TY - DATA
T1 - Soil organic carbon stock and uncertainties, 30cm and 1m depth, at 250m spatial resolution in Canada, version 3.0
PY - 2021/12/29
AU - Camile Sothe
AU - Alemu Gonsamo
AU - Joyce Arabian
AU - Werner A. Kurz
AU - Sarah Finkelstein
AU - James Snider
UR - https://data.4tu.nl/articles/dataset/Soil_organic_carbon_stock_and_uncertainties_1m_depth_at_250m_spatial_resolution_in_Canada_version_2_0/16686154/2
DO - 10.4121/16686154.v2
KW - soil carbon density
KW - soil carbon stock estimate
KW - soil carbon storage
KW - terrestrial ecosystem models
KW - machine Learning Methods Enable Predictive Modeling
N2 - This project aimed to produce the first wall-to-wall estimate of C stocks in plants and soils of Canada at 250 m spatial resolution. This dataset contains the map with the soil organic carbon (SOC) in kg/m² for entire Canada in 1m depth, and the uncertainty in SOC predictions. The SOC stock map was produced using 39,323 ground samples of soil organic carbon concentration (g/kg) distributed in 6,533 sites, 11,068 ground samples of bulk density (kg/dm3)
distributed in 2,157 sites, long-term climate data, remote sensing observations and a machine learning model. The soil samples containing the x and y coordinates, depth and SOC (in g/kg) information were overlaid with the stacked covariates (soil forming factors) to compose the regression matrix. Random forest models were trained using a recursive feature elimination scheme and a cross-validation assessment. The best model was used for spatial prediction of SOC over Canada in intermediate depths between 0 and 1 m (0cm, 5cm, 15cm, 30cm, 60cm, 100cm). Afterwards, the SOC stock of each depth increment was computed using SOC concentration and bulk density maps, and corrected with coarse fragment information. The depth increments have been added to compose the 0-30cm and 0-1m depth intervals multiplied by root depths fraction to discount shallow soils. Water and ice/snow areas were removed using a mask based on the Land Cover of Canada map. Ground ice in permafrost areas was discounted according to ice abundance using the ground ice map of Canada. The SOC stock uncertainty map was built using the first and third quantiles of quantile regression forest approach of SOC concentration and bulk density prediction (90% confidence interval).
ER -