Suggested Locations for the Placement of Environment Canada's New Hydrometric Stations in the Georgia Basin
Title Page
Introduction
Data
Methodology
Spatial Analysis
Problems & Errors
Conclusion

SPATIAL ANALYSIS

Weighted Linear Combination

I now have four factors and one constraint to use for my multicriteria evaluation of suggested locations for Environment Canada's new hydrometric stations.

To create my weighted linear combination of these, I first used the WEIGHT module to assign relative weights to the four factors. Click here to see the pairwise comparisons. I assigned the hydrologic zones and watersheds coverages to be equally strongly more important than the helicopter sites and roads coverages. I did this because watersheds and hydrology are important criteria in determining the data content acquired by a hydrometric station whereas the transportation coverages are merely infrastructure. I assigned the roads coverage to be moderately more important than the helicopter coverage since one would prefer to travel by road to a station, if one could, since this is a cheaper form of transportation (usually). Finally, I specified that the watersheds stream order coverage is moderately less important than the zonation coverage since I believe that the need to place more stations in the snowmelt hydrologic zone outweighs the need to place a station in a particular stream order watershed. After running the module, the following weights were obtained. These weights, factors, and the constraint were then input into the MCE module, specifying "Weighted linear combination". Figure 21, (drum roll please), illustrates the result. Figure 22 shows the ORTHO version of this coverage.

The most striking thing to notice about Figure 21 is that the snowmelt hydrologic zone really stands out with its high scores. This is not surprising considering that the snowmelt zone was given a high ranking (233) in the HYDROLRANK coverage and that HYDROLRANK was given the largest weight, by far, of all the coverages. It does not disturb me that, clearly, the most preferred locations for new stations will be located solely within the snowmelt zone because this satisfies what my co-op supervisors and I saw greatly lacking in Environment Canada's hydrometric station network. Within this region, the eye can detect stream buffers that are more lightly colored (have higher ranks) than the rest. To precisely pinpoint these locations, I RECLASSed my weighted linear combination image, GBWLC, to give the very highest scores (220 to 230) a value of 1 and all other scores a value of zero. I then took this raster image of a few points and vectorized it. I then overlaid (not OVERLAYed) these points over GBWLC to see where these potential station locations were situated. Figure 23 illustrates these points as well as existing stations.

Click here to see the cartographic model of this process.

I now began to explore the points in 220VECT to see what their characteristics were and if they were close to any existing stations (and therefore redundant). I selected seven non-redundant locations in the summer snowmelt hydrologic zone that also tried to address the underallocation of stations to particular stream orders. The following table gives links to the seven potential hydrometric stations.
 
Station 1
Station 2
Station 3
Station 4
Station 5
Station 6
Station 7

Finally, here is the map showing the seven suggested locations for the placement of Environment Canada's new hydrometric stations.