Forecasting Alpine Vegetation Change Using Repeat Sampling and a Novel Modeling Approach

Abstract

Global change affects alpine ecosystems by, among many effects, by altering plant distributions and community composition. However, forecasting alpine vegetation change is challenged by a scarcity of studies observing change in fixed plots spanning decadal-time scales. We present in this article a probabilistic modeling approach that forecasts vegetation change on Niwot Ridge, CO using plant abundance data collected from marked plots established in 1971 and resampled in 1991 and 2001. Assuming future change can be inferred from past change, we extrapolate change for 100 years from 1971 and correlate trends for each plant community with time series environmental data (1971-2001). Models predict a decreased extent of Snowbed vegetation and an increased extent of Shrub Tundra by 2071. Mean annual maximum temperature and nitrogen deposition were the primary a posteriori correlates of plant community change. This modeling effort is useful for generating hypotheses of future vegetation change that can be tested with future sampling efforts.