FLOOD10102 - Laserfiche WebLink

Home Browse Search

The correlation coefficient, root mean square error, and slope of the regress- ion line between recorded and synthetic peak discharges for the calibration of each site, all of which are measures of the relative success of the calibration, also are given in table 6. The relation between recorded and synthetic peak discharge for all 141 floods used in the calibration process is shown in figure 5. This relation has a correlation coefficient of 0.90 and the slope of the regression I ine is 0.93. These statistics indicate the overall model-cal ibration phase was successful and unbiased. However, stations 07126450 and 07129100 had statistically unsuccessful calibrations and were not used in the synthesis phase of the model- ing process. Station 07126450 had calibration statistics indicating a low correla- tion coefficient (O.56), a high root mean square error (J100 percent), and the slope of the regression line was considerably less than 1.00 (0.47). Based on only three storms, the slope of the regression line for station 07129100 was con- siderably greater than 1.00 (3.20). . Synthesis of Long-Term Annual Flood Series Recorded rainfall-runoff data were used to "calibrate" the infiltration, soil- moisture, and flow-routing parameters in the model for the 15 basins. Long-term daily rainfall, daily evaporation, and unit (5-min time interval) storm rainfall were then used as input to the calibrated rainfall-runoff model of each basin to gener- ate synthetic peak discharges and flood volumes for each historical storm period. The following National Weather Service climatological stations provided the climatic data required by the synthesis phase of the modeling process: Denver, Colo. {station 2220)------------------------rainfal I, 1898-1970 John Martin Dam, Colo. (station 4388)------------evaporation, 1950-1975 Pueblo, Colo. (station 6740)------------------------rainfal I, 1900-1969 Pueblo City Reservoir, Golo.(station 6743)-------evaporation, 1950-1975 Amari 110, Tex. (station 23047)----------------------rainfall, 1914-1974 Wichita Falls, Tex. {station 13966)--------------evaporation, 1950-1974 Rainfall data consisted of both daily rainfall for the indicated period and unit storm rainfall for the three to five greatest storms occurring in each of these years. It was assumed that one of these three to five storms would produce the annual peak discharge, although this peak discharge would not necessarily be associated with the greatest total rainfall. Evaporation data consisted of seasonal daily pan evaporation for the indicated period. To be compatible with the long- term rainfall data, daily pan evaporation for the period prior to 1950 was synthe- sized based on 3-day moving averages calculated from the observed evaporation data for the 1950 through 1975 period. Denver and Pueblo long-term rainfall was used along with either John Martin Dam or Pueblo City Reservoir evaporation, while Amarillo long-term rainfall was used along with Wichita Falls evaporation. , 14