Laserfiche WebLink
<br />O!' 'If.! ~.' 1" <br />t) t,..... \; .:'. <br /> <br />23. <br /> <br /> <br />ord on which they are based oShould be long enough to encompass a represen- <br />tative sample of floods, droughts and long-term trends in flow. (Z) All curves <br />in a region should be based on the same period of time to permit unb'lased <br />comparisons of stream characteristics and the development of synthesized <br />curves at ungaged sites. (3) Tl'ie;curves should reflect the integrated expe- <br />rience in a :region .rather .'than: the m-a-re'Hmifed .exp'eriences: 'at 'individual <br />sites. <br /> <br />Selection .bLBase Period <br /> <br />Streamflow records available in Kansas are much too short to furnish an <br />adequate sample of past experience for fully reliable forecasts of future flows. <br />But to provide inforrna tion now when it is needed, it was necessary to select <br />from the available records a period as long and as representative as possi- <br />ble as a base for analysis. Table 1 shows that coverage definitely was in- <br />adequate prior to 1920. Since then, nearly complete records have been col- <br />lected at 21 gaging stations, fairly well distributed throughout the State, that <br />are adequate for the establishment of a base period extending from October 1, <br />1920 to Sep,tember 30, 1956. or 36 water years, 1921 to 1956. This 36-year <br />period experienced two extensive droughts in the 1930's and 1950's as com- <br />pared witha major drought in the 1860's and lesserones in the 1870's, 1890's <br />and 1910's. Flood-frequency analyses show that in the period 1921-56 in <br />Kansas the rare floods have occurred more frequently than in the period <br />1895 to 1920 but that ordinary floods are equally frequent during the two pe- <br />riods. Eight Weather Bureau stations in Kansas have been maintained since <br />1878 and their records show that rainfall has averaged only 2 percent less <br />in the period 1921-56 than in the period 1878 to 1920. Thus, although record- <br />ed past experiences are not fully definitive, they do imply that the period <br />1921-56 is a representative base period. The following section describes <br />the methods used for correlating and extending to this base period the flow <br />duration relations based on available records collected at 122 stations. <br /> <br />Pivot Station Method <br /> <br />Evidence available to date indicates that floods and droughts are somewhat <br />random in their occurrence. An individual event will strike one stream ba- <br />sin more intensely than another. However, in Kansas, topographic and cli- <br />matic factors are sufficiently uniform so that over an extremely long period <br />of time each stream should eventually receive its equal quota of floods, low <br />flows and droughts. Thus an extremely long streamflow record, uninfluenced <br />by man-made changes, should define a flow-duration curve nearly in accord- <br />ance with the characteristic flow pattern of that stream. But even the long- <br />est records in Kansas are too short to define reliably the average experience <br />of the respective stream. The records cannot be extended in;time but they <br />can be extended in areal representation by considering the combined experi- <br />ence of several streams to be a better measure of the average condition than <br />that afforded by the individual experiences at anyone of them. In short-term <br />