USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5159

Back to Table of Contents

Re-evaluating the Historic Peak Discharges

A SAC run using the 2003 and 2006 flood data to calculate the discharge for various roughness coefficients calculated a velocity of 12.1 ft/s for XS3, but the velocities calculated in the subareas for XS2 were 14.3 ft/s for the left main channel and 4.9 ft/s in the right high-water bench. With the obstruction of the forested island/bar and such a large difference in the velocities in the two subareas, flows during peaks are not expected to be parallel to the banks of the reach because they would be going around the forested island/bar. Non-parallel flows would introduce some head losses that are not accounted for in the calculation of the energy equation, and for these reasons introduce some error in the calculation that are impossible to quantify. A detailed two-dimensional flow model could possibly account for these losses, but one is not currently available for this analysis. The n-verification study done for the 1949 flood did not have this problem because the island/bar was practically barren and flows were more likely to be parallel to the bank. Due to these hydraulic complications in the 2003 and 2006 peak flow n-verification calculations and the fact that the study reach in 1949 was more similar to the reach in 1921 than it was in 2003 or 2006, there is more confidence in the n verification and subsequent 1921 peak-flow recalculation based on the 1949 peak flow than the calculations using only the 2003 and 2006 peak-flow data.

Novak (1985) discusses the USGS criteria for revisions of peak flows as follows:

“Extremes and peaks greater than base discharge should be revised if the difference in discharge between old and new data is more than about 10 percent. Revisions may be made for errors less than 10 percent if they are needed to maintain the correct relationship between the annual maximum discharge or to keep annual maximums in the proper order of magnitude.”

The revised discharge values reported by Flynn and Benson in their unpublished report from1952 is less than 10 percent different than the published value, and because the magnitudes of the three other historic peaks that occurred in 1897, 1909, and 1917 are dependent on the magnitude of the 1921 peak discharge, Bodhaine, the USGS area engineer at the time, felt that a revision would not be needed (G.L. Bodhaine, U.S. Geological Survey, written commun., May 1954). Since that time, two new independent n-verification studies as described in this report came up with a similar conclusion that the currently published peak flows are too high. Together with the straight-line extension of the stage-discharge relation, there are now three independent estimations that all are lower than the published values.

The accuracy of the peak discharge of the four historic floods at this site is essential to the current flood-reduction study by the U.S. Army Corps of Engineers and the Skagit County Flood Insurance Study Update by FEMA. In order to provide the best data available for the current flood studies, the USGS revised historic floods at Skagit River near Concrete despite the fact that their estimates are all less than 10 percent from the currently published discharges (table 1). The 1921 peak discharge is determined from the recalculation using Stewart’s data, subdivision of cross section 2, and the n value calculated from the n verification using 1949 peak-flow data as described in this report. The other historical peak discharges are revised based on the rating extension described in this report and rounded to the nearest 5,000 ft³/s (table 1).

Back to Table of Contents