Principal component analysis of weed communities
Fig 10. PCA of weed composition near study trees averaged across 2011 and 2012
A principal component analysis of the weed communities averaged through 2011 and 2012 and grouped by site visually shows the composition and abundance of weed species and their presence in each of my study sites. The points resulting from the analysis represent the vegetation community around each tree for the duration of the study, and also the size of the point is directly proportional to the amounts that the tree grew in stem volume. See the results in the ordination graph in Fig 10 at right.
Initial results of the ordination returned vectors for all species, but only five species had abundances high enough for the specie vectors to be of any significant length in the ordination - those species can be seen in the ordination figure.
From Fig. 10 we can first see that each of my sites has a distinct vegetation community. We can also see that the between the four sites that the "Back40" and "Homestead" sites are more productive than the "Farmyard" and "Rockpile" sites. Something that the unproductive sites also have in common with each other is a high abundance of quackgrass (Elytrigia repens). It is interesting to note however that in one area of the ordination we can observe a similiar vegetation community existing between the "Homestead" and "Rockpile" sites but the tree growth is drastically different. This suggests that another factor, such as microclimate or soil resources, may be influencing the ability of the trees to grow and compete with their neighbors.
Initial results of the ordination returned vectors for all species, but only five species had abundances high enough for the specie vectors to be of any significant length in the ordination - those species can be seen in the ordination figure.
From Fig. 10 we can first see that each of my sites has a distinct vegetation community. We can also see that the between the four sites that the "Back40" and "Homestead" sites are more productive than the "Farmyard" and "Rockpile" sites. Something that the unproductive sites also have in common with each other is a high abundance of quackgrass (Elytrigia repens). It is interesting to note however that in one area of the ordination we can observe a similiar vegetation community existing between the "Homestead" and "Rockpile" sites but the tree growth is drastically different. This suggests that another factor, such as microclimate or soil resources, may be influencing the ability of the trees to grow and compete with their neighbors.
Regression tree analysis of weed composition, abundance, and proximity
To get a further picture of the relationship between the characteristics of the neighboring weed community and tree growth I performed a regression analysis using package Rpart with total stem volume growth as the response with weed cover values as the predictor variable. Each weed species was classified into its functional group, which was then further categorized by quadrat and year. Included in the analysis was site ("SITE") and initial stem volume size ("INITIAL"). Also included are the results of variable importance returned by the 'summary' command. See Fig 11 and 12. Alternative splits are presented below the regression tree
Alternative splits for multiple regression tree
|
Discussion of regression tree results
Only a single regression tree is shown here because additional trees separated by sites did not yield additional information from what is already presented in Fig 11.
It is interesting to note that before anything else that site is our most important variable. This suggests that other factors (most likely abiotic) have a stronger influence on tree growth than anything else that was measured in this study. Of the measured biotic factors we still do see a pattern resulting from the regression analysis and variable importance results strongly implies that perennial grasses in quadrat 1 have a significant inhibitory effect on tree growth. Contradictory, we see that any presence of perennial grass in one of our productive sites (Back40 or Homestead) and located in quadrat 3 - which is about one meter away from the tree bole - is associated with an increase in tree growth. This can either be a form of facilitation or is simply correlation with the presence of perennial grasses being connected to a resource pool of increased soil water or nutrients and thus the nearby tree is connected to those resources as well.
Additional implications that can be gathered from the results suggest that annual forbs have a moderate impact on tree growth, where perennial forbs have very little influence. What appears to be more important than both of them though is the initial size of the tree as the start of the study which has a positive relationship with growth over the next two years regardless of the neighboring vegetation.
It is interesting to note that before anything else that site is our most important variable. This suggests that other factors (most likely abiotic) have a stronger influence on tree growth than anything else that was measured in this study. Of the measured biotic factors we still do see a pattern resulting from the regression analysis and variable importance results strongly implies that perennial grasses in quadrat 1 have a significant inhibitory effect on tree growth. Contradictory, we see that any presence of perennial grass in one of our productive sites (Back40 or Homestead) and located in quadrat 3 - which is about one meter away from the tree bole - is associated with an increase in tree growth. This can either be a form of facilitation or is simply correlation with the presence of perennial grasses being connected to a resource pool of increased soil water or nutrients and thus the nearby tree is connected to those resources as well.
Additional implications that can be gathered from the results suggest that annual forbs have a moderate impact on tree growth, where perennial forbs have very little influence. What appears to be more important than both of them though is the initial size of the tree as the start of the study which has a positive relationship with growth over the next two years regardless of the neighboring vegetation.
Analysis of vegetation control treatments
Fig 12. The effects of site interact with the effects of treatment
The first test required in this analysis was a test of covariance to determine if my fixed effect (TREATMENT) had an interaction effect with my random effect of SITE. From Fig 12 at right you can see that they do and hence further analysis must proceed as a one-way ANOVA on a site-by-site basis. Fig. 13 below has the final results for the vegetation control methods separated by site.
The results of the vegetation control part of the study appear to suggest that the business-as-usual approach of management is the most effective in controlling competing weeds and thus allowing the trees to grow more. However, early treatments of spot herbicide to weeds near to the tree bole was also effective in controlling competing vegetation and may be cheaper for the plantation manager to implement which could lead to greater profits even if tree growth is not as great as in the business-as-usual management approach. Additionally, the business-as-usual approach was quite similar in effectiveness to the other treatments in the Back40 site where competing vegetation was already limited in abundance and composed primarily of perennial forbs - which we see from previous analyses is not very competitive with the trees to begin with. This suggests that less intensive vegetation control techniques can be implemented on naturally productive sites with a low competition index without a significant loss in tree biomass.
To conclude: the results of this study provide interesting insight into the competition dynamics in hybrid poplar plantations and can be used to improve existing weed management plants. However, these results were only observed in juvenile stands and cannot be inferred to apply to all poplar plantations at all ages. To more fully understand how the ecological interactions between weeds and trees shifts as a plantation matures would require a longer term observation study with collection of abiotic data in addition to biotic assessments.
To conclude: the results of this study provide interesting insight into the competition dynamics in hybrid poplar plantations and can be used to improve existing weed management plants. However, these results were only observed in juvenile stands and cannot be inferred to apply to all poplar plantations at all ages. To more fully understand how the ecological interactions between weeds and trees shifts as a plantation matures would require a longer term observation study with collection of abiotic data in addition to biotic assessments.