We present a global 1-km albedo dataset derived from Advanced Very High Resolution Radiometer (AVHRR) satellite remote sensing data, a global land cover classification and field observations of typical land surface bidirectional reflectance distribution functions (BRDFs). There is a clear need for such a global albedo map for use in atmospheric radiative transfer models. A land cover classification based on surface structure is employed and class correspondences used to generate a structural class map from an existing ecological global land classification. Representative field measurements of BRDF are assigned to each of the structural classes allowing us to retrieve BRDF shape parameters and albedo for each class using a semiempirical model. These parameters were adjusted using global AVHRR data to account for intra-class variations of surface reflectance. Albedos were calculated from the adjusted parameters and summarised by land cover type, continent and season.