Data from: Thermal segregation drives patterns of alder and willow expansion in a montane ecosystem subject to climate warming

1.Tall-shrub expansion into low-statured communities, a hallmark of recent
vegetative change across tundra ecosystems, involves three major genera:
Alnus, Betula, and Salix. Which genus expands most into tundra landscapes
will determine ecosystem properties. 2.We show that Alnus and Salix shrubs
segregate thermal space (elevation x insolation) and colonize tundra
landscapes differently in response to climate warming, thereby replacing
different tundra types. 3.Vegetative change estimated from repeat
photography should account for hill-slope. Methodologically, slope
determines surface area estimated from orthophotos as projected pixel area
times secant of pixel slope. Ecologically, the change in
thermally-responsive vegetative area is sensitive to terrain steepness,
scaling as the cosecant of hill-slope, so that studies should expect more
shrub expansion in areas of shallow slopes than steep slopes. 4.Repeat
aerial photography in Alaska's Chugach Mountains from 1972-2012
orthorectified on high-resolution lidar DEM indicated tall Salix was rare
in 1972 and colonized warmer slopes by 2012. Tall Alnus colonized steeper,
cooler slopes both by 2012 and by 1972. Salix and forest colonized similar
thermal space. Colonization probability for both shrub genera was
maximized at intermediate elevations. 5.Alnus colonization adjacent to
dwarf-shrub tundra was twenty-times as likely as Salix colonization. Salix
colonization adjacent to low-shrub/herbaceous tundra was three-times as
likely as Alnus colonization. Replacement of dwarf-shrub tundra by Alnus
and of low-shrub/herbaceous communities by Salix will affect herbivores
and soil properties. 6.Good agreement between observations of plant
functional type and multinomial predictions in a thermal space defined by
elevation and insolation suggested that these two variables were
sufficient for forecast modeling. Spatially explicit, climate-driven GLM
multinomial and random forest classification models in available thermal
space forecast surface areas of forest, Alnus, Salix, and tundra over a
range of warming, modeled as upward shifted isotherms, including expected
IPCC scenarios. Both modeling approaches indicated that shrubs may respond
non-linearly to warming. 7.Synthesis The provision of taxon-specific
coefficients for climate-driven, spatially-explicit models using high
resolution digital elevation models is necessary for accurately
forecasting vegetative change due to climate warming in montane and arctic
regions.