Bayesian Spatial Modeling for Loblolly Pine Seed Source Movements

• Activity Number: 264
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Alfredo Farjat* and Brian J. Reich
• Companies: North Carolina State University and North Carolina State University
• Keywords: Spatial statistics ; Bayesian modeling ; climate change ; plant breeding ; MCMC ; matern

Abstract

The establishment and analysis of provenance tests are generally intended to identify superior seed sources for planting at specific locations. They usually consist of replicated experiments comparing trees grown at several locations that were collected from different geographical regions of a species distribution. The possibility of utilizing spatial information from the sampled trees and the climatic differences between their origins and the test site locations provide meaningful information for assessing the response of genotypes to environmental change. We developed a spatial Bayesian approach for predicting the expected height of seed sources using a set of climate variables and the growth potential of the planting site as predictor variables. The proposed modeling technique introduces a separable covariance structure to describe and estimate effects associated with the origin and planting site locations. The model is tested using a large replicated provenance test of loblolly pine (Pinus taeda L.). The statistical model can be used as a seed deployment tool aimed to identify the location of seed sources suitable for plantations under a particular climate scenario.