Laserfiche WebLink
<br />Neverlheless, Ihe results for Neff's Canyon are com- <br />parable wilh those obtained in an independent anal- <br />ysis made by the Corps of Engineers (1969b). The <br />Corps study predicls Ihe peak flow at the moulh of <br />Neff's Canyon for a "IOO.year" storm to be 1500 <br />cfs. This model predicls 1490 cfs, al an antecedent <br />soil moislure level of II inches. <br /> <br />Hourly-time-increment model. <br /> <br />It was also necessary to calibrate the hourly mod- <br />el on Mill Creek and 10 transfer the resulting parameler <br />values, with appropriale adjuslment, to the drainage <br />areas of Neff's Canyon. The lower Mill Creek sub. <br />watershed was selected for this calibration process. <br /> <br />Some parameter values for the hourly.time- <br />increment model \VOre taken to be Ihe same as Ihose <br />found for Ihe daily lime incremenl model. For <br />example, soil field capacity remains unchanged for <br />both time increments. Other parameters, such as <br />the interflow time delay constant, are a function of <br />the model time increment. Such parameter values <br />were selected from Ihe results of Ihe simulation of <br />the lower Mill Creek subwatershed on an hourly <br />basis. The same calibralion procedure was used for <br />both the daily and hourly models. <br /> <br />The size of the Mill Creek watershed is not <br />suilable for simulalion on an hourly basis. Given <br />the existing data network for hourly precipitation, <br />it is difficult to say what proportion of Ihe water- <br />shed is covered by the stonn, and is, therefore, con. <br />tributing 10 Ihe gaged watershed outflow. <br /> <br />In transferring parameter values from Mill <br />Creek to Neff's Canyon, and later 10 Ihe urban. <br />izing area of Olympus Cove, the effects of differ. <br />ences in areas were taken into account, where nec~ <br />essary. Parameter values, such as the time delay <br />constant in the surface water routing equation were <br />adjusted for the decrease in size of the watershed. <br />The fmalized parameter values used for the hourly <br />time increment simulation of the rural portions of <br />the study area (Figure 1.1) are shown in Table <br />3.2. <br /> <br />Model Testing for tbe Rwal Areas <br /> <br />While model calibratioa is a fitting process to <br />estimal.e values for Ihe model parameters, model <br />testing involves using a second and independent sel <br />of data from the same hydrologic unit in order to <br />determine the level of agreement between the ob- <br />served and computed oulput functions. Thus, model <br />testing is simply an independent tesl of results <br />achieved under the calibration phase. <br /> <br />Table 3.2 Parameter values used in the hourly time <br /> Increment model of the ruml portions of <br /> the study area. <br />Parameter Description Value; <br />SS Sahuated soil level (inches) 13.0 <br />SFC Field capacity of soil (inches) 6.0 <br />FO Maximum infiltration capacity rate 1.0 <br /> (inches/hr) <br />FC Minimum inmtration capacity rate .20 <br /> (inches/hr ) <br />DKT Decay constant in infLItration equation 2.0 <br /> (hIhr) <br />TAUSW Decay constant in surface water .50 <br /> routing (hrhr) equation <br />TGW Decay constant in interflow routing .01 <br /> (hrm) equation <br />QK The fraction of outflow from soil .30 <br /> moisture storage that becomes <br /> interflow <br />SI Upper limit of interception and de- .20 <br /> pression storage qnches) <br /> <br />Daily time incremen t model. <br /> <br />Following calibration for the two sub watersheds <br />of Mill Creek, the parameter values given by Table <br />3.1 were used 10 simulate five years of record at each <br />of Ihe two stations. Sample comparisons with ob- <br />served hydrographs are shown by Figures 3.6 and 3.7 <br />for the upper and lower sub watersheds, respectively. <br />AI the bottom of Table 3.1, the mean value (over the <br />five years of simulation) of the objective funclion is <br />given for the two Mill Creek subwalersheds. The table <br />also includes Ihe ralio of the mean yearly objective <br />function to mean yearly streamflow, a quantity term- <br />ed the relative objective funclion. <br /> <br />While without data it was nol possible to test <br />the model for Neff's Canyon, the runoff predic- <br />tions were comparable wilh those of the lower Mill <br />Creek and are shown by Figure 3.8. Runoff is <br />computed al the points of discharge indicated on <br />Figure 1.1. <br /> <br />Hourly time increment model. <br /> <br />This model was tested by generating runoff hy- <br />drographs associated with several short duration, high <br />inlensity rainfall events on the lower Mill Creek sub. <br />walershed. The compuled and observed hydrographs <br />for two of these evenls are shown by Figure 3.9. <br /> <br />Model Calibration for the <br />Urban Watersheds <br /> <br />As previously indicated, Ihe time increment <br />adopted for the urban model is 30.minules. As for <br />the rural models, the precipitation dala available for <br />calibrating the urban model are daily total from non- <br />recording gages and d~ta from a recording gage (Fig- <br /> <br />32 <br />