A Two-Stage GIS-Based Suitability Model for Siting Biomass-to-Biofuel Plants and its Application in West Virginia, USA

Woody biomass has been considered of low value because the cost of removal generally exceeded market price. New, valued-added markets to offset removal costs are necessary for utilization to be effective. In recent years the use of biomass as feedstock for biofuel production in the United States has been on the rise. A variety of liquid fuels can be produced from woody biomass; ethanol is one of the most promising. This study presents a two-stage approach to selecting woody biomass-based biofuel plants using Geographical Information System (GIS) spatial analysis and the multi-criteria analysis ranking algorithm of compromise programming. Site suitability was evaluated to minimize direct cost for investors and potential negative environmental impacts. The first step was to create a site suitability index using a linear fuzzy logic prediction model. The model involved 15 variables in three factor groups: (1) general physical conditions, (2) costs, and (3) environmental factors. The weights of the cost factors were determined using pairwise comparisons in the Analytical Hierarchy Process (AHP). The value of site suitability was reclassified into three categories (non-suitable, low-suitable, and high-suitable) using different classification methods. With a feasible plant location defined as an industrial site within the most suitable area, the second stage of the analysis used compromise programming to compare the potential sites. The criteria used to rank the potential sites included fuzzy distance to woody biomass, highways, railways, commercial airports, communities, and available parcel size. The AHP was used to compute the relative importance of each criterion. The top ten suitable sites were determined, and sensitivity analyses were conducted to derive the most preferred sites. The approach was successful in taking a large amount of non-commensurate spatial data and integrating a site-based ranking algorithm to find the top locations for biomass plants. It also has great potential and applicability to other suitability and site selection studies.