Laserfiche WebLink
<br />Flood-Frequency Equations <br /> <br />Initially, 13 independent variables were included in the regression <br />computations for each T-year flood. Results of the analyses showed that <br />three of the independent variables, drainage area, coefficient of imper- <br />viousness, and the length-slope factor (table 2), would satisfactorily <br />explain more than 98 percent of the variation in the dependent variable. <br />Benson (1964) showed that after inclusion of three or four independent <br />variables, additional variables usually did not appreciably decrease the <br />regression standard error of estimate. <br /> <br />Drainage area (A), which is the most important variable, accounts <br />for at least 90 percent of the flood-discharge variation and is statis- <br />tically significant at the 5 percent level for all seven T-year floods. <br />The coefficient of imperviousness (K), which is an index to urbanization, <br />and the length-root slope ratio (L/I8), which is an index to lag time, <br />account for another 8 percent of the variation. The length-slope factor <br />is also useful in determining the effect of future straightening of the <br />main channel. <br /> <br />The regional flood-frequency equations, the regression standard <br />errors of estimate, and a minimum combined error are summarized in table <br />10. These equations were derived using simulated peak discharges based <br />on the same long-term rainfall record for all sites. Therefore, the <br />standard errors of estimate associated with the equations are smaller <br />than the total errors. An estimate of the minimum combined error (table <br />10) is obtained by combining the regression standard error of estimate <br />with the standard error of the rainfall-runoff calibration. <br /> <br />A flood-frequency curve for an ungaged site can be computed if the <br />variables are determined by the methods used in this report and are not <br />outside the observed range. The best results will be obtained for T- <br />year floods between 1.25 and 50 years for drainage areas between 4 and <br />15 square miles (12.9 and 38.8 square kilometers), L/ ;S:values between <br />0.7 and 1.5, and K values between 1.1 and 1.6. <br /> <br />Flood-Frequency Nomograph <br /> <br />To aid in application of the flood-frequency equations, a nomograph <br />(fig. 8) was developed by using simplified equations for the four highest <br />T-year discharges (Q,0, Q2S, Qso, Q,00). The simplification involved de- <br />leting L/ l'ifrom the equations given in table 10, leaving A, K, and the <br />regression constant C. To prepare the nomograph, it was assumed that the <br />variation of the exponents of the variables was insignificant above the <br />10-year flood discharge so that average exponents could be used for A and <br /> <br />-42- <br />