Methods

Site selection

In May 2023, a wildfire occurred approximately 15.7 km northeast of Fox Creek, Alberta. The fire generated a gradient of burn severities across the landscape. We selected sampling sites based on burn severity rather than fixed plot dimensions. Burn severity was classified using a combination of remote sensing (to assess crown damage) and ground-based assessments. Sites were sufficiently large to accommodate the target number of trees per species and severity class (see below). These methods were adapted from the Fire Effects Monitoring and Inventory System (FIREMON) (Lutes et al., 2006).

Sampling design

In September of the same year, we sampled three tree species across three severity classes: lodgepole pine (Pinus contorta), white spruce (Picea glauca), and trembling aspen (Populus tremuloides). This sampling was repeated the same time of the following year (2024).

Per site we sampled:

|----> species (pine/ spruce/ aspen)

|-------------->severity (control/ low/ medium/ high), except no control in 2024

|-------------------------------->n=9 trees

|-----------------------------------------------> sampling from cross-sections at 2 vertical heights (1.3 and 12m) n=18

Figure 1. Satellite map of the wildfire area and sampling design. Red dots indicate sampling sites. Sampling years happened in both 2023 and 2024. Burn severity classes are indicated as H = high, M = medium, L = low, C = control.

Sample processing

Wood samples were ground prior to fungal DNA extraction. Extracted DNA was amplified, sequenced, and processed to get amplicon sequencing variants (ASVs) (also known as unique fungal sequences). This data was used to assess fungal community diversity and composition.

Statistical analysis

Using the raw fungal sequences, differences in community composition were quantified using the Bray-Curtis dissimilarity index and visualized with Principal Coordinates Analysis (PCoA). Statistical significance was tested by perMANOVA.

After taxonomy assignment of these sequences, compositional differences at the fungal order level were compared among treatments using stacked bar charts of relative abundance.

To assess the effects of fire severity, samples from low, medium, and high burn severity classes in 2023 were compared to unburned controls from the same year. To evaluate changes in fungal communities one-year after fire, samples collected in 2024 were compared to those from 2023 across all fire severity classes. All analyses were conducted separately for each tree species.

All analyses and visualizations were performed in RStudio (v. 5.1.513; R Core Team, 2025). Statistical significance was assessed at a 95% confidence level (α = 0.05).