Seeding, Herbicide, and Fungicide Impact on Perennial Grass Establishment in Cheatgrass Infested Habitats

Bromus tectorum (downy brome, cheatgrass) is a winter annual grass that infests millions of hectares of cropping, forage, and rangeland systems in western North America (Rice 2005). Bromus tectorum invasion into rangelands has been associated with a loss of desired perennial species’ seedbank and recruitment (Humphrey and Schupp 2004), but successful management has increased desired perennial species establishment (Else-road and Rudd 2011). However, niches created by control efforts may be filled by B. tectorum or other undesirable species present in the propagule pool (Svejcar et al. 2014). Consequently, attention has focused on integrating management strategies and revegetation (DiTomaso et al. 2010).

Revegetation of native species can fail because of pathogenic factors that affect seedling recruitment (Svejcar et al. 2014). Soil pathogens are ubiquitous and can have saprophytic or pathogenic effects on desired species’ seeds, thereby reducing their survival (O’Hanlon-Manners and Kotanen 2004). One such soil-borne pathogen is Pyrenophora semeniperda, a generalist grass pathogen. Pyrenophora semeniperda infects B. tectorum seeds in the seedbank (Beckstead et al. 2007), providing a useful management tool. However, P. semeniperda can also infect and kill co-occurring native grasses (Mordecai 2013) and research is need to evaluate if the non-target effects on these species can be mitigated.

Seed treatments such as fungicides have been suggested to improve desired species establishment (Jacobs et al. 1998), and they could protect seeded grasses from P. semeniperda mortality. While fungicides have successfully increased tree sapling survival in forests (O’Hanlon-Manners and Kotanen 2004), to our knowledge no such research has been reported in B. tectorum-invaded systems treated with P. semeniperda. Previous studies assessing integrated strategies to control B. tectorum have focused on integrating herbicides and revegetation (Sbatella et al. 2011, Orloff et al. 2015), with few studies incorporating other tools such as biological control (Dooley and Beckstead 2010, Ehlert et al. 2014), seeding rates, and seed treatments. Accordingly, our objective was to assess the potential of improving B. tectorum management by integrating the herbicide imazapic with P. semeniperda and varying rates of perennial grass seeding and a fungicide seed treatment.

This field study was conducted over two years (Year 1: September 2013 to August 2014, Year 2: September 2014 to August 2015) at a site located 10.5 km northwest of Amsterdam, MT, (45°45′ 25.088″ N, 111°25′ 35.5944″ E) where a naturally occurring B. tectorum infestation existed. Plots (1 m × 2 m) were established in fall 2013 as a randomized complete-block design with four blocks. A factorial combination of two herbicide treatments (non-sprayed, sprayed), two perennial grass seeding rates (recommended rate [1×], twice recommended rate [2×]), two P. semeniperda treatments (non-inoculated, inoculated), and two seedfungicide treatments (non-treated, treated) was randomly assigned to plots.

Imazapic [Plateau^®, 120 g ai * ha⁻¹] plus a non-ionic surfactant (Penetrator^®, Helena Chemical Company, 0.10% volume/volume) was applied to treated plots using a CO2 backpack sprayer delivering 210 L * ha⁻¹ water at 294 kPa. Non-treated plots received an equal amount of water. Application occurred November 2013 when B. tectorum was at the 4-leaf stage. Pseudoroegneria spicata (bluebunch wheatgrass) and Koeleria macrantha (prairie Junegrass) (Western Native Seed, Coaldale, CO) were hand seeded as a mix in April 2014 at the rate of either 13.5 or 27.0 kg * ha⁻¹ (1× or 2×) and 2.2 or 4.4 kg * ha⁻¹ (1× or 2×), respectively. Prior to seeding, difenoconazole and mefenoxam (Dividend Extreme^®, 12 g ai * 100 kg⁻¹ seed difenoconazole and 3 g ai * 100 kg⁻¹ seed mefenoxam) were applied to seeds receiving the fungicide treatment. Drill seeding was simulated with garden hoes (2.5 cm depth) for each of four rows planted within the center of each plot, with 30 cm row spacing.

Pyrenophora semeniperda was applied using a CO2 backpack sprayer delivering 250 mL * m⁻² of 5,000 conidia mL⁻¹ at 294 kPa in September 2013 and we collected soil samples from all plots in August 2014 and October 2015, approximately one and two years after. Fifty B. tectorum seeds per sample were collected with tweezers, visually inspected for P. semeniperda stromata, and scored as infected or noninfected. We sampled vegetation density July 2014 and June 2015 in each plot with two random samples using a 20 cm × 50 cm frame. Pseudoroegneria spicata and K. macrantha were not separated by species because of difficulty in distinguishing them at the seedling stage. In each sample, we measured the abundance of three functional groups (B. tectorum, seeded perennial grasses, and non-seeded perennial grasses [Hordeum jubatum, Elymus trachycaulus, Poa compressa]). At the 2015 sampling, we clipped aboveground biomass from the frames by functional group, dried it at 65°C for 72 hours and then weighed it.

We used the program R (R version 3.5.0, R Foundation for Statistical Computing, Vienna, Austria) to analyze data.

We tested P. semeniperda soil prevalence with a repeated measures logistic regression model using the function glmer in the lme4 package (Bates et al. 2015). We tested the effects of herbicide application, perennial grass seeding rate, fungicide application, year, and their interactions on vegetation density using a repeated measures analysis of variance (RM ANOVA). We tested the effects of herbicide application, seeding rate, fungicide, and their interactions on biomass collected in Year 2 with an ANOVA. When interactions occurred, means separations were conducted using the lsmeans package (Lenth 2016).

There were no differences in the proportion of B. tectorum seeds infected with P. semeniperda between inoculated and non-inoculated plots (Repeated measures logistic regression model, Chi-square = 0.0225, DF = 123, p = 0.88). Average infection of collected seeds was 22.1 ± 3.6% and 20.9 ± 3.3% in non-inoculated and inoculated plots, respectively. Consequently, data were pooled across the P. semeniperda treatments for density and biomass analyses. Biomass of B. tectorum, seeded perennial grass, and non-seeded perennial grass was unaffected by treatments (p > 0.05).

The effect of herbicide on B. tectorum density varied between years (RM ANOVA, F_1,56 = 5.164, p = 0.03). In Year 1, herbicide reduced B. tectorum density by 60% in sprayed plots compared to non-sprayed plots (Figure 1). In Year 2, B. tectorum densities were higher than Year 1 and similar between sprayed and non-sprayed treatments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Bromus tectorum density (tillers * m⁻²) as affected by herbicide and year. Black bars represent the non-sprayed treatment and gray bars represent the sprayed treatment. Error bars indicate the 95% confidence interval around the mean. Bars with different letters indicate means that are different from each other (α = 0.05).

Seeded perennial grass density was affected by herbicide across both years (RM ANOVA, F_1,53 = 20.316, p < 0.01) and was 25 ± 7 tillers * m⁻¹ in sprayed plots compared to 13 ± 6 tillers * m⁻¹ in non-sprayed plots. The effect of herbicide on non-seeded perennial grass density varied between years (RM ANOVA, F_1,56 = 7.156, p = 0.01). Non-seeded perennial grass density was similar between non-sprayed and sprayed plots in Year 1. However, in Year 2 density was lower with 30 ± 30 tillers * m⁻² in the non-sprayed plots compared to sprayed plots with 291 ± 202 tillers * m⁻², an order of a magnitude greater (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Non-seeded perennial grass density (tillers * m⁻²) as affected by herbicide and year. Black bars represent the non-sprayed treatment and gray bars represent the sprayed treatment. Error bars indicate the 95% confidence interval around the mean. Bars with different letters indicate means that are different from each other (α = 0.05).

Integrating multiple management approaches could help overcome common causes of revegetation failure, including unsuccessful control of the target weed, residual effects of herbicides, insufficient propagule pressure of desired species, and exposure of desired species to soil pathogens (Svejcar et al. 2014). In this study, imazapic and P. semeniperda were integrated with two rates of perennial grass seeding and a fungicide seed treatment as an approach to revegetate a B. tectorum infested site.

Imazapic efficacy on B. tectorum can be site and year specific (Orloff et al. 2015), and its effect on seeded perennial grasses may be affected by the relative timing of imazapic application and seeding of desired species. Degradation of imidazolinone herbicides is primarily due to microbial activity and is slowed by low moisture and temperatures (Prostko et al. 2005). The imazapic rate we used did not affect emergence of perennial grasses seeded 166 d after application and resulted in a two-fold increase in seeded perennial grass density. In contrast, in Idaho, Shinn and Thill (2004) observed injured perennial grasses that were sprayed 1 d after seeding with imazapic at 140 g ai * ha⁻¹. In Utah, Morris et al. (2009) applied four imazapic rates (35, 70, 105, and 140 g ai * ha⁻¹) to B. tectorum and seeded perennial plants 12 d later; successful B. tectorum control and increase in perennial grass establishment occurred at only one site with select rates. In Montana, Ehlert et al. (2015) found that fall-applied imazapic (80, 160, and 240 g ai * ha⁻¹) persisted into the spring and all rates similarly degraded 160 d after application. Our study indicated that imazapic applied at 120 g ai * ha⁻¹ likely degraded to concentrations non-lethal to perennial grasses by the time of seeding.

Plant competition may explain the absence of a seeding rate effect on perennial grass establishment. Mangla et al. (2011) studied intra- and inter-specific competition between invasive annual grasses (B. tectorum and Tae-niatherum caput-medusae [medusahead]) and two native species in greenhouse settings. Their results revealed that intraspecific competition in native species during initial growth stages contributed to smaller seedlings, probably because individuals shared similar resource requirements.

In our study, seedbed preparation and planting method may have contributed to successful establishment of perennial grasses. We simulated drill seeding using garden hoes and hand seeding, which allowed accurate seed depth and good seed-soil contact. James and Svejcar (2010) investigated different seeding techniques of desired grass and forb species and found seedling density increased from 12.2 ± 6.1 plants * m⁻² with a rangeland drill to 18.4 ± 4.6 plants * m⁻² with hand seedling. In our study, perennial grass densities of 13 to 25 tillers * m⁻¹ were achieved across herbicide treatments two years after seeding. Such densities are more promising than densities reported in other annual grass revegetation studies where drill seeding or broadcast seeding was implemented (Sheley et al. 2006, Mangold et al. 2015) and support the need for improvements in seeding technology (James and Svejcar 2010).

Perennial grass seeds were treated with a fungicide prior to planting, because emergence can by inhibited by soil-borne fungal pathogens. However, we observed no effect of the fungicide seed treatment, which may be because temperature and moisture conditions were favorable for perennial grass establishment and emergence. Similarly, Orrock et al. (2012) observed that a fungicide did not increase seedbank survival of two native species in the presence of an invader, Lonicera maackii (Caprifoliaceae). Land managers can therefore save resources by not treating seeds prior to planting, unless known pathogenic soil microbes exist.

We controlled B. tectorum for one year with a single imazapic application and there was strong perennial grass establishment two years after treatment. Future research should test P. semeniperda against B. tectorum from a range of environments and test seeding technologies to secure desired species establishment and B. tectorum control.