Activity Number:
|
322
- Analyses in Ecology, Epidemiology, and Environmental Policy
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
|
Sponsor:
|
Section on Statistics and the Environment
|
Abstract #318525
|
|
Title:
|
Comparing the Use and Implications of Stationary Confidence Limits to Weighted Bootstrap and Bayesian Credible Intervals in Environmental Monitoring Applications
|
Author(s):
|
Marie-Agnes S. Tellier* and Timothy J. Robinson
|
Companies:
|
Trihydro Corporation and University of Wyoming
|
Keywords:
|
Confidence limit;
Remediation;
Area-weighted bootstrap estimates;
Geostatistics;
Credible interval;
Integrated Nested Laplace Approximation (INLA)
|
Abstract:
|
Industrial site assessment and monitoring are often required by State and Federal agencies when assessing corporation compliance to local or national health standards and for determining whether remediation is needed. Quite often, methods such as confidence limits (CLs) of the mean are used and assume spatial stationarity (i.e. no spatial variation in the parameter of interest) across the site when many sites contaminated by industry are characterized by an outward gradient in concentration from the contaminant source or hot spots. Additionally, samples may not be homogeneously distributed throughout the geographical region of site; some samples may represent larger areas than others. CLs not accounting for the spatial distribution of the response are exploratory at best, if not misinformative. Upper CLs, often used in risk assessment to determine if the site is safe for human access, should be stable and reliable estimates. This paper compares CLs obtained through traditional methods to CLs obtained using alternative approaches accommodating nonhomogeneous spatial distribution of the response such as area-weighted bootstrap CLs and spatial Bayesian credible intervals.
|
Authors who are presenting talks have a * after their name.