Populus euphratica is a key community-building species in the desert riparian forests of Northwest China, exhibiting exceptional resistance to stress and playing a vital role in soil and water conservation as well as maintaining ecological balance in arid regions. To investigate the ecological processes underlying the composition of P. euphratica communities and to identify their community construction mechanisms, this study analyses the species diversity and phylogenetic diversity of 58 P. euphratica communities, exploring their assembly processes and key influencing factors. This research aims to elucidate the relationship between community structure from the perspective of species evolution and analyse the construction mechanisms of P. euphratica communities across different clusters in arid environments. The results show that the species diversity of P. euphratica clusters in Northwest China is relatively low, and a significant correlation is noted with phylogenetic diversity (PD). The Shannon–Wiener and Margalef indices exhibit similar trends, whereas Simpson’s index show the opposite trends. Pielou’s index range from 0.7 to 0.85. Notably, the PD and species diversity of the P. euphratica–Haloxylon ammodendron association group (Group 4) is significantly higher (p < 0.05) compared to that of the other groups. Additionally, net relatedness index (NRI) and nearest taxon index (NTI) peaked in the P. euphratica–H. ammodendron association group (Group 4) and the Populus pruinosa–Tamarix ramosissima–Phragmites australis association group (Group 1) (p < 0.05). A Pearson correlation analysis indicated that PD was significantly positively correlated with Margalef’s index, Shannon–Wiener’s index, and Pielou’s index, but was significantly negatively correlated with Simpson’s index, while also being associated with environmental factors. Key factors influencing the diversity of P. euphratica communities in Northwest China include total phosphorus, pH, soil moisture content, total potassium, the mean temperature of the coldest quarter, precipitation of the wettest month, and precipitation seasonality. Soil factors primarily affected the Pielou and Simpson indices of species diversity, whereas climatic factors mainly influenced the Margalef and Shannon–Wiener indices. PD and structure were mainly influenced by climatic factors. The combined effects of soil and climatic factors play a crucial role in sustaining the diversity and ecological adaptation of these plant communities. In summary, P. euphratica communities may exhibit a significant ecological niche conservation in response to environmental changes, and competitive exclusion might be the primary process shaping community structure. Climatic factors were shown to be important regulators of community diversity and phylogenetic structure.

Dryland riparian forests depend on river-derived water in arid landscapes, but the relative associations of regional climate, local soil moisture, and biodiversity with ecosystem multifunctionality (EMF) remain uncertain. We analysed data from 516 plots in Populus euphratica forests across northwestern China. EMF was quantified from eight vegetation and soil attributes using averaging and multi-threshold approaches. We combined species richness, Rao’s quadratic entropy, community-weighted trait axes, climate principal components, and soil principal components in mixed-effects models, hierarchical partitioning and partial least-squares path models. EMF declined from near- to far-riparian plots, whereas soil water content did not differ significantly among distance bands. Hydroclimatic stress was negatively associated with EMF, especially in far-riparian plots and at higher thresholds, while soil moisture was positively associated mainly at low-to-intermediate thresholds. The relative explanatory contributions of biodiversity varied across riparian positions and predefined functional thresholds: RaoQ contributed most near rivers at lower thresholds, species richness was more important at intermediate and far positions, and community-weighted traits became prominent under selected high-threshold conditions. Species richness and community-weighted trait axes also changed with riparian distance, whereas RaoQ was comparatively stable. In the PLS-PMs, modelled direct abiotic associations were generally stronger than indirect associations involving biodiversity. Given the observational study design and correlational analyses, these findings indicate context-dependent associations among climate, soil moisture, biodiversity and ecosystem multifunctionality rather than causal relationships.