FLOOD09736 - Laserfiche WebLink

Home Browse Search

...i"1.~1.""'''C~-~-.--~ -~~-~_.~....-....-. ." ../ , I ~F.. ,.-;:- '. ESTIMATING FLOOD FLOWS FROM MIXED POPULATIONS IN COLORADO By William P. Stanton,l and Wen-Sheng Liang2 ABSTRACT Floods in Colorado may be caused by intense thunderstorms, broad general rains, melting snow, rain on melting snow, ice jams, dam failures, or some combination of these, When flood frequency curves are derived from the annual series of peak recorded flows for each year regardless of the cause of flooding, the data are said to come from a "mixed population," A flood flow frequency analysis based on a mixed population of snowmelt and rainfall peak discharges, which comprise two distinct and generally inde- pendent populations, will tend to underestimate flood magnitudes for rare events in Colorado, Flood frequency analysis using stream gauge data will be improved if a separate annual series from rain and snowmelt events can be identified from the record. The frequency-discharge curves may then be sta- tistically combined to give a composite frequency-discharge curve for the gauge site, For an ungauged site, several gauged subbasins which comprise a homoge- neous region may be analyzed individually to determine statistical parameters from each event at each station, By regression analysis, the statistical parameters may be related to measurable subbasin and climatic variables. The statistical parameters should be given weights equal to the station's length of record in the regression analysis, As a practical example, the mean logarithm, coefficient of variation of logarithms, and skew coefficient of logarithms of peak flows were used to define a log-Pearson type III distribution for 22 stream gauging stations in the Upper San Juan River basin in Colorado. Drainage basin area and mean watershed elevation were found to have an influence on the frequency curves. If the basin or climatic variables in a given site are known, the regional statistical parameters can be determined from the regression equations, and discharge-frequency curves for. each independent population can be computed, After these curves for the .independent events are statistically combined, the results will be compatible. with other records in the region. The discharge-frequency curve may be adjusted to reflect long-term records in the vicinity of the site if present, ., '. . INTRODUCTION Flood flow estimates have long played an important role in land use planning, Knowledge of peak flood discharges at various frequencies is an essential element in assessing flood hazards. Both structural and non- structural approaches to reducing flood damages must be based on sound hydrologic principles. Special conditions unique' to, the West and the Rocky Mountain Region become apparent when estimating flood flows in Colorado. 1 Supervising Water Resources Engineer, Colorado Water Conservation Board 2 Associate Principal Engineer, Simons, Li & Associates, Inc.