Mineral control of soil carbon turnover in a savanna lithosequence in Kruger National Park, South Africa
Carbon-mineral aggregates sequester much of the carbon in soil and in some cases are responsible for slowing the rate of carbon return to the atmosphere. We do not have full understanding of how mineral control of carbon turnover plays out in soils dominated by differing mineral assemblages. Mineral-carbon bonding is facilitated by dehydration reactions and probably by hydrophobic attractions. The former is strongly favored by metastable short-range-order (SRO) minerals such as allophane and ferrihdydrite. These chemically active and hydrated minerals have received much research focus because it is clear that they are highly effective at sorbing and protecting carbon from decomposition. There has been less focus on crystalline silicate clays and metal oxy(hydr)oxides. Here we evaluate the role of different crystalline minerals in increasing the turnover time of carbon. Mixed shrub and grass savanna soils were sampled on basic to felsic lithologies with otherwise similar soil forming factors. The soils were sampled from an ancient craton and associated volcanics in South Africa. We separated carbon into light (LF) and heavy (HF) fractions and measured their isotopic compositions. In addition we measured the isotopic composition of clay-associated carbon in HF. Clay mineral distributions were quantified by post-processing of x-ray diffraction patterns.
The clay-size fraction was <10% of the <2-mm soil on granite and rhyolite and between 35 and 50% on basaltic rocks. The sum of smectite, halloysite+kaolinite, chlorite+mica, and crystalline iron oxides made up >90% of the clay fraction and there were almost no SRO minerals. Smectite made up >90% of the clay in soils forming on base-rich rocks but <10% in soils on felsic rocks. Kaolin minerals were identified in all soils, and dominated the clay fractions on the felsic rocks but were minimal on base-rich rocks. Heavy fraction (>1.7 g cm-3) carbon made up >85% of soil carbon, but except for smectite-rich basaltic soils the silicate clay-associated carbon was a small portion of HF carbon. In smectite-rich soils, HF turnover times (based on a one-pool model) were ~103 yr whereas for all other soils they were ~102 yr. Among the silicate clay fraction only the expandable 2:1 clays sorb and hold carbon for appreciable time. Crystalline iron minerals sorb more HF carbon than most silicate clays, possibly because they coat silicate surfaces, providing readily available surface area. There is strong spatial patterning to mineral controlled carbon storage and turnover across this landscape that likely exists in many others.