Data exploration
Data was explored by investigating environmental variables and PLFAs across the four site types. Aspen canopies were present in natural deciduous dominated sites (NAT.DDOM), clear-cut deciduous dominated sites (CC.DDOM), and clear-cut coniferous dominated sites (CC.CDOM) while white spruce canopies were present in the natural coniferous dominated (NAT.CDOM) site type.
Moisture and pH can affect soil microbial community composition. Moisture content was similar across all site types with NAT.CDOM sites having the highest median moisture content on the sampling date (Figure 9). NAT.CDOM sites had the greatest variability in moisture content, indicated by the whiskers in Figure 9. Median pH was higher in deciduous dominated stands compared to coniferous dominated stands regardless of clear-cutting. Even though clear-cut spruce stands currently have an aspen canopy, there appears to be a legacy effect of spruce in terms of pH. In both deciduous and coniferous dominated stands, pH was higher in clear-cut compared to natural stands.
Site leaf area index (LAI) is a measure of vegetation density and could be related to microbial community composition. Site LAI was highest in clear-cut stands, with CC.DDOM having highest median LAI, and lower in natural stands (Figure 10). Harvest treatment appears to affect site LAI more strongly than stand type; however, the difference between clear-cut and natural stands was greater for deciduous dominated stands than coniferous dominated stands.
|
Carbon and nitrogen are important elements in soil. Figure 11 shows that total carbon was similar across the four site types. In terms of total nitrogen, it appears that NAT.CDOM sites had less soil nitrogen than the other site types. Microorganisms may be more nitrogen limited in natural spruce stands. Because microbes degrade soil organic matter to access nutrients, specifically nitrogen, organic matter decomposition may be occurring more readily in NAT.CDOM soils. PLFA results will be critical in determining the likelihood of this process. For example, if NAT.CDOM has a higher proportion of fungi and gram negative bacteria, soil organic matter decomposition has the potential to be greater than under other site types. This would be an interesting result because as white spruce are replaced by aspen with climate change, NAT.CDOM sites will become less prevalent.
Figure 11. Bargraphs comparing average total carbon (C) in g/kg and average total nitrogen (N) in g/kg in the bulk soil for different site types (CC.CDOM = clear-cut, coniferous dominated; CC.DDOM = clear-cut, deciduous dominated; NAT.CDOM = natural, coniferous dominated; NAT.DDOM = natural, deciduous dominated). Error bars represent one standard deviation from the mean.
Analyzing PLFA data required non-parametric statistics. Lab analysis identified 110 PLFAs; PLFA 16:1 w3c was selected as representative. Data for PLFA 16:1 w3c is not normally distributed (Figure 12) and transformations to make the data normal were unsuccessful because of the large amount of zero values within many of the PLFAs observed means. For all analyses, normality was checked by plotting residuals and non-parametric statistical analyses were used to explore the data for significant relationships.
Figure 13 and Table 2 show that PLFAs in groups 5, 7 and, 9 were quite common across site type and sample type. NAT.CDOM.RHIZ and NAT.CDOM.BULK1 appear similar to each other, but very different from the other site types. This makes sense, because NAT.CDOM was the only site with a spruce canopy (other site types had an aspen canopy). The cluster tree (Figure 13) indicates that for CC.CDOM, CC.DDOM, and NAT.DDOM site types, PLFA communities differed depending on whether samples were taken from the rhizosphere or bulk soil.
Figure 13. Heatmap showing relative proportions of phospholipid fatty acids (PLFAs) on the x axis in each site type and sample type (y axis). Site types are as follows: CC.CDOM = clear-cut, coniferous dominated; CC.DDOM = clear-cut, deciduous dominated; NAT.CDOM = natural, coniferous dominated; NAT.DDOM = natural, deciduous dominated. Sample types are rhizosphere (RHIZ) or bulk soil (BULK1). PLFAs have been scaled so that rare PLFAs are considered as important as non-rare PLFAs.
Table 2. Phospholipid fatty acids (PLFAs) grouped based on heatmap results (Figure __) with site type in which the group was relatively important indicated below. Site types are as follows: CC.CDOM = clear-cut, coniferous dominated; CC.DDOM = clear-cut, deciduous dominated; NAT.CDOM = natural, coniferous dominated; NAT.DDOM = natural, deciduous dominated. Sample types are rhizosphere (RHIZ) or bulk soil (BULK1).
Figure 14 is a MetaMDS. The vectors were obtained by summing the PLFAs from Table 2 for each sample. Comparing Table 2 with Figure 14 shows that the heatmap was effective in defining groups of PLFAs associated with site types. However, because the PLFA groups are so variable in terms of the types of PLFAs present (eg. saturated, straight chained), it will be more useful to associate PLFAs with specific microbial groups. Investigating microbial groups will help answer whether changes in microbial communities with vegetation shifts will affect carbon flux, as fungi and gram negative bacteria are important for this process.
Figure 14. MetaMDS ordination of microbial community composition. Each point represents the microbial community of a specific site, as determined by phospholipid fatty acid (PLFA) analysis. Site types are as follows: CC.CDOM = clear-cut, coniferous dominated; CC.DDOM = clear-cut, deciduous dominated; NAT.CDOM = natural, coniferous dominated; NAT.DDOM = natural, deciduous dominated. Rhizosphere are bulk samples are indicated by different shapes. Th vectors correspond to the PLFA groups determined through heatmap analysis.