The Upper South Platter River watershed provides 80% of the water for the two million people in the Denver area. Erosion after large wildfires in 1996 and 2000 severely degraded water quality and reduced reservoir storage capacities. In response, the USDA Forest Service is conducting large-scale thinning and prescribed burning to reduce the risk of catastrophic wildfires. The initial goal of this study was to evaluate the effects of thinning, prescribed burning, and the associated forest roads on runoff, sediment production, and channel conditions at the hillslope and small catchment scales. The study area is about 45 km southwest of Denver, Colorado. Elevations range from 2300-2750 m. Mean annual precipitation is 410 mm, and about 40% of this falls as snow. The predominant vegetation type is ponderosa pine, and the soils are shallow, coarse-textured (sandy-skeletal, Kassler and Sphinx series) derived from the Pikes Peak granite. In mid-2001 sediment fences were installed to measure erosion rates from 20 pairs of convergent hillslopes (“swales”) and 24 road segments in three study areas. The contributing area, slope, surface cover, and soil particle-size distribution were measured for each swale and road segment. Channel morphology and monthly water quality data also were collected from each study area. In 2001, just 3 of the 40 swales produced sediment. In contrast, sediment was generated from 17 of the 24 road segments. Summer 2002 was much drier, and in June 2002, the Hayman wildfire burned all 20 swales in one of the study areas. In the sites burned at high severity, the mean ground cover decreased from 90% to 6%, and the soils were strongly water repellent to a depth of 6 or 9 cm. The first two storms of 11 and 17 mm caused extensive rilling and channel incision, and an 11 mm, 45 minute storm resulted in a mean erosion rate of 0.75 kg m^-2. On a storm-by-storm basis, the amount of sediment eroded from the severely-burned swales was strongly correlated with rainfall intensity (R²=0.92). None of the unburned swales produced any sediment in 2002. The lower summer rainfall in 2002 caused the mean sediment production rate for the 24 road segments to drop from 2.0 kg m^-2 in 2001 to 0.7 kg m^-2 in 2002. On average, the surface soils in the swales averaged 50% gravel, 40% sand, 9% silt and only 1% clay. The composition of the eroded sediment from the severely-burned swales was quite similar, as gravel and sand comprised approximately 90% of the eroded sediment. The proportion of sand and gravel increased rapidly with increasing rainfall erosivity. Contributing area times slope explained approximately three-quarters of the variability in erosion for the burned swales immediately after the fire. Subsequent data from the second year after burning indicates that changes in the cross-sectional area of the central rill explained approximately 64% of the variability in sediment production. Since the contributing area is effectively a surrogate for the amount of runoff and the slope reflects the energy available for sediment detachment and transport, this result is consistent with the observation that 80% of the sediment is being generated by continuing incision of the central rill rather than sheetwash on the hillslopes. When separated by particle size, area times slope also explains most of the variability in the amount of gravel and sand, half of the variability in silt, but only about one-third of the variability in the amount of clay. On average, the road surface had slightly more sand and slightly less gravel than the surface soils in the swales. The particle-size distribution of the sediment eroded from the road surface was very similar to the texture of the road surface. However, the eroded sediment had slightly more sand, silt, and clay than the road surface, indicating that the finer particles were being preferentially eroded. Contributing area times slope explained approximately half of the variability in sediment production for the road segments. The proportion of gravel in the sediment eroded from the unpaved roads tended to increase with increasing storm erosivity while the proportion of silt tended to decrease. These results suggest that the detachment and transport of sand and gravel is more tightly coupled to rainfall intensity than the amounts of silt and clay. The implication is that erosion of the finer particles is supply limited, and erosion of the coarser particles is more transport limited.