Heterogeneities in fluvial sandstones can be classified in outcrop using a sixfold hierarchy of lithosome-bounding surfaces. Individual lithofacies units or multistory units are bounded by first- and second-order surfaces, respectively. These surfaces may be readily identified in core, but can rarely be correlated reliably between wells because of their limited areal extent (less than 10 ha. or 25 ac). Outcrop studies provide information on typical geometries and size ranges for input into reservoir model studies.

Macroforms are complex, compound bars (e.g., point bars) bounded by fourth-order surfaces. Accretionary phases may be separated by internal, low-angle third-order surfaces. These depositional units are 10²-10³ m (328-3,280 ft) across and, because of their limited size and internal complexity, may require well spacings of 32 or 16 ha. (80 or 40 ac) for reliable mapping.

Fifth-order surfaces define major channel bodies, ranging from ribbon to sheetlike in geometry. Sheet sandstones may be mappable using closely spaced well data, but ribbon sandstones are difficult to correlate except in the most well-developed field. Three-dimensional seismic is a powerful new tool for mapping fifth-order surfaces. The largest scale represents units at the member or submember level, bounded by sixth-order surfaces. These units are mappable using conventional wireline log correlation.

The area of a shale bed depends on the scale of the lithosome with which it is associated. For example, shales associated with sixth-order bounding surfaces may represent basin-wide periods of low fluvial energy or lacustrine flooding, whereas those lying on second-order surfaces are the result of mud drape over bed forms in small abandoned channels, and may have areal extents of only a few square meters. Such data need to be built into computer models of reservoir flow behavior.