Service Public de Wallonie, Direction Générale de l'Agriculture, des Ressources Naturelles et de l'Environnement (DGO3), Département de l'Etude du Milieu Naturel et Agricole, Direction de la Coordination des Données, Cellule SIG
Type of resources
Available actions
flanderskeyword
Provided by
status
Topics
Keywords
License
Contact for the resource
Years
Formats
Representation types
Update frequencies
Scale 1:
Resolution
INSPIRE themes

The Belgian soil organic carbon (SOC) stock map for topsoils (030 cm) at the resolution of 40m x 40m was composed of 2 regional SOC stock maps. For the regional maps a different approach was used for agricultural land as compared to forest. The maps are based on digital soil mapping approaches using empirical models calibrated to predict the SOC stock and using covariates that are available at a sufficient resolution at the regional scale. All maps are strongly dependent on the Belgian Soil Map (texture and drainage parameters). The regional maps were compiled at a fine resolution (10m x 10m grid cells for Flanders and 40m x 40m grid cells for Wallonia). Next they were joined (40m x 40m grid cells). The regional maps for Flanders were scaled up tot the 40m x 40m grid of Wallonia. Given the different origin of the individual maps, the uncertainty varies between maps. For instance, a map of the 90% confidence interval of the SOC stocks was produced for agricultural soils in Wallonia based on a Monte Carlo Approach taking into account both the measurement and the model uncertainties. For Flemish forest soils, spatial and analytical uncertainties were taken into account using bootstrapping techniques. For Flemish agricultural soils, the uncertainty reported is the model uncertainty on point estimates for each data point, in which the estimated model parameters are simulated 1000 times as being independent normal distributed variables using their model estimation and standard error as distribution parameters. No additional uncertainty is taken into account for the conversion functions that use the stochastic variables "bulk density". The metadata are available and allow assessing the uncertainties of the stock estimates in the different component maps. The SOC stock map (resolution of 40m x 40m) is the first comprehensive map for Belgium integrating grasslands, croplands and forests. Based on this map another SOC stock map at a resolution of 1km x 1km in the coordinate reference system WGS84 was created.

The Belgian soil organic carbon (SOC) stock maps for topsoils (030 cm) were composed of 2 regional SOC stock maps. For the regional maps a different approach was used for agricultural land as compared to forest. The maps are based on digital soil mapping approaches using empirical models calibrated to predict the SOC stock and using covariates that are available at a sufficient resolution at the regional scale. All maps are strongly dependent on the Belgian Soil Map (texture and drainage parameters). The regional maps were compiled at a finer resolution (10m x 10m and 40m x 40m grid cells). Next they were joined (40m x 40m grid cells) and finally scaled up to the required 1 by 1 km grid cells. This was done using the following tools: block statistics (mean), mosaic to new raster (mean), project to raster, block statistics (mean), resample (nearest neighbour) and project raster. Given the different origin of the individual maps, the uncertainty varies between maps. For instance, a map of the 90% confidence interval of the SOC stocks was produced for agricultural soils in Wallonia based on a Monte Carlo Approach taking into account both the measurement and the model uncertainties. For Flemish forest soils, spatial and analytical uncertainties were taken into account using bootstrapping techniques. For Flemish agricultural soils, the uncertainty reported is the model uncertainty on point estimates for each data point, in which the estimated model parameters are simulated 1000 times as being independent normal distributed variables using their model estimation and standard error as distribution parameters. No additional uncertainty is taken into account for the conversion functions that use the stochastic variables "bulk density". The SOC stock maps are the first comprehensive maps for Belgium integrating grasslands, croplands and forests. There are two versions of the SOC stock maps for Belgium: 1) resolution of 40m x 40m in the coordinate reference system Lambert72 and 2) resolution of 1km x 1km in the coordinate reference system WGS84. The metadata are available and allow assessing the uncertainties of the stock estimates in the different component maps.

The Belgian soil organic carbon (SOC) stock map for topsoils (030 cm) was composed of 2 regional SOC stock maps. For the regional maps a different approach was used for agricultural land as compared to forest. The maps are based on digital soil mapping approaches using empirical models calibrated to predict the SOC stock and using covariates that are available at a sufficient resolution at the regional scale. All maps are strongly dependent on the Belgian Soil Map (texture and drainage parameters). The regional maps were compiled at a finer resolution (10m x 10m and 40m x 40m grid cells). Next they were joined (40m x 40m grid cells) and finally scaled up to the required 1 by 1 km grid cells. This was done using the following tools: block statistics (mean), mosaic to new raster (mean), project to raster, block statistics (mean), resample (nearest neighbour) and project raster. Given the different origin of the individual maps, the uncertainty varies between maps. For instance, a map of the 90% confidence interval of the SOC stocks was produced for agricultural soils in Wallonia based on a Monte Carlo Approach taking into account both the measurement and the model uncertainties. For Flemish forest soils, spatial and analytical uncertainties were taken into account using bootstrapping techniques. For Flemish agricultural soils, the uncertainty reported is the model uncertainty on point estimates for each data point, in which the estimated model parameters are simulated 1000 times as being independent normal distributed variables using their model estimation and standard error as distribution parameters. No additional uncertainty is taken into account for the conversion functions that use the stochastic variables "bulk density". The SOC stock maps are the first comprehensive map for Belgium integrating grasslands, croplands and forests. There are two versions of the SOC stock maps for Belgium: 1) resolution of 40m x 40m in the coordinate reference system Lambert72 and 2) resolution of 1km x 1km in the coordinate reference system WGS84. The metadata are available and allow assessing the uncertainties of the stock estimates in the different component maps.

The Belgian soil organic carbon (SOC) stock map for topsoils (030 cm) was composed of 2 regional SOC stock maps. For the regional maps a different approach was used for agricultural land as compared to forest. The maps are based on digital soil mapping approaches using empirical models calibrated to predict the SOC stock and using covariates that are available at a sufficient resolution at the regional scale. All maps are strongly dependent on the Belgian Soil Map (texture and drainage parameters). The regional maps were compiled at a finer resolution (10m x 10m and 40m x 40m grid cells). Next they were joined (40m x 40m grid cells) and finally scaled up to the required 1 by 1 km grid cells. This was done using the following tools: block statistics (mean), mosaic to new raster (mean), project to raster, block statistics (mean), resample (nearest neighbour) and project raster. Given the different origin of the individual maps, the uncertainty varies between maps. For instance, a map of the 90% confidence interval of the SOC stocks was produced for agricultural soils in Wallonia based on a Monte Carlo Approach taking into account both the measurement and the model uncertainties. For Flemish forest soils, spatial and analytical uncertainties were taken into account using bootstrapping techniques. For Flemish agricultural soils, the uncertainty reported is the model uncertainty on point estimates for each data point, in which the estimated model parameters are simulated 1000 times as being independent normal distributed variables using their model estimation and standard error as distribution parameters. No additional uncertainty is taken into account for the conversion functions that use the stochastic variables "bulk density". The SOC stock maps are the first comprehensive map for Belgium integrating grasslands, croplands and forests. There are two versions of the SOC stock maps for Belgium: 1) resolution of 40m x 40m in the coordinate reference system Lambert72 and 2) resolution of 1km x 1km in the coordinate reference system WGS84. The metadata are available and allow assessing the uncertainties of the stock estimates in the different component maps.

The Belgian soil organic carbon (SOC) stock map for topsoils (030 cm) at the resolution of 40m x 40m was composed of 2 regional SOC stock maps. For the regional maps a different approach was used for agricultural land as compared to forest. The maps are based on digital soil mapping approaches using empirical models calibrated to predict the SOC stock and using covariates that are available at a sufficient resolution at the regional scale. All maps are strongly dependent on the Belgian Soil Map (texture and drainage parameters). The regional maps were compiled at a fine resolution (10m x 10m grid cells for Flanders and 40m x 40m grid cells for Wallonia). Next they were joined (40m x 40m grid cells). The regional maps for Flanders were scaled up tot the 40m x 40m grid of Wallonia. Given the different origin of the individual maps, the uncertainty varies between maps. For instance, a map of the 90% confidence interval of the SOC stocks was produced for agricultural soils in Wallonia based on a Monte Carlo Approach taking into account both the measurement and the model uncertainties. For Flemish forest soils, spatial and analytical uncertainties were taken into account using bootstrapping techniques. For Flemish agricultural soils, the uncertainty reported is the model uncertainty on point estimates for each data point, in which the estimated model parameters are simulated 1000 times as being independent normal distributed variables using their model estimation and standard error as distribution parameters. No additional uncertainty is taken into account for the conversion functions that use the stochastic variables "bulk density". The metadata are available and allow assessing the uncertainties of the stock estimates in the different component maps. The SOC stock map (resolution of 40m x 40m) is the first comprehensive map for Belgium integrating grasslands, croplands and forests. Based on this map another SOC stock map at a resolution of 1km x 1km in the coordinate reference system WGS84 was created.