Soil Organic Carbon vs. Bulk Density Following Temperate Grassland Afforestation

Abstract

Afforestation is part of a worldwide strategy to mitigate CO2 emissions. However, afforestation in grassland soils may have the opposite effect by promoting the loss of native carbon. Potential effects of this land use change on the flow of organic carbon to and from the soil can be described through bulk density (Db). Nowadays the suitability of Db for this purpose is being questioned. In order to bring new elements to the discussion, we carried out a comparative study of soil in the western region of Uruguay. Based on the background information and our own data, collected for over a decade, we evaluated the fitness of Db as proxy soil organic carbon (SOC) stocks in grassland converted to tree afforestation. These data were also related to soil pH values. The sampling consisted of five plots afforested with Eucalyptus grandis paired with control plots under grassland. All samples were taken at depth (0–10 cm and 10–20 cm) except for Db samples (0–10 cm). In afforested sites, Db increased (1.62 vs 1.53 g/cm3; p ≤ 0.01) and SOC decreased (0–10 cm: 0.90 vs. 1.22%; p ≤ 0.08). Db values were not significantly correlated with SOC content. As with SOC, pH values decreased after afforestation at both depths (0–10 cm: 4.92 vs. 5.62; p ≤ 0.01; 10–20 cm: 4.76 vs. 5.54; p ≤ 0.01). The high acidity generated in soils following afforestation, is enough to affect the interaction between mineral and organic fractions and, with them, the original Db values. According to a previous study in the same location, there is a change in the predominance of different clay minerals in the topsoil (0–20 cm). This qualitative change in the mineral fraction can affect the ability of the soil to retain organic carbon, and not be reflected in the recorded Db values. The non-reciprocity recorded between Db and SOC values warns about the need for restriction of the generic use of Db in calculation of SOC stocks estimation. In view of these results, we present a discussion of possible causes that explain the disparity between Db values and SOC measurements.