Habitat zonation shapes ecosystem multifunctionality and functional trade-offs in farmland across an oasis–desert ecotone in Xinjiang, China

Farmland ecosystems within the oasis–desert ecotone of arid regions play a crucial role in sustaining regional ecological stability and environmental functioning. However, the spatial distribution patterns of ecosystem multifunctionality (EMF) and the key factors associated with its variation across different crop types and habitat zones remain inadequately understood. This study investigated jujube orchards and cotton fields across two contrasting habitat zones, namely the oasis core and the oasis–desert transition zone, in the Alar region of southern Xinjiang Uygur Autonomous Region, China, during the growing season from June to September 2024. Eight ecosystem functional indicators, including three soil nutrient pools, soil water-holding capacity, three herbaceous α-diversity indices, and dust retention per unit leaf area, were quantified and standardized to calculate EMF and subsequently integrated into five ecosystem functional dimensions representing carbon sequestration, nutrient cycling, water storage, plant diversity, and air purification. An integrated analytical framework combining random forest analysis, two-stage partial least squares structural equation modeling (PLS-SEM), and trade-off–synergy analysis was employed to identify dominant predictors associated with EMF, clarify potential pathways linking habitat zone and crop type to EMF, and characterize functional interactions across different habitat zone–crop type systems. Along the gradient from the oasis core to the oasis–desert transition zone, soil organic carbon, total nitrogen, and total phosphorus stocks in both jujube and cotton systems declined by 14.8%–37.2%, 25.0%–51.0%, and 16.9%–24.9%, respectively. Soil water-holding capacity also decreased by 5.3–20.8 mm, whereas dust retention per unit leaf area increased markedly by 16.0%–73.0%. Accordingly, EMF was approximately 10.0% lower in the oasis–desert transition zone than in the oasis core, but remained consistently higher in jujube orchards than in cotton fields within the same habitat zone. Random forest analysis and two-stage PLS-SEM consistently showed that habitat zone regulates EMF primarily through a soil fertility–biodiversity pathway, whereas crop type exerted a weaker effect mainly through soil water dynamics. Trade-off–synergy analysis revealed predominantly synergistic functional relationships in the oasis core, with all pairwise relationships among the five ecosystem functional dimensions in oasis-core cotton fields exhibiting synergy. In contrast, air purification in oasis-core jujube orchards showed clear trade-offs with the other four functional dimensions. Trade-offs became more frequent and pronounced in the oasis–desert transition zone, particularly in transition-zone cotton fields, where reductions in carbon sequestration and water storage are accompanied by enhancements in plant diversity, nutrient cycling, or air purification. Overall, this study elucidates the mechanistic basis of EMF decline and the intensification of functional trade-offs under conditions of farmland marginalization in oasis–desert ecotones. These results provide important scientific support for zone-specific management strategies that emphasize multifunctional co-benefit optimization in the oasis core, while prioritizing the conservation of carbon sequestration and water storage functions in environmentally vulnerable oasis–desert transition zones.