Abstract
Understanding the variables that determine the variation in forest species composition and diversity in tropical montane systems remains a topic for discussion in plant ecology. This is especially true in areas where the topography is complex and forests are vulnerable to human activity. In this study, a set of topographic, soil, and climatic variables were used to determine their effects on the composition and diversity patterns of two forests in the West Usambara Mountains (Tanzania). Two-phase systematic sampling was used to collect vegetation data from 159 sample plots distributed across the forests. An agglomerative hierarchical clustering method was used for forest community classification, and indicator species analysis was used to determine the species significantly associated with forest communities. The influence of environmental variables on forest communities was analysed using canonical correspondence analysis (CCA). Finally, we evaluated diversity patterns by comparing diversity indices (Shannon-Wiener diversity index, evenness, and richness) and beta diversity processes. In total, 7,767 individual trees belonged to 183 species, 132 genera, and 66 families were quantified. We found that (i) the forests of West Usambara can be divided into three different forest communities; (ii) each forest community has a specific set of topographical, soil, and climate variables; (iii) there are significant differences in Shannon diversity and richness indices among communities; and (iv) community composition is mostly influenced by species turnover than by species nestedness. Our study revealed the importance of considering a set of environmental variables related to climate, soil, and topography to understand the variation in the composition and diversity of forest communities in tropical montane forests.