Year
1994
Abstract
In many areas of environmental engineering, such as site characterization of potential nuclear waste repositories and hazardous waste site remediation, it is often necessary to estimate the cumulative distribution function (CDF) of a particular variable. When data are known to be correlated, a case that often arises with spatial and temporal data collected in support of nuclear management programs, the usual empirical distribution function estimator may be unsuitable because it weights each datum equally without regard to the correlation structure. Weighted distribution function estimators that incorporate the spatial or temporal correlation in the data offer a more realistic alternative. In this paper, several such weighted distribution function estimators are presented in a spatial context. Differences between these estimators and the usual empirical distribution function estimator are illustrated using transmissivity data collected as part of one of the Department of Energy's nuclear waste disposal projects.