Data Collection & Manipulation
Data for this research was acquired mainly through the acquisition of secondary data sources based at provincial or national scales and modified for use in this analysis. A major portion of data was obtained from the Department of Geography at Simon Fraser University in vector format. Metadata and attribute data of these files was reviewed to determine its suitability for this project. Before any stage of analysis could take place, the data needed to be manipulated and was then clipped from the provincial or national levels to incorporate only those spatial entities located within the study area. Table 3.1 outlines the original datasets utilized for this research.
Additionally, data was manually digitized from paper sources and point location data geocoded and spatially referenced for GIS input. The locations of the 14 possible municipalities containing sufficient populations in the four regional districts were georeferenced using absolute decimal degrees (Table 3.2). The primary BC Hydro power line extending from Fort Nelson to Prince George and east and west along the provincial highway #16 was also digitized from BC Hydro maps (BC Hydro, 2000). This data geographically shows the layout of the main electricity grid and can be compared with proposed site locations of nuclear facilities. Additionally, the locations of the main electricity generating stations in northern British Columbia (the W.A.C. Bennett and Peace River Dams) were geocoded, both of which are located along the Peace River and currently power the majority of the needs to central and northern British Columbia (BC Hydro, 2000). An earthquake threat zone was also established demarcating areas in close proximity to the central and northern coast mountain range which pose a moderate threat risk for seismic activity. The northern wood bison and pink mountain plains bison habitat protection area of the Northern Rockies regional district was also examined (Figure 3.1) (Harper et al, 2000). This data was not available in digital format and had to be manually digitized from supplementary paper sources (Harper et al, 2000).
Upon review of secondary data sources, errors and the incompleteness of the data was discovered and had to be manually modified before use in this analysis. For example, due to impossible values found within the point elevation dataset, elevation values of less than 200 metres were discarded, often in the form of zero and one anomalies. A random sample test was then completed and confirmed that the remainder of the elevation points were accurate. Additionally, the provincial park dataset was also incomplete as eight provincial parks had not been added to the dataset and had to be digitized. These manual additions proved important in influencing the multi-criteria analysis mainly because of their proximity to sufficient water bodies and municipalities.
The Universal Transverse Mercator (UTM) Zone 10 North projection was utilized for this research as this projection succinctly encompasses all of the regional districts assessed in the analysis. Additionally, the Canada Albers Equal Area Conic projection was used at the provincial scale as a reference to define the geographical location within British Columbia of the regional districts examined.
A digital elevation model (DEM) was created in order to determine slope percentages within the study region. This model was created in ArcGIS using elevation point values as a reference source from which the 3D Analyst function could be used to create a raster interpretation. A sample of 12 addition points based on mean elevations were digitized outside of the regional district range in order to construct a square raster tessellation with which to analyze the entire geographical study area. A 200 metre spatial resolution was chosen to allow for relatively precise distance measurements considering the separation of the elevation values and to manage storage sizes (Figure 3.2). In addition, the spatial extent of the study area of the four regional districts examined covers an approximate total area of 329,005.30 km² requiring a relatively small scale analysis.
The final step in the data acquisition process was to convert all datasets from ESRI Shapefile format suitable in ArcGIS to IDRISI vector format before rasterization could be accomplished. Using the “SHAPEIDR” module in IDRISI, followed by “POINTRAS,” “LINERAS,” and “POLYRAS” modules all vector files were converted into raster IDRISI file format (Figure 3.3). The exception to this was the interpolated raster DEM which was directly transferred from ArcGIS to IDRISI using the “GRID2IDRISI” function. This DEM was then used as the raster reference image using the “INITIAL” module allowing the rasterization of all vector files.
Figure 3.3: Cartographic model created in IDRISI's Macro Modeler showing the operations taken in the rasterization of vector data.