Explainable AI for Mapping Mature and Old-Growth Forests (MOFG) of Italy

The dataset provides an explainable AI workflow to identify and map Mature and Old-Growth Forests (MOGF) across Italy using high-resolution Earth Observation data. The code integrates machine learning and spatial ecology to shift from traditional inventory-based assessments to continuous, wall-to-wall mapping. The predictive model uses a Random Forest classifier trained on six predictors. To address the black-box nature of machine learning, the script incorporates Shapley Additive Explanations to quantify the global importance and local influence of each forest-structure variable on classification. The evaluation phase includes an analysis of classification errors. The output is a 30m spatial resolution MOGF probability map at the Italian level. Furthermore, the UNESCO-listed beech forests (Fagus sylvatica) of Italy are included in the dataset.