AbstractAn Erratum has been published for this article in Hydrological Processes 16(5) 2002, 1131–1133. Wildfires in the Colorado Front Range can trigger dramatic increases in runoff and erosion. A better understanding of the causes of these increases is needed to predict the effects of future wildfires, estimate runoff and erosion risks from prescribed fires, and design effective post‐fire rehabilitation treatments. The objective of this project was to determine whether runoff and sediment yields were significantly related to the site variables of burn severity, percent cover, soil water repellency, soil moisture, time since burning, and slope. To eliminate the variability due to natural rainfall events, we applied an artificial storm of approximately 80 mm h−1 on 26 1 m2 plots in the summer and fall of 2000. The plots were distributed among a June 2000 wildfire, a November 1999 prescribed fire, and a July 1994 wildfire.For 23 of the 26 plots the ratio of runoff to rainfall exceeded 50%. Nearly all sites exhibited strong natural or fire‐induced water repellency, so the runoff ratios were only 15–30% larger for the high‐severity plots in the two more recent fires than for the unburned or low‐severity plots. The two high‐severity plots in the 1994 wildfire had very low runoff ratios, and this probably was due to the high soil moisture conditions at the time of the simulated rainfall and the resulting reduction in the natural water repellency. Sediment yields from the high‐severity sites in the two more recent fires were 10–26 times greater than the unburned and low‐severity plots. The plots burned at high severity in 1994 yielded only slightly more sediment than the unburned plots. Percent ground cover explained 81% of the variability in sediment yields, and the sediment yields from the plots in the 1994 wildfire are consistent with the observed recovery in percent ground cover. Copyright © 2001 John Wiley & Sons, Ltd.