GIS and Large-scale Linear Programming: Evolution of a Spatial Decision Support System for Land Use Management

Forest management and planning is a major task faced by private timber companies and public agencies. As an example, the U.S. Forest Service is responsible for managing over 190 million acres of forests and grasslands in the U.S., an area that is almost twice the size of Germany. Thus, it is not surprising that the U.S. Forest Service uses many types of analysis tools including the use of input-output models to assess the impact of forest operations on local and regional economies. The most common approach to plan activities for a forested region over decades involves decomposing the decision making process into three hierarchically-defined components: strategic, tactical, and operational (Weintraub and Cholaky, 1991; Nelson, Broadie, and Sessions, 1991). Each level of management involves the development of decision support models, either optimization or simulation. These models aid in exploring alternatives and generating tradeoffs among objectives. Strategic models help identify overall targets for harvesting and conservation activities over a long planning horizon and over large tracts of land. Tactical models help specify where the activities should be accomplished in order to meet strategic targets. Finally, operational models help translate tactical level decisions into detailed operations plans. This type of hierarchical decomposition has been used in forest planning by governmental agencies and by private industry (Church, Murray and Barber, 1994).