The wild olive, the thorny ancestor of one of the Mediterranean’s most iconic crops, is projected to lose more than 80 percent of its highly suitable habitat by the final decades of this century under a high-emission climate scenario. That is the central finding of a new species distribution modelling study published in the journal Web Ecology, which mapped where the wild olive, Olea europaea subsp. europaea var. sylvestris, can grow today and where its climatic niche is likely to move as the region warms. The results point to a steady northward retreat of favourable conditions, with the sharpest losses concentrated late in the century under the most extreme of the two emissions pathways the researchers tested.
The wild olive, also known as the oleaster, is far more than a botanical curiosity. It is a keystone species for Mediterranean biodiversity and an important genetic resource for cultivated olives, which can interbreed freely with it wherever wild and cultivated populations coexist. Understanding how its range may shift under future climatic conditions matters both for conservation planning and for the resilience of olive cultivation itself. Yet while previous studies of the genus have mostly been confined to national or regional scales, this new work covers the entire Mediterranean Basin, from Spain to the Levant, using a standardised modelling framework.
To build the model, the team, led by Wisse Zijlstra of the University of Málaga, assembled 558 verified presence records for the wild variety. Of these, 479 came from the Global Biodiversity Information Facility, filtered down from 1,283 raw records by keeping only herbarium-preserved specimens and excluding observational entries, while 79 survived cleaning from 422 records supplied by 13 regional herbaria contacted through the Index Herbariorum network. Each record was checked to confirm it represented a wild, non-cultivated individual, using location, habitat type and collection notes, and records from botanical gardens or obviously cultivated settings were removed. The researchers deliberately did not apply spatial thinning, a decision they acknowledge leaves the model sensitive to sampling bias, with a risk of overestimated suitability near densely sampled areas such as the Iberian Peninsula.
The occurrence data were combined with six bioclimatic variables drawn from the WorldClim database at roughly one-kilometre resolution, representing historical averages from 1970 to 2000. The six predictors, selected for low mutual correlation and ecological relevance, were mean diurnal range, minimum temperature of the coldest month, mean temperature of the warmest quarter, annual precipitation, precipitation seasonality and precipitation of the warmest quarter. Future projections were generated from the CMIP6 climate model EC-Earth3-Veg under two Shared Socioeconomic Pathways: SSP2–4.5, an intermediate scenario with moderate mitigation, and SSP5–8.5, a fossil-fuel-intensive pathway associated with warming exceeding 4 degrees Celsius by 2100. Projections were made for two periods, 2041–2060 and 2061–2080.
Rather than relying on a single algorithm, the researchers used an ensemble approach that combined generalised linear models, generalised additive models and maximum entropy modelling within the biomod2 framework. A pool of 10,000 pseudo-absences was randomly generated across the study area to stand in for the absence data that presence-only records cannot provide. Model performance was assessed through repeated split-sample cross-validation, with 75 percent of records used for training and 25 percent held out for testing across ten replicates. The ensemble achieved a mean calibration AUC of 0.9 and a true skill statistic of 0.7, with a sensitivity of 87.4 percent and a specificity of 79.3 percent, figures the authors describe as indicating good discrimination with little overfitting.
The response curves reveal a species living on a narrow climatic ledge. Minimum temperature of the coldest month emerged as the single most influential variable, with a permutation importance of 64.9, and the model predicts unsuitable conditions where winter minima fall below roughly 0.5 degrees Celsius, with suitability peaking near 9 degrees. Mean temperature of the warmest quarter was the second most important predictor: suitability begins to rise around 16 degrees, peaks at 20 degrees, then declines steeply beyond 24 degrees and becomes unsuitable above 25 degrees. Annual precipitation showed maximum suitability around 500 millimetres, with unsuitable conditions below 300 and above 1,250 millimetres. Precipitation of the warmest quarter rose to a peak near 30 millimetres before settling into a low-suitability plateau.
Under present-day climate, the model identifies extensive areas of high to very high suitability across southern and central Spain and Portugal, southern Italy, western coastal Greece, northern and central Morocco, and parts of Algeria and Tunisia, with smaller patches on Sicily, Corsica and Crete and along coastal fringes of the Levant. When the model was projected forward, every future scenario pointed in the same direction: a reduction in total suitable area and a general northward shift in optimal habitat, particularly across northern Spain, southern France, northern and central Italy, and parts of Greece and Türkiye. Under SSP2–4.5, the mid-century and late-century patterns were broadly similar, with moderate contraction of highly suitable areas, while SSP5–8.5 for 2041–2060 showed a comparable pattern before diverging sharply.
The quantitative results are stark. Highly suitable area, defined as a suitability score of 0.6 or above, declined from 29.7 times 10 to the fourth square kilometres under current conditions to 5.3 times 10 to the fourth square kilometres under SSP5–8.5 for 2061–2080, a reduction of more than 80 percent. Many regions classified as highly or very highly suitable today are downgraded to lower categories or become unsuitable altogether, especially in the southern Mediterranean. Yet the picture is not uniformly bleak. Some coastal zones, including parts of Portugal, northern Morocco and parts of Spain, persist as isolated hotspots of high suitability across all future projections, likely because coastal regions warm less than inland areas and often receive more summer precipitation, buffering the extreme summer temperatures and droughts that the model identifies as the species’ main climatic constraints.
The authors are careful to frame these coastal zones as areas of projected climatic stability rather than validated refugia, noting that confirming a refugial function would require field data on local microclimates and demographic or genetic monitoring of coastal versus inland populations. They also flag substantial limitations. The future projections rest on a single climate model, and incorporating additional models would allow between-model uncertainty to be quantified. Several Mediterranean countries, including Algeria, Tunisia, Palestine, Israel, Lebanon, Syria and especially Türkiye, contributed no occurrence data at all; 28 Turkish herbaria were contacted but none responded, and no herbarium-preserved records were found in GBIF for the country. Because the wild olive is known to occur naturally throughout the eastern Mediterranean, widely considered its centre of origin, the unsuitable predictions there should be read as an artefact of incomplete sampling rather than a genuine ecological signal.
Taxonomic uncertainty adds a further layer of caution. The wild variety is frequently recorded in databases simply as Olea europaea without specification, and it is currently treated as a synonym of the subspecies in major taxonomic sources, while hybridisation with cultivated olives is widespread in marginal or abandoned groves. Despite these caveats, the study’s core message aligns with earlier regional work: a projection for cultivated olive in Türkiye found a similar northward and upslope redistribution, and an assessment of olive varieties in southern Spain reported marked declines for the wild variety under mid-century and later projections. The authors conclude that reducing greenhouse gas emissions and promoting sustainable land management remain essential to slowing habitat loss and supporting the long-term persistence of wild olive populations across the basin.
Subject of Research: Projected impacts of climate change on the distribution of climatically suitable habitat for the wild olive across the Mediterranean Basin
Article Title: Impacts of climate change on the wild olive (Olea europaea subsp. europaea var. sylvestris) in the Mediterranean Basin
Article References: Zijlstra, W., Solakis-Tena, A., & Hidalgo-Triana, N. (2026). Impacts of climate change on the wild olive ( Olea europaea subsp. europaea var. sylvestris ) in the Mediterranean Basin. Web Ecology, 26(2), 251-269. https://doi.org/10.5194/we-26-251-2026
Image Credits: AI Generated
Keywords: wild olive, Olea europaea var. sylvestris, Mediterranean Basin, species distribution model, climate change, habitat suitability, SSP2–4.5, SSP5–8.5, bioclimatic variables, ensemble modelling, CMIP6, habitat loss
News Source: Sloane Callahan. (November 3, 2026). Climate change could strip the wild olive of most of its Mediterranean habitat by 2080. Scienmag.
