Rapid delineation of preliminary Ecological Sites applied to forested northern Appalachian landscapes

The fundamental assumption underpinning Ecological Sites is that soils, topographic, and climatic variables can be correlated with composition and productivity of plant communities with sufficient precision to subdivide landscapes by their inherent ecological potential and management needs. If Ecological Sites are to form the basis of management plans across the United States, then methods need to be standardized and refined, especially outside of rangeland ecosystems where this conceptual framework was pioneered. We present an approach to rapid delineation of preliminary Ecological Sites by using readily accessible geospatial data in a geographic information system (GIS) and common multivariate statistical methods. We applied this methodology to ∼815,000 ha of publically owned terrestrial forests in Pennsylvania. We generated random sampling points through a data layer containing ∼58,000 polygons representing mapped occurrence of 19 types of terrestrial forest stands. We used these points to sample 11 environmental variables spanning the soils (depth, bulk density, permeability, and pH), topographic (elevation, topographic position index, slope, curvature, and folded aspect), and climatic (mean annual temperature and precipitation) data categories. We used hierarchical cluster analysis of z-score-standardized means for each environmental variable to group forest stand types into preliminary Ecological Sites. We then used principal components analysis to explore the distribution of forest stand types within and among preliminary Ecological Sites in ordination space and to understand the influence of each environmental variable on the occurrence of forest stand types. This approach led to logical groupings that we hypothesize represent Ecological Sites and constituent states or phases.