Vegetation patches in semiabandoned olive groves: using generalised linear mixed models to determine the effect of area on community composition of woody plants

Aim of study: The existence of distinct vegetation habitats within semiabandoned olive groves provides a unique chance to study the plant community within semiabandoned Mediterranean landscapes. We investigated changes in community composition of woody plants across a gradient of patch sizes by providing an example of a novel statistical technique. We also aimed to determine if commonness, life form and dispersal mechanisms of woody plants are key factors influencing species presence at vegetation patches of different sizes. Study area: Three traditional, partially managed mountain olive groves (La Soledad, Las Ni??as and Piqu??n) were selected within Sierra Morena de C??rdoba, in Central Southern Spain. Materials and methods: The woody vegetation within patches at the three groves was sampled in July 2020 following a stratified ran-dom approach. All woody plants were identified and recorded. Variation in community composition across patch area was examined using generalised linear mixed models (GLMMs) with a binomial distribution. Main results: There were significant changes in the community composition of woody plants as patch area increased. There was turnover of species with increasing area, characterised by the gain of species. This was observed both as a general and site-specific trend. Patterns in presence across area showed clear among species variation. Including dispersal strategies and life form variables improved model fit, revealing these are important factors influencing the community composition within the patches. Research highlights: The GLMM analysis demonstrated that patches of larger areas support higher richness without incurring in any loss of species. Thus, maintaining large patches is important for woody plant conservation within semi-abandoned groves.