| Across landscapes, the availability of suitable habitat controls the distribution of plant species. This study tested the predictive capability of site and landscape-level data for identification of potential rare plant habitat. Four rare endemic plant species were chosen to include a range of environmental variability, community associations, and elevations. Herbarium records were used to locate known occurrences. These sites were visited during the 1996 season. Coordinates, environmental attributes, and associated species data collected at each site were used to develop two predictive models: a field key and a probability-of-occurrence map. The field key was developed from environmental attribute and associated species data collected at the sites. Maps were developed from the coordinate data collected at each site by locating the position of each site in a geographic information system (GIS) containing slope, elevation, aspect, soil, and geologic data. A tree-classification-based dichotomous key was used to generate the field keys and the maps of occurrence probabilities. The map was field validated during the 1997 field season. Development of the models is an iterative process, where data collected during field validation were incorporated into subsequent predictive models. Results suggest potential habitat can be identified, by combining elevation, slope, aspect, rock type, and geologic process data. Soil data rarely contributed to predictions. |
