In the Joarcam-Beaverhill Lake area of Alberta, there are three long (40 km), narrow (5 km) sandbodies in the late Albian Viking Formation. They are surrounded by offshore open marine mudstones and have previously been interpreted as offshore bars. However, the sandbodies are both underlain and overlain by erosion surfaces. Joarcam is the only field with abundant cores, and the sandbody shows a sandier-upward and coarsening-upward succession. All three fields have recently been reinterpreted in the literature as lowstand shoreface deposits, but the order in which they formed remains controversial. Our well log cross sections suggest that the Lindbrook shoreface was the first to form at a time of maximum lowstand of sea level. It was truncated as sea level began to rise, and the Lindbrook transgressive surface of erosion (TSE) has been traced into the erosion surface that underlies Joarcam. Joarcam can therefore be interpreted as a shoreface incised during a pause in an overall transgression. Both the Lindbrook and Joarcam shorefaces appear to be truncated by the regressive surface of erosion (RSE) that underlies Beaverhill Lake. This suggests that Beaverhill Lake is the youngest of the shorefaces and implies a second major fall of relative sea level. The lowstand shorefaces tend to be thinner because of reduced accommodation space during sea level fall. They tend to prograde farther (Beaverhill Lake, about 23 km) because of abundant sediment supply at lowstand, but consist largely of sandstone directly overlying the RSE. At Beaverhill Lake, the shoreface sediments prograded at least 10 km before any mud in the offshore-to-shoreface transition was preserved. By contrast, the transgressive shorefaces tend to be thicker because of accommodation space created during transgression. However, they tend not to prograde so far (Joarcam, up to 13 km), possibly because of reduced supply of sediment to the shoreface during transgression. The two proposed episodes of lowering of relative sea level are tentatively correlated with the two separate valley incisions documented at Crystal, about 90 km to the southwest. These incisions represent type 1 sequence boundaries, but the boundaries are not continuous regressive surfaces of erosion from Crystal to the Joarcam-Beaverhill Lake area. In both cases, the falling-stage RSEs have been eroded by the subsequent TSEs; these TSEs are the only surfaces preserved between the incised valleys and the lowstand shorefaces.