Assessing the Relative Utility of Models of Vegetation Dynamics for the Management of Sagebrush Steppe Rangelands.

Hosten, Paul

This PhD dissertation lists in Appendix B common plants from study sites used by him or others. Study sites included in part Carey Kipuka, Little Park, and Laidlaw Park, Idaho, all within CRMO as areas administered by NPS or BLM. The literature, long-term temporal data sets, and spatial data points surrounding livestock watering points within rangeland communities were used to examine common assumptions and concepts used in models of vegetation dynamics. Of the stability concepts examined, the constancy concept was invalidated using long- XIV term relict area data, whereas long-term data following disturbance indicated that sagebrush-dominated areas could be considered as resilient under the circumstances examined. Piosphere as well as relict and postdisturbance data indicated that the individualistic notion of vegetation change was favored. Species showing the most predictability (based on repeatability of abundance curves, and spatial and temporal serial correlations) were sagebrush and cheatgrass, specifically in their response to fire. Bunchgrasses were characterized by their persistence in abundance under all circumstances (grazing disturbance and fire disturbance) except long-term absence of fire. This, together with the support of the individualistic notion of plant community change, suggests examining change on a vital attributes basis within a kinetic framework. It was found that most models of vegetation dynamics could be validated, because of the multivariate nature of the data examined. A closer examination of the requirements of range management for a model to be used as a basis for range condition and trend analysis determined that three basic model types are needed by range managers: First, a conceptual model for marrying divergent perspectives of different rangeland user groups; second, mechanistic models for the exploration of vegetation dynamics; and third, expert systems for infonnation transfer. Further emphasis was placed on deriving a suitable conceptual model subscribing to the objective of linking vegetation dynamics, range management, and monitoring within sagebrush-dominated areas. Important characteristics (derived from the data sets examined) included in this model are the recognition of static and dynamic plant communities, swathes of change rather than predefined pathways, the individualistic nature of change, and the recognition of spatial heterogeneity. Predictability at various scales was incorporated by focusing on cheatgrass and sagebrush, since these species indicated the most repeatable dynamics observed in the data sets, and were "common denominator" species (thus overcoming the problem of spatial heterogeneity). The undesired effects of cheatgrass and sagebrush on other growth-forms suggest that monitoring and managing these undesired species may be a proactive way of managing for a desired plant community incorporating a balance of growth-forms.