Flash floods in dryland landscapes often begin with hydraulic bores characterized by rapidly increasing flow depths. It is not known how well bedload transport rates during floods with bores can be predicted using existing bedload transport equations, which generally assume steady, uniform flow. We conducted laboratory flume experiments with five different reproducible flash flood hydrograph shapes to relate gravel bedload transport rates to hydraulic forcing. Bed shear stresses were calculated using the Saint Venant assumptions for shallow water flow. For rapidly changing hydrographs with bores over initially dry beds, bedload transport rates fall within the variability of existing predictive equations. Our data are consistent with a scaling exponent of 1.5 (i.e., q(s)* proportional to (tau(b)*)(1.5) , where q(s)* is the nondimensional bedload transport rate and tau(b)* is the nondimensional bed shear stress). Thus, a key result is that previous bedload relations developed for quasi-steady flow may reasonably predict transport during rapidly changing hydrographs, provided that shear stress calculations account for unsteady flow. Critical nondimensional shear stresses in our experiments (tau(c)* <= 0.017 ) were lower than found in most previous studies, which we primarily attribute to the loose and unstructured bed state in these experiments. In contrast to initially dry beds, bores propagating over shallow flowing water had higher peak shear stresses but lower overall bedload transport rates, which we attribute to turbulent energy from the bore being dissipated through the flow in the vicinity of the bed, considerably reducing peak shear stresses acting on the actual bed.