Predictive Modeling of Soil Types and Their...
E-book
Vasyl Cherlinka - Michal Gallay - Yuryi Dmytruk
The textbook is a practical guide to predictive mapping of soil types and properties in R. It bridges digital soil mapping theory with hands-on practice, covering the full workflow from data acquisition and preprocessing to validation and uncertainty quantification. Early chapters introduce the tidyverse ecosystem and the sf and terra packages for spatial data handling. Subsequent parts address modelling of categorical variables (soil types/WRB) and continuous variables (soil organic carbon) using Random Forest, Cubist, and Quantile Regression Forests. Step-by-step code examples show training/testing, feature selection, accuracy metrics, and model interpretation. Reproducibility and cartographic best practice are emphasised. A Slovakia case study demonstrates real-world application while outlining limits and generalisability. The book targets bachelor’s, master’s, and PhD students in geography, geoinformatics, environmental and agricultural sciences, as well as practitioners. Written in English, it fits international courses and mobility programmes.
