In an effort to bridge the gap between theory and practice, we explore
the impacts of spatial correlation on parameter estimation and model
selection. With regards to parameter estimation, we investigate the
influence of the strength of spatial dependence on maximum likelihood
and restricted maximum likelihood estimates of covariance parameters.
Spatial correlation can also impact model selection, but is often
ignored. We show that using AIC for a geostatistical model is
superior to the more traditional approach of ignoring spatial
correlation in the selection of explanatory variables. Much of this
work includes consideration of the effects of different sampling
designs and sampling intensities. This is joint work with several
research groups including Richard Davis, Alix Gitelman, Kathryn
Irvine, Andrew Merton, and Sandra Thompson.
Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.