Forest Herb‐Layer Species Richness in the Western Balkan Diversity Hotspot

Questions Southeastern Europe harbours a high diversity of vascular plants at a regional scale. However, not much is known about the local diversity of forests in this area and the underlying factors. Therefore, we studied a wide range of forest vegetation types in various environmental conditions to (1) explore the patterns of vascular plant species richness in forest herb layers, (2) identify the most species‐rich forest types and (3) determine the key factors that shape these patterns. Location Montenegro (Balkan Peninsula). Methods We sampled 342 vegetation plots of 100 m 2 across the whole country and its main forest types. We fitted a negative binomial generalised linear model to test the relationship between the herb‐layer species richness and abiotic and biotic predictors. Abiotic predictors included soil characteristics, topographic and climatic variables. Biotic predictors included the cover of trees, shrubs and evergreen species. We also tested the differences between forest types defined by the dominant tree species. Results Forest herb layers had a broad range of vascular plant species richness from 7 to 149 species per 100 m 2 . Herb‐layer species richness showed a significant unimodal relationship with precipitation in the warmest quarter and a negative relationship with tree‐layer cover and terrain ruggedness. In addition, soil pH had a significant positive effect on herb‐layer species richness, but only in areas with intermediate values of mean annual temperature. Mixed deciduous, pine and riparian forests were the richest, while evergreen and beech forests were the poorest. Conclusions Herb‐layer species richness varies considerably and predictably in the forests of the western Balkan Peninsula. Thermophilous oak forests on base‐rich soil often harbour more than 90 vascular plant species per 100 m 2 . Mean annual temperature, summer precipitation, terrain ruggedness, soil pH and tree‐layer cover are the main factors shaping herb‐layer species richness patterns.