FLOOD CHARACTERISTICS OF URBAN WATERSHEDS IN THE UNITED STATES by M.A. Benson Prepared by the U.S. Geological Survey 1952 TABLE OF CONTENTS Introduction Base data Mean annual flood Short periods of record     Periods used     Drawing of frequency curves     Arithmetic vs. graphical mean annual flood     Reliability of mean annual flood values Composite flood-frequency curves     General     Periods combined     Results from composite graphs     Effect of number of stations combined     Median vs. average ratios     Theoretical distribution of composite curves Conclusions References FIGURES 1. Theoretical frequency curve 2. Frequency curves for 10-year periods 3. Frequency curves for 25-year periods 4. Frequency curves for 50-year periods 5. Frequency curves for 100-year periods 6. Range in value of arithmetic and graphical mean annual flood with length of record 7. Range in value of graphical mean annual flood with length of record, for 80, 95, and 100 percent of time 8. Recurrence intervals of 2.33-year flood 9. Computing range for mean annual flood 10. Range in value of 10-year flood with length of record for 80, 95, and 100 percent of time 11. Range in value of 25-year flood with length of record for 80, 95, and 100 percent of time 12. Range in value of 50-year flood with length of record for 80, 95, and 100 percent of time 13. Range in value of 100-year flood with length of record for 80, 95, and 100 percent of time 14. Recurrence intervals of 10-year flood 15. Composite frequency curves based on ten 10-year periods 16. Composite frequency curves based on ten 25-year periods 17. Composite frequency curves based on ten 50-year periods 18. Range in value of 10-year flood ratio with length of record 19. Range in value of 25-year flood ratio with length of record 20. Range in value of 50-year flood ratio with length of record 21. Variation of flood ratios with number of records combined - Trial No. 1 22. Variation of flood ratios with number of records combined - Trial No. 2 23. Theoretical distribution of composite curves TABLES 1. Flood-frequency data, 1,000-year record 2. Comparison of true values with averages of computed values of mean-annual flood ABSTRACT This report contains a summary of the procedures followed and the results obtained in analyzing a "perfect 1,000-year record" of maximum annual peak floods. The investigation was started with the object of studying the possible variations in frequency curves computed from short periods of record, each taken from a long-term record whose frequency characteristics were exactly determinable. There is no "theoretically correct" frequency distribution of flood events, although claims have been made for one or another system either based on empirical evidence or (with necessary assumptions) on statistical theory. For this reason, we are not sure of the "true" shapes of the frequency curves of even the longest flood records we have. It was, therefore, decided to start from an assumed list of 1,000 peak floods, defining exactly a simple frequency curve. Separate portions of the array could be analyzed independently, and the results compared with the known characteristics of the base curve. This could demonstrate the variations, due to chance alone, in frequency curves from short records. LizardTech's Djvu plug-in is needed to view these reports. CLICK HERE TO VIEW THE ENTIRE REPORT For questions or comments, contact K. Van Wilson.