Design criteria for riprap at bridge piers in rivers is based on the specification of a size, gradation, and cover that does not fail under an appropriately chosen flood flow. Experimental tests of riprap performance at bridge piers to date have relied on a configuration for which the ambient bed is not mobilized, that is, clear-water conditions. In the field, however, riprap is, as a rule, subjected to mobile-bed conditions during floods. Recent experiments by three cooperating research groups (University of Auckland, Nanyang University, and St. Anthony Falls Laboratory) indicate a heretofore unrecognized mechanism for riprap failure under mobile-bed conditions. When the flow is in the dune regime, the passage of successive dunes causes riprap that is never directly entrained by the flow to sink and disperse. Pier scour is realized as a consequence of these processes. In some cases, the depth of scour realized is not significantly less than that which would occur without riprap. When the riprap is fully underlain by a geotextile, edge effects can cause local removal of riprap, upturning of the geotextile, and general failure. When the riprap is underlain by a partial geotextile (i.e., one that covers an area less than the riprap), edge scour causes local sinking that anchors the geotextile. The sinking and dispersion of the rest of the riprap are greatly limited, and the riprap fails only when flow velocities are sufficient for direct entrainment. The experiments suggest improved design criteria for the installation of riprap in the field.
