FLOOD10363 (2) - Laserfiche WebLink

Home Browse Search

I I I I I I I I I I I I I I I I I I I Tomichi Creek has a larger percent of watershed in high mountain valley and as a result, eight out of sixty-three years of peak values were caused by rain on snow event flows. For Tomichi Creek, two analyses were performed: one with a data set including the rain on snow peaks, and a second with rain on snow peaks removed using only the snowmelt peaks for the corresponding years. Peaks that occur as a result of rain on snow runoff were small - on the order of the 99 - 84 percent chance exceedence flow (approximately 1 - 6 year return period) - in comparison to the snowmelt runoff peak values for the other years. The rainfall on snow events peaks were replaced with the next largest non rain on snow event in the record. In addition, the 3-, and 7-day runoff were also replaced since the magnitude of the rain on snowmelt event caused the 3-day and 7-day averages to be higher than the replacement peak and 1-day value for some events. It was not necessary to replace any values to eliminate rain on snow events for runoff periods longer than a 7-day. This method was continued throughout the process for consistency. The removal of the rain on snow events had a notable effect on the resulting skew values. The altered data set was more negatively skewed than that of the corresponding raw data set. According to the guidelines of Bulletin 178, page 16, "If the flood events that are believed to comprise two or more populations cannot be identified and separated by an objective and hydrologically meaningful criterion, the record shall be treated as coming from one population." Based on the above guideline and the ability to identify rain on snow events that show a notable impact on the resulting statistics, the use of the 0.5-inch rainfall threshold was used throughout the remainder of the study as the "objective and meaningful hydrologic criterion" to identify events influenced by rainfall. 4.2 REACHES WITH INSUFFICIENT STREAMGAGE DATA For all but Reaches 5 and 8, the data from the stream gage records are inadequate for completing the flood frequency analyses. For these six reaches regression and correlation analyses are completed using the REGFREQ program to expand the available records by correlating them with longer records available in the region. The REGFREQ program was chosen for its ability to determine the correlation of overlapping data between two gages. The resulting correlation coefficients were used to determine the best gage to supplement the record. The REGFREQ program produces a reconstituted data set with missing data points estimated based on the correlation between the gages and computes the associated frequency statistics (mean, standard deviation, and skew). The reader is referred to the REGFREQ Users Manual (USACE, 1972) for more information on the methodology used in the program. The REGFREQ program was used to determine the reconstituted gage data frequency statistics (mean, standard deviation, and skew) for the annual peak, 1-,3-, 5-, 7-, 15-, 30- and 60-day family of curves. The family of curves and associated gage data (median plotting positions) were plotted on a log- probability graph. In order to provide a consistent family of curves for each reach, the standard deviation and skew values were manually smoothed. The smoothing prevents the crossing of any of the curves and provides a consistent set of curves. In addition, smoothing was done to improve the fit of the curves to the raw gage data. This was needed since no correlation was so close to 100 percent that the estimated data provided a perfect fit with the original data. The resulting smooth statistics are provided in Chapter 5. 9