The U.S. Geological Survey and the Alaska Department of Transportation and Public Facilities undertook a cooperative multiphase study of streambed scour at selected bridges in Alaska beginning in 1994. Of the 325 bridges analyzed for susceptibility to scour in the preliminary phase, 54 bridges were selected for a more intensive analysis that included site investigations. Cross-section geometry and hydraulic properties for each site in this study were determined from field surveys and bridge plans. Water-surface profiles were calculated for the 100- and 500-year floods using the Hydrologic Engineering Center?s River Analysis System and scour depths were calculated using methods recommended by the Federal Highway Administration. Computed contraction-scour depths for the 100- and 500-year recurrence-interval discharges exceeded 5 feet at six bridges, and pier-scour depths exceeded 10 feet at 24 bridges. Complex pier-scour computations were made at 10 locations where the computed contraction-scour depths would expose pier footings. Pressure scour was evaluated at three bridges where the modeled flood water-surface elevations intersected the bridge structure. Site investigation at the 54 scour-critical bridges was used to evaluate the effectiveness of the preliminary scour analysis. Values for channel-flow angle of attack and approach-channel width were estimated from bridge survey plans for the preliminary study and were measured during a site investigation for this study. These two variables account for changes in scour depths between the preliminary analysis and subsequent reanalysis for most sites. Site investigation is needed for best estimates of scour at bridges with survey plans that indicate a channel-flow angle of attack and for locations where survey plans did not include sufficient channel geometry upstream of the bridge.