Using long-term ecosystem experiments to study contemporary climate change: Results and comparisons from a 23 year warming experiment
Ecosystem responses to climate change can be mediated by slow-moving factors, such as shifts in vegetation community composition or soil carbon content. Long-term experimental manipulations help address these slow changes, but also present another opportunity: to infer the extent to which contemporary climate change is responsible for changes in ecosystems observed under ambient conditions. In this study, a 23-year warming experiment was used to demonstrate convergent effects across both experimentally warmed and ambient (untreated), plots. In 1989, ten 3×10m plots were established in ungrazed alpine meadow. Above every other plot, overhead infrared radiators have cast a year-round downward heat flux onto the soil surface since January, 1991 (~15 watts m−2 prior to June 1993, and 22 watts m−2 subsequently). Four response variables were monitored over time and are considered here: date of snowmelt, aboveground biomass (AGB) of forbs and shrubs, and soil organic carbon. We sought to describe not only how these variables had changed over time, but how experimental warming reflected changes observed in ambient plots under real-world climate conditions.
Date of snowmelt advanced significantly in both experimental and ambient plots (7.7 days yr-1, P<0.04 heated; 0.74, P<0.0002 ambient). In both types, forb AGB declined significantly over time (−3.49 g m-2 since 1990, P<0.0001 heated; −2.23 g m-2 since 1990, P=0.012 ambient). In contrast, shrub AGB in both plots increased over time (5.69 g m-2 since 1990, P<0.001; 1.78 g m-2, P<0.054). Soil organic carbon (SOC) decreased in the experimental plots after the onset of heating and then, after reaching a minimum, slowly increased; ambient plot SOC slowly decreased. Previously, effects of experimental warming on SOC have been explained as the interaction of rapid decreases in abundant, labile forb biomass and slow increases in less abundant but more recalcitrant shrub biomass (Harte et al, 2006). Based on analyses of the ambient plots, we here propose that background climate conditions are slowly decreasing forb and increasing shrub biomass, leading to a downward trend in ambient plot SOC similar to that seen in experimental plots during the early stages of the experiment. Findings demonstrate the realism of an experimental manipulation and demonstrate how long-term study of ambient and experimental responses may help identify mechanistic drivers behind changing ecosystems.