VISUALIZING SPATIAL PROBABILITY FIELDS REFLECTING SIGNIFICANT DIFFERENCES IN NEIGHBORING VALUES

Abstract

 In this decade, hazard maps have been developed and utilized for disaster prevention planning in communities. The values shown in the hazard maps are inherently uncertain. It is very important to reflect the probability information in the spatial resolution as well as the average value. In this study, given a probability distribution, we propose a new visualization method (Uniform uncertainty mapping) that adjusts the resolution according to the significant difference at neighboring points. We first introduce the theoretical framework of this method and discuss its performance by applying it to numerical problems. It is then applied to the problem of estimating 1D velocity structure in microtremor array exploration as an example of the inverse problem, and to the problem of displaying 2D ground amplification as an example of the seismic hazard map, respectively.