No abstract

Understanding long‐term changes in ecological communities during global change is a priority for 21st‐century ecology. Deserts, already at climatic extremes, are of unique interest because they are projected to be ecosystems most responsive to global change. Within a 500‐km2 landscape in the Mojave Desert, USA, we measured perennial plant communities at 100 sites three times (1979, 2008, and 2016) during 37 yr to evaluate six hypotheses of community change. These hypotheses encompassed shifts in community measures (e.g., diversity, cover) and species elevational distributions, biotic homogenization, disproportionately large change at the highest elevations, relationships between turnover and species’ responses to disturbance and drought, and that environmental refugia (e.g., moist topographic positions) would receive species during climatic warming and drying. Most community measures changed temporally, such as species density (species/600 m2) increasing 23% and plant cover doubling between 1979 and 2016. There was no increase in nonnative species and minimal evidence for biotic homogenization. High‐elevation communities did not display greater change than low‐elevation communities. Moreover, environmental refugia factored little in species shifts. While species distributional changes were unrelated to affinity for disturbance, the six most persistent species (persisting on >80% of sites) were long‐lived shrubs not associated with disturbance. Overall, seemingly paradoxically, climatic warming and drying was associated with increasing plant abundance. Comparing the 1970s to 2007–2016, precipitation in the study area declined 16% from 17 to 14 cm/yr and average daily minimum temperature rose 13% (1.2°C). The current climate with fewer freezes, together with reduced grazing, could be among the most optimal for desert perennials in the past century, although potential response lags to continuing warming and drying are uncertain. This study of long‐term elevational shifts in communities during global change is among few in deserts, and the average upward elevational shift of 6 m/decade for species in our study is within the range reported for temperate biomes. However, the 41% of species moving downslope is unusually high. We propose that dynamics within desert perennial communities follow a core‐transient species model where a site's species are either highly persistent or transient in approximately equal proportions.

Aim: Modelling and quantifying habitat changes using three historical field surveys for four Mojave Desert species and projecting future scenarios. Location: Newberry Mountains, southern Nevada, USA. Methods: Three vegetation field surveys were conducted, ending in 1979, 2008 and 2016, respectively. Field data collection across three time steps resulted in a unique dataset with 100 re-surveyed 0.06-ha plots. Using Maxent ecological niche modelling and 800-m resolution Parameter Elevation Regressions on Independent SlopesModel (PRISM) temperature and precipitation data, habitat was assessed for four high-elevation species. Recognizing that missing location data is a common challenge in species distribution modelling, sensitivity analysis was conducted utilizing three high-elevation plots that species inhabited during the first two field surveys but that were unable to be sampled in 2016. Area under the curve for three species' model runs exceeded 0.95, while one wider-distributed species exceeded 0.79 for all runs. Future scenarios were also modelled under temperature increases and 16% and 0% reductions in precipitation. Results:The number of presence locations and densities for Juniperus californica, Pinus monophylla, Quercus turbinella and Yucca schidigera decreased from 1979 to 2016, while their average elevational distribution increased. Using the 100 plots measured each of the three study years, models for all four species indicated decreases in potential suitable habitat ranging from 10% to 45%. When including the three highest elevation plots not re-sampled in 2016, but likely still containing the modelled species, modelled habitat loss ranged from 0% to 29%. Future model projections show drastic reductions in the suitable habitat of all four modelled species.Main conclusions: While many Maxent ecological modelling studies have been conducted, few have incorporated repeat surveys of identical plots as in our study. Model projections using a unique 100-plot dataset suggest that suitable habitat area is contracting for the species studied, and this correlated with climatic warming and drying in the Mojave Desert. Despite model projections that show nearly complete loss of suitable habitat in 2053, care should be taken given uncertain estimates of the velocity of climate change and the slower pace of habitat contractions measured in the field.

No abstract