青藏铁路格拉段（以下简称青藏铁路）， 特别是从西大滩至安多穿越550 km多年冻土区的路段， 沿线有着复杂的地形地貌和冻土环境。近年来随着气候变暖和多年冻土退化， 对于建造在多年冻土之上的青藏铁路的维护需求剧增， 为了能充分捕捉青藏铁路沿线地区复杂的地形地貌对局地气候变化的影响， 以期对青藏铁路的运行维护提供理论支持。本文利用铁路沿线站点观测数据， 并以ERA5数据驱动WRF模式进行网格距离为10 km×10 km的动力降尺度模拟。结果显示， 青藏铁路沿线的六个站点自1998 -2020年普遍呈现增温趋势， 增温率最低的站点为0.27 ℃·（10a）^-1， 最高为0.56 ℃·（10a）^-1； WRF模式对表面气温的模拟在年平均气温结果与观测数据略有差异， 在夏季和秋季的模拟结果较好， 夏季相关系数达到0.95以上， 秋季在0.80以上， 春季和冬季较差； WRF降水模拟的结果显示， Nudging方法有效改进了青藏高原夏季降水的湿偏差， 青藏铁路北段至铁路南段降水逐渐增加， 且在铁路中部降水出现极大值。WRF模式在青藏高原的温度和降水模拟上仍存在着一定的冷偏差和湿偏差， 寻找新的方法或利用高质量驱动数据来驱动模式将可能对青藏高原地区的降尺度模拟结果有更进一步的改善。

The Qinghai-Xizang Railway （QXR）， especially the 550-kilometer segment from Xidatan to Anduo， traverses a region characterized by complex terrain， diverse landforms， and a permafrost environment.In recent years， climate warming and permafrost degradation have significantly increased maintenance demands for the QXR， which is constructed atop permafrost.To understand the impacts of the intricate terrain and landforms along the QXR on local climate changes and to provide theoretical support for its operation and maintenance， this study utilizes station observation data along the railway and employs the Weather Research and Forecasting （WRF） model driven by ERA5 data to conduct high-resolution simulations with a grid resolution of 10 km×10 km.The results indicate that the six stations along the Qinghai-Xizang Railway exhibit a general warming trend from 1998 to 2020， with the lowest temperature increase rate at 0.27 ℃·（10a）^-1 and the highest at 0.56 ℃·（10a）^-1 years.The WRF model's simulation results show some discrepancies with observed annual mean temperature data.The simulation results are more accurate in summer and autumn， with correlation coefficients above 0.95 in summer and above 0.80 in autumn， but less accurate in spring and winter.Regarding precipitation simulation， the Nudging method effectively reduces the wet bias in summer precipitation on the Qinghai-Xizang Plateau.Precipitation gradually increases from the northern to the southern section of the railway， peaking in the central section.However， the WRF model still exhibits cold and wet biases in temperature and precipitation simulations on the Qinghai-Xizang Plateau.Exploring new methods or utilizing higher-quality driving data may further enhance the downscaled simulation results for this region.