Habitat Quality and Urban Development Level in Arid Xinjiang, China: Improved Coupling Coordination and the Spatiotemporal Impacts of Urban Development

Understanding the coupling coordination between habitat quality (HQ) and urban development and their spatial interactions can inform coordinated ecological and socioeconomic development in Xinjiang, China. This study established an integrated evaluation framework for habitat quality and urban development across 14 prefecture-level units in Xinjiang, using observed data from 2012 to 2024 and projected data for 2027. Entropy-weighted TOPSIS, non-seasonal exponential smoothing, CA–Markov modeling, and an improved coupling coordination model were combined with bivariate spatial autocorrelation, geographically and temporally weighted regression (GTWR) and XGBoost–SHAP were employed to assess development patterns, spatial relationships, and influencing factors. During the observation period, habitat quality remained relatively low but showed a fluctuating upward trend, with significant differences between northern and southern Xinjiang. The average coupling coordination degree across the region steadily increased from 0.317 to 0.381. High-value clusters centered on Urumqi and Ili Prefecture and declined toward surrounding areas, while Turpan remained severely uncoordinated. Urban development and habitat quality showed stable negative spatial agglomeration, and GTWR indicated that their centers of influence shifted from Aksu toward Bayingolin. Analysis of the results incorporating the projected data showed that the mean coupling coordination degree across Xinjiang is expected to reach 0.39 in 2027. The dual-system association network exhibited a hierarchical transmission pattern comprising four block types: core output, linkage transit, marginal association, and bidirectional interaction. GTWR revealed significant spatial bidirectional effects, with the strongest negative effects in Kizilsu and Kashgar, while Aksu and Hotan showed relatively stable positive reciprocal feedback. XGBoost–SHAP analysis identified educational resource concentration as the primary negative driver, whereas rational urban land-use planning provided positive support across regions. The driving factors also differed between northern and southern Xinjiang. Accordingly, habitat quality should be incorporated into development performance assessments, differentiated management should be implemented according to block functions, and the allocation of educational resources and land use should be optimized to promote high-level mutual reinforcement between urban development and habitat quality.