There are many areas of uncertainty, error and subjectivity present in the design of our MCE and within the data that was utilized. Any conclusions or recommendations that are drawn from our MCE and its results would have to be directly tested before any action is taken to provide renewable energy to Haida Gwaii’s North Grid.

The subjectivity used in our Advised Against layer simply reflects attitudes we encountered in our research. Ecologically and culturally important areas were cited more often than economic factors, so they were given precedence in the normalization of the layer. Furthermore, the classification and selection of commercially active attributes in fishing and transportation layers derived from BCMCA excluded all but the most densely reported areas. This classification was done arbitrarily to save only areas used extensively. 

The wave and wind energy potential data layers provided by the BCMCA exist at low 2 km by 2 km resolution. Such low resolutions may cause errors that are unaccounted for in our MCE analyses. For example, while the low spatial resolution of the wave energy potential raster layer may be a major source of error for waves in the open ocean, it’s plausible that this layer may not be giving us an accurate representation of how wave energy varies in areas close to the coastline, where bathymetric features (which influence wave height) are more rapidly changing. These resource estimates were made almost a decade ago, thus it may be worthwhile for renewable energy proponents to provide more detailed resource estimates with increased spatial accuracy. The wind potential layer is set at 30m height, though wind turbines will be set higher. This may actually result in higher average annual wind speeds, though this is uncertain.

The raster layer representing the direct incoming solar radiation for each pixel in our DEM (created by the Area Solar Radiation tool) also represents a large area of error/uncertainty present in our solar results. First off, the shading effects caused by Haida Gwaii’s forested areas were modeled simply by adding 40 meters to the DEM for each DEM pixel that overlapped with a tree polygons in our vegetation layer. Of course in reality, all forested areas on Haida Gwaii are not 40 meters in height and thus the shading effects caused by the trees on Haida Gwaii are subject to high degrees of uncertainty and error. Furthermore, our use of the Area Solar Radiation tool did not specifically take into account the effects of reduced incoming solar radiation caused by cloud coverage. To expand, the Area Solar Radiation Tool requires an input transmissivity value (between 0 and 1), to calculate the amount of direct incoming solar radiation over a period of time. We used the default value of 0.5. In reality, clouds are likely to cover the island for significant portions of the year, which would decrease the transmissivity of the lower atmosphere and thereby decrease the amount of incoming solar radiation received at each pixel over the course of a year. These sources of error could explain why our solar results predicted a very large power capacity of ~10MW for a solar farm of area 350 meters by 800 meters on Haida Gwaii.

The weights which were chosen to be applied to our normalized factor layers when conducting our weighted overlay as well as the scores that were applied to our advised against areas layer also represents a major source of uncertainty and error which is brought about by the inherent subjectivity of our MCEs. In many ways, it is naive and far-fetched for two undergraduate students at UBC to decide the relative importance of each factor without any consultation with the people and governing bodies on Haida Gwaii. While our selection of what scores/weights would be applied across our MCE was not arbitrary, they were simply speculative and in no way certain.