Understanding and prediction of man's environment rests most heavily on observation. Since revolutionary developments were made in communication science (telegraphy and radio) in the mid to late nineteenth century there has been a dramatic upsurge in the demand for observational data of the Earth. At first this demand was served by the development of conventional monitoring programmes. Unfortunately, economic, practical and political factors have led to a growing shortfall of such data as the twentieth century has advanced (Barrett, 1980a). Earth-observation satellite systems, effectively inaugurated by Tiros-I in 1960, have given a tremendous fillip to broad areas within the environmental sciences, due especially to the areal extent and continuity of the resulting environmental data, and its complementarity to in situ (conventional) observations. During the first two decades of Earth-observation satellite activities two types of orbital configurations have been dominant, namely: a) Near-polar sunsynchronous orbits, from which individual satellites commonly observe given areas on the Earth every 12 hours at the same local times; and b) Equatorial geosynchronous (i.e. geostationary) orbits, in which satellites remain over a selected locality on the Earth's surface, and observe the visible disc much more frequently, commonly every 30 minutes. It is these satellites which form the subject of this review. In recent years it has become fashionable to split Earth-observation satellites into two broad categories on another basis also, related to the principal objects of their attention. Thus we can speak of: a) Environmental satellites, characterized by relatively low resolution data (at best c.0.5 km linear resolution) obtained relatively frequently (at least twice per day from any single locality); all geostationary Earth-observation satellites to date have fallen into this group. b) Earth Resource satellites, which provide relatively high resolution data (e.g. the nominal 80 m linear resolution of Landsat imagery), but relatively infrequently (commonly every 18 days from a single satellite in the case of Landsat). Although such a distinction is both apt and helpful in some conceptual and organizational contexts, it has one unfortunate effectit has tended to polarize the use of the data from those two categories into atmospheric and broad-scale at SETON HALL UNIV on April 1, 2015 ppg.sagepub.com Downloaded from 160 oceanic applications in the first case, and terrestrial and detailed oceanic applications in the second. However, it has been shown that Earth Resource satellites (e.g. Landsat) have provided data of considerable value in atmospheric science (e.g. the report by Barrett and Grant, 1977), and that Environmental satellites (e.g. Meteosat) can be employed with value for Earth Resource applications (e.g. the report by Barrett and Hamilton, 1980). Interest in this possibility has been enhanced by the establishment of a working group of EARSeL (the European Association of Remote Sensing Laboratories) to consider '...