THE NATURE OF SEISMICITY PATTERNS BEFORE LARGE EARTHQUAKES
Hiroo Kanamori
Seismological Laboratory
California Institute of Technology, Pasadena, California 91125
Abstract. Various seismicity patterns before quake prediction; measurements of other physical major earthquakes have been reported in the liter parameters such as the spectra, the mechanism and ature. They include foreshocks (broad sense), the wave forms of the background events should be preseismic quiescence, precursory swarms, and made concurrently. doughnut patterns. Although many earthquakes are preceded by all, or some, of these patterns, their Introduction detail differ significantly from event to event. In order to examine the details of seismicity Spatio-temporal variations of seismicity before patterns on as uniform a basis as possible, we major earthquakes have been studied by many made space-time plots of seismicity for many large investigators in an attempt to understand the earthquakes by using the NOAA and JMA catalogs. physical mechanism of earthquakes and to use them Among various seismicity patterns, preseismic as a tool for earthquake prediction. In this quiescence appears most common, the case for the paper, we review the recent progress in this 1978 Oaxaca earthquake being the most prominent. field, add some new data, and propose a simple Although the nature of other patterns varies model which facilitates the understanding of the from event to event, a common physical mechanism nat..:.re of these seismicity patterns. may be responsible for these patterns; details of Since these patterns have not been defined the pattern are probably controlled by the tecton unar,1biguously, we first discuss some representa ic environment (fault geometry, strain rate) and tive patterns b:,r using a schematic diagram shown the heterogeneity of the fault plane. Here a by Figure 1. This figure includes, following simple asperity model is introduced to explain Kogi (1976), the pattern of foreshocks, precursory these seismicity patterns. In this model, a fault swarms, precursory quiescence and doughnut plane with an asperity is divided into a number of patterns. subfaults. The subfaults within the asperity are, on the average, stronger than those in the sur Foreshocks rounding weak zone. As the tectonic stress increases, the subfaults in the weak zone break in Although there is no widely accepted definition the form of background small earthquakes. If the of foreshocks, some earthquakes (e.g., 1974 frequency distribution of the strength of the sub Haicheng, China earthquake; 1963 Kurile Islands f aults has a sharp peak, a precursory swarm occurs. earthquake) were preceded by a very remarkable By this time, most of the subfaults in the weak short-term increase in the number of small events zone are broken and the fault plane becomes in the epicentral area so that little ambiguity seismically quiet. As the tectonic stress exists in calling them the foreshocks. In other increases further, eventually the asperity breaks cases, however, ambiguity arises because of either and sympathetic displacement occurs on the entire too small number of events, too spreadout time fault zone in the form of the main shock. Fore interval, or both. Yet these events may be shocks do or do not occur depending upon the causally and/or physically related to the main distribution of the strength of the subfaults shock, and may be called the foreshocks. within the asperity. Since the spatio-temporal Sometimes small events which preceded a main change in the stress on the fault plane is most shock and occurred in, or in the neighborhood of, likely to dictate the change in seismicity the mainshock rupture zone are called preshocks. patterns, detailed analysis of seismicity patterns In this paper, we will use the term foreshocks would provide a most direct clue to the state of in a rather loose sense of the word to include stress in the fault zone. However, because of the both "obvious" foreshocks and preshocks. large variation from event to event, seismicity According to Jones and Molnar (1976), about 44% pattern alone is not a definitive tool for earth- of large shallow earthquakes in the world were
KAi'IAMORI 1 Seismicity_ Pattern island had not experienced a large earthquake for 70 years, and suggested that a large earthquake Main Space was imminent there. The Tonankai (Er = 8.0) and Shock the Nankaido (Ms = 8.2) earthquakes indeed occurred there in 1944 and 1946 respectively. • Fedotov (1965) and Mogi (1968a) studied seismicity Background / in the Kamchatka, Kurile and Japan regions and identified several zones which had not experienced a large earthquake for a long t1me. These zones were considered to be candidate sites of major earthquakes in the future. In fact, several major earthquakes including the 1968 Tokachi-Oki, Japan earthquake (Mw = 8.2) occurred in these zones subsequently. These results were furhter extended to the concept of seismic gaps, and have been used 0 more globally by many investigators (Kelleher, Time 1970; Utsu, 1970; Kelleher et al., 1973; Sykes, E le men ta rY-_ Patterns 1971; Ohtake et al., 1977). Kelleher and Savino (1975) demonstrated that gaps in seismicity for CD Foreshocks (Preshocks) great earthquakes are also gaps for smaller magni tude activity and such gaps commonly persist until ® Quiescence the time of the mainshock. A very comprehensive @ (+cg)) Precursory Clustering (Swarm) review can be found in McCann et al. (1980). Usually these seismic gaps refer to a spatial gap @(+cg)) Doughnut of seismic activity, particularly of large earth Fig. 1. Schematic space-time diagram showing quakes. various seismicity patterns. (Modified from Mogi, 1977). Quiescence Inouye (1965) found that seismicity in the epi preceded by foreshocks of their definition. A central area of several large earthquakes in Japan very useful summary of foreshock activity in Japan (e.g., 1952 Tokachi-Oki and 1964 Niigata earth can be found in Mogi (1963). Among the best quakes) became very low before the mainshock. This documented foreshock sequences are those of the quiescence was followed by increased activity for 1974 Haicheng, China earthquake (Wu~ al., 1978), several years before the mainshock. Mogi (1968a) the 1978 Oaxaca, Mexico earthquake (Ponce, et al., showed that before several large earthquakes 1977-1978) and the 1963 Kurile Islands eqrthquake (e.g., 1944 Tonankai and the 1946 Nankaido earth (Santo, 1964). quakes), the focal region became very calm. Occasionally, a very tight clustering of activi These studies suggested that seismic activity in ty occurs before the mainshock. Mogi (1968b) and the eventual rupture zone of a large earthquake Kelleher and Savino (1975) demonstrated that decreases more or less abruptly sometime before seismic activity prior to a great earthquake the mainshock. In this regard this pattern may tends to cluster around the epicenter of the be called a temporal gap. Perhaps the most eventual mainshock. More recently, Ishida and pronounced of the temporal gaps is the one before Kanamori (1978) and Fuis and Lindh (1979) found the 1978 Oaxaca, Mexico earthquake (Mw = 7.6) a very tight clustering of activity before the reported by Ohtake et al. (1977). Mogi (1979) 1971 San Fernando, California and the 1975 Galway called the spatial gap and the temporal gap, the Lake, California earthquakes, respectively. gap of the 1st and 2nd kind respectively. In Engdahl and Kisslinger (1977) found a clustering any case, a preseismic quiescence of seismic of small events before a magnitude 5 earthquake activity in the epicentral area of a large earth in the Central Aleutians. Although these events quake appears very common to many large earth are not usually called typical foreshocks, they quakes. can be considered to be f oreshocks in a broader sense, or preshocks. Precursory Swarm
McNally (1977) found that distinct clusters of small earthquakes occurred in the near-source Imamura (1928) investigated historical data on region of several moderate size earthquakes in large earthquakes in southwest Japan (Tokaido Central California 2 to 10 years before the main Nankaido region), and found that large earthquakes shock. Sekiya (1977) and Ohtake (1976) reported in this region had occurred repeatedly at approx that anomalous seismic activity occurred about imately the same location with a repeat time of 10 years before the 1974 Izu-Hanto-Oki earthquake about 100 to 150 years. He ,iointed out that the in the epicentral area which had generally been area southeast of the Kii peninsula and Shikoku quiet before the earthquake. Sekiya (1977) 2 KANAl10RI reported further examples for aho~t ten other used to identify a possible physical mechanism. Japanese earthquakes. Evison (1977a,b) found Once a physical mechanism is identified, other such precursory activities before the 1968 means such as monitoring temporal variations of Borrego Mountain, California earthquake and source mechanism, spectra and wave forms may be several earthquakes in New Zealand. Evison con used for prediction purposes. In view of the sidered that a burst of seismic activity marks the large degree of non-uniformity of the presently start of a precursory sequence, and called it the available seismicity catalogs and of the methods precursory earthquake swarm. Brady (1976) found used by various investigators, we will be primarily a clustering of seismic activity before the 1971 concerned with the second approach in this paper. San Fernando, California earthquake and inter preted it as a "primary inclusion zone" of the Examples of Seismicity Patterns impending failure. Various seismicity patterns which have been Doughnut Pattern reported so far are summarized in Table 1 in terms of the elementary patterns described in the Mogi (1969) found that before several large Introduction. Although these results provide a earthquakes in Japan, the region surrounding the fundamental data base for the present study, we focal region became very active, while the focal made a global survey of seismicity patterns region was quiet at the same time. This pattern associated with large earthquakes by using space is often called a doughnut pattern. A similar time plots of seismicity in order to clarify the doughnut pattern has been reported for a magnitude nature and regional variation of seismicity 6 earthquake in Kyushu, Japan by Mitsunami and patterns. Kubotera (1977) and for a magnitude 6.1 earthquake We used the NOAA catalog for all of the regions in the Shimane prefecture, Japan by Yamashina and except for Japan. Since the uncertainty in the Inoue (1979). locaticn is probably about 30 to 50 km, we will For many earthquakes, these elementary seismi be primarily concerned with the patterns for city patterns described above appear either by earthquakes larger than magnitude 7.5 (The long itself or as combinations so that the actual period magnitude, Mw, is preferred whenever pattern often becomes very complex. Furthermore, available. If it is not available, the surface identification and classification of the patterns wave magnitude, Ms, is used.) whose rupture depend upon the catalog used for the study, the length is 70 km or larger. magnitude threshold, the time period, the depth The non-un~formity of the catalogues results range, and the judgement of the investigators, so from combinations of many factors which include: that entirely different patterns have of ten been 1) temporal variations in the number and distri identified for the same event. Because of this bution of stations, 2) changes in the practice of ambiguity, the reported seismicity pattern should the magnitude deternination both at the individual not be regarded as a unique feature of the earth stations and the central agency, 3) changes in the quake, but should rather be regarded as a mani instruments, 4) changes in the personnel and the festation of the physical process leading to an operation practice at the individual stations, and earthquake. 5) changes in the location procedure. It is not In general, two approaches are possible in clear how we can remove the e=f ect of all of these using seismicity patterns for earthquake predic factors to extract the real spatio-temporal tion. The first is represented by the works of variations of seismicity. Hence, we will plot the Keilis Borok et al. (e.g., 1980), Wyss et al. raw data from routinely available catalogues with (1978) and Habermann (1980). In this method, out heavy processing. Symbols with different various seismicity patterns are treated as sizes are used for different magnitude ranges to rigorously as possible in a statistical framework facilitate visual inspection of the patterns. As to establish an empirical algorithm for earthquake will be shown later, it is encouraging that some prediction. Essential to this approach are the patterns are discernible from the space-time plots uniformity of earthquake catalogs and rigorous produced in this manner from routinely available definition of the seismicity patterns. catalogues. We first illustrate the method for In the second approach, various seismicity Mexico. patterns are used as a clue to the physical mechanism of earthquake failure process. Although Mexico the observed seismicity patterns are very complex, the fundamental physical mechanism may be simple. We extracted all the events shallower than 60 km The complex structures of the fault zone may be which occurred within the box shown in the index primarily responsible for the variation of the map (Figure 2). Then the distance to the individ observed seismicity patterns. In this approach, ual epicenter from the pole shown in Figure 2 was rigorous definition of seismicity patterns and measured and the events were plotted as a function uniformity of the catalo~s are less important than of time in the form of a space-time plot as shown in the first approach. Although seismicity pat in Figure 3. The pole is placed on the approximate terns thus somewhat loosely defined may not be extention of the strike of the region considered, used directly for prediction purposes, they can be and at a distance comparable to the total length
KANAMORI 3 -1'- TABLE 1. OBSERVED SEISMICITY PATTERNS*
EVENT PRECURSORY QUIESCENCE FORE SHOCK DOUGHNUT REFERENCE NOTE ~ SWAR'1 (PRESHO CK) PATTERN ~ H Cape Kamchatka,1971 1969 during the 1962-1971 No No Wyss et al. See alao "" (M = 7. 8) quiet period ( 19 7 8) Fedotov et. al. s (1977) Friuli, 1976 See the column No activity Clustered activity No Wyss et al. (Ms = 6.4) under Foreshock prior to 1960 in March 1975 (19 78) Assam, 1897 During 28 years Khattri & (Ms = 8.2) before the main Wyss (1978) shock Kalapa::ra, Hawaii No About 2 years Wyss et al. 1975 (M = 7.2) from May,1972 (19 78) s San Fernando, 1961 "" 1964 1965 to 1969 From Sept.,1970 No Ishida & See also Brady California, 1971 Kanamori(l978) (1976,1977) (M = 6.5) s Quiet period 211 days before No Ohtake et al. began 597 days the main shock (1978 before the main shock. Kamchatka, 19 52 During 3 years Not Kelleher & From 1920 (M = 9.0) just before the obvious Savino (1975) w main shock Alaska,1964 From 1944 1954 to 1964 Not Kelleher & (M = 9. 2) increased obvious Savino (1975) w activity near the edge of the rupture zone Sitka,Alaska No Quiet everywhere No No Kelleher & 19 72 (M = 7. 2) near the Savino (1975) s rupture zone Alaska, 19 58 No Increased No Kelleher & (M = 7.9) activity during Savino (1975) s about 10 years before the main shock Kern County, 1939"vl941 Much of the Clustered No Wesson & California, 1952 rupture zone activity near Ellsworth(l973), (M = 7. 7) quiet for at the epicenter Kelleher & s leat 20 years Savino(l975), Ishida & Kanamori(l980) Parkfield, No Quiet in the Small foreshocks No McEvilly et al. California,1966 rupture zone 8 days before (1967), (~ = 5.5) about 9 months the main shock Kelleher & Savino(l9 75) Oh take et al. (19 78) Table 1. cont'd.
EVENT PRECURSORY QUIESCENCE FORES HOCK DOUGHNUT REFERENCE NOTE SWAR-M (PRESHO CK) PATTERN
Chile, 1960 Quiet for at Immediate No Kelleher & (MW = 9. 5) least about foreshocks Savino ( 19 7 5) 5 to 8 years (33 hours before the main shock) Tokachi-Oki,1952 No 1943 to 1951 Yes Mogi (1969) (MW = 8.1) 1926 to 1951 Utsu (1968) 1934 to 194 7 Increased activity Inouye (1965) during 1948 to 1952 1937 to 1949 2 years before the Katsumata (1973), main shock Katsumata & Yoshida (1980) Tonankai,1944 No 20 years No Yes Mogi (1969) (M = 8.1) and Nai°ikaido,1946 (MW = 8.1) Sanriku,1933 12 years No Yes Mogi (1969) (MW = 8.4)
Tokachi-Oki,1968 No 1961 to main Yes Mogi (1969) (MW= 8.2) shock except 1965 1948 to 1963 Increased activity Katsumata & near the epicenter Yoshida (1980) 1964 to 1968 1962 to May 1968 Habermann(l980) Shimane, Japan No 5 months before No Yes Yamashina & 1978(MJMA= 6.1) the main shock Inor.!e(l979) Oaxaca, Mexico No About 5 years Yes No Oh take et al. 1978(M = 7 .6) before the (1977), Ponce w main shock et al. (1977-78)
Oaxaca, Mexico No 1966 to 1968 Increased activity No Oh take et al. 1968(M = 7.1) (1977) s just before main shock Oaxaca, Mexico No 1964 Preshocks No Oh take et al. 1965(M = 7 .6) 1 year before (19 77) s main shock Milford Sou..'!d, 1968 1969 to 1975 Evison (1977a) New Zealand, 1976(ML= 7.0) ~ Borrego Mountain, 1965 1965 to 196 7 Yes No Evison (1977a) ~ California,1968 •\bout 1 year Increased activity Oh take et al. 0 (~ = 6.4) for 395 days (1978) H '° Mendocino Ridge, 1959 1959 to 1960 No No Evison (1977b) California, 19EO Vl (~= 6.2) O' Table 1. cont'd.
EVENT PRECURSORY QUIESCENCE FORE SHOCK DOUGHNUT REFERENCE NOTE ~ SWARM (PRESHO CK) PATTERN ~ ::
Inangahua, 1962 1963 to 1967 No No Evison ( 19 77 c) New Zealand,1968 (M = 7 .1) 1
Hastings, March 1972 April 1972 to No No Evison (1977c) New Zealand,1973 Jan 1973 cl\ = 5. 7)
Off Izu Peninsula 1963 to 1965 1965 to 1974 No Sekiya (1977) Japan, 1974 1961 to 1966 1967 to 1974 No Ohtake (1976) (MJMA= 6.9)
Central Gifu, About 5 years Sekiya (1977) Japan,1969 before the (MJMA= 6.6) main shock
Choshi, Japan 3 years and 5 Some foreshocks No Seki ya ( 19 77) 1974 months before (MJMA= 6.1) main shock
Fukui, Japan 19 years and 3 Sekiya (1977) 1948 months before (MJMA= 7.3) the main shock
N. Miyagi, Japan 1956 to 1958 Sekiya (1977) 1962(MJMA= 6.5
Shizuoka, Japan 4 years before Sekiya (1977) 1965 (MJMA = 6 .1) the main shock
Ebino, Japan About 15 days Sekiya (1977) 1975(MJMA= 4.1) before the main shock Kanta, Japan About 82 years Sekiya (1977) 1923(~ = 7.9) before the main shock L
EVENT PRECURSORY QUIESCENCE FORE SHOCK DOUGHNUT REFERENCE NOTE SWARM (PRE SHOCK) PATTERN
Central Aleutian, 4-1/2 months 6 foreshocks Engdahl & 1976 (1\ = 5) prior to the during 5 week Kiss linger main shock period (1977) Markansu, About 7 years 1968 to 1974 Kristy & Central Asia before main shock Simpson(l980) 1974(M = 7.4) s
Zaalai, About 2 years Kristy & Central Asia after main shock Simpson ( 19 80) 19 7 8 (M = 6. 7) s Imperial Valley, About 4 months About 3-1/2 Yes Johnson & California before the months before Hutton(l980) 1979(M = 6.9) s main shock the main shock Off Fukushima, Quiet at least Increased Inouye Japan,1938 from 1926 to activity from (1965) (MJMA= 7.7) 1933 1934 to 1938
Niigata, Japan 1946 to 1961 Slight increase Inouye 1964 (MlJ = 7 ,6) in activity from (1965) 1962 Aso, Japan 3 days before 1-1/2 days During 30 hours Mi tsunami 3-dimensional 1975 (MJMA= 5.9) the main shock before the before the & Kubotera feature main shock main shock (1977) Assam, 1950 > 30 years Khattri & Similar 2.2 years before Tanaka et al. the main shock Kurile Is., 1967 to 1969 Katsumata and 1969(Mw = 8. 2) Precursory activity Yoshida (1980) Nemuro-Oki 1961 to 1971 1972 to Katsumata main shock Yoshida(l980) ~0 Wakayama, 1968 :;o 2 to 3 years A precursory Mizou"' et al. H & 1977 event (1978) (M = 4.8,4. 7) ___, Pole H ,....._ ctl r-- • r-- +.J Cll °'H :> 1000 km Mexico 0 ~ U),-._ H Cll C'"l H (/) r-- ctl Fig. 2. Index map for Mexico. All the events µ z CllH°' cJ shallower than 60 km which occurred in the box are z '-' :;;:: shown in Fig. 3. The asterisks show the location of the 1973 Colima earthquake and the 1978 Oaxaca earthquake. The location of the pole is arbitrar~ and the scale refers to the middle of the figure. of the region. The pole position would not have a drastic effect on the overall pattern. The magni tude ranges for the larger symbols are indicated in the figure; the magnitude ranges for the events smaller than 6 are not indicated but the +.J i:: size of the symbols is approximately proportional Cl) :> to the magnitude. The dashed curves indicate the Cl) period of quiescence before the 1973 Colima and 0 +.J 1978 Oaxaca earthquakes. No rigid criterion is +.J used for drawing these curves; they are drawn ffi mainly to indicate the region being discussed ~ rather than to define it. Therefore the estimate s of the length of the quiet period and the size 0 µ 4-l µ.:i µ u Cl) + + 4-l M 2 8 8>M 27 7>M 26 4-l NW Mexico d :0::60 km 5 5 5 ~ .,., Ul u:i '"dH µ.:i .,., H Ul ctl p i:: +.J '+ O' µ Cl) 320 Cl) '"d +.J +.J µ ctl 0 !." P-<4-l + Cl) Ul .+t 1- ·~ :. ,.c: aJ 0 + ,, +.J C) 270 + ' u) +i· 4-l ffi u 0 µ c aJ i:: 4-l 2 0 aJ .,., µ 0 +.J .,., H 220 i:: ctl .,., i:: 4-l .,., aJ bl) '"d .,., µ Cl) 0 ,.c: +.J aJ ,.c: 70 80 4-l +.J 0 Time, year 0 Ul +.J ,...; Fig. 3. Space-time plot of seismicity for Mexico .,., µ obtained from the NOAA catalog (See Fig. 2 for the ctl aJ +.J 4-l location). Dotted curves encircle the period of aJ aJ i:::. ~ quiescence. The magnitude ranges for the larger symbols are indicated in the figures. The magnitude ranges for the events smaller than 6 are not indicated but the size of the symbols is approximately proportional to the magnitude. 8 KAN AMO RI Colima earthquakes, no obvious precursory swarms or doughnut patterns are seen. Japan-Kurile - Kamchatka Ohtake et al. (1977) reported a pattern of quiescence before the 1965 (Ms= 7.6) and 1968 (Ms= 7.1) Oaxaca earthquakes. Although there is some indication of reduced seismicity prior to these earthquakes, the pattern is not obvious on the scale of this plot (see Figure 3). As seen in Figure 3, there are many quiet zones (Kurile) which are not followed by a large earthquake. It is important to note that while large earthquakes appear to be preceded by a period of quiescence, the mere existence of a quiet period does not necessarily point to an impending large earthquake. We will proceed with a similar analysis for other seismic zones. The analysis method and the basic philosophy of interpretation will be the same un less noted otherwise. Kurile Figure 4 shows the index map for the Kurile, Kamchatka and Northern Japan regions. The result •Pole (Kamchatka) for the Kurile Islands is shown in Figure 5. During the period from 1960 to 1978, there were three events larger than 7.5, the 1963 event 1000 km being the largest. •Pole ( N. Japan) The 1963 event (Mw = 8.5) was preceded by a distinct period of quiescence. During 1961, the Fig. 4. Index map for the Kurile Is., Kamchatka seismic activity became very high along a sub and Northern Japan. For details, see caption for stantial length of the arc. This increase may be an artifact of increased number of reports from a Fig. 2. regional network. However, it consists of many events with a magnitude larger than 6, and is of the zone indicated by the dashed curves should unlikely to be entirely due to nonuniformity of not be given too much significance; no attempt is the catalog. For example, the numbers of events made here to use it for statistical arguments. with mb ~ 6.0 which occurred within this box are Since the detection capability, the location 2, 3, 15 and 2 for 1959, 1960, 1961 and 1962 accuracy, the reporting procedure for the indi respectively. For events with mb ~ 6.3, the vidual station, and the data reduction procedure corresponding numbers are 1, 2, 8 and 2. The 1963 are not uniform during this time period, the earthquake was preceded by remarkable f oreshock result shown here is inevitably nonuniform. For activity (e.g., Santo, 1964). Also, an increased example, the sudden increase in the number of activity during about a one year period before small events in 1963 is probably due to the the mainshock is seen in Figure 5. establishment of the World Wide Seismographic Station Network (WWSSN). Also, the number of small events seems to have decreased abruptly in 1969. This sudden change could be an artifact of Ku rile d :0:60 km the reporting and the cataloguing procedures. Despite this nonuniformity, the quiescence before these two earthquakes, particularly the 1978 0axaca earthquake, appears very obvious. The i: 130 quiescence before the Oaxaca earthquake was Q first noted by Ohtake et al. (1977) and was one of a) () the basis of their forecast -of this earthquake. c 0 The quiescence before the Colima earthquake is in 100 less pronounced, but the activity in the encircled 0 area seems to be lower than that during the pre t ceding time period. As seen in Figure 3, the Oaxaca earthquake was preceded by several fore 70 shocks which were located by the world-wide net work. At a smaller magnitude level, more than 10 Fig. 5. Space-time plot of seismicity for the foreshocks were located by a local network (Ponce Kurile Is. region. See the caption for Fig. 3 et al., 1977-1978). For both the Oaxaca and for details. KANAMORI 9 A similar pattern is seen for the 1969 Kurile Islands earthquake (Mw = 8.2). This earthquake Chile d $. 60 km was preceded by several foreshocks during the 30 ?------minute period before the mainshock. The pattern . r, for the 1973 Nemuro-Oki earthquake = 7.8) is (Mw E r more or less similar to those for the 1963 and Q530~ " ' ? - --- . 1969 events, though the precursory swarm is not I very distinct. The 1973 event was preceded by I several small foreshocks. ' . Thus, for all three major earthquakes which occm-red in the Kurile Islands since 1963, the swarm-quiescence-foreshocks pattern can be seen. Although the location accuracy is not good enough 60 Time. year to investigate such patterns for smaller events, Fig. 7. Space-time plot of seismicity for Chile. there is an indication that such patterns preceded See caption for Fig. 3 for details. smaller events in 1961, 1975 and 1976, as shown by dashed curves in Figure 5. period started. The 1960 Chilean earthquake was South America preceded by remarkable foreshock activity during the 33 hour period just before the mainshock. Figure 6 shows the index map for South America. For the Peruvian subduction zone (see Figure 8), Figure 7 shows the space-time plot for Chile. three events larger than 7.7 occurred during the Although the data prior to 1960 are, in general, period since 1960. The 1970 event (Mw = 7.9) rather poor, it is very clear that Stismic occurred at a depth of about 70 km and probably activity in the rupture zone of the 1960 Chilean represents failure within the down-going slab earthquake (Mw = 9.5), the largest event in this (Abe, 1972; Isacks and Barazangi, 1977). Since century, had been lower than in the adjacent the location of this event is very close to that segment to the north during the preseismic period. of the 1966 event, the pattern of seismicity for This low activity has already been pointed out by this event cannot be studied very well with the Kelleher and Savino (1975). Because of the poor space-time plot used here. For both the 1966 quality of the data, it is unclear when the quiet (Mw = 8.1) and 1974 (Mw = 8.1) events, a period of seismic quiescence seems to have preceded the mainshock, as shown in Figure 8. However, since the total number of events is relatively small, the statistical significance of these patterns is considered marginal. The pattern of precursory swarms and doughnut patterns are not evident. Neither of these events had f oreshock activity detectable by the world-wide network. For the Colombia-Ecuador subduction zone, the data are too sparse to study seismicity patterns. Alaska-Aleutians Figure 9 shows the locations of the seismic zones in the Alaska-Aleutians region studied here. Figure 10 shows the space-time plot for the 1964 Alaska earthquake (Mw = 9.2) and the 1957 Fox Island earthquake (Mw = 9.1). •Pole (Colombia - Ecuador) In constrast to the examples shown above, both of these earthquakes were preceded by a distinct increase in seismic activity which may be called a preshock activity during about 10 years before the mainshock. Although this increase may be due •Pole (Peru) to increased detection capability, the fact that the commencement of the increased activity for the 1964 event differs from that of the 1957 event 1000 km suggests that it represents a real seismicity change. This kind of increase in activity before large earthquakes concentrated near either end of the rupture zone, has already been pointed out by Kelleher and Savino (1975). This preshock Fig. 6. Index map for South America. See caption activity was preceded by a relatively quiet period, for Fig. 2 for details. although it is not very distinct compared with the 10 KANAHORI + + + M .2:8 8>M ?:.7 7>M 2:_6 Peru d S.60 km 5 5 5 N E .'<'. 425 Q QJ () c 0 "lil 375 0 70 Time, year Fig. 8. Space-time plot of seismicity for Peru. See caption for Fig. 3 for details. Fig. 10. Space-time plot of seismicity for the rupture zones of the 1964 Alaskan and the 1957 Fox Is. earthquake. For details, see caption for adjacent regions. During the period from 1928 to Fig. 3. 1936, an increased level of seismic activity is seen in the rupture zone of both the 1964 and 1957 events. This activity may be considered to be a Kamchatka precursory swarm. Figure 11 shows the pattern for the 1965 Rat The area to be considered is shown in Figure 4, Island earthquake(~= 8.7). The main feature and the results are shown in Figures 12 and 13. of the seismicity pattern is similar to that of Figure 12 shows the result for the 1952 Kamchatka the 1964 and 1957 earthquakes. A relatively earthquake (Mw = 9.0). Unfortunately, the data quiet period from 1957 to 1961 was followed by are too incomplete to investigate the pattern. an increased level of activity for about 4 years, As Kelleher and Savino (1975) pointed out, a and there is some indication of an increased higher seismic activity than during the previous activity which may be considered to be a precur period is seen near the epicenter for about 3 sory swarm around 1956. years. Although the quiescence is not very clear, Thus, all three major earthquakes in the Alaska the seismic activity during the 15 year period Aleutian region have a common feature which is from 1935 to 1950 appears lower than the preceding not observed in the other regions discussed above. period. Figure 13 shows the space-time plot for the period 1960 to 1978. During this period only one event larger than 7.6 occurred (Ms= 7.8, 1971). However, the rupture zone for this event appears very small and no obvious pattern is seen on the scale shown here. Wyss et al. (1978) made a detailed analysis of this event and found a quiet period from 1962 to 1971 with a short period of increased activity in 1969. For a smaller event (Ms= 7.2) which occured in 1973, a period of .:'.': Aleutian d <::60 km M5 8 8>M~):7 7>M5 26 w - ----;--- ) -~ -~---T -~, ,)15- E ~ Pole Q (Aleutian) ~ • 0 ~ 0 E j______J_~- 52 56 60 64 68 .Pole (Alaska) 1000 km Time, year Fig. 11. Space-time plot of seismicity for the Fig. 9. Index map for Alaska and the Aleutians. rupture zone of the 1965 Rat Is. earthquake. For See caption for Fig. 2 for details. details, see caption for Fig. 3. KAN AMO RI 11 Petatlan earthquake, Meyer et al., 1980). Thus Kamchatka d ~60 km Ms2'8 8>M5 27 7>M5 26 N ,----,~,-~,--.,~---,-~,----.---,---,----,--,--,~~0,------, there appears to be a fairly systematic regional variation in the pattern for large earthquakes along various subduction zones. Precursory -y::------swarms and doughnut patterns are not always obvious in the space-time plot on regional scales, but detailed studies by other investigators have identified such patterns for some of the events. Asperity Model 40 Time, year 50 Fig. 12. Space-time plot of seismicity for the The coseismic motion on earthquake faults is rupture zone of the 1952 Kamchatka earthquake. often irregular as evidenced by complex wave forms For details, see caption for Fig. 3. of seismic waves generated by large earthquakes. This observation suggests that the fault plane is irregular either geometrically or in its physical quiescence may be identified (Figure 13). Wyss or mechanical properties. The strength of the and Habermann (1979) examined seismicity within contact zone between the two sides of the fault is 100 km radius from the epicenter of this event, larger at some places than elsewhere. Such places and concluded that a 50% decrease in seismicity of increased strength, either of geometrical rate began in mid 1967. This period of quiescence origin or of some other causes, are generally can be. identified also in the space-time plot called asperities. The importance of asperities shown in Figure 13. in various failure processes was recognized in laboratory studies (Byerlee, 1970; Scholz and Northern Japan Engelder, 1976) and the concept of asperity has been frequently used in seismology, either Figure 14 shows the seismicity pattern for explicitly or implicitly, to explain non-uniform Northern Japan (see Figure 4 for the location). se:lsmicity along fault zones (Wesson and Ellsworth, For this plot, the earthquake catalog compiled 1973; Bakun et al., 1980) complex events (Wyss and by the Japan Meteorological Agency (JMA) is used, Brune, 1967; Nagamune, 1971, 1978; Kanamori and and the earthquakes with MJMA > 5 are shown. The Stewart, 1978; Lay and Kanamori, 1980a; Das and largest earthquake during this time period is the Aki, 1977; Aki, 1979), seismic clustering 1968 Tokachi-Oki earthquake CMw = 8.2). The (Ishida and Kanamori, 1978, 1980), and certain activity during about 3 years just before the main aspects of seismicity patterns (Mogi, 1977, shock is considerably lower than during the pre Tsumura, 1979; Katsumata and Yoshida, 1980; ceding period. No obvious foreshocks were reported Lay and Kanamori, 1980b). for this earthquake, although Nagumo et al. (1968) Kanamori (1978) interpreted preseismic cluster recorded a number of very small events during ing of events near the main shock epicenter in several days before the mainshock by a ocean- terms of stress concentration around a strong bot tom seismograph which had been deployed in the asperity due to failure of weaker asperities epicentral area. surrounding it. Jones and Molnar (1979) explained Mogi (1969), Katsumata and Yoshida (1980) and observed time dependence of f oreshocks by using Habermann (1980) have made detailed analyses of a fault model with inhomogeneous contact planes seismicity patterns associated with this earth on which asperities fail by static fatigue. Ebel quake. Summary + + Both the results summarized in Table 1 and those described above show that many large earth quakes were preceded by a period of quiescence. Some events have a pronounced foreshock activity during a period of hours to weeks before the 25 280 mainshock. Examples include earthquakes in the c Kurile Islands (1963 and 1969), Nemuro-Oki (1973), 0 ~ and Chile (1960). Many large events in the 0 Alaska-Aleutians and the Kamchatka regions tend to have an increased seismic activity during 220 several years before the mainshock. However, some s events do not have obvious forshocks or preshocks; 1960 64 68 72 76 BO examples are the large earthquakes in Peru and Time, year Northern Japan. Large earthquakes along the Mid Fig. 13. Space-time plot of seismicity for the America Trench were of ten preceded by moderate Kamchatka region for the period 1960 to 1978. foreshocks (e.g., 1978 Oaxaca earthquake, 1979 For details, see caption for Fig. 3. 12 KANAMORI N N Japan d~60 km 226 \++-, f .. I I t t ~ *+t+ I t H: -t .t , If -/- + ~ -t H ! +I'-+ '1 + + I- I I 202 "+I' +* E t ' .cs:. I ++ ~ ~ j ' j[ _, Q + 178 (j) t u I c I t "' I I " 1 1 I E . -~+ (f) -.1--l- -++--+ 11 . .:-++ 0 154 ~t } +, s 130 + -rt 1940 50 Time, year 60 70 Fig. 14. Space-time plot of seismicity for Northern Japan obtained from the JMA catalog. Earthquakes smaller than 5 are not plotted. (1980) interpreted a foreshock-main shock-after ity, a Gaussian distribution with the average l shock sequence in the New Hebrides Islands in and the standard deviation L to represent the terms of loading and subsequent failure of variation of the strength (see Figure 15b). The asperities on the fault plane. details of the form of this distribution are un Mogi (1977) explained the pattern of temporal important for the present purpose. seismic quiescence and doughnut patterns in terms An asperity is introduced as a region within of a heterogeneous stress distribution on the this unit fault where the strength is higher than focal zone. As the tectonic stress increases, in the surrounding region (Figure 15a). Let ha be small earthquakes occur at high-stress spots in the strength of the subfaults located in the the focal region of an impending large earth asperity. We assume that ha follows another quake. When all the high stress spots are broken, Gaussian distribution with the average la the focal region becomes seismically quiet, but (l > h) and the standard deviation La. Thus the the activity in the surrounding region increases. ov~rall distribution of the strength of the sub Tsumura (1979) argued that various seismicity f aults on the fault surface is given by a bimodal ·patterns including foreshock patterns can be ex distribution as shown by Figure 15b. Although plained by introducing fault surf aces with the actual fault is more complex and may be more variable strength. Katsumata and Yoshida (1980) adequately represented by a multi-modal distribu proposed a model in which the coupling conditions tion, the bimodal distribution shown by Figure between the lithospheric plates control the 15b is introduced to isolate the effect of an temporal variation of seismicity patterns. asperity, and represents the simplest case. , Brune (1979) discussed the importance of We consider a loading stress o0 which varies heterogeneous stress distribution (asperity) on linearly in time: the fault plane as a controlling factor of various premonitory phenomena. (1) In order to explain the nature of various seismicity patterns presented in the previous 0 a where t is the time and 00 and are constants. section, we propose a very simple asperity model. When the stress at a grid point (i,j) exceeds the As mentioned above, this type of asperity model strength of the subfault there, the subfault fails has been used by various investigators in the and the stress there drops to 0. For simplicity, past; the main emphasis here is to parameterize we assume that once a subfault fails, the fault the model and relate it to the variation of ob surface there is decoupled (i.e., no healing takes served seismicity patterns. place), and the loading stress o 0 is held uniform Figure 15 illustrates the model. The rectangular ly by the remaining subfaults. Thus, under this box represents all or part of the rupture surf ace. assumption, the stress at the subfault at (i,j) is of a large earthquake, and hereafter is called a given by: unit fault. The unit fault is divided into smaller subfaults. Leth be the strength of the o(i,j) = o /[l-(~/N)] (2) subfaults. In general, the strength of the fault 0 surface is not uniform. Here we use, for simplic- where N is the total number of subfaults in the KAN AMO RI 13 s (l, j) Strength swarm-like activity, and 0 exceeds Z in a relatively short time. This stage (stage 2, Sa (1; ;) Sfrengfh ( As;)(zr/fy) Figure 16) corresponds to a precursory swarm. Loading Stress When this stage is passed, most subfaults outside (a) ~=%0 +at the asperity are broken so that few subfaults fail as the tectonic stress increases, resulting Stress at (t; ;) in seismic quiescence (stage 3, Figure 16). At this stage, stress is concentrated on the asperity N=IOO a =0.001 k bar /yr. OkfZ~~t>z±L!Ll'ZlLLlLJ.::Lbl'Zf:'.1A~ 1955 57 59 61 63 cr 00=0 s = 0.1 kbar Year c =0.3 Main Shock 2:=0.06, sa=0.45, L:a=0.025. :z::=o.06, ·vo.45, :z::a =0.015 2:=0.1, sa =0.45, Ia =0.025 c 40 d e Q z" 0 50 100 150 Time, year 2:=0.06, sa=0.35, L:a=0.015 f Nu~mber { Ia Controls foreshocks ~ Controls quiescence 2 ,2: 22:a : L.. Controls Precursory 0 50 100 150 0 S Sa Strength swarm Fig. 17. Comparison of temporal variation of number of events between the 1963 Kurile Is. sequence and the asperity models. asperity and the surrounding area is small (i.e., Among the important consequences of this model small Xa/X), the duration of the quiet period in terms of observable seismological parameters decreases (Figure 17f). This situation is are the clustering of events and increasing (in similar to that discussed by Tsumura (1979). A time) stress drops as schematically shown in large L results in a spread-out precursory swarm Figure 18. The overall loading stress increases activity as shown in Figure 17e. linearly with time. When a weaker subfault breaks, In the model presented here, we did not consider a stepwise increase in o occurs on other subf aults. any physical failure criterion, healing mechanism, Thus the stress drop of small earthquakes would anelastic time dependent mechanism or dynamic increase as a function of time. Since the in situ response of the medium. The primary purpose of condition is far more complex than is modeled here, this model is to provide the simplest possible one would expect a considerable variation in model with which the variation and complexity of stress drops at a given time. Nevertheless, if observed seismicity patterns can be reproduced. the asperity model is correct, the stress drop At present, our knowledge of the nature, distri should increase, on the average, as the final bution and regional variation of asperities on failure of the asperity approaches. the fault is too limited to fully test this model. Since foreshocks and preshocks occur as a result However, if this model proves useful for inter of local stress concentration within the asperity, preting seismicity patterns, more physical and they would be tightly clustered in space and dynamic models such as the one developed by probably have the same mechanism. As a result, Mikumo and Miyatake (1979) need to be introduced they would have approximately the same wave form, to study further details of seismicity patterns. if the magnitude of the events is about the same. The mechanism of the preshocks may be different Discussion from the background mechanism. Hamaguchi and Hasegawa (1975) studied a large Although it is not possible at present to test number of aftershocks of the 1968 Tokachi-Oki this model directly, it is desirable to investi earthquake having similar wave forms, and con gate whether this model is reasonable in the light cluded that these similar events occurred at of available seismological data other than approximately the same location under the same seismicity patterns. mechanical condition. Geller and Mueller (1980) KANAMORI 15 'CT (i, il suggest that the stress distribution near the epicenter of an impending earthquake can be very heterogeneous. Failure S (i,j)------~of asperity Spectrum and Stress Drops 0 (Main Shock) Whether foreshocks and preshocks are higher stress drop events than the earlier events is as yet unresolved. Although several examples Failure of Quiescence Foreshock indicate that foreshocks had higher frequency weak subfault content than other events, the quality of the High stress data was somewhat limited. s (i, j) Swarm drop Ishida and Kanamori (1980) analyzed the wave forms of small events which occurred near the -f~ Background Low stress seismicity epicentral area of the 1971 San Fernando and the drop 1952 Kern County earthquakes and found that foreshocks had, on the average, more high 0 Time frequency energy than the earlier events. Fig. 18. Temporal variation of stress cr at a Although these results are obtained with one subfault. A step-wise increase is caused by station data, they are based on very uniform failure of another weaker subfault. broad-band data (Wood-Anderson seismograms) collected over a very long period of time, 10 and 18 years for the San Fernando and the Kern County found four small earthquakes with similar wave earthquakes respectively. Since the characteris forms on the San Andreas Fault in Central tic time scale of earthquake loading cycles is at California, and suggested that they may represent least 10 to 100 years, it is extremely important breakings of an asperity where stress is to have a long-term data base for this kind of repeatedly concentrated and released. study. Some studies indicated, however, that the Waveform identification of foreshocks is not very straight forward. Tsujiura (1977) observed that some There are not many high-quality data sets of foreshocks had higher stress drops, but it was not wave forms for foreshocks or preshocks. Ishida always the case. Bakun and McEvilly (1979) and Kanamori (1978) found that the wave forms of examined several foreshocks and aftershocks of the all of the five events which occurred in the 1966 Parkfield, California earthquake and conclud proximity of the epicenter of the 1971 San ed that the difference between the foreshocks and Fernando earthquake during about 2 years before the aftershocks in terms of their frequency the main shock were nearly identical to each content is extremely subtle. other. Ishida and Kanamori interpreted this clustering and similarity of the wave forms in terms of stress concentration around an asperity whose failure led to the San Fernando earthquake. AMS A 0 62 km 1979 For the 1974 Haicheng earthquake, Jones et al. (1980) found that most foreshocks can be classi fied into two groups, each having approximately the same wave form. For the 1979 Imperial Valley, California earthquake, the main shock was preceded by three foreshocks whose wave forms are very similar to each other, as shown in Figure 19 (James Pechmann, Personal communication, 1980). These examples strongly indicate that a very 1979 I 71 I - tight clustering of activity which occurred more or less under the same stress preceded the main GLA Ll 0 6Bkm 1979 shock in the close proximity of the main shock 7 rupture zone. However, Tsujiura (1979) found that, while the wave forms of small events which preceded several earthquake swarms were very similar, those which occurred before several distinct main shock aftershock sequences varied considerably. Although the distinction between swarms and main shock aftershock sequences is not very obvious and Fig. 19. Wave forms of three foreshocks of the whether the wave form is similar or not depends on 1979 Imperial Valley earthquake recorded at three the frequency band used, Tsujiura's observations stations. 16 KANAHORI According to the numerical experiment described benefited from comments by anonymous reviewers above, a ratio of 4 of the stress drop of fore concerning several problems arising from non shocks to other events would be enough to yield uniformity of the catalog. the observed seismicity patterns. In view of the This work was partially supported by U.S. large error in the measurements of stress drops, Geological Survey Contract No. 14-08-0001-18371. particularly for small events, detection of Contribution No. 3508, Seismological Laboratory, possible temporal variation of stress drops of this California Institute of Technology, Pasadena, magnitude would be very difficult. Nevertheless, California 91125. with a better (wider dynamic range, broader frequency band with digital recording) instrumen References tation, it will eventually be possible to make more accurate stress drop measurements for Abe, K., Mechanisms and tectonic implications of monitoring stress variation on a fault plane. In the 1966 and 1970 Peru earthquakes, Phys. Earth this regard, Archambeau's (1978) approach (Ms/mb Planet. Interiors, 5, 367-379, 1972. ratio), House and Boatwright's (1978) analysis Aki, K., Characterization of barriers on an earth (use of local strong-motion record), and Mori's quake fault, J. Geophys. Res., 84, 6140-6148, (1980) study (use of short-period WWSSN data) would 1979. have a very important potential for monitoring Archambeau, C., Estimation of non-hydrostatic seismic gaps. stress in the earth, by seismic methods: Lithospheric stress levels along Pacific and Conclusion Nazca plate subduction zones, Proc. of Conference VI; Methodology for identifying Although various seismicity patterns have been seismic gap and soon-to-break gaps, U.S. reported for many earthquakes, the nature of the Geological Survey Open-File Report 78-943, patterns varies substantially from event to e\ent. p. 47-138, 1978. A global survey of seismicity patterns before Bakun, W. H., and T. V. McEvilly, Are foreshocks major subduction-zone events indicates significant distinctive? Evidence from the 1966 Parkfield regional variations of the nature of seismicity and the 1975 Oroville, California sequence, patterns. It appears that the heterogeneity and Bull. Seismol. Soc. Am., 69, 1027-1038, 1979. the complexity of the individual fault zones are Bakun, W. H., R. M. Stewart, C. G. Bufe, and responsible for the observed variations. The S. M. Marks, Implication of seismicity for fundamental physical process leading to an earth failure of a section of the San Andreas fault, quake may be common to most events, but its Bull. Seismal. Soc. Am., 70, 185-201, 1980. manifestation as seismicity patterns may vary Brady, B. T., Theory of earthquakes, Pageoph., significantly depending upon the regional and 114, 1031-1082, 1976. local variations of the fault-zone structure. A Brady, B. T., Anomalous seismicity prior to rock very simple asperity model is presented in this bursts: Implications for earthquake prediction, paper to reproduce this situation. The basic Pageoph., 115, 357-374, 1977. physical process in this model is gradual stress Brune, J. N., Implications of earthquake trigger concentration at an asperity on the fault zone. ing and rupture propagation for earthquake This stress concentration followed by failure of prediction based on premonitory phenomena, the asperity manifests itself as a variety of J. Geophys. Res., 84, 2195-2198, 1979. seismicity patterns depending upon the strength Byerlee, J. D., Static and kinetic friction of and heterogeneity of both the asperity and the granite under high stress, Int. J. Rock Mech. surrounding area. Although it is not presently Min. Sci., 7, 577-582, 1970. possible to test this model directly, it serves Das, S., and K. Aki, Fault planes with barriers: as a useful working model for a better under A versatile earthquake model, J. Geophys. Res., standing of earthquake precursors. 82, 5658-5670, 1977. Although seismicity patterns provide important Ebel, J. E., Source processes of the 1965 New information on the earthquake preparatory process, Hebrides Islands earthquakes inferred from its usefulness for prediction purposes is somewhat teleseismic waveforms, Geophys. J., in press, limited because of the substantial variations 1980. from event to event. However, the asperity model Engdahl, E. R., and C. Kisslinger, Seismological would suggest use of other seismological data such precursors to a magnitude 5 earthquake in the as wave forms, spectra and mechanism of preshocks central Aleutian Islands, J. Phys. Earth, 25, for monitoring the state of stress on the fault 5243-5250, 1977. plane, the key information for earthquake predic Evison, F. F., Fluctuations of seismicity before tion. major earthquakes, Nature, 266, 710-712, 1977a. Evison, F. F., The precursory earthquake swarm, Acknowledgments. I thank Carl Johnson and Phys. Earth Planet. Int., 15, 19-23, 1977b. Bernard Minster for teaching me how to plot earth Evison, F. F., Precursory seismic sequences in quakes by computer. I also thank Larry Ruff, New Zealand, N.Z. Journal of Geology and Thorne Lay, Don Anderson and James Pechmann for Geophysics, 20, 129-141, 1977c. many helpful comments on the manuscript. I Fedotov, s. A., Regularities of the distribution KANAMORI 17 of strong earthquakes in Kamchatka, the Kurile aspects of the Guatemala earthquake of February islands and northeastern Japan, Trans. Acad. 21, 1976, J. Geophys. Res., 83, 3427-3434, Sci. USSR, Inst. Phys. Earth, 36, (203), 66-93, 1978. 1965 (in Russian). Katsumata, M., and A. Yoshida, Change in Fuis, G. S. and A. G. Lindh, A change in fault seismicity and development of the focal region, plane orientation between foreshocks and Papers in Meteorology and Geophysics, 31, 15-32, aftershocks of the Galway Lake earthquake, 1980. ~ ML = 5.2, 1975, Mojave Desert, California Keilis-Borok, V. I., L. Knopoff and I. M. Rotvain, (abstract), Tectonophysics, 52, 601-602, 1979. Burst of aftershocks, long-term precursors of Geller, R. J., and C. S. Mueller, Four similar strong earthquakes, Nature, 283, 259-263, 1980. earthquakes in central California, Geophys. Kelleher, J. A., Space-time seismicity of the Res. ·Lett., 7, 821-824, 1980. Alaska-Aleutian seismic zone, J. Geophys. Res., Habermann, R. E., Precursory seismicity patterns: ]J_, 5745-5756, 1970. Stalking the mature seismic gap, This volume, Kelleher, J., and J. Savino, Distribution of 1980. seismicity before large strike slip and thrust Hamaguchi, H., and A. Hasegawa, Recurrent type earthquakes, J. Geophys. Res., 80, occurrence of the earthquakes with similar wave 260-271, 1975. forms and its related problems, J. Seismal. Soc. Kelleher, J., L. Sykes, and J. Oliver, Possible Japan, 28, 153-169, 1975 (in Japanese). criteria for predicting earthquake locations House, L., and J. Boatwright, Investigation of and their application to major plate boundaries two high stress-drop earthquakes in the of the Pacific and the Caribbean, J. Geophys. Shumagin seismic gap, Alaska (abstract), EOS, Res., 78, 2547-2585, 1973. 59, 1124, 1978. Khattri, I., and M. Wyss, Precursory variation of Imamura, A., On the seismic activity of central seismicity rate in the Assam area, India, Japan, Jap. J. Astron. Geophys., 6, 119-137, Geology, 6, 685-688, 1978. 1928. Kristy, M. J., and D. W. Simpson, Seismicity Inouye, W., On the seismicity in the epicentral changes preceding two recent central Asian region and its neighborhood before the Niigata earthquakes, Submitted to J. Geophys. Res., earthquake, Kenshin Jiho, 29, 31-36, 1965 (in 1980. Japanese). Lay, T., and H. Kanamori, Earthquake doublets in Isacks, B. L., and M. Barazangi, Geometry of the Solomon Islands, Phys. Earth Planet. Int., Benioff zones: Lateral segmentation and down 21, 283-304, 1980a. ward bending of the subducted lithosphere, in Lay, T., and H. Kanamori, An asperity model of Talwani, M. and W. C. Pittman (eds.), Island great earthquake sequences, in this volume, Arcs, Deep Sea Trenches and Back-Arc Basins, 1980b. Maurice Ewing Series I, 99-114, 1977. McCann, W., R. S. P. Nishenko, L. R. Sykes, and Ishida, M., and H. Kanamori, The foreshock activ J. Kraus, Seismic gaps and plate tectonics: ity of the 1971 San Fernando earthquake, Seismic potential for major boundaries, Pageoph., California, Bull. Seismal. Soc. Am., 68, 117, 1087-1147, 1980. 1265-1279, 1978. McEvilly, T., W. Bakun, and K. Casady, The Park Ishida, M., and H. Kanamori, Temporal variation of field, California earthquake of 1966, Bull. seismicity and spectrum of small earthquakes Seismal. Soc. Amer., 57, 1221, 1967. ~~- preceding the 1952 Kern County, California McNally, K., Patterns of earthquake clustering earthquake, Bull. Seismal. Soc. Am., 70, 509-527, preceding moderate earthquakes, central ann 1980. southern California, abstract, EOS, 58, 1195, Johnson, C. and L. K. Hutton, The 15 October, 1979 1977 (full text in preprint form). Imperial Valley earthquake: A study of after Meyer, R. P., w. D, Pennington, L.A. Powell, shocks and prior seismicity, to appear in the W. L. Unger, M. Guzman, J. Havskov, s. K. Singh, U.S. Geological Survey Professional Paper on the C. Valdes, and J. Yamamoto, A first report on 1979 Imperial Valley Earthquake, 1980. the Petatlan Guerrero, Mexico earthquake of Jones, L., and P. Molnar, Frequency of foreshocks, 14 March 1979, Geophys. Res. Lett., 7, 97-100, Nature, 262, 677-679, 1976. 1980. Jones, L. M., and P. Molnar, Some characteristics Mikumo, T., and T. Miyatake, Earthquake sequences of foreshocks and their possible relationship on a frictional fault model with non-uniform to earthquake prediction and premonitory slip on strengths and relaxation times, Geophys. J. R. faults, .J. Geophys. Res,, 84, 3596-3608, 197:L Astron. Soc., 59, 497-522, 1979. Jones, L., Wang Biquan, and Xu Shaoxie, The Mitsunami, T., and K. Kubotera, On the activity Haicheng foreshock sequence, This volume, 1980, of the earthquake swarm in the northern part of Kanamori, H., Use of seismic radiation to infer Aso caldera, 1975 - Interpretation of hypo source parameters, Proc. of Conference III; central migration, J. Seismal. Soc. Japan, 30, Fault mechanics and its relation to earthquake 73-90, 1977 (in Japanese). prediction, U.S. Geological Survey Open-File Mizoue, M., M. Nakamura, Y. Ishiketa, and N. Seto, Report 78-380, p. 283-317, 1978. Earthquake prediction from micro-earthquake Kanamori, H. and G. S. Stewart, Seismological observation in the vicinity of Wakayama city, 18 KANAMORI Northwestern part of the Kii peninsula, central on spatio-temporal pattern of preceding seismic Japan, J. Phys. Earth, 26, 397-416, 1978. activity and main shock relocation, Geofisica Mogi, K., Some discussions on aftershocks, fore Internacional, 17, 109-126, 1977-1978. shocks and earthquake swarms - The fracture of Santo, T., Shock sequences of the southern Kurile a semi-infinite body caused by an inner stress Islands from October 9 to December 31, 1963, origin and its relation to the earthquake Bull. Int. Inst. Seismal. Earthquake Eng., 1, phenomena, Bull. Earthquake Res. Inst. Tokyo 33, 1964. Univ., 41, 615-658, 1963. Scholz, C. H. and J. T. Engelder, The role of Mogi, K., Some features of recent seismic activity asperity indention and ploughing in rock in and near Japan, (1), Bull. Earthquake Res. friction, I. Asperity creep and stick slip, Inst. Tokyo Univ., 46, 1225-1236, 1968a. Int. J. Rock Mech. Min. Sci., 13, 149-154, 1976. Mogi, K., Source locations of elastic shocks in Sekiya, H., Anomalous seismic activity and earth the fracturing process in rocks (1), Bull. quake prediction,J. Phys.Earth, 25, 585-593, 1977. Earthquake Res. Inst. Tokyo Univ., 46, 1103-1125, Sykes, L. R., Aftershock zones of great earth 1968b. quakes, seismicity gaps, and earthquake Mogi, K., Some features of recent seismic activity prediction for Alaska and the Aleutians, J. in and near Japan, (2) Activity before and after Geophys. Res., 76, 8021-8041, 1971. ~ great earthquakes, Bull. Earthquake Res. Inst. Tanaka, K., T. Ohuchi and T. Santo, Formation Tokyo Univ., 47, 395-417, 1969. patterns of seismic gaps before and after great Mogi, K., Seismic activity and earthquake predic shallow earthquakes, Program and abstracts, tions, in Proceedings of the Symposium on Earth Seismol. Soc. Jap. Meeting, No. 1, 24, 1979 (in quake Prediction Research, 203-214, 1977 (in Japanese). Japanese). Tsujiura, M., Spectral features of foreshocks, Mogi, K., Two kinds of seismic gaps, Pageoph, 117, Bull. Earthquake Res. Inst. Tokyo Univ., 52 1-15, 1979. 357-371, 1977. Mori, J., Effective stress drops of moderate Tsujiura, M., The difference between foreshocks earthquakes in the eastern Aleutians (abstract), and earthquake swarms, as inferred from the EOS, 61, 294, 1980. similarity of seismic waveform, Bull. Earth Nagamune, T., Source regions of great earthquakes, quake Res. Inst. Tokyo Univ., 54, 309-315, 1979 Geophys. Mag., 35, 333-399, 1971. (in Japanese). Nagamune, T., Tectonic structures and multiple Tsumura, K., A model for temporal variation of earthquakes, Zisin: J. Seismal. Soc. Japan, 31, seismicity and precursors, abstract of the 457-468, 1978 (in Japanese). paper presented at the Spring meeting of the Nagumo, S., H. Kobayashi and S. Koresawa, Seismological Society of Japan, Tokyo, p. 159, Foreshock phenomena of the 1968 Tokachi-Oki 1979. earthquake observed by ocean-bottom seismographs Utsu, T., Seismic activity and seismic observation off Sanriku, Bull. Earthquake Res. Inst. Tokyo in Hokkaido in recent years, Rep. Coord. Univ., 46, 1355-1368, 1968. Committee Earthquake Pred., 2, 1-2, 1970 (in Nersesov, I. L., A. A. Lukk, V. S. Ponomarev, Japanese). T. G. Rautian, B. G. Rulev, A. N. Semonov, and Wesson, R. L. and W. L. Ellsworth, Seismicity I. G. Simbireva, Possibilities of earthquake preceding moderate earthquakes in California, prediction, exemplified by the Garm area of the J. Geophys. Res., 78, 8527-8546, 1973. Tadzlink SSR, in Earthquake Precursors, Acad. Wu, Kai-tong, Ming-sheng Yue, Huan-ying Wu, Shin Sciences, USSG, eds. M. A. Sadovsky, I. L. ling Chao, Hai-tong Chen, Wei-Qong Huang, Kang Nersesov, and L. A. Latynina, 72-99, 1973. yuan Tien and Shau-de Lu, Certain characteris Ohtake, M., Search for precursors of the 1974 Izu tics of the Haicheng earthquake (M = 7.3) Hanto-Oki earthquake, Japan, Pageoph, 114, sequence, Chinese Geophysics, 1, 289-308, 1978. 1083-1093, 1976. Wyss, M. and J. Brune, The Alaska earthquake of Ohtake, M., T. Matumoto, and G. V. Latham, 28 March 196~: A complex multiple rupture, Seismicity gap near Oaxaca, Southern Mexico as Bull. Seismol. Soc. Am., 57, 1017-1023, 1967. a probable precursor to a large earthquake, Wyss, M. and R. E. Habermann, Seismic quiescence Pageoph, 115, 375-385, 1977. precursory to a past and a future Kurile Island Ohtake, M., T. Matumoto, and G. V. Latham, earthquakes, Pageoph, 117, 1195-1211, 1979. Patterns of seismicity preceding earthquakes in Wyss, M., R. E. Habermann, and A. C. Johnston, Central America, Mexico, and California, Proc. Long term precursory seismicity fluctuations, Conference Methodology for identifying seismic in Methodology for Identifying Seismic Gaps and gaps and soon-to-break gaps, U. S. Geological Soon-to-Break Gaps, Conference VI Nat. Eq. Raz. Survey Open-File Report 78-943, 585-610, 1978. Red. Prog., 869-894, 1978. Ponce, L., K. C. McNally, V. Sumin de Portilla, Yamashina, K., and Y. Inoue, A doughnut-shaped J. Gonzalez, A. Del Castillo, L. Gonzalez, pattern of seismic activity preceding the E. Chael, and M. French, Oaxaca Mexico earth Shimane earthquake of 1978, Nature, 278, 48-50, quake of 29 November 1979: A preliminary report 1979. KANAMORI 19