Ecological drivers of dissimilarity in multidimensional plant composition of early-stage plantations in a loess hilly-gully region

Plant diversity is an important link between forest ecosystems' structural composition and functional processes. The simple composition of young plantations is weakening the ecological function of restored forest ecosystems in arid regions. The ecological drivers of the plant composition and dissimilarity in plantations remain unclear, especially in the early restoration stage. To fill the gap in knowledge of the differences in species composition in early restored ecosystems, and to reveal the influencing factors of similar species numbers with different compositional structures, based on species' presence-absence, abundance, and conspicuousness, we explored dissimilarity in the plant composition of young plantations in the loess hilly-gully region. Our results showed that differences in alpha diversity among plantations were minimal, while the differences in beta diversity were significant. Even if the species diversity of a plantation was low, communities with similar species richness used to have large dissimilarities in their species composition. Turnover components and balanced variation often contributed significantly to species compositional dissimilarity. Co-inertia analysis results showed species conspicuousness had the greatest consistency with observed environmental variables (RV = 0.65). The redundancy analysis and generalized diversity modeling results showed differences between the drivers of species diversity and dissimilarity. The primary influencing factors for alpha diversity were: soil water content, soil texture, soil total phosphorus, and soil organic carbon. Primary influencing factors for dissimilarity (beta diversity) were: soil nitrogen to phosphorus ratio, breast-height diameter, soil water content, and stand density. Collectively, plant taxa dissimilarity was influenced by soil nutrient ratio and community structure and plant taxa abundance was influenced by soil nutrient content. Our findings improve the understanding of the plantation species composition and facilitate the prediction of ecological functions.