Data Collection The study site is located near Robb, Alberta. There are three blocks at this site, each with four plots where different simulated MPB attack treatments were applied (Fig 5). MPB attack was simulated by stem injections with glyphosate, which induced chemical girdling of the trees resulting in a pattern of mortality that closely resembles that following MPB attack. Treatments included: high mortality (100% of trees treated), medium mortality (30% of trees were treated; called 50% kill), simulated salvage logging operation (clearcut), and an untreated control (Fig 6).
Fig 5. The location of the study site (left) near Robb, Alberta and an overview of the three Blocks (NRH, PSH, HTL), each of which had 4 Treatment Plots (Salvage Logged, Control, 100% Simulated kill, and 50% Simulated kill; right)
Fig 6. Treatments included control plots (top left), medium mortality (top right), high mortality (bottom left), and salvage logged (bottom right).
Fig 7. The grid design of each of the twelve 0.48 ha treatment plots, with nine sampling points nested inside.
Each of the twelve treatment plots contain nine nested sampling points (Fig 7), each of which were sampled for understory vegetation, soil moisture, and percent canopy cover one year prior to disturbance (2008), the year treatments were applied (2009), one-year post-treatment (2010) during red attack stage, and a variety of years post-treatment (2012, 2013, 2014, and 2016). Percent canopy cover was recorded using hemispherical photographs and soil moisture was measured with a soil moisture probe (Fig 8). Percent cover of species was visually estimated in 1-m x 1-m quadrats for all vascular plant species and the common forest floor mosses.
Fig 8. A 1-m x 1-m vegetation quadrat used for visually estimating cover for understory vegetation species (left), a hemiphotograph used to determine canopy cover (middle), and a soil moisture probe used to determine Volumetric Water Content (right).
Analysis Repeated measure techniques account for the lack of independence in observations in the same locations over multiple years. Repeated measure mixed effect models were used to examine a variety of variables including vegetation richness, total percent vegetation cover, and percent cover for the four main functional groups (graminoids, bryophytes, forbs, and shrubs). Analysis of Variance (ANOVA) analyses were carried out on the repeated measure models to determine if the variance among data was due treatments and/or years. Pairwise lsmean post-hoc tests were used to further examine significant effects of treatments among each year, as well as the effects over time for each treatment.
A Principle Coordinate Analysis (PCoA) is an unconstrained distance-based method that can use either Euclidean or non-Euclidean distance measures to plot all points in relation to distances from each other. Distances are projected onto a Euclidean space in up to n-1 dimensions for n data points. The first point is placed at the origin, the second point is added along the first axis in relation to the first point, the third point is added in relation to the first two, and so on until all points are added. The points are rotated along the axes to maximize the variance from the n-1-dimensional space to create a two-dimensional ordination. The first axis contains the most variation possible, and each successive axis explains less variation than the previous one.
A PCoA was used to examine responses of understory vegetation to canopy changes that arose due to tree mortality induced by the treatments over time. Vegetation community composition for each treatment plot in each year was projected with a Curtis-Bray distance matrix and then plotted in the ordination. Trajectory arrows were calculated based on scores from the distance matrix and were included to show changes in vegetation composition over time. Vectors for functional vegetation groups (forbs, shrubs, graminoids, and bryophytes) and vegetation species were created and plotted to see how treatments and time impacted vegetation communities and individual species.