Interpretation (and subsequent prediction) of the lithofacies geometry of ancient river-channel deposits requires full understanding of the formative processes. This is ideally gained by linking channel geometry and hydraulics with sediment erosion, transport, and deposition, using generalized (quantitative) physical models. Such models exist for high-stage deposition in single curved channels of simple planform: they are capable of approximating the three-dimensional variation of mean grain size and internal structure of point-bar deposits in channels with differing geometries and flow characteristics. However, such models cannot presently predict processes operating on point bar tops (e.g., sheet floods, chute channel and bar formation, scroll bars, flow separation zone ) or the nature of low-flow deposits. Lateral lithofacies variation due to meander-loop evolution and cutoff is also inadequately understood.

Generalized physical models of braided and anastomosed river deposition are particularly poorly developed, and need urgent attention. Single-channel and braided rivers can be distinguished on the basis of their water discharge, slope, width/depth ratio, and sinuosity; quantitative analysis of ancient alluvium is required for reconstruction of these parameters. Although braided river deposits should typically have a high proportion of coarse-grained channel fills relative to lateral accretion deposits, coarse-grained channel fills are also common in sinuous rivers with cutoffs. It appears that presently available qualitative facies models do not adequately represent the range of lithofacies geometries expected from different kinds of rivers, and therefore do not allow thorough and uneq ivocal interpretation of paleochannel geometry, flow characteristics, and evolution.

End_of_Article - Last_Page 429------------