Application of the Maxent model in habitat suitability assessment for Rosalia alpina (Linnaeus, 1758) in Montenegro

Abstract

This study applies the Maxent model to predict the potential habitat suitability of the endangered saproxylic beetle Rosalia alpina in Montenegro. The model identified temperature (TEMP) and near-surface relative humidity (NSRH) as the dominant predictors influencing the species’ distribution, followed by vegetation type (VEG) and topographic variables such as slope (SLP) and aspect (ASP). The model achieved good predictive accuracy, with an average test AUC of 0.798 ± 0.012 across five replicates, confirming its robustness and reliability. The resulting potential habitat suitability map indicates that the central and north mountainous regions of Montenegro represent the most suitable areas, corresponding closely with known field observations. These findings highlight the importance of preserving structurally diverse deciduous forests and provide a valuable basis for spatially targeted conservation and sustainable forest management strategies for R. alpina in Montenegro. However, spatial overlap analysis showed that only about 18% of highly suitable (potential) habitats are located within nationally protected areas, and around 20% within the proposed Emerald network. This indicates that a considerable portion of potentially suitable habitats remain unprotected, emphasizing the need to strengthen and expand the existing conservation frameworks and to prioritize these areas for inclusion in the future Natura 2000 network, ensuring effective long-term protection of the species and its key forest habitats.