Design and Analysis for Modeling and Predicting Spatial Contamination (1998)

Abstract:

Sampling and prediction strategies relevant at the planning stage of the cleanup of en­vironmental hazards are discussed. Sampling designs and models are compared using an extensive set of data on dioxin contamination at Piazza Road, Missouri. To meet the assumptions of the statistical model, such data are often transformed by taking log­arithms. Predicted values may be required on the untransformed scale, however, and several predictors are also compared. 

Fairly small designs turn out to be sufficient for model fitting and for predicting. For fitting, taking replicates ensures a positive µieasurement error variance and smooths the predictor. This is strongly advised for standard predictors. Alternatively, we propose a predictor linear in the untransformed data, with coefficients derived from a model fitted to the logarithms of the data. lt performs well on the Piazza Road data, even with no replication. 

Keywords:

Best lineal unbiased prediction, dioxin contamination, Gaussian stochastic process, lognormal kriging, ordinary kriging, spatial statistics. 

Author: 
Markus AbtWilliam J. WelchJerome Sacks
Publication Date: 
Sunday, March 1, 1998
File Attachment: 
PDF icon tr82.pdf
Report Number: 
82