Stacking Different Spatial Statistics in a Novel Recursion Algorithm to Improve the Design of Waste Management Regions in Saskatchewan

Abstract

AbstractCanadians disposed 25 million tonnes of waste in 2018. Some Canadian provinces have implemented regionalized waste management systems such as Alberta and Nova Scotia. In Saskatchewan, the idea of regionalization has been discussed since the 1990s, however, transition to regional systems has been difficult due to the autonomous nature of prairie communities. Currently, the Government of Saskatchewan intends to investigate and encourage regional collaboration among municipalities. Previous work on regionalized waste management systems introduced an algorithm capable of improving and optimizing regions for waste management in various jurisdictions. It was theorized that regions could be optimized when the number of landfills, populated places, and roads across regions was spread evenly; mathematically, regions were optimized when the standard deviation of these parameters was reduced across all regions. Successful application of the tool yielded reductions in the standard deviation these parameters by 4.9–46.1% in Saskatchewan. In more recent work, different spatial statistics such as central feature, mean center, and median center have been substituted into the proposed Centroidal Voronoi Tessellation (CVT) algorithm with varying success. The objectives of this study are to: (i) stack different spatial statistics (mean and median center) on top of the initial CVT algorithm and (ii) compare the results to those using only the CVT algorithm to determine if the stacking method proposed in this study can further improve the results of the CVT algorithm. The results from this study may help to further develop data driven regions for waste management in Saskatchewan and Canada.