This raster dataset models soil organic carbon stocks at 0-30 cm depth for the Casanare flooded savannas of Colombia at a 12-meter spatial resolution. The model was trained with SOC measurements from 80 sampling sites and estimates total SOC stocks in the study area at 55.07 million tonnes, with a mean density of 83.1 t/ha. It was produced by J.M. Martín-López and published in 2023.
Use Cases
- Analyze the spatial variability of SOC stock density (t/ha) across four land-cover-based mapping units.
- Model the relationship between SOC stocks and environmental covariates derived from vegetation cover and topography.
- Estimate the proportion of total carbon stocks (12.3%) stored in semi-seasonal savannas experiencing prolonged flooding.
- Validate and compare DSM approaches using the 12m resolution raster and the underlying 80-site sampling data.
Strengths
- High spatial resolution of 12 meters for detailed landscape analysis.
- Model trained on field measurements from 80 specific sampling sites.
- Provides quantified total SOC stocks (55.07 Mt) and mean density (83.1 ± 24.3 t/ha) for the study area.
- Identifies specific environmental covariates (vegetation, topography, soil properties) as key explanatory drivers.
Limitations
- Spatial coverage is limited to the specific 664,752-hectare study area in the Casanare region.
- Model is based on 2019 data and may not reflect current conditions.
- The underlying sampling density is 80 sites, which may limit fine-scale validation.
Provenance
- Source
- The Alliance of Bioversity International and CIAT Dataverse, authored by J.M. Martín-López.
- Collection Method
- Modeled using an Expert Knowledge-based Digital Soil Mapping approach implemented through a Soil Inference Engine.
- Time Range
- 2019
- Freshness
- Data represents conditions from 2019; the dataset record was last updated in 2026.
- Geography
- Flooded savannas of Casanare, in the Colombian Llanos region (study area: 664,752 hectares).