Data
Data table processing
The raw data used for this project was collected in the field and entered into a combined database with separate sections for each plot recorded. Therefore, some processing needed to be completed in order to combine the data into one table suitable for analysis (Table 4). R Statistical Software (v4.3.1; R Core Team, 2021) was used to reformat and analyze the data. "ID" indicates the opening number, plot number, and location (north or south) of each 1 m2 quadrat used for plant identification (Table 4). "TREATED" indicates how many years have elapsed since herbicide treatment was applied. "SUBREGION" contains the subregion code for each cut-block. The "SHANNON" and "SIMPSON" columns are Shannon's and Simpson's diversity indices that were calculated using the percent (%) cover data of each species identified within the quadrat. Shannon's index focuses on the presence of rare species, whereas Simpson's index emphasizes dominating cover groups, both of which can provide valuable information regarding overall species composition. Percentage (%) of any water, wood, forest floor (FF), or rock/mineral soil (ROCKSOIL) that was present was also recorded for each quadrat. "SLOPE" and "ASPECT" were recorded at each plot. All columns following these indicate % cover data for each species identified in the quadrat.
Table 4. First few rows and columns of the simplified data table used for analysis.
Exploratory graphics
The % cover data for my project is heavily zero-inflated (Figure 6), so a distance-based ordination was used to analyze the plant communities. Using an NMDS ordination with Bray-Curtis distance (Figure 7) did not separate the data points very well, making it difficult to discern shifts in plant communities over time, so instead a PCoA was used for my results (Figure 8). The Bray-Curtis distance was used to build the PCoA because it is better suited to comparing plant communities.
Figure 6. Examples of histograms for a few of the species identified. Data is heavily zero-inflated as not every species was found in every plot.
Figure 7. Test of an NMDS using Bray-Curtis distance that did not separate my groups enough to see overall shifts in plant community.
The initial result of the PCoA had many overlapping arrows, making it hard to determine the overall trajectory of each plant community from year 1 to year 2 (Figure 8). To make the differences easier to see, I split the communities up by subregion, and added points for the untreated plots.
Figure 8. Initial graph created using the PCoA ordination. Arrows indicate direction the plant communities are moving from 1 year after treatment to 2 years after treatment ("0" indicates untreated plots), vectors indicating species of interest have not been added yet at this stage. Arrows are coloured by subregion. "CM" is central mixedwood, "LF" is lower foothills, and "UF" is upper foothills. Only subregions with corresponding untreated reference plots are shown in this figure.
In order to examine the effect of subregion on glyphosate residual concentration I also created a figure to indicate how many plant samples contained glyphosate one year after treatment and two years after treatment (Figure 9 and 10). Interestingly, in the central mixedwood subregion the number of samples containing glyphosate residuals was greatly reduced from years one to two (Figure 9).
Figure 9. Percentage of samples that contained glyphosate residuals one year after treatment and two years after treatment in each subregion.
Figure 10. Percentage of samples that contained glyphosate residuals one year after treatment and two years after treatment for each species of interest.
I also needed to examine the occurrence of species in each subregion to determine which species of interest were recovering more frequently than others (Table 3). This table will be converted into a frequency figure for the final version of my project.
Table 5. Frequency of occurrence of species in each subregion. Frequency is shown here as a percent (%) of total plots in each subregion. "0" indicates untreated sites. "1" and "2" represent one year after herbicide application and two years after respectively. Green highlights the species that are more frequent in the untreated or year one plots. Blue highlights the species that were more frequent after the two-year recovery period. Orange highlights the species that did not change in frequency from year one to year two. "CM" = central mixedwood. "DM" = dry mixedwood. "LBH" = lower boreal highlands. "LF" = lower foothills. "SA" = subalpine. "UF" = upperfoothills.
