The Threat of the Death Cap in Southwest British Columbia

Limitations and Sources of Error

 

Limitations of Distribution Mapping

• Some limitations in mapping the distribution of the Death Cap in Vancouver using ArcGIS online were mainly related to the conversion of addresses into georeferences. Indeed, in some instances, addresses which were georeferenced in ArcMap were portrayed to be in random places in the United States. This needs to be fixed.

• Some challenges arose due to the fact that the Betulus tree data file was too large to be uploaded on the online version of ArcGIS. Moreover, the excel file with the Amanita phalloides sites combined the civic number and street address together manually, whilst this could have been done by georeferencing the table directly in ArcMap.  

 

Limitations of the MCE Analysis

 

• Buffer of 50m for all host tree species, even though the age of the tree and the tree species affects the root distance

• Vancouver tree data set only consists of boulevard/verge trees, no park tree data, or private tree data (e.g. trees in backyards). To address this, the park factor was considered by overlaying the 50m host tree buffer layer with the known park location polygons. Parks with higher overlay were considered to be at greater risk of Death Cap presence.

• Lawn watering factor was determined by attaining Vancouver land-use data, with this data, percentage residential zoned land cover was calculated per census tract area. Higher residential zoned land was thought to have greater lawn watering and therefore higher risk of Death Cap presence. As discussed in the results section, further analysis is required to determine the effect of lawn watering.

• Tree age was not considered in the MCEs, this is an important factor that should be included in future studies. As tree age increases, the likelihood of Death Cap presence increases.

• As the GIS analysts, we are not mushroom or health experts. With this in mind, it should be noted that the weighting regimes for each factor are arbitrary. The Analytical Hierarchy Process is a platform to aid decision making by creating a structure for factor weights, but at the end of the day, it is the user’s judgment that determines the weighting. Further discussions with mycologists and health professionals would result in more accurate factor weights.

• Only nine factors were considered. There are certainly more than factors that affect Death Cap presence and potential risk.

 

Limitations of the Maxent Analysis 

• Sample size: Maxent makes predictions about species distributions based on the locations of sample species and environmental layers. The sample species train the model, showing it what environmental factors are a suitable habitat. A small number of sample species can deprive the model of this information causing it to over or underestimate the distribution of potential habitats.

• Environmental Data: The suitable habitats that Maxent predicts for a species are contained in the environmental layers entered into the model. The availability, amount, and resolution of this data is a limitation. More localized data, finer resolution data, or more environmental layers to show more habitat detail could change the model output.

 

Conclusion 

This project highlights the processes required to map the potential zones for Death Cap presence. The techniques applied here can be used for any geographic location. Although there is significant variability in the results depending on the input data, the factors considered and the weighting of factors, the results presented show that presence of the Death Cap mushroom is certainly an issue in Vancouver, Mission, and Victoria. The risk associated with Death Cap presence can be expected to increase as the mushroom spreads. In Vancouver, known host trees can be found all over the city, yet only 66 Death Cap sites have been recorded. Based on this study further investment into analysis of how Death Cap spreads and potential mitigation techniques are recommended.

​