Search

Land surface albedo plays a critical role in climate, hydrological and biogeochemical modeling, and weather forecasting. It is often assigned in models and satellite retrievals by albedo climatology look‐up tables using land cover type and other variables; however, there are considerable differences in albedo simulations among models, which partially result from uncertainty in obsolete albedo climatology. Therefore, this study introduces a new global 500 m daily blue‐sky land surface albedo climatology data set under both actual and snow‐free surface conditions utilizing 20‐year Moderate Resolution Imaging Spectroradiometer products from Google Earth Engine. In situ measurements from 38 long‐term‐maintained sites were utilized to validate the accuracies of different albedo climatology datasets. The root‐mean‐square error, bias, and correlation coefficient of the new climatology are 0.031, −0.003, and 0.96, respectively, which are more accurate than the Global Land Surface Satellite, GlobAlbedo, and 16 model datasets. Data intercomparison suggests that ERA5 exhibits better performance than Modern‐Era Retrospective Analysis for Research and Applications Version 2 (MERRA2) and 14 Coupled Model Intercomparison Projects Phase 6 models. However, it contains positive biases in the snow‐free season, while MERRA2 underestimates the snow albedo. Global albedo variation associated with basic surface plant functional types was also characterized, and snow impact was considered separately. Temporal variability analysis indicates that traditional climatology datasets with coarser temporal resolutions (≥8 days) cannot capture albedo variation over areas with distinct snow seasons, especially in central Eurasia and boreal regions. These results confirm the high reliability and robustness of the new albedo climatology in model assessment, data assimilation, and satellite product retrievals. Plain Language Summary: Surface albedo represents the Earth's surface ability to reflect solar energy, and thus directly influences atmospheric and water circulations by controlling available energies. As an essential variable estimated in current climate models, surface albedo is assigned by searching look‐up tables based on the land cover and soil types. However, there are large differences in simulated albedo among various models, partially because the albedo climatology in these tables was generated in the 1980s and needs to be updated. As global surface albedo has been accurately retrieved from satellite data for decades, we were able to produce a new albedo climatology data set under actual and snow‐free conditions with high spatiotemporal resolutions. Ground validation and data comparison indicate that the new data set performs with higher accuracy compared with other datasets. Moreover, assessment by the new climatology data set shows that albedo from advanced climate models and reanalysis datasets has clear uncertainties, especially over high latitudes and the Tibetan Plateau. Temporal variation in global albedo is better characterized at basic vegetation types. The new albedo climatology data set shows great potential for use in model assessment, modeling improvement, and satellite retrieval of other variables. Key Points: A global 500 m daily blue‐sky land surface albedo climatology data set in actual and snow‐free surface conditions was produced from Moderate Resolution Imaging SpectroradiometerSite validation and data comparison reveal its reliability and robustness in model assessment, data assimilation, and satellite retrievalsSimulated albedo data by state‐of‐art reanalysis and Coupled Model Intercomparison Projects Phase 6 models still have a clear bias in polar regions and the Tibetan Plateau [ABSTRACT FROM AUTHOR]