Open AccessThe use of sequential indicator simulation to characterize geostatistical uncertainty; Yucca Mountain Site Characterization Project
Author(s) -
K.M. Hansen
Publication year - 1992
Language(s) - English
Resource type - Reports
DOI - 10.2172/138892
Subject(s) - variogram , computer science , kriging , quality (philosophy) , geostatistics , sample (material) , uncertainty analysis , characterization (materials science) , data mining , statistics , mathematics , simulation , spatial variability , machine learning , philosophy , chemistry , materials science , epistemology , chromatography , nanotechnology
Sequential indicator simulation (SIS) is a geostatistical technique designed to aid in the characterization of uncertainty about the structure or behavior of natural systems. This report discusses a simulation experiment designed to study the quality of uncertainty bounds generated using SIS. The results indicate that, while SIS may produce reasonable uncertainty bounds in many situations, factors like the number and location of available sample data, the quality of variogram models produced by the user, and the characteristics of the geologic region to be modeled, can all have substantial effects on the accuracy and precision of estimated confidence limits. It is recommended that users of SIS conduct validation studies for the technique on their particular regions of interest before accepting the output uncertainty bounds