H13C-1545
Use of Streamflow Data in Limiting Rainfall-Runoff Model Parameters

Monday, 14 December 2015
Poster Hall (Moscone South)
Wen Chiang Wang and David R Dawdy, Multech Engineering Consultants, Inc., San Jose, CA, United States
Abstract:
Watershed rainfall-runoff models must be calibrated to produce realistic storm runoff hydrographs. In optimizing such models, one should optimize both to volumes and peaks. Otherwise, parameters can work against each other to curve-fit the peak or volume alone. Many rainfall-runoff models used in practice have a time of concentration as a parameter. The widely used Clark unit hydrograph is such a model. The time of concentration in that and other similar models can be determined from streamflow records. In order to check on the time of concentration, you should look at the streamflow recession when rainfall is no longer recorded. Thus, no runoff potential is present and only the time of concentration is effective. Analysis of the streamflow records can put strict bounds on the estimate of the time of concentration and precludes 'optimizing' to unrealistic values. An example from Coyote Creek in California is used to illustrate the principle.