PARENT SESSION
Wednesday, August 9, 1:30-5:00 pm
COS 71 - Landscape ecology
Sultana, Mezzanine Level, Cook Convention Center
Presiders: G Benson

Comparing conceptual and empirical range vegetation classification systems for the Intermountain West.

Adair, William^*,1, Ramsey, R Douglas¹, Lowry, John¹, ¹ Utah State University, Logan, UT

ABSTRACT- Most classification systems used to describe Intermountain rangelands, including SRM Range Cover Types and NVC Ecological Systems, Alliances, and Associations, are based on a priori concepts of how rangeland communities are organized. Unfortunately, these systems frequently include class definitions with dissimilar emphases, varying information content, inconsistent taxonomic resolution, undefined spatial scale, and conflicting temporal models. These shortcomings became apparent to us during the course of the Southwest Regional GAP Analysis Program (SWReGAP) land cover mapping process. This experience encouraged us to investigate an empirical approach to partitioning landscapes, which we believed could provide insight into how to improve existing class definitions. To achieve this objective, we applied a cluster analysis to SWReGAP field training data collected in the northern Colorado Plateau and eastern Great Basin. Our search for the optimal configuration of classes focused on identifying where differences in degree transition into differences in kind. We also created alternative datasets to address concerns typical of vegetation classification systems in general (e.g., dominant species vs. whole communities) and specific to Intermountain rangelands (e.g., include or exclude substrate measures). Although the classes that emerged from our analysis share much in common with existing systems, our results also suggest new ways for resolving problems associated with defining land cover classes and plant communities, particularly those associated with information content.

Key words: classification, landscape ecology, rangelands