Activity Number:
|
162
|
Type:
|
Invited
|
Date/Time:
|
Tuesday, August 13, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
WNAR
|
Abstract - #300366 |
Title:
|
Forecasting Migratory Bird Settling Patterns with Non-Gaussian Hierarchical Spatiotemporal Dynamical Models
|
Author(s):
|
Christopher Wikle*+ and J. Royle
|
Affiliation(s):
|
University of Missouri and U.S. Fish and Wildlife Service
|
Address:
|
, Columbia, Missouri, 65211, USA
|
Keywords:
|
spatio-temporal ; hierarchical ; dynamical ; waterfowl ; settling patterns ; spectral
|
Abstract:
|
Understanding the abundance and distribution of waterfowl populations is essential for wildlife conservation. Extensive monitoring programs over the last half century have provided substantial data concerning population abundance over time. Although there is significant ongoing research on the biological mechanisms for yearly settling patterns of waterfowl, there is still substantial uncertainty. We model temporal and spatial variation in historical waterfowl settling patterns using hierarchical spatiotemporal dynamical models. The data are inherently non-Gaussian but can be successfully modeled conditional on underlying Gaussian dynamical processes. Given the high dimensional nature of the data and forecasting area (a large fraction of North America), spectral methods are utilized in the underlying hierarchical framework. We demonstrate significant skill in forecasting the spatial distribution of waterfowl abundance on annual time scales.
|