This project focuses on extracting and processing national-scale raster data layers (datacubes) for Sweden to support digital soil mapping of key soil properties, including soil organic carbon. The work involves preparing covariate layers derived from environmental, climatic, and remote sensing datasets, which will serve as input variables for predictive machine learning models. These models aim to generate high-resolution national maps of soil properties such as soil organic carbon, texture, and pH. Compute resources are required for large-scale raster processing tasks (e.g., mosaicking, resampling, format conversion), as well as for training and validating machine learning models. The processed datacube layers and prediction outputs will contribute to ongoing efforts in environmental monitoring and sustainable soil management across Sweden.