Limited protection of macro-aggregate-occluded organic carbon in Siberian steppe soils

Abstract

Abstract. Macro-aggregates especially in agricultural steppe soils are supposed to play a vital role for soil organic carbon (OC) stabilization at a decadal timescale. While most research on soil OC stabilization in steppes focused on North American prairie soils of the Great Plains with information mainly provided by short-term incubation experiments, little is known about the agricultural steppes in southwestern Siberia, though they belong to the greatest conversion areas in the world and occupy an area larger than that in the Great Plains. To quantify the proportion of macro-aggregate-protected OC under different land use as function of land use intensity and time since land use change (LUC) from pasture to arable land in Siberian steppe soils, we determined OC mineralization rates of intact (250–2000 µm) and crushed (< 250 µm) macro-aggregates in long-term incubations over 401 days (20 °C; 60 % water holding capacity) along two agricultural chronosequences in the Siberian Kulunda steppe. Additionally, we incubated bulk soil (< 2000 µm) to determine the effect of LUC and subsequent agricultural use on a fast and a slow soil OC pool (labile vs. more stable OC), as derived from fitting exponential-decay models to incubation data. We hypothesized that (i) macro-aggregate crushing leads to increased OC mineralization due to an increasing microbial accessibility of a previously occluded labile macro-aggregate OC fraction, and (ii) bulk soil OC mineralization rates and the size of the fast OC pool are higher in pasture than in arable soils with decreasing bulk soil OC mineralization rates and size of the fast OC pool as land use intensity and time since LUC increase. Against our hypothesis, OC mineralization rates of crushed macro-aggregates were similar to those of intact macro-aggregates under all land use regimes. Macro-aggregate-protected OC was almost absent and accounted for < 1 % of the total macro-aggregate OC content and to a maximum of 8 ± 4 % of mineralized OC. In accordance to our second hypothesis, highest bulk soil OC mineralization rates and sizes of the fast OC pool were determined under pasture, but mineralization rates and pool sizes were unaffected by land use intensity and time since LUC. However, at one chronosequence mean residence times of the fast and slow OC pool tended to decrease with increasing time since establishment of arable use. We conclude that the tillage-induced breakdown of macro-aggregates has not reduced the OC contents in the soils under study. The decline of OC after LUC is probably attributed to the faster soil OC turnover under arable land as compared to pasture at a reduced plant residue input.