You are here:
STATISTICAL ESTIMATION AND VISUALIZATION OF GROUND-WATER CONTAMINATION DATA
Citation:
Boeckenhauer, R. K., D. D. Cox, P. B. Bedient, AND A. W. Holder. STATISTICAL ESTIMATION AND VISUALIZATION OF GROUND-WATER CONTAMINATION DATA. U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-00/034 (NTIS PB2001-104380), 2000.
Impact/Purpose:
Information
Description:
This work presents methods of visualizing and animating statistical estimates of ground water and/or soil contamination over a region from observations of the contaminant for that region. The primary statistical methods used to produce the regional estimates are nonparametric regression and geostatistical modeling (kriging). Nonparametric regression can be used as a more "rough and ready" method to produce surface estimates with little outside intervention, whereas geostatistical modeling produces prediction errors.
Finally, a method is proposed for estimating the total amount of contaminant present in a region. The proposed method models the data as a realization of a lognormal stochastic process and then capitalizes on conditional simulation to generate realizations of the modeled process from which the distribution of the total contaminant (or integral of the process) is estimated.