Articles | Volume 11, issue 4
Research article
20 Feb 2014
Research article |  | 20 Feb 2014

The impact of extreme summer drought on the short-term carbon coupling of photosynthesis to soil CO2 efflux in a temperate grassland

S. Burri, P. Sturm, U. E. Prechsl, A. Knohl, and N. Buchmann

Abstract. Along with predicted climate change, increased risks for summer drought are projected for Central Europe. However, large knowledge gaps exist in terms of how drought events influence the short-term ecosystem carbon cycle. Here, we present results from 13CO2 pulse labeling experiments at an intensively managed lowland grassland in Switzerland. We investigated the effect of extreme summer drought on the short-term coupling of freshly assimilated photosynthates in shoots to roots as well as to soil CO2 efflux.

Summer drought was simulated using rainout shelters during two field seasons (2010 and 2011). Soil CO2 efflux and its isotopic composition were measured with custom-built chambers coupled to a quantum cascade laser spectrometer (QCLAS-ISO, Aerodyne Research Inc., MA, USA). During the 90 min pulse labeling experiments, we added 99.9 atom % 13CO2 to the grass sward. In addition to the isotopic analysis of soil CO2 efflux, this label was traced over 31 days into bulk shoots, roots and soil.

Drought reduced the incorporation of recently fixed carbon into the shoots, but increased the relative allocation of fresh assimilates below ground compared to the control grasslands. Contrary to our hypothesis, we did not find a change of allocation speed in response to drought. Although drought clearly reduced soil CO2 efflux rates, about 75% of total tracer uptake in control plots was lost via soil CO2 efflux during 19 days after pulse labeling, compared to only about 60% under drought conditions. Thus, the short-term coupling of above- and below-ground processes was reduced in response to summer drought. The occurrence of a natural spring drought in 2011 lead to comparable albeit weaker drought responses increasing the confidence in the generalizability of our findings.

Final-revised paper