Predictive Modeling of Soil Types and Their Characteristics

Why this book?
Soils underpin food security, climate regulation, biodiversity, and landscape resilience. Mapping soils and their properties at useful spatial scales is therefore a central task in environmental management. Over the last two decades, Digital Soil Mapping (DSM) has matured into a reproducible, data-driven practice that combines field observations, covariates from remote sensing and terrain analysis, and modern statistical learning. This textbook introduces DSM from first principles and shows how to implement the full workflow—data preparation, modelling, validation, and uncertainty communication—using opensource R.

Who this book is for?
The book is intended for graduate and postgraduate students in soil science, geography, geoinformatics, environmental science, and related fields, as well as for researchers and practitioners who need a practical, reproducible route into DSM.
What you will learn?
After completing the book, readers will be able to (i) set up a reproducible R/RStudio project for spatial analysis; (ii) assemble and clean soil point data and spatial covariates; (iii) fit, tune, and interpret local and global predictive models (e.g., regression, tree-based and ensemble methods); (iv) assess predictive performance and quantify uncertainty with appropriate diagnostics; (v) generate soil maps and related products (e.g., SOC, texture classes) and communicate their limitations.