Conclusions

The highest populations of economically and ecologically valuable wild salmon were found in the North Coast rivers and Bulkley river watersheds. Additionally, given population data observations, we might expect a loss in wild salmon revenues of approximately 0.3% per year, or about $150,000/ year. 

 

Discussion: 

A. Data Quality

 

The primary purpose of the study was to demostrate a potential application of hybridized economic and GIS analysis, and in that aim the project should be considered successful. However, the data used wasn't perfect. All data was estimated to be the most accurate expected values given the economic and GIS analysis tools and the data at hand. Given the inclusion of major British Columbian river watersheds, such as the Fraser and the potential inconsistency of the salmon observations, we perhaps might expect different potential impacts on salmon populations, we might not consider the data and its use to be of the highest quality. Several clues suggest this possibility, as I explain below. 

 

The salmon observations data from data.bc.gov.ca was collected, as stated, not primarily for use in population modelling but rather in charting geographic ranges. Individual salmon observation data points comprised a small fraction of the larger data set which included points indicating the presence of salmon in streams and other waterbodies. In fact, individual salmon observations comprised roughly a tenth of the dataset, waterbodies comprised the other 90%. Redundant observations were minimized, suggesting that this data must likely represent 10% or less of the general populations present.

 

As a rough data quality test, salmon observations and waterbodies layers (from Data.gov.bc.ca, available here, and on the Citations page) were overlaid and intersected in ARCMap. Of over 100,000 major waterbodies in BC, salmon observations were only present in 140.

 

Finally, The BC Ministry of the Environment's salmon observations do not appear equate proportionally (by species) with the BC Ministry of Agriculture's average catch data, as is apparent in the ranked species data below.  

 

 

 

 

 

 

 

 

 

 

There a few reasons the data used might be accurate and a few which suggest this might not be reflective of reality. First, salmon do not spawn in all waterbodies, many are damed, the data could be accurate. Second, many observations were made outside of the waterbody polygons in lesser streams, suggesting the waterbodies layer was not comprehensive, but the salmon observations could be. There is, however, the possibility that observations were not made at every waterbody fairly, that observations were concentrated in certain areas, that figures underrepresented salmon populations. 

 

Between all of the variables at play, and as an economist, I am not equipped to suggest which potential explanation, if any of these presented, is more likely. In the future, and given similar projects, I would likely consult specialists at Marine Fisheries and Oceans at BC's Ministry of the Environment, or fisheries academics for better estimations. Until then, let us consider the method involved in this study, the potential for such modelling, rather than the final data. 

 

B) Range of spill.

Oil spills likely will not effect every salmon which swims up any river to spawn. The Fraser river for example, is roughly , 1300 km long (according to wikipedia). Accordingly, one last map was compiled (displayed below), and that was to test the model presented in its ability to assess a spill which had geographic dimensions, a sort of sensitivity analysis for the model. Several buffers were considered. A 2009 study completed in this course  looked at an impact of 50km from the 843,444 gallon Marshall Michigan Enbridge spill.  Another impact of 30 km from the 100,000 gallon Pine River Spill, quite near Enbridges route, was also considered. In the end I used a 50 km buffer of the watersheds and salmon observations and repeated economic aspect of the study with the limited geographic scope. As in the analysis of entire watersheds, the North Coast Rivers watershed remained the most valuable. However, due to significantly smaller numbers of salmon observations within this buffer, I felt conclusions from the study to be beyond acceptable range, and the use of this final data to be inappropriate in the final project. With better, more comprehensive population data, such a buffered study should prove valuable.