Research Findings: Objective 3
Apply the Model to the City of Montreal to Identify Census Tracts with the Least Tree-Cover and Greatest Heat Vulnerability
Results from the three model segments are outlined below.
(1) Canopy Cover
The model isolates trees from the LiDAR point cloud and develops a shapefile of canopy coverage for the City of Montreal (Figure 5). This shape-file is then aggregated to the census tract and the 25% of tracts with the lowest tree canopy coverage are identified (Figure 6). Canopy cover among the tracts with lowest coverage ranges from 4.6 to 14.3%. Canopy cover in all populated tracts ranges from 4.6 to 58.0%.
Figure 5. Tree canopy cover for the Island of Montreal derived from LiDAR data.
Figure 6. Map of Census Tracts Identified as High Priority for the Canopy Cover Variable.
(2) Socioeconomic Variables
Each tract is indexed based on percent prevalence of variables 1-7, and the 25% most vulnerable census tracts are identified for each socioeconomic variable. Figures 7-13 show high- and low-vulnerability tracts within the City of Montreal based on each socioeconomic variable: Age (Figure 7), Income (Figure 8), Education (Figure 9), Ethnicity (Figure 10), Social Isolation (Figure 11), Height of Housing Structure (Figure 12), and Age of Housing structure (Figure 13). In each figure the high priority tracts are identified in pink, and low priority tracts are shown in grey.
(3) Proximity to cooling centers
Average proximity to cooling infrastructure is derived for each census tract. Figure 14 shows the top 25% most vulnerable census tracts based on proximity to cooling infrastructure; residents of these tracts have the greatest average Euclidean distance to travel to access their nearest cooling center.
Final Model Output
A bivariate heat vulnerability map shows the distribution of combined vulnerability throughout Montreal based on the indexed multi-variable vulnerability scores (Figure 15). Those tracts with greatest social heat-vulnerability and lowest tree cover make up the highest-priority tracts.
Figure 15. Bivariate heat map of the Island of Montreal. Tracts are ranked on a spectrum of both Canopy Cover and Socioecnomic Vulnerability. Tracts with High Vulnerability and Poor Cover (highest priority) are designated dark green while tracts with Low Vulnerability and Good Canopy coverage are denoted in light green.
The choropleth resulting from Equation 1 is used to conduct the Hot Spot Getis-Ord Gi* analysis. This analysis identifies areas of greatest vulnerability (hot spots) and least (cold spots) vulnerability to heat. The model identifies three areas of high vulnerability with a 99% confidence level as shown in Figure 16. These areas are extracted from the map and overlaid onto the bivariate heat map of tree cover and socioeconomic vulnerability to find the most vulnerable tracts in Montreal. Figure 17 shows the 10 high-priority tracts located within two of the three high-priority areas. These ten tracts are highly vulnerable to heat and have minimal canopy cover. These are the tracts that require increased canopy cover and should be the focus of future tree planting efforts in Montreal.
Figure 16. Heat map of census tracts in Montreal that represent areas of highest and lowest heat vulnerability. The three areas that are highly vulnerable with 99% confidence are Ville-Marie (centre-right), Montreal-Nord/Saint Leonard (top), and LaSalle (bottom).
Figure 17. Map of the ten most heat vulnerable census tracts in Montreal. The model has determined that these tracts would benefit the most from increased tree planting to mitigate heat vulnerability.