We examined reach-scale change in riparian stands, tree mortality, wood recruitment, shade, and soil disturbance associated with riparian prescriptions for harvest adjacent to perennial headwater streams in the western Washington State Forest Practice Rules. We compared unharvested reference sites (REF) with three treatments: unbuffered clear-cut harvest (CC), 50-foot wide buffers (BUF), and buffers around perennial initiation points (PIP). Ten years post-harvest, cumulative change in live basal area was positive in REF stands (+3%) and negative in BUF (-14%) and PIP (-39%) stands. Cumulative mortality (percent basal area) was higher in PIP and BUF stands than REF stands (48, 27 and 9%, respectively). Post-harvest mortality rates in buffers were highest before year five and decreased dramatically between years five and ten; wind was the dominant mortality agent. Cumulative recruited wood volume in BUF and PIP reaches was double and four times REF volumes, respectively. Many fallen trees in buffers were uprooted and fell spanning over narrow channels. One year after harvest, canopy closure was lower in BUF (76%) and PIP (52%) reaches compared to REF reaches (89%), but both increased to over 85% by year ten. Canopy closure in the CC reaches was 12% one-year post-harvest, increasing to 37% by year five, and 72% by year ten due to growth of shrubs and saplings. Harvest-related soil disturbance occupied 6% of surface area in 30-foot wide equipment limitation zones adjacent to CC reaches but was minimal in buffers. Twice as many uprooted trees delivered sediment to streams in BUF reaches as in REF reaches, but the percentage of root-pits that delivered sediment was greater in REF reaches. Mean horizontal distance to the stream for root-pits that delivered sediment was 8 feet. Overall, 50-foot buffers reduced change in shade and wood input regimes compared to unbuffered reaches.