CarboSOIL model and climate outputs from two GCMs (GISS and HadCM3), three time horizons (2020, 2050, 2080), and two emission scenarios (A2 and B2) according to IPCC were used to study the effects of climate change on soil organic carbon (SOC) stocks in a Mediterranean region (Northeast Sardinia, Italy). CarboSOIL is an empirical model based on regression techniques and developed to predict SOC contents at standard soil depths of 0-25, 25-50 and 50-75 cm. The area is characterized by extensive agro-silvo-pastoral systems, and six land uses with different levels of cropping intensification were compared: tilled vineyards (TV), no-tilled grassed vineyards (GV), hay crop (HC), pasture (PA), cork oak forest (CO), and semi-natural systems (SN). The main objectives were: (i) to validate the model predictions with the measured SOC stocks, and (ii) to predict SOC stocks in future climate projections for the different land use types. The model proved its ability to predict SOC stocks at different soil depths, and can be used as a tool for predicting SOC stocks under different climate change scenarios. The results suggest that future climatic scenarios can have a negative effect on SOC stocks in the upper sections of the soil profile, mainly due to a very low increase in the 0-25 cm section and a sharp decrease in the 25-50 cm soil section, in particular in a long term perspective (2080) and under the emission scenario A2. Important decreases of SOC stocks were found in the upper soil sections of the vineyards.
