2019-Modeling distributed soil depth maps using different methods

This repository contains three scripts to produce distributed soil depth maps. The Ordinary_regression_kriging.R script interpolates soil depth based on ordinary and regression kriging, the Random_forest.R script uses the random forest classification method and the GIST.py script predicts soil depth by using a geomorphological index approach based on Catani et al. ( 2010). Users need to provide their own input data, which requires soil depth observation points for the model calibration and validation and certain covariates in a raster format. A detailed description of the methods as well as an application of the methods in a Swiss mountainous catchment are found in the Documention.pdf file.