Scour at bridge sites in Delaware, Maryland, and Virginia

Scour data were obtained from discharge measure- ments to develop and evaluate the reliability of constriction-scour and local-scour equations for rivers in Delaware, Maryland, and Virginia. No independent constriction-scour or local-scour equations were developed from the data because no significant relation was deter-mined between measured scour and streamflow, streambed, and bridge characteristics. Two existing equations were evaluated for prediction of constriction scour and 14 existing equations were evaluated for prediction of local scour. Constriction-scour data were obtained from historical stream discharge measurements, field surveys, and bridge plans at nine bridge sites in the three-State area. Constriction scour was computed by subtracting the average-streambed elevation in the constricted reach from an uncontracted-channel reference elevation. Hydraulic conditions were estimated for the measurements with the greatest discharges by use of the Water-Surface Profile computation model. Measured and calculated constriction-scour data were used to evaluate the reliability of Laursen's clear-water constriction-scour equation and Laursen's live-bed constriction-scour equation. Laursen's clear-water constriction-scour equation underestimated 21 of 23 scour measure- ments made at three sites. A sensitivity analysis showed that the equation is extremely sensitive to estimates of the channel-bottom width. Reduction in estimates of bottom width by one-third resulted in predictions of constriction scour slightly greater than measured values for all scour measurements. Laursen's live-bed constriction- scour equation underestimated 10 of 14 scour measurements made at one site. The error between measured and predicted constriction scour was less than 1.0 ft (feet) for 12 measure-ments and less than 0.5 ft for 8 measurements. Local-scour data were obtained from stream discharge measurements, field surveys, and bridge plans at 15 bridge sites in the three-State area. The reliability of 14 local-scour equations were evaluated. From visual inspection of the plotted data, the Colorado State University, Froehlich design, Laursen, and Mississippi pier-scour equations appeared to be the best predictors of local scour. The Colorado State University equation underestimated 11 scour depths in clear-water scour conditions by a maximum of 2.4 ft, and underestimated 3 scour depth in live-bed scour conditions by a maximum of 1.3 ft. The Froehlich design equation under- estimated two scour depth in clear-water scour conditions by a maximum of 1.2 ft, and under- estimated one scour depth in live-bed scour conditions by a maximum of 0.4 ft. Laursen's equation overestimated the maximum scour depth in clear-water scour conditions by approximately one-half pier width or approximately 1.5 ft, and overestimated the maximum scour depth in live-bed scour conditions by approximately one-pier width or approximately 3 ft. The Mississippi equation underestimated six scour depths in clear-water scour conditions by a maximum of 1.2 ft, and underestimated one scour depth in live-bed scour conditions by 1.6 ft. In both clear-water and live-bed scour conditions, the upper limit for the depth of scour to pier-width ratio for all local scour measurements was 2.1. An accurate pier- approach velocity is necessary to use many local pier-scour equations for bridge design. Velocity data from all the discharge measurements reviewed for this investigation were used to develop a design curve to estimate pier-approach velocity from mean cross-sectional velocity. A least- squares regression and offset were used to envelop the velocity data.