Hierarchical Spatial Models for Predicting Forest Variables Over Large Heterogeneous Domains

• Activity Number: 52
• Type: Invited
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Sudipto Banerjee*+ and Andrew Finley
• Companies: University of Minnesota and Michigan State University
• Keywords: Biomass ; Coefficient processes ; Hierarchical spatial models ; Markov chain Monte Carlo ; Multivariate spatial process ; Spatial predictive process

Abstract

We are interested in predicting one or more continuous forest variables at a fine resolution across a specified domain. Given a definition of forest/non-forest, this prediction is typically a two step process. The first step predicts which locations are forested. The second step predicts the value of the variable for only those forested locations. Rarely is the forest/non-forest predicted without error. Here we envision two latent processes generating the data: a continuous spatial process to model the forest attribute conditional upon a binary spatial process indicating ``measurable'' attribute. Flexibility is achieved by modeling the regression coefficients as unknown functions of space. The proposed models are motivated using geo-referenced National Forest Inventory (NFI) data and coinciding remotely sensed predictor variables. Computing issues for large datasets are also addressed.