Aims: Gebel Elba is an arid mountain range supporting biological diversity that is incomparable to any other region of Egypt. This mountain has a vegetation structure and floristic community similar to the highlands of East Africa and the southwestern Arabian Peninsula. We aimed to provide the first classification of the vegetation units on Gebel Elba and identify the environmental factors controlling their distribution. Study area: Wadi Yahmib and its tributaries, which drain the north-western slopes of Gebel Elba, south-eastern Egypt. Methods: On the basis of 169 relevés, we used TWINSPAN to classify the perennial vegetation. We calculated separate GAMs for the deciduous and evergreen species to describe the patterns for each leaf strategy type with elevation. We used CCA to quantify the relationship between the perennial vegetation and the studied environmental factors. To estimate diversity and our sampling strategy, we used rarefaction curves for species richness. Results: We identified seven communities along the elevational gradient of Wadi Yahmib and its tributaries. We found that each community was restricted to a confined habitat depending on its drought resistance ability. Deciduous Vachellia woodland was the main vegetation type on Gebel Elba, while evergreen Olea woodland appeared in small fragments at higher elevations. We analysed the distribution patterns of deciduous and evergreen trees along the elevational gradient. We found a turnover at 500 m, indicating a potential ecotone between the Vachellia and Olea woodlands that was occupied by a Ficus community. CCA revealed the importance of altitude and soil quality in determining the vegetation structure of Gebel Elba. The species richness increased with elevation as a result of reduced stress and increased water availability at the upper wadis. Conclusions: This study identified seven vegetation units in the study area and showed the importance of orographic precipitation, soil quality and the complex topography in determining the habitats.<br> Taxonomic reference: Boulos (2009); names updated according to POWO (2019).<br> Abbreviations: CCA = Canonical Correspondence Analysis; GAM = Generalized Additive Model; TWINSPANTWINSPAN = Two Way Indicator Species Analysis.
