Universal approach for albedo estimation from multiple fine spatial-resolution satellite data: Preliminary consistency evaluation with Landsat 8, Sentinel-2 and GaoFen-1 albedo data

Surface albedo is a fundamental parameter in quantitative remote sensing and surface energy budget. The creation of fine spatial resolution spatiotemporal surface albedo products with high availability holds great significance in solar industry, particularly over heterogeneous landscapes. Despite the estimation methods being developed to generate surface albedo data for different sensors over recent years, it is crucial to develop a universal approach for albedo estimation and evaluate the consistency of their estimates, particularly with fine-resolution satellite data. Therefore, we conducted a study that involved estimating surface albedo data from top-of-atmosphere (TOA) reflectance using the direct estimation method for Landsat 8, Sentinel-2, and GaoFen-1 sensors. In this study, we assessed their consistency based on observations from SURFRAD, AmeriFlux, and HiWaterWSN ground stations and regional intercomparisons with existing satellite products. The results indicate that the validation root mean square error (RMSE) for the three sensors on SURFRAD stations are 0.0270, 0.0356 and 0.0409, respectively. Moreover, our study revealed that Landsat 8 and Sentinel-2 are consistent with ground observations and MODIS MCD43 products concerning time trends, and GaoFen-1’s multiangle observations can compensate for the vacancy of approximate nadir observations of Landsat 8 and Sentinel-2. Overall, the study demonstrates the universal applicability of the direct estimation method for multi-sensor fine-resolution albedo estimation, and lays the necessary theoretical foundation for producing seamless time-series fine-resolution surface albedo products over time.