Abstract
We describe an expert system that will operate on continuous data and aggregate a landscape into biophysically defined units. General knowledge of ecosystem processes is coded as rules in order to aggregate data themes into ecologically similar areas. An ecosystem process model simulating daily forest stand fluxes of carbon and water is applied on the aggregated database for regional estimates. Aggregations preserve essential information while reducing data to a manageable size. The expert system classifies the landscape by evaluating which of the three key landscape attributes (leaf area index, soil water holding capacity, or climatic efficiency) limits forest growth. Adjacent units in the same class are aggregated. Uncertainty existing between discrete rules and continuous data, and uncertainty originating from conflicting evidence in the classification process are integrated and used in the final classification and aggregation. -Authors
Original language | English |
---|---|
Pages (from-to) | 35-43 |
Number of pages | 9 |
Journal | AI Applications in Natural Resource Management |
Volume | 3 |
Issue number | 4 |
State | Published - 1989 |