Laserfiche WebLink
<br />analysis. <br /> <br />The model error (co) is estimated by <br /> <br />Co = maximum [0; (SE)2 -c1 (E~L]' <br /> <br />where SE is the standard error of the <br />estimate from the OLS equation; <br /> <br />Cj is a constant; and <br /> <br />ERL is the mean effective record <br />length, in years, for all gaging <br />stations used in each respective <br />regression-model data set. <br /> <br />The time-sampling error (ti) is estimated by <br /> <br />'i = Cl(ER~(i)). <br /> <br />where ERL(i) is the effective record length, in <br />years, for the ith gaging station <br />used in each respective <br />regression-model data set. <br /> <br />I <br /> <br />~ <br /> <br />The constant, Cj, is related to the recurrence <br />interval of the response variable and to the <br />weighted skew coefficient (g) of the observed <br />annual-peak discharges. It is determined by <br /> <br />~ ( -2 )~ <br />. -2 k 3 -2 - <br />c1 = maxImum 0;8 1+2(1+4g) +kg , <br /> <br />where s <br /> <br />is the mean standard deviation of <br />the logarithms (base 10) of the <br />observed annual-peak dis- <br />charges; <br /> <br />Ii <br /> <br />is the mean standardized <br />Pearson Type-III deviate for <br />selected T -year recurrence <br />interval and mean weighted <br />skew coefficient g (IACWD, 1982, <br />p. 3-1 - 3-27); and <br /> <br />g <br /> <br />is the mean weighted skew <br />coefficient of the logarithms (base <br />10) of the observed annual-peak <br />discharges (IACWD, 1982, p. <br />12-15). <br /> <br />The values <br />determined <br /> <br />sand jj are statewide estimates <br />by the averages of the 188 <br /> <br />(5) <br /> <br />streamflow-gaging stations analyzed using <br />either the drainage-basin or channel-geometry <br />flood-estimation techniques. These estimation <br />methods are based on the assumption that sand <br />jj are approximately constant for all the gaging <br />stations in the State. <br /> <br />The effective record length (ERL) of a gaging <br />station is based on an empirical analysis made <br />by Gary D. Tasker (U.S. Geological Survey, <br />written commun., March 1992) of results <br />reported in Tasker and Thomas (1978) and <br />Stedinger and Cohn (1986). It is determined by <br /> <br />ERL = LS+ (HST-LS)a, <br /> <br />(8) <br /> <br />where LS <br /> <br />is the systematic record length of <br />a gaging station, in years, the <br />number of water years during <br />which the gaging station was <br />operated; <br /> <br />(6) <br /> <br />HST is the historic record length of a <br />gaging station, in years, as used <br />in a Pearson Type-III historical <br />flood-frequency analysis; if a <br />systematic flood-frequency <br />analysis was performed, HST = <br />LS; if (HST-LS) > 200, set <br />(HST- LS) = 200; and <br /> <br />(7) <br /> <br />a = 0.55 - O.lllog ( ph )] . <br />L ' (I-ph) <br /> <br />(9) <br /> <br />In the last equation, ph = 1.0 - (np / HSTJ, and <br />np is the number of historic and extremely large <br />discharge (high-outlier) peaks. <br /> <br />The ERL used in the weighted least-squares <br />regression analyses for each gaging station is <br />listed in table 8 (at end of this report). <br /> <br />ESTIMATING DESIGN-FLOOD <br />DISCHARGES USING <br />DRAINAGE-BASIN <br />CHARACTERISTICS <br /> <br />The drainage-basin flood-estimation method <br />uses selected drainage-basin characteristics to <br />estimate the magnitude and frequency of floods <br />for stream sites in Iowa. Multiple-regression <br />equations were developed by rel8ting significant <br />drainage-basin characteristics to Pearson <br />Type-III, design-flood discharges for 164 <br /> <br />ESTIMATING DESIGN-FLOOD DISCHARGES USING DRAINAGE-BASIN CHARACTERISTICS 9 <br />