be visualized in Figure 2. The solid line gives the population average (fixed) predicted cumulative percent seedling emergence for the 11 site-years in Kansas, while the dashed line shows the specific prediction for each site-year, which is in good agreement with the observed values and site-year-specific parameter estimates shown in Table 3. In general, seedling emergence at Hays (4 site-years) was more rapid (dashed line to the left of fixed line) than that observed in Garden City (3 site-years), which was slower (dashed line to the right of fixed line).

Figure 2 The site-year-specific predicted cumulative percent of emergence (dashed line) and fixed population predictions (solid line) by fitting nonlinear mixed-effects Weibull model of Kochia seedling emergence for 11 site-years in Kansas. GDD, growing degree days.

Table 2 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia cumulative seedling emergence following a Weibull response function to cumulative GDD across 11 site-years in Kansas.

Table 3 Weibull model parameters (Asym and Drop) for the predictive seedling emergence model, calendar date for 10% emergence, and estimated GDD for 10%, 50%, and 90% cumulative seedling emergence for each of 11 site-years across field types in Kansas.

a Abbreviations: Crop, cropland; NC, non-cropland; NT, no-tillage system; T, tilled system.

Seedling emergence patterns for kochia populations observed in Wyoming and Nebraska (6 site-years) were also modeled using NLME, and based on comparison among a pool of candidate models using AIC, the most parsimonious NLME Weibull model was constrained by equality of Asym and Drop parameters, i.e., Ø1i 2i and only Ø1i (Asym) and Ø4i (pwr) had random effects in Equation 1 (Table 4). The quality of the fitted NLME model for 6 site-years across Wyoming and Nebraska can be visualized in Figure 3. The solid line gives the population average (fixed) predicted percent cumulative seedling emergence, while the dashed line shows the site-year specific prediction, which is in good agreement with the observed values. In general, seedling emergence across Wyoming and Nebraska in 2010 was slower and variable as compared with 2011 (Table 5; Figure 3).

Figure 3 The site-year-specific predicted cumulative percent of emergence (dashed line) and fixed population predictions (solid line) by fitting nonlinear mixed-effects Weibull model of Kochia seedling emergence for 6 site-years in NE and WY. GDD, growing degree days.

Table 4 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia cumulative seedling emergence following a Weibull response function to cumulative GDD across 6 site-years in Wyoming and Nebraska.

Table 5 Weibull function parameters (Asym and pwr) for the predictive seedling emergence model, calendar date for 10% emergence, and estimated GDD for 10%, 50%, and 90% cumulative seedling emergence for each of the 6 site-years in Wyoming and Nebraska.

a Abbreviations: Crop, cropland; NC, non-cropland.

The predicted cumulative GDD needed for 10%, 50%, and 90% cumulative kochia emergence in Kansas, based on the average (fixed) model, were 168, 308, and 692 GDD, respectively, while for the Nebraska and Wyoming populations, the GDD needed were 90, 156, and 230 GDD. Moving north, fewer GDD were required for kochia seedling emergence. This may indicate a lower critical temperature for kochia in more northern latitudes and would need to determined. In its more northerly range in Manitoba, field emergence of kochia was observed to begin at 50 cumulative GDD (Tbase 0 C, soil temperature at 2.5-cm depth); this was typically accumulated by April 17, and kochia emergence continued throughout the growing season, although in small numbers after the initial flush (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008). Werle et al. (Reference Werle, Sandell, Buhler, Hartzler and Lindquist2014) found that kochia was the first summer annual that emerged in their study in Iowa conditions, with a short rapid window of emergence beginning in early April through mid-May. Anderson and Nielsen (Reference Anderson and Nielsen1996) observed kochia beginning to emerge in the field at Akron, CO, in early April, with 80% of the seedlings emerging between April 11 and June 20 from 1987 to 1990.

Although kochia is a C4 plant that thrives under hot temperatures, its seed readily germinates very early in the spring when snow is still on the ground, and kochia plants in the early seedling stage are very cold tolerant, usually surviving spring freezing nighttime temperatures (Chepil Reference Chepil1946; Dyer et al. Reference Dyer, Chee and Fay1993). As soon as there is sufficient moisture in the early spring, kochia will initiate germination and emergence (Werle et al. Reference Werle, Sandell, Buhler, Hartzler and Lindquist2014). Often, this is less moisture than required for soil-applied herbicides to be effective for kochia control (Sebastian et al. Reference Sebastian, Nissen, Westra, Shaner and Butters2017).

Cropland sites had an emergence pattern that was later and more prolonged compared with non-cropland sites, in general. For example, when comparing the two sites in Hays, KS, kochia emergence was 2 wk later in the cropland site compared with the non-cropland site in 2010 (Table 3). Kochia is capable of adapting to chosen weed control practice, as was demonstrated in western Nebraska, where kochia biotypes were selected for delayed emergence until after the soil-applied herbicide isoxaflutole had dissipated compared with plots without a residual herbicide (Sbatella and Wilson Reference Sbatella and Wilson2010). In cropland environments with previous aggressive weed management, such as both years at Garden City, the overall emergence was low, with maximum of 22 to 193 seedlings m−2, but required more than 1,450 GDD in a tilled environment and more than 2,350 GDD in no-tillage cropland sites to reach 90% cumulative emergence (Table 3), demonstrating a very prolonged emergence pattern and opportunity for selection of later-emerging individuals.

Seed Burial

In this study, multiple cohorts of kochia plants whose seed were harvested, buried, and exhumed as separate cohorts were evaluated for 10 site-years. These included two from Colorado (dryland and irrigated sites with seed buried in 2010), four from Nebraska (Mitchell and Scottsbluff with seed buried in 2010 and 2011), three from Kansas (Garden City no-tillage 2010, and Stockton 2010 and 2011), and one from Wyoming (2010). For these data, after comparison among a pool of candidate models based on AIC and BIC, the most parsimonious NLME Weibull model for percent viable seed as function of length of seed burial had a fixed shape parameter value, Ø4i (pwr)=1, and only Ø1i (Asym) and Ø3i (lrc) had random effects in Equation 1. Parameter estimates for this best model are given in Tables 6 and 7. The depth of burial (0, 2.5, or 10 cm) had no effect on seedbank persistence, and type of field with respect to cropland or non-cropland did not contribute enough information to be included in the model. The quality of the fitted NLME model can be visualized in Figure 4. The solid line gives the population average (fixed) predicted percent viable seed for the 10 site-years and at each burial depth with multiple cohorts, while the dashed line shows the specific prediction for each site-year and burial depth, which is in good agreement with the observed values. By the first exhumation at 0.5 yr, there was variability in percent viable seed, but by the fourth exhumation at 2.0 yr, percent of viable seeds from all site-years and burial depths was very low (<2%) (Figure 4).

Figure 4 The predicted percent viable seed (dashed line) and fixed population predictions (solid line) for the nonlinear mixed-effects Weibull model of kochia populations across site-years and burial depths with multiple cohorts. Site and year abbreviations: NEMit, Mitchell, NE; NESco, Scottsbluff, NE; CoIri, Colorado-irrigated; KSSto, Stockton, KS; CoDry, Colorado-dryland; KSGar, Garden City, KS; WYLin, Lingle, WY; 10, buried seed in 2010; 11, buried seed in 2011.

Table 6 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia seedbank persistence following a Weibull response function to length of seed burial across 10 site-years and multiple cohorts of kochia seed in Wyoming, Nebraska, Colorado, and Kansas.

Table 7 Weibull function parameters (Asym and lrc) for the predictive percent viable seed for each seed burial depth across 10 site-years and multiple cohorts in Wyoming, Nebraska, Colorado, and Kansas.

A total of 6 site-years had single cohorts of kochia seed buried, including four in Kansas (Hays crop and non-cropland in 2010 and Garden City tilled and no-tillage environments in 2011) and two in South Dakota (Brookings, 2010 and 2011). For these data, the most parsimonious NLME Weibull model for percent viable seed as function of length of seed burial had a fixed shape parameter value, Ø4i (pwr)=1, and Ø3i (lrc) had random effects in Equation 1 (Table 8) as well as significant covariates for site, field type, and year of burial as described by Equation 2. The quality of the fitted NLME model can be visualized in Figure 5 (Table 9).

Figure 5 The predicted percent viable seed (dashed line) and fixed population predictions (solid line) for the NLME Weibull model of kochia populations across site-years and burial depths with a single cohort. Site-year abbreviations: KSha10, Hays, KS, 2010, non-cropland; KSha10c, Hays, KS, 2010, cropland; SDak10, Brookings, SD, 2010; SDak11, Brookings, SD, 2011; KSGa11T, Garden City, KS, 2011, tilled system; KSGa11, Garden City, KS 2011, no-tillage system.

Table 8 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia seedbank persistence (percent viable seed) following a Weibull response function in response to length of seed burial across a single cohort of kochia seed using Equations 1 and 3.

Table 9 Weibull function parameters for lrc [ln(rate constant)] for each of the seed burial depths across 6 site-years for kochia populations with a single cohort.

a Abbreviations: Crop, cropland; NC, non-cropland; NT, no-tillage system; T, tilled system.

Seeds buried in the fall and exhumed in March (0.5 yr, within 6 mo) were viable, intact, and had high germination capacity, but there was some variation among sites-years and burial depths (Figures 4 and 5). This first exhumation period corresponds to the time when kochia typically emerges in the Great Plains and highlights why such large populations can be observed each year in the early spring.

The majority (>95%) of seed buried in these experiments did not persist for 2 yr (Figures 4 and 5). By the second exhumation period in October (nearly 1 yr after burial), remaining seeds were significantly less viable. There was much seed mortality over the first summer, as demonstrated by these low levels of viable seed. The share of remaining viable seed decreased through the last two exhumation dates (to 2 yr after burial), with <2% of seed remaining viable. Some site-years had no viable seed remaining in any of the burial packets, as at Garden City, KS, in 2010, or in six of nine cohort-by-burial depth packets in the irrigated environment at Fort Collins, CO, in 2010 (unpublished data). The range in percent viable seed also decreased with time.

The lack of a seed burial effect on kochia seedbank persistence in this study is not typical for most annual weed species. Burial of seed is often considered a mechanism for seeds to maintain viability and persist until returned to the soil surface for germination and emergence. These results did not show that burial depth increased or decreased seed persistence. Zorner et al. (Reference Zorner, Zimdahl and Schweizer1984) observed that initially nondormant kochia seed had slightly longer persistence when buried at 10, 15, or 30 cm, but after 1.5 to 2 yr of burial, 3% or less seed was still viable across all depths. Other studies have shown that even if kochia seed germinated at these depths, seedlings cannot emerge from greater than 5 cm in the soil (Schwinghamer and Van Acker Reference Schwinghamer and Van Acker2008; Zorner et al. Reference Zorner, Zimdahl and Schweizer1984).

Weed Management Recommendations Based on Germination Ecology of Kochia

An integrated kochia management approach is needed. Predictions of kochia emergence based on GDD as presented here could be used in determining timing of control practices as well as crop planting dates. Early and rapid kochia emergence may make this weed susceptible to a stale seedbed approach, where weeds are controlled prior to crop planting and emergence. For soil-applied herbicides to be effective, fall or very early spring pre-planting applications are required. Kochia can germinate and emerge under very low temperature and low-moisture requirements and, therefore, identifying herbicides that can be activated and effective with low-moisture conditions are critical (Sebastian et al. Reference Sebastian, Nissen, Westra, Shaner and Butters2017).

The first flush of kochia is usually very dense and can appear rapidly in late February and very early March in the central Great Plains. Soil- and foliar-active herbicides applied prior to crop planting and emergence may control a majority of the kochia seedlings, resulting in a much lower-density kochia population that would need to be controlled with subsequent soil-residual herbicides or selective herbicide applications after crop emergence. In many cases, more diverse herbicide options may be available for controlling weeds prior to crop emergence as compared with after crop emergence for crops such as corn (Zea mays L.), grain sorghum [Sorghum bicolor (L.) Moench], or soybean [Glycine max (L.) Merr.]. Few options are available for sugar beet (Beta vulgaris L.) production systems. Unfortunately, single and multiple herbicide–resistant kochia biotypes are already present across the Great Plains (Thompson et al. Reference Thompson, Thill and Shafii1994; Varanasi et al. Reference Varanasi, Godar, Currie, Dille, Thompson, Stahlman and Jugulam2015; Waite et al. Reference Waite, Thompson, Peterson, Currie, Olson, Stahlman and Al-Khatib2013). Consequently, there is a limited opportunity to diversify herbicide modes of action to minimize further selection for herbicide resistance or to control kochia biotypes already resistant or tolerant to herbicides that are available for a crop after it emerges.

The rapid emergence patterns may be exploited to develop effective cultural and mechanical management strategies. Early tillage may be effective, particularly given the results from this research indicating that kochia seed deeper in the soil that may be exhumed by tillage may have a very low percent viability. Tillage options, however, may be limited in many areas where kochia is common because of no-tillage cropping systems being used to conserve soil moisture in the Great Plains.

The very low dormancy and short persistence of kochia seed as demonstrated in this research suggests that it may be possible to quickly deplete seedbanks if kochia is aggressively managed using cultural and mechanical practices. Including rotational crops that are highly competitive with kochia may effectively reduce seedling survival and fecundity and thus help to reduce new additions to the seedbank (Esser Reference Esser2014) and permit crops with limited kochia control options to be added back into the rotation sequence. For example, monocot crops such as wheat (Triticum aestivum L.) and corn enable the use of some highly effective herbicide options, whereas dicot crops such as sunflowers, soybeans, sugar beet, and pulse crops. have fewer effective herbicide options, particularly after crop emergence. Including cover crops in the rotation sequence can sometimes be an effective component of an integrated weed management program. Spring-sown cover crops would be at a competitive disadvantage, since the crop and kochia are likely to emerge together (Petrosino et al. Reference Petrosino, Dille, Holman and Roozeboom2015). In the Great Plains, fall-established cover crop systems were shown to effectively suppress kochia density and reduce biomass, as the winter annual cover crops were able to develop an aggressive canopy earlier in the spring before kochia emerged (Petrosino et al. Reference Petrosino, Dille, Holman and Roozeboom2015).

Kochia has become a significant weed problem across the Great Plains because it has developed herbicide-resistant biotypes, emerges very early in the spring before other summer annual weed species, and disperses seed by tumbling across the landscape. PRE weed control, including tillage, herbicides, or cover crops, needs to be initiated in the fall or very early spring (February) to reduce or eliminate seedling emergence and thus reduce seed production. Short persistence of seed in the soil would assist in depleting seedbanks, and a community-wide effort to minimize kochia plant and seed movement would reduce the impact this weed species has in the Great Plains region.

Acknowledgments

Financial support was provided by the Monsanto Company, St. Louis, MO. Thank you to anonymous reviewers for their feedback. This is Contribution no. 16-287-J from the Kansas Agricultural Experiment Station.

Footnotes

a

Current address of eighth author: University of Wyoming, Powell, WY 82435

b

Current address of eleventh author: Field Scientist, Dow AgroSciences, Toronto, SD 57268.

Associate Editor for this paper: Adam Davis, USDA–ARS

References

Literature Cited

Anderson, RL, Nielsen, DC (1996) Emergence pattern of five weeds in the central Great Plains. Weed Technol 10:744749 Google Scholar
Baker, DV, Beck, KG, Bienkiewicz, BJ, Bjostad, LB (2008) Forces necessary to initiate dispersal for three tumbleweeds. Invasive Plant Sci Manag 1:5965 CrossRefGoogle Scholar
Becker, DA (1978) Stem abscission in tumbleweeds of the Chenopodiaceae: kochia. Am J Bot 65:375383 Google Scholar
Bell, AR, Nalewaja, JD, Schooler, AB (1972) Response of kochia selections to 2,4-D, dicamba, and picloram. Weed Sci 20:458462 Google Scholar
Blackshaw, RE, Larney, FJ, Lindwall, CW, Watson, PR, Derksen, DA (2001) Tillage intensity and crop rotation affect weed community dynamics in a winter wheat cropping system. Can J Plant Sci 81:805813 CrossRefGoogle Scholar
Borza, JK, Westerman, PR, Liebman, M (2007) Comparing estimates of seed viability in three foxtail (Setaria) species using the imbibed seed crush test with and without additional tetrazolium testing. Weed Technol 21:518522 CrossRefGoogle Scholar
Chepil, WS (1946) Germination of weed seeds. I. Longevity, periodicity of germination, and vitality of seeds. Sci Agric 26:307346 Google Scholar
Clay, SA, Davis, A, Dille, A, Lindquist, J, Ramirez, AHM, Sprague, C, Reicks, G, Forcella, F (2014) Common sunflower seedling emergence across the U.S. Midwest. Weed Sci 62:6370 CrossRefGoogle Scholar
Davis, AS, Clay, S, Cardina, J, Dille, A, Forcella, F, Lindquist, J, Sprague, C (2013) Seed burial physical environment explains departures from regional hydrothermal model of giant ragweed (Ambrosia trifida) seedling emergence in U.S. Midwest. Weed Sci 61:415421 CrossRefGoogle Scholar
Derksen, DA, Anderson, RL, Blackshaw, RE, Maxwell, B (2002) Weed dynamics and management strategies for cropping systems in the Northern Great Plains. Agron J 94:174185 Google Scholar
Dyer, WE, Chee, PW, Fay, PK (1993) Rapid germination of sulfonylurea-resistant Kochia scoparia (L.) Schrad. accession is associated with elevated seed levels of branched chain amino acids. Weed Sci 41:1822 CrossRefGoogle Scholar
Eberlein, CV, Fore, ZQ (1984) Kochia biology. Weeds Today 15:57 Google Scholar
Esser, AR (2014). Maternal Environmental Factors Influencing Kochia (Kochia scoparia) Seed Characteristics. M.S. thesis. Manhattan, KS: Kansas State University. 62 pGoogle Scholar
Forcella, F (1985) Spread of kochia in the northwestern United States. Weeds Today 16:46 Google Scholar
Friesen, LF, Beckie, HJ, Warwick, SI, Van Acker, RC (2009) The biology of Canadian weeds. 138. Kochia scoparia (L.) Schrad. Can J Plant Sci 89:141167 Google Scholar
Godar, AS (2014). Glyphosate Resistance in Kochia. Ph.D dissertation. Manhattan, KS: Kansas State University. 148 pGoogle Scholar
Godar, AS, Stahlman, PW (2015) Consultant’s perspective on the evolution and management of glyphosate-resistant kochia (Kochia scoparia) in western Kansas. Weed Technol 29:318328 Google Scholar
Heap, I (2017). The International Survey of Herbicide Resistant Weeds. www.weedscience.org. Accessed: March 21, 2017Google Scholar
Mengistu, LW, Christoffers, MJ, Lym, RG (2005) A psbA mutation in Kochia scoparia (L) Schrad from railroad rights-of-way with resistance to diuron, tebuthiuron and metribuzin. Pest Manag Sci 61:10351042 Google Scholar
Petrosino, JS, Dille, JA, Holman, JD, Roozeboom, KL (2015) Kochia suppression with cover crops in southwestern Kansas. Crop Forage Turfgrass Manag 1, 10.2134/cftm2014.0078Google Scholar
Phillips, WM, Launchbaugh, JL (1958) Preliminary studies of the root system of Kochia scoparia at Hays, Kansas. Weeds 6:1923 CrossRefGoogle Scholar
Primiani, MM, Cotterman, JC, Saari, LL (1990) Resistance of kochia (Kochia scoparia) to sulfonylurea and imidazolinone herbicides. Weed Technol 4:169172 Google Scholar
R Core Team (2016) R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org Google Scholar
Ratkowski, DA (1983) Nonlinear Regression Modeling: A Unified Practical Approach. New York: Marcel Dekke. Pp 135153 Google Scholar
Sawma, JT, Mohler, CL (2002) Evaluating seed viability by an unimbibed seed crush test in comparison with the tetrazolium test. Weed Technol 16:781786 Google Scholar
Sbatella, GM, Wilson, RG (2010) Isoxaflutole shifts kochia (Kochia scoparia) populations in continuous corn. Weed Technol 24:392396 Google Scholar
Schwinghamer, TD, Van Acker, RC (2008) Emergence timing and persistence of kochia (Kochia scoparia). Weed Sci 56:3741 Google Scholar
Sebastian, DJ, Nissen, SJ, Westra, P, Shaner, DL, Butters, G (2017) Influence of soil properties and soil moisture on the efficacy of indaziflam and flumioxazin on Kochia scoparia L. Pest Manag Sci 73:444451 Google Scholar
Stallings, GP, Thill, DC, Mallory-Smith, CA, Lass, LW (1995) Plant movement and seed dispersal of Russian thistle (Salsola iberica). Weed Sci 43:6369 CrossRefGoogle Scholar
Thompson, CR, Thill, DC, Shafii, B (1994) Growth and competitiveness of sulfonylurea-resistant and -susceptible kochia (Kochia scoparia). Weed Sci 42:172179 Google Scholar
Ullrich, SD, Buyer, JS, Cavigelli, MA, Seidel, R, Teasdale, JR (2011) Weed seed persistence and microbial abundance in long-term organic and conventional cropping systems. Weed Sci 59:202209 Google Scholar
Varanasi, VK, Godar, AS, Currie, RS, Dille, JA, Thompson, CR, Stahlman, PW, Jugulam, M (2015) Field-evolved resistance to four modes of action of herbicides in a single kochia (Kochia scoparia L. Schrad) population. Pest Manag Sci 71:12071212 CrossRefGoogle Scholar
Waite, J, Thompson, CR, Peterson, DE, Currie, RS, Olson, BLS, Stahlman, PW, Al-Khatib, K (2013) Differential kochia (Kochia scoparia) populations response to glyphosate. Weed Sci 61:193200 Google Scholar
Werle, R, Sandell, LD, Buhler, DD, Hartzler, RG, Lindquist, JL (2014) Predicting emergence of 23 summer annual weed species. Weed Sci 62:267279 Google Scholar
Wicks, GA, Martin, AR, Hanson, GE (1997) Controlling kochia (Kochia scoparia) in soybean (Glycine max) with postemergence herbicides. Weed Technol 11:567572 Google Scholar
Zorner, PS, Zimdahl, RL, Schweizer, EE (1984) Effect of depth and duration of seed burial on kochia (Kochia scoparia). Weed Sci 32:602607 Google Scholar
Figure 0

Table 1 Mean kochia density and observed range (low and high) for total emergence (plantsm−2) observed across sites and field types identified as cropland or non-cropland in 2010 and 2011.

Figure 1

Figure 1 Cumulative kochia seedling emergence, by day of year, for 17 site-years across the Great High Plains region of the United States. NC, non-cropland; C, cropland; NT, no-tillage system; T, tilled system

Figure 2

Figure 2 The site-year-specific predicted cumulative percent of emergence (dashed line) and fixed population predictions (solid line) by fitting nonlinear mixed-effects Weibull model of Kochia seedling emergence for 11 site-years in Kansas. GDD, growing degree days.

Figure 3

Table 2 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia cumulative seedling emergence following a Weibull response function to cumulative GDD across 11 site-years in Kansas.

Figure 4

Table 3 Weibull model parameters (Asym and Drop) for the predictive seedling emergence model, calendar date for 10% emergence, and estimated GDD for 10%, 50%, and 90% cumulative seedling emergence for each of 11 site-years across field types in Kansas.

Figure 5

Figure 3 The site-year-specific predicted cumulative percent of emergence (dashed line) and fixed population predictions (solid line) by fitting nonlinear mixed-effects Weibull model of Kochia seedling emergence for 6 site-years in NE and WY. GDD, growing degree days.

Figure 6

Table 4 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia cumulative seedling emergence following a Weibull response function to cumulative GDD across 6 site-years in Wyoming and Nebraska.

Figure 7

Table 5 Weibull function parameters (Asym and pwr) for the predictive seedling emergence model, calendar date for 10% emergence, and estimated GDD for 10%, 50%, and 90% cumulative seedling emergence for each of the 6 site-years in Wyoming and Nebraska.

Figure 8

Figure 4 The predicted percent viable seed (dashed line) and fixed population predictions (solid line) for the nonlinear mixed-effects Weibull model of kochia populations across site-years and burial depths with multiple cohorts. Site and year abbreviations: NEMit, Mitchell, NE; NESco, Scottsbluff, NE; CoIri, Colorado-irrigated; KSSto, Stockton, KS; CoDry, Colorado-dryland; KSGar, Garden City, KS; WYLin, Lingle, WY; 10, buried seed in 2010; 11, buried seed in 2011.

Figure 9

Table 6 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia seedbank persistence following a Weibull response function to length of seed burial across 10 site-years and multiple cohorts of kochia seed in Wyoming, Nebraska, Colorado, and Kansas.

Figure 10

Table 7 Weibull function parameters (Asym and lrc) for the predictive percent viable seed for each seed burial depth across 10 site-years and multiple cohorts in Wyoming, Nebraska, Colorado, and Kansas.

Figure 11

Figure 5 The predicted percent viable seed (dashed line) and fixed population predictions (solid line) for the NLME Weibull model of kochia populations across site-years and burial depths with a single cohort. Site-year abbreviations: KSha10, Hays, KS, 2010, non-cropland; KSha10c, Hays, KS, 2010, cropland; SDak10, Brookings, SD, 2010; SDak11, Brookings, SD, 2011; KSGa11T, Garden City, KS, 2011, tilled system; KSGa11, Garden City, KS 2011, no-tillage system.

Figure 12

Table 8 Summary of fixed and random effects for the most parsimonious nonlinear mixed-effects model of kochia seedbank persistence (percent viable seed) following a Weibull response function in response to length of seed burial across a single cohort of kochia seed using Equations 1 and 3.

Figure 13

Table 9 Weibull function parameters for lrc [ln(rate constant)] for each of the seed burial depths across 6 site-years for kochia populations with a single cohort.