Skip to content

Methodology

To achieve our goal of finding suitable habitats we needed to conduct research on the habitat conditions for each of the three fungi species of interest. These conditions were the basis for following the classic Multi-Criteria Analysis for our project and acted as our criteria for our AHP analysis. These habitat conditions differed in both importance and type for each species : 

Morel mushroom habitat required the consideration of the following criterion or constraints to our analysis:

  • Burned areas and the intensity of the fire
  • Elevation
  • Aspect

Chanterelle mushroom habitat criteria included:

  • BEC zones
  • Vegetation density
  • Aspect

and Late Fall Oyster Mushroom habitat includes the consideration of:

  • Broadleaf forest cover
  • Elevation

Before we were able to Normalize in the suitability modeler we had to prepare and clean our data so that it was usable for our goals. 

Our Methodology will be discussed in the following order: 

  1. Collection of Data and Data Preparation
  2. Multi-Criteria Evaluation Analysis (MCE)
  3. Sensitivity Analysis
  4. Accessibility Layers

Collection of Data and Data Preparation 

  • Initially, we wanted to focus our project on a provincial scale, however this was proven difficult not due to data limitations, but due to our computer’s capabilities. So we had to create a smaller study area that would better perform with the hardware available to us. To create this new boundary we created a KML file of a polygon that we made on Google Earth Pro, with a new focus – Southern BC, which we then transferred over to ArcGISPro. 
  • Because we were running the suitability modeler we had to take that into account while preparing our data and convert certain datasets that we need for the mushroom habitats into rasters. For example, we converted the fire polygon dataset that we used into a raster dataset along with the BEC zone polygon dataset and the Broadleaf forest polygon dataset. We did this using the Polygon to Raster tool specifying the value field as “Burn Severity” for our fire polygon dataset, the “Zone Name” for our BEC zone dataset, and “Shape Area” for our broadleaf forest dataset. We then had to clip these new Rasters to our study area boundary. 
  • We needed aspects for 2 of our mushroom species. To create an Aspect map we used the terrain aspect raster function on our DEM dataset of the province.
  • Chanterelle mushrooms are found in densely vegetated areas, to account for this we created an NDVI map with Landsat 8 imagery on Google Earth Engine so we could place a focus on areas with high NDVIs. A map of the NDVI imagery created in GEE is represented below in Figure 1.
  • Once our rasters were the same size and under the same projections, we were able to start our Multi-Criteria Evaluation Analysis.

    Figure 1

Multi-Criteria Evaluation Analysis

  • Beginning with determining the weights of each criterion we ran an Analytical Hierarchical Process through an online application called 123ahp listed in our sources. With the weights we got from our AHP, we inputted the rasters into our Suitability Modeller to begin the MCE Analysis. 
  • We determined the following weights and transformations for each mushroom species: 
  • Morel Mushrooms
    • Burned areas = 78.70 – Gaussian with the midpoint to equal the value we classified to mean moderately burned areas, as Morels need areas that burned well enough to activate, but not burn the mycelium. 
    • Elevation = 16.73 – We chose MSLarge for this to favor higher elevations, with a lower threshold of 900 and a midpoint of 1200.
    • Aspect = 4.57 – For Aspect, we used Gaussian as well, You are more likely to find Morels on South and West facing slopes due to orographics and sunlight, so we placed the midpoint to be 250 degrees to favour South-West slopes the most and fan out from that point in favourability. 
    • Consistency Ratio  = 0,4188
  • Late Fall Oyster Mushrooms:
    • Elevation = 20.00 – Exponential was chosen for this variable to ensure we are favoring higher elevations only for this species. 
    • Deciduous forest cover = 80.00 – Large was chosen for this raster. Our thoughts on this were higher surface areas of deciduous forest cover would include the likelihood of finding these mushrooms.  The area with the highest surface area in the raster was chosen for upper threshold 
    • Consistency Ratio = ∞
  • Pacific Golden Chanterelles:
    • Aspect = 27.97 – Gaussian was used as a transformation function here to ensure West slopes are favored with a midpoint of 270. West slopes get more precipitation making them more green than Eastern and therefore more densely vegetated and favorable habitat for these mushrooms. 
    • NDVI = 62.67 – MSLarge was the function used for NDVI as we wanted to favor areas with higher NDVIs, and 0 was the value for the lower threshold. 
    • BEC Zones = 9.36 –  We did Gaussian for this raster based on the assigned values of each BEC zone. The 2 zones needed were values 9 (Coastal Western Hemlock) + 11 (Coastal Douglas Fir) so we made out midpoint 10 lower threshold 2 and upper threshold 12. 
    • Consistency Ratio = 0,0824
  • The suitability modeler was run with each of the above values three separate times; one for each species. After running our model with the weights and transformations we moved to the locate pane where we identified that we wanted 30 different regions located with a minimum size of 5km. In our models, one raster cell of suitable habitat equals this minimum size of 5km. After running our suitability modeler with the locate pane filled out the suitable habitat locations appeared on our map.
  • Below in Figure 2, the green areas show suitable habitat locations for one of our mushroom species (Late Fall Oyster) in a premature map.
Figure 2

Sensitivity Analysis 

  • Before we were done with our MCE we had to do a Sensitivity Analysis for each species’ MCE analysis. We ran the suitability modeler again for each species separately with the same rasters needed for the habitat but instead of at the weights decided through the AHP, setting them to equal weights, there neither of the three weights held a higher preference. And with the same normalization functions and parameters in the locate pane. 
  • After completion of the sensitivity Analysis, we completed our Multi-Criteria Evaluation Analysis.

Accessibility Layers

    • Before we could create our deliverable maps we had to take into account accessibility to the areas identified by our MCE Analysis. To do this we created one layer that included Private Lands, Provincial Parks and Ecological Reserves, Indigenous Lands, and Government Lands (all of the areas you cannot pick from) and overlaid them on top of our suitable areas. To put all of these datasets together we merged them into one layer and then dissolved the boundaries between them all for aesthetics. 
    • we also overlaid our map with roads for additional accessibility help
    • Figure 3 is another premature map that shows the “no-foraging” areas coloured in grey. This “Non-forage Lands” layer includes provincial parks, ecological reserves, privately-owned lands, government-owned lands, and Indigenous reserves; essentially all lands where you cannot forage, any areas not highlighted under this layer are crown lands where you are able to forage

 

Figure 3

Written by Amy Manuel & Isabela Hrehorsky

 

Spam prevention powered by Akismet