Laserfiche WebLink
<br />tiIl~.!U'_OPLjj,.i\ T I ON <br />Flood frequency anal ysi s has been based on the assLlmpti on that an annLlal <br />flood series can be considered a sample from a single or homogenous <br /> <br />population. When the popLllation is not homogenous it has been recognized as a <br /> <br /> <br />mixed population problem. Both rainfall-runoff methods and methods based on <br /> <br /> <br />streamflow gaging records have made the assumption that the peak discharges <br /> <br /> <br />originate from the same homogenous population. <br /> <br /> <br />The mixed population problem can be addressed by analyzing the <br /> <br /> <br />population's source of moisture, a hurricane versus non-hurricane generated <br /> <br /> <br />runoff, floods occurring in a particular month, or rainfall versus snowmelt. <br /> <br /> <br />They can be handled separately or combined to create a compound probability <br /> <br /> <br />distribution using weighting coefficients. The two meteorologic causes <br /> <br /> <br />identified in this paper are rainfall and snowmelt. <br /> <br /> <br />Peak flows in the foothills region of Colorado are caused primarily by <br /> <br /> <br />one of two meteorologic sources -- snowmelt or rainfall. In Jenkins (1960), <br /> <br /> <br />it was intimated that rain storms rarely occur at the higher elevations with <br /> <br /> <br />the opposite being that the major floods at lower elevations are caused by <br /> <br /> <br />rainstorms. However, in many watersheds at higher elevations, the yearly peak <br /> <br /> <br />flow is produced from snowmelt. Previously no distinctions were made between <br /> <br /> <br />snowmelt- and rainfall-produc~d peak flows in the same year (Elliott and <br /> <br /> <br />others, 1982). <br /> <br />10 <br />