Premium
Assessment of uncertainty using geostatistics
Author(s) -
Rossi Mario E.
Publication year - 1992
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.3170030105
Subject(s) - geostatistics , sampling (signal processing) , computer science , kriging , data mining , uncertainty analysis , range (aeronautics) , spatial analysis , grid , statistics , spatial variability , mathematics , machine learning , simulation , engineering , geometry , filter (signal processing) , computer vision , aerospace engineering
Geostatistical techniques applied to site characterization require modelling of the spatial dependence between attribute values at different locations throughout the site. The availability of this model allows for quantification of the uncertainties associated with site characterization from the sampling stages, through data analysis, and up to the three‐dimensional spatial model of the contaminant occurrence. Uncertainty models are built at each location of the grid thus providing probabilities of occurrence, probabilities of exceeding certain threshold(s) (critical value(s)), or probabilities of the true contamination value being within a specified range. Geostatistical modelling of uncertainty depends on the quality and quantity of the information available, and thus it is a useful tool for QA/QC programs as well, since a quantification of the performance of the sampling and other characterization work is readily available. If the desired level of uncertainty is not reached, the information gathering process can be modified and refined to meet the required standards. This paper outlines the assumptions behind the methodology, and provides a basic example on sampling optimization and on uncertainty analysis of the site.