WSP03621 - Laserfiche WebLink

Home Browse Search

18 that region is considered homogeneous in some phases of flood-frequency char- acteristics. It follows that the average shape of the frequency graph defined by the gaged streams will also apply to ungaged streams in that region. Thus by combining records on the basis of geometric similarity of the frequency graph, regions for which the shape of the frequency graph is common to all streams may be identified. The shape of the frequency graph is dimensionless and as such cannot be used to determine flood magnitudes in absolute terms at specific points on streams. This procedure of combining records locates the frequency graph only with respect to the scale of recurrence intervals. To be of practical use, the graph must also be located with respect to the scale of magnitudes, that is, the dimensions must be restored. Dimensions are restored to the frequency relations by the second opera- tion in which the records are combined on the basis of similarity in their fre- quency graphs along the scale of magnitude, Experience has shown that the mean annual flood is indicative of the flood characteristics of a drainage basin and provides a good index on which to base comparisons. Hydrologically sim- ilar areas are defined by combining stations that have similar relations between their mean annual floods and the physical characteristics of their basins, The mean annual flood defines one point on the frequency graph with respect to both the scale of magnitude and of recurrence, If one point on the graph is thus defined, and the shape of the graph has been determined, the complete flood- frequency graph can be drawn. It should be pointed out that the geographical boundaries of the areas de- fined by the two procedures of combining records need not, and generally do not, coincide. To distinguish between the two classes of areas, those defined by the first operation are referred to in this report as "flood regions" and those defined by the second as "hydrologic areas". The method of regional analysis is not applicable to streams for which the floods are appreciably influenced by change in storage in manmade reservoi rs unless proper allowance is made for the storage effects. In this report, the an- nual floods thus affected have all been adjusted to natural conditions except those for the main stem of the Arkansas River upstream from the Pawnee River. Flood-frequency relations for this reach of the Arkansas River will be derived in a subsequent section entitled "Regulation". There are four general steps in the method of regional analysis: (1) ad- justing records to a common or base period; (2) testing the homogeneity of the records to be used; (3) combining homogeneous records to produce a composite frequency curve; and (4) correlating the mean annual flood with characteristics of the drainage basins.