Laserfiche WebLink
<br />In the computer program the sequence of computations is the same <br />as that described above except that three intermediate points are used <br />between each pair of input frequency values. For example, if the <br />frequency input data values were those shown in Table 1, the computer <br />would interpolate between .1 and .2 and compute three intermediate <br />points, .125, .150 and .175. Each of these points would also be used <br />to compute damage values. A "spline-fit" routine uses the input and <br />intermediate data to compute additional points prior to integration. <br />The use of intermediate points allows for a more accurate determination <br />of the damage-frequency relationships. This point is illustrated in <br />Figure 4 which shows both the input data and the intermediate points <br />computed by the computer. <br /> <br />SINGLE EVENT FLOOD DAMAGE <br /> <br />Using Figure 4,f100d damage can be estimated for any flood event <br />of known frequency simply by picking the damage off the curve at known <br />frequency values. For example, the 100-year event (1.0 exceedance <br />frequency in events per 100 years) would cause an estimated $2.39 <br />million damage with 1980 conditions. The 50-year event (2.0 exceedance <br />frequency) would cause an estimated $1.5 million damage. The computer <br />program can compute damage for any single event uSing stage or flow as <br />its input values instead of frequency. <br /> <br />EXPECTED ANNUAL FLOOD DAMAGE <br /> <br />Expected annual damage is the damage which can be expected to occur <br />in anyone year assuming conditions remain unchanged. It is computed by <br />weighting each damage value in Figure 4 according to its probability of <br />occurrence. Graphically this amounts to finding the area beneath the <br />damage-frequency curve over the entire range of damaging events. It <br />needs to be emphasized that the correct computation of expected damage <br />includes the full range of probabilities from initial threshold to zero. <br />Any truncation, that is, not going to 0 or to the threshold will result <br />in an error which can be significant depending upon the shape of the <br />function. The hand integration or weighting of the damage values is <br />commonly performed by planimeter or rectangular area computation. The <br />latter approach is illustrated in Table 2. <br /> <br />EQUIVALENT ANNUAL FLOOD DAMAGE <br /> <br />The expected annual damage computed from Figure 4 is for 1980 <br />'conditions. That is, the damage, stage, flow and frequency data <br />are characteristic of the flood plain and flood hazard in 1980. <br />If these conditions were to remain the same in the future, then the <br />expected annual damage value would be the damage which could be expected <br />to occur during anyone year. However, it is common for conditions to <br />change--damageab1e property in the flood plain may increase or decrease, <br />urbanization upstream may cause increased runoff, or the channel itself may <br /> <br />EXHIBIT 3 <br />3 of 12 <br /> <br />~I <br />