All fire perimeters from the 2023 season were taken from the Alberta Wildfire open-source dataset. This study focused on large fires, defined as > 10,000 ha in size. Fires under 10,000 ha were discarded, and fires over 10,000 ha were split into Early and Late season categories, with the cutoff date being 05/15/2023. That left 18 Early-season and 18 Late-season fires that fit these metrics (Figure 1). The study area was restricted to fires with AB fire numbers, meaning mutual aid fires were mostly left out. 8 fires, 4 from each time of year, were chosen at random for this study (Figure 1).
Figure 1. Fire perimeters >10000ha chosen for this study
The fire perimeters for these large fires were overlayed onto a map of fuel types in Alberta obtained through the Wildfire Analytics lab to analyze the links between fire perimeter formation and fuel type (Figure 2).
Legend
Figure 2. A map of MWF023 overlayed on the fuel type data for that area. Different colors represent different fuel types. C=Conifer, D=Decidious, S=Slash, O=Grass, M=Mixed-wood.
One-kilometer transects were generated perpendicular to the fires' perimeter, spaced five kilometers from one another. Each transect's midpoint landed on the perimeter. Each transect had points added to it every 200m, including each endpoint (Figure 3).
This meant there were 6 location classifications per transect:
Internal- 500I, 300I, 100I
External- 500E, 300E, 100E
Transects were designed this way to ensure changes in fuel type were noticed, as documented perimeters or fuel type classifications can be off by a few hundred meters. The different distances could then be assessed to see if one distance pair offered stronger results one way or another. For example, if a fire perimeter was off by 400m to the true perimeter then only the 500m pair may pick up a change in fuel type.
Figure 3. A stretch of the perimeter of MWF023 with transects and points is shown. The right image shows the same transect overlayed on the fuel type map from Figure 2. In the case of this transect, all observations would be fuel type C1: spruce lichen woodland
The spatial join feature in ArcGIS Pro was then used to pair each point along the transect with its associated fuel type and burn code. Each point on a transect was considered a fuel type "count" going forward. For example, if the 500E in Figure 3 sat on C2, then there would be one "count" of C2, unburnt, at the 500E location. Time of year was also paired to each point based on the fire; all points on MWF023 in Figure 3 were Early, for example.