Laserfiche WebLink
<br />v <br /> <br />For convenience, a statistical model may be displayed as a frequency <br />curve. The magnitude of the event is the ordinate. Probability of exceedance is the <br />abscissa. For hydrologic engineering studies, the abscissa commonly shows "percent <br />chance exceedance." This is exceedance probability multiplied by 100. <br /> <br />1.3. SUMMARY OF STREAMFLOW FREQUENCY ANALYSIS <br />TECHNIQUES. Techniques for selecting and calibrating streamflow frequency models <br />may be categorized as graphical or numerical. With graphical techniques, historical <br />observations are plotted on specialized graph paper and the curves are fitted by visual <br />inspection. Numerical techniques infer the characteristics of the model from statistics <br />of the historical observations. The procedures for both graphical and numerical <br />analysis are presented in detail in EM 1110-2-1415 and are summarized herein for <br />ready reference. <br /> <br />2. FREQUENCY ANALYSIS CONCEPTS. <br /> <br />2.1. DATA REOUIREMENTS. Statistical models of streamflow frequency <br />are established by analyzing a sample of the variable of interest. For example, to <br />establish a statistical model of annual peak discharge, the sample will be a series of <br />annual peaks observed throughout time. The procedures of statistical analysis require <br />the following of any time series used in frequency analysis: <br /> <br />2.1.1. The Data must be Homogeneous. That is, the data must <br />represent measurements of the same aspect of each event. For example, daily <br />discharge observations should not be combined with peak discharge observations. <br />Furthermore, all sample points must be drawn from the same parent population. For <br />example rain-flood data and snowmelt flood data should not be combined if they can <br />be identified and analyzed separately. Likewise, discharge data observed after <br />development upstream should not be combined with pre-development data. <br /> <br />2.1.2. The Data must be Spatially Consistent. All data should be <br />observed at the same location. Data observed at different locations may be used to <br />develop probability estimates. However, these data must be adjusted to represent <br />conditions at a common location. <br /> <br />2.1.3. The Time Series must be Continuous. Statistical analysis <br />procedures require an uninterrupted series. If observations are missing, the missing <br />values must be estimated, or techniques for analysis of broken records must be used. <br /> <br />2.2. PROBABILITY ESTIMATES FROM HISTORICAL DATA. Streamflow <br />probability is estimated from analysis of past occurrence. The simplest model of the <br />relationship of streamflow magnitude and probability is a relative frequency model. <br />This model estimates the probability of exceeding a specified magnitude as the <br /> <br />7-91 <br />