作物残茬等非光合植被在干旱、半干旱地区生态系统物质循环、能量流动过程中承担着不可替代的角色，同时在阻抑土壤侵蚀、保持土壤水分、促进土壤发育等方面也具有重要作用。张家口坝上地区位于首都两区建设及京津风沙源治理的核心区域，利用遥感手段估算该地区作物残茬生物量，对区域风蚀状况评估、生态环境评价及碳、氮循环研究具有重要意义。基于实地测量的作物残茬生物量、Sentinel-2光学影像、Sentinel-1雷达影像构建作物残茬光学遥感指数和雷达遥感指数，采用最优指数归一化相乘和多元线性逐步回归分析方法，建立结合光学与雷达遥感的作物残茬生物量估算模型，计算并分析2017—2023年张家口坝上地区作物残茬生物量。结果表明：（1） 光学遥感指数中，由Sentinel-2的短波红外波段（B11和B12）构建的RI_(11,12)指数与作物残茬生物量的相关性最高，模型决定系数（R²）为0.744。雷达遥感指数中，交叉极化（VH）后向散射系数与作物残茬生物量的相关性最高，R²为0.409。（2） 结合光学与雷达遥感估算模型中，多元线性逐步回归模型精度最高，R²为0.796，均方根误差（RMSE）为8.84 g·m^-2，可较好地预测作物残茬生物量。（3） 构建的作物残茬生物量估算模型精度较单纯使用光学遥感高约9.72%，较单纯使用雷达遥感高约66.74%。（4） 2017—2023年张家口坝上地区年均作物残茬生物量为23.74×10⁴ t，呈波动下降趋势，作物残茬生物量年际变化受气温和降水的影响，近年来，因土地流转政策造成的种植结构变化是导致该地区作物残茬生物量下降的重要因素。

Non-photosynthetic vegetation, such as crop stubble, plays a crucial role in material cycling and energy flow in arid and semi-arid ecosystems. It also significantly contributes to inhibiting soil erosion, retaining soil moisture, and promoting soil development. The Bashang region of Zhangjiakou Hebei Province, China is a core area for the ecological construction of Beijing-Tianjin sandstorm control and the development of the two capital areas. Estimating crop stubble biomass in this region using remote sensing is essential for evaluating regional wind erosion, the ecological environment, and the carbon and nitrogen cycles. This study utilized measured crop stubble biomass, Sentinel-2 optical images, and Sentinel-1 radar images to construct optical and radar remote sensing indices of crop stubble. Using optimal index normalization and multiple linear stepwise regression analysis, an estimation model combining optical and radar remote sensing was developed to calculate and analyze crop stubble biomass in the Bashang region from 2017 to 2023. The results show that: (1) Among the optical remote sensing indices, the RI_(11,12) index, derived from Sentinel-2 short-wave infrared bands (B11 and B12), showed the highest correlation with crop stubble biomass, with a determination coefficient (R²) of 0.744. For radar remote sensing indices, the cross-polarization (VH) backscattering coefficient had the highest correlation with crop stubble biomass, achieving an R² of 0.409. (2) The multivariate linear stepwise regression model demonstrated the highest accuracy, with an R² of 0.796 and a root mean square error (RMSE) of 8.84 g·m^-2, making it a reliable predictor of crop stubble biomass. (3) The estimation model incorporating both optical and radar remote sensing indices improved prediction accuracy by approximately 9.72% compared to optical remote sensing alone and by 66.74% compared to radar remote sensing alone. (4) From 2017 to 2023, the average annual crop stubble biomass in the Bashang region was 23.74×10⁴ t, exhibiting a fluctuating downward trend. Annual variations in crop stubble biomass were influenced by air temperature and precipitation, while changes in planting structures driven by land transfer policies were a significant factor contributing to the decline in recent years.