Bias‐Corrected Climate Projections for Xinjiang: Decomposing Future Trends and Uncertainties in Temperature and Precipitation

ABSTRACT Xinjiang is highly sensitive to climate change and faces escalating challenges from global warming. However, Coupled Model Intercomparison Project (CMIP) projections still exhibit notable biases and uncertainties in this region. To address these limitations, this study applies the quantile delta mapping (QDM) approach to correct biases in temperature and precipitation simulations and further decomposes the contributions of model uncertainty, scenario uncertainty and internal variability. The results show that the QDM approach substantially reduces the simulation biases, especially in mountainous regions and the Taklamakan Desert. Based on QDM‐corrected projections, CMIP6 multi‐model ensembles project continuous increases in mean annual temperature and precipitation. By the end of the 21st century, under the SSP5‐8.5 scenario, temperature is projected to increase by 6.8°C (uncertainty range: 5.7°C–7.7°C) and precipitation by 40.0 mm (18.8–67.8 mm) relative to the 1981–2000 baseline. Uncertainty analysis reveals that model uncertainty is the dominant contributor (over 60%) to both temperature and precipitation projections in the near term, whilst scenario uncertainty becomes increasingly significant as anthropogenic forcing intensifies over time. Moreover, the signal‐to‐noise ratio (S/N) suggests that the warming signal exceeds the uncertainty in all future periods. Conversely, the S/N ratios for precipitation are below unity in the near to mid‐term but strengthen in the far‐term, especially in southeastern Xinjiang. By improving the reliability of climate projections and clarifying the sources of uncertainty, this work provides a more robust foundation for formulating effective climate adaptation and mitigation strategies in Xinjiang.