Alaska Earthquake and Tsunami, II — Furumoto 
315 
at 0.0044 cps for R 3 /R 2 . These two peaks prob- 
ably coincide, and the apparent lag between the 
two is due to inadequate resolution of the Fou- 
rier analysis at these frequencies. There are 
opposing patterns at 0.0068 cps. 
The best-fitting curve of the directivity func- 
tion with the R 3 /R 2 spectrum is plotted on the 
upper section of Figure 3. In this curve the pa- 
rameters are: B = 800 km, V = 3-0 km/sec, 
and 0 = 15°. R 3 /R 4 fits the curve also, except 
for the mismatch in the neighborhood of 0.0067 
cps. 
The epicenter determined by the U. S. Coast 
and Geodetic Survey (1964) was 61.05 °N, 
147. 5°W. The coordinates of the Kipapa Sta- 
tion are 21°25'24"N and 158°00'54"W. The 
direction of the station from the epicenter is 
Sl5.3°W. This defines the direction of the rup- 
ture line from the epicenter as S30°W. 
In Figure 4, the rupture line, as obtained 
from the present study, is superimposed on a 
map prepared by the U. S. Coast and Geodetic 
Survey (1964) which shows the epicenters of 
the main shock and the aftershocks of the 
Alaska earthquake. In general, the aftershock 
area defines the area of rupture. In the present 
case, the rupture line obtained from Rayleigh 
wave analysis extends 100 km beyond the after- 
shock area. 
Surveys of elevation changes after the Alaska 
earthquake show positive changes in the Prince 
William Sound area, and negative changes in 
the Kodiak Island area. In Figure 5, the calcu- 
lated line of rupture is superimposed on the 
map of elevation changes as prepared by Pa- 
raras-Carayannis (see his Fig. 1, on p. 302 of 
this issue). The rupture line runs diagonally 
across the section of positive changes. In this 
calculation the direction of the rupture line may 
vary about 5°. (This value is determined by 
the resolving power of the Fourier analysis.) If 
the direction of the rupture line is turned 5° 
clockwise, with the epicenter as the pivotal 
point, the rupture line will agree with the line 
of zero displacement from field observations. 
An inspection of the directivity function 
D(f) in equation (1) shows that the period- 
icity in terms of frequency of the peaks and 
troughs of the function is controlled by the 
length B of the rupture line. The peaks and 
troughs of R 3 /R 2 and R 3 /R 4 in Figure 3 are 
such that a length of B = 800 km fits the data 
best. The superimposition of the rupture line 
on the elevation-change map shows that the 
rupture line extends to the south 100 km be- 
yond the zone of elevation changes. On the 
other hand, if the total area of the observed ele- 
vation changes is considered, the zone has a 
length of 700-800 km (Plafker, 1965). 
The present analysis shows a discrepancy be- 
tween the direction of the calculated rupture 
line and the direction expected from field sur- 
vey, but the discrepancy is within the limits of 
error of the calculation. The length of the cal- 
culated rupture line agrees with that from field 
data. 
DISCUSSION 
The results of the field survey by the United 
States-Japan Cooperative Team (Berg et al., in 
preparation) have heavily influenced the anal- 
ysis presented here since the author was a mem- 
ber of the survey team. Perhaps because of this 
bias, the analysis should not be considered as 
an independent study but, rather, as additional 
evidence to strengthen the results proposed by 
the field survey. The rupture zone of the Alaska 
earthquake of 1964 has now been outlined con- 
sistently by four different methods: (a) field 
survey of elevation changes (Berg et al., in 
preparation; Plafker, 1965); (b) plot of epi- 
centers of aftershock (U. S. Coast and Geodetic 
Survey, 1964); (c) tsunami refraction diagrams 
(Pararas-Carayannis, p. 301-310, in this issue) ; 
and (d) seismic surface wave method (this 
paper) . 
REFERENCES 
Ben-Menahem, A. 1961. Radiation of seismic 
surface waves from finite moving sources. 
Bull. Seism. Soc. Am. 51:401-435. 
and M. N. Toksoz. 1962. Source mech- 
anism from spectra of long period seismic 
surface waves, 1. The Mongolian earthquake 
of December 4, 1957. J. Geophys. Res. 
67:1943-1955. 
1963^. Source mechanism from spectra 
of long period surface waves, 2. The Kam- 
chatka earthquake of November 4, 1952. 
Ibid. 68:5207-5222. 
1963^. Source mechanism from spectra 
