|
Abstract:
|
There has been substantial interest in the past decade related to spatio-temporal dynamic models, particularly as applied to environmental processes. However, many of those processes are best described by nonlinear interactions over time. The literature includes some specific examples of nonlinear spatio-temporal models as suggested, for example, by nonlinear differential equations. Similarly, there is a long-standing literature in time-series related to nonlinear models, both parametric and nonparametric. One encounters difficulties in the blind application of these methods to environmental processes given the prevalent curse of dimensionality. This talk presents a class of nonlinear spatio-temporal models that can accommodate realistic dynamics and can be implemented in a reduced dimensional hierarchical framework.
|
