Detecting changes in arctic tundra plant communities in response to warming over decadal time scales

Detecting the response of vegetation to climate forcing as distinct from spatial and temporal variability may be difficult, if not impossible, over the typical duration of most field studies. We analyzed the spatial and interannual variability of plant functional type biomass from field studies in low arctic tussock tundra and compared these to climate change simulations of plant community composition using a dynamic tundra vegetation model (ArcVeg). Spatial heterogeneity of peak season live aboveground biomass was estimated using field samples taken from low arctic tundra at Ivotuk, Alaska (68.5degreesN, 155.7degreesW) in 1999. Coefficients of variation for live aboveground biomass at the 1 m(2) scale ranged from 14.6% for deciduous shrubs, 18.5% for graminoids and 25.3% for mosses to over 57% for forbs and lichens. Spatial heterogeneity in the ArcVeg dynamic vegetation model was simulated to be greater than the field data, ranging from 37.1% for deciduous shrubs to 107.9% for forbs. Disturbances in the model, such as caribou grazing and freezing-thawing of soil, as well as demographic stochasticity, led to the greater variability in the simulated results. Temporal variances of aboveground live biomass over a 19-year period using data from Toolik Lake, AK fell within the range of field and simulation spatial variances. However, simulations using ArcVeg suggest that temporal variability can be substantially less than site-scale spatial variability. Field data coupled with ArcVeg simulations of climate change scenarios indicate that some changes in plant community composition may be detectable within two decades following the onset of warming, and shrubs and mosses might be the key indicators of community change. Model simulations also project increasing landscape scale spatial heterogeneity (particularly of shrubs) with increasing temperatures.