Methods
Site Description
In 2018, field experiments were established at two sites near the Alberta-Montana border in the Dry Mixed Grass Natural Subregion, AB (Figure 1). Site 1 was near Aden, AB on Ross Ranch and Site 2 was approximately 20 km away in the Pinhorn Provincial Grazing Reserve, AB. The soil types of the sites have not been documented yet.
Figure 1. Map indicating study sites.
Experimental Design and Treatments
A randomized complete block design was implemented at each of the two sites, with four reps at each site, each of which contained six herbicide treatments and two nontreated control plots in a factorial (randomized) design (Figure 2). Each plot was 7 x 3 m in size and the site was fenced from grazing (Figure 3). The 6 herbicide treatments compared seasonality (spring or fall application) and rate [40 g ai/ha (0.5x recommended rate), 80 g ai/ha (1x recommended rate), and 160 g ai/ha (2x recommended rate)] of indaziflam.
Figure 2. Representative site diagram of site 1.
Figure 3. Site 1 exclosure with plots inside.
Indaziflam treatments were applied with a Kubota RTV 1100 four-by-four vehicle using TeeJet AIXR 11003 nozzles at 275 kPa (Figure 4). Fall treatments were applied October 16, 2019 and spring treatments April 9, 2020.
Figure 4. Indaziflam being sprayed on plots in the fall of 2019.
Vegetation Assessment
At each site, plant species composition was assessed by visually estimating the foliar cover of every plant species in three 0.25 m2 quadrats per plot at the 1 m, 2 m and 4 m mark on the west side of the transect. Measurements were taken prior to treatment in early July of 2019, and then post treatment in late June of 2020, 2021, and will be taken in 2022.
Figure 5. Estimating species cover on the right side of transect.
Figure 6. 0.25m² quadrat placed on transect. Lots of sandberg's bluegrass (Poa sandbergii) tufts in this plot, one of the dominant native perennial grasses at these sites.
Statistical Analysis
Analyses were separated by year after treatment (year as a repeated measure). The data from the 2 sites were pooled for analyses.
Factors affecting brome cover were assessed with controls removed using mixed model ANOVA, with rate and time as fixed effects, and site and replication as random effects, and brome cover in the pretreatment year as a covariate.
Pairwise comparisons were conducted using a Tukey test with a bonferroni correction in the package multcomp.
1 sided t-tests were done using the glht function.
NMDS was used to visualize the effect of fall rates on plant composition. Due to zero inflated data the bray-curtis distance matrix was used. Using a vector scaling cut-off value of r>0.2, species frequencies and other variables that were not part of the ordination (shannons diversity (H), simpsons diversity (1-D), richness, perennial grass functional group, rate of indaziflam) were overlaid on the final ordination.