Forecasting climate change impacts on plant populations over large spatial extent

Andrew T. Tredennick, Utah State University
Mevin B. Hooten, Colorado State University
Cameron L. Aldridge, Colorado State University
Collin G. Homer
Andrew R. Kleinhesselink, Utah State University
Peter B. Adler, Utah State University


Plant population models are powerful tools for predicting climate change impacts in one location, but are difficult to apply at landscape scales. We overcome this limitation by taking advantage of two recent advances: remotely sensed, species-specific estimates of plant cover and statistical models developed for spatiotemporal dynamics of animal populations. Using computationally efficient model reparameterizations, we fit a spatiotemporal population model to a 28-year time series of sagebrush (Artemisia spp.) percent cover over a 2.5 × 5 km landscape in southwestern Wyoming while formally accounting for spatial autocorrelation. We include interannual variation in precipitation and temperature as covariates in the model to investigate how climate affects the cover of sagebrush. We then use the model to forecast the future abundance of sagebrush at the landscape scale under projected climate change, generating spatially explicit estimates of sagebrush population trajectories that have, until now, been impossible to produce at this scale. Our broadscale and long-term predictions are rooted in small-scale and short-term population dynamics and provide an alternative to predictions offered by species distribution models that do not include population dynamics. Our approach, which combines several existing techniques in a novel way, demonstrates the use of remote sensing data to model population responses to environmental change that play out at spatial scales far greater than the traditional field study plot.