1. Data: 

• Individual salmon observations in BC: BC Ministry of the Environment via data.gov.bc.ca,  taken from a data set surveyed BC generally for the presence of specific fish species in inland waterbodies. This data set was not intended to be used directly for population modelling, but if we assume that data points are roughly randomly sampled, evenly dispersed accross the province, observations should be relatively accurate of species populations in the province. These assumptions will be discussed in depth in the conclusions/ discussion section of the project. 

• BC's Major Watersheds: From Ministry of Forests, Lands and Natural Resource Operations via data.gov.bc.ca.

• Enbridge Northern Gateway Pipeline Proposed Route: Via email correspondence with Chris Darimont, Assistant Prof. at U. Vic, and co-authoer of a paper that inspired this analysis, available from raincoast.org. 

• Basemap from ArcGIS.

​

2. GIS analysis was used to isolate useful data.

The pipeline route layer was manually edited to omit the route through Alberta: this analysis was primarily interested in impacts to BC due to temporal constraints. 

Major watersheds not along the pipeline were omitted through a select by spatial attriute query.

Fish observation data points were assigned inclusion within the boundaries of the watersheds through a spatial join with the watersheds on route layer.

 

3. ARCGis's select by attributes query tool was used to quantify the number of salmon observations in each watershed, as identified by unique watershed and species' IDs.

 

4. With species observations in hand, MS Excel was used to perform the economic analysis: observed populations as a proxi for salmon revenue, and the compilation of multi criteria evaluation. This analysis used the Risk = Hazard * Probability model, where hazard was determined to be composed of economic and ecolgical value equally, and as detailed below:

 

A. Salmon values: I used the following ratio to approximate an expected salmon revenue, per year, for each species in each watershed. This allowed me to control for varying compositions of economically valuable and less valuable salmon catchs and to approximate roughly accurate salmon values. It was not a perfect proxi, but a reasonable one, potentially valuable in anticipating expected values. This proxi ratio is illustrated graphically as follows: 

 

 

 

B. The ecological index was estimated by quantifying species at risk categorization by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC): an independent committee of wildlife experts and scientists whose "raison d’être is to identify species at risk" in Canada for the advising of SARA, the federal species at risk [protection] act. COSEWIC was established in 1977 to provide a single, scientifically sound classification of wildlife species at risk of extinction. Schedule and Sara Status are determined by the Minister of the Environment and the Governor in Council (both political), based upon the recomendations of COSEWIC. COSEWIC's rankings are: 

 

 

 

 

 

COSEWIC has two additional categories: nominated for review, and not nominated. I quantified these rankings as follows:

 

 

 

 

 

 

 

 

 

 

 

 

This assumes that the ranking is linear, which may not allways be the case, but more detailed analysis was not possible.

Thus each salmon species was assigned an ecological index as follows and according to this COSEWIC database:

 

 

 

 

 

 

 

 

 

 

 

Each watersheds salmon species' observations were multiplied by their ecological index to produce an ecological index for the watershed, to then be mapped. 

 

C. Probabilities were then estimated using the legnth of the pipeline in each watershed as a proxi for the probability of a spill in each watershed. This again assumes that the potential probility of a spill is equivelent accross BC's pipeline route landscapes. Given the precedent set in the UVIC/ UC study and my inability to formulate more complex probability, this assumption was accepted. 

 

5. Finally the MCE was completed by multiplying the revenue at risk by the probability of a spill, and the ecological index at risk by the probability of a spill for each watershed, and adding the two values together, to determine a final salmon value at risk index for each watershed.

 

6. This final MCE map was compiled by manually adding a field to the watersheds layer in ArcMap, entering these calculated indices, converting the layers to raster (to show continuous gradients, and finally by mapping them. .

 

All maps and raw data/ results can be seen in the following sections of this website, with additional steps in analysis (sensitivity analysis and data quality considerations) described in the conclusions/ discussions section.