Aim: Montane environments are sentinels of global change, providing unique opportunities to assess impacts on species diversity. Multiple anthropogenic stressors such as climate change and atmospheric pollution may act concurrently or synergistically in restructuring communities. Thus, a major challenge for conservation is untangling the relative importance of different stressors. Here, we combine long-term monitoring with multivariate community modeling to estimate the anthropogenic drivers shaping forest tree diversity along an elevational gradient. Location: Camels Hump Mountain, Vermont, USA Methods: We used Generalized Dissimilarity Modelling (GDM) to model spatial and temporal turnover in beta diversity along an elevational gradient over a 50-year period, and tested for spatiotemporal shifts in density and elevational distribution of individual species. GDMs were used to predict community turnover as non-linear functions of changes in elevation, climate and atmospheric pollution. Results: We observed significant shifts in elevational range and density of individual species, which contributed to an overall reduction in the elevational gradient in beta diversity through time. GDMs showed the combined effects of sulfate deposition and temperature as drivers of this temporal reduction in beta diversity. Spatiotemporal changes differed among species, with shifts observed both up and downslope. For example, in a reversal of a previous upslope range contraction, red spruce (Picea rubens Sarg.) increased in density and shifted downslope since the 1990’s, occupying warmer, drier climates. Main conclusion: Our results demonstrate that global change is impacting the stratification of forest tree diversity along elevational gradients, but the responses of individual species are complex and variable in direction. We suggest abiotic drivers may directly impact individual species while also indirectly altering species interactions along elevational gradients. Our approach modelling the drivers of compositional turnover quantifies the rate and amount of change in beta diversity along environmental gradients, and serves as a powerful complement to studying species-specific responses.

In 1964, a long-term forest monitoring study was established on Camels Hump, a 1248m peak in the Green Mountains of Vermont. Inventory stands were established at intervals of approximately 60m along an elevational gradient from 549 to 1,159m on the mountain’s western slope. At each stand (N=11), five to ten permanently marked survey plots, each 3.0 x 30.5m, were placed parallel to each other and roughly perpendicular to the contour line. Within each plot (N=85), species identity and diameter at breast height (dbh; 1.37m above ground) of all woody stems > 2cm dbh were recorded during each of nine census years (1965, 1979, 1983, 1986, 1990, 1995, 2000, 2004, 2015).

This is a comma separated file containing counts of forest trees by species for each census year.

There are no missing values in the dataset.

diam: Diameter at breast height (dbh) of each tree

elevation: elevation of plot measured in meters above sea level (a.s.l.)

rounded_ele: elevation of plot measured in meters above sea level (a.s.l.) rounded to nearest ten meters

species: unique value assigned to each species in dataset