The goals of energy independence and sustainability have motivated many countries to consider biomass-based energy sources. The United States has substantial and increasing forest resources that could be used to produce both electricity and liquid fuel. However, these forest resources are highly heterogeneous in terms of the wood's properties, the logging cost, the spatial distribution, and the value to other industries. These factors make predicting costs and selecting plant locations particularly challenging. When dealing with forest biomass, feedstock cost and location have frequently been highly simplified in previous studies. This paper presents a methodology for combining highly resolved forest inventory and price data with records of competing industries to develop detailed maps of feedstock availability. The feedstock sourcing strategy of the proposed bioenergy plants is modeled by a cost-minimizing linear program, as is the feedstock selection of the competing mills. A case study is performed on the southeast United States. (C) 2010 Elsevier Ltd. All rights reserved.