Plant community response to climate change will be influenced by individual plant responses that emerge from competition for limiting resources that fluctuate through time and vary across space. Projecting these responses requires an approach that integrates environmental conditions and species interactions that result from future climatic variability. Dryland plant communities are being substantially affected by climate change because their structure and function are closely tied to precipitation and temperature, yet impacts vary substantially due to environmental heterogeneity, especially in topographically complex regions. Here, we quantified the effects of climate change on big sagebrush (Artemisia tridentata Nutt.) plant communities that span 76 million ha in the western United States. We used an individual-based plant simulation model that represents intra- and inter-specific competition for water availability, which is represented by a process-based soil water balance model. For dominant plant functional types, we quantified changes in biomass and characterized agreement among 52 future climate scenarios. We then used a multivariate matching algorithm to generate fine-scale interpolated surfaces of functional type biomass for our study area. Results suggest geographically divergent responses of big sagebrush to climate change (changes in biomass of -20% to +27%), declines in perennial C₃ grass and perennial forb biomass in most sites, and widespread, consistent, and sometimes large increases in perennial C₄ grasses. The largest declines in big sagebrush, perennial C₃ grass and perennial forb biomass were simulated in warm, dry sites. In contrast, we simulated no change or increases in functional type biomass in cold, moist sites. There was high agreement among climate scenarios on climate change impacts to functional type biomass, except for big sagebrush. Collectively, these results suggest divergent responses to warming in moisture-limited vs. temperature-limited sites and potential shifts in the relative importance of some of the dominant functional types that result from competition for limiting resources.

We applied STEPWAT2, a plant community, gap dynamics model that integrates an annual time step, individual-based plant simulation model (STEPPE) and a daily, process-based soil water balance model (SOILWAT2) to understand plant community response to climate change in 200 big sagebrush plant communities. STEPWAT2 simulates establishment, competition, growth, and mortality of multiple plant species and functional types based on life-history traits and on soil water availability in multiple soil layers, which is frequently the limiting resource in dryland plant communities.We used version 1.0.0 of an R program called rSFSTEP2 (https://github.com/DrylandEcology/rSFSTEP2/releases/tag/v.1.0.0) to run STEPWAT2 simulations for all 200 sites for 300 years and 200 iterations. Simulations were conducted for 300 years because STEPWAT2 initiates with no vegetation, and it takes ~100 years for plant communities to reach a steady-state. Thus, all analyses here focus on characterizing biomass in the last 200 years after steady state conditions were reached but where biomass fluctuated due to stochastic (i.e. establishment) and deterministic (i.e. mortality due to low resource availability) processes. We simulated each site under current climatic conditions (1980-2010) and future climatic conditions derived from 13 Global Climate Models (GCMs) for representative concentration pathways RCP4.5 and RCP8.5 for both mid-century (2030-2060) and end-century (2071-2100) for a total of 52 future climate scenarios. We simulated plant communities under a light grazing regime where grazing occurred annually and was implemented by removing a fraction of the current year’s growth for each functional type.

We calculated the mean biomass and density of each functional type over the last 200 years of the simulations for each site and RCP/GCM/time period combination (N = 53). This yielded 200 means (one for each site) for each functional type/RCP/GCM/time period combination (for a total of 424 means per site). Thereafter, for each site/functional type/RCP/time period combination, we sorted mean biomass and mean density (N = 13) for the 13 GCMs and identified the median value. We used the median as a measure of central tendency rather than the mean to moderate the impact of extreme values from individual GCMs. For each site, functional type, RCP, and future time period (2030-2060, 2070-2100), we calculated the absolute change in median biomass and median density from current to future conditions and percentage change scaled to the maximum biomass simulated under current conditions.To investigate if climate change will alter the biomass and composition of plant functional types, we quantified absolute and percentage change in median biomass for each functional type and examined whether the relative importance of functional types changed from current to future conditions by ranking median biomass under current conditions and examining whether that ranking changed in the future. To assess how changes in functional type biomass and composition vary along environmental gradients dictated by climate and ecohydrology, we quantified relationships between change in median biomass (g/m2) for the dominant plant functional types (big sagebrush, perennial C3 grasses, perennial C4 grasses, and perennial forbs), and climatic and ecohydrological variables: (annual precipitation (mm), precipitation falling as rain (mm), rain/precipitation ratio, and mean annual temperature (°C) snowpack (SWE, mm), the number of wet days in surface (0-30 cm) and subsurface (30-160 cm) soil layers, and the transpiration from surface and subsurface soil layers). 

Data are mean values over the last 200 years of the simulation for current conditions and each GCM-RCP-time period combination (N=52) for each of 200 sites. There are no missing values - the data are complete. Values presented in Palmquist et al. (2021) are primarily the median of these values across the 13 GCMs. Variables represent plant functional type biomass and density, climatic, and ecohydrological variables described in Palmquist et al. (2021).