Tuesday, August 5, 2008

PS 18-8: Describing ecosystem shifts with data-driven state and transition models in northwestern Colorado rangelands

Emily J. Kachergis, Colorado State University and Maria Fernandez-Gimenez, Colorado State University.

Background/Question/Methods
State and transition models are important tools to describe non-linear ecosystem dynamics and to aid in land management.  While many are created using expert knowledge, there is increasing interest in developing data-driven models and standardizing field and statistical methods.  We created a state and transition model to capture plant community response to different disturbances and rangeland management practices in a Northwestern Colorado watershed.  This model focuses on the Claypan Ecological Site, an NRCS classification of soil and climate factors that may result in similar plant community response to management practices.  Species cover, aboveground production, and current year's herbivory (percent utilization) were estimated on 35 plots located in this ecological site and stratified by management history.  Management history factors included historic grazing intensity (high, medium and low) and shrub treatment (mechanical, aerial spraying or none).  Measured environmental variables included slope, aspect and soil depth.  States were determined using non-metric multidimensional scaling of cover data together with herbivory, management history, and environmental variables.

Results/Conclusions
Preliminary results describe three alternate states:  lightly grazed sagebrush steppe, heavily grazed sagebrush steppe, and chemically managed grassland.  Removal of shrubs by aerial spraying of herbicide was strongly correlated with increased cover of grasses.  Herbivory was also strongly correlated with differences in plant community composition, suggesting that grazing may be an important mechanism driving transitions between alternate states.  Historic grazing intensity had a similar, but weaker influence than current herbivory.  The small number of alternate states suggests that this is a resilient ecosystem.  Future work will compare this model to one created from local knowledge, and relate water balance to herbivory as a mechanism for observed ecosystem shifts.