Reduced temperature sensitivity of soil respiration after a 17-year climate change experiment

In 1994, a reciprocal soil transplant experiment was initiated between two elevations (310 m, warmer and drier, and 844 m, cooler and wetter) on Rattlesnake Mountain in southeastern Washington, USA, testing whether the microbial and biochemical dynamics that developed under cool, moist conditions would be destabilized under hot, dry conditions. In March 2012 we resampled the original transplanted soils to study longer-term changes in microbial community composition, soil C and N dynamics, and soil physical structure. These resampled cores were randomly assigned to climate-control chambers simulating the lower or upper site climates. We measured respiration throughout a 100-day incubation, coupled with biogeochemical analyses, to examine how these soils had responded to environmental changes over 17 years. Temperature and soil moisture were the primary drivers of CO2 evolution, but transplant source and destination both exerted significant effects. Most strikingly, respiration from cores originally from the hotter, low-elevation site that spent 17 years at the upper site exhibited almost no temperature sensitivity (Q10=1.07, 13-33 °C). Cores from the upper site had more carbon (~1.1% versus 0.8%), but equivalent C:N ratios, while soils incubated in the 'upper' chamber had greater N-acetylglucosaminidase and β-glucosidase potentials. Tomographic reconstructions revealed that porosity, moisture content, grain size distribution, and organic C were highly heterogeneous, consistent with the observed macro-scale variability. These results suggest that the upper-site soils were more resilient to the 1994 transplant, but that there is a significantly altered microbial community in the transplanted soils, particularly the lower-to-upper cores, that has not recovered almost two decades after the original experiment. This raises more general questions of how current climate change will affect soil resistance to future perturbations, and how confidently we can model this response at larger scales.