AbstractSurface radiation data were collected at seven sites in the Riviera Valley in southern Switzerland during the special observing period of the Mesoscale Alpine Programme's Riviera project from August until October 1999. During the daytime on selected ‘valley wind’ days, strong site contrasts occurred with respect to solar and downward long‐wave radiation fluxes. Site‐to‐site differences of incoming global radiation were caused by the different orientations of valley surfaces relative to the direct solar beam component and by shading from surrounding topography. Diurnal cycles of albedo, determined from horizontal and slope‐parallel measurements of reflected and incoming global radiation at three eastern side‐wall sites, show that reflection of the direct radiation component from the inclined surfaces is predominantly anisotropic diffuse radiation. Downward long‐wave radiation decreases with elevation in the valley because the higher sites have an increased angular exposure to the radiating sky (i.e. sky view factor). Contrasts between the valley‐floor and ridge‐top sites are largest on valley wind days but also occur on overcast days. Finally, strong site‐to‐site variations of the net all‐wave radiation occur during daytime due to variations in the received direct beam component of incoming global radiation. During night‐time, net outgoing all‐wave radiation increases with elevation, again because sky view fraction increases with elevation. On valley wind days, significant contrasts are found in mean diurnal cycles and daily sums of available radiation energy between south‐west and north‐east facing slopes. When the direct beam component is absent, inter‐site contrasts in radiation components are reduced, especially short‐wave radiation components.A simple semi‐empirical method is described to estimate the net all‐wave radiation in a slope‐parallel frame of reference from measurements with horizontally exposed net radiometers. This method employs the theoretical ratio of short‐wave radiation on a slope to the horizontal equivalent. The method is tested successfully at one site where simultaneous measurements were made with inclined and horizontal sensors. Copyright © 2003 Royal Meteorological Society.