Predicting Site Locations for Biomass-using Facilities with Bayesian Methods

Logistic regression models combined with Bayesian inference were developed to predict locations and quantify factors that influence the siting of biomass-using facilities that use woody biomass in the Southeastern United States. Predictions were developed for two groups of mills, one representing larger capacity mills similar to pulp and paper mills (Group II), and another group of smaller capacity mills similar to the size of sawmills (Group I). “Median Family Income,” “Road Density,” “Slope,” “Timberland Annual Growth-to-Removal Ratio,” and “Forest Land-Area Ratio” were highly significant in influencing mill location for Group I. “Slope,” “Urban Land Area Ratio,” and “Number of Primary Wood Processing Mills” were highly significant in influencing mill location for Group II. In validation the sensitivity of the model for Group I was 86.8% and specificity was 79.3%. In validation the sensitivity for Group II was 80.9% and specificity was 84.1%. The higher probability locations (> 0.8) for Group I mills were clustered in the southern Alabama, southern Georgia, southeast Mississippi, southwest Virginia, western Louisiana, western Arkansas, and eastern Texas. The higher probability locations (> 0.8) for Group II mills were clustered in southeast Alabama, southern Georgia, eastern North Carolina, and along the Mississippi Delta. Key-words: Biomass-using facilities, woody biomass, site location, prediction, Bayesian logistic models