FLOOD07643 - Laserfiche WebLink

Home Browse Search

points more judiciousiy, and using visual smoothing resulted in changes (If up to 6.0 percent in the computed value fmm using only fourteen points and t(llerating some slight visual bumps in the data. With the 10-year event at the first floor there were no appreciable differences in the computed values. It was also found that inconsistencies could occur in computed values of expected annual damage because of the lack of symmetry in the family of elevation-frequency relationships used (Figure A-1 curves extrapolated to the annual event). The smoothing and using additional points, described in the second paragraph of this section, eliminated most of the inconsistencies but not all. The smoothing and additional points caused the computed values of total damage to increase from a flood hazard factor of 2.0 to 12.0 feet but decrease slightly for a flood hazard factor of 14.0 feet. It was noted by visual observation that for a flood hazard factor of 14.0 feet (Figure A-1) that the curve was slightly asymmetric near its lower end. By adjusting this curve slightly (0.8 feet at the annual event) the curve was made symmetric with the other curves. This adjustment eliminated the remaining inconsistency. The eighteen points selected to represent the depth-damage relationship were selected at breaks in linearity of the function. Thus, between points, a linear relationship was assumed. These points remained the same for all analyses described in the preceeding paragraph. The integration routine used in the analysis is especially sophisticated and was selected to overcome as many of the integration problems which are normally encountered as is possible. In addition to the data input to represent the elevation-frequency relationship the integration routine internally generates three exceedance frequency values between each pair of input values. These internally generated values help to define the nonlinearity of the elevation- frequency function. The additional points are used throughout the integration process. A detailed description bf this rouiine may be found in Exhibit 2 of Reference 3. The family of elevation-frequency curves shown in Figure A-1 were developed by the Federal Insurance Administration and used in the insurance rate study of Reference 4. It appears they were originally drawn with a french curve at a relatively small scale. The tabulated data in the reference appear to have been points picked-off the drawn curves. Plotted to a much larger scale these data showed some minor irregularities (bumps and dips) which caused the inconsistencies in computed values described previously. These inconsistencies were most pronounced with frequent flooding at the first floor. The maximum difference is less than 6 percent. While this percentage is not significant for the conclusions and results of this study it does cause the ,graphical presentations to appear inconsistent. Sensitivity to FHF and Event at First Floor An analysis of expected annual damage and its sensitivity to FHF and event at first floor was made for four type structures and the results are presented in Figures A-5 throughA-8. The expected annual damage on the ordinate axis is total annual damage expected to occur to structure and contents expressed as a percentage of structure value. For example, a .10 percent value means the expected annual damage is estimated to be 10 percent of the structure value for the conditions assumed. These Figures show that the increase in expected annual damage with a larger FHF is relatively small beyond a FHF of about 8.0. That is, for a FHF greater than 8.0 the change in expected annual damage is relatively insensitive to change in FHF. This is true for all type structures and all events at the first floor especially those events greater than 5 years. For a A-7