Introduction
Conservation agricultural systems—those that include residue retention, reduced tillage, and diverse crop rotations—have received considerable attention for their potential soil-improving benefits (Hobbes et al. Reference Hobbes, Sayre and Gupta2008; Reicosky Reference Reicosky2015). However, adoption of these systems is often constrained by the increased difficulty and expense associated with weed management (Hoyt et al. Reference Hoyt, Monks and Monaco1994; Luna et al. Reference Luna, Mitchell and Shrestha2012). Although many studies have characterized shifts in weed seedbank diversity and abundance under conservation agriculture (e.g., Brainard et al. Reference Brainard, Peachey, Haramoto, Luna and Rangarajan2013; Buhler Reference Buhler1995; Cardina et al. Reference Cardina, Regnier and Harrison1991; Gomez et al. Reference Gomez, Liebman and Munkvold2014; Pollard and Cussans Reference Pollard and Cussans1981), the mechanisms responsible for those shifts are poorly understood. These mechanisms may include changes in seed production and dispersal, as well as mechanisms of loss from the weed seedbank, including germination, predation, and decay. The persistence of weed seeds in the soil depends on their inherent resistance to fatal germination or death and on their exposure to environmental conditions that influence those fates (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015). Differences in seedbank persistence and their causes are challenging to quantify, but are important for predicting shifts in weed species communities and identifying efficient weed management strategies that target weak points in weed life cycles (Cousens and Mortimer Reference Cousens and Mortimer1995; Gomez et al. Reference Gomez, Liebman and Munkvold2014).
Conservation agricultural practices, including reduced tillage and cover cropping, may influence weed seed persistence through changes in soil biology that affect seed mortality (Gallandt et al. Reference Gallandt, Fuerst and Kennedy2004; Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015; Schafer and Chilcote Reference Schafer and Chilcote1970) and species-specific genetic differences in seed resistance to decay (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015). For example, Fennimore and Jackson (Reference Fennimore and Jackson2003) found shepherd’s-purse [Capsella bursa-pastoris (L.) Medik.] emergence and seedbank density were lower in treatments with higher levels of cover cropping and compost addition, and speculated that changes in microbial populations associated with higher levels of organic matter may have been responsible. In contrast, Davis et al. (Reference Davis, Anderson, Hallett and Renner2006) found that seed mortality of giant foxtail (Setaria faberi Herrm.) and velvetleaf (Abutilon theophrasti Medik.) was lower in long-term cropping systems with higher levels of cover crop and compost addition and that mortality of these species was associated with shifts in microbial community composition. However, Ullrich et al. (Reference Ullrich, Buyer, Cavigelli, Seidel and Teasdale2011) and Gomez et al. (Reference Gomez, Liebman and Munkvold2014) found only small and inconsistent differences in persistence of weed seeds in soils from long-term cropping systems that differed in organic matter inputs and soil microbial biomass. Such contrasting responses of weed seed persistence to broadly similar organic matter additions are not surprising, given the diversity of potential factors influencing persistence and the variability in those factors and their impacts across soil types, weather conditions, and weed species (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015).
Differences in weed seed persistence across cropping systems may be explained in part by differences in fungal communities that influence seed decay (Chee-Sanford et al. Reference Chee-Sanford, Williams, Davis and Sims2006; Davis et al. Reference Davis, Anderson, Hallett and Renner2006; Gallandt et al. Reference Gallandt, Fuerst and Kennedy2004; Gomez et al. Reference Gomez, Liebman and Munkvold2014; Kremer Reference Kremer1993; Pitty et al. Reference Pitty, Staniforth and Tiffany1987) and species differences in seed resistance to decay agents (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015). Many saprophytic and pathogenic fungi are present in the soil and isolated from seeds (Gomez et al. Reference Gomez, Liebman and Munkvold2014; Mitschunas et al. Reference Mitschunas, Filser and Wagner2009), and several studies have used soil sterilization or fungicide seed treatment to evaluate the role of microorganisms as a cause of seed mortality (Gallandt et al. Reference Gallandt, Fuerst and Kennedy2004). In a review of 14 such studies, fungicide treatments increased persistence in most cases, with results varying by study, weed species, and duration of burial (Wagner and Mitschunas Reference Wagner and Mitschunas2008). Several researchers have used this approach to evaluate whether differences in fungal decay varied based on tillage or cover crop practices. For example, Gallandt et al. (Reference Gallandt, Fuerst and Kennedy2004) detected no interaction between tillage and fungicide treatment on decay of wild oat (Avena fatua L.). Similarly, Kumar et al. (Reference Kumar, Brainard, Bellinder and Hahn2011) found that although fungicides increased overwinter persistence of Powell amaranth (Amaranthus powellii S. Watson) and barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.], there were no differences in these fungicide effects in buckwheat (Fagopyum esculentum Moench) cover crop treatments compared with the no cover crop control. On the other hand, higher populations of fungal pathogens such as Pythium, Rhizoctonia solani, Fusarium, and Thielaviopsis spp. were found following incorporation of cover crops (Dabney et al. Reference Dabney, Schreiber, Rothrock and Hohnson1996; Mohler et al. Reference Mohler, Dykeman, Nelson and Ditommaso2012; Rothrock et al. Reference Rothrock, Kirkpatrack, Frans and Scott1995; Toussoun et al. Reference Toussoun, Patrick and Snyder1963), and these have been associated with reductions in emergence (Mohler et al. Reference Mohler, Dykeman, Nelson and Ditommaso2012) or persistence (Davis et al. Reference Davis, Anderson, Hallett and Renner2006) of common weed species.
Differences in seed persistence across tillage and cover-cropping systems may also be strongly influenced by effects on dormancy loss and fatal germination (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015; Wagner and Mitschunas Reference Wagner and Mitschunas2008). For example, tillage-induced exposure of seeds to light, fluctuating soil temperatures, oxygen, and inorganic nitrogen may contribute to dormancy release and stimulation of germination of seeds that fail to successfully establish (Baskin and Baskin Reference Baskin and Baskin1998; Finch-Savage and Footitt Reference Finch-Savage and Footitt2017). Many short-term studies have shown lower rates of weed emergence under reduced-tillage systems (e.g., Brainard and Noyes Reference Brainard and Noyes2012; Hendrix et al. Reference Hendrix, Young and Chong2004; Wang and Ngouajio Reference Wang and Ngouajio2008), suggesting that—other things being equal—fatal germination may also be reduced in conservation tillage systems. However, this relationship is complicated by differences in the vertical distribution of seeds in the soil and other biological factors influencing recruitment following germination (Haramoto and Brainard Reference Haramoto and Brainard2017; Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015). Tillage redistributes seeds by depth in the soil, which alters light, temperature, and moisture conditions that affect seed dormancy, germination, and persistence (Long et al. Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015; Mohler Reference Mohler2001). Despite the well-established observation that seeds are more shallowly distributed in reduced-tillage systems, surprisingly few studies have evaluated the effects of tillage on germination, emergence, or persistence independent of depth of seed burial. Roberts and Feast (Reference Roberts and Feast1972) found that for a given depth of weed seed burial, the emergence of weeds was lower and the persistence of weed seeds higher in undisturbed compared with cultivated soils. However, when controlling for seed depth, Haramoto and Brainard (Reference Haramoto and Brainard2017) found that emergence of A. powellii was sometimes higher in undisturbed soil with surface cover crop residue, presumably due to higher moisture retention at the soil surface.
The importance of tillage-induced light exposure on germination of annual weeds is well known, but its relative importance in explaining differences in persistence across tillage systems has not been extensively explored in previous studies (Wagner and Mitschunas Reference Wagner and Mitschunas2008). Light exposure increases seed germination for many weed species (Baskin and Baskin Reference Baskin and Baskin1998; Wesson and Wareing Reference Wesson and Wareing1967), including southern crabgrass [Digitaria ciliaris (Retz.) Koeler] (Tang et al. Reference Tang, Hamayun, Khan, Shinwari, Kim, Kang, Lee, Na, Nawaz, Kang and Lee2010), common lambsquarters (Chenopodium album L.) (Milberg et al. Reference Milberg, Andersson and Noronha1996), large thornapple (Datura ferox L.) (Scopel et al. Reference Scopel, Ballaré and Sanchez1991), and redroot pigweed (Amaranthus retroflexus L.) (Gallagher and Cardina Reference Gallagher and Cardina1998a). As a consequence of this light sensitivity, cultivation or tillage events occurring at night result in lower rates of emergence for many important weed species (Botto et al. Reference Botto, Scopel, Ballaré and Sanchez1998; Buhler Reference Buhler1997; Fogelberg Reference Fogelberg1999; Scopel et al. Reference Scopel, Ballaré and Sanchez1991, Reference Scopel, Ballaré and Radosevich1994). However, the relationship between light-induced increases in emergence, fatal germination, and seed persistence is not well established.
Previous studies evaluating tillage effects on seed persistence rarely control for light exposure, so it is often unclear whether observed tillage effects are due to long-term changes in soil conditions influencing seed decay or predation or short-term changes in other factors—including light exposure—that strongly influence dormancy release, fatal germination, and persistence. For example, studies evaluating seed persistence in tilled systems using buried mesh bags often remove bags temporarily during tillage operations to avoid destruction of the seed bags (e.g., Hill et al. Reference Hill, Renner and Sprague2016; Ullrich et al. Reference Ullrich, Buyer, Cavigelli, Seidel and Teasdale2011). While often necessary, this approach may expose seeds to light, higher oxygen levels, and other stimuli that their ambient seed counterparts may not experience, depending on the timing and type of tillage used.
The objectives of this study were to evaluate the effects of tillage and cereal rye (Secale cereale L.) cover cropping on weed seed persistence and to investigate the extent to which these effects are mediated by fungal pathogens and exposure to light. We hypothesized that (1) weed seed persistence is influenced by historic cover-cropping and tillage practices; (2) tillage and cover crop effects on seed persistence are mediated in part by fungal pathogens; and (3) tillage effects on weed seed persistence depend on the level of light exposure experienced by those seeds during tillage. To test these hypotheses, persistence was first evaluated in a seed burial study (Experiment 1) within a long-term cropping systems trial with two levels of tillage (strip tillage vs. full-width tillage) and cover cropping (none vs. cereal rye). To better understand the possible mechanisms responsible for observed differences in persistence, we designed a second set of experiments (Experiments 2A and 2B) with more frequent exhumations and additional treatments to evaluate the potential role of fungal pathogens (Experiment 2A) or light (Experiment 2B) in mediating tillage and cover crop effects.
Materials and Methods
Long-Term Trial Experimental Treatments and Design
Three weed seed burial experiments (Experiment 1, Experiment 2A, and Experiment 2B) were conducted within a subset of treatments in a long-term tillage trial initiated in September 2008 on Oakville fine sand (mixed, mesic Typic Udipsamments) at the Southwest Michigan Research and Extension Center (SWMREC) in Benton Harbor, MI (42.085244°N, 86.358736°W). These treatments consisted of all combinations of two factors: tillage (strip tillage [ST] vs. conventional full-width tillage using a moldboard plow [FWT]) and cover crop (no cover crop [no cover] vs. cereal rye [rye]). Treatments were imposed in the same plots each year with crops following a 3-yr rotational sequence of sweet corn (Zea mays L.)–snap bean (Phaseolus vulgaris L.)–cucurbit crop (butternut squash (Cucurbita moschata Duchesne) in 2011 and pickling cucumber (Cucumis sativa L.) in 2014. Plots were arranged in a split-split plot design with tillage as the main plot factor and cover crop as the subplot factor. Tillage main plots measured 11.4 m by 18.3 m and were arranged in a randomized complete block design with four replications. Cover crop split plots were 3.8 m by 18.3 m with either 2 (winter squash) or 5 rows (all other crops) per plot.
Field Management
Cereal rye was drilled at 125 kg ha−1 in mid-September of each year using a grain drill with 19-cm between-row spacing. The following spring of each year, the entire experimental area was treated with glyphosate (0.84 kg ae ha−1) in mid-May to terminate both cereal rye and winter annual weeds before tillage. Immediately before glyphosate application, mean aboveground shoot dry weight of both rye and weeds was estimated from tissue sampled from two 0.25-m2 quadrats per plot. Dry shoot weight at the time of termination ranged from approximately 300 to 700 g m−2 for rye and 0 to 250 g m−2 for weeds depending on the year and treatment, with far lower weed biomass in rye treatments (see details in Brainard et al. 2016). Tillage occurred in late May or early June depending on the crop. FWT consisted of moldboard plowing followed by disking and field cultivating. ST was accomplished using either a Hiniker 6000 strip-tiller (for sweet corn, snap beans, and cucumbers) or an Unverferth 120 subsoiler (for winter squash). Both strip-tillage implements were equipped with a row cleaner (to remove cover crop residue), a shank, offset disks, and a rolling basket. ST resulted in an approximately 25-cm-wide by 30-cm-deep zone of disturbed soil into which crops were planted.
Experiment 1: Tillage and Cover Crop Effects on Seed Persistence
Experimental Design
This experiment evaluated the effects of cover crop (no cover vs. rye) and tillage (ST vs. FWT) on seed persistence of A. powellii collected in 2011 from agricultural fields at the Kellogg Biological Station in Hickory Corners, MI, and D. sanguinalis, collected in 2011 from SWMREC, Benton Harbor, MI. Seeds were separated from chaff, counted, and stored in glass jars in a laboratory at 20 C until the time of burial (~1 mo after seed collection). Before burial, seed viability (determined via germination with gibberellic acid [GA3]) was 84% and 92% for D. sanguinalis and A. powellii, respectively.
Seeds were buried following winter squash harvest in each cover crop subplot of the long-term trial described earlier (Table 1). Therefore, for each weed species, the design was a split-plot design with tillage as the main plot factor and cover crop as the subplot factor.
Table 1 Trial design and treatments.
Seed Preparation and Burial
Mesh bags (~8 cm by 10 cm) were constructed from 20D polyester “no-see-um” permeable mesh (Outdoor Wilderness Fabrics, 123 E Simplot Blvd, Caldwell, ID 83605, USA) using nylon thread. Bags were filled with 100 g of sand that had been sieved through a 500-micron sieve (smaller fraction kept) mixed with either 100 seeds of A. powellii or 200 seeds of D. sanguinalis. Silica sand was used for ease of subsequent seed separation and because it mimicked the soil texture (Oakfield fine sand; 94% sand; 1% to 2% soil organic matter) at our experimental site. One drawback of this approach is that buried seeds were not initially in direct contact with field soil and hence may have experienced different edaphic conditions than seeds in the ambient seedbank. To minimize such differences, we used a small volume of sand and flattened each bag such that seeds were in close proximity to ambient soil. Flattened bags were only 1.5-cm thick, so colonization by soil microbes was likely. Seed densities were chosen to roughly mimic the proportion of these species in the ambient weed seedbank. Bags were buried 15-cm deep in all four cover crop by tillage (FWT+no cover, FWT+ rye, ST+no cover, FWT+rye) combinations in November 2011 (Table 2). Seed bags in ST treatments were buried in the between-row zone and therefore were not subject to disturbance from the ST operation. Bags were buried by using a golf-cup hole cutter to first remove a cylindrical soil core (10.2-cm deep with an 11.4-cm diameter) and then restore the core into the hole.
Table 2 Timing of relevant field operations and experimental procedures.
Seed Retrieval and Viability Assessment
One bag of each species from each plot was exhumed in June 2012, May 2013, and June 2014 (7, 19, and 31 mo after burial [MAB]). Bags were stored at 4 C until viability testing. Seeds were removed from bags by sieving through a 500-micron sieve. Seeds of A. powellii were then germinated in petri dishes with 2 µM GA3, while D. sanguinalis seeds were germinated with deionized water. These seeds were placed in a 16-h day/8-h night growth chamber set at 30/25 C with incandescent bulbs that provided approximately 28 µmol m−2 s−1 of light. These treatments and conditions were chosen based on the stimulation of high germination rates in preliminary studies with nonburied seeds from the same seed lots. Ungerminated seeds were assessed for viability by a combination of squeeze testing with forceps and using a 0.1% 2,3,5-tretrazolium chloride (TZ) solution in accordance with methods outlined by the Tetrazolium Subcommittee of the Association of Official Seed Analysts (Peters and Lanham Reference Peters and Lanham2000). Seeds were characterized as germinated, dormant (did not germinate but were TZ viable), or dead; germinated and dormant seeds were considered viable.
Seed Reburial
In FWT treatments, all bags in each plot were removed from the soil before spring tillage and reburied following tillage and planting the next day. During this removal period, no attempt was made to exclude seeds from light, although seeds were mixed with sand, which may have limited light exposure. For FWT treatments with rye residue, bags were supplemented with rye residue that had been collected before termination, dried down, and coarsely ground into 5- to 10-mm-long segments. Each bag received a quantity of rye residue chosen to reflect actual dry matter yields produced in the field that year, divided by the assumed depth of incorporation. Using this approach, 0.30 g of residue was added per bag in 2012 and 0.13 g of dry rye residue was added per bag in 2013.
Statistical Analysis
The total number of viable seeds for each seed bag for each exhumation date t (N v,t ) was calculated according to the equation:
$$N_{{v,t}} {\equals}N_{{g,t}} {\plus}N_{{tz,t}} $$
where N g,t is the total number of seeds retrieved at a time t that germinated in the growth chamber postexhumation, and N tz,t is the total number of seeds retrieved at time t that did not germinate, but tested TZ positive. The persistence of seeds at a given exhumation date t (P i,t ) was defined as:
$$P_{{i,t}} {\equals}N_{{v,t}} \,/\,N_{{v,i}} $$
where N v,i is the total number of viable seeds initially buried (t=initial).
For the 7-, 19-, and 31-month exhumation dates, the effects of tillage and cover crop on N v,t and P i,t were analyzed using the PROC GLIMMIX procedure in Statistical Analysis System v. 9.4 (SAS Institute, Cary, NC). For these exhumation dates, data were analyzed as a split-plot design with tillage as fixed main plot factor, cover crop as fixed subplot factor, and replicate as a random effect. Where main or interactive effects were significant, treatment mean separation occurred using Fisher’s Protected LSD at α=0.05.
Experiment 2A: Tillage and Cover Crop Effects on Seed Persistence within a Year
Experimental Design
This experiment evaluated the effects of tillage (ST vs. FWT), cover crop (no cover vs. rye), and seed treatment (fungicide treated vs. untreated) on seed persistence of A. powellii collected in 2011 from Hickory Corners, MI, and D. sanguinalis collected in 2012 from Benton Harbor, MI. Both fungicide-treated and untreated seeds were buried in cover crop subplots of the long-term trial described earlier (Table 1). Therefore, for each weed species, the design was a split-split-plot design with tillage as the main plot factor, cover crop as the subplot factor, and fungicide as the sub-subplot factor.
Seed Preparation
During the winter of 2012, a subset of seeds of both species were coated with a triple-fungicide treatment used in previous studies to protect weed seeds against fungi including Rhizoctonia, Fusarium, Pythium, and Phytophthora (Kumar et al. Reference Kumar, Brainard and Bellinder2008, Reference Kumar, Brainard, Bellinder and Hahn2011). This coating contained captan (Captan Fungicide, 71 mg ai per 100 g seed, Southern Agricultural Insecticides), trifloxystrobin (Flint®, 10 mg ai per 100 g seed, Bayer CropScience), and metalaxyl (Apron XL® 350ES, 15 mg ai per 100 g seed, Syngenta Crop Protection). Untreated seeds were from the same seed lot but received no fungicide coating. Both treated and untreated seeds were stored in glass vials in a laboratory at 20 C (±2 C) until the initiation of the experiment. Despite the relatively long storage period under these conditions, seed viability at the time of Experiment 2 (determined via germination in petri dishes) was approximately 75% and 90% for D. sanguinalis and A. powellii, respectively (compared with 84% and 92%, respectively, in Experiment 1). Germination testing at this time also confirmed that fungicide treatment did not affect germination of D. sanguinalis . However, fungicide treatment increased A. powellii germination by approximately 11% in light, but had no effect in darkness (unpublished data).
Seed Burial
For each species and fungicide treatment combination, 100 seeds were mixed with 125 g of sieved silica sand and placed in no-see-um mesh bags. In October of 2015, after cover crop planting (Table 2), seed bags were buried at a depth of 10.2 cm in all four cover crop by tillage treatments (FWT+no cover, FWT+rye, ST+no cover, FWT+rye). Seed bags in ST treatments were buried in the between-row zone and therefore were not subject to disturbance from the ST operation. For bags containing seeds without fungicide treatment, 8 bags of each species were buried in each of the 16 cover crop subplots so that 2 bags per plot could be removed for each of four subsequent exhumation dates. For bags containing fungicide-treated seeds, 4 bags of each species were buried in each plot so that 2 bags could be pulled at two subsequent exhumation dates to assess fungicide effects. Only two exhumation dates were evaluated for fungicide effects because of the likely limited persistence of the fungicides themselves. Bags were buried by using a golf-cup hole cutter to first remove a cylindrical soil core (10.2-cm deep with an 11.4-cm diameter) and then restore the core into the hole. In treatments containing rye surface residue, the residue was carefully removed before burial and replaced after burial. In FWT treatments, all seed bags (other than those used in Experiment 2B) were retrieved in the morning before tillage, placed in paper bags, stored in a cooler at 4 C, and reburied that same afternoon after tillage operations were complete. This process was necessary to avoid disturbance of the seed bags during tillage. The day of tillage was partly cloudy with a high temperature of 26 C, and the duration of bag storage was 6 h.
Seed Retrieval and Viability Assessment
Two bags each of both fungicide and non–fungicide treated seeds were retrieved 1 and 4.5 MAB for D. sanguinalis, and after 1 and 7 mo for A. powellii. Additional sets of 2 bags of untreated seed were retrieved at 7 and 9 mo for D. sanguinalis, and 9 and 10.5 mo for A. powellii. Following the recommendation of Wagner and Mitschunas (Reference Wagner and Mitschunas2008), earlier retrieval dates for D. sanguinalis seeds were used to better characterize seed decay, which Experiment 1 demonstrated occurred more rapidly for D. sanguinalis than A. powellii. Upon retrieval, all bags were placed in cold storage at 4 C until being processed to evaluate seed viability. Seeds were separated from silica sand using a 500-micron sieve and tested for germination in either deioized water for D. sanguinalis or 2 mL 2 μM GA3 for A. powellii. These seeds were placed in a 16-h day/8-h night growth chamber set at 30/25 C with incandescent bulbs that provided approximately 28 µmol m−2 s−1 of light. After 2 wk, all ungerminated seeds were tested for viability as described for Experiment 1.
Statistical Analysis
The total number of viable seeds for each seed bag for each exhumation date t (N v,t ) was calculated according to Equation 1. The persistence of seeds at a given exhumation date t (P 1,t ) was defined as:
$$P_{{1,t}} {\equals}N_{{v,t}} \,/\,N_{{v,1}} $$
where N v,1 is the total number of viable seeds at the December exhumation time (t=1) and N v,t is the number of viable seeds at exhumation date t. Persistence was calculated using December as the initial time point, because seeds buried in October experienced warm soil temperatures that resulted in fatal germination that was likely not reflective of natural populations of A. powellii or D. sanguinalis (which would have had a higher level of dormancy at this time). Mean values of these responses from the 2 bags recovered from each sub-subplot were used for subsequent analysis.
For the 1-, 4.5- (D. sanguinalis only), and 7-month (A. powellii only) exhumation dates, the effects of tillage, cover crop, and fungicide treatment on N v,t and P 1,t were analyzed using the PROC GLIMMIX procedure in Statistical Analysis System v. 9.4 (SAS Institute). For these exhumation dates, data were analyzed as a split-split-plot design with tillage, cover crop, and fungicide treated as fixed effects, and replicate, replicate by tillage, and replicate by tillage by cover crop as random effects. For the 7- (D. sanguinalis only), 8.5-, and 10.5-mo (A. powellii only) exhumation dates, where fungicide treatments were not included, data were analyzed as a split-plot design with tillage and cover crops as fixed effects and replicate and replicate by tillage as random effects. Where main or interactive effects were significant, treatment mean separation occurred using Fisher’s protected LSD at α=0.05.
Experiment 2B: Role of Light on Tillage-mediated Seed Persistence
Experimental Design
To evaluate the potential impact of light on seed persistence in FWT treatments, two exhumation treatments were evaluated: (1) exhumation as described for Experiment 2A (ambient light); and (2) exhumation in darkness (no light) (Table 1). To accommodate these treatments, 4 additional bags of untreated A. powellii and D. sanguinalis seeds were buried in all four FWT+no cover plots to allow for subsequent evaluation of 2 bags per light treatment at one additional sampling date.
Exhumation, Storage, and Reburial
For seed bags without exposure to light, an opaque plastic box with gloved openings (Figure 1) was placed upside-down over the soil. The edges of the box were buried carefully to exclude all light while bags were excavated by hand and placed in sealed tins contained within the box. Seeds were stored in tins at 4 C for 8 h in the same location as seed exposed to light. Reburial was managed by again excluding light using the box.
Figure 1 The outside (A) and underside (B) of the opaque plastic box with gloved openings that was placed upside-down over the soil to facilitate seed bag exhumation without light exposure.
Retrieval and Viability Testing
All bags for this experiment were left in the field for 1.5 mo after reburial, at which point they were removed and tested for viability as described in Experiment 2A above.
Statistical Analysis
The effects of exhumation procedure on N v,t and P v,t were analyzed using PROC GLIMMIX in Statistical Analysis System v. 9.4 (SAS Institute) with exhumation procedure as a fixed effect and replicate as a random effect. The persistence of seeds from time period 3 (June) to 4 (July) was calculated according to:
$$P_{{3,4}} {\equals}N_{{v,4}} \,/\,N_{{v,3}} $$
where N v,3 and N v,4 are the total number of viable seeds at the June (t=3) and July (t=4) exhumation times, respectively. Persistence for Experiment 2B was defined based on the June time point to better evaluate seed persistence specifically after light exposure during the tillage event. Mean values of these responses from the 2 bags recovered from each sub-subplot were used for subsequent analysis.
Results and Discussion
Experiment 1: Tillage and Cover Crop Effects on Seed Persistence
The overwinter (7 MAB) persistence of A. powellii and D. sanguinalis was 77% and 20%, respectively, regardless of cover crop or tillage (Figure 2). By 19 MAB, A. powellii seeds in FWT treatments had only 5% persistence, while counterparts in ST treatments had 51% persistence, with no significant differences between cereal rye and the no cover treatment (Figure 2A). At 31 MAB, A. powellii seeds in FWT had only 0.6% persistence compared with 29% persistence in ST treatments. At 19 and 31 MAB, persistence of D. sanguinalis seeds was 15% and 2%, respectively, regardless of tillage or cover crop treatment (Figure 2B).
Figure 2 Effects of tillage (full-width tillage [FWT] or strip tillage [ST]) and cover crop (no cover or rye) from Experiment 1 on the persistence (mean±SE) of (A) Amaranthus powellii and (B) Digitaria sanguinalis seeds between November 2011 and June 2014. Persistence is defined as the total number of viable seeds on a given date (N v,t ) divided by the total number of viable seeds initially buried (N v,i ). Asterisks indicate significant differences between tillage treatments within each species: *P<0.05; **P<0.01. The effects of cover crop and cover crop * tillage interactions were not significant for either species at any sampling date. The main effect of tillage was significant only for A. powellii at the May and June sampling dates (P<0.05).
There are several possible explanations for the large effect of tillage on the persistence of A. powellii in this study. One explanation is that seeds in FWT treatments were temporarily removed from the field during tillage operations and were therefore exposed to stimuli (e.g., light, oxygen) that may have triggered fatal germination following reburial. In contrast, seeds in ST treatments remained continuously buried and did not experience such stimuli. Seeds of species closely related to A. powellii are well known for their germination response to light (Gallagher and Cardina Reference Gallagher and Cardina1998a, Reference Gallagher and Cardina1998b). In contrast, D. sanguinalis germination may be less affected by light, which may account for the lack of observed tillage effect on D. sanguinalis persistence, although the related species southern crabgrass [Digitaria ciliaris (Retz.) Koeler] is light sensitive in its germination response (Tang et al. Reference Tang, Hamayun, Khan, Shinwari, Kim, Kang, Lee, Na, Nawaz, Kang and Lee2010). In addition, tillage and cereal rye effects on D. sanguinalis seeds may have been difficult to observe in this experiment due to both the high variability and the rapid decline in D. sanguinalis viability that occurred before the first exhumation. An alternative explanation for reduced persistence of A. powellii under FWT in this study is that fungal pathogens responsible for seed decay were more prevalent in FWT treatments; this was tested in the subsequent experiment.
Experiment 2A: Tillage and Cover Crop Effects on Seed Persistence and the Potential Role of Fungal Pathogens
Effects of Tillage on Persistence
Consistent with Experiment 1, A. powellii seed persistence in Experiment 2A was influenced by tillage but not the cereal rye cover crop (Table 3). This tillage effect was only evident after at least 8.5 mo of burial and following removal and reburial to accommodate spring tillage in the FWT plots. The persistence of A. powellii seeds recovered in ST treatments did not differ between June and September following tillage. In contrast, the proportion of persistent A. powellii seeds in FWT treatments declined by approximately 60% during the same time period. Because the tillage effect on persistence did not occur until 8.5 MAB, and fungicide treatments were only evaluated at the 7-mo exhumation date, we could not evaluate the potential role of fungal pathogens in mediating this effect. This shortcoming could have been overcome by burying sufficient numbers of additional bags of fungicide-treated seeds to allow seed recovery and testing for a longer period, although the lack of persistence of the fungicides themselves over such longer periods limits the value of this approach (Kamrin Reference Kamrin1997; Krieger Reference Krieger2001). Another alternative would have been to bury fresh bags of untreated and fungicide-treated seeds in the spring and evaluate differences in seed persistence over a shorter period following tillage. However, interpretation of results from this approach would be complicated by the fact that the seeds had not undergone stratification in the field.
Table 3 Effects of tillage, cover crop, and fungicide treatment on the percentage of Amaranthus powellii seeds persisting at 7, 8.5, and 10.5 mo after burial (N v,t /N v,1*100) from Experiment 2A.
a na indicates that the given effect or interaction is not applicable for that date.
b FWT, full-width tillage; ST, strip tillage.
c P-values are statistically significant at the following α: +<0.10; *<0.05; **<0.01; ***<0.001; or not significant (NS).
Digitaria sanguinalis persistence was influenced by tillage (Table 4). As with A. powellii, the persistence of D. sanguinalis seeds was greater in ST compared with FWT treatments, but only following removal and reburial to accommodate spring tillage. Specifically, seed persistence declined by 63% between December and July in ST plots compared with 88% in FWT treatments.
Table 4 Effects of tillage, cover crop, and fungicide treatment on the percentage of Digitaria sanguinalis seeds persisting at 4.5, 7, and 8.5 mo after burial (N v,t /N v,1*100) from Experiment 2A.
a na indicates that the given effect or interaction is not applicable for that date.
b FWT, full-width tillage; ST, strip tillage.
c P-values are statistically significant at the following α: +<0.10; *<0.05; **<0.01; ***<0.001; or not significant (NS).
Our finding that seeds of both A. powellii and D. sanguinalis had greater persistence under ST compared with FWT is consistent with several other studies demonstrating greater persistence of summer annual weeds under no-tillage conditions compared with conventional FWT (Davis et al. Reference Davis, Cardina, Forcella, Johnson, Kegode, Lindquist, Luschei, Renner, Sprague and Williams2005; Roberts and Feast Reference Roberts and Feast1972; Steckel et al. Reference Steckel, Sprague, Stoller, Wax and Simmons2007). Although weed seed persistence and weed seedbank dynamics under ST are in theory different from those under no-till due to greater spatial heterogeneity (Brainard et al. Reference Brainard, Peachey, Haramoto, Luna and Rangarajan2013), in practice they may be similar, especially in the between-row zone where seeds were buried in our experiments.
Effects of Cover Crops on Persistence
We did not find a significant cereal rye cover crop effect on the persistence of A. powellii seeds (Table 3). There was, however, an effect of rye cover cropping on the proportion of persistent seeds of D. sanguinalis, independent of tillage (Table 4; no tillage by cover crop interaction). Contrary to our hypothesis, D. sanguinalis seed persistence was greater in cereal rye treatments compared with the no cover crop control at the March and June retrieval dates, although this effect had dissipated by July. For March and June sampling dates, D. sanguinalis viability was 2-fold greater in cereal rye compared with no cover crop treatments.
Our results are generally consistent with those of several other studies that also evaluated the impact of cereal rye on persistence of various summer annual weeds. Mohler et al. (Reference Mohler, Taylor, DiTommaso, Hahn and Bellinder2018) also found that the persistence of A. powellii was unaffected by rye cover crop residue. However, they found that in the grass species S. faberi, seed persistence (in one of two burial studies) was twice as high in the presence of rye compared with a bare-soil treatment. Similarly, Hill et al. (Reference Hill, Renner and Sprague2016) found that cereal rye increased seed persistence of S. faberi and A. theophrasti.
In our bare-soil treatments, the decline in viability of D. sanguinalis seeds occurred most rapidly between December and February, with little change occurring between February and June. In contrast, the viability of D. sanguinalis seeds in ST treatments was relatively stable until June, declining most rapidly between the June and July sampling dates. Digitaria sanguinalis seeds had greater overwinter persistence in rye treatments regardless of tillage, implying that cereal rye residue affected viability regardless of whether the residue was left on the surface (ST) or incorporated into the soil (FWT) during the previous growing seasons.
Effects of Fungicide Treatment on Persistence
Fungicide treatment did not affect overwinter persistence of either A. powellii (Table 3) or D. sanguinalis (Table 4). Findings from other studies on the interactive effects of tillage or cover cropping on fungal-mediated seed persistence are mixed. Pitty et al. (Reference Pitty, Staniforth and Tiffany1987) found greater fungal colonization of green foxtail [Setaria viridis (L.) P. Beauv] and S. faberi seeds at 0- to 7.5-cm depths of plowed plots as compared with reduced-tillage plots, presumably due to crop residue placement and organic matter accumulation. However, Gallandt et al. (Reference Gallandt, Fuerst and Kennedy2004) found no difference in fungal-mediated persistence of A. fatua in soils that differed in their historic tillage intensity.
We also found no evidence that the increased persistence of D. sanguinalis seeds in rye cover crop treatments (Table 4) was due to differences in fungal decay agents in rye versus bare-soil treatments. In particular, the effect of cereal rye on D. sanguinalis was independent of fungicide seed treatment. We cannot rule out the possibility that the fungicides used in our study were ineffective at providing protection from specific fungal pathogens, or strains of pathogens, responsible for decay of these species. When applied to soil, both captan and trifloxystrobin have reported half-lives ranging from only a few days to about 2 wk (Kamrin Reference Kamrin1997; Krieger Reference Krieger2001), while metalaxyl has a reported half-life ranging from 7 to 170 d (Kamrin Reference Kamrin1997). However, fungicides on seeds buried in the soil may have a longer half-life than those applied directly to the soil (Griffith and Mathews Reference Griffith and Matthews1969), and many previous studies have shown positive effects of fungicides on seed persistence over periods greater than 4.5 mo, including those relying only on captan (reviewed in Wagner and Mitschunas Reference Wagner and Mitschunas2008).
Given the lack of evidence for pathogen-mediated effects of rye on seed persistence, the mechanisms responsible for rye effects on persistence remain unclear. Perhaps rye changed soil edaphic conditions to disfavor decay agents of D. sanguinalis. For example, the presence of actively growing rye during the winter (regardless of tillage system) may have reduced soil moisture through transpiration, resulting in drier soil conditions that were less conducive to decay. Rye is also a well-known source of allelochemicals (Barnes et al. Reference Barnes, Putnam, Burke and Aasen1987) that may be toxic to decay agents of weed seeds. Another potential mechanism, suggested by both Hill et al. (Reference Hill, Renner and Sprague2016) and Mohler et al. (Reference Mohler, Taylor, DiTommaso, Hahn and Bellinder2018), is that cover crops with high C:N ratios, like cereal rye, reduce fatal germination of seeds by lowering their exposure to soil nitrate. Because nitrate is well known to stimulate germination of many weed species, including both A. powellii (Brainard et al. Reference Brainard, DiTommaso and Mohler2006) and D. sanguinalis (Gallart et al. Reference Gallart, Verdú and Mas2008), nitrogen immobilization following incorporation of high C:N cover crops may reduce fatal germination and hence increase persistence of these species.
Experiment 2B: Role of Light on Tillage-mediated Seed Persistence
Effects of Light on Persistence
In FWT treatments, seeds exposed to light during the spring-tillage exhumation event had a lower proportion of persistent seeds at a subsequent sampling date than seeds kept in darkness (Figure 3). After 1 mo of burial following tillage, A. powellii seeds that had been exposed to light had 48% lower persistence than those kept in darkness. Similarly, D. sanguinalis seeds exposed to light had 62% lower persistence than those kept in darkness.
Figure 3 Effects of tillage (full width tillage [FWT] or strip tillage [ST]) and light condition during exhumation on seed persistence (mean±SE) of Amaranthus powellii and Digitaria sanguinalis seeds 1.5 mo after tillage (8.5 MAB). Data are from Experiment 2B. Persistence is defined as the total number of viable seeds for each seed bag in July (N v,4) divided by the total number of viable seeds at the June exhumation time (N v,3). All seeds were placed within the no cover treatment. Different letters indicate significant differences between exhumation treatments within each species. Means were separated within each species using Fisher’s protected LSD at α=0.05.
These results suggest that light exposure that occurred in FWT treatments while seed bags were exhumed during tillage triggered subsequent fatal germination upon reburial, resulting in lower persistence. Petri dish germination trials supported this concept, with D. sanguinalis seeds showing a positive response in germination following exposure to light (unpublished data). However, the effect of light on A. powellii germination was less clear, with seeds tending to have higher germination in darkness when tested in a petri dish (unpublished data). Previous studies have shown that light exposure increases germination of Digitaria spp. (Tang et al. Reference Tang, Hamayun, Khan, Shinwari, Kim, Kang, Lee, Na, Nawaz, Kang and Lee2010) as well as species in the Amaranthus genus closely related to A. powellii (Gallagher and Cardina Reference Gallagher and Cardina1998a; Liebman et al. Reference Liebman, Mohler and Staver2001). The discrepancy between our petri dish and field observations for A. powellii may be explained by the fact that light sensitivity of Amaranthus species changes with burial and prechilling (as would occur during overwinter burial and stratification) (Wesson and Wareing Reference Wesson and Wareing1967, Reference Wesson and Wareing1969), perhaps due to seasonal phytochrome sensitivity (Gallagher and Cardina Reference Gallagher and Cardina1998a, Reference Gallagher and Cardina1998b; Taylorson Reference Taylorson1972).
We observed no difference in persistence between tillage types when we controlled for light exposure during FWT (Figure 3). This result suggests that light exposure was the only factor influencing tillage effects on seed persistence. In other words, long-term changes in soil conditions due to historic differences in tillage had no detectable effect on persistence of either species.
Our work has several important methodological implications for studies of this kind. First, to understand the influence of edaphic factors on seed persistence resulting from historic differences in management, it is important to control for short-term impacts of those practices. In the case of tillage, short-term exposure to light is clearly an important germination cue for many agricultural weed species and—as we have shown—may be the primary factor explaining tillage effects on persistence. However, mimicking the light exposure experienced by ambient weed seeds during tillage presents a challenge. Tracking the fate of known species and densities of seeds in mesh bags is a useful and well-established method for evaluating seed persistence, but removal of these bags during tillage—to minimize their destruction or loss—may result in light exposure that differs from that experienced by ambient seeds during tillage. Although many ambient seeds are likely to experience sufficient light exposure to trigger germination, others may be embedded in soil aggregates or insufficiently disturbed to experience such light cues (Pareja et al. Reference Pareja, Staniforth and Pareja1985; Terpstra Reference Terpstra1986). Our approach—removing and temporarily storing seeds in both light and darkness during tillage—provides a useful method for defining the range of plausible responses to light exposure.
With respect to cover crop effects on seed persistence, our results are consistent with several recent studies suggesting that residues with high C:N ratios may increase seed persistence relative to residues with low C:N ratios or bare soil (Hill et al. Reference Hill, Renner and Sprague2016; Mohler et al. Reference Mohler, Taylor, DiTommaso, Hahn and Bellinder2018). We found no evidence that cover crop–mediated changes in fungal pathogens were responsible for the observed increase in persistence of D. sanguinalis, but the mechanism and significance of this effect remains unclear. Future studies evaluating potential mechanisms responsible for the cover crop effects observed in this study would be valuable for understanding and managing D. sanguinalis, a major weed problem in both conventional and conservation agricultural systems.
In general, our results suggest that two key practices associated with conservation agriculture—reduced tillage and retention of crop or cover crop residue—may contribute to the persistence of important summer annual weed species and help explain observed increases in problematic species in these systems, including D. sanguinalis. Continued efforts to identify mechanisms responsible for these effects should help to identify complementary cover crop, tillage, or fertilization strategies that promote seed predation, decay, or fatal germination and lower weed management costs associated with these problematic weed species.
Acknowledgments
Funding for this work was provided by the North Central Region-Sustainable Agriculture Research and Education (NCR-SARE) program as well as by Michigan State University’s Project GREEEN (Generating Research and Extension to meet Economic and Environmental Needs). Many thanks to the NE-1026 project group for guidance and insight, as well as Alan Taylor at Cornell University for fungicide seed treatment. No conflicts of interest have been declared. We also acknowledge support from USDA National Institute of Food and Agriculture and Michigan State University AgBioResearch.
