陆地稳定目标的地表反射率是利用反射率基法其对卫星传感器进行辐射定标的关键参量。但对于国产历史遥感卫星的光学载荷，同步观测的稳定目标地表反射率数据的缺乏，是对其进行历史再定标工作的主要难点。本文基于长时间序列的MODIS BRDF模型参数产品，采用统计分析的方法，建立了全球典型沙漠稳定目标的逐月方向反射率参考模型，并基于独立的样本数据进行了模型精度验证及辐射定标应用示范研究。验证结果显示，国外目标的方向反射率参考模型不确定度优于国内目标。国外的13个沙漠目标中，有12个目标的模型误差在±3%以内，其中Libya 4可低于±1%；在国内，有6个目标与国外这12个目标的模型精度相当。辐射定标应用示范结果显示，本文构建的稳定目标方向反射率参考模型可以代替实时的MODIS BRDF模型参数产品用于遥感仪器的辐射定标中。一方面，可以解决国产卫星光学载荷历史数据再定标中缺乏地面反射率同步观测的问题；另一方面，对于在轨遥感仪器的辐射定标中，可避免频繁下载MODIS BRDF模型参数产品的问题。

Objective: Accurate and consistent calibration for the long-term satellite remote sensing data is critical to accurately estimate multi-decadal climate variability. However, for the historical optical satellite sensors, the lack of or limited number of synchronous in situ observation of surface Bidirectional Reflectance Distribution Function (BRDF) model data is the main difficulty to recalibrate their radiance products. The goal of this paper is to build a surface directional reflectance reference model for some typical stable earth targets on the worldwide, which can provide a uniform surface reflectance reference for the historical remote sensing satellites to perform a consistent recalibration of their long-term TOA radiance products. Method: The long-term data products form MODIS satellite instrument, well known by its high calibration accuracy and stability, are used to establish the surface directional reflectance reference model for the selected stable earth targets. To be specific, for the desert sites, the MODIS surface BRDF products (MCD43A1) from the year 2008 to 2012 are used to build a monthly climatology of the BRDF parameters as the surface directional reflectance reference. Then, independent MODIS products, i.e., the data time is different from that used in the model construction, are applied to verify the accuracy of the established directional reflectance reference model. At last, in order to check the applicability of this model in the radiometric calibration of satellite instruments, this model was used to carry out calibration experiments on the Visible Infrared Radiometer (VIRR) aboard on the FY-3C, and compared with the calibration results obtained based on real-time MODIS BRDF products. Result: The uncertainty estimation of the surface directional reflectance reference model over desert sites shows that the model uncertainty of foreign desert targets is lower than that of domestic desert targets. The reference model uncertainty of foreign targets is below 3%, and the domestic desert sits are below 4% (except TKLM_3). The model accuracy verification of the radiometric reference model show that among the 13 foreign desert targets, except Tinga_Tingana in Australia, the relative errors are basically within ±3%， and that in Libya 4 can be less then ±1%. There are 6 domestic desert sites, i.e., BDJL_1, BDJL_2, TNGR_1, TNGR_2, WULBHE and LBPO_W, can achieve a sound accuracy of within ±3%, which is comparable to the stable earth targets frequently used by the international research. The radiometric calibration test to FY-3C VIRR shows that using the directional reflectance reference model of the stable target are in good agreement with the radiometric calibration results based on the real-time MODIS BRDF product, and the difference between the two calibration results is very small, i.e., the average relative bias is basically within ±0.6%. Conclusion: The above results indicates that established directional reflectance reference model over the stable earth targets can be used to in the application of radiometric calibration of the optical satellite instruments and can be helpful to solve the problem of consistent recalibration of historical data from domestic optical satellite instruments.