The primary input of energy to the Earth's climate system occurs at the surface and can be highly sensitive to the surface albedo. Albedo changes have been proposed as one cause of climatic variation, but results from climate models are not yet consistent. It is very difficult to establish an agreed global data set with which to initiate comparative climatic simulations. Albedo observations must be spectrally resolved because reflexion of solar radiation is a strong function of wavelength and incident and reflected beams are modified by the atmosphere. Parametrization of system albedos in energy-balance models draws on satellite data. The use of satellite observations is less easy in general circulation climate models. The removal of atmospheric distortion is particularly difficult. The establishment of a surface albedo data set generally follows one of two approaches: geographical categorization or remote monitoring. Surface albedo specification in current general circulation models is diverse. This paper reviews the ways in which remotely derived albedo measurements are used now and may, in the future, be improved for climate research.