Climate-driven futures of olive (Olea europaea l.): Machine learning-based ensemble species distribution modelling of northward shifts under aridity stress

Abstract

With its millennia-long agricultural history, Olive (Olea europaea L.) is one of the most strategic crops of the Mediterranean basin and a key component of the Turkish economy. This study assessed the effects of climate change on the potential distribution of olive in T & uuml;rkiye using machine learning-based species distribution models (SDMs). Analyses were conducted using the 1970-2000 reference period and future projections for 2041-2060 and 2081-2100 under the SSP2--4.5 and SSP5-8.5 scenarios, incorporating bioclimatic variables as well as topographic factors such as elevation, slope, and aspect. The model showed strong predictive performance (AUC = 0.93; TSS = 0.77) and identified elevation, winter precipitation (Bio19), and mean temperature of driest quarter (Bio9) as the primary variables influencing the distribution of olive trees. Model results predict a significant shift in suitable areas for olive cultivation, both northward-from the traditional Aegean and Mediterranean coastal belt toward the Marmara and Black Sea regions-and upward in elevation into higher-altitude inland areas. High-suitability areas, which accounted for 4.4% of T & uuml;rkiye's land area during the reference period, are projected to decline to 0.2% by the end of the century under the SSP5-8.5 scenario. UNEP Aridity Index analyses indicate increasing aridity pressure on olive habitats. While 87.2% of suitable habitats were classified as sub-humid in the reference period, projections for 2081-2100 under SSP5-8.5 suggest that 40.1% of these areas will shift to dry sub-humid and 26.4% to semi-arid conditions.