Digital map of soil organic carbon stocks (t/ha) at 0-30 cm in the Casanare flooded savannas of the Colombian Llanos (2019)

This raster dataset represents the spatial distribution of soil organic carbon (SOC) stocks (t/ha) at 0–30 cm depth in the flooded savannas of Casanare, modeled using Digital Soil Mapping (DSM) at a spatial resolution of 12m. The map is part of the study “Modeling the spatial distribution of soil organic carbon and carbon stocks in the Casanare flooded savannas of the Colombian Llanos” (https://doi.org/10.1007/s13157-023-01705-3). Total SOC stocks in the study area were estimated at 55.07 Mt, with a mean stock density of 83.1 ± 24.3 t ha⁻¹ in the upper 30 cm of soil. Of this total, 12.3% is stored in areas experiencing prolonged flooding (semi-seasonal savannas), which account for only 7.9% of the study area (664,752 ha). Covariates derived from vegetation cover, topography, and soil properties were identified as the main drivers explaining the spatial variability of SOC stocks. Methodology:The spatial distribution of soil organic carbon (SOC) stocks at 0–30 cm depth was modeled at a spatial resolution of 12 m using an Expert Knowledge (EK)–based Digital Soil Mapping (DSM) approach implemented through the Soil Inference Engine (SIE). The model was trained with SOC measurements from 80 sampling sites at two depth intervals (0–10 and 10–30 cm) and applied across four land-cover–based mapping units. Fuzzy membership functions were used to represent SOC–landscape relationships. Environmental covariates were selected based on field observations, prior knowledge, and exploratory analyses, while non-informative categorical variables were excluded. For a detailed description of the methodology and results, see: https://doi.org/10.1007/s13157-023-01705-3.