Estimation of potential scour at bridges on local government roads in South Dakota, 2009-12

Abstract

In 2009, the U.S. Geological Survey and South Dakota Department of Transportation (SDDOT) began a study to estimate potential scour at selected bridges on local government (county, township, and municipal) roads in South Dakota. A rapid scour-estimation method (level-1.5) and a more detailed method (level-2) were used to develop estimates of contraction, abutment, and pier scour. Data from 41 level-2 analyses completed for this study were combined with data from level-2 analyses completed in previous studies to develop new South Dakota-specific regression equations: four regional equations for main-channel velocity at the bridge contraction to account for the widely varying stream conditions within South Dakota, and one equation for head change. Velocity data from streamgages also were used in the regression for average velocity through the bridge contraction. Using these new regression equations, scour analyses were completed using the level-1.5 method on 361 bridges on local government roads. Typically, level-1.5 analyses are completed at flows estimated to have annual exceedance probabilities of 1 percent (100-year flood) and 0.2 percent (500-year flood); however, at some sites the bridge would not pass these flows. A level-1.5 analysis was then completed at the flow expected to produce the maximum scour. Data presented for level-1.5 scour analyses at the 361 bridges include contraction, abutment, and pier scour. Estimates of potential contraction scour ranged from 0 to 32.5 feet for the various flows evaluated. Estimated potential abutment scour ranged from 0 to 40.9 feet for left abutments, and from 0 to 37.7 feet for right abutments. Pier scour values ranged from 2.7 to 31.6 feet. The scour depth estimates provided in this report can be used by the SDDOT to compare with foundation depths at each bridge to determine if abutments or piers are at risk of being undermined by scour at the flows evaluated. Replicate analyses were completed at 24 of the 361 bridges to provide quality-assurance/quality-control measures for the level-1.5 scour estimates. An attempt was made to use the same flows among replicate analyses. Scour estimates do not necessarily have to be in numerical agreement to give the same results. For example, if contraction scour replicate analyses are 18.8 and 30.8 feet, both scour depths can indicate susceptibility to scour for which countermeasures may be needed, even though one number is much greater than the other number. Contraction scour has perhaps the greatest potential for being estimated differently in replicate visits. For contraction scour estimates at the various flows analyzed, differences between results ranged from -7.8 to 5.5 feet, with a median difference of 0.4 foot and an average difference of 0.2 foot. Abutment scour appeared to be nearly as reproducible as contraction scour. For abutment scour estimates at the varying flows analyzed, differences between results ranged from -17.4 to 11 feet, with a median difference of 1.4 feet and an average difference of 1.7 feet. Estimates of pier scour tended to be the most consistently reproduced in replicate visits, with differences between results ranging from -0.3 to 0.5 foot, with a median difference of 0.0 foot and an average difference of 0.0 foot. The U.S. Army Corps of Engineers Hydraulics Engineering Center River Analysis Systems (HEC-RAS) software package was used to model stream hydraulics at the 41 sites with level-2 analyses. Level-1.5 analyses also were completed at these sites, and the performance of the level-1.5 method was assessed by comparing results to those from the more rigorous level-2 method. The envelope curve approach used in the level-1.5 method is designed to overestimate scour relative to the estimate from the level-2 scour analysis. In cases where the level-1.5 method estimated less scour than the level-2 method, the amount of underestimation generally was less than 3 feet. The level-1.5 method generally overestimated contraction, abutment, and pier scour relative to the level-2 method, as intended. Although the level-1.5 method is designed to overestimate scour relative to more involved analysis methods, many assumptions, uncertainties, and estimations are involved. If the envelope curves are adjusted such that the level-1.5 method never underestimates scour relative to the level-2 method, an accompanying result may be excessive overestimation.