Objective 5: To evaluate the strengths and limitations of the GIS based models
Overview of Models Used
GIS is a useful tool for complex spatial decision making allowing users to work with and manipulate multiple layers of spatial data in a manageable way. This allows for the determining of optimal sites based on a variety of desired characteristics and to discover easiest routing through varying geographic and economic variables (Newhard, et al., 2008). However, limitations within both models exist in the analysis. First, the MCE and LCP analysis both boast obvious bias through the subjective choosing of constraint and evaluation criteria despite presence in literature. Further, the weights assigned to the evaluation criteria in the final suitability analysis are arbitrarily assigned and again reflect the interests and preferences of the researchers (Bell, Wilson, & Wickham, 2002). As well, due to lack of research on UAV technology, the MCE model operates under a very roughly determined relative cost of operation variable that may be quite inaccurate but the best option for use in this study considering current state of UAV research. Further, data limitations for deriving slope may exist based on resolution, as a lower resolution will have a more compressed range and will lack precision (Bell et al., 2002). Ultimately, though, MCE and LCP analysis provide a beneficial way to solve complex spatial problems regarding ideal site locations and least-cost access routes that would otherwise remain extremely difficult to resolve (Bell et al., 2002; Ballis, 2003; Newhard et al., 2008; Latinopoulos & Kechagia, 2015; Ahmed et al., 2016).
A limitation experienced was the tool of Euclidean distance not working for the data being used. Through the research of this project and specifically the development of the MCE, it was discovered that the main data used, Remoteness from Statistics Canada, had difficulties with Euclidean distance. After this discovery, multiple buffers substituded Euclidean distance with buffers from 10 - 1000 km, in 10 km intervals. These final buffers were combined together to great the distance from Thunder Bay, roads and each remote city in North-Western Ontario. Another limitation experienced was how heavily weighted Thunder Bay was in the MCE. This produced a final best suitable site to be very close to Thunder Bay, and not as north as previously hoped for.