The occurrence of two wildfires separated by 31 yr in the chaparral-dominated Arroyo Seco watershed (293 km2) of California provides a unique opportunity to evaluate the effects of wildfire on suspended-sediment yield. Here, we compile discharge and suspended-sediment sampling data from before and after the fires and show that the effects of the postfire responses differed markedly. The 1977 Marble Cone wildfire was followed by an exceptionally wet winter, which resulted in concentrations and fluxes of both fine and coarse suspended sediment that were ∼35 times greater than average (sediment yield during the 1978 water year was 11,000 t/km2/yr). We suggest that the combined 1977–1978 fire and flood had a recurrence interval of greater than 1000 yr. In contrast, the 2008 Basin Complex wildfire was followed by a drier than normal year, and although suspended-sediment fluxes and concentrations were significantly elevated compared to those expected for unburned conditions, the sediment yield during the 2009 water year was less than 1% of the post–Marble Cone wildfire yield. After the first postfire winters, sediment concentrations and yield decreased with time toward prefire relationships and continued to have significant rainfall dependence. We hypothesize that the differences in sediment yield were related to precipitation-enhanced hillslope erosion processes, such as rilling and mass movements. The millennial-scale effects of wildfire on sediment yield were explored further using Monte Carlo simulations, and these analyses suggest that infrequent wildfires followed by floods increase long-term suspended-sediment fluxes markedly. Thus, we suggest that the current approach of estimating sediment yield from sediment rating curves and discharge data—without including periodic perturbations from wildfires—may grossly underestimate actual sediment yields.