The climate system is driven, primarily, by energy absorbed at the surface. Surface albedo sensitivity is incorporated into all types of climate models, and changes can lead to large feedback effects. For example, alterations in the extent and/or state of the cryosphere and large‐scale modification of vegetation cause significant perturbations in climate model results. The specification of surface albedo in general circulation climate models (GCM's) differs. An improved and agreed surface albedo data set is urgently required for climate modeling. It is likely that the most appropriate means of achieving consistent and credible surface albedos is by using well‐designed satellite surveillance to augment global inventories of soils and vegetation. However, retrieval of surface albedo values for all sky and surface conditions from satellite observations is difficult. Atmospheric distortion is especially hard to remove. Some of the sensitivity of GCM's to surface albedo values may be the result of inadequate parameterization of other climatic components. The accuracy of information demanded by climate modelers could be reduced and made more consistent. Recommendations are made for the implementation of a new global scale observational program with the aim of providing surface albedo data at an accuracy of ±0.05 within 5–10 years. Immediate initiation is urged.