(i.e., a flood level with a recurrence interval of 100 years) will not 

 occur in any period of 100 years. Therefore, a 100-year flood level 

 predicted from a 100-year period of record may be too low. Also, there 

 is a 9.5-percent probability that a 1,000-year flood level will occur 

 at least once in any period of 100 years. Therefore, the predicted 100- 

 year flood level, based on a 100-year period of record, may be too high. 



Confidence limits for the predicted flood levels can be obtained 

 using methods similar to those used for river flood levels. However, 

 rivers have a seasonal variation in flow, so a 1-year time increment is 

 significant in that case. In the case of tsunamis, the 1-year time 

 increment is a convenient means of measuring time, but there is no par- 

 ticular relationship between this time increment and the generation of 

 tsunamis. Methods used for obtaining confidence limits for tsunami 

 flood levels should give the same results, regardless of the chosen 

 time increment. 



Beard (1962) notes that there is a 5-percent probability that the 

 magnitude of the difference between the real flood level and the pre- 

 dicted flood level will be greater than or equal to twice the standard 

 error. Assuming there is an equal chance of the real flood level being 

 either greater than or less than the predicted value gives +2.5- and 

 -2.5-percent confidence limits. 



Where no historical data are available, data may be constructed 

 entirely from a computer model by assigning magnitudes to various tsu- 

 namis in the mathematical model, and by determining the probability of 

 generation for each tsunami magnitude. However, the results will not 

 have the same degree of accuracy. 



An exact relationship between tsunami magnitude and earthquake 

 magnitude has not been determined. Iida (1961) proposed that tsunamis 

 could be assigned a magnitude based on their energy (the energy of the 

 generated waves), with an increase in magnitude of 0.5 being equal to a 

 doubling of the energy. He also related the tsunami magnitude to the 

 maximum runup height in meters at the shoreline area experiencing the 

 strongest tsunami action (Iida, 1970) . The relationship between the 

 runup height fy^^ and the tsunami magnitude m is shown in Figure 2. 

 The dashlines show the range of the expected maximum runup, based on 

 Iida's data, due to differences in the characteristics of the individual 

 tsunamis and coastal areas . 



Soloviev (1970) revised the definition of tsunami magnitude by relat- 

 ing it to the average runup height R (in meters) at the shoreline area 

 experiencing the strongest tsunami action. This tends to average out 

 any high runup heights related to a particular coastal feature, and should 

 be more representative of the actual tsunami energy. Soloviev does not 

 indicate the length of coastline to be used in the average, but does pro- 

 vide an equation for the magnitude as 



m = log 2 (/?R) (3) 



22 



