USGS Home
Contact USGS
Search USGS

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

Back to Table of Contents

Introduction

The peak discharge record at the U.S. Geological Survey (USGS) gaging station Skagit River near Concrete, Washington, is a key record being used by the U.S. Army Corps of Engineers to determine the 100-year flood discharge for the Skagit River Flood Study. The four largest annual peak discharges of record that are used to determine the 100-year flood discharge for this station occurred in 1897, 1909, 1917, and 1921. A peak discharge of 240,000 ft³/s for the flood on December 13, 1921, was determined in 1923 by USGS hydrologist James Stewart by averaging the results of a contracted-opening measurement and slope area. The peak discharges of the other three largest floods were determined from a stage-discharge rating after extending it on the basis of the slope-area peak discharge of the 1921 flood. The accuracy of the peak discharge of the 1921 flood has been questioned, thus putting into question the accuracies of the other three largest peak discharges.The peak discharge of the 1921 flood was first checked in the 1950s with a channel roughness coefficient (n value) verification study based on the 1949 peak discharge of 153,000 ft³/s that included a survey of high-water marks (HWMs) and four cross sections in the same reach used by James Stewart for his original slope-area measurement. H.C. Riggs and W.H. Robinson (U.S. Geological Survey, written commun., 1950) used this data to calculate an n value for the reach. In 1952, F.J. Flynn and M.A. Benson (U.S. Geological Survey, written commun., 1951–52) revised the n-verification study and used the revised n value to recalculate the 1921 peak discharge with channel geometry data compiled by James Stewart. They calculated a peak discharge of 225,000 ft³/s for the 1921 peak discharge. G.L. Bodhaine (Area engineer, U.S. Geological Survey, written commun., 1954) agreed with the analysis of F.J. Flynn and M.A. Benson, and reasoned that the other historical peak discharges also should be lower than the published values. Based on extension of the stage-discharge rating for the gage that was constrained to pass through the 1921 revised discharge point, he lowered peak-discharge values for all other historical peak discharges except for the 1815 flood, which remained the same. However, he noted that because none of the revisions were more than the 10 percent guideline used to decide if revisions are needed, the published values should not be changed.

The validity of the original slope-area measurement of the 1921 flood was recently re-evaluated by recalculating the peak discharge using an n value based on an n-verification analysis done following the flood on October 21, 2003. The peak discharge associated with the 2003 flood was determined to be 166,000 ft³/s from the current stage-discharge rating (Mastin and Kresch, 2005). However, a considerable amount of variation in the elevation of surveyed HWMs for the flood on October 21, 2003, precluded the determination of a single definitive water-surface slope for use in the n-verification analysis. The variability in surveyed HWMs for the flood on October 21, 2003, was attributed to the length of time between the event and the actual field survey, which was not completed until 9 months later in July–August 2004. Consequently, a range of water-surface slopes produced a range of possible n values (Mastin and Kresch, 2005).

A peak discharge of 145,000 ft³/s on November 6-7, 2006, determined from the current stage-discharge rating provided an opportunity to survey a peak water-surface slope soon after the flood using high-water marks. Access to the left bank of the study reach had to be made by boat, which was unavailable at the time, so only the HWMs on the right bank of the slope-area reach were surveyed, but the survey provided an accurate water-surface slope between cross sections 2 and 3 (James Stewart and Mastin and Kresch [2005] both used the same cross-section locations and identification; cross section 2 is upstream of the most downstream cross section, cross section 3). The field survey of the 2003 flood included HWMs on both sides of the river and detailed cross-sectional data; therefore, the water slope surveyed in 2006 was used in combination with the channel-geometry information for the 2003 flood to compute an n value for the 2003 flood. Using this recalculated n value, the discharge for the 1921 peak discharge was recalculated and re-evaluated. It was thought at the time of the 2006 survey that this data could be used to improve on the previous n-verification studies and provide a better verification of the 1921 peak discharge calculation.

Back to Table of Contents