355-374. April1974. UPPER MANTLE SHEAR-WAVE VELOCITY STRUCTURE IN THE JAPAN. REGION. BY K. L. KAILA, V. G. KRISHNA, AND HARI NARAIN.
Bulletin of the Seismological Society of America. Vol. 64, No.2, pp. 355-374. April1974
UPPER M A N T L E SHEAR-WAVE VELOCITY S T R U C T U R E IN T H E JAPAN REGION
BYK. L. KAILA,V. G. KRISHNA, AND HARINARAIN
ABSTRACT
The upper mantle shear-wave velocity structure in the Japan region has been determined from S travel times of 101 earthquakes with focal depths varying from 40 to 600 km, using a new analytical method given by Kaila (1969). In southwestern Japan, the S velocity obtained as 4.35 km/sec at a 40-km depth remains almost constant to a depth of about 170 km. The shear velocity in northeastern Japan increases linearly from 4.42 km/sec at a depth of 45 km to 4.62 km/sec at a depth of 145-km. For central Japan, the S velocity determined as 4.41 km/sec at a 40-km depth increases linearly to 4.55 km/sec at a 170-km depth, followed by a slight decrease in the velocity gradient, with velocity still increasing linearly to 4.68 km/sec at a depth of 345 km. At this transition depth, there is a first-order velocity discontinuity, the velocity increasing from 4.68 to 4.92 km/sec. Below this depth, velocity again increases linearly from 4.92 to 5.04 km/sec at a depth of 600 km. The shear velocities at depths between 440 to 640 km in Japan are found to be extremely low in comparison to those of Jeffreys (1939), Gutenberg (1959), and Arnold (1967). These low S velocities can explain satisfactorily the late S arrivals from shallow earthquakes between A = 20 ° to 30 ° as observed in the Japanese region. Graphs have been drawn to show the variation with depth of A*, the epicentral distance to the inflection point, A1, A2, (A2-A1), Ptrue d T / d A , and a s = (T-p A) at the inflection point as obtained from the S-wave travel-time analysis. :
INTRODUCTION
Most of our knowledge pertaining to the structure of the Earth's interior is derived from seismic waves traveling through the Earth, especially body waves generated by natural earthquakes or artificial explosions. The increased quantity and quality of bodywave data have led to a detailed and more precise picture of the structure of the Earth's crust and mantle. Several studies have been made over the last few years from the analysis of P travel-time data to derive the P-wave velocity structure in different regions of the Earth. From such studies, it is now well established that there are significant regional differences in the velocity structure of the Earth's crust and upper mantle extending to a depth of at least 1,000 km. Similar S-wave studies, however, are comparatively much less due to the inherent difficulties in dealing with them. The S arrivals have periods between 10 and 20 sec, which do not match the critical periods of most seismographs. All high-magnification seismographs have selective-frequency response which enhances the P-wave motion but discriminates against the low-frequency S waves. Moreover, the S arrivals are commonly interfered with by other phases and, consequently, picking their onset times on the seismograms is generally a difficult problem. These difficulties lead to a relatively large standard deviation (about 3 times greater) in the S travel times in comparison to P waves. 355
356
K . L . KAILA, V. G. KRISHNA, AND H. NARAIN
Gutenberg (1959) determined the velocity structure for the upper mantle beneath Japan from deep-focus earthquakes by making use of his graphical method (Gutenberg, 1953). His velocity model for the Japan region shows a decrease of velocity with depth with a minimum at about 100 km for P waves and 150 km for S waves. Jeffreys (1962), from his study of some normal earthquakes, found that the velocities of P and S at short distances for Europe, Central Asia, and northeast America are decidedly greater than those used in the 1940 tables, which depended mostly on Japanese deep-focus earthquakes. Some later studies have also shown that the seismic velocity in the mantle beneath Japan is considerably lower than average, and large lateral variations of seismic velocity exist in the Japan region. Nuttli (1963) summarized the P and S velocities over various regions of the Earth and concluded that there are lateral variations of P- and S-wave velocities in the upper mantle. Recently, Kaila et al. (1971) determined the detailed upper mantle P-wave velocity structure in the Japan region, making use of Kaila's (1969) analytical method. Their study clearly brought out the lateral variations of the upper-mantle velocity structure under southwestern, central, and northeastern Japan. The velocity structure determined by them also revealed a sharp first-order velocity discontinuity for P waves at a depth of 365 km in central Japan. In the present investigation, we have analyzed the S travel times in the Japan region to determine the detailed upper mantle shear-wave velocity structure, using Kaila's (1969) analytical method. Upper mantle S-wave velocities in southwestern, central, and northeastern Japan are found to be quite consistent with their P-wave velocity structure. The upper mantle shear-wave velocities in the Japan region, as determined in the present study, are also compared with the results reported by other investigators.
DATA USED
For the period 1957 to 1963, 101 shallow, intermediate, and deep earthquakes (their epicentral data listed in Table 1 and their locations shown in Figure 1) are selected from the Japan region for velocity determination by this method. S-wave travel times and the epicentral distances A are taken from the International Seismological Summaries (ISS) for these earthquakes. The usual criterion of rejecting S travel-time points whose residuals exceed 8 sec with respect to the J-B tables is not used as such, because it is likely to bias the velocity structure to be determined to that of Jeffreys-Bullen, corresponding to their travel-time tables. Therefore, in the present study, we have used S travel times with residuals of up to 20 sec, which enabled inclusion in our analysis of the majority of the data. This was decided after plotting the J-B residuals for the Japan region, which indeed showed larger residuals that could not be rejected without assigning any definite reason, and without the possibility of biasing the upper mantle structure to be determined.
RESULTS
For determining the upper mantle velocity structure in the Japan region, we have analyzed the S travel times of all of the earthquakes listed in Table 1 by making use of Kaila's (1969) analytical method. The apparent and true p(= OT/OA) and velocity values are obtained at various depths. The graphs for the determination o r a l and Az limits for all of the earthquakes with various focal depths are not given here due to paucity of space. However, the least-squares fit obtained for T-A points between A1 and A2 for S waves of one specimen earthquake with focal depth of 439 km is shown in Figure 2.
UPPER MANTLE SHEAR'WAVE VELOCITY STRUCTURE IN THE JAPAN REGION
357
TABLE 1 EARTHQUAKE EPICENTRAL DATA
Time
Earthquake No.
Date
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
Jul. 18. 1961 Jul. 18, 1961 Feb. 26, 1961 Nov. 7, 1960 Aug. 14, 1961 May 18, 1960 Sep. 8, 1958 May 10, 1960 Mar. 4, 1960 May 9, 1960 Apr. 22, 1962 Jan. 16, 1961 Feb. 23, 1961 Jan. 16, 1961 Jan. 16,1961 Mar. 24, 1961 Jan. 16, 1961 Mar. 19, 1961 Aug. 10, 1961 Jan. 20, 1961 Feb. 31 1961 Mar. 29, 1961 Jan. 16, 1961 May 26, 1961 Jan. 24, 1959 Mar. 18, 1959 Jul. 25, 1959 Oct. 31, 1957 Nov. 25, 1960 Jul. 15, 1962 May 3, 1960 Apr. 16, 1962 Feb. 5, 1962 Dec. 4, 1960 Jun. 9, 1961 Mar. 4, 1959 Jan. 14, 1962 Apr. 1, 1963 Aug. 27, 1962 Jun. 26, 1959 Oct. 17, 1962 May 13, 1957 Jan. 26, 1962 Jan. 17, 1957 Jul. 22, 1957 Apr. 10, 1962 Oct. 14, 1958 Jun. 22, 1958 Sep. 22, 1958 May 13, 1957 Sep. 1, 1958
Continued.
(h) (m) (s) 14 03 15 16 18 10 13 23 22 05 06 35 14 53 23 17 03 53 00 11 19 15 07 20 04 16 12 12 11 19 22 57 15 41 09 18 12 03 22 34 13 31 18 10 14 04 22 49 05 08 07 26 21 20 02 37 21 54 06 47 22 22 13 20 22 55 16 20 15 43 23 00 13 34 04 28 02 18 05 03 12 39 02 21 05 22 22 26 10 16 23 54 21 05 05 29 08 37 15 19 15 29
40 17 52 12 06 12 18 59 11 27 31 12 20 30 40 12 17 50 26 49 44 20 02 55 38 47 34 00 12 22 42 10 52 46 03 52 09 44 59 58 13 00 53 51 37 14 17 33 24 37 34
Epicenter Latitude N Longitude E (deg) (deg) 29.74 29.83 31.84 32.24 31.59 29.35 33.75 33.90 31.15 29.90 32.16 36.16 38.41 36.41 36.19 35.88 36.61 36.80 37.33 37.13 36.41 36.83 36.47 38.79 37.42 36.68 36.79 37.67 38.19 39.62 32.32 30.32 35.93 32.77 30.35 37.55 44.47 44.74 40.27 44.98 33.45 43.68 32.26 33.14 34.73 37.46 33.76 37.39 27.38 32.36 38.00
131.73 131.58 131.56 132.04 131.44 129.98 131.85 131.82 130.39 129.65 130.45 141.71 143.00 141.79 141.46 141.06 141.53 141.02 137.15 141.49 141.08 141.60 141.80 142.86 141.17 141.09 140.70 140.82 140.81 141.08 140.08 140.97 139.32 141.37 140.52 138.70 140.94 141.34 137.98 141.35 137.86 135.68 138.18 137.58 136.19 135.41 137.12 135.03 140.12 137.72 134.51
Focal Depth (kin) 39 48 56 65 70 81 84 93 141 170 171 36 37 39 50 51 52 54 60 61 63 64 67 80 86 90 94 94 103 106 119 121 135 146 154 219 242 256 270 271 342 350 360 364 364 365 366 369 393 404 405
358
K. L. KAILA~ V. G. KRISHNA, AND H. NARAIN
TABLE 1--Continued Time
Earthquake No.
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
Date
Jun. Dec. Aug. Feb. Jul. Aug. Sep. Dec. May Jan. Apr. Feb. Nov. Feb. Apr. Jun. Apr. Sep. Sep. Aug. Aug. Apr. Jul. Nov. Jul. Feb. Apr. Jan. Dec. Aug. Jan. Jan. Aug. Jun. Mar. Aug. Jan. May Jan. Sep. Apr. Apr. Nov. Jan. Mar. Mar. Jan. Feb. Jan. Oct.
26,1959 17, 1958 19, 1960 11, 1962 18, 1957 1, 1957 13, 1960 7, 1962 20, 1959 18, 1963 18, 1960 11, 1961 17,1957 23, 1958 9, 1957 29, 1960 9, 1957 28,1957 14,1960 13, 1963 15,1963 2, 1957 2, 1960 6,1958 1, 1959 25, 1959 3, 1960 28,1959 3, 1958 6, 1960 3, 1957 3, 1957 11, 1961 12, 1957 30, 1963 21, 1961 23, 1961 27, 1961 6, 1961 8, 1959 15, 1960 26, 1958 27,1960 3, 1960 5, t959 26, 1963 12, 1959 7, 1961 23, 1958 19, 1957
Epicenter
(h) (m) (s)
Latitude N (deg)
08 08 12 02 12 17 03 14 11 03 08 06 17 09 00 05 10 00 03 23 02 08 12 21 02 11 05 13 16 03 13 12 23 08 16 17 04 07 01 19 11 17 15 21 14 19 14 21 02 21
30.67 32.85 26.96 29.65 30.04 29.91 27.23 29.22 32.65 33.12 28.44 28.95 30.41 28.64 30.54 30.08 30.57 30.93 30.23 27.38 28.00 30.07 41.32 27.81 28.11 28.43 27.84 28.89 28.85 28.06 43.87 43.85 42.91 41.18 44.20 41.00 43.20 41.32 42.44 42.34 41.00 42.44 42174 45.07 44.05 44.31 43.98' 44.45 44.32 44.20
35 57 41 42 06 53 09 03 26 12 07 12 55 12 24 14 35 27 49 00 17 33 44 46 27 19 10 59 01 05 43 48 33 28 51 00 48 18 20 19 39 17 17 20 09 47 16 01 34 42
52 15 47 39 47 17 13 40 30 07 11 34 07 27 43 59 34 33 52 30 20 13 22 05 47 09 34 57 01 57 32 30 51 38 58 41 27 06 38 39 00 40 17 13 50 47 35 52 10 00
Longitude (deg)
Focal Depth (kin)
138.93 137.12 •40.32 139.32 138.94 139.31 140.28 139.38 136.67 135.96 139.75 139.46 138.40 139.57 138.39 138.66 138.24 137.85 138.26 140.16 •39.76 137.22 131.96 139.57 139.51 138.81 139.52 138.02 138.10 139.19 130.55 130.63 145.47 142.83 148.11 139.42 145.68 142.21 143.50 142.89 141.60 143.60 143.45 148.19 146.93 147.30 145.84 147.09 146.04 146.08
413 414 423 425 430 433 436 439 441 444 445 448 449 450 457 462 468 468 486 490 503 520 522 525 533 537 538 539 545 547 579 593 45 45 47 55 65 65 68 70 94 99 107 108 109 123 128 131 139 146
359
UPPER MANTLE SHEAR-WAVE VELOCITY STRUCTURE IN THE JAPAN REGION
I
135"
130° E
140~
15d~E
145"
~N
odN
4~
4~--
×39 4~
24 X Eli3
"°L
@46
~ t, Z~ Z~ Z~
.• 061
047 04 I @44
X34
60" ~3 @5043
34
73@ 681~64 70@~67
I
AX35 X32 -
2 q~l
79--
130°E
J 135"
-
3
o
°
65@ 2
140 °
145"
25°N 150°E
FIG. 1. Map of Japan region showing locations of observatories and earthquake epicenters numbered as in Table 1. Various symbols used are: A, observatories; [], shallow earthquakes; x , intermediate earthquakes and O, deep earthquakes.
360
K.L.
K A I L A , V. G . K R I S H N A , A N D H. N A R A I N
In the same figure, ( T - p A ) versus A is also plotted. It is noticed, for some of the earthquakes in the Japanese region, that the S travel times reported in the I S S do not correspond to the same cycle of the S phase throughout the A range reported. This fact is clearly brought out in our Figure 3 for one earthquake specimen with focal depth of 94 kin, which depicts such a feature. The travel-time points between A 1 and A 2 are separated as if by one cycle (open circles and solid circles) as shown in this figure, and for this reason least-squares fits are obtained separately for the two plausible sets of traveltime points. In the same figure, ( T - p A ) versus A corresponding to two sets of T-A points are also plotted. It can be seen that the two sets of travel-time points fit to two almost parallel straight lines shifted from each other by about 12 to 15 sec, which is equal 1560
I h=4-59Km I
520
280
240
200
/ 160
120
8O
!
T= (20.914-.30)~+39.5
1
4- 3.3
I I
CO I'-
/
I I 1 1 I ! I l
4O
I
[ 4
i
I
8
i
I
i
12
I
16
i
I
20
~°~
"1~, r 24 " ~
B°
l
4o
"~
o
A (Deg) --40
FIG. 2. S-wave travel-time curve T versus A for earthquake 59 in central Japan, A1 and A2 of epicentral distance. Reduced travel-time curve (T-pA) versus A is p = OT/OA at the inflection point.
O Cq I F-
b e t w e e n t h e limits also shown where
to the period of the S waves. The two sets of travel-time points for this earthquake yield almost identical velocity, as is clear from the slopes obtained for the two straight lines fitted to them, indicating thereby that they belong to the same S phase but to two different cycles. Obviously, such a misidentification of the onset of the S phase (first cycle) is quite possible in the presence of other phases and more so because, unlike P waves, the S wave is not recorded as a first arrival. The travel times of S waves of all of the earthquakes used in the present study are scrutinized in the light of the above observation, and necessary care is taken so as not to mix two different cycles of the same S phase in the velocity determination. The final results obtained f o r p ( = OTIs?A) values and the corresponding velocities with their standard deviations are given in Table 2. The earthquakes in the three regions of Japan, central, southwestern, and northeastern, as in our 1971 paper were treated separately
UPPER MANTLE SHEAR-WAVE VELOCITY STRUCTURE IN THE JAPAN REGION
361
for obtaining the velocity structure of the upper mantle. The true velocities thus obtained at various depths for the three regions of Japan are plotted as IA, 2, and 3 in Figure 4. The velocity-depth points are shown by circles, and the horizontal bars with them indicate the standard deviations of those velocity values. The thick straight lines are the least-squares fits made separately to the velocity-depth points in the three regions, and the dashed straight lines bounding them show the 95 per cent confidence limits of their slopes. The velocity-depth functions for the three regions of Japan, as can be seen from Figure 4, show quite different pictures. In southwestern Japan, the S velocity obtained as 4.35 km/sec at a 40-km depth remains almost constant to a depth of about 170 km. 280
--
h=94Km
240
200
,4 160
,,
o~
.~
120
~
8o
]d k
/1 40
//
-
'/
1 ~
0
40
4
l
,o
I
8
12 /x (Deg)
- ~16- ~.
o
24
"~
L.-.="~"o..
",,
-40
~.
\
