The development and analysis of future agricultural landscapes

Graduation date: 2003Alternative futures studies provide a way for policy-makers and stakeholders to\ud investigate the future impacts of different management strategies by providing\ud representations of possible outcomes. An alternative futures study was undertaken by the\ud Pacific Northwest-Ecosystem Research Consortium for the Willamette River Basin,\ud Oregon, USA, to determine how this regional system would respond to various drivers of\ud change. To support this effort required the development of a simulation model that\ud generates future agricultural landscapes in response to these drivers of change and the\ud evaluation of the resulting landscapes to derive information useful for policy analysis and\ud community discussions. These goals were achieved by creating a characterization of the\ud initial agricultural system, developing an agricultural land-cover change model, and finally,\ud integrating this model with scenario elements to produce the final agricultural landscape\ud evolution model. An important part of the initial landscape characterization was\ud determining the crop suitability class of the basin's agricultural soils for 14 commonly\ud grown crops. Suitability rankings were generated by using rough-set rule induction to\ud create predictive if-then rules that related a soil's biophysical characteristics to crop production and then applying these rules to all the agricultural soils in the basin. After\ud characterizing the agricultural system, a spatially explicit, multi-attribute decision model\ud was developed to simulate a grower's crop-selection decision. This model used attributes\ud describing a field's biophysical conditions and each crop's agronomic, economic, and\ud management characteristics to select a field's preferred crop from a list of crop alternatives.\ud Next, rules and constraints were formulated for each of the three policy scenarios: a\ud continuation of current trends, an increased reliance on market forces to determine land\ud use, and an increased emphasis on environmental restoration programs. The initial\ud agricultural landscape depiction, crop selection model, and scenario rules were used to\ud generate three future landscapes, each depicting an alternative state of the agricultural\ud system in the year 2050. Then, landscape metrics and screening models were used to assess\ud the agronomic and environmental condition of each agricultural landscape. Finally, this\ud information was placed into the context of the Willamette River Basin, with suggestions\ud for policy development and analysis