Framework for Spatial Predictive Modelling of Soil Properties Using Machine Learning
Soil is essential to life. To preserve soil quality, we need to understand and map its spatial variability. Machine learning is becoming an increasingly widely used tool for predictive soil properties modelling, which can efficiently capture complex non- linear relationships. The main outcome of this project will be a spatial machine learning framework for spatial predictive modelling which benefits from spatial autocorrelation. The new framework will be used to create gridded soil properties datasets for Europe and for Estonia. Through ML we will explore new knowledge on the interplay between environmental covariates with target soil properties.
Funding agency: Estonian Research Council
Principal investigator: Alexander Kmoch, 1.01.2023−31.12.2027
