|
Activity Number:
|
422
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #304158 |
|
Title:
|
Spatio-Temporal Modeling of Air Pollutants Using a Process Convolution Approach
|
|
Author(s):
|
Jenise Swall*+
|
|
Companies:
|
U.S. Environmental Protection Agency
|
|
Address:
|
109 TW Alexander Drive, Research Triangle Park, NC, 27711,
|
|
Keywords:
|
spatio-temporal modeling ; spatial models ; air pollution ; environmental monitoring
|
|
Abstract:
|
In environmental analyses, we often depend on a limited number of samples or monitoring locations to yield information about pollutant levels across an entire region or time period. This presentation discusses the development of a spatio-temporal model, based on a Bayesian process convolution approach, which can be used to estimate air pollutant levels across the eastern US and through the time period of interest. The approach allows us to avoid modeling covariance structure directly, while accommodating non-stationarity and anisotropy. We demonstrate this approach using ozone concentrations recorded by sparse networks of monitors in the eastern portion of the United States.
|
