Table 1.
Spatially continuous covariates used in the deep learning model for topsoil Hg distribution on a European scale.
| Feature group | Environmental feature | Covariate | Source | #layers/classes | Covariate type |
|---|---|---|---|---|---|
| Land cover | Land cover | CORINE land cover type | CORINE | 44 | Categorical |
| Natural | Climate | Seasonal rainfall and air temperatures | Worldclim | 12 | Numerical |
| Surface temperature | Landsat | 1 | Numerical | ||
| Soil properties | Soil parent material | European Soil Database | 50 | Categorical | |
| Soil chemical and physical parameters | LUCAS | 8 | Numerical | ||
| Topography | DEM derived topographic features | SRTM/EU DEM | 13 | Numerical | |
| Vegetation | EVI, MODIS reflectance | MODIS | Numerical | ||
| Sources of Hg | Coal power plants | Coal power plants density | Enipedia | 1 | Numerical |
| Global Hg emissions | Global Hg emissions into atmosphere | AMAP/UNEP | 1 | Numerical | |
| Mining activity | Distance to Hg and Au/Ag mines | Mines4EU | 3 | Numerical |