FLOOD09736 - Laserfiche WebLink

Home Browse Search

in western Colorado, is shown in Figure 4, The combined curve is seen to give higher values than either of the separated curves. A discharge-frequency curve based on a mixed population is also shown for comparison. It can be seen from this example that the curve based on mixed events significantly underestimates the flood discharge for events which occur at intervals less frequent than about 20 years. All curves for the example station exhibit some degree of positive skew. . UNGAUGED SITES One of the most reliable methods for determining flood flows at un gauged sites, or sites where the gauge record is relatively short, is to derive a series of regional regression equations which relate flood flows to measurable basin and climatic parameters. The definition of a homogeneous region is a critical element in the pro- cedure. It is suggested from previous work that study regions should not cross major river basin boundaries. Each subbasin selected for possible inclusion in the study region must satisfy certain statistical tests until a region is defined. Each basin or climatic parameter such as drainage area, basin slope, channel slope, mean watershed elevation, and annual precipitation should be tested for their significance before inclusion in the regression equations. Regression equations for the statistical parameters as functions of the basin and climate variables are obtained from analyzing individual gauging station records. The three statistical parameters which usually define the log-Pearson type III distribution include the mean, standard deviation, and skew of the logarithms of the annual peak flows, The coefficient of variation, defined as the ratio of the standard deviation to the mean, has been found to be a more useful statistical parameter than the standard deviation alone in developing the regional regression equations. The degree of accuracy of the statistical parameter estimates increases with the number of observations at each individual station, assuming all other factors are equal. Statistical parameters for each individual gauging station used to derive the regression equations are therefore assigned weights equal to their length of record, A brief outline of the steps recommended in regional analysis is pre- sented below, Details of the procedure are included in the example which follows the outline, - On the basis of engineering judgement, select a trial region con- sisting of small basins having similar watershed characteristics and hydrometeorlogic conditions, Inventory climatic or basin variables in all gauged basins in the trial region including: (a) period of record (b) drainage area (c) mean watershed evaluation (d) other measurable variables 9