1. Does fire severity affect the regeneration of aspen stands?
Overall, there tends to be greater aspen regeneration in stands that experienced moderate severity fires, however there is a lot of variability (Figure 1). In order to tell whether or not there was a significant difference an ANOVA was performed on the aspen regeneration data. Before executing the ANOVA the data needed to undergo a square root transformation in order to satisfy the assumptions of ANOVAs, homogeneity of variance and normally distributed data. Figure 1 shows us that there is not homogeneity among the variances across the different fire severities. The top left histogram in figure 2 shows that the aspen data did not have a normal distribution, as it was skewed to the left. The data went through 3 different transformations (log, squared, and square root) in order to see which one satisfied the ANOVA assumptions best (Figure 2). After running the ANOVA it was found that there was no significant differences between the amount of aspen regeneration between the different fire severities (p>0.05).
Figure 1. Number of aspen seedlings regenerating after different levels of fire severity, in previously pure aspen stands.
|
Figure 2. Histograms of different data transformations on the response variable of the number of aspen seedlings regenerating.
|
Figure 3. Residual plot and q-q plot of square root transformed responding variable (number of aspen seedlings regenerating) relating to the fire severity classes.
2. Does habitat type (i.e. bog or upland) have an effect on regeneration of black spruce stands?
From a basic boxplot (Figure 3) we can see that aspen regenerated more on upland black spruce sites, compared to bogs, which was expected. However, when it comes to black spruce regeneration, the habitat type doesn't seem to make much difference (Figure 3). In order to test for significance the same steps were taken to check the black spruce response variable as were taken to check the aspen response variable (above). The black spruce data was transformed with a square root transformation in order to meet assumptions. After running a MANOVA it was found that there was only a significant difference between the habitat types for aspen regeneration, with upland habitats producing significantly more aspen seedlings (p<0.01).
|
Figure 4 Number of black spruce and aspen seedlings growing after fire on different habitat types, in previously pure black spruce stands.
|
3. Do stand characteristics (i.e. age, height, basal area, competition) have an effect on regeneration in aspen and black spruce stands?
After creating scatter plots comparing the amount of regeneration to pre-fire stand characteristics, such as age, height, and basal area, it seemed none had an effect. ANOVAs were used to confirm this and none of the characteristics had a significant effect (p>0.05). However, the total percent understory cover does seem to have a negative effect on the amount of aspen regeneration (Figure 4). This was confirmed with an ANOVA and was found to be significant (p<0.01). This competition effect, however, has an opposite effect in the black spruce stands (Figure 5). Increased levels of total understory percent cover corresponds to increased black spruce regeneration (p<0.01).
Figure 5. Comparisons of the number of aspen regenerating across a range of understory percent cover levels.
Figure 6. Comparisons of the number of black spruce regenerating across a range of understory percent cover levels.
4. Does fire severity and habitat type have an effect on the regeneration of ground cover vegetation?
Overall, fire severity and habitat type don't seem to produce much difference in understory vegetation (Figures 6 & 7). In order to test for significance, a MANOVA was performed for each stand type. Assumptions were checked using the boxplots and histograms, and it was determined that a square root transformation was needed in order to create homogeneity among variances. The results only showed a difference between tree percent cover among the fire severities (p<0.001). While there was a difference in tree, shrub, and bryrophyte percent cover among the different black spruce habitat types (p<0.05).
Figure 7. Amount of vegetation functional groups present after different fire severities, in previously pure aspen stands.
|
Figure 8. Amount of vegetation functional groups present in different habitat types, in previously pure black spruce stands.
|