FLOOD00918 - Laserfiche WebLink

Home Browse Search

I I I I I I I I I I I I I I I I I I I produce significantly different results, as compared to NOAA Atlas. The results of this study show that the 1997 values has a pronounced impact on the results. When using either the GEV regional analysis or the Log Pearson III "at-site" analysis, the effects of the longer period of record and the July 1997 storm have a considerable effect on the results, as compared to currently used design values. It was the recommendation of technical advisory committee that all of the available data be used with an GEV distribution based on the "at-site" data set rather than the regional analysis. The results obtained by CSU using the GEV distribution and L Moments compared very closely to the results obtained using Log Pearson III for the "at-site" analysis. D. CONCLUSIONS: As was explained in the results, it was felt by many members of the committee that the City of Fort Collins appears to be dissimilar from the other gage sites in the region and most likely experiences "preferred storm tracks", producing more frequent extreme rainfall events in comparison to nearby areas. In the regional analyses, 7 of the biggest regional events occur above the 20-year return period. In all cases, either 3 or 4 out of the 7 events are from the Fort Collins gage (CSU, 1998). This observation supports the use of an "at-site" analysis using the data set available through 1997 for the City of Fort Collins. The number of extreme events will have a significant impact on the results obtained from the analyses since the majority of the data points are lower values. This posses somewhat of a problem with plotting position. The current data set for Fort Collins has relatively few recorded large values that influence the fit to the data for the upper region of the frequency curve. Therefore, the 1997 storm will have more of an impact on the analysis since it is the most extreme value recorded and there are few data points in that range of magnitude. As more data is collected, the plotting positions for the 1997 event will likely be shifted to the right. As a result, the predicted values for the less frequent return periods, based on this analysis. may be slightly conservative and may decrease as more data is added to the analyses. It was the consensus of technical advisory committee that the inclusion of the 1997 value was appropriate, even if it may be a high outlier, because of the importance of such a large magnitude event to the record. -9-