Intermediate GIS Processing
As discussed in Objective 3, there are various intermediate steps and map outputs that contribute to the complete MCE model. The creation of standardized criteria and contraint layers provide insight into habitat suitability for each factor independently. Additionally, they aid in visualizing how multiple layers are combined using the MCE algorithm. Listed below are several intermediate standardized layers including land cover and slope criteria, hazardous forest type, and water constraint. As the output for proximity to water and distance from roads are the products of the euclidian distance tool, it is difficult to visually observe these values in a visual output format and therefore intermediate maps are not included below but were still components of the MCE algorithm.
The overall purpose of this research is to develop a GIS model that predicts areas of suitable habitat for Algonquin wolves within Algonquin Provincial Park in order to determine the proportion of viable wolf habitat at low risk for loss due to potential wildfires. The initial analysis comprised the criteria and constraint layers (refer to Figure 10 and Figure 11 in Appendix I), excluding any wildfire risk factors, to determine suitable habitat within the park based on environmental variables. The MCE model analysis produces a map delineating habitat suitability within Algonquin for wolf habitat, based on the criteria and constraints laid out in Objective 1 (Figure 6). The most suitable habitat is observed in regions of light yellow and green while less suitable habitat is represented by dark blue, with suitability scores ranging from 0 to 72.7. Visually, the map depicts the majority of suitable area exists in the northern extent of the park. The final model is created with the same criteria and constraints as the initial habitat suitability analysis, with an imposed wildfire risk layer (refer to Figure D in Wildfire Risk Layer). This analysis creates a final map delineating areas of suitable wolf habitat within Algonquin with scores adjusted to reflect wildfire risk. The final model considers environmental preferences (as outlined in Objective 1) while incorporating forest fire risk based on fuel types. The completed output is a map with high suitbaility represented by light yellow/green and less suitable habitat represented by dark blue. As seen in Figure 7, suitability scores range from 0 to 72.7. The overall distribution of suitable habitat does not change drastically with introduced forest fire risk, however it is significantly decreased as more area is covered by dark blue cells with lower suitability values that are at higher risk for disturbance.
Figure 6. MCE model output of habitat suitability.
Figure 7. Model output with imposed wildfire risk.
Acceptably suitable habitat is considered to have a score equal to or greater than 60. The intial MCE model outputs a total of 21.8% of available land as suitable habitat. Specifically, there is a large area of reasonably suitable habitat in the North West sector of the park, as seen in green. However, small segments of suitable habitat are dispersed throughout the entire study area. The division of suitable and non-suitable habitat can be seen in Figure 8 and Figure 9, displaying the data according to the suitability threshold value of 60. The final map shows that with the inclusion of wildfire risk values into the model, suitable wolf habitat decreases significantly to 8.7% of the total park area as expected (Figure 9).
Figure 8. Habitat with suitability scores greater than or equal to 60 based on the initial model.
Figure 9. Habitat with suitability scores greater than or equal to 60 based on the final model.
As stated in Objective 4, the nature of this model renders it difficult to evaluate using an existing model as reference. Therefore, Google Earth is used to visually examine suitable den sites given by the model. This method of evaluation is very susceptible to human error, as it involves the inspector making judgement calls based on what is seen by that inspector. To combat this, each team member is to individually inspect the model until consensus is reached regarding whether the model agrees with what is seen on Google Earth. Upon inspection, it can be seen that the model agrees with Google Earth. Google Earth shows roads, water bodies, and slope, leaving out two of the factors outlined in Objective 2 (the land cover type and fire risk in an area). The only road displayed by Google Earth is Highway 60, as it is a large road which is the only type of road the model takes into account (leaving out small roads, such as the roads used for logging purposes). On the model, Highway 60 is defined as an unsuitable area, as it should. The water bodies displayed on both the model and Google Earth are shown in the same spots and with the same shapes, meaning the model displays the bodies of water properly. Lastly, the slopes shown on both the model and on Google Earth through shading agree with each other. These three factors (roads, water, and slope) are consistent in the model and in Google Earth. The factors land cover and fire risk cannot be evaluated using Google Earth, as these data are additional layers and do not exist on Google Earth. Unfortunately, it is not possible to quanititatively identify suitable habitat sites through Google Earth, so the results of this evaluation are based on qualitative observations and represent a best estimation. Overall, the model seems to identify suitable habitats in the same areas each of the team members thought suitable habitat would exist based only on inspecting Google Earth.