Skip to content

Our team developed geospatial tools to enhance your geospatial analysis, streamlining processes for greater speed, accuracy, and automation.

Purpose and use: This script enhances soil property mapping accuracy by optimizing terrain attribute scaling using Random Forests (RF). It demonstrated the effectiveness of optimized scaling for nine soil properties.

The script resamples a digital elevation model (DEM) at multiple resolutions and finds the optimal scales for each terrain attribute specific to each soil property.

Inputs:  DEM and geolocated points with soil properties measurements (.shp file).

Outputs: Raster file (.tif) of each terrain attribute optimally scaled for each soil property, and table with the associated optimal scales.

Programming environment: R

Status of work: Public domain

Reference: Dornik A, Cheţan MA, Drăguţ L, Dicu DD, Iliuţă A, 2022, Optimal scaling of predictors for digital mapping of soil properties, Geoderma (IF 6.1; AIS 1.221; Q1), 405:115453, DOI: 10.1016/j.geoderma.2021.115453

Inputs: Slope raster or curvature raster; any other variables with skewed or long-tailed distributions.

Outputs: Normalized slope or curvature raster; a text file informing about transformation used and new skewness/kurtosis.

Programming environment: Python

Status of work: Public domain

Reference: Csillik, O., Evans, I.S., Dragut, L., Transformation (normalization) of slope gradient and surface curvatures, automated for statistical analyses from DEMs, Geomorphology (2015), doi: 10.1016/j.geomorph.2014.12.038