DETECTION OF HOTSPOTS FOR THREE-DIMENSIONAL SPATIAL DATA AND ITS APPLICATION TO ENVIRONMENTAL POLLUTION DATA

Abstract

We address the problem of detecting areas with markedly high values (hotspots) in spatial lattice data. Spatial lattice data are observations that include several spatial areas supplemented by neighboring information. The spatial scan statistics is an effective tool for hotspot detection. Echelon analysis is an analytical method to investigate the phase-structure of spatial data both systematically and objectively. In this paper, we describe the structure of spatial lattice data using echelon analysis and detect hotspots based on the echelon structure. As an application, we apply this method to simulation data of leachate accidents in final-disposal sites and detect considerably high-density pollution areas.