Answers are grounded in this paper and cite their sources.
Plant diversity serves as the foundation for maintaining ecosystem functions and services. Existing research has primarily focused on areas with high vegetation cover, such as forests and grasslands, whereas community assembly mechanisms and distribution patterns in desert regions remain poorly understood. This study used data from 160 sample sites in the Tarim Basin, combined with 40 factors, including topography, soil, climate, and remote sensing, using a random forest algorithm to build a plant diversity model to systematically analyze the spatial patterns and coupling relationships of multidimensional α diversity in plant communities in the Tarim Basin. The results show the following. (1) The plant communities in the study area exhibit adaptive characteristics, including low species richness, prominent dominant species, functional convergence, and phylogenetic clustering. (2) Under the influence of environmental filtering, phylogenetic conservatism promotes multidimensional coupling of species, functional, and phylogenetic diversity over large areas, while decoupling occurs in local regions due to convergent evolution. (3) Soil sand content is the top environmental predictor of plant diversity variation across the desert-to-oasis transition. (4) The constructed RF_ZI_Winsor model optimizes the random forest algorithm by integrating zero-inflated and winsorization techniques, and compared to the original model (Adj. R 2 : 0.227–0.494), the new model shows better predictive performance (Adj. R 2 : 0.379–0.532). This study confirms that optimized machine learning methods effectively simulate multidimensional α-diversity patterns in desert ecosystems, offering theoretical and methodological support for ecological restoration and vegetation monitoring in arid regions.
No takes yet. Share an insight, caveat, or question.
Xin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: