Further Analysis

OLS

Variables: Education, Immigrants, Mining Industry, Natural Resources and Agriculture

AICc: 10719.546

Adjusted R2: 0.087857

Geographically Weighted Regression

Variables: Education, Immigrants, Mining Industry, Natural Resources and Agriculture

AICc: 9828.938

Adjusted R2: 0.4613155

Crude-R2 fixed

Figure 8. GWR Local R2

Figure 7. The local R2 values for this model range from 0.003 to 0.328, with the highest R2 values located on the western portion of Census Division 11. This also happens to be the area with a high density of crude oil spills. Compared to our first GWR model, the addition of labour force variables explain the dependent variable better in the western area of the Census Division while our previous model performed poorly outside of the city.

Figure 9. Raster Coefficient surface for Mining Industry

Figure 9. It is expected that areas with the highest incidents would be positively correlated to areas with a large percentage of the population working in the mining, quarrying, and oil and gas extraction Industry. While this may seem obvious, it is important to note that this may explain why census variables such as income may not explain environmental justice in Alberta. We commonly expect those with lower incomes to be the most affected, when in the case of Alberta, people working in the mining industry typically have a higher income and live near oil facilities. This seems to be true for the western portion of CD no. 11, where there is a large percentage of labour force by industry in mining and oil extraction.

In Edmonton (small inset map on top left hand corner of figure 9), there is a slightly negative relationship. The % of labour force in the mining industry increases as the density of crude oil spills decreases. But the coefficient values here are close to zero, indicating a negligible relationship between crude oil releases and labour force in mining.

Figure 10. Raster Coefficient surface for Natural Resources and Agriculture

Figure 10: For our further analysis, we expected that the demographic affected by crude oil spills would be those outside of the city working in agriculture. The variable % Labour Force by Industry: Agriculture, forestry, fishing and hunting was not found to be significant however % Labour Force by Occupation: Natural resources, agriculture and related production occupations was significant. The red areas are locations with a positive relationship between incidence of crude oil spills and labour in natural resources(NR) and agriculture(AGRI). These areas are also where our model fit best. The western area in CD no.11 and the eastern area outside of Edmonton (CSD) both have a positive relationship and a high R2 value relative to other areas. Similar to labour force in mining, the red areas in figure 10 are also areas with a high percent of labour in natural resources and agriculture. This area does not have the highest percentage band ( 21% – 50%), but ranges from about 5% to 20%.

In the eastern perimeter of Edmonton, there is a small red area showing a positive relationship. This areas also has a low % of people working in natural resources and agriculture. We found that select areas of the city contained the highest % of labour in NR and AGRI and a negative coefficient. However these are also areas where the GWR performed poorly. The positive relationship between crude oil release incidents and labour force in natural resources is stronger in the western portion of census division no. 11.

Leave a Reply

Your email address will not be published. Required fields are marked *

Spam prevention powered by Akismet