A machine learning framework for predicting agricultural terrace suitability using positive and unlabelled datasets: an application for the Troodos Mountains, Cyprus

Terraces are a defining feature of Mediterranean mountain landscapes, enabling agriculture on steep slopes while providing multiple ecosystem services. Land suitability analysis (LSA) can guide authorities and land users to sustainably manage and expand these environments. It typically requires fully labelled datasets, but in many real-world applications only a fraction of positive examples is available, with the rest unlabelled. This study aims to present an integrated predictive modelling framework that combines GIS with data-driven Machine Learning (ML) techniques, capable of learning from Positive and Unlabelled (PU) datasets for LSA. The proposed framework was applied to develop a terrace suitability map for Cyprus’ Troodos Mountains. A 5-m DEM was processed to extract the mountain area, with elevation ≥ 500 m and slopes ≥ 15%, defining the study area. Crop plots registered under the Single Area Payment Scheme of the European Common Agricultural Policy were used to classify the study area into Terrace-Present (TP) and Terrace-Absent (TA) cells, with TP serving as labelled positive and TA as unlabelled samples. A two-step ML approach was applied, first identifying reliable negatives from TA cells, then using these with TP cells for suitability prediction. The developed PU-classifier was evaluated under the selected completely-at-random (SCAR) assumption, achieving a Recall of 84.6%, Precision of 81.5% and an F1 score of 83%. Feature importance analysis identified land cover, terrain slope and tree cover density as the most influential parameters for terrace suitability. Comparative analysis between 2017 and 2024 revealed abandonment of terraced agricultural land (29% decrease) as well as revitalisation (12% increase). The resulting suitability map and accompanying data layers are accessible through a Google Earth Engine application, aiming to support informed decision-making for sustainable landscape planning.