Modeling Bark Beetle Outbreaks Using Generalized Method of Moments Estimation

• Activity Number: 140
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Kimberly Kaufeld*+
• Companies: University of Northern Colorado
• Keywords: Generalized Method of Moments ; Autologistic model ; Spatio-Temporal ; Monte Carlo ; Bark Beetle

Abstract

Recent bark beetle outbreaks in western forests has dramatically changed the forest composition. Spatio-temporal models can be used to explain the spread of bark beetles over space and time by accounting for both spatial and temporal dependence between and among geographic locations. A particular model, the centered spatial-temporal autologistic regression model (Wang & Zheng, 2012) directly accounts for the spatial and temporal dependence in the data. In this research, we propose generalized method of moments (GMM) estimation to obtain estimates of model parameters. A set of estimating equations with a specified working correlation structure is constructed from neighborhoods to account for spatial and temporal dependence of the data. We compare parameter estimates from GMM, Monte Carlo Maximum Likelihood and Monte Carlo expectation-maximization using a study of bark beetle damage using aerial survey data in the Rocky Mountain Region from 2001-2010 on a regular and irregular lattice. In addition, we compare results of a small simulation study with varying levels of spatial dependence.