ML20150G071

Cascadia Subduction Zone:Evaluation of Earthquake Potential Implications to Washington Nuclear Plant 3

ML20150G071
Person / Time
Site:	Satsop
Issue date:	06/30/1988
From:	WASHINGTON PUBLIC POWER SUPPLY SYSTEM
To:
Shared Package
ML20150G047	List: ML20150G044 ML20150G061 ML20150G071
References
NUDOCS 8807190196
Download: ML20150G071 (216)

Text

. . __ . _ . - _ _ - . . . _. . ._ _ _ _ _ _ .

CASCADIA SUBDUCTION ZONE An Evaluation of the Earthquake Potential and implications to WNP-3 1

l l

Washington Public Power Supply System )

Response to l NRC Questions 230.1 and 230.2 June 30,1988 A PNU

l l

1 I

QUESTIONS 230.1 and 230.2 (SRP 2 5,2,g, g g ,4, 2 5 2.6)

TABLE OF CONTENTS P_ag I. INTRODUCTION 2 II. CAPABILITY OF PO'".NTIAL SUBDUCTION ZONE SOURCES 3 The Juan de Fuca Plate and the Cascadia Subduction Zone 4 Geographic Definition 4 Present Juan de Fuca (JdF)/ North American (NA) Plate Interaction 6 Special Characteristics of the Juan de Fuca Plate and the Cascadia Zene 8 Seismicity 8 Geometry of the Subducted Plate 12 Geologic factors 14 Predicting the Magnitude of Subduction Zone Earthquakes Using Plate Convergence Rate and Age of the Subducting Oceanic Plate 15 Validity of Heston and Kanamori's (1984) Analysis as Applied to the Cascadia Subduction Zone 19 Heaton and Kanamori's Estimate of Plate Age at Trench (t) for the Juan de Fuca Plate 19 Very Young Occanic Plates and Seismic Coupling 20 Estimating the Interplate Seismic Potential of Subduction Zones: An Alternative 23 Physical Explanations for Aseismic Convergence, Cascadia Subduction Zone 25 Introduction 25 The Role of Sediments in Influencing Aseismic Subduction 26 Clay-Rich Marine Sediments 26 High Temperatures of Cascadia Basin Seditents 28 High Fluid Pressures 28 Stable Sliding Along the Juan de Fuca/ North American Plate Interface 30 This document is a response to NRC Questions 230.1 and 230.2 (SRP 2 5 2.2, 2 5 2.4, 2 5 2.6) for the WNP-3 sito, Satsop, Washington.

The following individuals contributed to the development of this response:

Dr. Robert R. Youngs Geomatrix Consultants Dr. Kevin J. Coppersmith Geomatrix Consultants Dr. Gregory A. Davis University of Southern California Mr. Donald O. West Golder Associates Dr. Robert S. Crosson University of Washington Mr. David D. Tillson Consultant Mr. William A. Kiel Washington Public Power Supply System l

i l

I

)

1 l

1 1

TABLE OF CONTENTS (continued)

Pagg Comparisons of the Cascadia Subduction Zone With Other Zones 32 l

Strain Classification of Upper Plates of Subduction Zones 33 l

Jarrard's (1986) Classification 34 Cascadia Zone Strain Class 35 Southwestern Japan (Nankai Trough) 36 Southern Chile (Southern Nazca) 39 Southwestern Colombia-Northern =ost Ecuador 41 Rivera 43 Northern Cocos (Southwest Mexico) 46 Southernmost Chile (Tierra del Fuego) 48 Conclusions from Subduction Zone Comparisons 48 Geodetic and Geologic Evidence for and Against Great Cascadia Zone Earthqu akes 49 Strain and Stress Within the North American Plate. Pacific Northwest 49 Coastal Geomorphology and Neotectonics 55 Evidence for Holocene Coastal Subsidence 57 Washington 57 Oregon 58 Discussion 59 Evidence for Holocene Coastal Uplift 63 Discussion 63 Turbidite Deposits at the Base of the Continental Slope 68 Holocene Faulting Near the Mendociro Triple Junction. l Northern California 69 Indian Legends Regarding Pacific Northwest Seismic Pheno =ena 71 1

Great Floods (= Tsunamis?) 71  !

Great Earthquakes 72 Conclusions Regarding Capability of Subduction Zone Sources 73 I

-it-

TABLE OF CONTENTS (continued)

Page III. MAXIMUM EARTHQUAKE MAGNITUDES 77 l Intra-Slab Source 77 l Plate Interface Source 78 l

l Rupture Length and Seg=entation of the Cascadia Subduction l

Zone 78 Evidence for Seg=entation of the Cascadia Subduction Zone 79 Conclusions Regarding Segmentation of the Cascadia Subduction Zone 84 Rupture Width 86 Maxi =u= Magnitude Esti= ate 89 i

IV. GROUND MOTION ATTENUATION RELATIONSHIPS FOR SUBDUCTION ZONE i SOURCES 91 )

i Strong Motion Data 92 l Attenuation Relationships for Peak Horizontal Acceleration on Rock 94 Attenuation Relationships for Spectral Velocity 99 l Comparison with Published Empirical Attenuation Relationships 102 Vertical Ground Motions 103 Ground Motions for Earthquakes of Magnitude Greater than 8 105 Attenuation Relationships for M > 8 Earthquakes 109 Comparison with other Numerical Modelling Resu);ts 110 V. SITE GROUND MOTIONS FROM POSTULATED SUBDUCTION ZONE EARTHQUAKES 111 Comparison of SSE Spectra with Response Spectra for Interface Earthquakes 112 Co=parison of SSE Spectra with Response Spectra for Intraslab Earthqu akes 113 l

VI. CONCLUSIONS 114 s VII. REFERENCES 116

-111-

1 l

l QUESTIONS 230.1 and 230.2 (SRP 2 5 2.2, 2.5 2.4, 2 5 2.6) l Question 230.1 1

The work by Ruff and Kanamori (1980) and others appears to support the view l that the subduction of the Juan 4: Fuca plate creates a potential for large magnitude ear thquakes in the subduction zone beneath WNP-3 In addition:

a) Kanamori (1983) has publishe a equation relating the age of the sub- l ducting plate, convergence selocity, and the largest expected magnitude ,

event. Does this equation apply to the Juin de Fuca plate and if not, why not? Alternatively are there other convincing models that allow the estimation of the magnitude of subduction zone earth-quakes under the site to values lower than would be predicted by the Kanamori (1983) relationship. l l

b) Are there specific examples of aseismic subduction zones which share the following features with the Juan de Fuca subduction zone: young sub-ducted lithosphere, low convergence rate, no back-arc basin, similar maximum depths of seismicity, shallow oceanic trench, lcw free-air l gravity anomaly, small variation in surface topography of the subducted plate and, particularly, complete seismic quiescence down to the magni- i tude 5 level? '

c) Crustal uplift rates of approximately 2mm/yr were observed in the region from 120 to 220 km inland of the Nankai Trough for the 50 years pre-ceding the 1944, M=8.0 Tonankai and 1946, M=8.2 Nankaido earthquake. l Why shouldn't the crustal uplift and NE-compressive strain reported by Savage (1981) for western Washington be considered consistent with a similar preseismic deformation? How is the Juan de Fuca subduction zone any different from the subduction zone in the Nankai Trough and the subduction zone associated with the Rivera plate?

d) What is the magnitude of the largest shock in the plate or along the plate interface that could occur beneath the site without exceeding the SSE acceleration? Specify the attenuation and dirtance used in the discussion. Assign a confidence level to your magnitude estimate, or estimate a range of magnitudes and corresponding confidence levels.

Question 230.2 a) What is the magnitude of the mt ximum credible earthquake that could occur on the subduction zone benea;h the WNP-3 site? This magnitude may be described by a range of values with associated probabilities and a best-estimate value, b) Estimate response spectra at the site assuming the occurrence of the maximum subduction zone earthquake beneath the site, for both vertical

- c .- - .- -

l and horizontal components of motion. _Specify all assusmptions about hypocentral depth and attenuation. The spectra should be' calculated on a deterministic basis. If,.in addition, probabilistic response spectra are presented, describe the t eatment of uncertainty in the magnitude of the maximum earthquake, the attenuation relation, and the hypocen-tral depth. Justify the SSE spectrum in light of your deterministic (and probabilistic) results, for both vertical and horizontal ground motion.

Response

I. INTRODUCTION Questions 230.1 and 230.2 relate to the issue of the earthquake potential' of the Cascadia subduction zone and the possible associated ground motions at the WNP-3 site. We view these questions as the deterministic counterpart to Q230.6, which asked for the probability of exceeding the SSE. Our response to Q230.1 and 230.2 is an integrated response that addresses all of the critical elements of the question: capability '(earthquake potential or activity), source location, maximum magnitude, ground motion attenuation, and site ground motions. We make the assumption that questions 230.1 and 230.2 are being asked and a response is being given in order to evaluate the adequacy of the previously-defined SSE, particularly in light of'the possi-bility that subduction-related earthquakes might be credible. Therefore, we view our response as one mechanism (others include probabilistic seismic hazard analysis) by which the NRC is evaluating the design values in light of alternative tectonic hypotheses having various levels of credibility in the scientific community.

The format of the responses to Q230.1 and 230.2 follows the key elements of a deterministic ground motion analysis, namely: capability, maximum magni-tude and ground motions. The integrated response given here is quite lengthy for the following reasons:

1) the issues related to the above elements are very complex. For example, the capability of the interface between the Juan de Fuca and North American plates is difficult to assess because it has been historically aseismic and because it is buried beneath the continental margin and is not amenable to fault-specific geologic analysis.

2) Question 230.1 was submitted by the Nuclear Regulatory Commission to the Washington Public Power Supply System in 1983, at a time of emerging speculation among the earth science community that subduction of the Juan de Fuca oceanic plate along the Cascadia (or Juan de Fuca) subduc-tion zone might be seismogenic. Since that submittal, much has been learned and written regarding the potential for a great (M > 8) subduc-tion zone earthquakes along the interface between the Juan de Fuca and l

the North American plates. The body of available knowledge pertaining to

! the basic question -- i.e., the "potential for large magnitude earth-quakes in the subduction zone beneath WNP-3" (230.1)'-- is now so great that specific answers to questions 230.la, b, and c (above) would be too narrow in scope.

3) Ground motion attenuation relationships for various subduction zone sources (e.g. , plate interface ar.d intra-slab sources) have been updated in light of recent data and an extensive discussion is given. Approaches to estimating ground motions have entailed both empirical and numerical methods.

4) Despite the inability of deterministic approaches to incorporate uncer-tainty, we are providing here a full discussion of uncertainties in order to specify the technical basis for the deterministic characteris-tics finally selected. By doing so, we intend to assist the reader in understanding the scientific issues and levels of conservatism involved.

II. CAPABILITY OF POTENTIAL SUBDUCTION ZONE SOURCES i l

l In this section of the response, we address the potential earthquake sources associated with the Cascadia subduction zone and we evaluate their capabil-ity. Because of the rather unique type of potential earthquake sources that we are dealing with here, 10CFB Part 100. Appendix A criteria are somewhat difficult to apply. For example, the plate interface has not been associated with seismicity in the historical period. Although young (late Quaternary) fe lting is essociated with plate convergence in the offshore region, this 63fctmation is occurring in the young, v-'er-saturated sediments of the outer l accretionary wedge and is seismically qu.escent based on the historical record and comparisor.s with other subduction zones. Obviously, the plate interface is not exposed at the land surface, which would allow for fault-specific types of evaluations of capability.

1 1

Despite these problems, we here broaden the concept of "capability" to mean j seismogenic or active within the present tectonic stress regime and we evaluate capability using any indications that a potential source can y, , , _ --~-- ., ,

r-- --- . -, y . _

generate earthquakes. Within this context, a wide range of' scientific arguments can be brought to bear on the problem. These arguments are discussed below.

As a note of explanation to the reader, we will in this response be using the terminology "seismically quiescent" and "aseismic" to describe the plate interface. "Seismically quiescent" means that the interface has not been associated with earthquakes during the historical period of observation but does not necessarily imply anything about longer-term behavior. "Aseismic" means that the long-term deformational behavior is sich that earthquakes are not generated by differential slip across the plate interface.

The Juan de Fuca Plate and the Cascadia Subduction Zone Geographic Definition The Juan de Fuca oceanic plate is a remnant of the formerly extensive Farallon plate, one of the major eastern oceanic plates of the Pacific Basin in Mesozoic and Cenozoic time (cf., Engebretson and others,1984) . Most portions of the Farallon plate have been subducted beneath western North America, but portions of its spreading ridge still survive as the Juan de Fuca. Explorer, and Gorda Ridges offshore of the Pacific Northwest (Fig. 1).

The latter two ridges lie closer to the continental margin than the Juan de Fuca and produce small, youthful (= hot) "platelets" respectively the Explorer /Winona, west of northern Vancouver Island, and the Gorda (or South Gorda), wests.? northernmost California and southwestern Oregon. Both platelets exhibit considerable internal deformation and seismicity (unusual features for most oceanic plates) and their interactions with the continental margin along the Cascadia subduction zone are complex. For example,'a l subducted portion of the Gorda plate is inferred to underlie northwestern California (Jachens ane Griscom, 1983), but active subduction of the plate i 1

appears to have given way to internal, offshore convergent deformation (Riddihough, 1980; 1984; Wilson, 1986). Earthquakes within the southern half of the Gorda plate produce fault plan solutions for strike-slip fault-ing with the axis of principal compression being horizontal and north-south l

l

. _ . _ . ~ _____ _ ._. .__ _

in orientation (Wilson, 1936). similarly, subduction of the very slow moving xplorer /Winona platelet into the asthenosphere has also ceased, according to Riddihough (1984), although it and North America still have a convergent j

relationship to each other. Tectonic overriding of the Explorer platelet by

~

! the more rapidly moving North American plate (Vancouver Island) appears to be occurring (Riddihough, 1984), perhaps by as much as 70 km (ibid.). This tectonic scenario may lead to an underplating of northern Vancouver Island by the overriden platelet. Riddihough (1984, p. 6992) believes that such tectonic overriding produces a higher degree of oceanic plate (Explorer)-

North Americar. . . 'te "stress coupling" than is seen farther south where the Juan de Fuca plate is still being subducted beneath North America. He cites differences in seismicity, topography, and recent vertical movement histories between northern and southern portions of Vancouver Island to support his contention that the northern Explorer platelet is behaving independently of the more southerly Juan de Fuca plate.

The atypical behavior of the Gorda and Explorer /Winona platelets, south and north of the intervening Juan de Fuca plate, establishes a logical basis for seg=entation of the Cascadia "subduction" zone. The central and longest segment of the zone -- and apparently the only segment along which active subduction continues -- separates the Juan de Fuca and North American plates.

Subsequent references in this response to the Cascadia subduction zone refer only to the Juan de Fuca-North American plate boundary that underlies south-ern Vancouver Island, western Washington and most of western Oregon. It extends approximately 900 km from the Nootka fault zone on the north to the northern boundary of the Gorda plate. south of Cape Blanco. Figure 1 illus-trates the location of that plate boundary according to Riddihough (1984).

Intra-Gorda plate faulting and seismicity is pronounced south of the boundary where it meets the base of the continental slope (Fig. 2). A concluding discussion in this reponse of possible segmentation within the Cascadia sub-duction zone refers only to the Juan de Fuca-North American plate segment; it is not concerned with segmentation of the greater Cascadia zone based on the differing behavior of the Explorer /Winona, Juan de Fuca, and Gorda plates.

l l

Present Juan de Fuca (JdF)/ North Am~ e rican (NA) Plate Interaction There is widespread agreement within the earth sciences community that the Juan de Fuca plate is currently being subducted-beneath southern Vancouver Island, western Washington, and western Oregon, although the current rate of subduction is not well constrained. Direct physical evidence for present

~

convergence between the two plates is best documented by the folding, fault-ing, and dewatering of latest Pleistocene and Holocene (7) sediments (less than 0 3 Ma) in the vicinity of the boundary between the Cascadia basin and the base of the continental slope (Kulm, 1983, p. 31; Kulm, L. D. and others, eds., 1984, Atlas 1, Ocean Margin Drilling Program). Seismic reflection lines across the Nitinat Fan off northern and central Washington reveal youthful features and multiple faults that "appear to extend to the surface cutting the most recently deposited sediments of the fan" (Fig. 3; Kulm, 1983, p. 17; Kulm and others, 1984). Kulm (op. cit.) believes that such deformation must have occurred within the "past few thousand cr, at the O

most, tens of thousands of years." On land, marine terrace deformation, leveling studies, and tide gauge data indicate Holocene and/or historic distortion of the continental margin compatible with ongoing subduction of the JdF plate (e.g. , Ando and Balazs,1979; Adams,1984) .

Adams (1984) in a review and enalysis of diverse geodetic measurements con-cludes that contemporeneous shortening of the continental margin by permanent deformational mechanisms (folding, faulting, and tilting) may approximate 25 mm/yr, 80% of which probably occurs within 40 km of the offshore plate bound-ary (the base of the continental slope). This conclusio., if correct, pro-vides a minimum rate of latest Cenozoic convergence between the Cascadia Basin of the Juan de Fuca plate and North America (a component of total convergence could occur without structural manifestation in upper plate sediments and rocks).  ;

1 Most investigators believe that the direction of convergence' between the JdF l and NA plates is approximately N50'E, and that the rate of present conver-gence between the two plates, although somewhat uncertain, has been decreas-ing thrcughout latest Cenozoic time (Riddihough, 1977, 1984; verplanck and

Duncan, 1987; spence, 1987). Rates of JdF/NA convergence across the Cascadia subduction zone are determined indirectly, for example by relying on Pac-NA and Pac /JdF relative motions (the latter requiring a correct assessment of present spreading rates on the Juan de Fuca ridge ) or determining plate l

motions relative to "fixed" global hot spots. The most thorough analyses of JdF/NA plate interactions are those of Riddihough (1984) and Nishimura and others (1984); both analyses propose that rates of convergence are slower for the Oregon portion of the Cascadia zone than for the Washington portion.

Riddihough (1984, Figs.11,12, 46*N) has estimated that the rate of oblique convergence between the two plates 500,000 years ago and at the latituce of the Columbia River was 42 mm/yr (* 7 mm/yr). This convergence rate is equivalent to a rate perpendicular to (i.e., orthogonal to) the continental margin between 30 and 35 mm/yr. (op. cit., Fig. 12). Nishimura and others determine oblique rates at 48'N (NW Washington) and 45'N (N Oregon) to be 43 19 and 38

• 19 mm/yr, respectively. They state that the orthogonal rate of subduction between 42*N and 49'N ranges from approximately 20 to 40 mm/yr. Verplanck and Duncan (1987) estimate that the orthogonal rate of present day convergence at Cape Blanco, Oregon, is only 16 mm/yr, less than half the rate for the Oregon coast reported by Nishimura and others (1984).

In this response to Question 230.1, oblique and orthogonal JdF/NA convergence rates at the latitude of the Columbia River are taken as 40 and 34 mm/yr, respectively. The latter or orthogonal value, taken from Jarrard's compre-hensive study of subduction zones (1986), is probably the most important in i

terms of what the western edge of the North American plate "feels" in its l 1

interaction with the subducting oceanic plate. It is generally believed that oblique subduction along a continental margin is physically resolved into two components: orthogonal and trench parallel (Jarrard, 1986a). The orthogonal component may lead to compressional deformation with the overlying 1

The contemporary spreading rate along the JdF ridge is unknown, although direct observations of its axial rift by submersible vessel ("Alvin", Normark and others, 1987) and by side-looking sonar imagery ("Sea MARC", e.g. Kappel and Normark, 1987) clearly reveals ongoing spreading activity. Because there have been no magnetic reversals in the past 700,000 years, the rate of spreading generally assigned to the ridge (ca. 60 mm/yr) is an average rate for the past 700 Ka.

l plate, whereas the trench-parallel component may lead, if large enough, to strike-slip deformation between the trench and volcanic arc. Jarrard (1986) uses convergent rates perpendicular to trenches in his. intensive global compaiisons of 39 subduction zones, including the Cascadia zone (34 mm/yr).

As discussed above, the contemporary rate of convergence may be significantly

)

l less than 34 mm/yr; it is unlikely to be higher. Riddihcugh (1984) calcu-l lates that the velocity of JdF/NA convergence decreased by about 60% between ]

6.5 and 0 5 Ma. He hints that continued slowing since 0 5 Ma is likely, but he does not extrapolate declining rates of past convergence to the present.

It is worth noting that by another measure of plate motion, the Juan de Fuca plate is anomalous. Nishimura and others (1984) and Riddihough (1984) have determined the motion of the Juan de Fuca plate with respect to a global hotspot (HS) reference frame. Because hotspots are essentially fixed in position relative to the mantle, the motion of the JdF plate with respect to them is an absolute motion with respect to the earth. The JdF/HS pole of rotation lies in northern California (39'N, 120*W), very close to the JdF plate. The JdF plate has a very small absolute velocity, only 10-20 mm/yr, making it "the slowest moving of the oceanic plates" (Nishimura and others, 1984, p. 10,228). Higher relative convergence rates of this plate with North America (e.g., 34 mm/yr) require that JdF/NA convergence has two absolute components -- subduction of the oceanic plate beneath North America, and oserriding of the oceanic plate by North America (Riddihough, 1984).

Special Characteristics of the Juan de Fuca Plate and the Cascadia Zone e Seismicity Most convergent (subduction) boundaries between oceanic and oceanic and continental plates are characterized by well-defined Benioff (Wadatti-Benioff) -zones of intraplate seismicity wi: Sin the subducting plate and by the occurrence of thrust-type earthquakes along upper (shallow) levels of the plate interface. The Cascadia subduction zone between the JdF and NA plates is clearly unusual in both regards. No thrust-type earthquakes have been documented along the plate interface, and Benioff zone development is i

1

pronounced only beneath the Puget trough of Washington. Even the existence of a Benioff zone associated with JdF plate subduction was questioned until 1983, when Crosson documented the occurrence of a deep (35-70 km) subplanar zone of earthquake hypocenters below the Puget trouFh . The zone is dis-tinctly separate from the diffuse zone of shallower seismicity (< 30 km depth) that characterizes the upper, North American plate in this region (Fig. 4). Crosson (op. cit.) implied that.the deep zone of seismicity lay within the subducted JdF plate, a relationship since supported by seismic refraction data (Taber,1983; Taber and Smith,1985; Taber and Lewis,1986; Smith and Weaver, 1983) and broadband teleseismic waveform analysis (Crosson and Owens,1987; Owens and others,1987) .

Most active subduction zonec are characterized by: (1) the episodic or periodic occurrence of large (M = 7 to 8+), shallow (<50 km depth) thrust-type earthquakes; and (2) the intervening frequent occurrence of smaller meriitude (4 to <7) interplate events. The Cascadia zone is characterized ,

by neither as the following statements testify

• l l

I "One of the most striking features of the Cascadia subduction zone is l the remarkable paucity of shallow earthquake activity between the trench axis and the coastal mountain ranges." (Heaton and Hartzell, 1986, p. 679)

"One feature often noted about the Cascadia subduction zone is the absence of thrust earthquakes on the subduction interface in historic time .... Any large thrust event on the subduction interface in the past 200 years is unlikely to have gone unnoticed and certainly could not have escaped detection in the past 150 years. ... Even  ;

with the recent expansion of seismic arrays lowering the detection threshold over much of the subduction interface to the microcarth-quake level, no thrust events have been detected." (Rogers, in press)

"On a world-wide scale (magnitude >4), the entire shallow dipping zone in the Northwest is presently aseismic and even at the local network level (magnitude >2), there is a complete lack of interplate thrust events during the last 3 years." (Taber and Smith, 1985,

p. 247)

The Cascadia subduction zone is an unusual one by any global standards because its plate interface is so seismically quiescent. Historic records

for the past 150 years and perhaps 200 years (Heaton and Snavely,1985) provide no evidence of large subduction events, nor does research based on modern seismograph networks reveal evidence for any earthquakes of thrust type having occurred along the JdF/NA plate interface. However, seismic quiescence alone cannot be used to rule out the possibility of large earth-quakes, since quiescence is a normal state during pre-seismic intervals along some activa faults. The "remarkable" seismic quiescence of the shallow JdF/NA plate boundary has two end-member explanations. Either displacement along the plate interface is ongoing and characterized by aseismic deforma-tional mechar. isms (e.g. stable sliding, aseismic creep, ductile flow), or the plate interface is locked, is accumulating elastic strain across it, and will inevitably rupture in an earthquake or earthquakes of large magnitude.

Heaton and Kanamori (1984) believe it likely that the seismic quiescence along the Cascadia zone represents a seismic gap. They point out that segments of the San Andreas fault which ruptured in major earthquakes .in 1906 and 1857 are still remarkably quiescent today. Rogers (in press),

referencing Robinson (1986), cites similar historic quiescence along a subduction interface in the Wellington area of New Zealand that generated a major earthquake in the past century. Heaton and Hartzell (1986), referen-cing Duda (1963), report that for at least 17 years prior to 1960, earthquake activity was anomalously low along the segment of the Peru-Chile trench which ruptured in the 1960 Mw 9 5 earthquake.2 2

Seventeen years is a very short period of time when assessing seismic behavior of any fault zone. Astiz and Kanamori (1986) list seven large (M, 7.0-7.4) earthquakes that occurred prior to the 1960 event within the lati-tudinal range of its rupture: 1919 (2), 1920, 1927, 1934, 1940, and 1949 The 1940 earthquake is best interpreted as a shallow-dipping thrust event that "probably occurred on the interplate boundary" (ibid., p. 1617). Fault plane solutions for the 1934 and 1949 events are consistent with normal faulting and down-dip extension within the subducting Nazca slab. Kaizuka and others (1973) reported that three earthquakes with probable magnitudes >

7 5 occurred prior to 1900 in the latitudinal range of the 1960 earthquake:

1575 (est. magnitude 8.5); 1737 (est. mag. 7 5-8.0; 1837 (est. mag. 8+) .

Both the 1960 and 1837 events were accompanied by coseismic elevational changes along the Chilean coastline.

1 However, Acharya (1985, p. 889) argues that.the complete absence of inter- )

plate seismicity along the Cascadia zone sets it apart from other subduction zones: "Examination of seismicity changes prior to great earthquakes in the circum-Pacific area ... shows no rupture zone so devoid of small earthquake activity during the interval in which the Juan de Fuca zone shows quies-cence." Even zones such as the Marianas trench, considered "aseismic" when some global comparisons of subduction zones are made, exhibit occasional magnitude 7+ earthquakes and instrumentally detectable "earthquake activity at magnitude > 4.5 level" (ibid., p. 889).

We believe that the quiescence within a seismic gap that preceeds'a great earthquake differs distinctly from the quiescence currently observed along the Cascadia subduction zone. There is no question that quiescence may extend to extremely low levels in a seismic gap that precedes a great earth-quake, but there is also no question that seismic gaps have consistently been characterized by moderate or heavy activity in adjoining areas. For example, the Michoacan seismic gap just filled by the 1985 Mexico City earthquake was identified precisely because the quiescence within the pending rupture zone contrasted sharply to the ebundance of activity in adjoining segments of the larger zone. Every subduction-related seismic gap identified by Fedotov (1965), Mogi (1968), and Kelleher and others (1973) was selected not because of quiescence alone, but rather because there existed a zone of quiescence within a larger zone of activity. The same is true for seismic gaps along strike-slip boundaries such as the San Andreas fault.

Such arguments may be discounted by the simple expedient of stating that the entire Cascadia zone is a seismic gap, but neither we nor Spence (in review) know of any other convergent boundary between any two plates which has rup-tured in a great or giant earthquake along its entire length. The 1000 km-long ruptured zone' in the 1960 Chilean earthquake broke one-fif th of the 5000 km-long Nazca-South American plate interface, two-thirds of the southernmost morphotectonic segment in which it occurred (33-45'S; Schweller and others, 1981; Nur and Ben-Avraham , 1981), and all of the southernmost sub-segment south of a major, offshore fracture zone (38-46*S; Herron, 1981). The fact that the entire 900-km-long Cascadia subduction zone is quiescent, despite l

l

'i

variations along it in convergence rate (see above), slab dip (see below),

and thickness of sedicents being subducted, indicates that it may be charac-terized by a fundamentally aseismic mode of plate convergence. We will return to possible reasons for such a mode.

o Geometry of the subducted plate Much new information is available or. the geo;3try of the subducted Juan de Fuca plate between its sediment-buried "trench" and the Puget Sound-Georgia Straits lowland. Analysis of onshcre-offshore refraction data along an erst-west line through the Grays Harbor area reveals that the offshore, subhorizontal Juan de Fuca plate bends down slightly to the east 100 km west of the coastline (ca 50 km east of the base of the continental slope), and then maintains a 9' dip beneath the Washington Coast Ranges and the western Puget Lowland (Taber and Lewis, 1986). The seismically-defined subducted slab extends to depths of about 80-90 km beneath eastern portions of the Puget trough.

Crosson and Owens (1987) and Weaver and Baker (1988), using high quality hypocenter data and data from analysis of broadband teleseismic waveforms (Owens and Crosson,1988; Owens et al,1938), haue shown that the subducted JdF slab has a broad,10-12* ENE-plunging archlike configuration beneath-western Washington and the Puget lowland. Crosson and Owens (1987) have constructed a contour map at the top of the "deep" hypocentral zone of earthquakes which they assign to the subducted slab, and assume that it corresponds to the Moho in the subducted oceanic slab (Fig. 5). Based on estimates of oceanic crustal thickness for the JdF plate (Johnson and others, 1984), the plate interface would lie 5 to 6 km above the contoured horizon (Figure 6). The arch is inferred by Crosson and Owens (op. cit.) to be a structural ac'conmodation for the change in plate geometry produced by the change in trend of the Cascadia "trench" from N-S, south of 47.5*N latitude, to N30*W to the north.

Most seismicity in the state of Washington, and in the subducted Juan de Fuca plate beneath it, is areally coincident with the broad crest of the i

arch in the Puget Sound region. The subducted slab has an eastward dip of 9-12* beneath the Washington Coast Ranges west of Puget Sound, with eastward steepening likely beneath the Sound as discussed below (Figs 4, 5; Crosson and Owens, op. cit.: Weaver and Baker, op. cit). To the north, beneath Vancouver Island (Spence and others, 1985; Rogers, 1987), and to the south, beneath southwestern Washington and western Oregon (Crosson and Owens, op.

cit.; Weaver and Baker, op. cit.), the more planar (unarched) segments of the subducted dip more steeply, up to 18' to 20'.

l The downdip length of the seismically active (Benioff-defined) subducted JdF slab is about 280 km accordin3 to the data of Crosson and Owens, 1987; cf.

Fig. 4). As such, it is among the 4 shortest of the 35 slab lengths for subduction zones or zone segments studied by Jarrard (1986); the other three: Philippine, 170 km; N. Cocos, 230 km; and Nicaragua, 280 km. The remaining 31 subducted slabs have seismically-defined downdip lengths that average 620 km and range from 310 to 1480 km (Jarrard, 1986 Table 1). The most commonly accepted explanation for the short length of the seismically active JdF subducted slab is that it is anomalously hot and becomes inelastic at relatively shallow depths (Riddihough, 1978).

Severinghaus and Atwater (1987) have recently proposed that the subducted JdF plate should be essentially aseismic at the present time because of its young age and slow subduction rate. They report that modern subducting  !

slabs characteristically become aseismic after a period of time roughly equivalent to 10% of the age of the subducting lithosphere. Given an 8 mil- I lion year age for the JdF plate at the time of its Holocene subduction, the slab should become largely aseismic after 800 Ka (even less if the effective l thermal age of the slab is less than 8 Ma as discussed below). The length of slab subducted over this period is only 27 2 km (3.4 cm/yr x 800 Ka), a length which places it beneath the offshore aseismic accretionary prism (see below) when it itself becomes largely aseismic. l Evidence from several sources indicates that the subducted JdF plate steepens eastward beneath the Puget Sound lowland from its shallow, 9-12' dip benesth western Washington (Fig. 6). Such steepening can be inferred from the

l l

l locations of the deepest intraslab earthquakes in the subducted plate (Fig.

4). Weaver and Baker (1988, p. 274) propose that an eastward change in dip from 11' to more than 25' "occurs suddenly, over a few tens of kilometers of horizontal distance" in the vicinity of the epicenter of the 1965 Seattle earthquake in the eastern Sound region. A downward steepening "knee-bend" l in the subducted plate is also required by the presence of the active Cascade I volcanic arc. Active volcanic arcs typically lie 100-125 km above subducting plates (Dickinson, 1970, 1975; Gill, 1981). Dickinson (1970, 1975), using-potash-silica ratios in Quaternary Cascade lavas, drew contours (100-120 km) on the top of the subducted JdF plate beneath the range. The slab depths contoured by Dickinson require that the shallow dipping plate (9-12*) beneath Puget Sound steepen eastward (30-50*) in order to achieve depths greater than 90 km beneath the arc (cf. , Dickinson,1975; Davis,1977, Fig. 2R C-7) .

McKenzie and Julian (1971) were the first to attempt to define slab dip by looking at teleseismic arrival times through the high-velocity slab. They concluded that it dips beneath the Washington Cascades at an sngle of 50*.  !

In a much more detailed analysis of teleseismic P wave arrivals, Michaelson and Weaver (1986) deduced that the JdF plate dips at 45' beneath the central Washington Cascades to depths of about 200 km. An unpublished seismic tomography study by E. Humphreys (personal communication to G. A. Davis, 11/87) appears to significantly modify the findings of Michaelson and Weaver (1986). Humphreys' analysis of teleseismic data demonstrates that a high-velocity slab underlies the Washington Cascades and that it dips very steeply (ca 75*) eastward to depths in excess of 400 km.

o Geologic factors Apart from its remarkable history of seismic quiescence and the short, seismically-defined slab length of the subducted Juan de Fuca plate, three geologic factors combine to set the Cascadia zone apart from others:

1) the extreme youthfulness (8 Ma) of the oceanic plate being subducted; none of the 39 subduction zones or zone segments studied by Jarrard (1986) have younger lithosphere descending along them (except locally, as at the southernmost end of the S. Chile plate and the northernmost

~- ._ _ - . _ _

end of the N. Cocos plate; the Rivera plate was not studied by Jarrard);

as discussed below, the effective thermal age of the sediment-covered Juan de Fuca plate is even younger than 8 Ma (Sammis et al, 1988);

2) a relatively slow convergence rate; taking a 34 mm/yr orthogonal rate of JdF/NA convergence at the latitude of the Columbia River, only 8-10 of the 39 subduction zones analyzed by Jarrard have slower rates; and

3) the presence at this convergent boundary of the Pacific Rim's most voluminous Quaternary sediment "trench" accumulation (Schc11 and Marlow, 1974) as measured in volume of sediment per length of trench.

Two of North America's greatest rivers (Columbia and Fraser) have carried sediments to the subduction site since at least middle Miocene time. These sediments are not confined to the filled trench, but cover the floor of the Cascadia Basin, blanketing newly-formed oceanic lithosphere within a few tens of kilometers of the Juan de Fuca spreading center. Sediment thickness at the western edge of the Cascadia Basin is estimated at 300 m by Moran and Lister (1987). To the east at the base of the continental slope, sediment thickness including fan deposits ranges up to 2 5-2.8 km (Kulm, 1983; Moran and Lister, 1987). Scholl and Marlow (1974) estimate that the total volume

]

of Cascade Basin turbidites deposited within just the past 1 Ma is approxi-mately 140,000 km3 This sediment cover impedes the advective hydrothermal cooling experienced by almost all other oceanic plates soon after their formation (including, the western or Pacific flank of the JdF ridge).

We believe that this unusual sedimentational history may account for the unusual seismic quiesence of the JdF/NA plate interface when combined with  !

other factors, such as high plate temperature (young age) and slow plate  !

convergence (cf. Byrne and others, in press; Sammis et al. ,1988) . These I interrelationships are discussed in a subsequent chapter, "Physical explanations for aseismic convergence, Cascadia zone".

Predicting The Magnitude of Subduction Zone Earthquakes Using Plate Convergence Rate and Age of the Subducting Oceanic Plate Ruff and Kanamori (1980) proposed that a worldwide correlation exists between the observed seismic moment-related magnitude (M,') of subduction zone

earthquakes and two principal variables -- the age (t) of the subducting oceanic lithosphere and the rate of plate convergence (v). According to their analysis, increased seismic coupling (and larger resultant earthquakes) occurs along zones when oceanic lithosphere is young (hot and buoyant) and convergence rates are high. M,' is a modification of the maximum observed magnitude to reflect the overall seismicity of the subduction zone. Kanamori (1986) later concluded that M,' is more appropriately interpreted as a parameter measuring seismic moment release, rather than the maximum magnitude of the characteristic earthquake for a specific subduction zone. He gives the relationship between moment release rate (MMR) and M,' as:

logio(MMR) = 1.2 M,' + 18.2 dyne-em/100 km/100 yrs (1)

Heaton and Kanamori (1984) extended the analysis of Ruff and Kanamori to the Cascadia zone as a predictive tool for a possible future great earthquake along it. They concluded that Ruff and Kanamori's global analysis for "maximum" earthquake magnitude along subduction zones is "well fit by the following relationship":

M, ' = -0.00889t + 0.13v + 7 96 (2) where t is the age of the subducting plate in million years (Ma), v is the convergence rate in em/yr, and the standard deviation of the observed M,'

around the predicted value is 0.4". Their plot of worldwide t vs. v rela-tions is presented in Figure 7 Inserting values appropriate for the Juan de Fuca plate (t = 10-15 Ma [their age estimate and one with which we disa-gree, as discussed below]; v = 3-4 cm/yr), equation (2) predicts a future earthquake along the seismically quiescent Cascadia subduction zone with a maximum M,' of 8.3

• 0 5.3 In light of Kanamori's (1986) subsequent assess-ment of the Ruff and Kanamori (1980) definition of M,', the use of Heaten and Kanamori's (1984) relationship for estimating "maximum earthquake magnitude" can be questioned.

3Note that the value of M'w in this equation is controlled by the constant 7 96. An event of M,' = 7 96 or 8 is predicted even if t = 0 and v = 0.

Although Heaton and Kanamori did not rule out tha possibility that subduction of the JdF plate is being accomplished by aseismic creep, they pointed out l

that the Cascadia zone shares many features associated with zones along which great historic earthquakes have occurred. Among these features:

youthful subducting oceanic lithosphere, shallow dip of the Benioff zone

(< 15'), absence of back-arc basins supposedly (indicating a high degree of seismic coupling), a topographically featureless oceanic plate, and seismic (pre-seismic?) quiescence.

Heaton and Hartzell (1986,1987) later introduced another factor into esti-mating the magnitude of a potential great earthquake along the aseismic Cascadia subduction zone -- the length of the historically aseismic zone.

They compare the overall length of the Cascadia zone (1200 km, including Gorda and Explorer plates) with the length of rupture (1000 km) of the 1960 Chilean carthquake. They consider the magnitude of that event (M = 9 5),

the largest in the historic seismic record, to represent the "largest earth-quake feasible in the Pacific Northwest", although most workers, including apparently Heaton and Hartzell themselves (1986, p. 702-703), would consider simultaneous rupture of the entire Cascadia zone improbable (cf., Spence, in review). Comparisons between the two subduction zenes, South Chile and Cascadia, are made in a subsequent section; it is our opinion that these comparisons do not lend credence to using the Chilean earthquake as an analog to a possible giant earthquake along the Cascadia zone.

There is some controversy in the literature regarding the importance of convergence rate as influencing seismic coupling between convergent plates.

Peterson and Seno (1984; their Fig. 9) discount its significance, whereas Jarrard (1986) defends its use as a predictive parameter in estimating M,.

Peterson and Seno considered factors that influence coupling in subduction zones, and calculated seismic moment (M,) release rates for 24 worldwide subduction zones (p. 10,247): "The definition of coupling in our study is the seismic moment release rate (M,), while in Ruff and Kanamori's work it is the maximum [ observed -- added here] earthquake magnitude." They con-cluded (p. 10,233) that the "moment release rate decreases as the age of the 1

.= - .__

i i

I subducting lithosphere increasss, when the zones belonging to a single sub-ducting plate are considered." Within single plates (ibid.) "age is the dominating factor affecting the strength of seismic coupling but . . . each plate as a whole has a characteristic moment release budget. ... The coment release rate does not Ancrease with convergence velocity; no simple relation- I l

ship was found between these two parameters. The moment release rate depends j most clearly on the age of the subducting lithosphere and the absolute velocity of the upper plate."

l Jarrard (1986) has conducted an exhaustive study of 26 subduction-related parameters (variables) for 39 subduction zone segments [ including the Cascadia zone].* He found (p. 276) "that most of the dependent variables are accounted for by one or more of only three independent variables: conver-gence rate (V, or V,3, ) , slab age ( A, or A, ), and intermediate dip (DipI)",

the average dip of the subducting slab between 0 and 100 km depth. With respect to predicting maximum earthquake magnitude in subduction zones he agrees with the Ruff and Kanamori (1980) analysis that convergence rate and slab age are the two key independent variables in defining magnitude of the largest earthquakes in subduction zones. However, he found that slightly improved empirical relationships between t, v and magnitude were achieved

• when cumulative earthquake moment (M,') was used rather than M,, arid when v components perpendicular to the trench were used, as opposed to total, oblique convergence rates.

Parameters include quantifiable geometric variables (e.g. , slab dip, trench  !

depth, slab dimensions, arc-trench distance), kinematic variables (e.g. ,

convergence rate, absolute motions of overriding and underriding plates), j and age variables (e.g. , slab age at trench, age at subducted tip, arc i age). Geologic variables that may influence subductive behavior, e.g.

accretionary prisms above the subducting plate and the subduction of aseismic ridges, are not easily quantified and were not discussed systematically or in detail; the possible effects of sediment load on the subducting plate, generally believed to be important for the Cascadia zone, were not treated at all by Jarrard.

l l

D l

Validity of Heaton and Kanamori's (1984) Analysis as Applied to the Cascadia Subduction Zone 1

As discussed above, Kanamori (1986) has subsequently reinterpreted M,' as used by Ruff and Kanamori (1980) and Heaton and Kanamori (1984) to be equiva- '

lent to moment rate rather than to maximum observed earthquake magnitude.

As such, the "prediction" by Heaton and Kanamori (1984) of a maximum earth-quake along the seismically quiescent Cascadia zone with an M, of 8.3 (2 0 5) l seems no longer appropriate. The applicability of the Ruff-Kanamori (1980) and Heaton-Kanamori (1984) relationships to the Cascadia zone can be ques-tioned from several other standpoints as well, among them: (1) the very youthful age (t) of the Juan de Fuca plate at the subduction zone, and (2) the implications of the young age of the plate with respect to the state of seismic coupling along the plate interface. Heaton and Hartzell (1986, 1987) appear to discount the analysis of Heaton and Kanamori (1984) by introducting a third parameter, apart from t and v, in comparing the Chilean and Cascad-ian zones. This third parameter. length of potential rupture of the Cascadia zone (ca 900 km) as compared to observed length of rupture of the, Chilean zone in 1960 (ca 1000 km), is used by Heaton and Hartzell (1986, p. 703) to conclude that a "reasonable upper bound for a hypothetical Cascadia subduc-tion earthquake" is M, 9 5 e Heaton and Kanamori's Estimate of Plate Age at Trench (t) for the Juan de Fuca Plate We question the validity of using only t and v parameters to assess the earthquake potential of the Cascadia zone, because for that zone these parameters lie at the extreme edge of the data base (Fig. 7) used by Heaton and Kanamori to predict M,'. Our placement of the Cascadia subduction zone in the Heaton-Kanamori diagram differs from theirs. Heaton and Kanamori l (1984) assumed an age for the subducting Juan de Fuca plate of 10 to 15 Ma, l although analysis of magnetic anomslies (e.g. Connard and others, 1984) and I the results of the DSDP drilling program (e.g. von Huene and Kulm, 1983) indicate that the oldest unsubducted oceanic lithosphere is approximately 8 Ma. Moreover, from a rheological point of view, the subducting plate has l

l l

an effective thermal age even younger than 8 Ma because it is significantly  !

hotter than its 8 Ma-old counterpart on the Pacific (western) flank of the spreading ridge (Scheidegger, 1984: Moran and Lister, 1987). The thinly-sedimented western flank oceanic lithosphere has cooled relatively quickly by advective flow of sea water through it - a normal behavior for most oceanic plates. However, thick Quaternary terrestrial sediments in the l Cascadia Basin on the eastern flank of the ridge inhibit hydrothermal cooling of the underlying Juan de Fuca lithosphere and permit it to cool only by conduction (Fig. 8). Thus, the Juan de Fuca plate remains hotter for any given magnetic anomaly age, e.g. 8 Ma, than its advectively cooled counter-part west of the Juan de Fuca Ridge. The fact that heat flow measurements in Cascadia Basin sediments are 15-35% below theoretical predicted values for cooling oceanic lithosphere (Moran and Lister, 1987) is not at odds with this statement. Such low measured heat flow values are explained by the thermal blanketing effects of rapid Pleistocene sedimentation, up to 200 m/my in the eastern Cascadia Basin (Carson, 1971: Kulm and others, 1973)

[ note: as an analogy, the cool lid of a thermos bottle gives little hint of the high temperature of the hot coffee beneath it]. Moran and Lister (1987, i

p. 11,431) state that the "addition of extra material to the cold surface reduces the overall te=perature gradient and raises the temperature of the {

pre-existing basement material." The atypically hign temperature of the 8 million year old oceanic lithosphere now being subducted beneath the base of the continental slope means, in effect, that the subducting Juan de Fuca plate is thermally younger than 8 Ma.  ;

\

e Very Young Oceanic Plates and Seismic Coupling i

l The youthful age of the Juan de Fuca plate raises a problem when the applic- l ability of the Heaton and Kanamori equation, (2) above, is considered.

Although it makes intuitive sense that hot, youthful oceanic lithosphere may i resist subduction because of its greater buoyancy relative to cider litho-sphere, the material behavior of such hot lithosphere during subduction is i uncertain. Might, for example, very hot lithosphere have mechanical proper-ties that would reduce the degree of seismic coupling across a subduction zone interface and favor aseismic components (steady sliding, creep, ductile

~20-

. . . = . - - .

flow) along the interface? This question has been addressed by many workers, among them Ruff and Kanamori (1980), Peterson and seno (1984), Heaton and Kanamori (1984), and Kanamori and Astiz (1985) . Heaton and Kanamori (1984,

p. 933) noted, with respect to the subducting Juan de Fuca plate, that "strong seismic coupling implies that slip occurs only during earthquakes, whereas weak seismic coupling implies that slip occurs mainly in the form of aseismic creep." We will discuss intraplate coupling in a different context in a subsequent section of this response, but an alternative cencept to that proposed by Heaton and Kanamori (immediately above) needs introduction here.

A ple'. ioundary exhibiting physical or mechanical coupling to some degree is .. ..ecessarily a boundary characterized by seismic coupling and stick-slip behavior. As one example, the deformation of sediment within an active accretionary prism clearly indicates transmission of stress across the underlying interface with the subdacting plate, yet this interface is almost universally aseismic (Chen et al., 1982). Coupling between plates should not unequivocally be equated with seismic coupling -- a point overlooked apparently by some seismologists.

Kanamori and Astiz (1985) specifically addressed the question of whether or not Ruff and Kanamori's empirical relation between t, v, and M, is valid for the Cascadia subduction zone, given the youthfulness of its subducting plate and its slow rate of convergence with North America. Essentially, they conclude that it is by drawing upon a large (M, = 7 7) earthquake that occurred along the western coast of northern Honshu in 1983 (Fig. 9). They interpret this event, the Akita-Oki (or Nihonkai Chuba) earthquake, to be the consequence of the subduction of young (t = 20 Ma) oceanic lithosphere eastward beneath Honshu at a rate of only 1.1 cm/yr. Using these parameters, the Heaton-Kanamori equation predicts a M, = 8.0 event versus the M, = 7.8 actually observed (although as pointed out previously, any use of the equa-tion will result ih a M, no smaller than 7 96). Because of the Akita-Oki earthquake Kanamori and Astiz conclude that the Heaton-Kanamori equation is valid for subduction zones of Cascadia type with very small t and v.

We disagree. A serious obj9ction to the use of the Akita-Oki earthquake as a test for predicting M, for subduction zone earthquakes is the fact that no

well-developed subduction boundary dips eastward beneath northern Honshu and Hokkaido (Fig. 9; Tamaki and Honza,1985; seno,1985) . Convergence certainly exists along the boundary of the Sea of Japan with the western edge of north-een Honshu and Hokkaido, but Tamaki and Honza (1985) describe the convergent zone as characterized by alternating segments in which active thrust faults dip both eastward (e.g., Akita-Oki, 1983) and westward (e.g., Niigata, 1964). They interpret geologic and seismologic relations along this nascent plate boundary as due to both incipient subduction and obduction of oceanic crust; 1.e. , oceanic crust forms the lower plate along segments of the com-pressional zone, but along other segments it is being faulted over continen-tal crust (obduction). It is thus geologically inappropriate to compare the seismogenic behavior of the northern Honshu and Hokkaido convergence zone, with its alternating directions of tectonic vergence, to the long-established subduction of the Juan de Fuca (formerly Farallon) plate beneath western North America.

Peterson and Seno (1984) have documented that alpha, the ratio of seismic slip rate to plate convergence rate (= relative plate velocity) and, there-fore, a measure of seismic coupling, is age dependent; in general, older 1 plates (100 Ma er more) have lower alpha values. The same appears to be true for plates younger than 20 Ma. Kanamori and Astiz (1985) present an empirical curve in which alpha appears to vary between 0 and 1 for subducting plates with ages between 0 and 20 Ma. They estimate that alpha for the Cascadia subduction zone may be about 0 3 We believe that this estimate is suspect because it is based on an erroneous (too old) age for the Juan de Fuca plate at the time of subduction (i.e. 10-15 Ma) and is poorly con-strained by values of alpha for the subduction of other young plates. A lpha I for the Cascadia zone, given an effective thermal age for the subducting Juan de Fuca plate < 8 Ma, could thus be significantly less than 0 3 In considering the effects of variable coupling ratios, Kanamori and Astiz (1985, p. 309) conclude that the M, = 8.4 earthquake predicted by Heaton and Kanamori (1984) for the Cascadia zone would have "a repeat time of 126 years if slip at the plate boundary is completely seismic, or 420 years if only 30 percent of the plate motion is taken up by seismic slip."

_.~.. _ - _

Their assumption, and one which we question, is that the same deformational mechanisms prevail along plate interfaces with high alpha as along those with low alpha. It simply takes longer for strain sufficient for the generation of great earthquakes to build up if the coupling ratio is less.

An alternative effect is possible. Given that the degree of seismic coupling l is related to subducted plate age, i.e. plate temperature, it seems likely 1 to us that the basic slip mechanisms or the rheology of materials in the thrust interplate zone may be changed under conditions where very small alpha (high t) prevails. Because alpha going to zero effectively extends the repeat time of major earthquakes to infinity, there may well be some range of small coupling values for which plate interactions change from stick-slip to stable sliding behavior, or for which ductile relaxation processes along the plate interface prevent elastic strain from accumulating.

Physical reast u for possible aseismic subduction of the Juan de Fuca plate along the Cascadia zone have recently been presented by various authors, e.g.

Pavlis and Bruhn (1983), Byrne and others (1987; in press), Sykes and others (1987), and Sammis et al., (1988). They are discussed in a subsequent section.

l Estimating the Interplate Seismic Potential of Subduction Zones: An Alternative I

We believe that the Ruff and Kanamori (1980) and Heaton and Kanamori (1984) empirical relationship for estimating the seismic potential of subduction zones places an unreasonable emphasis on the two parameters of plate age and convergence rate by ignoring geologic factors that are widely believed to influence the seismic behavior of subduction zones. Among such factors are (1) sediment cover on the subducting plate, (2) the presence or absence of asperities on the 'ubducting s plate such as seamounts, fracture zones, and aseismic ridges, and (3) the downdip width (W) of the seismogenic interface that lies between the deepest extent of the typically aseismic accretionary prism (Chen and others, 1982) and the depth along the plate interface (typi-cally ca. 50 km for plates with average thermal characteristics) of the brittle-ductile transition (Fig. 10).

Byrne and others (in press) believe that the Ruff and Kanamori l

l "method of estimating seismic potential is not very direct nor '

physically related to earthquake processes. Because rupture area ultimately determines the size of an earthquake, we think that {

methods that try to estimate the maximum rupture area are a more direct way of estimating'the size of future or potential large earthquakes . For large to great earthquakes this area is defined by the downdip width, W, of the seismogenic zone, and the length along strike, L. ... Given an improved estimate of W for large thrust earthquakes, better estimates of their seismic moment are in turn possible because moment is related to W through M, = u ULW where u is rididity, U is the average slip over the rupture zone, and LW is the rupture area (length times downdip width) . ...

From an estimate of W, we can calculate the maximum seismic moment of s'uture thrust events. In addition, from knowledge of W and the plate convergence velocity it should be possible to estimate the average repeat time of large earthquakes. ... We think that this method of calculating maximum earthquake size has a better physical and mechanical basis than mathods like those of Kanamori [1971) and Ruff and Kanamori [1980,1983] that attempt to qualitatively describe maximum earthquake size in terms of seismic coupling. The two parameters that they use, rate of plate convergence and age of the downgoing plate, may be useful in certain situations for rough estimates of maximum size, but may well fall and give potential seismic moments that are too large for margins where large thick-nesses of young sediments underthrust the backstop" (italics added here for emphasis; "backstop" refers to the body of rocks stronger than the accretionary prism and against which the prism accumulates) .

There are ample reasons to consider that W, the downdip width of the seismo-

)

genic interface between the subucting Juan de Fuca plate and North America, should be narrow. Because the seismically-defined downdip width of the subducted plate (ba<sd on intraplate events) is among the shortest known (as l discussed above), the maximum width of the potential seismogenic interface above it should also be narrower than r.ormal or, possibly, eliminated alto- i gether by the unusual geologic factors associated with this zone (Fig.10) . j The large downdip width of the Cascadia accretionary wedge above the subduc-ting slab (Jarrard, 1986, p. 270) would shift the upper (shallowest) limit of W downward to an undetermined position beneath the continental shelf or l

l

1 i

i 1

l western Washington. The position is undetermined because the location of the upper-plate backstop, a rheological boundary between "soft" accreted l materials and more rigid, or "stiffer" rock assemblages has not been defined. l We attempt to define it subsequently in this response. Furthermore, as Byrne and others comment (op. cit., above), sediments subducted atop the Juan de Fuca plate may be so voluminous as to be carried down beneath the backstop of this subductive system, thus negating its normal role in defining W.

( Finally, the lower (deepest) limit of W for the Cascadia zone should be l abnormally shallow because of the abnormally high temperatures within the subducted plate; the depth of the brittle-ductile transition within the upper plate, which defines the lower limit of W, is, of course, strongly temperature dependent. Given the highly anomalous seismic quiescence of the Cascadia plate interface for earthquakes of any detectable magnitude, W - in light of mitigating factors discussed in this paragraph - could be expected to be very narrow, perhaps nonexistent. The issue of the width of the interface will be discussed subsequently in the context of maximum magnitude.

1 Physical Explanations for Aseismic Convergence, Cascadia Subduction Zone Introduction Heaton and Kanamori (1984) believe that "the best examples of seismically quiescent plate boundaries are ones that have experienced great earthquakes, but that could be considered as otherwise locked". Witis respect to the Cascadia zone they comment that "if slip is occurring aseismically on the shallow part of the subduction zone, then this particular example would have to be considered unique". In point of fact, the Cascadia subduction zone is unique, as Heaton later concedes in Heaton and Hartzell (1986, p. 676): l 1

"Unfortunately, no subduction zone is exactly the same as the Cascadia zone... .

No two' subduction zones are alike, and the Juen de Fuca plate  ;

with (a) its youthful age and high temperature, (b) its relatively slow I convergence with North America, (c) its very slow absolute velocity, (d) its very thick sediment cover, and (e) the short down-dip length of its seismi-cally defined subducted slab all contribute to the overall uniqueness of the Cascadia zone. It was initially argued intuitively that the slow subduction 25 i

. . .-~

of a-hot, sediment-laden plate, could give the Cascadia zone its unusual absence of shallow interplate seismicity (e.g., Atwater, 1970; to some extent, Riddihough, 1978). More recently, the case for aseismic subduction

- has been bolstered by rock mechanics-based appraisals of the significance of temperature, convergence rate (i.e., strain rate along the plate interface),

and thick sediment cover on the mechanical behavior of the JdF/NA plate interface (e.g. , Byrne and others,1987; in press; Sykes and others,1987; Sammis et al,1988) .

The Role of Sediments in Influencing Aseismic Subduction The thick sedimentary load (up to 2.8 km thick) of the. Juan de Fuca plate and the plate's very young age are the features that most set this plate apart from others that are actively undergoing subduction. Why might a hot, sediment-laden plate subduct aseismically? Several sediment-related factors have been proposed by recent authors as contributing to the likelihood of aseismic slip by stable sliding along the Cascadia plate interface. Among these factors are the clay-rich mineralogy of the sediments themselves, their elevated temperatures at the time of subduction, and the elevated fluid pressures within them. First, each factor is reviewed separately; then, their collective effect on stable sliding between the Juan de Fuca and North American plates is discussed.

e Clay-Rich Marine Sediments A feature which sets subduction fault zones apart from most other fault zones is the entrainment along the fault zone (plate interface) of l incompletely lithified marine sediments. Rock mechanics studies within the past decade (Summers and Byerlee,1977; Wang and Mao, .1979; Wang,1980; Scholz, in press) have demonstrated that most water-saturated clays and fault-generated "clay" gouge will, if present along movement surfaces (faults), facilitate stable sliding. This phenomenon occurs under confining pressures appropriate to the shallow depths of thrust faults beneath accre-tionary prisms, i.e up to 3 kb (ca 12 km) and may occur even deeper (up to 4 {

1 l

l

._. -. _. _ - - - . -- - ~ - - '

6.3 kb) if the clays are montmorillonite or vermiculite (Summers and Byerlee, 1977; Wang, 1980).

Clays, including montmorillonite and vermiculite, are abundant within, and dispersed throughout, the eastern Cascadia Basin sedimentary section (Ship-j board Scientific Party, 1973, Site 174, DSDP Leg 18; Hayes, 1973). At DSDP drillsite 174, on the southern distal edge of the Astoria fan, mixed-layer mica (illite)-montmorillonite and vermiculite clays make up 44-66% of the total clay fraction in 10 Late Pleistocene to Pliocene samples (Hayes, 1973.

Table 2). The samples were collected within the upper 761 m (83%) of a sedimentary section believed to have a total thickness of 911 m (Hayes, 1973; shipboard Scientific Party, op. cit.).

Byrne and others (in press) emphasize the role played by water-saturated, clay-rich marine sediments during their subduction in facilitating the buildup of high fluid pressures and the lowering of effet,tive stresses (confining pressure minus fluid pressure) across the plate interface.

However, another, perhaps more important role, can be attributed to clay-rich sediments during subduction. It has been established that the coeffi-cient of friction (0.6-0.8 for most rocks, Byerlee, 1978) is anomalously low for most water-saturated clays, e.g. , 0.08 and 0.15 for montmorillonite and kaolinite, respectively (Wang, 1980). Expandable clays, such as montmoril-lonite and vermiculite, contain water between silicate layers that is struc-turally bonded to the clay; their physical properties differ fron non-expand-able clays, such as kaolinite, that contain free water (water not structur-ally bound to the clay minerals). Wang (ibid.) believes that as clays of the latter group are subducted, they will gradually lose their trapped water by compaction, effective stresses across the plate boundary will rise, and stick-slip behavior will be enhanced. "On the other hand," according to Wang (1980, p. 531), "if the subducted sediments are composed largely of expandable clays such as montmorillonite, sliding between the plates may remain stable even after the free water is largely lost" (Wang, 1980,

p. 531).

o High Temoeratures of Cascadia Basin Sediments Sediments above the subducting slab at the buried Cascadia trench have higher temperatures than those being subducted along most convergent zones.

For example, the easternmost sediments of the Cascadia Basin are buried to depths of as much as 2.8 km and rest on a basaltic substratum that may be at -

temperatures of 200*C or more (Scheidegger, 1984: Moran and Lister, 1987).

Although such temperatures are not hot enough (ca 300*C; Brace and Kohlstedt, 1980; Sibson, 1984) to induce crystal-plastic flow of quartz in basal sedi-ments being subducted beneath shallower levels of the accretionary prism, they could contribute to stable sliding (aseismic) behavior at the JdF/NA plate interface, as opposed to stick-slip (seismic) behavior (Stesky and others, 1974; Tse and Rice, 1986; Sammis et al, 1988).

e High Fluid Pressures Not only are sediments of the Cascadia Basin being heated to unusually high te=peratures at the convergent boundary, but they are being impressively dewatered near the slope-basin interface, presumably because of horizontally-induced stresses related to plate convergence and deformation at the outer ridge (Kulm and Peterson,1984; Shi and Wang,1985). Sediments deformed in t the outermost marginal ridges at the base of the continental slope are being mechanically consolidated; their water content ranges bstween 17% and 30%

compared with 55% to 70% in similar, but undeformed sediments of the Cascadia Basin (Kulm, 1983). Dewatering is presumably accompanied by an increase in fluid pressure (and a lowering of effective stress) and a decrease in shear strength of the affected sediments. Both physical effects could be expected to contribute to aseismic slip along the JdF/NA plate interface if even a thin layer of such sediment with high fluid pressure is subducted. High fluid pressures in undeformed sediments presently being subducted beneath the Barbados Ridge complex are credited by Westbrook and Smith (1983) with permitting decoupling at the base of the thick accretionary prism.

The aseismic nature of accreted prism materials in forearc settings uround the world has been investigated by Chen and others (1982). They report that

'ths ccerotsd widgss (definsd as composed of material clearly removed from the subducted plate) of subduction zones deform aseismically. Materials of the wedge can probably not be expected to accumulate significant elastic strnins because of the - -;h internal fluid pressure (which contributes to low effective stress) and low strength. Possibly, Chen and others suggest (op. cit., p. 3688), "the mixture of soft, wet sediments which are sheared off from the surface of the rubducting slab behave like plastic or viscous materials us. der compressive stresses."

Given the aseismic nature of accretionary prisms, it is unlikely that the shallow thrust fault interface between them and the subducting plates will be seismogenic. Pyrne and others (1987; in press) believe that this aseismic portion of the plate boundary might move either continuously or episodically.

Its downdip limit along most subduction zones is the "seismic front" (Fig.

10), the transition along the plate interface between aseismic (stable sliding) behavior and deeper, seismic (stick-slip) behavior. The seismic front appears to be broadly coincident with the upper-plate contact between the aseismic acecetionary prism and the backstop, the lithified rock assem-blage against which the prism has accumulated. For most subduction zones, subducted sediments along the plate interface generally appear to have been added to the accretionary prism (presumably by underplating) before they reach the backstop. Rock (upper plate) against rock (lower plate) interac-tion along the plate boundary beneath the backstop leads to stick-slip (seismogenic) behavior until the interface reaches the depth of the brittle-ductile lithospheric transition. But, as Sykes and-others (1987) and Byrne er.d others (in press) argue, for those subduction zones where voluminous sediment is present, subducted, everpressured sediments may be carried down to levels beneath the backstop. In these cases, the width (W) of the seismogenic plate interface may be dramatically ns.rrowed or eliminated altogether. Quotihg from Sykes and others (1987, p.1468):

l "This paper emphasizes the Cascadia subduction zone where gt9at thicknesses of young sediments are present. We argue that the plate  ;

boundary has not reached the bese of those sediments by the time it encounters the backstop. Young sediments are thus carried beneath {

l the backstop, which results in the downdip width of seismic slip j being reduced or eliminated altogether. The maximum size of earth- l 1

i

)

l l

. - - - - . . . _ . . . . - -~ ~ , - -

l 1

quakes along the Cascadia zone appears to be much smaller than some recent estimates based on relat'tonships-of Ruff and Kanamort."

(italics added here for emphasis) l Stable Sliding Along the Juan do Fuca/ North American Plate Interface The seismic quiescence of the shallow-dipping JdF/NA plate interface can be explained either as an expression of a thoroughly locked plate boundary, currently in a period of pre-seismic elastic strain accumulation, or as the consequence of aseismic slip processes along it. What is most puzzling about the former explanation (i.e., stick-slip behavior) is the apparent -

complete absence of even small magnitude thrust-type events along the plate interface. Physical models for total locking of a several hundred kilometer wide, 500 km-long plate interface, given differences in convergence rate j (Oregon vs. Washington) and inclination of the subducting slab along it, are difficult to conjecture.

-l However, an explanation for the seismic quiescence of the plate interface does come from experimental studies of stable frictional slip (stable slid- l ing) and those environmental parameters that favor it; at deeper levels ,

where temperatures exceed 300*c, crystal-plastic (ductile) deformational wcharisms and low-grade metamorphic recrystalliization can account for aseismic interface behavior. Recent papers by Byrne and others (1987; in press) and Sykes and others (1987) emphasize the physical role played by  !

subducted sediments in facilitating stable sliding. They propose that the subduction of overpressured, clay-rich marine sediments lowers effective _

stresses across the plate boundary and leads ;.o stable sliding.

Sammis and others (1988) also propose that stable sliding is the most likely behavioral mode of the JdF/NA plate interface, but they emphasize that the elevated temperatures of subducted sediments along the plate boundary are the most important reasons for stable sliding - not just the presence of overpressured, clay-rich sediments. All three sediment-related factors -

high fluid pressure (= lower effective stress), clay-rich mineralogy (= lower friction coefficients), and elevated temperatures - will favor stable fric-tional sliding behavior as opposed to stick-slip behavior, but the dominance

of temperature among the three appears likely. It is the high temperature of the subducting JdF plate and the sediments that are carried down with it that most sets this plate apart from others. i l

What are likely temperatures of the sediments being subducted along the Cascadia zone? Te=peratures in the 2 to 3 km thick sedimentary section in l the eastern Cascadia Basin are believed to increase to approximately 200*C at the contact with basaltic basement (McClain, 1981; Scheidegger, 1984; Moran and Lister, 1987). At what depth within the sedimentary pile does tectonic separation occur between sediments that are off-scraped into the Cascadia accretionary prism and those that are subducted (underplated) with the Juan de Fuca plate? Some seismic reflection records across the base of the continental slope off fashington indicate accretion to the slope of as much as the upper kilometer or more of Pliocene (?) and Pleistocene sediments (McClain and others, 1984). Shi and Wang (1985) have shown that pore pres-sures in I sedimentary layer at the onset of subduction are such that the effective pressure is at minimum values near the center of the layer.

Decollement would be most expected at this horizon; for Cascadia Basin sediments, temperatures along the decollement would approximate 100*C if a uniform bottom-to-top temperature gradient is assumed. Decollement near the base of the sedimentary layer, as proposed by Davis and others (1986) for the Cascadia subduction zone offshore from Vancouver Island, raises the probability of even higher interface temperatures.

Sammis et al (1988) postulate that sediments being subducted along the Cascadia zone have an initial temperature of 100*C. They conclude:

"Comparison with the boundary between stick-slip and stable sliding in pressure-temperature space for Westerly Granite at strain rates of 10 3 sec. (the only experimental data with which the comparison can be made) suggests that these temperatures are high enough to place the entire upper-slab interface within the stable-sliding regime This conclusion is strengthened by considering more realis-tic pm ameters including fluid pressures, lower strain rates, and the lower brittleness for sedimentary rocks for which laboratory data is presently not available.

Previous empirical correlations used to assess the seismogenic potential of the Cascadia subduction zone (limited only to rate of

convergence and age of the subducting plate) have not taken into account the thermal and mechanical consequences of. the JDF plate's unique sedimentation history. The very young effective thermal age of the plate resulting from the early suppression of hydrothermal cooling implies such a low seismic efficiency (afpha) that it is reasonable to question whether any stick-slip thrust type earthquakes are to be expected during its subduction. The frictional analysis developed above suggests that they are not the entire shallow interface may deform by stable sliding."

Comparisons of the Cascadia Subduction Zone With Other Zones Given the present seismic quiescence of the Cascadia subduction zone, one approach to understanding its potential for great earthquakes is to compare it to other seismogenic subduction :ones along which plate convergence is slow, or along which young oceanic lithosphere is being subducted. In the preceding section we argued that the Cascadia zone is an atypical. if not.

unique, zone and that comparison of its deformational behavior with other zones may, therefore, not be valid. Nevertheless, since such comparisons have been made by others (Heaton and Hartzell, 1986: 1987: Rogers, in press),

we discuss them here. It is our conclusion from these comparisons, that they substantiate the position taken here that the Cascadia zone, in its totality of geophysical and geological parameters, is genuinely different than the seismogente zones with which its is sometim6s compared. Spence (in review) has reached a similar conclusion: "In general, each source zone for great earthquakes that has been termed analogous with Cascadia has its own unique properties that make the nature of stress accumulation there different from stress accumulation at Cascadia."

The most detailed, systematic comparison of global subduction zones with the Cascadia zone has been published by Heaton and Hartzell (1986: Woodward-Clyde Consultants, 1984, and Jarrard, 1986, have presented more extensive comparisons of major subduction zones on a worldwide basis) . Heaton and Hartzell (ibid.) deem the subduction zones most similar to the Cascadia zone to be those of southwestern Japan (Nankai Trough), southern Chile (southern Nazca plate, north of 46*S latitude), and Colombia (northern Nazca plate),

all of which have experienced historic great (M > 8) earthquakes. Other subduction zones that have been co= pared with the Cascadis (Heaton and

. _ _ _ _ . . _ ~ . . _ _ . _ , _ --- _ _ . - -

l I

Hartzell, 1986, 1987; Rogers, in press) are those of the Rivera and norther- -l

!~ nmost Cocos plates off western Mexico, and the southernmost Chilean margin (Tierra del Fuego) between the South American and Antarctic plates (south of I 46*S latitude). All six of the aforementioned zones are characterized by the subduction of very young to ycung oceanic lithosphere, but all differ 'in l one or more major respects from the Cascadia zone.

I A detailed comparison with the Makran subduction zone of southern Iran and Pakistan will not be made here, despite the recent suggestion in abstract that it may be an analog of the Cascadia zone (Laane and others, 1986).

Although the rate of convergence along this zone 3 7 cm/yr, is comparable to that of the Cascadia zone, the oceanic lithosphere being subducted along it is of Cretaceous age (ca 87 Ma; Jarrard, 1986) and is, therefore, much colder than the subducting Juan de Fuca plate. The Makran zone is clearly anomalous in the extremely wide development of its accretionary prism; its arc-trench width of 480 km is the largest of any of the 39 subduction zones characterized by Jarrard (1986).

Strain Classification of Upper Plates of Subduction Zonen I Before comparisons are drawn between the six zones and the Cascadia zone, the work of Jarrard (1986) in assi ning subdu,: tion zones to strain classes should be introduced. Jarrard has attempted to semiquantify the strain regime cf the upper plate for the 39 zones or zone segments analyzed by him.

He has arbitrarily established seven strain classes which define a continuum from strongly extensional (class 1) to strongly compressional (class 7) upper-plate strain regimes. The strain classes are geologically-based and according to Jarrard tell us something about (1) the degree of coupling between the subducting plate and the plate above it, and (2) the potential earthquake magnitude of interplate events.

We believe that some caution is advisable regarding the inferred relation-ships between coupling and potential earthquake magnitude. A plate interface in a stable sliding mode may still be coupled to some degree, although large earthquakes are not likely to occur along it. Stable sliding is a variant

. of frictional sliding and complete decoupling along the sliding surface is not implied by this mechanism. It is, therefore, erroneous to equate physi-cal coupling across a plate interface with seismic coupling that can only be relieved by future faulting and earthquakes. For example, some degree of physical coupling must exist between subducting plates and their offscraped and underplated accretionary wedges (e.g., see previous discussion regarding the conclusions by Adams, 1984, in this regard on the Cascadia zone), yet the globally observed aseismic behavior of plate interfaces beneath these wedges is perhaps the strongest evidence that stable sliding mechanis=s must occur there.

I Stress and strain indicators used by Jarrard (op. cit.) to indicate strain class include fault plane solutions, Quaternary structures, volcanic vent alignments, results from overcoring, and presence or absence of back-arc spreading. An admitted complication in the use of such disparate data is that some supply instantaneous stress or strain data (e.g. , earthquake focal mechanisms, overcoring), whereas others represent deformation throughout l geologic time (e.g. , Qucternary structures, volcanic alignments),

e Jarrard's (1986) Classification 1

"Compressional" and "extensional" environments designated by Jarrard refer to those respective cases where the maximum horizontal compression direction and the maximum horizontal extension direction are perpendicular to the trench, Jarrard warns that the strain classes do not represent equally spaced points in a continuum and that some subduction zones have a classi-fication uncertainty of one class. His classification:

Class 1: Active back-arc spreading; formation of oceanic lithosphere Class 2: Incipient extensional rif ting or very slow back-arc spreading Class 3: A tensional environment, but no significant or only minor extension Class 4a: Neutral, i.e., zones with little evidence of either compression or extension 4

. _ - , , , _ . , . . _ . .- _ . . , . ~ , - . - . _e-._ ...-.<.,e-,,...-,._.-...-...m., _-,r,, , ., . . . - . _ , ,, - . ,

Class 4b: Strain gradient perpendicular to the su'<Juction zone; most typically back-arc extension and arc and forearc compression Class 5: Mildly compressional, including gentle folding; some evidence for local extension is permissable Class 6: Moderately compressional with active folding and reverse l faulting; no cignificant extension Class 7: Very strongly compressional, especially in back-arc (foreland areas)

Jarrard's strain classification thus serves as an additional parameter for subduction zone co=parisons, and is cited below for each zone compared with the Cascadia.

e Cascadia Zone Strain Class Jarrard assigns, somewhat reluctantly, the Cascadia zone (his "Cascades" zone) to the "neutral" group of strain class 4a on the basis of focal mechan-isms in the Pacific Northwest that indicate north-south compression approxi-mately parallel to the "trench" (both strike-slip and thrust mechanisms, Zoback and Zoback,1980), and active north-south normal faulting in the Cascade Range of Oregon (Hammond,1979) . However, he admits (1986, p. 228) that the "Cascades may be more appropriately considered as mildly exten- l sional" (class 3), or alternatively, that "the subduction-related strain pattern may be dominated by the more pervasive strike-slip environment of the westermost United States and Canada."5 Jarrard (1966, p. 217) believes that "strain class is probably determined by a linear combination of conver gence rate, slab age, and shallow slab dip."

To the contrary, Rogers (1985), following Dewey (1980), argues that the tectonic regime in the North American plate above the subducting Juan de l

5The suggestion that the north-south compressional stress field in the Pacific Northwest may be an expression of transform motion between the very large Pacific and North American plates has been made before by others, among them Crosson (1972), Davis (1977), and sbar (1983). The implications of this possibility for the state of coupling between the North American and Juan de Fuca plates is discussed in a subsequent section on geodesy.

Fuca plate is determined by the difference in motion of the overriding NA plate (V,) and the oceanward migration or roll-back vector (V,) of the line of flexure (the "knee bend") in the subducting JdF plate at a depth of 40-80 km. The roll-back vector is a function of the degree of density contrast between the sinking plate and the surrounding asthenosphere; density differ-ences become accentuated at the knee bend where oceanic crust undergoes the basalt to eclogite transformation. According to Dewey (1980) V, is indepen-dent of both the subduction rate of the sinking plate (V,) and the total convergence rate (V, + V,). If V, > V,, then the upper plate experiences compression. If V, < V,, the strain regime of the upper plate is extensional.

Rogers (1985, p. 496) argues that the two movement vectors must be about equal for the Cascadia zone, given the absence of either back-arc spreading or "extensive contemporary mountain building". However, both he and Weaver ,

and Michaelson (1985) attribute the greater volume of Quaternary volcanism in the Oregon Cascades, as opposed to more northerly portions of the volcanic arc, to slight extension (ie., V, slightly greater than V,) of the Oregon upper plate perpendicular to the subduction zone. The segment of the arc in Washington and British Columbia is considured by Rogers to be either neutral or slightly compressional. Thus, the "neutral" classification applied by Jarrard to the upper plate of the Cascadia zone may be overly simplistic. j In reality, the zone might better be segmented from north to south into  !

strain classes 5 or 4 (north of Rainier), 4 (between Rainier and Hood), and 3 (south of Hood) (cf., Weaver and Michaelson). This topic is readdressed in the section of this response on segmentation of the Cascadia zone.

Southwestern Japan (Nankai Trough)

Heaton and Kanamori (1984) considered the Cascadia and southwestern Japan (Nankai Trough) subduction zones to closely resemble one another in terms of plate convergence rates (v), age of subducting crust (t), sud radiment-clogged trer ch. Both zones have apparent convergence ratss of 3 to 4 cm/yr, although the age of crust now being subducted differs by a factor of three (e.g. , SW Japsn, ca.17-24 Ma, Heaton and Hartzell,1986, Table 1 and Figs.

I l

l l

l 6, 7; Cascedia, 8 Ma with a still younger "thermal" age6 ). Great thrust-type earthquakes along the SW Japan zone, e.g., 1944 (Tonankai, M = 8) and 1946 (Nankaido, M = 8.2), have an average repeat time of approximately 120 years-(Kanamori, 1977), and are used to support the case that the Cascadia zone is itself capable _of generating great earthquakes. Seno and others (1987) estimate the convergence rate between central Honshu and the Phillip-pine plate at 2 9-3 7 cm/yr, a rate which they equate to the seismic slip rate (alpha = 1: Kanamori and Astiz, 1985, estimated alpha = 0.88).

Jarrard (1986) has assigned the Southwest Japan subduction zone to the class 5 category, i.e., mildly compressional, in large part because Naka=ura and Uyeda (1980) reported that the direction of principal horizontal compression in the upper plate we.s roughly parallel to the direction of NW-SE plate convergence. Ukawa (1982, p. 550), however, disputes this finding, clciming that sigma h,,, in the upper plate is nearly E-W, "significantly deviated

... from the moviig direction of the Philippine Sea plate relative to the Eurasian plate." He attributes the east-west compression to convergence between the Eurasian and nearby Pacific plate, not to Eurasian-Philippine Sea plate convergence. But to further co=plicate matters, Okano and others (1980) report that although sigma h,,, is nearly E-W for earthquakes in the upper plate beneath Shikoku island, it is N-S for depths between 23 and 50 km. It thus appears that Jarrard's strain classification for the upper plate of the SW Japan zone is not well established, despite seismological evidence for very strong coupling along the plate interface (alpha = 0.88-1),

Like the Colombia zone discussed below, the SW Japan zone is unusual in 6

Given the wide range of age for oceanic lithosphere (ca 0 - 160 Ma), the age difference between 8 and 17-24 Ma may seem insignific.nt, especially when plotted on a Heaton and Kanamori t vs. v plot (Fig. 7). However, there is considerable difference in the thermal characteristics of oceanic plates with these age differences. For example, the observed mean heat flows from the tops of oceanic lithosphere with ages of 8 vs. 24 Ma are approximately 2.8 vt,. 1 5 heat flow units, respectively (Fig. 11). The latter value approximates the mean heat flow for all oceanic lithosphere older than 30 t million years (Fig. J1). It is difficult to visualize, therefore, why  !

oceanic plate age differences from 30 to 160 Ma should influence the seismic behavior of such plates upon subduction as proposed by p- nd Kanamori.

It is not at all difficult to appreciate that very yount ces (< 10 Ma) 1 j

are physically (thermally) anomalous (Fig. 11).

l 1

that it is subducting an abandoned oceanic spreading ridge (ca. 17-24 Ma) )

that is oriented at right angles to the trench. It is conceivable that this geologic factor may influence the degree of coupling between the two plates.

Both the Nankai trough and the Cascadia zone contain thick sections of marine sediments, although the flat-floored Nankai trough is still a closed depression several hundreds of meters below the adjacent Shikoku Basin (Aoki and others, 1982; Fig. 12). In western portions of the trough the sedimen-tary fill varies from 100 to 1000 m thick. In the east-central trough the axial sediuentary thickness is 1300 m, of which the lower 500 m is Shikoku Basin sediment (ibid.). The Shikoku Basin, unlike the Cascadia Basin, is bathymetrically complex (Fig. 12). Numerous seamounts and an abandoned spreading ridge oriented at right angles to the Nankai trough could provide "windows" through the basin's sediment cover that would enable effective advective cooling of the eastern Philippine plate through hydrothermal circulation. Thus, the cooling histories of the Juan de Fuca and eastern Philippine plates may be significantly different, a difference that could accentuate their magnetic anomaly-defined age difference (8 vs.17-24 Ma) .7 Unlike subduction of the Juan de Fuca plate, subduction of the eastern Philippine plate beneath Japan not associated with Quaternary are volcanism.

Kanamori (1972) explained this absence of volcanism by concluding that sub-duction of the eastern Philippine plate beneath Shikoku and Honshu was a 7Yamano and others (1984) have recently co= pared the Nankai trough with the Cascadia subduction zone in terms of heat flow observed in sediments just seaward of the deformation front. Heat flow values as high as 130 M-/m2  ;

have been measured in the trough, only slightly below maximum values reported i in the eastern Cascadia basin (ca 140 Mw/m2; Scheidegger, 1984; Moran and i Lister, 1987). Contrary to the distribution of heat flow values on the Juan '

de Fuca plate, however, heat flow values in the Nankai trough are higher than j those measured on the oceanic plate farther offshore (Shikoku Basin). This  !

relationship suggests that the relatively high heat flow of the Nankai trough l compared with its offshore plate is related to local factors (eg. hydrother-mal circulation, trench magmatism, etc.), not to characteristic thermal l properties of the Philippine plate. )

i l

very youthful event, and that the downgoing plate has not yet reached levels deeper than 60 km (the lower Itmit of seismicity) . Sachs (1983), however.

l in reporting the work of others, especially Hirahara (1981), states that the top of the subducted oceanic place appears to dip downward to a depth of about 60 km below the island of Shikoku (ca.175 km from the trench), but then flattens horizontally beneath Honshu for an additional 280 km. Thus,

, the two lithospheric plates (continental and subducted oceanic) may remain in contact without an intervening asthenospheric wedge for perhaps as much as 450 km (Fig. 13). Not only could this geometry account for the absence .

of an active volcanic arc in southern Honshu, but it could account for the strong seismic coupling (alpha = 0.88-1) represented by the periodic (ca.120 years), great event seismic history of the shallow plate interface. Accord-ingly, a physical basis for a probable difference in coupling and maximum earthquake magnitude between the two subdection zones (SW Japan and the Pacific Northwest) is available that goes fac beyond similarity in conver-gence rate and general similarity in youthful age of subducted crust.

Southern Chile (Southern Nazca)

Heaton and Hartzell (1986, 1987) compare the Southern Chile (north of 46' S) l and Cascadia subduction zones on the basis of thick sediment fill in their buried trenches and similar continental margin physiography, including low coastal mountains, a central valley / inland sea, and a chain of active strato-volcanoes. This Chilean subduction zone ruptured in 1960 (Fig.14),

in the largest known historical earthquake (M, = 9.5) . Astiz and Kanamori (1987) report that seven large earthquakes (M range: 7 0-7.4) occurred between 1919 and 1949 within the latitudinal range of the 1960 event, but fault plane solutions could only be determined for the three youngest events (1934, 1940, 1949). They interpret the first and last of the three as resulting from down-dip extension within the subducting Nazca plate; the 1940 event appears to be a thrust-type earthquake that "probably occurred on the interplate boundary", but "could be an outer-rise event with a horizontal compression mechanism" (ibid., p. 1616). Kaizuka and others (1973) have reported three, still older large to great earthquakes that occurred in the latitudinal range of the 1960 event -- in 1575, 1737, and 1837 The nature 1

i 1

of the faulting responsible for these major earthquakes is obviously uncer-tain, but coastline coseismic elevational changes during the 1837 ovent suggests that it was an interplate, thrust-type event.

Jarrard (1986) assigns the upper plate of the Southern Chilean zone to the mildly compressional class 5, in large part because Andean volcanoes within one segment of the zone have flank vent distributions that indicate intrare compression crudely parallel to the direction of plate convergence. As in the case of SW Japan, the seismic coupling between the Nazca and South American plates is very high (alpha = 1.0, Kanamori and Astiz, 1985).

At least three major differences exist between the Southern Chilean and Cascadia zones. Firstly, although the 1000 km-long rupture zone for the 1960 earthquake propagated southward into an area of young (ca.15 < 4 Ma) subducting crust, it was initiated to the north on or near a major fracture zone and in an area of older (ca. 30-35 Ma) subducting crust (Fig. 14; Herron et al., 1981). It is not clear, therefore, what the behavior of this margin would have been, had very young crust been subducting along its entire length, as is the case for the Cascadia zone.

Secondly, the rate of plate convergence for this zone, ca. 9 cm/yr, is very fast when compared with that for the Pacific Northwest (ca. 3.2 cm/yr). Ex-3 cluding other seg=ents of the Nazca plate, only the Solomon and NE Japan subduction zones have higher convergence rates (12 and 9 9 cm/yr respec-tively; Jarrard, 1986, Table 1).

Thirdly, and perhaps the most import ant difference with respect to the Cascadia zone, the Southern Chilean margin appears to have a very narrow Cenozoic accretionary prism, despite the fact that the trench is sediment filled (Plafker, 1972; Scholl and others, 1970). According to Scholl and others (1970) crystalline continental rocks crop out on the sea floor within 100 km landward of the Peru-Chile trench (the segment which ruptured in 1960), and "undoubtedly underlie much of the continental shelf and slope."

It has been claimed that this continental margin has been affected by subduc-tion erosion, meaning that the downgoing plate rasps into the crystalline i

rocks of the continental plate rathee than accreting its sediment load onto that plate. If so, the "backstop" of this subduction zone (the landward limit at depth of the narrow accretionary prism) would lie at or very near the trench and the width, W, of the seismogenic plate interface between the backstop and the depths to the brittle-ductile transition would be anomalous-ly wide. Wang and others (1987, p. 1467) have developed a related concept by contrasting the thermal structure of the Cascadia and Peru-Chile continental margins:

We show that at the Oregon-Washington margin, where subduction accretion has been predominant, heat flow over the continental slope has been controlled by thickening of the accretionary prism, both through offscraping and underplating, and by the motion of the oceanic plate. Excess pore pressures in the thick subducted sedi-ments have probably cushioned the upper and the lower plates to reduce any frictional effects. At the Peru-Chile margin, on the other hand, subduction erosion was dominant until *i.e recent geolog-ical past. The absence of underplating and the presence of fric-tional heating at the decollement could be responsible for the observed high heat flow, giving a thermal structure greatly dif-ferent from that of the accretionary Oregon-Washington margin.

Thus, the anomalously large magnitude of the 1960 Chilean earthquake might be explained by special geologic factors, such as a very wide seismogenic plate interface (W), without relying solely on the Ruff and Kanamori empiri-cal relationship relating M,' to convergence rate and age of subducting crust.

Southwestern Colombia-Northernmost Ecuador Oceanic lithosphere of 8-17 Ma age is being subducted beneath southwestern Colombia and northernmost Ecuador (Fig. 15) at a rate comparable to that of southern Chile, ca. 8 cm/yr. Great earthquakes have occurred on this zone l in 1906 (M, = 8.8) and 1979 (8.2), with magnitude 7+ events in 1942 and 1958 l

(Heaton and Hartzell, 1986). The 1979 event was of thrust-type; the 1906 l

event was probably so, but not indisputably (Kanamori and McNally, 1982).

The physiography of the upper-plate Andean chain with three parallel mountain belts, the central of which has elevations up to 6000 m is dra=atically l

l different from that of western Washington and Oregon. This extreme Andean continental margin tcpegraphy, the shallow nature of the sediment-buried .

offshore trench (Jarrard,1986. Table 1), and the high level of seismicity of this region may be attributable to the ongoing subduction of seafloor asperities, one an abandoned late Miocene spreading ridge at 2 5' N that was active until 8 Ma. The ridge trends at right angles'to the trench with a '

geometry comparable to that described for the subducting eastern Philippines plate along the Nankai trough (Fig. 15). The tectonics of the Colombia subduction zone are further complicated by the subduction at it southern end, O' N of the aseismic Carnegie Ridge on the Nazca plate. For example, Kanamori and McNally (1982) believe that the 1906 earthquake, which may have had a rupture length of 500 km, propagated northeastward from the area of collision of the Carnegie Ridge with the subduction zone. Rapid subduction of both an abandoned spreading center and an aseismic oceanic ridge estab-lishes a distinctive tectonic setting for this area that should not be compared with the much slower subduction of an abyssal plain (Cascadia Basin) beneath coastal Washington and Oregon.

The Colombian subduction zone and two other Nazca plate segments to the south, Ecuador and Peru, are all assigned by Jarrard (1986) to strain class a

6, moderately compressional, The value of alpha for this zone is ca. 0.85 according to Kanamori and Astiz (1985). These indicators of a high degrao of interplate coupling are not surprising given the asperities on the sub-ducting Nazca plate described above. South American intraplate seismicity above these segments generally reveals east-west compression; back-arc deformation is characterized by the development of an east-vergent foreland fold and thrust belt between the Andes and the Brazilian shield (Burchfiel and Davis, 1976; Jarrard, 1986). Because of differences in plate convergence rate, continental margin physiography, and the occurrence of major asperities on the subducting Nazca plate, comparisons of the highly seismogenic Colombian zone with the Cascadia zone (Heaton and Hartzell, 1986; Rogers, in press) seem completely inappropriate to us.

1 4

I

l Rivera l

l

[ The Rivera plate, a small young plate like the Juan de Fuca, is being sub-l ducted beneath the western Mexican state of Jalisco (Fig. 16). With an esti-

mated cos, vergence rate of 2 3 0.6 cm/yr (Minster and Jordan, 1979) and an age of subducting ocean lithosphere of 8-10 Ma (Klitgord and Mammerickx, 1982), the plate has v and t parameters closely resembling those of the Juan de Fuca plate. Like the Cascadia zone, the Rivera segment of the Middle America subduction zone is seismically quiescent (Eissler and McNally, 1984),

although Singh and others (1985) attribute the 1932 Jalisco earthquake (M, =

8.2) to Rivera plate subduction. Singh and others (1981) believe that other major (M > 7) earthquakes occurred in the Jalisco region in 1837, 1875, 1900 (2 events), and 1911. They consider them to be interplate (presumably Rivera/ North American) events, but their hypocentral depths and exact loca-tions are not known. Given uncertainties (see below) regarding the epicen-tral location of even the 1932 Jalisco event, a definite assignment of the 1837-1911 earthquakes to subduction of the Rivera plate (rather than to sub-duction of the adjacent Cocos plate) seems inappropriate. As one exa=ple, the June, 1911, earthquake (M, 7 9) has recently been related to Cocos plate aubduction (UNAM Seismology Group, 1986).

Despite similarities in subduction parameters t and v. there are important geologic differences between the Rivera and Juan de Fuca plates. The Rivera plate has only a relatively thin cover of sediments, and the Middle America trench is topographically well-defined. The axis of the trench adjacent to the Rivera plate varies from a depth of more than 4000 m at its northern end to more than 4800 m at its southern; the height of the trench's outer slope similarily increases southward from about 700 m to 1500 m (Shipboard Scienti- I f'.c Party, 1973, DSDP Leg 18; Mammerickx, 1985). With the exception of Tres Marias basin at the northern end of the Middle America trench, the trench is "narrow and has only small accumulations of recent sediments in the botto="

(Shor, 1974, p. 595). At DSDP Site 473, only about 50 km west of the Middle America trench at Tres Marias basin (Fig. 17), approximately 250 m of late Cenozoic sediments overlie an irregular basement of diabase and basalt. The sediments are no older than 6.5-8 Ma (Late Miocene). Sediment accumulation

rates for 0-3 Ma-(130 m total) and > 3Ma (120 m) are 40m/Ma and 20m/Ma respectively (Shipboard Scientific Party, 1981, DSDP Leg 63). In contrast, at DSDP Site 174 in the eastern Cascadia Basin, the thickness of Pleistocene and Pliocene sediments is 911 m and the average rate of sedimentation is ca 200m/Ma (Shipboard Scientific Party, 1973. DSDP Leg 18); Barnard (1978) estimates Pleistocene rates of sedimentation in lower slope basins above the filled Cascadia trench to be 500-1000 m/Ma.

Unlike the topographically smooth-floor of the Juan de Fuca plate, which is at least in part a function of thick sediment cover within the Cascadia basin, the Rivera plate is topographically complex due to the presence on it of many seamounts (Fig. 17). These seamounts may play an important role in the thermal history of the Rivera plate by acting as conduits through the sedimentary cover for hydrothermal circulation and cooling of the oceanic lithosphere. Heat flow data for the Rivera plate is limited to measurements made at Site 473, where the thermal gradient in sediments was estimated to be 64* C/km and the estimated heat flow is 1 3 HFU (Shiboard Scientific Par-ty, 1981. DSDP LeF 63); the temperature at the base of the sedimentary sec-tion is believed to be less than 20' C. In contrast, the temperature at the base of sediments in the eastern Cascadia Basin may be as high as 200* C.

Although Rivera data are quite limited, the thermal history and present thermal state of the comparably young Juan de Fuca and Rivera plates seem to be quite different. In other words, despite comparable t values (ca 8-10 Ma) based on magnetic anomalies, the effective thermal age of the Juan de Fuca plate appears to be significantly younger than that of the Rivera.

Given this conclusion and expextable differences in the amounts of sediment being subducted along the Cascadia and Rivera subduction zones, there is no a priort reason to expect similar seismic behavior of the Cascadia and Rivera subduction zones (cf. Sammis et al, 1988).

Heaton and Hartzell (1986, p. 697) have argued that the 1932 Jalisco earth-quake is proof of the seismogenic nature of the Rivera/ North American plate interface

! "there is little doubt that these events" (the 8.2 main shock and an j 7.8 aftershock) "represent thrusting of the Rivera plate beneath the

, l

North American continental margin. ... Although there are clear dif-ferences in the physiographic features and the frequency of moderate-  :

ly latge coastal earthquakes between subduction of the Rivera and l Juan de Fuca plates, the earthquake history of the Jalisco regi n l clearly demonstrates that the slow subduction of a small, very young plate does not necessarily imply aseismic slip."

These conclusions seem over-stated. Although we do not deny the possibility that the Rivera subduction zone is seismogenic, there is indeed doubt that i

l j major historic thrust events along northern portions of the Middle American l subduction zone are related to subduction of the Rivera plate. The 1932 earthquake is located by Kelleher and others (1973) and Eissler and McNally (1984) at positions in western Mexico close to the projected boundary (Rivera fracture zone) at depth between the Rivera and Cocos plates (Fig. 17). The latter authors raise the possibility that the 1932 earthquake may have broken "the northernmost section of the Cocos-North American plate interface, as opposed to the Rivera-North American plate interface." Petroy and Wiens (1986) later concluded that the boundary between the Cocos and Rivera plates is so ambiguously located that subduction of either plate could have been  !

responsible for the Jalisco event. Singh and others (1985) portray the af tershock area of the 1932 earthquake as predominantly adjacent to the Rivera zone, but K. McNalley (personal conmunication to G. Davis, 1987) believes that the locations of the aftershocks are so poorly constrained as i I

to be nondefinitive.

Given uncertainties in the location of the 1932 earthquake, it is not reason-l able on the basis of existing data to conclude that the now nearly aseismic subduction zone of the Rivera plate experienced great earthquake rupture in 1932. Given also that the Cocos segment of the Middle America trench has l

experienced 42 earthquakes with M)7 this century ( Anderson and others, 1 1985), the association of the 1932 event with the northwestern edge of the l Cocos plate would not be surprising. In summary, despite similarities in  !

their magnetic anomaly ages (t) and rates of convergence (v), comparisons of the seismogenic potential of the subducting Juan de Fuca and Rivera plates j must take into account (1) important differences in their sedimentational l

and thermal histories, and (2) uncertainties in the seismic history of the I

l Rivera plate. We believe that these differences and uncertainties make the case for aseismic subduction of the Juan de Fuca plate no less convincing.

Northern Cocos (Southwest Mexico)

The northern Cocos plate (Fig. 16) is another plate with some similarities in its subduction parameters to the Juan de Fuca plate. Gecanic lithosphere being subducted directly south of the Rivera fracture zone is quite young, about 3-5 Ma (K11tgord and Mammerickx,1982), although northern Cocos litho-sphere along most of the Middle American trench is older (ca. 10-20 Ma, ibid.) The convergence rate with North America at 17' N is 5 9 cm/yr (Jarrard, 1986).

Subduction of the northern segment of the Cocos plate between the Rivera fracture zone (18 5* N) and the Tehuantepec Ridge (15' N) is characterized by greater seismic activity than any other subduction zone in the western ,

hemisphere (Anderson and others, 1985), including the 1985 Mexico City (or Michoacan) earthquake (M, 8.1). The possibility exists that subduction of young, hot Cocos lithosphere south of the Rivert fracture zone produced the M, 8.2 Jalisco earthquake in 1932, although some workers (see above) assign that event to Rivera plate subduction; alternatively, it is may have been localized along a subducted asperity -- the fracture zone separating the Cocos and Rivera plates. The descending northern Cocos slab, responsible for 42 large (M)7) interplate earthquakes since 1900, is defined by a shal-i low-dipping (20-30') Benioff zone that extends to a depth of 90-100+ km

(Jarrard, 1986; Burbach and Frohlich, 1986). Major earthquakes (M)7) have occurred opposite a given portion of the continental margin along the north-4 ern Cocos subduction zone every 30 to 75 years (Singh and others, 1981; Astiz and Kanamori,1984), except, until 1985, along the segment beneath j Michoacan province'. This seismic gap was terminated in 1985 by the great Mexico City earthquake; that event may have been localized along a subducting asperity, Orozco fracture zone (Spence, in review).

Jarrard (1986) "tentatively" assigns the upper plate of the subduction zone j to strain class 6 (moderately compressional), but his reasons for doing so i

! l

! l

)

l l

l i

- -,-.,-.--.---.----..-.------.c., . - , , . , , . - ~ , . . - , - , - , . , - . - , . , . - . - . , . - . - - . ,,,,. . , .,--...-.-.n.

~ _ . . . - . - . .. .. . -. - - -

are not clear. Kanamori and Astiz (1985) report that several investigators agree that alpha for the Mexican subduction zone is about 0 5. The high degree of seismicity along the 1200 km-long N. Cocos /NA subduction boundary is quite remarkable, given its broad (but not close) similarity in v and t parameters with the seismically quiescent JdF/NA plate interface along 800 km of Cascadia zone during the past 150 years. This profound difference in  ;

subductive behavior suggests the existence in the Juan de Fuca case of some' '

special factor or factors inhibiting the seismic character of plate conver-1 gence -- most likely, as discussed above, the effects of heavy sedimentation in the Cascadia basin. In contrast, sedimentation associated with the Middle America trench is sparse. The trench along the Northern Cocos plate is topo-graphically V-shaped with axial depths that vary southward from ca 4500 m to ,

5400 m, and an outer slope at its southern end with 1800 m of relief (Mammerickx, 1985; Connard and others, 1985). Based on seismic reflection profiles the trench contains only a discontinuous, relatively thin sedimen-tary fill, e.g., ca 250-500 m between 98.5-100* W longitude; the offshore ,

plate has an even thinner sedimentary section (e.g., < 180 m for sediments younger than 8 Ma, Site 487, DSDP Leg 66) . Given the paucity of offshore >

plate and trench sediments it is not surprising that the accretionary prism

, adjacent to the Northern Cocos plate is very narrow. Moore and others (1982)

report that crystalline basement rocks are shown, by magnetic, seismic, and drilling data, to underlie the mid-continental slope and te lie within 35 km of the trench axis. The extre=ely narrow accretionary wedge in this area is believed to be entirely of Neogene (late Tertiary) age; no Paleogene or Meso-zoic forearc or accretionary prism deposits appear to have been preserved along this margin. The narrow width of the Cocos accretionary prism clearly sets the Cocos subduction zone apart from the Cascadia zone in terms of probable width (W) of the seiscogenic interface between the prism backstop j and the brittle / ductile transition in the upper plate at greater depth). It is likely that, as in the case of the S. Chilean subduction zone. W for the

{

i Northern Cocos subduction zone is anomalously wide.

l i

}

l

~

l Southernmost Chile (Tierra del Fuego)

Southernmost Chile (46' 54' S. latitude) provides an exa=ple of a seismically quiescent subduction zone with a number of interesting similarities to the l Cascadia zone. This zone separates the Antarctic and South American plates j and lies south of the segmented East Chile Rise, the east-west boundary l t

between the Antarctic and Nazca plates (Fig. 18). Similarities with the l Cascadia zone include low convergence rate (2.1 cm/yr, Minster and Jordan, 1978), young subducting lithosphere (ca. 8-10 Ma), a sediment-filled trench, an absence of great historic earthquakes (Lomnitz, 1970; Kelleher, 1972) and thrust-type focal mechanisms (Forsyth, 1975), and a virtual absence of l seismicity north of 52' S (Forsyth, 1975). Major differences are that the subducting Antarctic plate is much larger than the Juan de Fuca plate, and I that the geologically recent subduction of spreading segments of the East Chile Rise along the northern part of this margin has undoubtedly complicated its thermal history. For example, there is no active volcanic arc opposite the northern part of this convergent zone (46*-49' S) where the spreading center has been subducted within the past 6 Ma, Unfortunately so little is know about the southernmost Chilean zone, either geophysically or geologi-cally (Jarrard, 1986, Table 1), that it offers us few insights into the aseismic vs. seismic nature of the Cascadia zone.

Conclusions from Subduction Zone Comparisons l l

We have compared th Cascadia subduction zone with the six subduction zones or zone segments selected by other workers (Heaten and Hartzell, 1986, 1987; Rogers, in press) as most similar to the Cascadia zone in terms of youthful age (t) of subducting crust and rate of plate convergence (v). Such compar-isons can neither prove nor disprove the potential for great earthquakes along the Juan de Fuca/ North American plate interface, but our comparison of geologic factors (other than t and v) indicates that there is a scientific basis for concluding that subduction of oceanic lithosphere along the Cascadia zone may be occurring in a manner unlike that of the other zones with which it has been compared. Specifically, geologic differences with l the six seismogenic zones supports tt.e conclusion that slow subduction of I

. . ~ . .. . .

the hot, topographically smooth, and sediment-laden Juan de Fuca plate could be occuring by stable sliding rather than . stick-slip behavior. Table 1 sum-marizes the comparative t and v parameters of the seven subduction zemes,

)

and cites major geologic differences between the Cascadia zone and each of the other six. It is interesting to note that values for two independently derived measures of coupling between converging plates, Jarrard's strain class numbers based largely on geologic characteristics of the upper plate, and alpha (the seismic slip rate / relative total plate convergence rate),

are both considerably lower for the Cascadia zone than for any of the other zones with which it is compared. Such low values are consistent with geo- i logic arguments proposed herein that W, the width of the seiscogenic plate interface, may be very narrow or nonexistent for the Cascadia zone.

l Geodetic and Geologic Evidence for and / gainst Great Cascadia Zone 1 1 Earthquakes Strain and Stress Within the North American Plate. Pacific Northwest Down-to-the-east crustal tilt along 1400 km of continental margin in south-western British Columbia, western Washington, and western Oregon has been documented by precise leveling, analysis of tide gauge records, and studies of tilted coastal marine terraces (Ando and Balazs, 1979; Reilinger and l Adams, 1982; Riddihough, 1982; Adams, 1984). Uplift of the cuter coastal areas (up to 2-3 mm/yr in Washington and British Columbia) is accompanied by subsidence (1-2 mm/yr) east of a hingeline that extends southward from Hecate Strait, through the Strait of Georgia Victoria, and the eastern edge of the Olympic Peninsula (Riddihough, op. cit.; Reilinger and Adams, op.

cit.; Adams, op. cit.). Ando and Balazs (1979) interpreted the tilting pattern as an expression of continuous aseismic subduction of the Juan de Fuca plate. Reilinger and Adams (1982) and Riddihough (1982) favored this interpretation (while admitting that alternative interpr4 stations were pos- l sible), but Adams (1984, p. 467) concludes that contemporary landward tilting )

t and shortening of the western continental margin in the Pacific Northwest  ;

provides no "definitive evidence as to whether the subduction causing the deformation is steady or episodic, aseismic or seismic."

1 49-I l 4

- -- . . . . - -. n,m...,. ,,, .m%__.y_ ..3, _ . c. .._,- - ---.f, e ,, - - . ,m.,~r_..,.w, ,-. ,

TABLE 1.

SUMMARY

OF SUBDUCTION ZONE COMPARISONS Zone (Jarrard. '86 Plate age Convergence strain classl at trench rate (cm/yr) Geologic differences with Ctscadia Cascadia 8 3.4 Alpha = 0 3 (Kanamori and Astir.

(3-4a) 1985); 0.0 to < 0.3 (this report).

SW Japan 17-24 2 9-3 7 Major topographic asperities on sub-(5) ducting plate (abandonad spreading ridge); no active volcanic arc; upper and lower plates in possible subhori-zontal concact beneath Shikoku and Honshu;. alpha = 0.88 1.0.

S Chile 5-35 9 Very high convergence rate; 1960 M 9 5 (5) event initiated in subduction zone where oldest crust is being subducted; accretionary prism very narrow, so width of seismogenic zone (W) may be very wide; alpha = 1.0.

SW Columbia 8-17 8 High convergence rate; zone is highly (6) seismogenic; major asperities (an aban- 1 doned spreading conter and Carnegie l Ridge) on subducting plate make this I a collisonal zone; upper plate under strong compression; alpha = 0.85 Rivera 8-10 23 Topographic asperities (seamounts) on (7) Rivera plate; subducting plate ha.s only thin sediment cover; trench is well-defined with only localized ponded l sediments; narrow (?) accretionary i

. prism =-wide (7) alpha = 7. I N Cocos 3-20 59 zone is exceedingly seismoganic (42 M (6) 7+ events since 1900); subducting plate has only thfn sediment cover; trench is well-defined with only localized ponded sediments; very nar-row accretionary prism = wide (7)

, alpha = 0 5 l

Tierra del 8-10 2.1 No active volcanic are opposite  !

Fuego (?) northern part of zone where an active spreading ridge was subducted during past 6 Mat therral characteristics of

this segment presumably ancmalous; alpha = ?.

Geodimeter surveys in the Puget Sound basin conducted between 1972 and 1979 by the U.S. Geological Survey were reported by Savage end others (1981) as indicating crustal shortening at a rate of about -0.13 2 0.02 u strain /yr in a N 71'E direction, with extension of about half that rate in an orthogonal (NNW-SSE) direction.8 They interpreted these measurements to be consistent with preseismic strain buildup due to convergence of the North American and Juan de Fuca plates and a locked condition of the shallow-dipping contact between them. Crosson (1986) pointed out that the USGS strain data have a significant component of non-linear time behavior that suggests that there was areal dilatation until 1975, followed by contraction on all lines meas-ured on the western side of the Puget basin. This time dependent behavior indicates a complex character of the strain field, and complicates the simple picture of uniform accumulation of compressional strain due to a locked subduction zone. .

In response to Crosson (1986), Savage and others (1986) attribute the areal dilatation to systematic observational error. They note that the engineering shear strain, which is insensitive to dilatational components, is consistent with compression in the direction of plate convergence. Morc recently, Lisowski and others (1987) and Lisowski and Savage (in prep.) have updated the measurements in the Puget Sound network, reoccupied triangulation sta-tions along the Strait of Juan de Fuca, and compiled data from additional Canadian observations. These measurements confirm very low ENE-WSW contrac-tional strain rates for the entire region (-0.03 ustrain/yr in both Puget Lowlands and Olympic Mountains), or alternatively, very low extension perpen-dicular to this direction. However, the newer strain rates are appreciably less than those originally reported by Savage and others (1981; see above).

i The newer strain values are said to be "roughly reproduced" by a dislocation model for the Cascadia subduction zone that has the plate interface locked, presumably up-dip, to a position near the Washington coastine.

sDue to oblique convergence and the orientation of the North American plate  ;

margin, the theoretically predicted direction of sigma hmax resulting from j the N 50o E convergence direction of the JdF and NA plates is approximately N 70o E; i.e., there is a tendency for the principal compressive stress axis to be rotated orthogonally with respect to the convergent plate margin l (Savage and others,1981; Jarrard,1986a) . i i

! Unfortunately, there is no way to determine if the strain accumulation l determined by geodimeter surveys is purely elastic or represents anelastic (permanent) deformation. Thus, it is difficult to utilize horizontal strain data alone to provide definitive evidence of the presence or absence of a j locked JdF/NA plate interface.

Melosh (1987) has constructed finite element models for the Pacific Northwest :

that incorporate both uplift data from western Washington and the compressive strain rates published by Savage and others (1981) for the Puget trough.

Assuming a downward bend in the subducting JdF plate at a depth of about 40 km, he concludes (p. 1240) that:

"None of the runs with steady slip were able to reprcduce both the observed uplift and compressive strain. Steady slip models that fit the uplif t data predict a compressive strain 20 times smaller than observed, whereas models that fit the strain data predict j uplift rates in excess of 50 mm/yr. Models in which the fault is presently locked fit the data best, although they predict a rather complex pattern of uplif t and strain. Such models may be tested i by further strain measurements to determine whether the predicted

! pattern is present."

i Melosh (1988) conducted a revised finite element study of strain and uplift in the Pacific Northwest using the revised, lesser rates of ENE-WSW contrac-tional strain reported by Lisowski and others (1987) across the Puget Low-i lands near Seattle (-0.03 + 0.01 p strain /yr vs earlier -0.13 s 0.02 p j strain /yr). Anomalously, Melosh concluded that a locked plate interface would most likely occur west of the Washington coastline and that th2 portion beneath Puget Trough and farther inland, may be slipping aseismically. This conclusion is in direct opposition to that of Savage and others (1981) who favored a lockad interface beneath western Washington only about as far west as the coastline. It is likely that Melosh's models are extremely sensitive to the configuration of the subduct.tng plate beneath the Puget Sound region, j In the most definitive study of the geometry of that slab to date, Crossen and Owens (1987, p. 826) conclude that the slab "may dip more steeply to the east after reaching depths of 60-70 km", considerably deeper than the 40 km

deep bend assumed by Melosh (1987) for modeling purposes. For reasons I

l l i ,

1

)

)

- . ~ - _ . , - - . -

summarized above, Melosh's conclusions regarding a locked plate interface  ;

should be viewed with considerable skepticism.

The geodimeter data of Savage and others (1981) and Lisowski and Savage (1987) suggesting ENE-subduction-related compression of the crust in the Seattle area are not easily reconciled with stress orientatica data obtained from fault plane solutions of upper-plate (North America) earthquakes in the Pacific Northwest (cf. Crosson,1972; Rogers,1979; Yelin and Crosson,1982:

Yelin, 1983; Crosson, 1983). Such solutions characteristically indicate an approximately N-S orientation (t 10*, R. Crosson, personal communication, 2/1988) for the P axis of fault plane solutions, which is widely believed to approximate the orientation of the principal compressional stress axis (sigma h= ax ) . 9 Most such solutions indicate either faulting along E-W-striking thrust faults or strike-slip faulting along NW-striking faults (dextral) or NE-striking faults (sinistral).

9A possible contradiction to the ubiquitous NS compression observed from upper plate focal mechanism solutions in the Pacific Northwest is the recent recognition of a 90 km-long, N 15' W-trending crustal earthquake zone in southwestern Washington -- the St. Helens seismic zone of Weaver and Smith (1983). Focal mechanism solutions for earthquakes in this zone (including the 1981 Elk Lake earthquake, M 5 5), indicate right-slip along planes orien-ted practically NS, corresponding to a P axis that is oriented NE-SW. Weaver and Smith interpret the northeast-trending P axes as indicating a locked JdF/NA plate interface southwest of the St. Helens seismic zone, although not necessarily extending as far to the west as the Coast Range. The zone itself is hypothesized to separate a western, locked plate interface from a sliding interface to the east. Weaver and Smith suggest that the St. Helens zone is a manifiestation of a previously unrecognized stress regime in the Pacific Northwest. We question this assertion. If earthquakes in the zone are occuring along a preexisting steep fault or faults, then the analysis of McKenzie (1969) tells us that sigma hmax (the principal axis of compressive stress) may lie virtually anywhere in the dilational quadrants of the strike-slip mechanisms presented by Weaver and Smith (cf. Jarrard, 1986, p. 225:

"motion on a preexisting fault is not a reliable indicator of exact stress direction."). Thus, sigma hmax could trend from nearly N-S to ENE-WSW.

While we cannot rule out the stress model presented by Weaver and Smith, a simpler model of more uniform NS compressive stress in the North American plate is not negated by the occurrence of right-slip along the St. Helens

seismic zone.

i t

i pr e< - = * *-wm- !z---z--->.---w---w -

-ge-we m%-- -------wyw - w w r= m ' --'N-'*--9-rr-

---rm., g-sq=-in-ry-----w-rww--

It has been argued that north-south- compression in portions of the North American plate above the subducting Juan de Fuca plate indicates a lack of 4

coupling between the two plates with their N 50' E direction of convergence.

A number of workers (among them Crosson, 1972; Davis, 1977: Sbar, 1982: Li and Crosson, 1988), although Spence (in review) takes the opposite view that the_ observed stress patterns are compatible with strong coupling between the offshore plate system and North America. There is-s) unfortunate absence of upper-plate focal mechanism solutions in the Oregon and Washington Coast Ranges above the shallow-dipping segment of the subducting plate (one N-S compressional mechanism is known from an onshore coastal earthquake west of Salem, Oregon; Dehlinger and others, 1970). Hence, the upper plate stress field in cortinental margin areas is not defined by fault plane solutions where the question of a locked or aseismically slipping plate interface is most pertinent.

Sbar (1983) has proposed that strains released by earthquakes develop over significantly longer periods of time than the geodimeter surveys of Savage and others (1981), and are probably more reliable for tectonic studies than

! strain measurements taken from such surveys. The latter, he believes, may i record short-term strain fluctuations that are averaged out during longer periods of pre-seismic strain accumulation, Adams (1984) presents an alter-native hypothesis. He proposes that the stress field along the continental i

margin of the North Acerican plate is probably one of NE-SW compression, but that farther east an attenuation of subduction-related stresses occurs and a N-S compressional field resulting from Pacific / North American plate interac-  ;

tion is saan. Adams discusses the problem of conflicting strain (and, there-l fore, stress) orientations from earthquake fccal mechanisms vs. geodimeter surveys at some length. Without reaching a conclusion as to the coupled versus uncoupled nature of the JdF/NA plate interface he states (1984,

p. 465-466):

"Is then the geodetic strain measured at Seattle episodic, accumu-lating and being released over a period of a few years as nearly smooth aseismic subduction occurs, as suggested by Sbar (1983]? Or is the accumulated NE-SW compressive scrain currently too low to 4

generate many earthquakes, because much previous strair. was ,

released in some large prehistoric earthquake? ... At the 2 x l

} ,

l 1

l

{

l l

I j 10-7 yr-1 rate of shear strain accumulation at' Seattle, earthquakes ,

that need -(40-100) x 10-ti of shear strain (as proposed by Sbar I (1983)) would require 200-500 years of accumulation." If great l earthquakes had such a long return period it would explain their l absence in the short historic record. ... In the absence of such  !

earthquakes, those earthquakes representing the release of north-south strain could dominate the recent seismicity."

L i

,' Coastal Geomorphology and Neotectonics Major historic thrust earthquakes along subduction zones are characteristi-cally accompanied by coseismic and long-term uplift and subsidence of the overlying plate. These deformations are largely the consequence of the

, release, by interplate thrust faulting, of elastic deformation accumulated during preseismic peric.ds of strain buildup during plate convergence. The most impressive examples of coseismic elevational changes are those that accompanied the giant 1960 Chilean and 1964 Alaskan earthquakes (Plarker, 1972). In both cases, linear zones of deformation (uplift and subnsidence) extended the length of the ruptured zones (ca 900-1000 km). The zones of uplift occurred closest to the trench axis, whereas subsidence occurred landward; depending on its distance from the trench, the coastline experi-enced either coseismic uplift or subsidence (Fig. 19). According to plafker ,

(ibid.) major uplift and subsidence extended over a region of 140,000+ km2 in Alaska and 85,000+ km2 in Chile. Maximum uplifts in Alaska and Chile, j respectively, were 11 3 Ind 5 7 mi maximum subsidence for the regions was j 2 3 m and 2.7 m, respectively. Figures 20 a and 20 b are selected profiles

! of the coseismic deformation that occurred in Alaska and southern Chile, f They also indicate the appropriate location of the Cescadia subduction zone i coastline for comporison with the zones of coseismic deformation accompanying )

these earthquakes.

I

" This geodetically-based estimate of major earthquake recurrence interval

! was based on Seattle area strain data from Savage and others (1981). civen the revised, lower NE"SW contractional strain rates for this area (Lisowski I and others,1987), a recalculation of recurrence interval should yield I significantly longer times, i

1 4

I I

I

The pattern and amount of coseismic elevation changes, both in map and cross-sectional view, cppear to be strongly sensitive to the geometry of the thrust-faulted plate interface, the structural characteristics of the accre-tionary wedge, and the magnitude of slip at the time of rupture. From a study of coastline coseismic elevational changes n elated to fourteen large magnitude thrust-type earthquakes along nine subduction zones worldwide

( Alaska, S. Chile, Colombia, New Hebrides, Nankai, Sagami, Kuriles, Mexico, and Makran), West and others (1987), and West and McCrumb (1988, unpublished daca) have compiled information on the geography of uplifted and subsided areas with respect to earthquake magnitude, distances between the adjacent trench and the coastline, and depth to the plate interface. These data indicate that:

(1) The maximum amount of coseismic uplif t and subsidence that occurs along the coastline roughly increases with increasing magnitude (Fig. 20c). For great earthquakes (M, 8+), the coastline may typically experience up to 4.0+ m or uplif t and/or 2.0 m of subsidence; (2) Coseismic coastline uplift has occurred from 26-60 km from trench axes; coseismic subsidence has occurred no closer than 108 km from the trenches (Fig. 20d). For reference Cascadia trench-coastline distances vary from 60 km at Cape Blanco, Oregon, to a maximum of 140 km at Willapa Bay, Washington; (3) In terms of depth to the interface at the coastline, coseismic uplift generally occurs 12-36 km above the interface, whereas subsidence appears to be restricted to areas where the coastline is 26-50+ km above the interfaced (Fig. 20e). Fer reference, the Oregon and Washington coastline vary from about 15-26 km and 19-30 km, respectively, above the Juan de Fuca/ North American plate interface; (4) Commonly, more than 50% of the coseismic deformation is preserved I as permanent uplif t or subsidence. Zones of historic, coseismic uplift are characterized by multiple, uplifted Holocene and Pleis- ,

tocene marine features (terraces, wave-cut platforms, strandlines) .

Zones of historic, coseismic subsidence have preserved submerged late Holocene stratigraphy, although even these zones ultimately undergo lens-term uplift. a l

l Given the common occurrence and characteristics of episodic coastline defor- I wacion above other subduction zones, the preservation in the geologic record of Holocene elevational changes along the Cascadia coastline, either positive I

i

or negative, could indicate the geologically recent occurrence of major, interplate thrust earthquakes along the shallow-dipping portion of the JdF/NA plate interface. Given evidence for coseismic vertical deformation, the present seismic quiescence of the plate '. aundary would -best be inter-preted as due to a strongly locked or couried interface. Recent coastal studies have yielded information regarding Holocene elevational changes and l their possible linkage to major seismic events. These studies, in part i

i contradictory, are discussed below.

e Evidence for Holocene Coastal Subsidence Washington The strongest evidence for Holocene subsidence possibly related to coseis=ic deformation comes primarily from the work of Atwater (1987), Atwater and others (1987), and Hull (1987) in the Willapa Bay and Grays Harbor areas of southwestern Washington (Fig. 21a). Atwater (1987) has discovered buried peaty deposits and lowland soils within coastal estuaries throughout.m1 area at least 40 km long and wide. Atwater and others (1987) believe that south-western Washington coastal areas have experienced five major episodes of regional subsidence in the past 3100 sidereal years. Peaty layers up to 0.2 m thick, are overlain by muddy deposits 0 5-1 m thick and, in some cases, by thin (< 10 cm) sandy interbeds. One sandy interbed extends 3 km up the Niawiakum River, Willapa Bay, becoming thinner and finer grained upstream --

characteristics that suggest a bayward (oceanward) source for the sandy I

sediment. Six buried peaty layers were cored along tht tidal estuary of Niawiakum River by Atwater (1987). These layers represa.nt former marsh -

j vegetation, including peripheral conifers, that grew under "nearly supratidal

)

coriditions". They are separated by muddy-deposits-containing intratidal plant remains (Triglochin marittma) . Eight buried "lowland" soils younger.

than 5000 radiocarbon years BP have been found in cores and exposures along South Fork Willspa River to depths of 5 5 m (Hull, 1987). The soil A hori-zons are generally composed of partially decomposed tree trunks and roots surrounded by forest litter. Sharp contacts with ovaalying estuarine muds suggest rapid burial.

1 Atwater (1987), Atwater and others (1987), and Hull (1987) interpret the estuarine Holocene sequences of southwestern Washington as resulting from:

(1) the rapid, presumably coseismic, subsidence of marsh vegetation; (2) their subsequent burial beneath intertidal muds; and (3) the gradual shoaling ,

of the intertidal deposits and the reestablishment of supratidal marsh i lands. Repeated subsidence of at least 0 5 m is said to be necessary to j produce the observed stratigraphic juxtapositions (Atwater, op. cit.). The ,

1 sandy deposits, which locally occur directly above three of the six Niawiakum l

River peaty layers (and above some peaty layers in other estuarios), are interpreted by Atwater (op. cit.) as probable tsunami deposits carried inland from offshore sand sources. Analysis of the sedimentology of the youngest sandy silt layer in the Niawiakum River's tidal marsh (ca 300 yr.

old) leads Reinhart and Bourgeois (1987, p. 1469) to conclude that it "was transported inland by a single large-scale process consisting of several pulses. Each pulse may have been separated by periods of several hours. Possible process models include storm surges, seiche, flooding (storms at flood times) and tsunamis generated by earthquakes at the Cascadia subduction zone."

Oregon Buried late Holocene marsh deposits along the northern Oregon Coast (four beneath Netarts Bay, one beneath the Salmon River estuary near Lincoln City, and another to the north at Nehalem Bay; Fig. 21b) are similarly interprc. d as evidence for rapid tectonically-induced subsidence (Darienzo and Peterson, 1987; orant and McLaren,1987) . Coring to 5 to 7 m depth in a small marsh l

in Netarts Bay near Tillamook encountered four sediment-defined burial cycles closely similar to those seen in southwestern coastal Washington l (Darienzo and Peterson,1987) . Buried peat layers contain diatom fauna that 1

record the transition from marine marsh to freshwater marsh conditiona. thus j establishing paleo-sea level horizons and indicating that the subsidence i events had 1 m dis ~ placements. Abrupt contacts between the tops of the peaty l

layers and overlying sand deposits 3-7 cm thick suggest rapid subsidence and '

burial. Marine silts and clays (0.2 1 m) overlie the sands. Radiocarbon dates from five cores through one correlated peaty deposit ranged in age from 750 ro 1240 + 80 years BP, a relationship described as possibly due to lateral progradation of the marsh (Darienzo and Peterson, 1987).

The Salmon River peaty horizon is overlain "abruptly" by a thin (0-5 cm) sandy deposit, that thins landward, and 0 30 .90 m of higher silt anc very fine sandy silt. Although admitting that the basal sand layer might have been deposited during a great storm, Grant and McLaren (p. 1239) prefer the interpretation that it is a tsunami deposit because it directly overlies a marsh deposit interpreted as subsiding coseismically during a "great Cascadia earthquake". [We believe that the reasoning here is somewhat circular.] Sand and silt deposits also overlie peaty soil of a buried marsh at Nehalem Bay.

Grant and McLaren (op. cit., p. 1239) claim that these clastic deposits

locally contain T. marttima which "probably [our italics] represents a lower position in the intertidal zone than does the peaty soil of the buried marsh."

In contrast to the conclusions drawn by investigators from the northern Oregon coast, the results of preliminary coring in five estuaries along the south-central Oregon coast suggest to Nelson-(1987, p.1240) "that relative sea-level rise during the late Holocene was gradual with no abrupt changes in the type or rate of sedimentetion." Nelson does report evidence for four "abruptly" buried marsh deposits in South Arm Slough of Coos Bay (Fig. 21b),

but the cored sites are near the axis of a syne v.e young enough to tilt Larine terraces on its western flank. He concludes that incal faulting or faulting cay be responsible for the buried marsh surfaces rather than regional coseismic subsidence accompanying a great Cac cadia zone earthquake.

Discussion The recent discoveries by Atwater and others of buried peaty horizons in the estuaries of coastal Oregon and Washington are scientifically very noteworthy if they constitute evidence for episodes of rapid and regional subsidence of the coastline on t'he order r f 0 5 to 1 m. If coseismic subsidence is called upon to explain the burial of the peaty horizons, four critical issues must be addressed: (1) the strength of evidence for rapid subsidence; (2) the con-sistency in age of the buried peat deposits from estuary to estuary across many tens of kilometers (and, hence, the synchroneity of subsidence events);

(3) corroborating evidence for strong shaking effects accompsnying local,

i.e. Cascadia zone, earthquakes; and (4) an apparent absence of subsidence of wave-cut platforms lying on the coastline (West and McCrumb,1988; unpub-lished).between estuaries. With respect to the latter issue, little evidence has been found in coastal sediments for the soft-sediment deformation that one would expect to be ubiquitous if the coastline has experienced multiple large, great, or giant earthquakes during the Holocene. Atwater and others (1987) report vented sand deposits on the uppermost buried lowland near Grays Harbor (age ca. 300 years) and attribute the sand to strong ground shaking, but the absence of similar occurrences elsewhere is striking.

The subsidence of tidal marshlands along the coast of the Pacific Northwest attributed to coseismic subsidence during Cascadia zone earthquakes has an alternative explanation. Rapid eustatic fluctuations in sea level due to cyclic fluctuations in climate have been proposed to account for tegular (and episodic) depositional events with both long periods, on the order of-ten thousand to one hundred thousand years (Goodwin and Anderson, 1985; Grotzinger, 1986; Chan and Langford, 1987), and short periods of a few hundred to a few thousand years (Fairbridge, 1976; Fairbridge and Hillaire-Marcel,1977; Ra=pino and Sanders,1981; Shennan,1986 a, b) . The cyclic depositional events are characterized by sharply-defined, non-depositional  ;

surfaces that are abruptly overlain by sediments of deeper, marine facies; deposits of each cycle are about 0 5 to 5 m thick. )

l Ra=pino and Sanders (1981) and Shennan (198f a, b) have suggested that estua -

rine or lagoonal marsh grasses may grow and die out along coastlines in response to repeated Holocene eustatic sea-level fluctuations. Rampino and Sanders (1981) have noted that the growth of marshes seems to be favored j during periods of slow submergence, and that marshes are killed when relative j submergence exceeds 1.8 mm/yr. Using the Holocene sea level curve of Fair-bridge and Hillaire-Marcel (1977), they were able to correlate sea level fluctuations of 3-5 m with six episodes of peat development, in the north-eastern United States, during the period 4.7 to 8.6 Ka. Oldale and Dillon (1981), however, have challenged this hypothesis on the basis of what they claim to be incomplete data on the presence of peat layers, and on their

_ . _ . . _ .._ _ . = _ - _ . _ _ . ._-__ _ _ _ _ _ . _ _ _ _ _ . _ _ _

belief that the Holocene eustatic rise lacked the oscillations needed to produce episodic peat development.

In contrast, Fairbridge and Hillaire-Marcel (1977) believe that geologically sudden wordwide drops or rises of sea level on the order of 2-5 m may have occurred within periods of 100-200 years at numerous times during the Holo-cene. They attribute such sea level changes to episodes of climatic cooling or warming and subsequent glacioeustatic effects; as one example, it is stated that the "great mid-Holocene glacial readvance" led to a 6 m drop in worldwide sea level (ibid., p. 416) ca 5053 BP (sidereal) years ago. Indeed, work by Shennan in Great Britain (1986 a, b) attributes cyclic changes in vegetation, lithology, and microfossil assemblages of Holocene coastal stratigraphy to at least 7 periods of positive sea-level tendencies and 6 periods of negative sea-level tendencies during the past 6.5 Ka. Recent studies by Golder Associates (1988) for the Supply System suggest that the eight buried marsh surfaces reported by Atwater fl987) and Atwater and others (1987) for southwestern Washington could possibly be the result of cyclic fluctuations in the overall late Holocene eustatic sea level rise.

This is based on observed coincidences in the interpreted ages of the marsh surfaces with the ages of late Holocene eustatic rise events of the Fairbridge and H111 aire-Marcel (1977) sea level curve.

1 l

Atwater (1987) states his belief that only rapid tectonic subsidence > 0 5 m '

can lead to the burial of the peaty layers beneath intratidal muds and silts. He concludes that burial of marsh vegetation under flood- or storm-related sediment deposits should lead to emergence of the coastal lowland, not to submergence. Subsidence related to shaking-induced settlement from large distant or close smaller earthquakes might occur in the sediment-filled estuaries, but Atwater (op. cit.) argues that subsidence is observed even where peaty l'ayers have lapped onto "stiff" Pleistocene deposits along valley sides. Darienzo and Peterson (1987) report that X-radiographs of sand layers above peaty deposits in Netarts Bay, Oregon, show sharp, irregu-lar peat / sand contacts that suggest rapid subsidence and burial processes.

However, the non-tectonic development of "puncteated aggradational cycles" caused by fluctuating sea level and rapid eustatic rise also can result

(Goodwin and Anderson, 1985) in sharp contacte between the nearly subaerial, i non-depositional surface (peaty layer) and the overlying deeper sediment facies (intratidal muds, silts, sands) .

Sand layers above some buried peaty deposits have been explained as the consequence of Cascadia zone-induced tsunamis, but sandy layers occur within intratidal sediments as well as atop some buried layers of marsh vegetation.

Studies at Willapa Bay along the Niawiakum River (Atwater,1987; Reinhart and Bourgeois, 1987) indicata that the youngest sandy silt layer above peat thins upstream for a distance of 3 km and that it extends across both a former wetland surface (the marsh) and, upstream, a former surface of forest growth. Catastrophic storms or tsunamis from distant seismic events could also transport sandy sediments inland from the coast across marshes and neighboring forests, and one would expect to find, as is the case, some such deposits interlayered with intratidal muds.

The timing of Holocene estuarine subsidences is imperfectly known at present, and the synchroneous death and burial of peaty layers in different estuaries over large distances has not yet been adequately demonstrated. Hull (1987,

p. 1469) claims that the youngest buried soils along the South Fork Willapa River (mean ages: 131, 1747, 2460, and 2860 radiocarbon years BP) "are less than 100 yrs. from mean ages of buried wetland surfaces 15 km farther northwest" (one soil horizon between 131 and 1747 years BP is undated)ts, Atwater, Hull, and Bevis (1987) report that the five youngest periods of rapid subsidence in the same Willapa Bay-Grays Harbor area at 300, 1600(?),

1700, 2700, and 3100 sidereal years BP (age differences with Hull [1987) appear to reflect, in part, differences between radiocarbon and sidereal time scales; cf. Fairbridge and Hillaire-Marcel, 1977, Fig. 2). However, their statement that radiocarbon ages from presumably correlative layers differ from the "means for individual episodes of subsidence" by as much as 250 years should raise cautionary flags when correlating peaty layers from 110bviously, the 131 yBP horizon cannot be related to coseismic subsidence during a great Cascadia zone earthquake; the historic record in the Pacific Northwest precludes such a recent event.

l l

I i

j

__ _ _ _ . . - - - - - - - - - - - - - - ~

i

I one estuary to another. Similarly, the large radiocarbon age variance of one layer collected in five adjacent cores from a Netarts Bay salt carsh, 750 to 1240 + 80 years BP (Darienzo and Peterson, 1987), indicates the potential difficulties in relying upon different ages in peaty layers as evidence for different times of subsidence and burial of marsh systems. l l

l The discovery of buried peat deposits and lowland soils alternating with  ;

intratidal mud and silt in Washington and Oregon estuaries is a major devel-opment in ongoing studies to evaluate the seismic risk of ongoing subduction within the Cascadia zone. Although the prevailing theory that sudden subsi-dence of the pests and related soils occurred at times of great coastal earthquakes in the Pacific ? 1rthwest is promising, it needs additional veri-fication. More data are needed on a variety of topics, including- (1) the time and amount of Holocene eustatic sea level changes; (2) the rate of estuarine subsidence events; (3) the synchroneity, or lack of it, between )

peaty layers in adjacent estuaries; (4) physical evidence for strong ground i

shaking during hypothesized great earthquakes; and (5) the apparent non- i subsidence of contemporary wave-cut benches alonT rocky headlands and coastal stretches between estuaries (see below).

e Evidence for Holocene Coastal Uplift Marine terraces of late Pleistocene age along the Oregon and Washington I coastlines clearly record the Quaternary uplift of the coastline adjacent to I the Cascadia zone (Adams, 1984; West and others, 1987; West and McCrumb, 1988). Erosional remnants of five major late Pleistocene high sea-level- t l

stand terraces ranging in age from ca 42 to 220 Ka can be traced along the I coast between La Push, Washington, and Cape Blanco, Oregon (West and others, op. cit.). The most continuous of these terraces is the next-to-youngest, j the Whiskey Run terrace with an age of approximately 82 Ka. That terrace generally ranges from 5 to 50 m above sea level (except at Cape Blanco where it has an elevation of 121 m) and has a characteristic average elevation of 20 m along 600 km of Cascadia coastline. Late Quaternary uplift rates for this terrace (relative to the eustatically rising sea level and averaged over the past 82,000 years) range from ca 0.2 mm/yr to 0.6 cm/yr, but are

_ . .-~ _- - . . _ . _ . __ ____ _ . - . _._

i l

relatively uniform at ca O.4 mm/yr (Fig. 22) regardless of varying coastline distance from the Cascadia trench (60 to 140 km; West and McCrumb, in i press).12 The Whiskey Run terrace is well-developed in southwestern  !

Washington in the area showing Holocene estuarine subsidence; its 13 to 15 m elevation at Bay Center along the eastern margin of Willapa Bay requires an average late Quaternary uplif t of 0 32 mm/yr (West and others,1987) . The l terrace at Bay Center, and a higher, older terrace (103 Ka?), are developed l in Late Pleistocene est eine deposits (Golder Associates, 1986, p. 31-32, Figs. 22-24).

I Given the documentable history for the uplift of Late Pleistocene (220 to 42 Ka) terraces along 600 km of Cascadia coastline, one would expect abundant l geomorphic evidence for uplif ted Holocene wave-cut platforms, terraces, or beach berms as well -- either as short-term coseismically uplifted counterparts to areas of inferred coseisic subsidence (see above), or as expressions of overall, long-term coastal uplif t and 2andward tilting above the Cascadia subduction zone. However, West and McCrumb (1988; West and others, 1987; Golder Associates, 1986) after a survey of the coastline from north of Cape Blanco to La Push could find no evidence for uplifted Holocene platforms, terraces, or beach berms:

There are no known uplifted Holocene shoreline features that would indicate repeated great earthquakes on the Cascadia subduction l zone interface. This is especially significant for the Oregon  !

coast which occurs well within the zone of expected coseismic uplif t and zone of multiple uplifted Holocene terraces common in most subduction zones. The present coastline configut lon - -

locally broad, modern wave-cut plaftforms situated ditvs'.ly below i I

the uplif ted Pleistocene terraces without intervening uplif ted Holocene terraces -- suggests relative tectonic stability during the late Holocene (last 2-4 Ka). (West and McCrumb,1988) 12Uplif t rates are anomalously high at Cape Blanco, which lies near the southern end of the Cascadia subduction zone as defined herein. Uplift rates at the Cape for the Whiskey Run terrace and the younger Cape Blanco terrace (42 Ka) are 1.6 and 2.4 mm/yr respectively [the 42 Ks *errance is geographi-cally restricted to the Cape area]. Cape Blanco lies at the smallest trench-coastline distance for the Cascadia zone, 60 km, and is just northeast of the Gorda plate with its complicated patterns of internal compressive deforma-tion (Wilson, 1986; stoddard, 1987).

Discussion The absence of uplifted Holocene shoreline features does not preclude' Cascadia zone coastal uplif t during the Holocene, but may indicate that uplift has cccurred at rates approximating the continuing eustatic rise of sea-level. It might be argued that the absence of uplifted shoreline fea-tures supports a case that the Cascadia coastline has been regionally affec-ted by Holocene subsidence (notwithstanding the data from late Pleistecene terraces documenting regional uplift). If so, an interesting geomorphologi-cal dilemna is raised by West and McCrumb (1988) that is especially appli-cable to the Washington coastline north of Willapa Bay and the entire Oregon coastline north of Cape Blanco (Fig. 21 a, b). If estuaries along the Washington coast (from north to south: Waatch River, Copalis River, Grays Harbor, Willipa Bay, Columbia River; Atwater, 1987) and along the Oregon coast (Netarts Bay, Nehalem Bay, Salmon River) preserve evidence for several meters of Holocene subsidence (up to 5 5 m in the past 4300 years at Willipa Bay; Hull, 1987), why is there no evidence for comparable subsidence of the modern wave-cut platform along the rocky headlands that lie between the estuaries?

Additional questions can be raised because of the absence of uplifted Holo-cene benches or terraces along the Washington and Oregon coasts. Coastlines above subduction zones that have spawned great to giant earthquakes are commonly characterized by uplif ted Holocene terraces, among them Alas!:a (Plafker,1972), Southern Chile (Kaizuka and others,1973), Nankai, Makran, Sagami, the New Hebrides, and northeastern New Zealand (Fig. 23). Such' Holocene terraces are not restricted to areas of coseismic uplift, but do tend to be highest in those areas (West and McCrumb, 1988). Why then are they absent from the coastline above the Cascadia zone?

If northern coastal parts of the zone have been subject to Holocene coseismic subsidence (Washington, northern Oregon), more southerly coastal areas in Oregon should presumably have experienced synchronous coseismic uplift if

___ _ _. _ . _ _-_ _ _._ _ _ . ,.,_,,.__,._.____,__..___f-

- - . . . - . - - _ - - . _ . .- - . -. .. -. . . - _ ~ . . .

l great earthqur.ke rupture (s) extended that far south.2 3 The southern third of the 600 km-long coastline between Cape Blanco and La Push, south of Otter l Rock (Fig. 21b), lies within 110 km of the offshore trench - the approximate minimum distance between the trench and the seaward boundary of coseismically subsided areas elsewhere (Fig. 20e; Atwater, 1987, Fig. 2; West and others, 1987). West and McCrumb ( op. cit.) speculate that the absence of uplif ted Holocene coastal features, especially along the Oregon coast between Cape Balanco and Yaquina Bay, argues for atypical behavior of the Cascadia subduc-tion zone. Among possible explanations for this absence are (1) aseismic behavior of the subduction zone, (2) unusually long recurrence intervals for great earthquakes along it (thousands of years?), (3) interplate earthquake activity with magnitudes too small to produce coseismic elevational changes

(< M 7 ?), and (4) unknown tectonic mechanisms during great earthquakes along this subduction zone that preclude uplif t in zones of expected coseismic uplift.

And finally, why should the rate of long term late Quaternary uplif t for the Cascadia zone (ca. 0.4 1 0.2 mm/yr over the past 82 Ka) be so low when compared with uplift rates for other zones along which historic great to ,

giant earthquakes have occurred? A compilation of uplif t data for such zones reveals that most have rates consistently > 0 5 mm/yr, and many have uplift rates > 1.0 mm/yr (Fig. 24). The answer does not appear to lie in faster convergence rates for zones with higher uplif t rates The seismogenic SW Japan (Nankai trough) subduction zone has a probable rate of convergence l

13Perhaps they did not. One speculative geologic / seismic scenario is that coseismic elevational changes related to major earthquakes on the Cascadia zone are confined to the northern Oregon and Washington coasts. Coseismical-ly uplifted areas would lie offshore and not be observed. The conclusions of Nelson (1987) that the south-central Oregon coast may have experienced only gradual sea-level rise during the Holocene raises the possibility that the seismic history of this region differs from that of coastal regions to the north. A gradual rise in sea level could explain not only the 4 m-thick peat deposit of one estuary (ibid.), but also the absence of uplif ted Holo-cene terraces and beach berms if tectonic uplift rates were equal to or slightly less than sea-level rise. These speculations are developed further in the se,ction below on possible segmentation of the Cascadia zone.

l quite similar to that of the Cascadia zone (ca 3-4 cm/yr), but its uplifted coastal .erraces indicate a relatively uniform rate for uplif t of 2.0 mm/yr throughout the late Pleistocene and Holocene (Thatcher, 1984). Once again, l the Cascadia zone stands out as being to malous.

l At the time of writing of this response to NRC Questions 230.la, t, and c, the data from studies of coastal geomorphology and neotectonics appear to us to be non-definitive in either establishing er refuting the occurrence during the Holocene of thrust-type, large to giant earthquakes along the JdF/NA plate interface. Although buried estuarine marsh deposits, particularly those of southwestern Washington and northwestern Oregon, are now being wide-ly interpreted as evidence for coseismic coastal subsidence during repeated major Cascadia zone earthquakes, several unanswered problems co=plicate this interpretation; such problems include: (1) a lack of compelling evidence for rapid submergence of the deposits; (2) a lack of evidence for regional syn-chroneity of subsidence along lengths of coastline (> several hundreds of km) co:patible with those observed to have subsided coseismically elsewhere (e.g. Chile, Alaska); (3) a lack of evidence for the strong ground shaking effects that coastal Oregon and Washington would have experienced during

=ajor Cascadia zone earthquakes; (4) the absence of uplifted Holocene shore-line features in Oregon, the central coastline of which largely lies at coastline-trench distances (< 110 km) comparable to those for worldwide seis-mogenic subd.ction zones that have experienced coseismic uplift (Fig. 20d);

and (5) the occurrence at present mean sea level in northwestern Oregon and southwestern Washington of well-developed modern wave-cut benches along rocky headlands; these headlands, which appear to have elevational stability with respect to Holocene sea level, separate estuaries believed by some workers to have experienced repeated "regional" coseismic subsidence. l 1

l Submerged estuarin'e marsh deposits of Holocene age are found elsewhere along passive continental margins in settings where coseismic subsidence is out of the question, e.g. eastern North America (Rampino and Sanders, 1981) and the east coast of England (Wales: Shennan, 1986 a, b). The submerged peaty deposits of these coastlines are generally attributed to periodic Holocene climatic fluctuations and the glacio-eustatic consequences of such fluctua-

l climatic fluctuations and the glacio-eustatic consequences of such fluctua-tions. Such an crigin for the buried marsh depotits along the coast of the Pacific Northwest needs to be carefully evaluated before a tectonic explana- J tion for these deposits (coseismic subsidence) is uncritically accepted.

l Turbidite Deposits at the Base of the Continental Slope i

Turbidity current deposits-are known to be present on the deep-sea floor of the Cascadia Basin off the mouth of the Columbia River (Griggs and Kulm, 1970). The average time interval between turbidity flows that are believed to postdate deposition of the Mazama Ash, ca 6600 radiocarbon years ago, is  ;

400 to 500 years (ibid.) -- although the ages of specific flow deposits are .!

not known. Adams (1984) has proposed that such flows may have been triggered by periodic great thrust earthquakes along the Cascadia zone, although he admits that the slumping of shelf-edge sediments could also occur periodi-cally when sediment thicknesses reach critical limits and become Eravitation- J ally unstable. l l

There is reason to doubt that subduction zone earthquakes are the principal triggering effect for periodic turbidity flows along the Cascadia margin. l while not denying that large earthquakes in the Pacific Northwest (e.g., tne 1949 South Puget Sound event) could by their strong shaking of continental shelf areas initiate slumps and subsequent turbidity flows. The unusually thick Pleistocene and Holocene sedimentary deposits of the eastern Cascadia Basin have arrived there largely by turbidity current flow. Huge quantities of sediment arrived at the continental margin via the Fraser and Columbia Rivers during the Pleistocene, especially during the last glacial period which ended only about 10,500 years ago (Barnard, 1978). Barnard (1978) estimates Pleistocene rates of sedimentation within lower slope basins of 0.5 to 1 m/1000 years, but he reports that two-thirds of the total sediment derived from upper slope canyons during the Pleistocene were routed by turbidity currents into the deep-sea fans of the Cascadia Basin. Scholl and Marlow (1974) estimate that the total volume of basinal turbidites deposited within f Tt the past 1.0 m.y. is approximately 140,000 km3! Given the initial deposition of such huge quantities of sediment so quickly upon the

continental shelf, no catastrophic reasons for their subsequent slumping and turbidity flow down canyons into the Cascadia Basin are required. Turbidity currents are generated by slumping of sediment fill at the heads of submarine canyons along continental margins the world over. The phenonemon is not restricted to seismogenic subduction zone margins, and is the most likely explanation for the turbidity flow deposits of the eastern Cascadia Basin.

Holocene Faulting Near the Mendocino Triple Junction, Northern California i

! Carver and Burke (1987 a, b) have described Holocene thrust faulting and l

folding in bay sediments and fluvial terraces on the northern California coast between Cape Mendocino and Big Lagoon (Fig. 2). Trenching across the western of two parallel strands of the Little Salmon fault, on the south side of Humboldt Bay revcals multiple Holocene fault displacements. A layer of charcoal with a radiocarbon age of 6200 years is displaced 33 m across the western fault strand. Three separate slip events, each of 4 to 5 m, are interpreted as having occurred along this fault trace within the past 1930 14 C years. Thus, the recurrence interval of large slip events on the Little Salmon fault zone may be about 600 years; the last slip event may have occurred about 300 years ago. The eastern strand of the fault zone has not been trenched, but has a morphology similar to the western strand; thus combined slip, if synchronous on the two strands, could significantly exceed the 4-5 m of western strand displacement and would imply seismic events of great magnitude. Carver and Burke (1987b) conclude:

"Such large coseismic displacements suggest great earthquakes are associated with Cascadia subduction, at least along the southern portion" (of the subduction zone). "Much of the deformation is accommodated by faulting in the accretionary margin landward from the plate boundary, similar to the coseismic deformation in Alaska in 1964."

The discovery of major, repeated Holocene faulting along the Little Salmon fault zone is of considerable importance to our understanding of present Gorda/ Pacific / North American plate interactions, but we do not regard this faulting as indicative of the mode of interaction between the Juan de Fuca and North American plates. The Cascadia zone, as we and others (e.g.

Riddihough, 1984; Rogers, in press) have considered it, does not include the boundary between the internally deforming Gorda microplate and North America.

The consensus among recent writers (Riddihough, 1980, 1984; Wilson, 1986; Spence,1987; Stoddard,1987) is that the hot, young Gorda plate is no longer being subducted beneath North America. Instead the plate is being internally deformed as it converges obliquely (southeastward) toward the offshore Mendocino transform zone boundary with the Pacific plate (Fig. 2).

Wilson (op. cit.) believes that no appreciable convergent deformation is occurring along the strike-slip Mendocino fracture zone, whereas Stoddard (op. cit.) believes that complex interplate deformation occurs there.

Specifically, he sugggests that upper brittle and lower ductile layers of the Gorda plate become decoupled along the fracture zone. The higher level (ca 2 km thick) is obducted onto and accreted against the high-standing Mendocino Ridge, whereas the lower level flows eastward parallel to and relative to the ridge.

Thus, the tectonic setting of the triple plate junction (Gorda, North America, Pacific) at Cape Mendocino is extremely complex and one, based on earthquake focal mechanisms (Fig. 2), of considerable Holocene north-south compressive deformation -- an orientation incompatible with the oblique, N 50* E direction of convergence between the North American and Juan de Fuca plates. The unusual state of tectonic activity in the triple junction region is clearly evidenced by dispersed seismicity within the Gorda plate (Fig. 2) and by studies of uplifted Holocene terraces near Cape Mendocino.

There, Lajoie and others (1983) document Holocene uplif t of multiple terraces and beach ridges of as much as 17 m and at a maximum uplif t rate of 3.6 mm/yr. It is not, therefore, surprising that significant onshore Holocene thrust faulting, accompanying folding, and rapid uplif t has now been docu-mented by Carver and Burke (1987 a, b) in this active triple junction region.

There is, however, no valid basis to extrapolate their conclusions about Holocene great earthquake activity in this complex triple junction region to the tectonically simpler subduction zone between the Juan de Fuca and North American plates; to do so would appear to us to be scientifically unjustifiable.

Indian Legends Regarding Pacific Northwest Seismic Phenomena e Great Floods (= Tsunamis?)

'l Some legends of Indian tribes in the Pacific Northwest describe major floods that affected Indian settlements. Heaton and Snavely (1985) report some of 1 these legends and suggest that they may be indicative of the effects of a large tsunani or tsunamis along the northwestern Washington Coast. One legend, attributed to Indians living near Cape Flattery, states that prior

, to a great flood that submerged the entire Cape and the surrounding country.

"the water suddenly recedad leaving Neeah Bay perfectly dry. It was four days reaching its lowest ebb, and then rose again without any waves or breakers. ... The water was four days regaining its accustomed level" (Swan, 1868). It is conceivable that this account records in a distorted fashion the occurrence of a tsunami at Neah Bay, but there is no certainty that such a tsunami was generated by a Cascadia zone great earthquake. (No occurrence of strong ground motion just before the "tsunami" was noted in J this legend.) Exotic tsunamis, those generated by distant Pacific Rim earthquakes, are often preceded by sea-level subsidence and withdrawal of l

the sea prior to the arrival of the first wave (and, in some instances, l prior to arrival of later waves). The 1964 Alaskan earthquake, for example, produced a minor tsunami at Neah Bay with a height of about 1 3 m (Cloud and Scott, 1972).

1 Peoples throughout much of the world, Christians included, have legends of past great floods. In the Pacific Northwest, such tales are not restricted to Indians living in coastal areas, but include accounts of great floods near Mt. Shasta, Jefferson, Baker, and Rainier, as well as in areas east of the Cascades (Clark, 1953). One account by Yakima area Indians indicates the similarity of flood legends from tribes living on both sides of the Cascade Range (Clark, 1953, p.45):

1 One of the good men told the others, "I have heard from the Land l Above, the land of the spirits, that a big water is coming a big water that will cover all the land. Make a boat for the good people. Let the bad people be killed by the water." ... Soon the flood came. It filled the valleys. It covered the hills and the

mountains. The bad people were drowned by the big water. The good ones were saved in the boat. We do not know how long the flood stayed. At last the canoe came down where it was built.

You can see it on the Toppenish Ridge, on the side toward the rising sun.

o Great Earthquakes Given the ubiquity throughout the Pacific Northwest of Indian legends regarding great floode and the obvious distortions in them of natural pheno-mena that may have formed the basis for those legends, suggestions that prehistoric subduction zone earthquakes may have been responsible for some Indian legends (Heaton and Snavely, 1985; Heaton and Hartzell, 1986) should receive little credence. What may be more important than numerous legends describing great floods is the apparent paucity of local legends describing effects (i.e., strong ground shaking) stemming from great (M > 8) or giant (M > 9) subduction zone earthquakes on the Cascadia zone. Surely such catastrophic events, had they occurred, would have made an impression on the impressionable natives. Clark (1955), for example, quotes a George Gibbs as I writing in 1865 that the Chinook Indians near Willapa Bay "have traditions of earthquakes that have shaken their houses and raised the ground." This isolated account could stem from either local (Cascadia) or distant (Puget Sound) earthquakes; the reference to ground uplift (not subsidence) is inter-esting. Given the propensity of Indian legends to describe unusual natural phenomena -- for example, the northern lights (Clark, 1953, p.160-1610), the warm Chinook winds of the Pacific Northwest (ibid., p. 169-171), the Spokane (Missoula) floods? (ibid., p. 112, 172-175), and the eruptions of Cascade volcanoes, including that of Mt. Mazama 6600 years BP (ibid. , p. 20-24, 53-55) -- the paucity of legends that could relate to great Cascadia zone _

l earthquakes (e.g. , strong ground shaking, sudden region subsidence), may be i significant. l Ella Clark, a leading authority on Indian legends of the Pacific Northwest comments (1953) on the role of such legends in Indian life. The first role was as "unwritten texts in history, geography, nature study, and ethics" (ibid., p. 129). ...

A second purpose in Indian storytelling, illustrated again and ageta in this book, was to explain the phenomena of nature. This has aeen true, of course, of many early peoples. "Where we propound a scientific. theorem," wrote John Fiske years ago in Myths and ifyth-Makers, "they construct a myth." Fiske defined a myth "cs, in its origin, an explanation, by the uncivilized mind, of some natural phenomenon; not an allegory, not an esoteric symbol,

... but an explanation." ...

Hundreds of explanations of r.atural phenomena are scatttered i through the tales of the Pacific Northwest Indians. ... The great l rocks and the many trees in the Columbia Gorge, the cut made by the big river through the Cascade Range, the lake in the deep crater on top of a mountain, the eruptions of volcanic peaks, the petrified trees and the bones of prehistoric animals ... these stimulated the imagination to answer the natural question "Why?"

One of the amazing and fascinating things about several of these explanatory myths of the Pacific Northwest is that in a fanciful wcy some details parallel modern discoveries and theories of scientists. The parallelism between Indian myths and geologists' theory about lakes east of the Cascade Range in what is now the Columbia River Basin has been pointed out ... . Several details in "The Origin of Crater Lake," a myth related in 1865 by an old Klamath chief, have striking parallels with the story that geolo-gists have unfolded concerning an ancient peak in southern Oregon.

They call it Mount Mazama . . . . (ibid., p. 130)

Why, if Northwest Indians have passed down for millenia legends regarding the eruption of Mt. Mazama 6600 years ago and stories possibly describing the even older Spokane floods, are there not stories that tell of great earthquakes and sudden regional coseismic subsidence in southwescern Washington, the area where Atwater and others (1987) believe such phenomena have occurred five times in the past 3100 years, or eight times in the past 4300 years (Hull,1987)? Just as Indian legends do not prove Cascadia zone-generated tsunamis, the absence of Indian legends about great earthquakes does not prove that such earthquakes were not generated along the Cascadia zone. But, the ab'sence is curious.

Conclusions Regarding Capability of Subduction Zone Sources The Cascadia subduction zone between the converging Juan de Fuca and North American plates is a unique zone, both seismically and geologically, when

compared with other subduction zones the world over. The JdF/NA plate interface is remarkably aseismic and this seismic quiescence extends along the entire 900 km length of the Cascadia zone. A Benioff zone has been defined only from earthquakes within the subducted plate north of the  ;

Columbia River, and then only to depths of ca 80-90 km. The down-dip, seismically defined subducted slab length, ca 280 km, is among the shortest l

observed worldwide. Geologic factors that set the Cascadia zone apart from others include the extreme youthfulness of the subducting plate, a relatively l

slow convergence rate, and the presence at the offshore convergent boundary of the Pacific Rim's most voluminous accumulation of Quaternary trench sediment.

It has been proposed by Heaton and Kanamori (1984) and Heaton and Hartzell (1986, 1987), following Ruff and Kanamori (1980), that the combination of slow convergence rate and youthful age of subducting crust make the Cascadia zone a likely candidate for a future earthquake of great (2 8) to giant (2 9) magnitude. They favor the proposition that the seismic quiescence of the plate interface is an expression of a seismic gap - an indication that the plates are strongly coupled until future stick-slip release of the coupled boundary occurs in the form of one or more major earthquakes.

We have attempted to assess this proposition by considering all available geological and geophysical data that bear on the seismogenic potential of the Cascadia zone. As discussed at some length in this report, comparisons of the Cascadia zone with seismogenic subduction zones where young oceanic ,

lithosphere is being subducted suggest to us that the Cascadia zone is truly atypical, and that the slow subduction of the hot, sediment-laden Juan de Fuca plate may indeed be occurring aseismically. There are valid scientific reasons to propose that the subduction of anomalously hot and abundant sedi-ments with clay-rich mineralogy and high internal fluid pressures may lead to stable sliding rather then stick-slip behavior along the plate interface.

)

However, while we are skeptical that the Cascadia zone is seismogenic in terms of major interplate thrust earthquakes, we cannot discount that possi-  !

bility given geodetic evidence for low value of NE-SW-directed compressive l

__ . . _ _ _ . ._ _ - _ - - _ _ __ _l

i l

l strain in the North American plate and geologic evidence for possible coseismic deformation along the Washington / Oregon coast. I l

I Our dilemna reflects several interesting and contradictory trends that appear to us to be occurring in earth science research today regarding the Cascadia zone : 1 l

(1) Some seismologists (e.g. , Heaton and Kanamori,1984; Heaton and Hartzell, 1986, 1987; Rogers, in press), drawing upon broad simi-larities in subduction parameters between the Cascadia zone and subduction zones elsewhere that have experienced major historic thrust-type earthquakes, interpret the present seismic quiescence of the JdF/NA plate boundary as a seismic gap; future ruptures of all or parts cf the zone are thought to be likely; (2) In contrast, some geologists and rock mechanicists have begun to question whether t and v parameters of plate convergence adequately define seismic hazard and magnitude of potential earthquakes along the Cascadia zone (given its special geological circumstances),

and appear to be collectively leaning toward stable-sliding along the JdF/NA plate interface as the probable mode of plate inter-action (e.g. , Pavlis and Bruhn,1983; Byrne and others,1987, in press; Sykes and others,1987; Wang and others,1987; Severinghaus and Atwater,1987; Sammis et al. ,1988) .

(3) contrarily, geologists studying Holocene sedimentation and geomor-phology alnng the Washington and Oregon coastlines appear to be collectively favoring the stick-slip behavior of the plate inter-face based on evidence for episodic, possibly rapid subsidence (coseismic't) of estuarine marshes (e.g. , Atwater,1987; Atwater and others,1987; Darienzo and Peterson,1987; Grant and McLaren.

1987). i l

Problems with the interplate behavior favored by the second research trend, i.e. stable sliding and aseismic subduction, come from several quarters --

among them (1) the seismic quiesence of some major subduction zones prior to the occurrence along them of great interplate earthquakes (although, it has not been established that any such zones were as seismically quiescent as  !

the Cascadia is today); (2) geodetic evidence for some degree of JdF/NA plate coupling; and (3) geologic evidence for multiple eposides of Holocene coastal subsidence in southwestern Washington and northwestern Oregon.

With respect to (2), it is important to stress, however, that stable sliding is frictional sliding and decoupling along the sliding plate interface surface is not implied by this deformational mechanism. There may be a tendency i

1 l

l l

1

among seismologists to equate coupled plate boundaries with stick-slip or seismic behavior, but stable sliding and/or viscous flow between plates can also be accompanied by some degree of coupling and stress transmission into the overriding plate.

Problems with the interplate behavior favored by the first and third research trends, i.e. stick-slip behavior and major thrust-type earthquakes, also exist -- among them, (1) absence of evidence along the central Oregon coast-line for Holocene coseismic uplift, (2) evidence for tectonically stable wave-cut platforms at present sea level between estuarine areas for which coseismic subsidence has been postulated, (3) the relatively uniform uplift of a late Pleistocene terrace along the entire margin (even in areas said to de=onstrate repeated Holocene subsidence), (4) uncertainties in dating the regional synchroneity of coastal subsidence by radiocarbon techniques, and (5) the paucity of legends among coastal Indian tribes telling of the effects (strong shaking, areal subsidence) of the multiple great Cascadia earthquakes that are inferred from estuarine studies.

1 At the present time, we consider the case for aseismic subduction versus seismic subduction along the Cascadia zone to be a scientific standoff.

While we favor the former, we cannot discount the latter. Therefore, for I purposes of evaluating the SSE of WNP-3, we shall consider the plate inter- l face to be capable. Note that this is not a substantial change from the f PSAR position. Technical arguments can be made that the plate interface is aseismic and not capable. However, recent geologic studies suggesting that the plate interface may be seismogenic have raised sufficient uncertainty that an evaluation of the SSE relative to this scenario is warranted. In this context, we provide an assessment of the ground motions associated with hypothesized plate interface earthquakes. l The subducted Juan de Fuca slab has been the source of several large earth-quakes in historical time (1946, 1949, 1965) and is associated with ongoing seismicity of small magnitude. Because of this association with seismicity, we consider the Juan de Fuca slab to be a capable seismic source (referred to here as the intra-slab source) .

III. MAXIMUM EARTHQUAKE MAGNITUDES The two subduction-related potential seismic sources considered here are_the l intra-slab source and the interface source. Standard practice in maximum magnitude assessment is to consider the likely dimensions of rupture, his-torical seismicity, analogies to similar seismic sources, and judgement. We have attempted to apply these approaches to the extent possible.

Intra-Slab Source As discussed previously.in this response (e.g., p. 12-14), the location and geometry of the subducted Juan de Fuca slab is determined from seismicity data, deep seismic reflection data, magnetotelluric data, and broadband teleseismic waveform analysis. These data show variations in the dip of the slab along strike, reflecting an arch in the slab beneath the Puget Sound region. The seismicity data also show a bend in the slab downdip (see' j Figure 6). Such increases in dip with depth along subducting slabs are a common occurrence (e.g., Jarrard, 1986) and is supported in the Cascadia zone both by studies of hypocenter distributions (Crosson and Owens, 1987:

Weaver and Baker,1988), and by the requirement that the slab attain magmatic generation depths ('100 - 125 km) beneach the Cascades (Dickinson, 1975; Gill, 1981). As discussed in the response to Q230 3c, d, the larger observed slab earthquakes (1946, 1949, and 1965) have all occurred in the vicinity of the downdip bend and are likely reflecting downdip tension within the slab.

Localization of these events near the bend is likely the result of stress j concentrations at the bend (e.g. , Weaver and Baker, 1988) and is comparable to intra-slab earthquake locations along other subduction zones (Spence, in review). It is not known if the downdip bend exists north or south of the -  !

Puget Sound arch, but it may not if the bend is related to the arch. The absence of the ben'd could explain the lack of large historical slab events north and south of the arch.

The largest intra-slab event in the historical record is the 1949 M 71 event. By convention, we would consider the maximum magnitude to be somewhat larger than that observed historically. Very.few intra-slab events around

the world have exceeded magnitude 7 5 (Spence, in review). This is probably largely because of the limited downdip width (thickness) of the seismogenic part of the subducted oceanic crust (recall that the intra-slab source would be high-angle extensional faults within the slab) . In the present case, the Juan de Fuca plate is younger than nearly all other subducted plates and, because of its youth, should have a relatively thin seismogenic layer. For these reasons, we conclude that a reasonable maximum magnitude for the intra-slab source is Mw 71.

l Plate Interface Source The historical quiescence of the interface precludes the use of historical seismicity data to estimate maximum magnitudes both,1.. terms of the maximum observed magnitude and in terms of the use of smaller thrust events to define the geometry and extent of the interface as a seismic source, l

In the absence of historical seismicity data, the estimated maximum magnitude  !

for the plate interface is estimated based on physical contraints. As discussed previously, although Heaton and Kenamori (1984) claim to be esti-mating "maximum magnitude" with their relationship between convergence rate, I slab age, and maximum observed magnitude, they are actually estimating the mo=ent rate magnitude (Mw'). We believe that constraints on the dimensions of rupture (e.g., rupture area) along the plate interface provide the best physical basis for estimating a maximum magnitude because rupture area is an important element of seismic moment. To estimate rupture area, we consider here both rupture length and downdip width.

Rupture Length and Segmentation of the Cascadia Subduction Zone It is observed globally that interface ruptures along subduction zones do not rupture the entire length of the subduction zone (e.g. , Spence, in review). This observation is not surprising because we also observe that shallow crustal faults do not rupture their entire lengths, even in the largest events along plate-boundary faults (e.g., the 1906, 1966, and 1857 earthquakes occurred along three separate segments of the San Andreas fault;

historical rupture along the North Anatolian fault has occurred in over ten events along the length of the fault zone).

Seg=ents along subduction zones are typically identified by the locations of historical earthquake ruptures (e.g. the seismic gap concept stems from an unruptured segment lying between two segments that have ruptured histori-cally), patterns of cmaller magnitude seismicity, and tectonic characteris-tics of the subducting slab and overriding plate. For example, Burbach and Frolich (1986) identify segments on the basis of characteristics including abrupt changes in the strike of the Benioff zone, abrupt changes in the dip of the Benioff zone, abrupt changes in the maximum depth of seismicity, gaps in seismicity, abrupt lateral end of the Benioff zone, or locations of apparent lateral strain caused by anomalous trench geometry. Habermann et al., (1986) identify "first-order" seguents along various subduction zones based on variations in the rate of interface seismicity and compare the ends of these segments with changes in trend of the arc, intersecting ridges, offsets in the volcanic arc, changes in age of volcanics, abrupt changes in trend sediment thickness, changes in dip of the Benioff zone, and changes in seafloor topography. Because the exact controls en the rupture of the plate interface are not known, empirical observation has shown that these tectonic and seismicity characteristics can provide indications of the location and extent of future ruptures. Accordinrly, we can use them to provide informa-tion on the likely locations of rupture of the interface along the Cascadia subduction zone, e Evidence for Segmentation of the Cascadia Subduction Zone Until relatively recently, it was widely assumed that oceanic lithosphere of the Juan de Fuca plate is being subducted as an unsegmented slab along its 900 km-long north-south-trending interface with North America (Gorda-North American and Explorer-North American plate interactions are not considered  !

here). However, several recent papers have proposed (or implied) that the Juan de Fuca plate is segmented beneath the Pacific Northwest and that such segmentation influences or controls upper plate volcanism and tectonics and the seismicity patterns in both plates. Because the plate interface has

- - - - - ----'l

l l been completely aseismic in the historical period, interplate seismicity -

l (either large historical rupture or small-magnitude earthquakes) is not available to define possible segmentation of the interface along strike.

However, several geologic and tectonic characteristics can be defined along the zone that can lead to confident assessments of segmentation. These characteristics are listed in Table 2 and are described below. I i

Based on the spatial distribution of the Cascade volcanoes and the petrology i of the volcanics, Dickinson (1970) indicated an anomalous bend in the trend' I of the volcanic chain at about 46*N and extending north to about 49*. Based on his studies, Dickinson suggested that an arch-like geometry of the sub- I ducted slab was present beneath Puget Sound. .

l Based on his analysis of gravity data, Riddihough (1979) noted that between 42*N and about 45 5'N, the width of the negative anomaly is relatively

{

narrow and coincides with the base of the continental slope. Between 45 5*N I and 47 5'N, the maximum negative anomoly decreases in amplitude and steps eastward beneath the continental slope structural trench. Riddihough (1979) suggests that these changes may be related to spatial variation in the i geometry of the subducted slab.

l i

Kulm (1983) and Kulm and Embley (1983) describe significant differences l between southern Washington and central Oregon in the bathymetry, morphology, I and structural evolution of the continental shelf and slope. The morphology of the lower continental shelf off southern and south-central Oregion is  ;

characterized by relatively steep escarpments. At about 44 5'N, it changes to prominent elongate north-northwest trending ridges and intervening basins, which extends the length of the lower continental slope off Washington.

1 Rogers (1983) and Keen and Hyndmann (1979) recognized that the severe 45' bend in the trend of the Cascadia subduction zone, combined with the plate convergence direction, leads to a space problem at the corner. Crosson (1983) concluded that the deep zone of seismicity beneath Puget Sound was occurring within the subduction slab and both he and Rogers ( J83) concluded that the abrupt cut-off of seismicity to the north of about 49'N and to the

TABLE 2. EVIDENCE FOR SEGMENTATION OF THE CASCADIA SUBDUCTION ZONE Location Fenture Description Reference 46* to 49' Cascade volcanics Benu in volcanic arc; change in Dickinson (1970) petrology 46.5* Continental slope Changes in deformation patterns across Barnard (1978) deformation northeast trending zone 42 5*. 45 5*. Gravity anomaly Change in amplitude and location of Riddihough (1979) and 47 5* negative anomaly, suggesting change in slab geometry 44.5* Seafloor morphology Change in morphologi of lower Kulm (1983); Kulm and continental shelf Embley (1983) 47* to 49' Send in subduction zone; Sharp 45* change in trend of zone must Rogers (1983).

localized deep lead to compression /deforma- Crosson (1983) i seismicity tion / phase changes in slab, which is reflected in intra-slab seismicity beneath Puget Sound and absence of seismicity to north and south 46' Upper plate tectonic Change from slightly extensional Rogers (1985) regime; Quaternary regi:ae to south to neutral to slightly volcanism compressive regime to north 44* and 45 5* Slab dip changes; P-wave delay patterns suggest Weaver and Michaelson Quaternary volcanism differences in slab dip coincident (1985); Michaelson with changes in volume of Quaternary and Weaver (1986) volcanics 46* to 49' Slab geometry: abrupt Arch in slab along strike: 10-20* dip Crosson and Owens dip changes; position of beneath Puget Sound. 15-20* to north (1987); weaver and volcanic front or 49* and to south of 46*; step in Baker (1988);'owens late Cenozoic volcanic front- et al. (1988) 48-5*. 47*. Seismicity slab dip. Variations in level of seismicity; Spence (in review) 45* . and 43' density change variations in p-wave velocity

south of about 47 N is real and is not due to an uneven seismograph station distribution. Rogers (1983) causally associated the Puget Sound seismicity with the bend-in-the-trend and postulated that the earthquakes were a response to shortening, which is being accomodated by internal deformation I of the slab or by phase changes within the slab (preferring the latter).

l i

Based on seismic refraction data and seismicity analysis, Taber and Smith (1985) concluded that the dip of the subducted slab beneath western Washington was about 11*. Weaver and Michaelson (1985) examined the distri-bution of seismicity and the volume of Quaternary volcanism in the northwest and concluded that the Juan de Fuca plate dips more steeply beneath south-western Washington than beneath northwestern Washington. The; postulate a sharp segmentation of the slab marked by, from south to north, a steeply dipping portion beneath Oregon and southern Washington, a possible "tear" fault, a shallower dipping slab segment from about 46*N to 49*, another possible "tear" fault, and a more steeply dipping segment to the north.

Michaelson and Weaver (1986) used teleseismic p-wave delay patterns to support this model of discontinuous segments of varying dip.

Weaver and Michaelson (op. cit.) attribute the greater volume of Quaternary volcanism and diminished seismicity above the southern segment to an exten-sional stress state in the upper plate; diminished volcanism and greater i intra-North American plate seismicity above the central (shallow-dipping) segment are said to be expressions of increased coupling between the two plates and horizontal compressive stress [ note: this explanation implies NE-SW compressive stress, not the N-S compressive stress observed in focal mechanism solutions from earthquakes in the Puget Sound area]. Rogers (1985) believes that marked differences in Quaternary volcanic history and volumes of extruded lava between Oregon volcanoes and those of Washington and British Columb'ia reflect changes in the trend of the offshore subduction zone at about 47' N latitude -- from to in Oregon and southern Washington to NW-SE in northern Washington and British Columbia. While not proposing that the volcanic chain is sharply segmented, he does believa (as do Weaver and I l

Michaelson, 1985) that the southern segment coincides with a slightly exten-sional upper-plate tectonic cegime and the central segment with a neutral or l slightly compressione.1 one. The two volcano /tector.ic regimes are said by Rogers to reflect different ratios between overriding plate velocity (NA) and roll-back velocity of the subducting plate (JdF), and it is possible that they reflect differences in the degree of coupling between the North American and Juan de Fuca plates.

l Taber and Lewis (1986), Crosson and Owens (1987), and Weaver and Baker (1988) have all published new data on the configuration of the subducted Juan de Fuca plate beneath western Washington and northwestern Oregon. The collective results of their studies indicate that the slab dips eastward at a low angle (9*, Taber and Lewis; 10-12*, Crosson and Owens) across western Washington, and has an archlike configuration beneath Puget Sound with the axis of the arch plunging gently (ca 10-12*) eastward (Figs. 4, 5). To the north and south of the arch, beneath Vancouver Island and western Oregon respectively, the subducting slab has more planar geometries and somewhat steeper dips (15-20*; Spence and others, 1985; COCORP, 1986, GSA San Antonio; Crosson and Owens, op. cit.). Crosson and Owens (op. cit.) note that most I of the shallow (NA) and deep (JdF) intraplate seismicity of the Puget Low-lands coincides areally with the location of the arch in the subducted plate. Reasons for localization of intraslab seismicity and upper-plate seismicity with respect to the arch are explored briefly by both Crosson and  ;

Owens and Weaver and Baker (1988), but remain speculative. The existence of the arch appears to be related to the northward change in trend of the offshore trench from N-S to NW-SE at about 47' N latitude, and the requisite bending of the JdF plate at depth to accomodate this change in trend of its subduction zone (cf. Dickinson, 1970; Rogers, 1983). The change in trend of the Cascadia trench is the same causative agent that Rogers (1985) believes is responsible for varying tectonic regimes in the North American plate from neutral or slightly compressive in Cascade areas of northern Washington and British Columbia t'o slightly extensional in Cascade areas to the south.

l The data of Crosson and Owens (1987) and Weaver and Baker (1988) appear to resolve previous speculation as to whether the southern termination of deep Puget Lowland seismicity (Fig. 4) might coincide with an abrupt, discrete "tear" in the subducted plate between a downward bending slab to the north l

i i

. -. .- -- -- - - J'

l l I

and a planar, more steeply-dipping slab to the south (Weaver and Michaelson, 1985). Although steeper dips are likely south of 47' N (Fig. 5), a flexcre l appears to exist in the dowagoing plate rathor than a discrete tear. The ocurrence of such a flexure is in good agreement with a recent circum-Pacific study by Burbach and Frohlich (1986) of the lateral structures in subducted lithosphere. They report that such lithosphere generally exhibits contor-tional (bending) behavior at segment boundaries and only rarely deforms by breaking or tearing, such as along a hinge. fault.

Weaver and Baker (1988) believe that a NE-SW line connecting Mt. Rainier and Portland, Oregon, and passing through Mt. St. Helens (cf. Fig. 25), defines a volcanic front for the westernmost occurrences of late Cenozoic volcanism in this latitudinal range. They propose that the trend of this line ca.n be used to define a similar trend for the strike of the southern flank of the arched Juan de Fuca plate. The essentially aseismic subducted slab to the south, beneath Oregon, is generally believed to have a steeper dip and to lack the downward bending that controls deep, Intra-slab seismicity beneath Puget Sound (Weaver- snd Michaelson,1985: Crosson and owens, 1987: Weaver and Baker,1988) . Weaver and Baker (1988, p. 272) state that "The plate geometry south of Portland, Oregon, cannot be inferred from earthquake data.

The volcanic front steps eastward to the axis of the Cascade Range, and con-tinues along the western edge of the range throughout Oregon and northern California. ... Thus, the change in the volcanic front position near Portland may indicate that the geometry of the Juan de Fuca plate to the south may also change."

e Conclusions Regarding Segmentation of the Cascadia Subduction Zone The physical division of the subducted Juan de Fuca plate into a central

~

arched slab and two neighboring planar slab elements Provides a basis for segmenting the Cascadia zone in terms of its seismotectonic behavior.

Burbach and Frolich (1986) have defined segments in other circum-Pacific subduction zones using features that are equally characteristic of the Cascadia zone. Among them (op. cit., p. 837): (1) abrupt change in strike of Benioff zone (which could equate here with the change in trend of the I

.1

Cascadia trench); (2) abrupt change in dip of Benioff zone (which would equate hera with the steepening of the tcp of the subducted JdF slab north and south of the Puget arch); (3) abrupt change in the maximum depth of seismicity; (4) gap in seismicity (deep, intermediate, or shallow); and (5) abrupt lateral end of Benioff zone. Some segment boundaries, they state (op. cit., p. 833), "occur where there is apparent lateral strain caused by anomalous trench geometry." About 68% of the segment boundaries that they have delineated "are associated with some noticeeble change in arc volcanism" (op. cit. , p. 851), although segment boundaries and "changes in volcanism seldom coincide exactly and are often separated by as much as 100 km" (ibid.)

[ note: this diffuseness would certainly be true of the Cascade arc].

The Puget arch provides mechanical explanation (s), albeit still uncertain, for the concentrations of seismicity in the Puget Lowlands in botn upper and lower plates (Crosson and Owens,1986; Rogers,1983) . However, the existence of the arch does not, in and of itself, cast new insights into either the extent of coupling of the plate interface acrose the arch beneath western Washington, or within more planar segments of the Cascadia zone to the north and south. In all three segments, the plate interface is remarkably seis-mically quiescent. What is perhaps most important is that a physical ba:is ,

for separating the Cascadia Lone into three geometric segments is now eatio-lished and the boundaries of these segments are coincident with other geologic / tectonic changes. Using the contour map of the subducted slab prepared by Crosson and Owens (1987; Fig. 5) and the kinematic analysis of Rogers (1995) as guf( =, these three segments might be geographically restricted and geologically characterized as tollows (Fig. 25; latitudes and lengths taken from approydmate positions along the trench axis):

1) Northern (British Columbia): ca 49 5'N, ?.o ca 47 5'N; length ca 275 km; slab dip, 15-20*; strain class (Jarrard, 1986) =

4a-57

2) Central (Washington): 47 5'N to ca. 45'N; lengd ca 250 km; slab dip, 11*; strain class = 4a-57

3) Southern (Oregon): 45'N ca. 42 7'N; length ca 330 km; slab dip,15-20'; strain class = 3

1 In summary, the various lines of evidence for the two segment boundaries separating the three segments are the following:

Northern Segment Boundary (at about 47 5'N at the trench) e Change in slab dip e Change in level of slab and shallow crustal seismicity e Bend in volcanic are Southern Segment Boundary (at about 45'N) i e Change in slab dip e Abrupt change in level of slab and shallow crustal seismicity e Bend or step in late Cenozoic volcanic front e Change in volume of Quaternary volcanics e Change in upper plate tectonic regime In terms of geometry, seismicity, and volcanic characteristics, the two I segment boundaries are essentially mirror images of each other. The southern boundary has been most extensively investigated, however, and defined by several characteristics that, upon further investigation, may also charac-terfze the northern boundary. In any event, the coincidence of a number of geometric / geologic characteristics at these two locations strongly supports the concept of segmentation along the zone at these locations. The WNP-3 site lies within the central segment (Figure 26). Based on the arguments for segmentation along the zone, we assess the maximum rupture length along the plate interface to be a rupture of the entire central segment, a length of about 250 km. Simple consideration of the magnitudes and distances asso-ciated with the adjacent segments show that the rupture of the northern or southern segments will not result in ground motions at the site as signifi-cant as those due to rupture of the central segment. l l

Rupture Width

)

Given the estimate of rupture length discussed above, we still require an estimate of the rupture width (W = downdip width of the seismogenic inter-face) to arrive at the rupture area associated with the maximum magnitude of I

1 I

i l

l an earthquake on the plate interface. Once again, the aseismic nature of l the interface precludes a direct estimate based on historical observation of rupture or patterns of smaller-magnitude thrust events. We must rely on other physical and tectonic argements.

It is well-established from observation and physical modeling that the seismogenic rupture of plate interfaces occurs updip at a position that is below unconsolidated materials of the accretionary wedge and downdip to depths where higher temperatures lead to non-brittle failure mechanisms.

Therefore, to assess the updip and downdip extent of seismogenic rupture on the Cascadia interface, we must consider the physical constraints on each.

The transition from aseismic deformation of the accretionary wedge to seismo-genic deformation along the interface is controlled by the quantity of sedi-ments being supplied to the accretionary wedge, the quantity of sediments that are actually carried down the subduction zone, and rheological con-straints such as pore pressure, temperature, and mineralogy (e.g., Byrne et al., in press). As discussed extensively in this response, the Cascadia subduction zone is being supplied with enormous quantities of sediments.

These sediments are water-saturated and are under low levels of effective stress because of high internal pore-fluid pressures. Because of thermal blanketing of the young, hot Juan de Fuca slab by these sediments, they are in turn hected by the slab. Further, the mineralogy of these sediments shows an abundance of saturated clays that exhibit stable sliding behavior at the expected temperature and pressure conditions of the accretionary wedge. All of these characteristics would suggest that a thick accretionary wedge should be present to significant depths along the Cascadia zone and that, in turn, the seismogenic part of the plate interface should lie below these sediments.

Several recent stu' dies ao in fact show that the accretionary wedge is quite extensive. For example, the results of LITH 0 PROBE transects through southern Vancouver Island, image a > 10 km thick zone of underplated sediments to depths of over 30 km (Clowes, et al., 1986, 1987), seismic refraction studies model a lower velocity (-4-5 km/sec) wedge of sediments that extend to depths of about 15 - 20 km (spence et al., 1985: Taber and Lewis, 1986).  !

l l

)

i l

l Magnetotelluric data across Vancouver Island (Kurtz et al. ,1987) and north-ern Oregon (EMCLAB Group, 1988) also support the model of a large sedimentary wedgu to depths of at least 20 km and further suggest that fluids related to dewatering sediments are present to depths of about 40 km.

Based on the above data that show the presence of a significant accumulation of sediments, indications that these sediments are water-saturated and heated, and the clear presence of these sediments as thick off3 craped units to depths of over 20 km and possible extension down the subduction zone to depths of 40 km (Fig.13: Clowes et al, 1986, their Figure 4), we here conclude that a seismogeni. , late interfaca can extend no shallower than about 20 km. This depth corresponds approximately to the position of the Washington coastline at the surface.

The downdip extent of the seismogenic interface is largely controlled by temperature (e.g. , Byrne et al. , in pross) . Two important aspects affecting the temperature at the interface are the temperature of the subducting slab and the thermal gradient of the continental (North American) plate. The Juan de Fuca slab is very young (ca. 8 my) at the interface and, as discussed previously, is thermally insulated by a thick blanket of sediments that cover the plate nearly to the ridge. The high temperature of the slab is further indicated by the extremely short downdip length of the Benioff zone (Jarrard, 1986) suggesting that the downgoing plate is close in temperature to the surrounding continental aesthenosphere and is able to quickly equilibrate to a higher-temperature, non-brittle state. It is not'ed by Crosson and Owens (1987) and Weaver and Baker (1988) that the slab appears to be seismogenic only in the region of the arch in the slab along strike. Crosson and Owens hypothesize that it is only this shallow-dipping part of the slab that l

exists at shallow enough depths to generate earthquakes. Perhaps to the north and south of' the arch "a weak or non-existent Benioff zone may result from relatively rapid heating of an already young, hot slab as it descends quickly past the brittle-ductile transition" (Crosson and Owens,1987) . The indications that the slab is relatively hot are further supported by the slow rate of convergence across the subduction zone. Slow rates of intrusion l I

- _ _ _ _ - _ _ _ _ . _ _ _ _ _ _ . _ _ . _ . - _ . _ _ _. , . _ _ _ _ _ _ _ _ _ . _ _ _ _ ,,,._a

l I

l of the slab into the mantle provide a better opportunity for slab heating and thermal equilibration. ,

The thermal ,,tructure of the North American plate can be inferred from ti e seismicity 6 utribution within the lithosphere. Detailed studies of crus al structure (Owens, et al. ,1988; zervas and crosson, 1986) indicate a depth to continental Moho of about 30 - 35 km at the latitude of the site. In generci, the maxi =um depths of crustal seismicity are less than 30 km (Figure 4b) with a few events to nearly 35 km. The maximum depth of seismicity marks the lower limit of '.he brittle-ductile transition within the continen-tal crust (Fig.10 and Bryne et al. , in press) . Using the indications of continental Moho at 30 - 35 km and conservatively assessing the brittle-ductile transition to be at 35 km, we will assume that 35 km depth marks the downdip extent of a seismogenic interface. Also note that Hartzell and Heaton (1988) postulate that rupture of the interface in the uppermost continental mantle may contribute to the seismic moment of an earthquake (through increased area) but this deeper rupture "is accompanied by little short-period radiation because of the non-brittle rheology or the cantle".

In conclusion, the physical and tectonic constraints on the maximum rupture width of fhe plate interface indicate that the updip extent should be no sha31ower than 20 km depth and no deeper than 35 km depth. To translate these depths to the implied width of the interface, a dip must be assumed.

Because the site is located on the arched segment of the slab and litt]e is known about the var;ation in dip along the arch, we conservatively assume the shallowest (11*) dip along the arch in calculating the width (che limbs of the arch would have a steeper dip). Accordi.ngly, the maximum rupture width along the plate interface is assessed to be about 75 km.

l Maximum Magnitude Estimate '

1 The previous discussion has resulted in an assessment of the maximum rupture length (250 km) and downdip width (75 km) that might be associated with a  !

l seismogenic rupture of the plate interface. The resulting maximum rupture area (18,750 km2 ) is translated into a maximum magnitude using e=pirical i

1 l

1 l

relations between observed rupture area Lnd moment magnitude. Pu'elished relationships include those by Abe (1975) and Kanamori (1977), which are developed specifically for interface subduction zone earthquakes. Their relationship can be written as M, = log A + 3 99 where A is area in km2 . In addition, an analysie of the subduction zone earthquake data listed by Wyss (1979) and some recent events such as the 1985 Mexico and 1985 chile earth-quakes resulted in a relationship identical to that given above and we, therefore, advocate its use for these purposes, Based on a maximum rupture area of 18,750 km , the maximum magnitude for the plate interface is assessed 2

to be 8.26 or about 81.

The uncertainty in this maximum magnitude estimate is a function of the uncertainty in the assessments of rupture length and rupture width. Q230.2a allows for a simple expression of the uncertainty in the maximum earthquake estimate. From an examination of the possible uncertainties in our assess-ments of rupture length and width, we conclude that the uncertainty in the magnitude estimate is about i magnitude units. We, therefore, arrive at a maximum magnitude estimate for the plate interface of M, 81.

We are aware that the maximum magnitude es. cate of 81 is less than some hypothesized earthquakes along the Cascadia zone. For example Heaton and Hartzell auggest that an earthquake of M, 9+ might be possible along the Cascadia subduction based on analogy to the 1960 chile earthquake. It should be noted, however, that these "giant" earthquakes are proposed based primarily on analogy to other zones (analogies shown previously in this response to be tenuous in detail) and for purposes of evaluating possible ground motions; they are not based on a rigorous consideration of the poten-tial dimensions of rupture associated with a maximum earthquake. Such an analysis shows that magnitudes of 9+ are simply not credible. For example, given a width of 75 km and the relationships of Abe and Kanamori dee tribed above an Mw 9 - 91 earthquake should have a rupture lenFth of about 'i400 -

4300 km. Even if the seismogenic width is assumed co be as large as 100 km, the rupture length ranges from about 1000 km to 3200 km for a M, 9 -91 earthquake. At the upper end, these rupture lengths imply rupture will extend beyond the subduction zone margin and into the adjacent San Andreas

l l

I l

fault zone to the south and/or into the Queen Charlotte zone to the north.

We find this scenario to be untenable. At the lower end, ruptures of 1000 to 1400 km imply rupture of the entire arc, including rupture into the adjacent Gorda and Explorer plates, which show very different seismic and tectonic behavior. A rupture length of about 1000 km entails the entire length of the are and nowhere (including short arcs such as Mexico and the Lesser Antilles) have we observed the rupture of an entire arc. Alternatively, if the 250 km rupture length defined by the segmentation of the zone is given, the downdip widths impled by M, 9 - 91 are about 400 - 1300 km. Such a down-dip dimension is physically implausible based on the thick accretionary wedge and the extremely short Benioff zone length reflecting equilibration of the slab into the mantle at depths significantly shallower than required for these widths. We conclude that our estimates of maximum magnitude are appropriate for the evaluation of site ground motions and that postulations of M, 9+ earthquakes along the Cascadia zone have no physical basis, and are therfore not credible.

IV. GROUND MOTION ATTENUATION RELATIONSHIPS FOR SUBDUCTION ZONE SOURCES The evaluation of ground motions at the WNP-3 site from potential subduction zone earthquakes require the estimation of ground motions fro = both interface thrust earthquakes and intraslab earthquakes (see Figure 10) . Existing published ground motion attenuation relationships for subduction zone earth-quakes (Iwasaki and others, 1978; Sadigh, 1979; Iwasaki, 1980; NOAA, 1982; Mori ano others,1984; Vyas and others,1984; Kawashima and others,1984; Crouse and others,1988) indicate that at d4=ter.ces greater than 50 km from the earthquake rupture ground motions from subduction zone earthquakes are substantially larger than those from shallow crustal earthquakes. However,  ;

the published relationships have been derived largely on the basis of soil 1

site recordings wh'ile the WNP-3 plant is founded on rock with shear wave '

velocities in excess of 3000 ft/sec (900 m/sec). In addition, the empirical data bases used in those studies consist primarily of recordings obtained at i distances greater than 50 km from events of magnitude s M 7 5, requiring {

extrapolation to the distances and magnitudes of interest. l l

l 1 i l l 1

l L j l

In light of the limited applicability of the published relationships for ,

estimating ground motions at the WNP-3 site, an analysis of the available ground motion data was performed for this study. An extensive set of peak ground acceleration data was collected, including data.from the magnitude M, 8 events in Chile and Mexico during 1985 Figure 27 shows a scattergram of the available subduction zone earthquake recordings on rock or rock-like material (shear wave velocity :t 750 m/sec) and on soil sites. As can be seen, the 1985 recordings have significantly expanded the data base for large magnitude, near field strong motion recordings on rock. Many of these recordings were obtained in the sa=e relative location with respect to earthquake rupture as the postulated situation at the WNP-3 site (i.e.,

directly or nearly directly abave the zone of rupture). These data provide a more appropriate basis for estimating near field motions for events of magnitude s M, 8 at the WNP-3 site than the existing published relationships.

Attenuation relationships for estimating pe-k horizontal accelerations and 5-percent dasped horizontal response spectral velocities on rock from subduc-tion zone earthquakes in the magnitude range 5 s M, s 8 were developed from ,

analyses of the collected strong motion data. The soil site data were used

to investigate the variance structure of the data and to aid in testing various hypotheses. The results of numerical simulations of ground motions from large plate interface thrust events provided the bases for extrapolation of the attenuation relationships to earthquakes of magnitude > M, 8.

Strong Motion Data The data set collected for this study is listed in the Appendix. Data were collected for both interface thrust earthquakes and intraslab earthquakes.

The primary sources of the data from various regions were: for Alaska, Beavan and Jacob ('1984); for Chile and Peru, Saragoni and others (1982, 1985); for Japan, Mori and Crouse (1981); for Mexico, Bufaliza (1984),  ;

Anderson and others (1986, 1987a, 1987b); and for the Solomons, Crouse and I others (1980). Earthquake size was characterized in terms of moment magni-tude, M,, as defined by Hanks and Kanamori (1979). If no seismic moment has been reported for an event, then the surface wave magnitude was used provided 1

V l l l 1

l i

l i-it fell in the appropriate range of M, 5 to 7 5, consistent with the defini- ji i tion of the moment magnitude scale. If only body wave magnitude was reported, then m, values in the range of 5 to 6 were converted to M, using the relationship M, = 1.8 m, - 4 3 proposed by Wyss and Habermann (1982) and  ;

the resulting value taken to be equal to moment magnitude. This magnitude l

conversion approach was used by Beavan and Jacob (1984) in developing their I catalog of strong motion data from Alaska. Source-to-site distance was characterized in terms of closest distance to the rupture surface. If no rupture surface has been defined for an event, then hypocentral distance was used.

The distinction betwcen soil and rock site conditions was made primarily on tne basis of site conditions listed in the various data sources. The record-ing station at the Geophysical Institute in Lima, Peru was classified as a r ek-like site on the basis of the reported subsurface shear wave velocities and evaluation = of site response and damage distributions during past earth- ,

quakes (Re,etto and others, 1980). The recording station at the School of Engineeririg in Santiago, Chile was also classified as a rock-like site as it is located on deposits similar in nature to those underlying the Lima site.

It should be noted that several of the recording stations for the 1985 Chile I earthquake listed as located on rock in Wyllie and others (1986) are actually located on soil deposits, notably the stations at Llolleo (Algermissen, l 1985) and Melipilla (Algermissen, personal communication). The recordings obtained on the very soft lake deposits such as those in the Mexico City area were not included in the analysis as they may represent soil sites with special amplification characteristics. (These recordings are designed by the site classification "SA" in the Appendix).

As discussed by Campbell (1987), there are strong component-to-component correlations for l'ndividual ground motion records that undermine the validity of using statistical tests under the assumption that the two horizontal I components of motion represent independent measurements. Therefore, for l

this study horizontal ground motion parameters (peak acceleration, spectral l velocity) were characterized in terms of the geometric average of the two l

horizontal components of motion. The results of the analyses indicate no 4

1 1

' \

- - - ------- , ..,,.--n, .,-.,-,,,.,,,..-,-,,,,,,.-,v.~. , ~ . - . , . , - , . . - . . , , - . , , , . - - ~ - - , , , , ,

l significant difference between the estimates of variance about the median relationships obtained using the average of two components and the values obtained using both components as independent data points, j The collected data set consists of the following groups of data.

l Subduction Number of Recordings Magnitude Distance ,

Zone On Rock On Soil Range Range Alaska 18 3 5 2 - 8.0 27 - 231 km Cascadia 0 5 67-71 55 - 80 km Chile 20 25 5 3 - 8.0 39 - 175 km Japan 0 59 5 2 - 7.6 42 - 390 km Mexico 45 -72 S.0 - 8.0 15 -'456 km Peru 13 1 5 3 - 8.1 68 - 260 km Solomons 9 35 5 0 - 8.1 38 - 418 km Approximately three quarters of the data were from interface thrust ecrth-quakes and the remainder from intraslab earthquakes. Data listed in the Appendix from events with estimated M, magnitudes less than 5 0 were not used in the analyses.

i l Attenuation Relationships for Peak Horizontal Acceleration on Rock Figure 28 shows the peak horizontal acceleration data for magnitude M, 5 5 to 8 earthquakes recorded on rock and soil sites. As can be seen, the soil i 1

recordings have on average greater peak motions than rock site recordings. I Although exhibiting a large degree of scatter, the data show a trend toward 1 j

near field distance saturation of ground motion levels for large magnitude j events that is well established for ground motions from shallow crustal earthquakes. Consequently, the general mathematical form used by many investigators for shallow crustal ground motions (e.g Campbell. 1981; Joyner and Boore, 1981: Sadigh, 1983) was employed in the analysis of the data.

The specific form of the relationship is:

In(a,n ) = C +CM 1 2 w - C 31n[R + C4 exp(C5M v

)] - Y R + E (1)

, , . . - ._ .. . _- .- - . . ~

l where R ~is closest distance to, the' zone 'of rupture in kilocoters, Y repre-

'sents an anelastic attenuation coefficient. O_ are coefficients determined g

from the data and E represents a normally distributed random error with zero mean.

l The ters TB in Equation 1 should have a negative sign to give the physically reasonable model of energy absorption as the seismic waves propagate away from the source. However, initial analysis of the data with Y unconstrained resulted in Y values less _than zero (positive YR term) for both soil and rock data sets. Constraining Y to be 2 0 resulted in fitted Y values very '

l near zero, indicating that the effects of geometric spreading and anelastic attenuation could not be separated in the data. Therefore, the term YR was dropped from the attenuation relationship.

The parameters of Equation 1 were obininad from the data sets shown in Figure 28 using nonlinear multiple regression techniques. The normalized residuals obtained from fitting the data are plotted against magnitude and distance in Figure 29 Inspection of the residuals indicated no trend with distance and a reduction in variance with increasing magnitude. Similar .

I dependence of the variance in ground motion on earthquake magnitude have - l been reported by Sadigh (1983), and Abrahamson (1987). Both Sadigh (1983) I and Abrahamson (1987) sugAest that the variability in the variance for peak i 1

ground motions can be modeled by a 11neer relationship between magnitude and standard error of log acceleration. The coefficients of such a relationship can be obtained by minimizing the expression (Gallant, 1987): l n

E(lel-/no*//2)2 g (2) i=1 where e, is the unnormalized residual for the it h data point and o' repre- )

sents the functional form for the standard error, assumed in this analysis to be:

1 o' = a + bM, (3) a 1

H

)

i .

The residuals for both rock and soil data sets were found to be significantly correlated with magnitude. Consequently a weighted-least squares approach-(Draper and Smith,1981; Gallant, 1987) was'used in all subsequent analyses, with weights inversely proportional to the variance as defined by Equation 3 As the differences in the variance estimates for the soil and rock data -

sets were not statistically significant, the residuals from two data sets were combined to estimate the parameters of o' . 6 As discussed in Section III (see also Figure 10), subduction zone earthquakes can be grouped into two basic types of events, low-angle thrust earthquakes occurring on plate interfaces and high-angle, predominantly normal faulting earthquakes occurring within the downgoing slab (see Figure 10) . As it has been suggested that the type of fault rupture may have an effect on median ground motion levels (e.g., McGarr 1984; campbell, 1987) possible differences between ground motions from interface and intraslab events were investigated.

The differentiation between interface and intraslab events was done on the basis of mechanisms, when reported, or on the basis of focal depth, with events below a depth of 50 km considered to be intraslab events. While it is unlikely that interface events would occur at depths greater than 50 km, intraslab events do occur at depths less than 50 km, and it is possible that some intraslab events have been misclassified as interface events.

, The residuals for interface and intraslab events are differentiated in Figure 29 As can be seen, the residuals for intraslab events tend to be J

greater than zero for both data sets. Application of the likelihood ratio  !

test for nonlinear regression models suggested by Gallant (1975a,b) indicates ,

I that the hypothesis that the coefficients of Equation 1 are the same for intraslab and interface events can be rejected at the 0.05 percentile level for both the rock and soil data sets. l To test the significance of the observed differences ire the residuals shown in Figure 3, equation 1 was modified to include a set of "dummy" variables (Draper and Smith,1981) to identify data from interface and intraslab events, yielding the relationship:

)

,-- ,m - _ . , - - . . , , _ . . . - , . , , - - - ,, ..---,,- ,-,v, . - - , _ ,--,-,,,------,,.--.+m. ,

I i

l In(a,,,) = C g + C2 M, + C3 In[R + C4 exp(C5 M,)] + BZ g + E (4) where 2, is zero for interface events and one for intraslab events. The coefficient B measures the average difference between the ground motions from interface and intraslab events. Equation 4 was fit to the data (using weighting based on equation 3) resulting in an averago value of B = 0 54 for the soil and rock data sets. Application of the likelihood ratio test

indicates that the hypothesis that B = 0 can be rejected at the 0.05 percen-tile level for both rock and soil data sets. Further extensions of the model to include a modifying effect of rupture type on the other parameters

)

of Equation 4 produced no further decrease in the estimated variance and were rejected.

As the intraslab events tend to be both deeper (Figure 10) and to produce higher ground motions (as indicated by the positive residuals), the possi-bility of including a term to make ground motion level proportional to depth of rupture in Equation 4 was exp.ored. When the depth term was included in the analysis no significant reduction in the standard error was cchieved ,

beyond that obtained by separation of the data into the two subsets of interface and intraslab.

The validity of the systematic difference between ground motions from inter-face and intraslab events was further investigated using the data from sites with multiple recordings of both types of earthquakes. Equation 4 was f modified to include a set of dummy variables, one for each site to remove the differences between the median ground motions at a site and the overall median over all sites - in essence removing the effects of systematic site 1

amplification. Coefficients C3 through C 5where held fixed at the values obtained from regression on the full data set and a linear regression was  !

performed to obtain the individual site terms and B, the rupture type term. j The resulting value of 0 55 agrees very well with the value obtained from I

the unconstrained regression using the full date set. l I

The residuals were also examined for evidence of systematic differences between different subduction zones. Systematic differences in source

i l

characteristics have been reported previously from examination of teleseismic records. Hartzell and Heaton (1985) found that earthquakes from a particular subduction zone tended to have similar characteristics in terms of their sotrce ti=e functions of energy release and that these average characteris-tics varied among different subduction zones. However, they were unable to identify an obvious relationship between the physical parameters of an individual subduction zone and the characteristics of the rupture process l for events occurring in the zone. Houston and Kanamori (1986) indicate that the source spectra for large earthquakes on the Mexican subduction zone tend to show lower levels of high frequency ground motion'than an ',' average" large event source spectrum. <

> Crouse and others (1988) found statistically significant differences in  :

response spectral ordinates for ground motions recorded on soil sites from different subduction zones. They wer unable to identify any correlation  ;

between differences in ground motions and physical characteristics of the various subduction zones. In contrast to the teleseismic observations described above, Crouse and others (1988) found that ground motions from Mexico recorded or stiff soil sites did not appear to be significantly lower than data from other zones.

Preliminary analysis of the initial data set developed for this study (reported by Youngs and others,1987) suggested that systematic differences may also exist in peak acceleration values from different subduction zones.

However, subsequent addition of data and reclassification of site conditions at some of the recording stations has yielded a data set that does not ex-hibit statistically significant difference in the peak accelerations among the different subduction zones.

The resulting median attenuation equations are:

l In(a,,,) = 19 16 + 1.045M, - 4.7381n[R+205 5exp(0.0968M,)] + 0 54Z, (5) for rock sites and 4

i

1 l

In(a,,,) = 16.75 + 1.045M, - 4.5651n[R+162 5exp(0.1309M.)] + 0 542, (6) i for soil sites. In the analysis, parameter C2 , which represents the far  !

field magnitude scaling term, and parameter B, the faulting type term were assumed to be the same for both soil and rock data. The resulting relation-ship for standard deviation of peak ground acceleration is o' = 1 55 - 0.125M. (7)

Figure 30 presents plots of the residuals about Equations 5 and 6 normalized by the standard error given by Equation 7 Comparison of these results with those shown in Figure 29 indicates that a homogeneous variance has been obtained and the bias in the intraslab residuals has been removed.

The median attenuation relationships specified by Equations 5 and 6 are com-pared with the recorded data in Figures 31 and 32 for rock and soil data, '

respectively. These figures clearly show the separation between'the data-for interface and intraslab events. '

l Attenuation Relationships for Spectral Velocity '

1 As only limited digitized ground motion recording obtained on rock sites ,

were available, attenuation relationships for spectral velocity (S,) for i

rock site motions were developed using the procedures employed by Sadigh i (1983, 1984). This involves developing relationships for the ratio S,/a,,,

as a function of magnitude and distance and then applying these relationships-to attenuation relationships for peak acceleration. The advantages of this j approach are that there is a much larger data base of peak acceleration data i than spectral response data for establishing magnitude and distance scaling of absolute levels' of ground motion and the use of spectral shapes results in attenuation relationships for various periods that are consistent over the full range of magnitudes and distances to which the relationships apply. l The procedure involves three steps: first, developing a spectral shape for a reference size event for which there is abundant data; second, developing i

i l

l

relationships to scale the shape to other magnitudes; and third, computing the standard error of the absolute spectral values about the attenuation relationship. For this analysis, a reference magnitude of M, 8 was used as the largest number of available rock site response spectra are from recordings of M, = 8 events.

Figure 33 presents median (mean of In[S,/a,,,]) spectral shapes for 5 percent de.mping developed from magnitude M, 7.8 to 8.1 ground motion data. The top plot shows the computed spectral shapes for the data from distances less than 150 km and greater than 150 km. As can be seen, there is a significant dif-ference in spectral shape for the two distances ranges. Because the interest is on near field ground motions, the spectral shape for the < 150 km distance data was used to develop the spectral velocity attenuation relationships.

The bottom plot of Figure 33 shows the statistical spectral shape in terms of spectral acceleration and the smoothed spectral shape used to define the attenuation relationships. The maximum spectral acceleration amplification for 5 percent damping is 2.25 at a period or 0.15 seconds.

The second step is the specification of the variation in spectral shape with earthquake magnitude. Figure 34 presents plots of the ratio [S,/a,,,(M,)]/

[S,/a,,,(M,=8)] derived from the available response spectra data for record-ings on rock sites for periods of vibration of 0.2 and 2.0 seconds. The ratios were obtained by dividing the spectral amplifications for individual events by the smoothed spectral shape for a magnitude M, event shown in l Figure 33 The three curves shown represent relative spectral a=plification I for shallow crustal events derived on the basis of empirical attenuation relationships (Joyner and Boore,1982; Sadigh and others,1986) and on the basis of numerical models employing theoretical source spectra and random vibration theory t'o estimate ground motions (Hanks and McGuire, 1981: Boore, 1983, 1986). As can be seen, the data for subduction zone earthquakes fol-lows the general trend defined by the relationships for shallow crustal ear thquakes . Accordingly, the form of the relationship for spectral amplifi-cation employed by Sadigh (1983) was used. Specifically:

, -100-

In(S,/a,,,) = C6 + C7 (Cs~M)' w (8)

Equation 8 was fit to the data for periods between 0.1 and 3 seconds. In conducting the regression Cs was fixed at 10 to provide for complete satura-tion at magnitude M, 10 and C, was fixed at 3 representing an average of the values obtained at longer periods where the data exhibit significant magni-tude effect on spectral shape. Applying these constraints, parameter C7 was found to vary linearly with the log of natural period and could be fit by the relatiorship:

C7 = -0.0145 - 0.00631n(T) (9) where T is natural period of vibration in seconds, For periods less than 0.1 second, parameter C7 was set to zero, resulting in spectral ordinates that are independent of magnitude at high frequencies. Parameter C6 was then constrained so that Equation 8 results in the spectral amplifications specified for magnitude 8 events by the spectral shape shown in Figure 33 The resulting relationships are compared with the empirical data and the relationships for shallow crustal earthquakes in Figure 34. i Attenuation relationships were developed for periods between 0.04 and 3 0  !

seconds. The spectral amplifications at these periods are given by the relationships:

T = 0.04 see in(S,/a,,,) = 1 960 T = 0.07 see in(S,/a,,,) = 2.845 T = 0.1 see In(S,/a,,,) = 3.431 T = 0.15 see in(S,/a.,,) = 3 985 - 0.0026(10 - M )3 T = 0.2 see In(S,/a,,,) = 4.278 - 0.0044(10 - M,)3 i

T = 0 3 see In(S,/a,,,) = 4.652 - 0.0069(10 - M )3 T = 0.4 see in(S,/a,,,) = 4.906 - 0.0087(10 - M )3 (10)

T = 0 5 see In(S,/a,,,) = 5.076 - 0.0101(10 - M )3 l T = 0.7 see in(S,/a,,,) = 5 154 - 0.0123(10 - M,)3 T = 0.8 see In(S,/a,,,) = 5 164 - 0.0131(10 - M,)3 T = 0.9 see in(S,/a.,,) = 5 167 - 0.0138(10 - M,)3

-101-l

(

. . _ - - - . _ _ . _ , , - . . - _ . , - , , , - , , - _ - , - , . - - , _ - - , - , - . , ~ . . , . _ , . . - , - - , - . , , , , , . - .

i l T = 1.0 see In(S,/a,,,) = 5.140 - 0.0145(10 - M.)3 j

T = 1 5 see in(S,/a...) = 5.059 - 0.0170(10 - M,)3 T = 2.0 see in(S,./a,,,) = 4 960 - 0.0189(10 - M,)3 T = 3 0 see in(Sy /a,,,) = 4.725 - 0.0214(10 - M,)3 The units of Equation 10 are em/sec/g.

l The third step is specification of the standard error in in(S,) . The stan-dard error was estimated by co=puting the residuals of the response spectral values about the median relationships for spectral velocity given by Equation

10. The resulting values were less than or equal to the standard deviation for peak acceleration defined by Equation 7 at all periods. Accordingly, Equation 7 was used to define the standard deviation in log spectral velocity.

Comparison with Published Empirical 1.ttenuation Relationships Figures 35 through 37 compare the attenuation relationships developed above with other empirically based relationships for subduction zone earthquakes.

Bufaliza (1984) developed attenuation relationships from recorded data in Mexico. He divided the data on the basis of site classification into two data sets, "fir:e", consisting of rock and (presu= ably) stiffer alluvial soil sites and "blando" consisting of deeper soil sites including lake depos-its. Figure 35 co= pares Bufaliza's (19841 telationships with Equations 5 and 6 for cagnit- : M, 7 and 8 events. As car, be seen, Bufaliza's relation-ships indicate similar differences between sail and rock ground motions as were obtained in this analysis. Bufaliza's relationships lie intercediate betveen the interface and intra-slab relatio0 ships developed in this study, reflecting the fact that he did not discriminate between the two types of I l

events in his anal' ysis.

l Iwasaki (1980) developed relationships for peak acceleration on rock and soil sites in Japan. He co= bines data from crustal as well as subduction zone earthquakes although the larger magnitude data are predominately from subduction zone events. As Iwasaki used only data from events with l

-102- '

. . . _ _ _ m . . _ _ _m hypocentral depths of 60 km or less, his relationships are more comparable with the interface relationship developed in this study. Figure 36 compares Iwasaki's rock and alluvial soil relationships with Equations 5 and 6 for magnitude M, 7 and 8 events. Iwasaki's relationships indicate a somewhat smaller differer.ce between soil and rock motions than were found in this  !

study.

Crouse and others (1988) present attenuation relationships for spectral velocity on soil sites based on data from Japan. They include a term propor-tional to focal depth in their relationships but did not find a statistically significant difference between shallow thrust (interface) and normal (intra-slab) events. Given the different depth ranges in which -the two types of events typically occur, the difference between interface and intraslab events found is likely cecounted for by their focal depth dependent relation-ships. Figure 37 compares the relationships of Crouse and others (1988) for focal depths of 30 and 70 km with Equation 6. These two depths are represen-tative of typical foal depths for interface and intraslab events. The peak  ;

acceleration curves labelled Crouse et al. were obtained from their spectral

, velocity relationship for a period of 0.1 seconds assuming a spectral ampli .

fication of 2 at that period. As can be seen, the differences between the 30-km and 70-km focal depth curves are comparable to the differences between ,

interface and it.traslab events found in this study.

The comparisons shown in Figures 35 through 37 indicate that the trends J found in this study with respect to effect of focal mechanism and site classification on ground motions from subduction zone earthquakes are similar to those observed by other investigators.

Vertical Ground Motions i

The attenuation of peak vertical acceleration was examined on the basis of the ratio of vertical to horizontal acceleration. Figure 38 shows the ratio of vertical to horizontal peak acceleration plotted versus distance for magnitude M, 5 5 to 8 events. Analysis of the data indicated that the ratio

)1 of vertical to horizontal accelerations is independent of magnitude and type

-103-

1 r

i of-rupture, and shows a weak tendency to increase with distance to rupture.

The only statistically significant dependence was on site classification with a,,,(vertical)/a,,,(horizontal) equal to 0.61 for rock sites and 0 51 ,

for soil sites. Accordingly, an attenuation relationship for peak vertical acceleration on rock was obtained by adding-the factor in(0.61) to Equation 5 The resulting median attenuation equation is:  !

In(a,,,) = 18.67 + 1.045M - 4.7381n[R+205 5exp(0.0968M,)] + 0 54Zt (11)

Equation 11 is compared with the recorded data in Figure 39 The dispersion of the vertical data about the median relationship was found to be very similar to that observed for the horizontal data and Equation 7 was used to specify the standard error as a function of magnitude for log a,,,.

Attenuation relationships for spectral velocity (S,) were developed using the procedure ecployed for horizontal motions. Figure 40 presents median (mean of in[S,/a,,,]) spectral shapes for 5 percent damping developed from magnitude M, = 8 vertical ground motion data. Significant differences were found between the spectral shape for the data from distances less than 100 km and greater than 100 km (shown at the top of Figure 40). Because of the interest in near field ground motions, the spectral shape for the < 100 km distance data was used. The bottom plot in Figure 40 shows the statistical spectral shape for the < 100 km distance data in terms of spectral accelera-tion and the smoothed spectral shape used to develop the spectral velocity attenuation relationships. The maximum spectral acceleration amplification- l is 2.28 at a period of 0.2 seconds. j l

The variation in vertical spectral shape with earthquake magnitude was assumed to be the same as was found for horizontal motions and Equations 8 and 9 together with the smoothed spectral shape shown at the bottom of Figure 40 were used to develop the following attenuation relationships for periods between 0.04 and 3 0 seconds.

Te 0.04 see In(S,/a,,,) = 2.103 T = 0.05 see In(S,/a,,,) = 3 033

-104-

l I

i T = 0.1 sec in(S,/a,,,) = 3 506 T = 0.15 see in(S,/a.. ) = 3 967 - 0.0026(10 - M,)3 T = 0.2 sec in(S,/a,,,) = 4.303 - 0.0044(10 - M.))

T = 0 3 see In(S,/a,,,) = 4.683 - 0.0069(10 - M,)3 T = 0.4 see In(Sy /a,,,) = 4.791 - 0.0087(10 - M )3 (12) i T = 0 5 see in(S,/a,,,) = 4.767 - 0.0101(10 - M )3 '

T = 0 7 sec in(S,/a,,,) = 4.850 - 0.0123(10 - M )3 ,

T = 0.8 see in(S,/a,,,) = 4.866 -.0.0131(10 - M,)3 i T = 0 9 see in(Sy/a,,,) = 4.874 - 0.0138(10 - M )3 T = 1.0 see in(S,/a,,,) = 4.882 - 0.0145(10 - M,)3 J

T = 1 5 see in(S,/a,,,) = 4.937 - 0.0170(10 - M )3 T = 2.0 see in(S,/a.. ) = 5 015 - 0.0189(10 - M )3 T = 3 0 see in(S,/a,,,) = 5 017.- 0.0214(10 - M,)3 The units of Equation 12 are cm/sec/g. Equation 7 was used to specify the standard error in in(S,).

Ground My.lons for Earthquakes of Magnitude Greater than 8 t The attenuation relationships defined above are considered applicable for estimating ground motions in the magnitude range M, 5 to 8 and for distances of 20 to 500 km. Extrapolation of the above attenuation relationships to larger magnitude events than have been recorded requires specification of the appropriate near field magnitude scaling law for ground motions. Past applications of the general form of the attenuation relationship defined by Equation 1 have typically followed two limiting cases. One approach has been to assume that the scaling of ground motions with magnitude is indepen-dent of distance, implying parameter C5 = 0. Examples of this approach are l the attenuation relationships developed by Joyner and Boore (1981, 1982) for ,

western U.S. strong motion data and most of the empirical relationships for l subductiou zone earthquakes referenced above. Attenuation relationships based on self-similar scaling of earthquake source spectra and random vibra-tion theory (Hanks and Meouire, 1981; Boore, 1983, 1986) also imply distance- j independent magnitude scaling, except for the modifying effect of anelastic l attenuation at large distances. The second approach has been to assume

-105-4

ground motions are independent of magnitude at zero distance, implying para =eter C 5 = ~C2 /C3 . Examples of this approach are the attenuation relationships developed by Campbell (1981, 1987).

It is likely that the true form of a near field magnitude scaling law is intermediate between the abovo limiting cases. The attenuation relationships developed by Seed and Schnabel (1980), Sadigh (1983, 1984) and Sadigh and others (1986) are examples of intermediate magnitude scaling laws developed from empirical data. These relationships have nearly distance-independent magnitude scaling for events below about magnitude M, 6.5 and nearly magni-tude independent peak accelerations at zero distance for events of magnitude greater than a, 't 6 5 Joyner (1984) has proposed that there is a critical earthquake above which the self-similar scaling of earthquake source spectra no longer applies. He suggests that the high frequency corner of the source spectrum becomes fixed for events that rupture the entire width of the seis-cogenic zone, resulting in a reduction by a factor of about 2 in the increase in ground motion amplitude per unit increase in magnitude for events above the critical size. Joyner (1984) estimates the critical magnitude to be approximately 6.5 for crustal events in the western U.S. Hartzell and Heaton (1988) propose that self-similar scaling breaks down for large thrust events greater than about M, 81 on the basis of both observations of teleseismic P-wave recordings and theoretical arguments similar to those of Joyner (1984). They propose that the scaling of the acceleration source spectrum with seismic moment for events above M, 81 be equal to one-half to two-thirds of the scaling on the basis of self-similarity. This should translate into comparable or greater reduction in the scaling of peak amplitude to those obtained with the Joyner (1984) model. These arguments suggest that the nearly distance independent magnitude scaling represented by Equations 5 and 6 overestimates the near field magnitude scaling of ground motions from very large (M, ) 8) earthquakes.

The characteristics of near field ground motions for evento larger than those that have been recorded can be investigated using ground motion simula-tions. An extensive set of simulations of ground motions from potential large interface earthquakes has been conduced for the WNP-3 site by S-Cubed

-106-

(1988). The simulated motions were obtained by the superposition of the l_

motions from a large number of subevents propagated to the recording site using ray theory (Day and Stevens, 1987). In the earthquake source model

! the rupture surface is represented by an assemblage of subregions. The radiation from each subregion slip episode is obtained numerically from a

! dynamic simulation of faulting based on three-dimensional finite difference I

solutions to propagating crack problems (Day, 1982a, b,; Stevens and Day, 1985). The radiated seismic pulses are scaled to the prescribed values of subregion dimension and local subregion stress drop. A large earthquake rupture is simulated by a kinematically prescribed superposition of subregion radiations with a stochastic element incorporated.

Figure 41 shows schematically how subregion contributions are combined.

Each frame at the top of Figure 41 is a snapshot of rupture at a given time.

A global rupture front sweeps the fault with a prescribed rupture velocity of 90 percent of the shear wave velocity. When a subregion is subsumed by the clobal rupture front, a subevent is triggered in that subregion. Shading in Fig,re 41 indicates subregions that are actively slipping and an arrow denotes aubregions in which a slip episode has been completed. The stress drop ant '4ource dimension of the subevents are selected randomly from a specified distribution. The mean subevent stress drop and source dimension were prescribed so as to optimize agreement between recorded and simulated motions for the 1985 Mexico and Chile M, 8 events.

Expansion of the rupture front and the consequent triggering of adjacent subregions will reload a subregion, and the model permits repeated failure of previously slipped regions. This is illustrated at the top of Figure 41 where, for example, subregion A is triggered at time t,, is locked at time

, t3 , but is then reloaded by expansion of the rupture front and triggers again at t . This secondary and other secondary subevents additionally load and trigger adjacent regions. For an overall fault dimension large compared to the subregion size, this retriggering may have to occur repeatedly to build up sufficient slip to accord with the prescribed seismic moment of the earthquake being simulated. The diagrsm at the bottom of Figure 41 illus-trates the resultant slip history at a representative point on a large fault

-107-

l rupture that has undergone five slip episodes. The number of slip episodes 1 i

required is just the average fault slip associated with the prescribed seis- l mic moment of the simulation divided by the average subevent slip. These slip events are distributed randomly over the rise time.

The seismic pulses were propagated from the subregions to the recording site using a horizontally stratified anelastic earth model. Two simplifying approximations were made in computing ground motions at the surface. First, j it was assumed that the Green's function of the earth model can be approxi-

)

mated by geometric ray theory. A consequence of this assumption is the neglect of near field contributions, leading to accurate computations only l for wavelengths that are small compared to the source-to-site distance. i Second, variation of the Green's function over.the subregion was-neglected, )

apart from corrections for travel time variations due to changes in the ray path length (equivalent to the Fraunhofer approximation, Aki and Richards, 1980). These approximations together with uncertainty of the source and attenuation models at high frequency limit the frequency range for reliable  :

computation of ground motions to 0.2 to 10 Hz.

l l

The model was first tested by simulating ground motions from the main shock I and af tershocks of the 1983 coalinga, california earthquake sequence (Stevens and Day, 1987). These analyses indicated that better agreement between recorded and simulated motions was obtained when a simple square subevent model with uniform stress distribution is used rather than a more complex model with a slip-weakening criterion of failure and a non-uniform stress distribution. The tests also indicated that much more consistent results were obtained when the Green's functions are "homogenized" by averaging the l horizontal components and averaging the radiation pattern coefficient over l the focal sphere. )

i The ability of the model to generate near field ground motions from large I

subduction zone thrust earthquakes was tested by simulating the near field j recordings from the M, 8.0 September 19, 1985 Michoacan, Mexico and March 3, l 1985 valparaiso, Chile earthquakes (Day and Stevens,1987; S-cubed,1988) . ,

1 The simulations were performed using a subregion size of 2 5 km, an average '

-108-i 1

4 i

_ , , . , - - , . . - . . - . _ , _ _ . , m. _ .

~ _ _ _ . . - _ _ _ _ _ _ .

_.-_.,_.m__-, -

_____,.,_____-..__m.._,

- __ _ =_ . . _ . _ . _m . _

i local stress drop of 38 bars and wave propagation characteristics estimated-from the P-wave velocity model given by Havskov and others (1983). Figure

'42 compares the response spectra for simulations of the ground motions at  !

the three rock sites located above the rupture of the Michoacan earthquake l and one rock site located above the rupture of the Valparaiso earthquake I l

(the only rock site response spectra from this event presently available) l l

with the response spectra for the recorded motions. As can be seen.-the i i

response spectra for the simulations are in good agreement with those for f

the recorded motions, i i The appropriate form of the near field magnitude scaling relationships were evaluated by examining the response spectra for a series of simulated ground motions at the WNP-3 site at distances of 30 to 40 km above the rupture surface of events in the magnitude range of M, 7 7 to 8.9 Figure 43 pre-

, o sents the scoothed median response spectra obtained from the simulated  :

motions. Each spectrum represents the median of four spectra representing two simulations and two horizontal components of motion. Figure 44 compares the relative amplitudes of spectral velocity for various magnitude events obtained from the simulations with the near field magnitude scaling relation- I ships derived from the empirical attenuation relationships defined by Equa-tions 5 and 10. Also shown in Figure 44 are the scaling relationships obtained by imposing the two limiting conditions of distance independent ,

magnitude scaling (C3 = 0) and magnitude independence at zero distance (C 5 "

-C,/C3 ) on the empirical data. There is considerable scatter in the simula-tion results but a linear fit to the data (dotted line) suggests that the magnitude scaling relationships for events above magnitude M, 8 should have a flatter slope than indicated by the empirical data from earthquakes of )

magnitude less than M, 8.

Attenuation Relationships for M_ > Earthquakes l

l The magnitude scaling relationships developed from the S-Cubed (1988) simula- j

tion results and from use of the Joyner (1984) source model were used to l l'

develop attenuation relationships for peak acceleration applicable to earth-quakes larger than M, 8. The "S-Cubed Model" was developed by constraining

-109- l l

l 1 1 4

j

. . -- . . - . _~ . .. .

i f

parameter C 5 in Equation 4 such that the resulting near field magnitude scaling relationship approximated that indicated by the results of the simulations shown in Figure 44 for events of M, ) 8 and matches the median values given by Equation 5 for M = 8. The resulting relationship is:

in(a,,,) = 19 16 + 1.045M - 4.7381n[R+154.7exp(0.1323M,)] (13) 4 The "Joyner Model" was developed by setting parameter C 3 in Equation 4 equal ,

to zero, consistent with distance independent magnitude scaling implied by l the random vibration approach to developing attenuation relationships, and adjusting coefficients C 1. C,, and C4 in Equation 1 such that the resulting magnitude scaling relationship equals that obtained using random vibration theory and the Joyner (1984) source model with the critical magnitude set equal to M, 8 (corresponding to a square rupture with a width of the seismo-genic zone of about 100 km) and matches the median values given by Equation 5 for the critical magnitude. The resulting relationship is: l In(a,,,) = 24.64 + 0 36M, - 4.7381n(R+445 8) (14) 4 The near field magnitude scaling relationships based on Equations 13 and 14 i are shown in Figure 45 The "Joyner" model results in a somewhat lower rate i of increase in ground motions with magnitude than implied by the "S-Cubed" model. Equations 13 and 14 are considered applicable to interface earth-quakes with magnitudes above the critical magnitude of M, 8. The dispersion  ;

about the median relationships defined by Equations 13 and 14 was assumed to be 0.55. equal to the value obtained from the empirical data for M, 8 events.

2 Comparison with other Numerical Modelling Results i

i l

Heaton and Hartzell (1986) have used an empirical Green's function technique l l to estimate the motions that would result from a postulated large earthquake on the interface between the Juan de Fuca and North American plates at a

! depth of approximately 30 km beneath a site in the coast ranges of Washington state. They used empirical Green's functions developed from strong motion l

l recordings of events in the magnitude range of M, 7 to 7 5 to simulate the 110-l l l

l 9--,--- , - - ., ,y

.,- -a-ww-n---%=g,.wwwymw---- e. gv'eww----+ e -e, g .---,wv--c--m=w-----, + - , mi-we-+wg--e-y- r ww . ey,- m y*.we,ir-+ymu-pre s g w -e<vewwg- -w-

. - . . . . .. ~ . - ~.

i l

ground motions above the rupture surface of earthquakes in the magnitude  ;

range of M, 7.25 to 9 5 Figure 46 compares their estimated spectra (Heaton and Hartzell,1986. Figure 17) with the response spectra predicted by the L

attenuation models developed in this study. The comparisons show very good l agreement except for motions at periods greater than about 0.8 seconds. The differences at longer periods are likely due to Heaton and Hartzell's use of I soil site recordings from Japan as Green's functions. Empirical observations i

of shallow-crustal earthquakes show that long period motionz on soil sites are typically a factor of 2 greater than those on rock sites.

V. SITE GROUND MOTIONS FROM POST'JLATED SUBDUCTION ZONE EARTHQUAKES In this section we present estimates of horizontal and vertical response spectra associated with the maximum magnitude subduction zone earthquakes  ;

described in Scction III. The maximum magnitude estimated for a shallow thrust interface eat thquake is M, 86 and the maximum magnitude for intraslab earthquakes is estimated to be M, 71.

The attenuation relationships for subduction zone earthquakes developed in Section IV use closest distance to earthquake rupture as the distance j measure. The smallest source-to-site distances for the two types of events l are estimated from the geometry of the subducting Juan de Fuca plate des-cribed in Section II and shown in Figure 6. i Figure 47 (reproduced from Figure 6) shows the down-dip cross section of the subduction zone in the site vicinity. Shown in the figure are the minimum distances to the plate interface and to the zone of flexure within the sub-

! ducting Juan de Fuca plate. The point on the interface directly beneath the

site is within the zone of possible brittle interface rupture described in

. 1 Section III (depth range 20 to 35 km) and the resulting minimum distance to  :

1 As discussed in Section III.

potential large interface earthquakes is 33 km.

large intraslab earthquakes are expected to occur in the vicintty of the downdip bend in the subducting Juan de Fuca plate. The location of the zone i

of flexure is shown in Figure 47 The minimum distance from the WNP-3 site to the zone of flexure within tne Juan de Fuca plate .ls approximately 70 km.

l -111-4

- - , . _ ,.. , ,,_ -~ __vv ,.v,.,, , . -,_-,---...1,.-r-.v,...,-.--.m -7 , . - - - -m---,m, - , _ . , . . , - ._---..y.--.

l l

Comparison of SSE Spectra with Response Spectra for Interface Earthquakes

) Figure 48 ccmperes 5-percent damped horizontal and vertical response spectra for a M, Si interface earthquake at a distance of 33 km with the SSE spectra._  ;

The median horizontal spectra were obtained using equations 13 and 10, and 3 the median vertical spectra were obtained using equation 13 reduced by a factor of 0.61 and 12. The mean and 84th percentile spectra were obtained assuming the peak motions are log normally distributed and using the magni-tude-dependent standard error defined by Equation 7 with a minimum value of 0 55 The associated median, mean, and 84th percentile peak horizontal accelerations are 0.20, 0.23, and 0 35 g, respectively. As can be seen, the horizontal SSE spectrum is only exceeded by the 84'h percentile interface event spectrum for frequencies greater than 25 Hz, with a inaximum exceedance -

of 9 percent at frequencies greater than 30 Hz, The associated median, mean, and 84th percentile peak vertical accelerations are 0.12, 0.14, and 0.21 g, respectively and the vertical SSE spectrum is not exceeded. On the basis of ,

this comparison, the SSE design spectra are considered to adequately envelope the response spectra for the postulated maximum interface earthquake.

Question 230.ld asks for an estimate of the largest interface earthquake 1

that could occur beneath the hWP-3 site without exceeding the SSE spectrum.

Figure 49 shows median responst spectra for events of magnituce M, 81 to 91 '

estimated using the two attenuation relationships for peak acceleration pre- >

sented w Section IV. T!.e "S-Cubed" scaling spectra were obtained using Equation 13 which is based extensive numerical simulations performed specifi-cally for the hWP-3 site and is considered the most appropriate relationship.

The "Joyncr" scaling spectra were obtained using Equation 14 which !s based on Joyner's (1984) model for the breakdown of s lf-similar scaling of earth-quake source spectra at large n.agnitudes. Similar scaling relationships j have been proposed' by Hartzell and Heattn (1988) for events greater than '-

4

81. The comparisons shown in Figure 49 indicate that the SSE spectrum is adequr.te at the median level for events up to M, 91.

i Figure 50 shows 84th percentile response rpectra for interfaca ever,tm in the magnitude range of M, 8.25 to 8.75 obtained t. sing the two secling relation-j -112-l l

l i

i

.--....-._ _ . - - . . . . - - - . - - _ . . _ - _ - , _ - . _ _ _ _ . . - . - . _ . - - - - _ . ,,-- ~ . . . _ . , _ - . - - - .

l ships discussed above. The comparisons indicate that significant exceed- i ances of the SSE horizontal spectrum by 84th peccentile ground motions from interface earthquakes occur for events larger than M, 81.

Figure 51 -~

', the horizontal SSE spectrum with the.respense spectra obta4.ned from g 'und motion simulations performed by Heaten and Hartzell (1986) and S-cub 388). The SSE spectrum adequately envelopes the simu-I lated motion sp ..a for events up to magnitude M, 9 On the basis of these comparisons, the SSE soectrum adequately envelopes the 84th percentile spectra for interface events up ;o magnitude M, 8) and the median spectra for events up to M, 9 to 91. The range for the median esti-mate reflects the uncertainty in assigning a quantitative statistical level, such as mean or median, to the results of the ground motions simulations.

Comparison of SSE Spectra Vith Response Spectra for Incraslab Earthquakes Figure 52 compares 5-percent damped horizontal and vertical spectre for a M, 7i intraslab earthquake occurring at a hypocental distance of 70 km from the WNP-3 site with the SSE spectra. The median horizontal spectra were obtained using equations 5 and 10, and the median vertical cpectra were obtained using equations 11 and 12. The mean and 84th percentile spSCtra WOPO obtained assuming the peak motions are log normall3 distri' outed and using the magni-tude-dependent standard error defined by Equation 7 The associated median, mean, and 84th per.:entile pc.k horizcatal accelerations are 0.16, 0.19, and 0.29 g, respectivuly. The associated median, mean, and 84" percentile peak vertical accelerations are 0.10, 0.11, and 0.18 g. The SSE horizontal and vertical spectro are not exceeded.

On the basis of the comparisons shown in Figures 48 and 52, the SSE design spectra are considered to adequately envelop the response spectra for the postulated maximum subduction zone earthquakes.

-113-

,~ _:

VI. CONCLUSIONS We have assumed that Q230.1 and 230.2 were motivated by an NRC desire to evaluate the adequacy of the SSE at the WNP-3 site relative to potential subduction zone earthquakes. Further, we view these questions as asking for a deterministic (Appendix A) approach toward site ground motion assessment, which serves to complement the probabilistic assessment provided in response to Q230.6. Our analysis of the earthquake potentia? of the Cascadia subduc- )

tion zone leads to two positions: a scientific position and a licensing position. Scientific arguments related to this issue are varied and lead to a broad range of interpretations. We believe that there is a considerable body of scientific reasons for expecting the Cascadia p ste interface to  !

-maintain its present aseismic behavior. We have ettempted to summarize these arguments in this response. However, the arguments supporting aseismic i behavior are not conclusive, and other geologic relations support the case of a seismogenic Cascadia none. Accordingly, the potential or lack of potential for the generation of large-magnitude earthquakes along the plate interface must be considered at the present time to be a scientific standoff.

While we favor the hypothesis that ti.e Cuacadia subduction . zone is aseismic, we cannot reasonably preclude arguments that the interface is capable of generating large-magnitude earthquakes. Therefore, from a licensing perspec-tive, we will consider the possibility of a large plate interface event in order to evaluate the adequacy of the SSE. j l

Given the possibility that the plate interface is capable, we have evaluated l

deterministic characteristics of the interface, and the subducting plate, as l potential seismic sources (i.e., geometry, maximum magnitudt, and resulting site ground motions) . In_these evaluations, we have adopted a reasonably conservative approach toward source characterization based on physical con-straints of the Cascadia subduction zone as well as compt isons to other sub-daction zones. The results of this analysis indicate that other deterministic conditions - possible intra-slab and interface events with maximum magnitudes of 71 and 8i respectively. -- are adequately enveloped by-the SSE response spectrum.

-114-

The response to Q230.1 and 230.2 represents the last in a series of analyses to evaluate the adequacy of the LSE at the WNP-3 site. Other analyses include an evaluation of the ground motions from the Olympia lineament (Q230 5), the largest historical earthquake within the tectoni province (Q230.4d), and a probabilistic assessment of all possible sources of signifi-cance to the site (Q230.6) . In the latter case, a concerted e ffort was made to incorporate the current scientific thinking regurding the earthquake potential of the Cascadia subduction zone by basing the hazard analysis on l the assessments of fourteen acknowledged experts on the seismology and tectonics of the Pacific Northwest. By employing this type of multi-faceted approach, we have analyzed the SSE from a variety of perspectives. The results of these analyses, including the deterministic analysis reported in this response, support and confirm the adequacy of the present seismic design basis for WNP-3

-115-

VII. REFERENCES Abe, K., 1975 Reliable estimation of the seitaic movement of large earth-quakes: Jciurnal Physics of t.te Earth, v. 23, p. 386-390.

Abrahamson, N.A. 1987, Some statistical properties of peak ground accelera-tions: Bulletin of the Seimsolgical Society of America (in press).

Acharya, H. ,1985, Comment on seismic.potentiel associated with subduction in the northwestern United States by T.H. Heatcn and H. Kanamori:

Bulletin of the Seismological Society of America, v. 75, p. 889-890.

Adams, J. ,1984, Active deformation of the Pacific Nothwest continental margin: Tectonics, v. 3, no. 4, p. 449-472.

Ahrlens and Schubert, 1975 Aki, K. . . and Richards, P.G. ,1980, Quantitative Seismology, Volume I, W.H. Freeman and Company, San Francisco, California, 512 p.

Algermissen, S.T. (ed.), 1985, Preliminary report of investigations of the central Chile earthquake of March 3,1985: U.S. Geological Survey .open File Report 85-542.

American Association of Petroleum Geologists, 1982, Plate-tectonic map of the Circan-Pacific region -- Pacific Basin sheet: American Association of Petroleum Geologists, Tulsa, Oklahora. l Anderson, J.G. , Brune , J.N. , Prince, J. , Mena, E. , Bodin , P. , -Onate. P., l Quaas, R., and S.K. Singh, 1986, As>ects of strong motion from the '

Michoacan, Mexico earthquake of Septamber 19, 1985: Proceedings of the 18th Joint Meeting on Wind and Seismic Effects, Washington D.C.,

May 12-16.

Anderson, J.G., Bodin, P., Brune, J.N., Prince, J., Singh, S.K., Quaas, R.,

and Onate M. ,1986, Sttong ground motion and source mechanism of .the Mexico earthquake of September 19, 1985 (Ms=8.1): Science, v. 233,

p. 1043-1049 Anderson, J., R. Qutus, D. Almora M., J.M. Velasco, E. Gueuara 0.,

L.E. de Pavia G., A. Gatierrez R., and R. Vazquez L., 1987a, Guerro, Mexico accelerogram array - Summary of data collected in the year 1985: l report prepared jointly by the Instituto de Ingenieria-UNAM anu the '

Institute of Geophysics and Planetary Physics-UCSD (in press) .

Anderson, J., R. Quaas, D. Almora M., J.M. Velasco, E. Gueuara 0.,

L.E. de Pavia G., A. Oatierrez R., and R. Vazquez L., 1987b, Guerro, Mexico accelerogram array - Sumnary of data collected in the year 1986:

report prepared jointly by the Instituto de Ingenieria-UNAM and the Institute of Geophysics and Planetary Physics-UCSD (in press).

-116-

Ando, M., and Balazs, E.J., 1979, Geodetic evidence for aseismic subduction of the Juan de Fuca plate: Journal of Geophysical Research, v. 89,

p. 3023-3027 Aoki, Y. , Tamano, T. , and Kato S. ,1982, Detailed structure of the Nankai trough from migrated seismic eactions: in Studies in Continental Margin Geology, Watkins, J.S. , and Drake, C.L. , eds. , American Association of Petroleum Geologists, Memoir 34, p. 309-324.

Armentrout, J.M. ,1980, Cenozoic stratigraphy of Coos Bay -and Cape Blanco, southwestern Oregion, field trip no. 9: Geologic Field Trips in western Oregon and southwestern Washington. K.F. Oles et al (eds.), Oregon Department of Geolegy and Mineral It.dustries Bulletin 101, p. 175-216.

Astiz , L. , and Kanamori , H. , 1984 ' An earthquake doublet ir, Ometepec, Guerrero, Mexico: Physics of the Earth and Plans *ary Interiors, v. 34,

p. 24-45 Astiz, L. , and Kanamori, H. ,1986, Interp1. ate coupling and temporal variation of mechanisms of intermediate-depth earthquakes in Chile:

Bulletin of the Seismological Society of America, v. 76, p. 1614-1622.

Astiz, L. , Kanamori, H. , and Eissler, H. ,1987, Source characteristics of earthquakes in the Michoacan seismic gap in Mexico: Bulletin of the Seismological Society of America, v. 77, p. 1326-1346.

Atwater, B.F. ,1987, Evidence for great Holocene earthquakes along the outer coast of Washington State: Science, v. 236, p. 942-944.

Atwater, B.F., Hall, A.G., and Bevis, K.A., 1987, A periodic Holocene recur-rence of widespread, probably coseismic subsidence in southwestern Washington: EOS, v. 68, p. 1468.

Atwater T., 1970, Implications of plate tectonics for the Cenozoic tectonic j

evaluation of western North America: Geological Society of America 1 Bulletin, v. 81, p. 3513-3536. )

Barnard, W.P., 1978, The Washington continental slope: Quaternary tectonics and sedimentation: Marine Geology, v. 27, p.79-114.  !

l Beavan, J., and K.H. Jacob, 1984. Processed strong-motion data from l subduction zones: Alaska: Report prepared by Lamont-Doherty Geological l Observatory of Columbia University, 250 p. I Boggs, S. , Jr. ,1984, Quaternary sedimentation in the Japan arc-trench system: Geological Society of America Bulletin, v. 95, p. 669-685 Boore, D.M. ,1983, Stochastic simulation of high-frequency ground motions based on seismological models of the radiated spectra: Bulletin of the Seismological Society of America, v. 73, no. 6A, p. 1865-1894.

-117-

Boore, D.M., 1986, Short period P- and S-wave radiation from large earth-quakes: Implications for spectral scaling relations: Bulletin of the Seismological Society of America, v. 76, no. 1, p. 43-64, 1 Brace, W.F., and Kohlstedt, D.L., 1980, Limits on lithospheric stress imposed by laborators experiments: Joarnal of Geophysical Research,

v. 85, p. 6248-6252.  ;

Bufaliza, M. A. ,1984, Atenuacion de intensidades sismicas con la distancia en sismos Mexicanos: Tesis Maestro en Ingenieria (Estructuras), UNAM.

I facultad De Ingenieria, 94 p.

Burbach, G.V., and Frohlich, C., 1986, Intermediate and deep seismicity and  !

lateral structure of subducted lithosphere in the Circum-Pacific reg!,'n:

Reviews of Geophysics, v. 24, p. 833-874.

Byerlee, J., 1>78, Friction of rocks: Pageoph., v. 116, p. 615-626.

Byrne, D.E. , Davis, D.M. , and Sykes, L.R. ,1988, Mechanics of the shallow region of subduction zones and the loei and maximum size of thrust earthquakes : Tectonics. I Campbell, K.W., 1981, Near-source attenuation of peak horizontal accelera-tion: Bulletin of the Seismological Society of America, v. 71, no'. 6, Campbell, K.W. ,1987, Stt ong motion attenuation in Utah, in evaluation of 1 l

urban and regional earthquake hazards and risk in in Utah: U.S. j Geological Society Erofessional Paper (in press).

Cande, S.C., and Leslie, R.B., 1986, Late Cenozoic tectonics of the scathern Chile trench: Journal of Geophysical Research, v. 91, p. 471-496.

Carson, B., 1971. Stratigraphy and depositional history of Quaternary sediment in north Cascadian Basin and Juan de Fuca abyssal plain, northeast Pacific Ocean: University of Washington, Ph.D. Dissertation.

Carson, R.J., 1970, Quaternary geology of the south-central Olympic Peninsula, Washin"ton: University of Washington Ph.D. Dissertation.

Carver, G.A., and Burke, R.M., 1987a, Late Pleistocene and Holocene paleo-seismicity of Little Salmon and Mad River thrust systems, N.W.

California - Implications to the seismic potential of the Caccadts subduction zone (abs.): Geological Society of America, Abstracts wich Programs, Phoenix, Arizona, v.19, p. 614.

Carver, G.A., and Burke, R.M., 1987b, Late Holocen paloeseismicity of the southern end of the Cascadia subduction zone: EOS, v. 68, p. 1240.

Chan, M. A. , and Langford, R.P. ,1987, Discrimination of eustatic and climatic events on Eolian deposition, Permian Cutler Group, southeastern Utah (abs.): Geological Society of America, Abstracts with Programs, v. 19,

p. 617

-118-

t Chen, A.T., Frohlich, C., and Latham, G.V., 1982, Seismicity of the forearc marginal wedge (accretionary prism): Journal of Geophysical Research,

v. 87, no. B5, p. 3679-3690.

Clark, E.E., 1953, Indian legends of the Pacific Northwest: . University of California Press, Berkeley, Califorr;ia, 225 p.

Clark, E.E. ,1955, George Gibbs' account of Indian mythology in Oregon and Washington Territories: Oregon Historical Quarterly, v. 56, no. - 4.

Cloud, W.K., and Scott, N.H., 1972, Distribution of intensity: in The Great Alaska Earthquake of 196b. Seismology and Geodesy; National Academy of Sciences, Washington, D.C., p.65-108.

Clowes, R.M., Brandon, M.T., Green, A.G., Yorath, C.J., Brown, A.S.,

Kanasewich, E.R. , and Spencer, C. ,1986. LITH 0 PROBE - Southern Vancouver Island: Cenozoic subduction complex imaged by deep seismic reflections:

LITH 0 PROBE Publication No. 4, Geological Survey of Canada.

Clowes, R.M., Spence, G.D., Ellis, R.M., and Woldron, D.A., 1986, Structure of the lithosphere in a young subduction zone: Results from reflection and refraction stud'es: Peflection Seismology: The Continental Crust, Geodynamics Series Vol. 14, p. 313-321.

Clowes, R.M. , Yorath, C.J. , and Hyndman, R.D. ,1987, Reflection mapping across the convergent margin of western Caaada: Geophysical Journal of the Royal Astronomical Society (in press).

Connard, G., Crouch, R.W., Roy, C., and Kulm, S., 1983, Heat flow: Atlas of ,

the ocean margin drilling program, western Washington - Oregon conti- '

nental margin and adjacent ocean floor: Region V, Joint Oceanographic Institutions, Inc., Marine Science International, Woods Hole, Massachussetts.

Connard, G., Crouch, R.W., Roy, C., and Kulm, S., 1985, Bathymetry, southern J

Mexico to Costa Rica margin, sheet 1: in Middle America Trench Off Western Central America, Atlas 7, Ocean Drilling Program, Regional  ;

Atlas Series, Marine Science International, Woods Hole, Massachussetts. '

Couch, R., Whitsett, R., Huen B., and Briceno-Guarupe L., 1981, Structures of the continental margin of Peru and Chile: Geological Society of America Memoir 154, p. 703-726.

Crosson, R.S. ,1972, Small earthquakes, structure, and tectonics of Puget Sound area: Bulletin of the Seismological Society of America, v. 62,

p. 1133-1171.

1 Crosson, R.S., 1983, Review of seismicity in the Puget Sound Region from i 1970 through 1978: A brief summary: in Yount, J.C., and Crosson, R.S.,  ;

eds., Earthquake Hazards of the Puget Sound Regd.on, U.S. Geological i Survey Open-File Report 83-15, p 6-18.

-119-

1 1

l I

' Crosson, R.S., 1986, Comment of "Geodetic Strain Measurements in Washington"  !

by J.C. Savage, M. Lisowski, and W.H. Prescott: Jour. 1 of Geophysical l

Research, v. 81, no. B7, p. 7555-7557  ;

l i Crosson, R.S. , and Crosson, E.L. ,1986, Freliminary analysis of Juan de Fuca plate seismicity using the Washington regional seismograph network EOS, v. 67, p. 1084.  !

Crosson, R.S. , and Owens, T.J. ,1986, Teleseismic P-waveform modeling for deep structure in a subduction zone environment: EOS, v. 67, p. 1116.

Crossen, R.S. , and Owens, T.J.1987 Slab geometry of the Cascadia subduction ,

zone beneath Washington from earthquake hypocenters and teleseismic '

converted waves: Geophysical Research Letters, v. 14, p. 824-827 Crouse, C.B. Hileman, J.A., Turner, B.E., and G.R. Martin, 1980, Compila-tion, assessment and Expansion of the strong motion data base: U.S.

Nuclear Regulatory Commission NUREG/CR-1660, 125 p.

Crouse, C.B., Vyas, Y.K., and Schell, B.A., 1988, Ground motions from subduc-tion-zone earthquakes: Bulletin of the Seismological Society of America, v. 78, n. 1, pp. 1-25 Darienzo, M. , and Peterson, C. , 1987 Episodic tectonic subsidence recorded in late Holocene salt marshes, northwest Oregion, EOS, v. 68, no. 44, 1

p. 1469 l l

Davis, C.A., 1977. Tectonic evolution of the Pacific Northwest, Precambrian to present: prepared for Washington Public Power Supply System, WNP 1/4, Preliminary Safety Analysis Report, Appendix 2R-C. )

l Davis , E. E. , Yorath , C.J. , Hyndman, R. D. , Currie , R.G. , Bornhold, B.D. , and '

Clowes, R.N. ,1986, The deformationn front of the northern Cascadia subduction zone in plan and cross section: EOS, v. 67, p. 906.

Davis, G.A., Coombs. H.A., Crosson, R.S., Kellecher, J.A., Tillson, D.D.,

Kiel, W.A., 1984, Juan de Fuca/ North American plate convergence: seismic or aseismic subduction? Prepared for Washington Public Power Supply System Nuclear Project No. 3, 40 p.

Day, S.M , 1982a, Three-dimensional simulation of fault dynamics:

rectangular faults with fixed rupture velocity: Bulletin of the l Seismological Society of America, v. 72, no. 3 p. 705-728.

Day, S.M. ,1982b, Three-dimensional simulation of spontaneous rupture propagation: Bulletin of the Seismological Society of America, v. 72, no. 6, p. 1881-1902.

Day, S.M., and J.L. Stevens, 1987, Simulation of ground motion from the 1935 Michoacan, Mexico earthquake (abs.): Eos, v. 68, no. 44, p. 1354

-120-

l Dehlinger, P. , Couch, R.W. , McMannis , D. A. , and Gemperle, M. ,1970, Northeast Pacific structure: in The Sea, Wiley Interscience, New York, v. 4,

p. 133-190.

Dewey, J.F., 1980, Episodicity, sequence and style at convergert plate boundaries: in Strangeway, D.W. , ed. , The Continental Crist and Its Mineral Deposits Geological Association of Canada Specir.1 Paper 20.

l Dickinson, W.R. ,1970, Relation of andesites, granites, and 6erivative sandstones to arc-trench tectonics: Reviews of Geophysics, v. 8, l

p. 813-860.

Dickinson, W.R., 1975, Potash-depth (K-h) relations in continental margin I

( and intra-oceanic magmatic arcs: Geology, v. 3, no. 2., p. 53-56.

Draper, N. , and H. Smith,1981, Applied Regression Analysis, Second Edition, John Wiley and Sons, New York, 709 p.

Duda, S.J., 1963, Strain release in the Circum-Pacific belt, Chile: 1960:

Journal of Geophysical Research, v. 68, p. 5531-5544. i I

Eissler, H.K. , and McNally, K.C. ,1984, Seismicity and tectonics of the Riveria plate and implications for the 1932 Jalisco, Mexico, earthquake:

Journal of Geophysical Research, v. 89, p. 4520-4530.

EMSLAB Group, 1988, The EMSLAB electromagnetic sounding experiment: EOS,

v. 69, no. 7, p. 89-99 Engebretson, D.C. , Cox, A. , and Gordon, R.G. ,1984 Relative motions between oceanic plates of the Pacific Basin: Journal of Geophysical Research,

v. 89, p. 10291-10310.

Fairbridge, R.W., 1976, Shellfish-eating Preceramic Indians in coastal.

Brazil: Science, v. 191, p. 353-359 Fairbridge, R.W., and H111 aire-Marcel, C., 1977, An 3000-yr Paleocll=atic record of the "Double-Hale" 45-yr solar cycle: Nature, v. 268,

p. 413-416.

Farrar, E., and Dixson, J.M., 1980, Miocene ridge impingement and the spawnin; of secondary ridges off Oregon, Washington, and British Columb.a: Tectonophysics, v. 69, p. 321-348.

Fedotov, S. A. ,1965, Regularities of the distribution of strong earthquakes in Kamehatka, the Juril Islands and northeast Japan: Trudy Inst. Fiz.

Zemli., Acod. Nauk. SSSR, v. 36, p. 66-93 Forsyth,1975, Fault plane solutions and tectonics of the South Atlantic and Scotia Sea: Journal of Geophysical Research, v. 80, p.1429 Gallant, A.R., 1975a Nonlinear regression: The American Statistician,

v. 29, no. 2, p. 73-81.

-121-

_. -. . - _ - . ~ . . _ - . - - _ - . _ _ . _ . , _ . . . _ . , . . . . . . _ - . . ~ . ,

=- - -- - _ - ~. . - . - -

Gallant, A.R., 1975b, Testing a subset of the narameters of a nonlinear regression model: Journal 30 the American Statistical Association,

v. 70, no. 352, p. 927-932.

i Gallant, A.R. ,1987, Nonlinear Statistical Models, John Wiley and Sons, New York, 610 p.

Gill, J.B., 1981, Orogenic Andesites and Plate Tectonics: Springer-Verlag, New York, 390 p.

Golder Associates, 1986, WNP-3 geologic support services: coastal terrace study: prepared for Washington Public Power Supply System, Richland, Washington.

Golder A.mociates, 1988, Letter Report, Task 2 - Analysis of Episodic Subsidence: prepared for Washington Public Power Supply System, Richland, Washington, 22 p.

Goodwin, R.W., and Anderson, E.J., 1985, Punctuated aggradational cycles: a general hypothesis of episodic stratigraphic accumulation: Journal of Geology, v. 95, n. 5, p. 515-533 Grant, W.C. , and McLaren D.D. , 1987, Evidence for Holocene subduction earthquakes along the northern Oregon coas : EOS, v. 68, no. 44,

p. 1239 Griggs, G.B., and Kulm, L.D., 1970, Sedimentatiobn in Cascadia deep-sea channel: Geological Society of America Bulletin, v. 81, p. 1361-1384.

Grotzinger, .T.P. ,1986, Cyc11 city and paleoenvironmental dynamics: Rocknest platform. Northwest Canada: Geological Society of America Bulletin,

v. 97, p. 1208-1231.

Habermann, R.E., McCann, W.R., and Perin, B., 1986, Spatial seismicity variations along convergent plate boundaries. Geophysical Journal of the Royal Astronomical Society, v. 85, p. 43-68.

Hammond, P.E. ,1979, A tectonic model for the evolution of the Cascade Range: in Armentrout, J.M. , Cole, M.R. , and Terbest, J., Jr, (eds,),

Cenozoic Paleogeography of the Western U.S., Pacific Coast Paleogeo-graphy Symposium 3, Society of Economic Paleontologists and Mineralo-gists, Pacific Section Los Angeles, California, p. 219-237 Hanks, T.C. , and Kanamori, H. ,1979. A moment-magnitude scale: Journal of Geophysical Research, v. 84, p. 2348-2350.

Hanks, T.C. , and R.K. McGuire,1981, The character of high frequency strong ground motion: Bulletin of the Seismological Society of America,

v. 71, no. 6, p. 2071-2095 1

-122-I

1 Hartzell, S.H , and Heaton, T.H., 1985a, Strong ground motion for large subduction zone earthquakes: Earthquake Notes Eastern Section, Seismological Society of America. v. 55, no. 1 Jrnuary-March.

Hartzell, S.H., 1985b, Teleseismic time functions for large shallow subduc-tion zone earthquakes: Bulletin of the Seismological Society of America, v. 75, p. 965-1004.

Hartzell, S.H. , and Heaton, T.H. ,1988, Failure of self-similarity for large (M, > 8-1/4) earthquakes: Bulletin of the Seismological Society of America, v. 78, n. 2, pp. 478-488. j Havskov, J., Singh, S.K., Nave, E., Dominguez, T., and Rodriguez, M., 1983, Playa Azul, Michoacan, Mexico earthquake of 25 October, 1981 (M, = l I

7 3): Bulletin of the Seismological-Society of America, v. 73, no. 2,

p. 449-457 Hayes, J.B. ,1973, Clay petrology of mudstones, Leg 18, Deep Sea Drilling Project: Ir1 Kulm, L.D., von Huene, R., et al., (eds.), Initial Reports of the Deep Sea Drilling Project, v. 18, U.S. Government Printing Office, Washington, D.C., p. 903-914.

Heaton, T.H., and Hartzell, S.H., 1986, Esti. nation of strong ground motions from hypothetical earthquakes on the Cascadia subduction zone, Pac'ific Northwest: U.S. Geological Survey Open-File Report 86-328, 69 p. ,

Heaton, T.H., and Hartzell, S.H., 1987, Earthquake hazards on the Cascadia l subduction zone: Science, v. 236, p. 162-168.  ;

Heaton, T.H., and Kanamori, H., 1984, Seismic potential associated with  !

subduction in the northwestern United States: Seismological Society of l America Bulletin, v. 74, p. 933-941.

Heaton, T.H., and Snavely, P.D., Jr., 1985, Possible tsunami along the 1 northwestern coast of the United States inferred from Indian traditions: '

Bulletin of the Seismological Society of America, v. 75, no. 5, p. 1455-1460.

l Herron, E.M., 1981, Chile margin near latitude 38'S: Evidence for a genetic '

relationship between continental and marine geologic features or a case of curious coincidence?: Geological Society of America Memoir 154,-

p. 755-769 ,

l Herron, E.M. , Cande, S.C. , and Hall, B.R. ,1981, An active spreading center collides with a subduction zone: a geophysical survey of the Chile i margin triple junction: Geological Society of America Memoir 154,

p. 683-701. '

Hirahara, K. ,1981, Three-dimensior,al seismic structure beneath southwest Japan: the subducting Philippine plates: Tectonophysics, v. 79, p. 144.

-123-l

-- - - . _ . _ ,, _. .- ,m - . , _ _ . . . _ _ _ . , _ . __ , . . , . , , , ~ _ ,

f l~

Houston, H. , and Kanamori, H. ,1986. Source characteristics of the 1985 Michoacan, Mexico earthquake at periods of 1 to 30 seconds: Geophysical Research Letters, v. 13, no. 6, p. 596-600.

Hughes, J.M. , Stolber, R.E. , and Carr, M.J. ,1980, Segmentation of the Cascade volcanic chain: Geology, v. 8, p. 15-17 Hull, A.G. 1987, Buried lowland soils from Willapa Bay, southwest Washington: further evidence for recurrence of large earthquakes during the last 5000 years: EoS, v. 68, no. 44, p. 1469 Isacks, B.L., and Molnar, P., 1971, Distribution of stresses in the descending lithosphere from a global survey of focal mechanism solutions of mantle earthquakes: Reviews of Geophysics, v. 9, p. 103-174.

Iwasaki, T. ,1980, Free field and design motions during earthquakes:

Proceedings of the 7th World Conference on Earthquake Engineering, Istanbul, Turkey, pp. 757-786.

Iwasaki, T. , Katayama, T. , Kawashima, K. , and M, Saeki,1978, Statistical analysis of strong-motion acceleration records obtained in Japan:

Proceedings of the Second International Conference on Microzonation for Safer Construction, Research and Application, November-December.

Jachens, R.C., and Griscom, A., 1983, Three dimensional geometry of the Gorda Plate beneath northern California: Journal of Geophysical Research, v. 88, no. Bil, p. 9375-9392.

Jarrad, R.D., 1986a. Relations among subduction parameters: Reviews of Geophysics, v. 24, p. 217-284.

Jarrad, R.D., 1986b, Terrane motion by strike-slip faulting of foreare slivers: Geology, v. 14, p. 780-783 Joyner, W.B., 1984, A scaling law for the spectra of large earthquakes:

Bulletin of the Seismological Society of America, v. 74, no. 4, p.

1167-1188.

Joyner, K.B. , and D.M. Boore,1981, Peak horizontal acceleration and velocity from strong-motion records including records from the 1979 Imperial Valley, California earthquake- Bulletin of the Seismological Society of America, v. 71, no. 6, p. 2011-2038.

Joyner, W.B., and D.M. Boore, 1982, Prediction of earthquake response spectra: U.S. Geological Survey open File Report 82-977 Kaizuka, S., Matsuda, T., Nagami, M., and Yonckura N., 1973, Quaternary tectonic and recent seismic crustal movements in the Arauco Peninsula and its environs, central Chile: Tokyo Metropolitan University Geographical Reports No. 8, 38 p.

-124-

Kanamori, H. ,1971, Seismological evidence for a lithospheric normal faulting in the Sanriku earthquake of 1933: Physics of the Earth and Planetary Interiors, v. 4, p. 289-300.

Kanamori, H. ,1972, Tectonic implication of the 1944 Tonankai and 1946 Nankaido earthquakes: Physics of the Earth and Planetary Interiors,

v. 5, p. 129-139 Kanamori, H. ,1977, Seismic and aseismic slip along subduction zones and their implications: in Island Arcs, Deep Sea Trenches and Back-arc Busins, Maurice Ewing series, M. Talwani and W.C. Pitman (eds.),

American Geophysical Union, Washington, D.C., p. 163-174.

Kanamori, H., 1983. Global seismicity: in Earthquakes: Observation, Theory and Interpretation, H. Kanamori and E. Boschi (eds.), North Holland, Amsterdam, p. 596-608.

Kanamori, H. , and As tiz , L. , 1985, The 1983 Akita-Oki earthquake (Mw=7.8) and its implications for systematics of subduction earthquakes:

Earthquake Prediction Research, v. 3, p. 305-317 Kanamori, H. , and McNally, K. ,1982, Variable rupture mode of the subduction zone along the Ecuador-Colombia coast: Bulletin of the Seismological Society of America, v. 72, p. 1241-1253 ,

Kappel, E.S. , and Normark, W.R. ,1987, Morphometric variability within the axial zone of the southern Juan de Fuca ridge: Interpretation from Sea Marc II, Sea Marc I, and deep-sea photography: Journal of Geophysical Research, v. 92, p. 11,291-11,302.

Kawashima, K. , Aizawa, K. , and K. Takahashi,1984, Attenuation of peak ground motion and absolute acceleration response spectra: in Proceedings of the Eighth World Conference on Earthquake Engineering, San Francisco,

v. II, p. 257-264.

Keen, C.E., and Hyndman, R.D., 1979, Geophysical review of the continental margins of eastern and western Canada: Canadian Journal of Earth Sciences, v. 16, p. 712-747 I I

Kelleher, J.A., 1972. Rupture zones of large South American earthquakes and some predictions: Journal of Geophysical Research, v. 78,

p. 2087-2103 Kelleher, J.A., Sykes, J.L., and Oliver, J., 1973, Possible cr$teria for predicting earthquake locations and their application to major plate boundaries of the Pacific and Caribbean: Journal of Geophysical Research, v. 78, p. 2547-2585 Kimura, T. ,1974, The ancient continental margin of Japan: in The Geology of Continental Margins, Burk, C.A., and Drake, C.L., eds., Springer-Verlag, New York, 1009 p.

-125-i

.--- - - . , . - . - , , , - , , , - - ,n-,,-, - . , ,- - - - -

Klitgord, K.D., and Mammerickz, 1982, Northeast Pacific rise: magnetic anomaly and bathymetric framework: Journal of Geophysical Research,

v. 87, p. 6725-6750.

Kulm, V., 1983, A study of the western Washington-Oregon margin and adjacent Juan de Fuca plate: report prepared for Washington Public Power Supply System, Richland, Washington.

Kulm, L.D., and Petersca, C., 1984. Western North American continental margin w.' .djacent ocean floor off Oregon and Washington, Atlas 1 Ocean Margin Drilling Program, Regional Atlas Series: Marine Science International, Woods Hole, Massachusetts, 32 sheets. >

Kulm, L.D., Von Huiene, R., et al., 1973, Deap sea drilling project: Initial reports of the Deep Sea Drilling Project, Volume 18. Washington, D.C.

(U.S. Government Printing Office).

Kurtz, R.D. , DeLaurier, J.M. , and Gupta, J.C. ,1986, A magnetotelluric sounding across Vancouver Island detects the subducting Juan de Fuca plate: Nature, v. 321, p. 596.

Laane, J.L. , Chen, W-P. , and Grimison, N.L. ,1986, The Makran carthquake of April 1983: A possible analogue to the Puget Sound event of 1965: EOS,

v. 67, p. 1115 Lajoie, K.R. , Sarna-Wojcicki, A.M. , and Ota, Y. ,1982 Emergent Holocene marine terraces at Ventura and Cape Mendocino, California - indicators of high tectonic uplift rates: Geologic Sc:1ety of America Abstracts with Programs, v. 14, p. 178.

Lisowski, M., and Savage, J.C., 1988, Geodetic strain in northwestern Washington: manuscript in preparation.

Lisowski, M., Savage, J.C., Prescott, W.H., and Dragert, H., 1987, Strain accumulation along the Cascadia subduction zone in western Washington:

EOS, v. 68, no. 44, p. 1240.

Lomnitz, C. ,1970, Major earthquakes and tsunamis in' Chile during the period 1535-1955: Geologische Ruedschau, v. 59, p. 938-960.

Mammerickx, J. ,1985, Bathymetry and topography: ~in Mammerickx, J., ed.,

East Pacific Rise Between 10' and 20 N, Atlas 8B, Ocean Margin Drilling Program, Regional Atlas Series, Marine Science International, Woods l

Hole, Massachussetts, sheet 12. 1 McClain, K.J. , Peper, J.S. , and Holmes, M.C. ,1984, Single channel seismic reflection records of western Washington and Cascadia basin: in Western North America Continental Margin And Adjacent Ocean Floor Off Oregon and Washington, Atlas 1, Ocean Margin Drilling Program, Regional Atlas Series, Kulm, L.D., and others, eds., Marine Science International.

Woods Hole, Massachussetts, sheet 26.  !

-126-

_ , _ . . . _ m _ . _ . _ . , _ , . . - - - ~ . - _ . . _ . , , . _ , - . - _ ,_

McClain, K.J., 1981 A geophysical study of accretionary processes on the Washington margin: Ph.D. Thesis, University of Washington, Seattle, Washington, 141 p.

McGarr, A., 1984, Scaling of ground motion parameters, state of stress, and focal depth: Journal of Ceophysical Research, v. 89, no. B8,

p. 6969-6979 l

McKenzie, D.P., 1969, The relation between fault plane solutions for earthquakes and the direction of the principal stresses: Bulle tin -.of the Seismological Society of America, v. 59, p. 591-601.

McKenzie, D.P., and Julian, B., 1971, Puget Sound, Washington, earthquake and mantle structure beneath the northwestern United States: Geological Society of America Bulletin, v. 82, p. 3519-3524.

Melosh, H.J., 1987 A finite element study of strain accumulation and release in the Pacific Northwest: EOS, v. 68, no. 44, p. 1240.

Melosh, H.J., 1988 A finite element study of strain and uplift in the Pacific Northwest: Holocene subduction in the Pacific Northwest: A symposium, Quaternary Research Center, University of Washington, Seattle, p. 16.

Michaelson, C.A., and Weaver, C.S., 1986 Upper mantle structore from tele-seismic P-wave arrivals in Washington and northern oregon: Journal of Geophysical Research, v. 91, p. 2077-2094.

Minster, J.B., and Jordan, T.H., 1978, Present-day plate motions: Journal l of Geophysical Research, v. 83, p. 5331-5354.

Mogi, K., 1968, Some features of recent seismic activity in and near Japan:

Bulletin of the Earthquake Research Institute, University of Tokyo,

v. 46, p. 1225 1

Moran, J.E., and Lister, C.R.B., 1987. Heat flow across Cascadic basin near  !

47'N, 128*W: Journal of Geophysical Research, v. 92, p. 11,146-11.432. I Mori, A.W., and C.B. Crouse, 1981, Strong motion data from Japanese earth-quakes: World Data Center A for Solid Earth Geophysics, Report SE-29, U.S. Department of Commerce, National oceanic and Atmospheric Administration, Bolder Colorado.

Mori, J. , Jacob, M.H. , and J. Beavan,1984, Characteristics of strong I ground motions from subduction zone earthquakes in Alaska and Japan:

presented at the 1984 Annual Meeting of the Seismological Society of America.  !

Nakamura, K., and Uyeda, S., 1980, Stress gradient in arc-back arc regions and plate subduction, Journal of Geophysical Research, v. 85,

p. 6419-6428.

-127-

Nelson, A.R., 1987, Apparent gradual rise in relative sea level on the south-central Oregon coast during the late Holocene - Implications for.

the great Cascadia earthquake hypothesis: EOS, v. 68m no. 44, p. 1240.

Nishimura, C., Wilson, D.S., and Hey, R.N., 1984 Pole of rotation analysis of present-day Juan de Fuca plate motion: Journal of Geophysical Research, v. 89, no. 10, p. 283-290.

NOAA, 1982, Development and initial application of software for seismic exposure evaluation, report by Woodward-Clyde Consultants, prepared for National Oceanic and Atmospheric Administration, Two volumes, May.

Normark, W.R. , Morton, J.L. , and Ross , S.L. ,1987. Submersible observations along the southern Juan de Fuca ridge: 1985 Alvin program: Journal of Geophysical Research, v. 92, p. 11,283-11,290.

Nur, A. , and Ben-Avraham, Z. ,1981, volcanic gaps and the consumption of aseismic ridgas in South America: Geological Society of America Memoir 154, p. 729-740.

Okano, K. , Kimura, S. , Konomi, T. , and Nakamura, M. ,1980, Focal mechanism in Shikoku, Japan inferred from microearthquake observations: Mem. Fac.

Sci., Kochi University, v. 1, Ser. B., p. 1-13 Owens, T.J. , and Crosson, R.S. , in review, Shallow structure effects on broadband teleseismic P-waveforms: Bulletin of Seismological Society of America.

Owens, T.J. , Crosson, R.S. , and Hendrickson, M. A. ,1988, Constraints on the subduction geometry beneath western Washington from broadband teleseis-mic waveform modelling: Bulletin of the Seismological Society of America, v. 78, no. 3, p. 1319-1334.

Pavlis, T.L. , and Bruhn, R.L. ,1983 Deep-seated flow as a mechanism for the uplif t of broad forearc ridges and its role in the exporure of high P/T metamorphic terranes: Tectonics, v. 2, no. 5, p. 473-497 Pavoni, N., 1966, Tectonic interpretation of the magnetic anomalies southwest of Vancouver Island: Pure and Applied Geophysics, v. 63, p. 172-178.

Peterson, E.T., and Seno, T., 1984, Factors affecting seismic moment release rates in subduction zones: Journal of Geophysical Research, v. 89, no. 10, p. 233-248.

Petroy, D.E. , and Wiens, D. A. ,1986, Historical earthquake location uncer-tainties from resampling and application to the 1932 Jalisco sequence:

EOS, v. 67, no. 44, p. 1115 Plafker, G., 1969, Tectonics of the March 27, 1964 Alaska earthquake: U.S.

Geological Survey Professional Paper 543-1, 74 p.

-128-

Plafker, G., 1972, Alaskan earthquake of 1964 and Chilean earthquake of 1960: Implications for are tectonics: Journal of Geophysical Research, I

v. 77, no. 5, p. 901-925 Plafker, G., and Savage, J.C. ,1970. Mechanism of the Chilean earthquakes of May 21 and May 22, 1960: Geological Society of America Bulletin, i

v. 81, p. 1001-1030.

Rampino, M.R. , and Sanders, J.E. ,1981, Episodic growth of Holocene tidal marshes in the northeastern United States: A possible indicator of eustatic sea-level fluctuations: Geology, v. 9. p. 63-67 Reilinger, Z.E. , and Adams, J. ,1982, Geodetic evidence for active landward tilting of the Oregon and Washington coastal rangee: Geophysical Research Letters, v. 9, p. 401-403 l

Reinhart, M.A., and Bourgeois, J., 1987, Distribution of anomalous sand at i Willapa Bay, Washington: Evidence for large scale landward-directed processes: EOS, v. 68, no. 44, p. 1469 ,

l Repetto, P. , Arango, I. , and H.B. Seed,1980, Influence of site l characteristics on building damage during the October 3,1974 Lima 1 earthquake: University of California, Berkeley Earthquake Engineering l Research Center Report EERC-80/41, 63 p.

l Riddihough, R.P., 1977, A model for recent plate interactions off Canada's I west coast: Canadian Journal of Earth Science, v. 14, p. 384-396.

Riddihough, R.P., 1978, The Juan de Fuca plate: EOS Tecnsactions, American Geophysical Union, v. 59, no. 9, p. 836-842.

Ridd1 hough, R.P. ,1979, Gravity and structure of an active margin - British l Columbia and Washington: Canadian Journal of Earth Sciences, v. 16, l

p. 350-363 l Riddhough, R.P. ,1980, Gorda plate motions from magnetic anomalv analysis:

Earth and Planetary Sciences Letter 51, p. 163-170.

Riddihough, R.P., 1982, Comtemporary vertical movements and tectonics on Canada's west coast: A discussion: Tectonophysics, v. 86, p. 319-341.

i Riddihough, R.P., 1984, Recent movements of the Juan de Fuca plate system:

Journal of Geophysical Research, v. 89, no. B8, p. 6980-6994.

Robinson, R. ,1986, Seismicity, structure, and tectonics of the Wellington rgion, New Zealand: Geophysical Journal of the Royal Astronomical Society, v. 87, p. 379-410.

Rogers, G.C., 1979, Earthquake fault plane solutions near Vancouver Island:

Canadian Journal of Earth Sciences, v. 16, p. 523-531.

-129-

Rogers, G.C., 1985, Variation in Cascade volcanism with margin orientation:

Geology, v. 13, p. 495-498.

Rogers, G.C., 1987, Vertical motions in the Vancouver Island region:

possible evidence for a megathrust earthquake environment: EOS, v. 68, no. 44, p. 1468.

Rogers, G.C. ,1988, An assessment of the megathrust earthquake potential of the Cascadia subduction zone: Canadian Journal of Earth Sciences, in press.

Ruff, L. , and Kanamori, H. ,1980, Seismicity and the subduction process:

Physics of the Earth and Planetary Interiors, v. 23, p. 240-252.

Sacks, I.S. ,1983, The subduction of young lithosphere: Jourr.a1 of Geophysical Research, v. 88, no. B4, p. 3355-3366.

Sadigh, K., 1979, Ground motion characteristics for earthquakes originating in subduction zones and in the western United States: Sixth Pan-American Conference, Lima, Peru.

Sadigh, K., 1983, Considerations in the development of site-specific spectra, Proceedings of Conference XXIII, Site-Specific Effects of Soil and Rock on Ground Motion and the Implications for Earthquake Resistant Design, U.S. Geological S rvey Open-File Report 83-845 Sadigh, K. ,1984, Character'stics of strong motion records and their implications for earthquake-resistant design EERI Seminar No.12, Stanford, California, July 19, 1984, EERI Publication No. 84-06.

Sadigh, K., Egan, J.A., and R.R. Youngs, 1986, Specification of ground motion for seismic design of long period structures (abs.): Earthquake Notes, v. 57, no. 1, p. 13 Sammis, C., Davis. G.A., and Crosson, R.S., 1988, New perspectives on the geometry and mecnanics of the Cascadic subduction zone: Report prepared for Washington Public Power Supply System, Richland, Washington.

Saragoni, R., Crempien, J., and R. Araya, 1982 Caracteristicas experimentales de los movimientos sismicos Sudamericanos:-Universidad de Chile, Revista Del Idiem, v. 21, no. 2, p. 67-88.

Saragoni, R., Fresard, M., and P. Gonzalez, 1985, Analisis de los acelerogramas del terremoto del 3 de Marzo de 1985: universidad de Chile, Departmento de Ingenieria Civil, Publicacion SES 1 4/1985(199).

Savage, J.C., Lisowski, M., and Prescott, W.H., 1981, Geodetic strain measurements in Washington: Journal of Geophysical Research, v. 86, no. B6, p. 4929-4940.

Savage, J.C., Lisowski, M., and Prescott, W.H., 1986, Reply: Journal of Geophysical 3esearch, v. 91, no. B7, p. 7559-7560.

-130-

._ _. .== ... . . . . -- . --.

Sbar, M.L. ,1982, Delinaations and interpretation of seismotectonic domains in western North America: Journal of Geophysical Research, v. 87,

p. 3919-3928.

Sbar, M.L. , '1983, An explanation for contradictory geodetic strain and fault plane solution data in western North America: Geophysical Research Letters, v. 10, no. 3, p. 177-180.

Sheidegger, K.F. ,1984 Thermal evolution of the Juan de Fuca plate: p1 Kulm, C.D. , and others, eds. , Western North America Continental Margin and Adjacent Ocean Floor Off Oregon and Washington, Atlas 1, Ocean Drilling Program, Regional Atlas Series, Marine Science International, Woods Hole, Massachussetts, sheet 8.

Scholl, D.W., Christensen, M.N., von Huene, R., and Marlow..M.S., 1970, Peru-Chile trench sediments and sea-floor spreading: Geological Society of America Bulletin, v. 81, p. 1339-1360.

Scholl, D.W., and Marlow, M.S., 1974, Sedimentary sequence in modern Pacific trenches and the deformed Circum-Pacific eugeosyncline: Society of Economic Paleontology, Minutes of Special Publication 19, p. 193-211.

Scholz, C.H. ,1988, The brittle-ductile transition and the depth of seismic faulting: Geologische Rundschau (in press).

Schweller, W.D., Kulm, L.D., and Prince, R.A., 1981 Tectonics structure, and sedimentary framework of the Peru-Chile trench: Geological Society of America Memoir 154, p. 323-349 Sclater, J., and Francheteau, J., 1970, The implications of terrestrial-heat-flow observations on current tectonic and geochemical models of the crust and upper mantle of the earth: Geophysical Journal, v. 20,

p. 309-542.

S-Cubed, 1988, Ground motion simulations.for thrust earthquakes beneath western Washington: Final report submittal to Washington Public Power '

Supply System. February.

Seed, H.B., and P. Schnabel, 1980, presented in Ground Motions and Soil.

Liquefaction During Earthquakes: Volume 5 in a series titled Engineering Monographs on Earthquake Criteria, Standard Design and Strong Motion Records, by Seed, H.B., and I.M. Idriss (1982): Earthquake Engineering Research Institute.

b Seno, T., 1985, Is northern Honshu a microplate?: Tectonophysics, v. 115, I

p. 177-196. )

Severinghaus, J. , and Atwater, T. ,1987, Age of lithosphere subducted beneath western North America: implications for the Neogene seismic history of the Cascadia slab: EOS, v 68, no. 44, p. 1467

-131-

_. ~ _ . _ . . .

Shennan, I. ,1986a, Flandrian sea-level changes in the Fenland I: the geographical setting and evidence of relative sea-level changes:

Journal of Quaternary Science, v. 1, p. 119-154.

Shennan , I . , 1986b, Flandrian sea-level changes in the Fenland II: tenden-cies of sea-level movement, altitudinal changes, and local and regional factors: Journal of Quaternary Science, v. 1, p. 155-179 Shepard, G.L. , and Moberly, R. ,1981, Coastal structure of the continental ~

margin, northwest Peru and southwest Ecuador: Geological Society of America Memoir 154, p. 351-391.

Shi, Y. , and Wang, C.Y. ,1985, High pore pressure generation in sediments in front of the Barbados ridge complex: Geophysical Research Letters,

v. 12, p. 773-776.

Shor, G.G. , Jr. ,1974, Continental margin of middle America: in Burk, C.A.,

and Drake, C.L. , eds. , The Geology of Continental Margins, Springer-Verlag, New York, p. 599-602.

Sibson, R.H., 1984, Roughness at the base of the seismogenic zone: Contribu-ting factors: Journal of Geophysical Research, v. 89, p. 5791-5799 Shipboard Scientific Party, 1973, Site 174: in Kulm, L.D., von Huene, R.,

and others, eds., Initial Reports, Deep Sea Drilling Project, V. 18 U.S. Government Printing Office, Washington, D.C. , p.97-167 Shipboard Scientific Party, 1981, Site 473, Rivera plate: in Yeats, R.S., ,

Haq, B.U., et al., eds., Initial Reports, Deep Sea Drilling Project, l

V. 63, U.S. Government Printing Office, Washington, D.C., p. 377-412.

1 Singh, S.K. , Astiz, L. , and Havskov, J. ,1981, Seismic gaps and recurrence l periods of large earthquakes along the Mexican subduction zone: A reexamination: Bulletin of the Seismological Society of America,

v. 71, p. 827-843 Singh, S.K., Ponce, L., and Nishenko, S.P., 1985, The great Jalisco, Mexico earthquakes of 1932: Subduction of the Rivera plate: Bulletin of the Seismological Socie. of America, v. 5, p. 1301-1314.

Snavely, P.D., Jr., and Wagner, H.C., 1982. Geologic cross section across the l continental margin of southwestern Washington: U.S. Geological Survey i Open-File Report 82-459, 10 p.

Spence, W. ,1987, Implications of the slowing of subduction at Cascadia:

EOS, v. 68, no. 44, p. 1467 Spence, W. , G.D. , Clowes , R.M. , and Ellis, R.M. ,1985, Seismic structure across the active subduction zone of western Canada: Journal of Geophysical Research, v. 90, no. B8, p. 6754-6772.

-132-

l-Stesky, R.K., Brace, W.F., Riley, D.K., and Robin, P.Y.F., 1974, Friction in ,

faulted rock at high temperature and pressure: Tectonophysics, v. 23, l

p. 177-20%  ;

Stoddard, P.R., 1987, A kinematic model for the evolution of the Gorda plate:

Journal of Geophysical Research, v. 92, p. 11,524-11,532.

Summers, R., and Byerlee, J.D., 1977, A note on the effect of fault gouge co= position on the stability of frictional sliding: International Journal of Rock Mechanical Mining Science, v. 94, p. 155-160.

Swan, J.G. ,1868. The Indians of Cape Flattery at the entrance to the i straight of Juan do Fuca, Washington Territory: Smithsonian l Contributions to Knowledge 220, 108 p. '

Stevens, J.L., and S.M. Day, 1985, The physical basis of m,, M, and vari-able frequency magnitude methods for earthquake / explosions discrimina-tion: Journal of Geophysical Research, v. 90, no. B4, p. 3009-3020.

S.vkes, L.R., Byrne, D.E., and Davis, D,M., 1987, Seismic and aseismic subduction, part 2: aseismic slip at zones of massive sediment supply and nature of great asperities at convergent margins: EOS, v. 68, no. 44, p. 1468.

Taber, J.J. ,1983, Crustal structure and seismicity of the Washington continental margin: Ph.D. Thesis, University of Washington, Seattle, Washington.

Taber, J.J. , and Lewis , B.T. ,1986, Crustal structure or the Washington I continental margin from refraction data: Bulletin of the Seismological Society of America, v. 76, p. 1011-1024.

Taber, J.J., and Smith, S.W., 1985, Seismicity of focal mechanisms associated with the subduction of the Juan de Fuca plate beneath the Olympic Peninsula. Washington: Seismological Society of America Bulletin, v. 75, no. 1, p. 237-249 Tamaki, K., and Honza, E., 1985, Incipient subduction and obduction along the eastern margin of the Japan Sea: Tectenophysics, v. 119, p. 381-406.

Taylor, F.W., Isacks, B.L., Jauannic, C., Bloom, A.L., and Dubois, J., 1980, Coseismic and Quaternary vertical tectonic movements, Santo and Malekula Islands, New Hebrides Island Arc: Journal of Geophysical Research, I

v. 85, no. B1-0, p. 5367-5381. l Thatcher, W., 1984, The earthquake deformational cycle at the Nankai Trough, southwest Japan: Journal of Geophysical Research, v. 89. no. B5,

p. 3087-3101.

Tse, S.T., and Rice, J.R., 1986, Crustal earthquakes instability in relation to the depth variation of frictional slip properties: Journal of Geophysical Research, v. 91, p. 9452-9472.

-133-l

Turcotte, D.L., and Schubert, G., 1982, Geodynamics: John Wiley and Sons, New York.

Ukawa, M., 1982, Lateral stretching of the Philippine sea plate subducting l along the Naukai - Suruga trough: Tectonics, v. 1, no. 6, p. 543-571.

UNAM Seismology Group, 1986. The September 1985 Michoacan earthquakes:

l aftershock distribution and history of rupture: Geophysical Research Letters 13, p. 573-576.

U.S. Navy Oceanographic Office,1969, Bathymetric atlas of the northwestern Pacific Ocean: H.O. Publication No. 1301, Washington, D.C., sheet 1 2306N.

U.S. Navy Oceanographic Office, 1971a, Bathymetric atlas of the northeastern Pacific Ocean: H.O. Publication No. 1303, Washington, D.C., sheet 1104N.

U.S. Navy Oceanographic Office, 1971b, Bathymetric atlas of the northeastern Pacific Ocean: H.O. Publication No. 1303, Washington, D.C., sheet 1309N.

Verplank, E.P. , and Duncan, R. A. ,1987, Temporal variations in plate conver-gence and emption rates in the western Cascades, Oregon: Tectonics,

v. 6. p. 197-209 von Huene, R. , and Kulm, L.D. ,1973 Tectonic summary of leg 18: Initial reports of the Deep Sea Drilling Project 18: U.S. Government Printing Office, Washington, D.C., p. 961-976.

Vyas, Y.K., Crouse. C.B., and Schell, B.A., 1984, Cround motion attenuation equations for Benloff Zone earthquakes offshore Alaska: Earthquake  !

Notes, v. 55 no. 1, p. 17 Wang, C-Y. ,1980, Sediment subduction and frictional sliding in a subduction ,

zone: Geology, v. 8, p. 530-533 Weaver, C.S. , and Baker, G.E. ,1988, Geometry of the Juan de Fuca plate beneath Washington and northern Oregon from seismicity: Bulletin of the Seismological Society of America, v. 78, p. 264-275 Weaver, C.S. , and Michaelson, C. A. ,1983 Segmentation of the Juan de Fuca plate and volcanism in the Cascade Range: (abs.) EOS, Transactions of the American Geophysical Union, v. 64, no. 45 l l

Weaver, C.S. , and Michaelson, C. A. ,1985, Seismicity and volcanism in the l Pacific northwest: Evidence for the segmentation of the Juan de Fuca plate: Geophysical Research Letters, v. 12, p. 215-218.

West, D.O., McCrumb, D.R., and Kiel, W.A., 1987 Geomorphology, convergent margins and earthquakes: Proceedings of the Fifth Symposium on Coastal and Ocean Management. Seattle, Washington, May, v. 4, p. 3320-3331.

-134-l l

i West, D.O. , and McCrumb, D.R. ,1988, Coastline uplif t in Oregon and ' {

Washington and the nature of Cascadia subduction-zone tectonics:

Geology, v. 16, p. 169-172.

Westbrook, G.K. , and Smith, M.J. ,1983, Long decollments and mud volcanoes:

Evidence from the Barbados ridge complex for the role of high-pore-fluid pressure in the development of an accretionary complex: Geology,

v. II, p. 279-283 Wilson, D.S. ,1986 A kinematic model for the Gorda deformation zone as a diffuse southern boundary of the Juan de Fuca plate: Journal of Geophysical Research, v. 91, p. 10,259-10,270.

Woodward-Clyde Consultants, 1984, Juan de Fuca plate comparison: Report prepared for Washington Public Power Supply System, Richland, Washington.

Wyllie, L.A., B. Bolt, M.E. Durkin, J.H. Gates, D. McCormick, P.D. Smith, N. Abrahamson, G. Castro, L. Escalante, R. Luft, R.S. Olsen, and J.

Vallenas,1986, The Chile earthquake of March 3,1985: Earthquake Spectra, v. 2, no. 2, p. 249-512.

Wyss, M., 1979, Estimating maximum expectable magnitude of earthquakes from fault dimensions: Geology, v. 7, no. 7, p. 335-340.

Wyss, M. and R.E. Habermann,1982, Conversion of m, to M, for estimating the recurrence time of large earthquakes: Bulletin of the Seismological Society of America, v. 72, no. 5, P. 1651-1662.

Yamano, M. , Honda, S. , and Uyeda, S. ,1984, Nankai trough: A hot trench?:

Marine Geophysical Research, v. 6, p. 187-203 Yeats, R.S., Haq, B.U., et al., 1981, Deep sea drilling project: Initial reports of the Deep Sea Drilling Project, Volume 63, U.S. Government Printing Office, Washington, D.C. l Yelin, T.S. ,1983, The Seattle earthquake swarms and Puget basin focal-mechanisms and their tectonic implications: M.Sc. Thesis, University of Washington, Seattle, Washington, 96 p.

Yelin, T.S., and Crosson, R.S., 1982, A note in the South Puget Sound basin ,

magnitude 4.6 earthquake of 11 March 1978 and its aftershocks: Bulletin of the Seismological Society of America, v. 72, no. 3, p. 1033-1038.

Youngs, R.R. , C.-Y. , Chang, J. A. Egan,1987, Empirical estimates of near field ground motions for large subduction zone earthquakes (abs.):

Seismological Research Letters, v. 58, no. 1, p. 22.

Zervas C.E., and Crosson, R.S., 1986, Pn observation and interpretation in Washington: Bulletin of the Seismological Society of America, v. 76, no. 2, p. 521-546.

-135-i

- _ . , _ _ , _ . .. _r.~,, . . - _ . ,... -. . . . _ . . . . . . - , . . . , _ .

Zoback M.L., and Zoback, M., 1980, state or stress in the conterminous United States: Journal of Geophysical Research, v. 85, p. 6113-6156.

1 l

I l

1

-136-

I E! .. i l

.- l

.Y 13 0' 125' 120' Queen QUEEN -1:

Chariotte CHARLOTTE 1 Fault ISLAND ';i;

+ Wils on @,.. .

\ Knolls ', , BRITISH

' ' Bro ok s Pen . COLUMBIA

.:u. .p;, g

' S O' \ i ';* 'I's E xplorer RWge

\s,y gy... . . ,

~ e.;. ., .g a

4 sot-Explorer \ e, .

,, . Y.. '.. ' i ,

Plate f 4 ,,Q , , ,

Sovanco F.Z.

/ ,\

.,,t. g

.O , .,(y.' -- ~~

i N oot ka F.Z.' . 7,N ' D' -

, ,. . :-n . . . I'.". . 6 1 S 'u1 \ ': .

W.' ~ .

• 9,\ T

, i.i,.j,i:'-

qS '\ s',: . *

. a

$1 [. SITE WASHINGTON c 810 6 A I .

J

, , , JUAN DE FUCA l ', 's - V 'v; PLATE 6l- I'i 0 Cl r ..s a +45 -

N og 7, O h, O4 nCo N8 y I

. OREGON O?

g .:

. g- x  :-

=-

mll Cape g ..

g ol Blanco o u.1 a

g- [ '

O I1 _

'N { ..T '

6w _ . I, tI l Gorda I o o South I i v' {,

O Plate 6

)

O'

~

$ Mendocino F+.

Me docino '

PACIFIC PLATE '

CALIFORNIA (

'* I, *

\

1;. * , ,

San Andreas Fault -*

San Franciso 130' 125' 5'- 120' I I (' 1 ' I Kilometers Mcdified from Riddihough,1984.

Figure 1. Tectonic Setting of the Cascadia Oubduction Zone

... + . t g,;, ,

. JUAN DE FUCA PLATE fli'

. p .

. 3.d. :

. e f, ,1 * ,, :h-Blanco Fracture , , * .S. QI '

Zone - . .

• s **** s

i. 2. +

+ . + . .e p.-

Cape 66,:

. *. #, B lan c o "lg. .,.

..

• 7f PACIFIC PLATE ..I,.: .. . ,

• O e .-

y 4

5.

., .gh,* . Humboldt N.

. i,&*-

GORDA g

• g Bay Q e

e'4Te. 3*9.y,.,

i

+ .

+ . .y

. .: g.ae .

e.

t 3

- ] fg*

• ' ' ~

l dy ,e** -

I~

Mendocino Fracture . .

Zone . . . . . .

Note: Seismicity (black oots) with size corresponding to magnitude (range 4 to 7),

and local mech anisms. ..

e p

+

f I

I Modified from Stoddard,1987.

s Figure 2. Seismotectonic Map of the Mendocino Triple Junction Area

.m u

28* 12 7 12 6* 125 24

! N$ .

h .[.h.: ,.,, k /' V A N C O U V E Rh, 7 g~. ;f , g L, ISLAND

.; y

--v 935 1-t 4. ~~

,, , t ;,,..ioo - .

f(f'

• • • * .. ,,\ p r 's c.

.%g . b :L,, Q*

... ,' - ~ l  %.

9 .

..w...w ...::.,

f,ri y.-

..t ..

y N -[ 'l

.jj l ^ ^_;..J - pt*&* % 1 9 '~

' *A : A \

.y p p % q-f,

\,{ &. .

V / \j 49 r M[.E ' //

s

/

.iw:w[a, ;.

t i i  ; =c i L/ / f' j . n.;, s u-  % .

~

Q \ ,i Itho* . be'k e

?{

L.' ! C!\,'%._f( _

i _

l w '-

, {'[Qjh,3Q .

4 '}

• ) [)

%q / I so !svJsu Lns \ u '.\,. , +L ,

f' \ l

//\ \  %. '

i ./ \ _, '/

t J" t SiiO L/(g \ s

[% '% f, ,

p

[o .) e.- t E' [ -

, ,.  %'(,,

~<

\/l~0j ', j . ; ' '

\ <

Wi lapa y . ,[

sf . . ..

N&r --

y,j,Q-( ' , ,, g. , s. , '..

y -

O

( Y n\ l y $

f

( T.?~

4

'~"

, . o y {.

~ \.j

,)l( l \ , l,.. ' ' Qi l' r.~

y(h

/ Anobisc beep .:. '

q ,

/

1

}l t

!1 V f *.b S

cisf

\

s N'\ 1 {q/ (fpl 1 i e

} V)

/

s kA, J ; %1 ,

( sc'qh .2 g, ,,

o \ ' "'::..

) ,, y ,

s l -

7 NdW' \> M )1.,f ',

( l/

.. q o ,

N s

7 1 1. &ym' ~w v.y y ,

\

,L, .1 ,

( m ' g' , '

- x\~ ',4 s '

<M, s, $hFkl

?, ,

128* 12 7

• 12 6* 125* 12 4*

Cm ******t

• 200 /****** ***ctrc'**

a,is to a too forman cavose ennt usee openca>e PRRfARED BT SCRi??$ INSTITVT10N01 QCRANOCRA?BT Modilled H.O. Pub.13 from map71.

0 f.19 1390N, Bathymotric Atlas of the Northeastern Pacific Ocean, U.S. Naval Oceanographic of fbe Figure 3. Cascadia Basin Bathymetry

n ,

\* 4 N .? f s No N ,\ o*f o o

+  !*

o x N g. .

o o

o

\ O s NJM b e

_. W

~

Y .&.ME.AT8 3 e . ,

o

)N N NN. EMam . o

- +

ia W"  : .

'E4:4M kV ~ *e N"N;WMW \ o -

~o ,,,

l T

,, \ o o

\/ 8,* M M Ve%_ li -

+~

l 3 . 7.O a cigny > soag

, y

,, \~. v. Tl,' $flibT, E )- J "b' 68? **~~

_ 'n o u ,.

) :

y.o

• oo e

,,,,. . L .' ..M&TY -

Why' / * *, /. * * "o

+ ,

O N/ V, 'i o -

• / @* O O

o g /O' o og O

~

/ og s  %

1 .v,f

• /

', l 'l /*

7 e,

! ,0 0

/ i w s=

l M +

1 u I+/ c / /+ ,

o o O  %

4+ ..

--30 f

40 l {

50 60 f5 h $

wona scu O ~ O = ""

O ~' ' O * ""

" A"** O ~ O " "' w ascr m x em m

-

• 8 A f -, o ~i' o ="n tulill"illickit:;it :::= n==:i,ta ea o g g g g g , J. AiH 29.19151518 h47.6% W1214M ogib S7 km ning lae M W W1.D O~'* O M "70 0 N9 3 o N 18M

. ~,2 o m um

. -si . m an o mi w/t Figure 4a. Map of seismicity in western Washington (1970-1986) showing locations of cross sections. Dashed lines are contours of depth ,

to subducted oceanic Moho from Crosson and Owens (1978).

i I

1

. l GMRST'Y T WA 961RWEXTA DATA (1970-1986) & (C(IP PRE 1970 DMA) poss xcus som m a ustMQ N m A S k 0.ffItIiLL13!If8IIIIIi3fLILLIIIiiIIIIIIIf8II!, IiiiI'.

o dI9f.I8II'III8IIIIIII

~

O-

[ . .O-- k 4~ -

' Un nh $ i

. H d' m,

+

n o  :

Y -

s: i :n

~

u M

!. i 4 i ~ liiiiii '7' s I i i i I 4 i i i i i i i i i i i i i i I i i i i i i i 1 I i i i i i i i i i I i i I i i 4 i i i i i i

' I Y h h h h N N h h h N Nh UKVD' STY & WA ESIPMN*M. . QUA (1970-1986) & (DOP PRE 1970 DATA) ooss rern nom a m ,,

35,. . m um o N P N g . if11 W gtti1IfIIitttIi S.

+

?+ t L t 1 i t 1 i i i 1 i t i 1 I t t t t I t 1 I I I f f2:nI l e i t I f f.M a y * ., -

e

- +=4-> ti i

+[+  :

<= * + k S q f.g + , - f

• 72 Ti

'." . . i iiiiill illi llilli lliiili lli s s t il li t : IIItiillillisili65ilini .

ut.UtsTY OF WA fl5TRUuDUE DATA (1970-1956) & (rt1P PRE 1970 DATA) ocSS KCYDe OfH m a 007E,

,2_1

- , DSfwa M m

< g . 8 d Y d l d 8 d l ffItiititiIf1tI N Nl g

1111ittttttttttilIIttIittIiiF11If11II18t Q

- 9 w 3 , ..

jq

a f,4%gp. .

p n! :n

• ^ '"

i- .... .1: iiiiii, .iiiiiiiiiiiiiiiiiiis iiiiiiiiiiiiiiiT1TTr.iiiiiiiis

• I F i E h 5 5 W E 3 5 E EE Hgure 4b. Hypocenter cross section (see Figure 4a for location).

ucARSN & WA estRIAKNTAL DATA (1970-1966) & (00P PRE 1970 MTA) ,

(noss scios so.w so a M tuSTANCE N a 08T 4 d $ N N R

,4 .

i d

. .$n n n ,

8, n , ,$, ., , , ,!, ?n n$,3,o n .

m ii i,4.$

. n o . .N.i n nNN.

l i. . 5 l ni .

')*.

• h,*i

n 7 * -

i .

. f. ..a.-

g fa sc.h q 3**hd + 3 so a ., -

nq :n 3 4- b r,.

.s. i ; . . . i i a i i i i i i i i e i i,i i i s,i i a i a,s i i i t,i i e i i,i e i i i i i i i i i i i i,i e i a i i i i i i,a i i e i i i

• E "

B P 5 W S U N P 5 P 55 utesn or wA estawonAt. cATA (i97o-1986) & (DEP PRE 1970 DArA) caoss acronom ao a Es7E, tot c,stm .s a c - n$, , , , ,t, , , o!i n , {, n n deg , , ,$, o o o N

, , ,E, , g .N. Na . . Ni n i ,No , o i n n .NN.

* .R. ,(<> +Ng

ee e qes :16 >M -

"f 3 -.

• i -

M &

f+s I -

!a i:

t

+ AT g,k i 9+ b4',M_ -

C

• +4@o

a

'3 Ears & sPECUL DUEM37

., i, na is.ie cair n.t 34 winess 7ii,. ,s u H w I

I -

^

== 1 a7ut 11104 t9 SS h47.130 121050 e m 54 are meg 7.hi!I H %e -

... *j & 684 29.,94S 1928 te47.400 W111400 et 67 trui sieg SJadt H W 4_L -g i;iisisiiiiiiiiiiiiisiiiiiiiiiiiiiiiiiiiisiiiiiiiiiiiiiiiiiiiiiiiii

" E I E S P W 7 5 5 5 5 $ U 55 1

1 us.tyssn v wA r<stawrwrAt prA (1970-1986) & (DGP PRE 1970 DATA) I re+- L DestaNCf M a F8 7E' e f $ $

e? * , , I , , , , f l . f I ,1. , l l i , I t i l l i ,1,1,1 1 1, , l i l l f . L t 1,1,1,3 f l . l l f i i l l I I .

~ -!  :

r.

t; 4 g

+ , * . ~4

-+--+- i t :2k

! 2 l +9 g O -

a..-

+ T k 4'41 :so

[ " :2

@ i I$t  :

,- O 4 -

+ :n

1. A3 1,5,16 EWNTS 4 &.sPECUL PUEREST

• 1- 1 ,. m -

., . 3 - m ,.44..t.w e.n.2. 0317, m M4

.i iu 7,.t.}2 .1213.a.5 6,ti., 71 tres me.g l.a.n.d.

.u-.m. .

o osio o o,o o o oiin o u oii o o o oiiiii o o oiiiiin.iiii

,,H v,t.

". # 8 E p s a N s D E 5 5 5 WW i

Figure 4c. Hypocenter cross section (see Figure 4a for location).

I l

i 12 8, 120*

l I 1 I

N.ANCOUVER

%, , t '

^' l,./

<7 ISL AN D Q.D: ,, . ,

h. ~[ BRITISH COLUMBI A i 49 _ ok 15k.

,p

\ 4..:.

'b i

-f-

-49'

!.{. . ( 'hll:: (

S;t .. \ . fj R. 'if ~ /

Approximate '

. j.l.c. ;t,, l ". A . ,,,, ./ l base of N ' ' ... i "' 5 continental slope N[

i .,

l i.

'?" . . , , v PUGET

, .... b

+ r y g %,. . h'.~~.' +

'. SO UN D g 'i;, OLYM PIC 9 aj .' (

MTS q , , .

.l.y , . >

g /

La f. , [,E j. $

'[;,fl .

~

O  :. ,' - $\

O "

\o W

.g

..:,stTE{%.'-y' + y +h- ~

\c Il, :;]Y /lp' E

/

{

g -

q T

g I

d4 \

Mt. St. Helens

\

g +7. .; +

\

l l ~ pl g q'

g+ W W (f-l l 4"> OREGON ,

30 40'50 60 \

45*~ l i i i (i ( _,3 i

12 6' 120' 0 200 Kilometers NOTE: Contours are depth in k;iometers.

W uified f rom Crosson and Owens 1987.

Figure 5. Contours of the Oceanic Moho Beneath Western Washington

. , . , , , . . . , - - . , - - . - - . , - , , _ , _ . . _ . . , , - - . - , . . , ,,,n.,,,,

4 aER^ ~nJ C 8 1 6 5 4 .

or0 00 0 0 0 0 4 1

1 - - N 0 9 W 2 .

1 i

7 0

. 1 F V 0 2 i

e

. i 0

g r 0 7 i i . .

u r

t i 0 c . .

0 e a l . .

6

. e .

x a

g . .

+ -

AoC g +

z ir o ar e r

a nt . 1 uq hs t

t s i o i e t u5J ".2E n

h us t

ae (1 kc

$2t 0.fE

+ + +

1 eti 0 0 so . w" e 4 o1h 1

' 1( n .

i

~-

t

)i7r9 h . .

s0o ". f d1 u o9g

. . I 4 ..: e S

w8h . .

8 I < . -

n )6 ht i 0% .

• T * .

-d. ,.* ;:

t hae 1 i

+ 2a 1

9E* * -

e ndi s . .

S . I -

r- ,

s .h.

d ip e t

,- i i

. 4 . + ,

4 .Y

}*

pos 1

, i osh mw i .s gk f ti i o -

r,y i

• .. }f.p.

4,,. v ' n ,.:. , -

how t .

^

i r 4 -

i .

r* 1 .

e ni =*

J r

• W t4 -

n <deI" [+ .. *

24. .e Jog .

T-2 0  : =

uf ap h . e -

0

.$i y K 7:)},E}[.' r.

nlap E.

M 1

~ g *f.t. . f

• t4.'a f-

- t do t

oc a g

\, .. - ,

o - .

i Fn i

e a

• t n ,; - ,

s+ - .

u crete ,

t r

af u4,,

p as * [

c o . .

a .e l

t f

$l 1 2

e .

• x "J m.T Ea 1 0 N

x " s' ; t 2

. 0

.@ . S 0m 0

0 4 5

4 9

0 0

. :. E 0 0

x *4 e 0 lill :JI1i lI t

)

12 l l l K8.5 l g C. Chile l

NE Japan 8.5 9" '

10 -

es Mc ma Peru S.o 9. 5 --

o o S. Chile

, , '8 . 3 e,l l c  ;

iengo C. America 9.1 8 -

7.1 Aleutians o Java 7.9 o u \7.2,

• Kerr dec Sumatra \ 92 8.8 C I mbia Q Izu -Bonin

, O 8.0 Alaska E6 U

7.8 Ryukyus o g

• New Zealand 9

• 6 e 7.2 o o Marianas 8.6 (r 4 -e SW. Japan 7.9 New Hebrides Juan de Fuca 7.5 S ,  !

Caribbean 2 --

• 7'9 l

• Scotia j O

l l l l l l k60 14 0 12 0 10 0 80 60 40 20 0 Age m.y.

Note: Relatonship of maximum energy magnitude Mw to convergence rate and age of subducted lithosphere for major subducten 3&s. The contours of Mware the predcted maximum earthquake magnitude against the other two variables. Open and closed circles are subduction zones with and without back-are basins, respectively.

Modified f rom Heaton and Kanamori,1984 Figure 7. Heaton and Kanamori Plot

i1l T 0 S 0 A

}_

E S_ 4

_0 1

_ . Ni 0

- l 0

_ - 3 N

I e _

S A _

B A

I D

A f

i m, l 0 0

2 ve i

C S _

A C

i 5 l i l00

- 1 t _

A - n e i CS mi i 2 )

UI id 6 FX Al e i 0

E A S 1 m

e 0 x i 0 DE t

( k G m 0 e ND 5 i g AI UR e 2 A i n

J s _

a -

, 0 B

i 0 1

i a

d a

c -

e 5 s s

p_i i ,d e t

n e

m I 0 0

2 f C

e o

s _ 0 e S i 1 r -

dnk e -

n c m v _

ai h - 0 5

. o _

yT r 1 0 0

8 4

C t 3 9 en t e 1 5 1 t n

me ym r

e e

g hi g m t d e T

ae id i

d S BS e e E - 0 0 h S

_ _, c

- - - - - - 4 S W 0 2 . 8 m .

E .ESo f o

r 8 d e e r i

f u id g o i M F I

12 0* 13 0' 14 0* 150$

i i i . i i 50*

StBEf tl A . l /

l

/

.i h Q' /

- C HIN A {-$*/ ) ,

A s ia n ,.c c / *p, I

, :/'

P l a t e[. sox'x Aioo'

, g*

JAPAN SEA

. 'c.

<. M;;;. .'

im

, q' '{o' .

M 1983 Ms7.7 Akita-Okl - 4 o.

i sugg ,< k y Earthquake

', 5'

'-:' . ,e n n

.. _ k OREA T e,

Hon ssy ,-

,. . . y 7 o f. Pacific Plate KYUSHU cf g CHINA

+ *%

? /

p '

_ so.

, e m y, 7AlWAN e/ -

&c .

1 Inactive spreading i centers )

{

Philippine Sea Plate .

\ .

- 20' I

\ '. I LUZON 3

s4%, . 3 5

'S . 8 9 '

N *s , , , T'# ,

IO*

Modified from Boggs,1984, Tamaki and Honza,1985, and AAPG 1982.

Figure 9. Tectonic Setting of the Western Pacific Arc-trench System.

M

% 0 C e @ E 5 M m

% a v

.c o a m U M @

C O @ O o en o o bV v

>- LaJ j -- -- - - - - ; Q

' o Juan de Fuca Plate eHcan Mate Brittle Accretionary o o o Ductile wedge o o .. Zone of flexture Interface Oo

, Oo Intraslab Seismicity Figure 10. Schematic cross section of subduction zone in site vicinity

lI\

+

0 IT _ 1 5 1

i 4 . i i

y 1 I

r o

o s e acn i ^ i0 0

l F

1 x a n luts a e f id t .

e g v"1 i

t s -

a ed h na c e

r -

)

s r O cA a f e eg e i e c y f o a gd _ 6 o ai s s n f

r r . > 0

. < 1 uf s oc i n

i (

e no oi de n a g i t t io ec A t a n c latc o _ i i

vu u n _

lc u mo eF a f r i05 D s C af

~ > -

dr a

~ < -

l a ts d n n

a e t m i

Se

" V . 2. r 8

9 d u s 1 n 1

t r

a a e e

- )

g

. b u

sM e

)

(

i h S

c l uw ao

)

(

d n Vl F

_ 0 a

_ _ - _ _ - _ _ nt e

_ 6 , , 0 t aa

$E 57 t

o c

ee r

u MH T

m 1

_ o r 1 f

e

_ d

_ e r

_ fi u -

_ id g

_ o i M F

\

ilIll 1,

l i

x,,

M; : g p a w 7 j

s*j- i

• N^ b x I. \1).';.Q W'? b. (\

l j'% '

' L ..- ?y/

pa r

,i Yq.s.sh % , ',h- !i

('QfK yu l' Y q

,,' a ,4

\ { j

,,'nQ.i%

f. - G

' ' ' s

'o-1/ sa 9 Y c ;e h Q ,cf iyQ s i ,

f' p

/ ' . .'s , ,' ~ .,

<x ., h' '

i

% /

,f' ..

'. , / \ ,( bY' $

-.; *>/ l8 s g' ( .n N *L r N, Q a- ~

' s, f) ,/

/' / i\ g. ,y w .o ,

.,  : ~\ y . .~ . ,r ~ .. .

~~ yj o f .* a ? pz ,/ .a \ q .

',-j - ' <;- f ?

,f b -- i3 #' ' k, /

'"s'  !

gi \ t

$$ / 75k s

'^ <

\ ' ( ~" v M)W, ; \ j

't o M..'N , } p$...A f ;>yV ,. ../ ' . - f, p .L - ~. /

m , ' .:. (

'1 , x t W

\. w' #

C f3 l \;,. , p. ' r-15 1 % $' y .\ z

• ~

' 'Y ./ _ .- / sj .'>

,~

s.

v y

. (' *

h. 'h.f. _'*. .)

= , g.. , s y' . \

N, e

- ~

$/!! ,' ,

.L  % /

\

g <-

..Y

• E A- {\ ll ' ' ,Y' 0UK Gm g;y  ; g,_myp yfy y9 y\fx 7,.

'y _

s. ' -

y - ;, I t ,

m S t . se /; ~

l / , y N. 'If )VO) /

lf- 'j' ~ , (_ lR'

/

y Note: Depths in f athoms Modififd from map 2306N. Bathymetric Atlas of the Northwesten Pacific Ocean.

U.S. Naval Oceanographic office H.O. Pub. No. 1301,1969 Figure 12. Bathymetry of the Nankal Trough i

l 100 km 200 300 T 40,04 i

i t 30 hse4 kp 60 90 A _

120 T i C Shikuku Honshu 3 i , , i QO

, ' " ** M 0 f, ' '.* W,,o m.Olb'5." . . e.

.ns*

. g.ob7.

b

{v{,k{eSll$,2l$ll{!?!'ltlll1 1, p vt g,:1vi , - ,

? , ,

$,,I,!ll if,

, lh,b,,'f,b,,lf;

,t- : np,1:@ ,11 <iz,t, , ,

{,1,1,,~,11,1,

,t-

,1,,

l $'., 'IE, ,' h ,'

i1 $'izl$

qi$  ? it,1 it h;'II',h, i$

l ik p f' h'

,1

'i

'@{?!{@?${l$((?!q?(({ll

- , : , , , , , ? - ,1 - , , :.1, i . ,' ,

'  ?{;l[?lll'{

,l{lll l{ l - '{'?{3

-B ___

~

VA T ,

Q Sil9 r- Puget Trough _ ,

y f A .,

2 od;3..g;fg2,o; LN% Ter

' 0 k ziW,i k; ei' ' i f'

,$ . /. ,

?  : <I i 'e'l'el e'idi l -1 k{'J

'-i t' ', ?t{E

-let t ' ;'i f 'i, ~

'h 0,*8f

• O p'***.

QAo '. . ,M C)* g 7 , ,

f

- s( 10 my sr -

im .

s?{?f?'ff'y'{p,,)[,'[ g'Q,'j  ; 30 g

Q 12 1 . s . ;s s,'p-'s'/ s zi ;< :;<,',s s

-e's' zozi 's ;'t,i s 's s' i,,,, c, ,y;,c,;,b ~ ,s '; -

~ sxtent of D'ORIma te age og 14 ' 'l< 'lz'il;ilz's , 'i t

l' .;

- subduction 'A'

i i

- for past 1 to 6 my subducted 7

'? '

D/ ate  ?

} at 3.5 cm/yr) g IC _

~

Modified from Davis et. al,1984 -

1

- )

1 i

l Not es: -

I A.Comparatrve profiles of the plate interfaces. The geometry of the Juan de Fuca/ North American interface is taken from Fgure 1; that of the Philippine Sea / Eurasian interface is largety from Hirahara,1981.

B.Crustd geology of the Eurason upper plate along a NW SE secton from the Nankai Trough (trench). T, to Honshu Island. Pre-Cenozoic ccm tinentil bthosphere extends to at least the southern coastline (C) of Shikoku Island (Kimura.1974). Cenozoc accreted deposas are shown by th) pattern between T and C. No contemporary volcanic arc is present in Honshu.

C. Crust:1 geology of the North Amercan plate along the tw. of secton illustrated in Fegure 1. The boundary between pre Cenozoic accreted t;rrnne and CSnozoc rocks accreted to North Amerca underlines the Puget Trough. The mapr low ang% fault near the coasthne (C) separates hanging watt Eocene basalts of the Crescent Formaton from Eocene to Pleistocene sediment of the und6ttying accreted wedge.

Loc"ton of this thrust fault is from Snavely and Wagner (1982). Fgures in the lower Juan de Fuca plate indcate that the age of oceanic lithosphere beneath the WNP 3 site is approximately 12 m y, and that the oceanic lithosphere beneath the sete was beneath the offshore trench (now filled) only 6 mhon years ago. VA = volcanc arc.

Figure 13. Comparative Geometry and Geology of the Juan de Fuca/

North American and Philippine Sea /Euraslan Plats interactions

• s s o. L n 0W * *

- 5 9

• z e a

=

op

• s - y c, ~

n* l a B o s** < o e a A

o, +.

g /vl

• 4 y

.n

'Io e

' /u o -

o * ,/ *.

p /*

/* m

• t" 4

O -/  ?./ > *

• f

$ -g / p s y

• 2 g o-e a *

• c 1 ?g,,8m 9 *

/*

/ + s' n so, . v /'m o* '

i * *B' '\* ',

i a /

ms

. o -

-g- /

U

• m

"* a a.

', o w sw o

• w

,m- by sy

a. ,.- so

' u p

T o/ .s

- -E. /m 4

Y l"

r' s

s ~

S. CHILE T Y g

~

SUBDUCTION ZONE T RENCH A IS y /.r g g .. , W k y n. -.,,:gy1: .

~

bOO V

3.a.

$S* . . :.f.

. M9.z.+.. . . *

• o

-e r ;y..c . .. 4-

• ~*

c ...

O

~, . ..

a 3 ...:. ..

Q.

C sg E mE

. SOUTH AMERICAN PLATE e $ ,

O S

a2 ( I 3

N

• • x-o E

a-

% h g o Y $. 3 *

• N @

i g $

2 s* - \

e * . .e g g a * $ e" _R *, E Y I o

o a o o  ?

o *1 9 9 ~. a o

• S.

a -

S =

I. -

o3E3o=

c m 53 g

g E*e*a 3n o I h._ '

eopI u o

a 5

o 3

fsg.e o o

E ea. 1 g

l l

l

. COSTA  :' :l ,

.s >.

' . , . " * ,RICA ,

~l . iPAN AM A, * : . '

'g Magnetic Anomaly

~

. .. : 3:, , .. > ... < . ' : .,.- V q. . ._

N 6V 3

o O Active Spreading ,

N/#d/ 6 l h  % .,6 i Center n $

y#* k ,ls . COLOMBI A

• Holocene Volcano I j .o tr

. g qo I

g /eb $ s.'

y

/

Subduction Zone gg j } /* g g ,,

. , laf I Trench Axis; solid k

j ep & T g s'. , barbs indicate 0 g C' ,, approximate rupture O [ /3 e 5V <ji.' .. .- length of 1906 M s.7 earthquake.

2h 3 $OUTH

) /2 //Q 6

i b 'l COCOS y; f

Vl .' '

.y s

f i

PLATE 2 .,,

,/  ;.c'

• . AMERICAN

- 5 . .

/s s 2 s .

e e e '

,s% j.;: s

/ A .

PLATE s / g]f 2 s * '

i

/ c- f "

sv'/j>,s*...tW

/e 2 >

d ~er,s. r*aa >

m. ..

( Nl NAZCA 7[ < >

. .c.

. :. i PLATE jI '.'"J ' l 6 ,.

o T PERU

/

CC <.

ut i W. . /, . -' . . . . .

y c.{ '

• yj; '.

Q:::. 'I -

6

/ ,e'..

,,./, g- i g ',

\ \ 80'

\ \  ;~ :

, \

N OT E:

Magnetic anomaly designations do not equal Modified from A APG.198 2, the age of the anomaly.

Figure 15. Configuration of the Peru-Colombia Subduction Zone

l l

l l l

'df4Xi. .

I I I I Bajk$- I l ^ ;,, :.L ' i N - U '-

- 2,,

a:,, .&  : 2 '! : _

';Cali a#!! . M EXICO

'::. : f.' '.~i_

$ .n:. 1 m..

~..

i

'e .

v d.. w,M ,

e)y.!

s

r. .

' e,.  ;. )

Peg +> ...

. 1 Inferred ,

-22' No4#of Rivera-Cocos Plate Boundary -

East Pacific 3/

Rise { *en ' f. - )

_ \ o* <

c -

s, o.

4 "

NORTH i

HlVERA ", 9.jj,' AMERICAN PLATE {

- 20 r ".- PLATE -

i Or 4 0,,0

,h:,k, ,

it ' v! .. *

-.- 4oft/ry 9 '..w, . \

i 2ot;e p,'n

~

94 ,?

'Qi*

- 18 *  ?,%,4 -

)

l PACIFIC I PLATE i 1

tomm yr -1 i

COCOS {

1-+1 East Pacific PLATE '

Rise  %  ;

- 16 '

l W 110* 10 8' 10 6' 4' 0 200 Kilometers Modified from Elssler and McNally,1984 Figure 16. Configuration of the Rivera Plate

,, po* 10 9' 108* 107* 106* 10 3* 304* 10.,* 102* lo t*

Ig, Q Y KW

$ e.D D

-* Q*7 = fr f...e,.,,4 k

';.5

,=c - ~$

\ , / / . f'- ,?

\ . i 7,=='ir 22* -

c

- + - M..

. f,.r w/

e?- r 1

-h t

-) /F. 1 I 3

-- y t  !

7 '/,#  ! & Y- a

- - - -- { 22*

e J%~~fy. DSGP Drillhole #473

+

g, ' f- /.'

9 / *j/ tw/(W f,y ,1)4 y Q>dx

. p 5 --+l!-- % - h '

i k

/h'7 1 :

H Y[Op.k=/M  !' _ f P A}. ' --i  ! l l' g/J.},M/p,p[u$/j'[E.-;Q

\

, l 2;. .-:l.2 L '

.L . [. ' f '

g O'I ) Wj '/.

.-- 1

h fk",2iEs%-,,lQ'~gqi~i I

\

- - 2 n.

bj

'* p ffy 4,  : f y.$~ %. ;

')

sy" lo 1

jV~

Possible Epicenters of 1932 Jalisco Earthquake

,.: 3 -i . N. pc %A i iL - ('

20* . d.1h h -4 -- ' --- A- - - -

kl 1Kelleher- --fet--al. (1973) 20*

IMs g-- L__. "f,4C _

o D oC S,,. (/D.k, '- ,

I I-nhg 1hv-Eid f 0 ca & , I Elsslerf and McNally (1984)

,;;- ( 'm,.y y(m 3 -

1 l

! j c

m#teg ~ =,

^~

~Q,i,

,  % __ j ,

i9.p-p')e. > 4d ., %_ . Ogj" ;_' -_s f-t

,s  ; i3 Approxirnate Rivera-Cocos n &b 4 S X], A j;,.ff j i }~ i 2#(> II#Sfe s mf M'i"Q*'4-_~

,4 S Plate Boundary

, l l

[M} -- W O

,  % f N _ ~- T b.,,'cN ya { { l L

9 l [J 7to,,-

4 p.,, .3%

hv d J i *'

vJ A-Nw w%'g 'll l i

-ae i

,e.t _ c- u r s' %g I "0 Qt' x L 3 4

0 pj g ?3d/$pl 2_}h)i q3  :

WMk c h g^

l 3 7. m m. n An ON/A daA '

_ O *- Z '

N._ ,

$d ** v7 18 0

• 109* 108* 107* 106* 105* f04* 303* 102* IOt*

~~~~~

10 0*

__ *---- No.1104 N Figure 17. Bathymetric Map of Rivera and Northern Cocos Plates

80 ' 5

f" " .o, N AZC A h$ l.h PLATE l $ ;-

'l

> ut 5

, :.h G i d0-..s . :C x t

>"'9 .a}i.' '

p-  : .!..

~

$\0 N.](g) 4 SOUTH AMERICAN i[. g J,).. PL ATE c, '

CU 6 '

sc :1 1'f,.-l < ;i

$ */ ..,;?

V 1. . .:u

'*~

6  % (7,'  ;, -

\ r -

'go. g\ 6A >n'g -

. +

BB ANTARCTICA sc \ ' E O . (.:

Y y

PLATE

} I 2cm/y 1 . g,-

6 .~^

. Faulkland

' .Q. f*' ' Y 6  % s 'N

! Usy a a" by

'3 X*

'8 Nx

\

y. 9 re e f *,-

60' fC 80* ,,

Af 1

6 - Magnetic Anomaly O

c/0'80* OM Active spreading center o

• Holocene Volcano Modified from A APG,198 2.

Figure 18. Tectonic Setting of Tierra Del Fuego Subduction Zone '

1 l

l l

I l

l l

l I *... i '. - 3 76* ,,., 72' l

..,f _

s Maln shock (M 8.4)

\* ,

.. 6

/

>v g 6 Aftershock ('A> 5.5) '

o ..- ..- .-. Uplift 3 . f- .c . .

• o , (. ..< ),-;

[ open pattern where inferred

..... ..n  %- . --,

, Subsidence i

....__; QQ ^ approximate axis shown 0,", IRF ; o* e o

......***%. Zero isobase

. . .n n;

~

o e

• e

-40

. .d y_,,, m 40._ *@ Horizontal Extension numeral indicates tensile strain

. " . '&OW

. . . ::'l' '. 4' - '.' ** (x10-5) y ..'..

• Volcano o o w

= .... .c M m;. * .

. . .. 1_ w .

o. , =

s s= .. B. m: - ,

m .- .

o w * . . a; . .

. . w s.e. ,y d ..- =

.i.';: o a

x ..

: :9,5% . -i . 2

? . . =+ . .: 6.n 3

o'".,' ._ .Or W

= . * .

h '

?* g g c

• z A

Ap B'8m ~e

. ' ', u-f J s v u * "-

7

. n 0%4'

• o, Isla Gcamblin -6 *

- gE Isla Guafo

~44' ...>g,'

. ,' . ,' g' . l

" .* ,*. rO-l . ' T 4.f., s' 4 3

4 4 *-

~

isla (pun ^

i_"

14,;

1  : - q 0 A ' ..' " ' . ' W ~'~W, ~

o." . o. t..l t ' ' .*. s, ,

Profilo showing vertical

'e o o . h. .' , '. ) . ,1 displacement along A- A', B-B'

...r..f.J.: --+-H .* - -

" ' '[ 9

-[

b:=!

.E.C'

• O 10 0

., - I

\

/, r?'# Mlles 1 7 6*( M'y% 'f W :.',,.O  ;

• O N

10 0 l I. .0 . '. Kilometers I

l Modified from Plafker,1972.

Figure 19. Coseismic Elevation Changes in Southern Chile f rom the 1960 Earthquake

Patton Bay Fault Fault?

10.0 -

~ I I) f' If O

.. ,'l E

I .

o

! ~

' . *I' b "h5.0-

,g , ' */

B i;

O

~

,y-M">-

... .3 -__6 --Q h

~ ' ',

- D-p-

f _

m s 0'0 _

k, kg u "

...............--: -o w

,a = -- _ -- _ _ .__ _a

- e.. * " " " ..... g ' -o - W C '.9 Cascadia S.Z.

~

~

Coastline

-5.0 , ,

I I l s i i I ' 8 0 100 200 l 300 400 500 Distance

• rom Trench Axis, km Profile data scaled from Plafker,1969; Plate -

Figure 20a. Sclected Profiles of Coseissaic Deformation -Alaska 1964 Mw 9.2 Earthquake I

5. 0-E -

q d -

2 W -

E _

• ,,, h h 2 A\ - - - -

8- -s% 14 W ~

~*N-n%..s.cfW

~

2 E -

ex>.o -

3O -

Cascadia S.Z.

o Coastline

- 5.0 - , , , . . , , , , .

g  ; g 0 100 200 300 Distance from Trench Axis, km Profile data scaled from Plafker and Savage,1973; Figure 3.

Figure 20b. Selected Profiles of Coselsn.ic Deformation-Southern Chile 1960 Mw 9.5 Earthquake

l l

l l 1 6 . 0- A SC ,

o o i S

4.0 - o oo E NH N g o oSC o

.o 2. 0-a N j

1 So o i C

O E

g 0.0 ___g_._________ _ _ _ _ __ ______ ___

} K g N

0 N

O

- 2.0 - C O O A SC l

- 4. 0 -

I I I I l 3 I I I l 1 I I I [ l f l l l 4 I 4 l l 7.0 8.0 9.0 Earthquake Magnitude M-Makran S-Sagaml See Figure 20c for other abbreviations Taken from West and McCrumb,1988.

Figure 20c. Earthquake Magnitude (type undifferentiated) of Shallow Subduction Zone Events vs. Amount of Coselsmic Coastline Deformation

10.0 -

E _

~

~

g 5 5.0 - . . * * * . . . . . . . . .. . . . . . ,

g .

o

= =. . r . . . . . . . . ... ..*' ., , . ,

O -

t 1 1 .

5 _

i.l , , ,

"I ,1. < >

-[ . t.

=

rI m

g ,, ,

'*y: s,As. g.I..,.

i_

h,,,db W, ,@'8 o  %

E k , ,o _ , ,_g , , , mi , ,,,,0 y ] c Oregon r e

Washington

- 5.0 , , , g , , g , , , ,

0 10 0 200 300 Trench / Coastline Distance, km i

Data Sources Subduction Zone Earthquake Subduction Zone Earthquake Nankal 194 6 M 6.2 New Hebrides 1971 (?)

1854 M 8.4 196 5 Ms 7.75 07 Mw 8.6 S. Chile 1960 Mw 9.5 Sagaml 192 3 M 8 + 183 7 M 8 +

) 17 0 3 M 8 + 1835 ML 8-8.25 Alaska 1964 Mw 9.2 Colombia 1979 Ms 7.9 Kurlie 1973 M 7.4 Makran 1945 M 8.3 Mexico 198 5 Ms 7.5. 8.1 c

Figure 20d. Amount of Coastline Coseismic Deformation vs. Distance to the Trench

10.0 -

8 E -

a ..........

3 5.e - .. .

O ..' .,

s 5

,, ..- ~

L j

o g*,". .

[*,.,,..'

o -

i-

. j, u'[{i*

gj.il ':. .' t ' c-c~

.g0.0 m

- g "O 9 o@d [ ~ ""' ' ~ ~ w '

o O o

)O -

Washington e r

~ ~ ~

Oregon _ '

- 5.0 i i , , i i i i i y . i , i i ,ig i , , i g i , i i g i i i i , , i i i g 0 10 20 30 40 50 60 70 80 Depth to Interface, km Data Sources Subduction Zone Earthquake SuDduction Zone Earthquake Alaska 1964 Mw 9.2 New Hebrides 1971 (?)

S. Chile If,60 Mw 9.5 1965 Ms 7.75 18 3 7 M 8 + Kurlie 1973 M 7.4 183 5 M 8-8.2 5 Makran 1945 M 8.3 Colombia i979 Ms 7 9 Mexico 19 8 5 M s 7.5. 8.1 Nankai 1946 M 6.2 Sagami l'- M8+

18 54 M 8.4 17 05 M 8 +

1707 Mw 8.6 Figure 20e. Amount of Coastline Coseismic Deformation vs. Depth to Subduction Zone Interf ace ,

I i

i t

Cape Flattery

(( , Neah Bay y N

/.Wa,

.. , %x .-

. y STRAIT OF JUAN DE FUCA I .. '.;L ,

Y Ml

\ ** ,

' I'['U. '.'J .* I,

? ; ...;... . - O Atwater's Study Locations Qulllay

} .

'- **'~;.* d A. !. . " ', ' '.I '-

,. A -Me

'0

.; 'l, ..' ... T-

. s. , .

' f.

f

• 1 :., d,

- Wa - Wantch River s .* ' ' ' ' ," ,; $ Co - Copalls River 1.a Push * . ./

R

"" .l .

'; .2 CC - Chenols Creek

\.,.

\*

d4 O L,Y M P1 C Bl CR - Blue Slough Chehalls River Jo - Johns River j ' e 'l, , ',

f. f WI, Wi-s - Willapa River

( S. Fk. Willapa Kalaloch O[,st. . MOUNTAINS e River Ni, Bc - Niawlakum River i [, . .

Bay Center k $'- No - North Oysterville Cape EI/zabeth i', '

? Ta - Tarlstt Slough

\,\-.'

\ .',. ' . *,

Pt. Grenville ~ ;. ' B e - 8 e Ar Rive r De, Gr - Deep River T *. ,. ' ,

Grays River OCo

.. CC r

, ,. Aberdeen i

(

Grays Harbor

;.t P "$S

• 1

[v;.;so{:".BI,CR Site %

i: .-

Cape Shoalsster li  ::sf. .:;

• w;* wi_,

.4 i Ni, Bc Willapa No Bap\y Bay Center * . C H E Hl ALlS )

1 Long Beach Y l Paninsula O, n .... LOWLANDS Il Ta' y. . ., ... -

l i t Be

' p*. . '. .

De, Gr i

.V ' O.

., .C  ;

S '

O( 0 to 20 30 I g... . .*

BlA Miles

\ -l'Astoria .E'a' .

i h. . ..s ,,.t.. 0 10 20 40

~

.a.-

• ' ' .'* ' . ' g". e '.15 Kilomete rs

\

Figure 21a, Location Map - Washington

l l

l 1 m I

g q/p v Ri 9 4:l. '.r. . ' . . . .

s

\. . , .V. .h. $$ ' OO h '

3As torl,a '.A /'Oh _.

.,5.; p y.  ;,gt. , g -

Q t:: - . .

,?

y.;.':. ',

7.. .

. " . ' ' . '-j p .- ,'

.f. Cannon Beach .

h a.

i 2 y".6. .l,:. . .

.ft m '. '

i' c. ::: , , . .

,2,..<

' , ,i. d~D . , , -

Nehalern Bay - * .. ,l,l,,

>.. . ~ . * * .

• NetartS Bay .n,g'gg]g. ,, .

ll .w. :: ,, nook PO RTL AND i[6 Cape L ookout t.p..  :. . . ; . ..:

J .*.p~. .... , .'.,' '. '.

Tlerra del Mar e Neskowin tu

.e

, y.,.'. - .

s Lincoln City e

2:

Otter Yaquina Head Rock g . [.,, \]gsf '

g 9 *. . . .

Yaquina Bay

• N**9*'t 6, ' ,:

• s .. 8

...i.. ... m e san Marine -

* '. 4 I .,.8Yachats O

Roosevelt Beach ?).c . ;

Heceta Head :.r ':. ' ' ' ' o Sea Lion Pt. **3 '.,.j. ..

Florence

'?.

No+

j. ; , . . -)

c

'e Re edsport '

\

tu

.:0*..

e

/ . ;. '. ..'

• O

(

\

Sunset Ba Cape Arago,Y  ; coos e . , *: . Bay  ;

i l'l :f . . . .C. harleston CoqulIIe Pt. :' . . 5*

\

e Bandon W. .

,J.'-

0 it) 20 40

,..'O'

, I Miles

'Ca o 30

\ . .pe.z.,Blanco 15 60 xitometers Figure 21b. Location Map - Oregon

Il' '

h s

u '\

r

,0 0 0 P 6 0

6 h a c , .

l o L . e o a Q

l l i n

l i a . .

l t

v K s n u ' a c

r o

G 0

0 0 0

C 5 5

. n t

P

. . o g

. . e r

J .

O m m -

k k n o

o t

o

.0 0 c 0 c g 4 n 0 n n l

a 4 ia i

. B B h e

e s _

. p p a _

a .

a W _

C C -

. . g m m n o

r .

o r o f f l

. 0 0 e 0 e A 3 in 0 3

n i

l l t t t f

. s . s i a a l p

i

o o c .

C .

C U o g g f R .

n .

n o r '% lo A

lo A e e s . .

t t e e a J

t 0 c 8 0 0 c R O 2 n e a t 0

2 n 8 9

. a 1

)

t s a .

t s .

a r iD R iD e m b k

c u hl_fI ^

t f

i l

t n r a Cc 8

(

2

^ p .

s M U i d y

. D an r y a o

0 r 0 e t n

o g 0

1 a 0 n se r c a n 1

i l W e r t t A~

n r e s 7 a c A V t a 8 9 u a o i .

C e o u

1 Q

o pa C .

o Q .

/ la t

p 1 h t e a c C . e . c t L

C t

a n s e

L _

e W

. _ - - - _ - _ ~ - _ _ T r .

0 0

) 0 o T m 2 1

a 4 1 ) o r

2 5 h 6%oc (  :

. 6E$5ro!r.

r_ b

(

f d e r

fie u id g o i M F

_ o 50 -

o o

40 -

~

E g 30-e -

O

[T] o o 20 -

[.l] O O E o

~

O O

_ .@ oE O 0 o X 10 -

~

Og --

[x 0 o

q-a O .

O OO O 00 0-I i l 6 3 I I I l [

0 60 100 14 0 200 Olstance from Trench, km Subduction Zone O Alaska * (1.0-7.6 5 kai O Makran* (23-8.0 kai O S. Chile * (1.99-3.93 ka) O Sagami* (e.0 kai X Nankal* (6.0 ka) O New Hebrides * (2.53-7.0 ka)

O Peru (3.C ka) O New Zealand (1.8-8.0 kal Asterisks indicate where coselsmic terrace is lowest in sequence.

Dates in parantheses indicate age of Holocene f eatures, Taken from West and McCrumb,1986.

Figure 23. Height of Holocene Terraces, Along Coa:stilnes Adjacent to Seismically Activa Subduction Zones, vs. Distance trom Trench

4.00-o l

l

- l I

3.00-

)

_ o o y _

o l

o o o i 2.00-E -

E "

S -

O o

_ o

,)

o o g o

_ ,, o o

'f No 1.00- ,,

_ o o o ll **o gi

_ o o

o . 0 0 "o

oe a ye. O o Ci e

• U

'5'. o$ ,,

,, I c669 g

• e 0.00 ' '

0 8

60 3

100 i i e i l i i i i q 14 0 200 300 Distance from.Tmnch, km Derived from raised Pielstocene terrace along coastlines adjacent to active subduction zones (southern Chile, twru, Nankal Sagami, Makrsn, New Hebrides, Java; open circles - Cascadia).

Taken f rom West and McC.%..b.1988.

Figure 24. Rate of Uplif t vs. Distance from Tranch l

I i

l l

So Notes: Dashed line west of coast is base of continental slope, Ses text for discussion of i

tr?,,

\,

((

s T

6

/G s

d~ h Md%

fineaments as follows: i f

A (Barnard,197 8); B (Kulm,1083); l

( *N' / h C (Hughes and others,1980); \ 't C M D (Wzaver and Michaelson,1985); \

E (Weaver and Smith,1983); ',

'f(

s., l[b' g?

F cnd G. zones of *of fset' in ',

/ N i offshore magnetic anamolles (Pavoni,1966; Atwater.1970).

L / 7' I

~*,M '

GS

'/

~ ,A NGOE arrow indicates relative { >

orientation of inferred North -\ .

' 7 American - Juan de Fuca plate i, /

convergence, Cascade volcanocs s

/

/-

are from north to south: ( V G = Giribaldi; B = Baker; 's /

GP = Glacier Peak; R - Rainiti; SH = St. Helens; A = Adams;

( so. - /

/

l ,O' / ,/ 4, H = Hood; J = Jef ferson; ', E A ,f

,, y TS - Three Sisters: 4T (

N - Newberry; CL = Crater Lake (l ', pD (Martma); M - McLoughlin; +S

,/ ',

/ j /

S = Shasta; L = Lassen. / ,3g ,{

/ 'n / ,

/

p l V

/k

,b \( /

' /

/ i 14 O A*

,/ J / i /r#0f tlan

/ / l /*

l\' h/

/ B D l

G

/ l'/ / N' j  ; /

8/

p -

% i[ /6 /

j/

/ /A ct

' /

s / / Au M

( /

/ l

/

\

34 t

/

I

- ~._ _,;

k / at 129 x _ __ ao 121 Modified from Cavis et al.,1984.

Figure 25. Map of the Pacific Northwest Showing Lineaments Postulated by Various Workers 1

I h .. I t i 13 0* 4'. - 125* 120' Queen L' QUEEN '.: i l

Charlotte i CHARLOTTE 7;J.

Fault ISLAND l

• - Wils on M,. ,

\ Knolls '

BRITISH i N .:

. Bro ok s Pen .: .. COLUMBIA

• ,W

);  ;?h

• y'.

i ~ s o. \ r ~;, . . jyc s

Explorer g -'N gjy g + . , ,.. ...

a

. u :;g ,. , . 3 s, 4 50t_

RWge Explorer b;., . . i

/,S .,j,!;. ,%j %7? C ,;a';.G '

^'

Plate -

u ,~. . f. . ..

i .C,.

Sovanco F.Z. g ~

' .A.g ...

he N oct ka F. Z. g h:by.}l'8 ,

\ }MD."..

f \g.gf D: 'y*5

/ k [' '

,i.;i.- a.

[e ,hl I

4. A ,g..

WASHINGTON e in

"' . gpg u ozt

' 4 5' J U AN DE F UCA -t IM ~3

[ O PLATE

% T i 4 5'-

l- .:

a <

to m l

~

?' .- . -

- , o a

/ o l . ... - A o S"Oc o g E m I w . OREGON

" ~

.c' zZl l .f '

e i Cape a Blanco..f lif uj a

xb" l 0 s I.4?

Q- p s .,

. o ,

p Gorda sI '

A l o South I i.: .

O Plate W '

1

\ Cape

~ 40* b

$ Mendocino F. Mewocino

_' 4 0* .

', e PACIFIC PLATE

. CALIF ORNI A (

I. 8

\

, n. . :. , ,

l San Andreas Fault-"

San Franciso 130'

. 3 200 125' 'yA.'.. 120' l I ' h 1 Kilemeters Modified f rom Riddihough,1984.

Figure 26. Juan de Fuca Plate Segments

i I

l

. . 8

' i 'OO O -

I

~

+

gg O O

_ j

• O O O -

k

% o0O O T O 1 0 0 g

O 0 0 O 00 O Oo _ 8T

-5 l

9 .

1 O

$ 00 O 9O O O O

o h0 0 O

O y

gj .

! 4 g OO O

~ ~ ~

~

0+ r O

o O ~ "

o Ln

, t , i , I i I '

9

' . . 8

~

l O O I

b o' .

D O@ .

)

o O o \

• O O

@ i o+

• O O j

+JO

% O g 8  % 30 gg l

SO+ O O O #

g I O O O + g g .

V . e*

$i '. f

• o ,890

s o O O + ;e -

8 .I, l

_ g O , o C '

+ ~

, o 0+

+

• 3 o l s . 8~

E l

+

, t , I , I - I '

9 e o o o o *

(MW) opnpu6on Figure 27. Distribution of strong motion data used in analysis

0 _

__ ~

=_ : _ O_ c:~ - yN - ,:~- ~

c~- '

0 5 _

O ' .

O , gO , O ,,h,$ ' 0

• 8 0 _

o *

' 2 ) _

' , m -

s 4 0 o S.'++

, k g

O

O O e ' 0 ( -

O s '

. p '

1 e

c O '

n .

O O @O 0 7

6k

' .O O

2 8k

' 0 a 5 t

. O 9 - co ao O + -

- xae '

is D

. 3. R s ' ,

8R s 7

6

+ 0 -

w

' , *

• w 2

M 4 O . M 4 O -

_ : - -  : :_ _ r: _ _ ! - - -  ::~- -

- : _ o t -

0 -

,::~- 0

= :_ _ c : ,-

0

_ E - -

C -Q - Q_ ' 5 0 '

,4 O + 0 O ' .

' ' 0 j 8O s

1. 2 )

m

% 4 @ 0 6 OO 0 k(

0 g

i W'o g,Qe '

" 1 e

g) G*0Oh

. e .

o 2 O * ' 7 c

n -

0

• 6k

' .

• 7 w

' 0 5

at 4e - ocao ' i O - xao '

is

. 8. R s ' ,

3. R s '

D 5 ' . 7 0 _

i w 6 O 0 w 6 O 2 -

M M _

_ : .- - :_ _ r: _E - -  :::~- ~ -  : _ 0 1

0

= :_ _

7:  : -

,::~-  : .- 0

. ' .E ~

' 5 _

O g 0e '

0 '

0 0

• 2 )

o O* , O m g

y( ' g q

) P Q ' " 01 0 (

k _

e i

s i

ps '

i

. so . O' '

e c

i

6) s 7 ' i 2 ' n _

5k OO O 7 k

' 0 5

a t s

O - co ao .

, - x i,o iD

. 3. R S ' . 8R s '

5 6 0

w w 2 M*O M*O

, __::- _ E - -  :: . - r:_ _ 0 1

I 5 2 1 5 2 1 5 2 1 5 2 i 5 2 1 5 2 -

0 0 0 0 0 '

0 0 0 0 0 0 0 0 0 g 8 o e*

• 8 < xoe

. 3 8i3* o< go E, P memx 70 NOSE mS*g'E Oo, e o g= E e1o 3 mms @ e $-

o

~O c e93xRcm e" u)$Es E  ! l c, 3* O 3 @ 3 o 3 e $ 2. Ou 3.m_

o0go03e3G.

l i '

l '

l 3 '

i '

i '

i I '

c - Rock e -

- Soit -

Q o hierface 9 o

- o interfoce $ o PO

. Intrastoo o [9 o -

2 0 Intraslob o -

o o, e o' o '

'O -

y 1 - o go I o

6 o

=z 3 O e

o

. e. ee, ~*

o

+ o, o e" g a a

,o . O o -

o oo P. o o ^- + - O 0g a3 ^* ~A ^^' '

? 0 O

-m u 69 9' o "6 .$ '

• o~ + +Y o- " "

hg j -

o , o to ooo ,g No og @

U O 'O *

-1 -

goo, o

-1 -

o 6 0 0

- m -

- e o o 6 4 8 * $s o gh gn C

-2 _

og P o _ _, _

+ o _

, l

[Q -

g"O, _3 i r i i .

_3 , t . e i ,

E 4 5 6 7 8 9 4 5 0 6 7 8 9 l m

o Magnitude (Mw) Magnitude (Mw) eo

{[ 3 . . .,,,, , . . , ,,,,. 3 . . . . ,

,,,3 . ,,

- Rock , -

- Soil o

{mR oo '

3 ~ o interfoce o+ ' - o interface e

9 OO

~

. intracco o intraslob O 3_

• 0

• c *

'o

-m c.

• O

++

3 ~ 8 Ea g 0. 6 .'. . % .

' 'o.

!' .- e , ,.

oo c- u* O o n

'u

?o  % .

0 n n o,v ,[o go..*Oo ' sfG OA

• M, e O ou do AI

• r v u

<o o d, m ga 4 0 o o e go 4, o o+ o o *gG opo -

6 g +7 0 aoo o o 3 -1 - O 6,

,g _ _j _ e oo do e po _

g 0 o o o c' b# k o

-]- o o 4 3 't ,

O g- -2 -

o g 0 -

-2 - -

eB ~

am .

3G _3 . . ....i . . . . , , ,

_3 . , , . ....t , , , ,i

? ~~

C 10 20 50 100 200 500 1000 10 20 0 50 100 200 500 1000

$ Distance (km) Distorice (km)

o

. - y , i g i

P CD . < , i l

l

.h o

OC 0 0 o O O O

- 0 000O O2c3 or C - 03 0 O CPh o ' CW g O O 0 $0$ O - O* g .* ,

g+O Om> OE * *e) .

o . o

- $ O - 6

+ ogo O

,O+ .. 0 O b o

• 0 9 8

] 7O O* o 82% .

O*O 8$

0 C 0

• j .

, ,0 g , . o 0 Oo+ CD C. - C .

OO *O O.. $

l ,

OO 00 . O c O .

. O .

O _ _ o oc g _

00 . .

~

~

h O - c ,, O s,3 -

35 ca -

R

. at

=O s s

=as OS5 W O. W O.

I , I f . I . . 1 i l i . I i 1 i O 9

n m - o 7 7 7 n u - o 7 y 7

, . , . cn . .

o 8

dh co y 0 e $ 0 .

,O -

0 O $ ~

0 00 " * *  !

0oO OOg) o g0 coco -

• O o g0 04 o j o,

m O

O O

o

= - 8 m

c. - s m. O

. O (8e -

,. O o o ., E C- 00 O "li ,

O

• 40% .% 6 l OO

. 3  ;- .O o

C*c h88 l

• 2 . 0  % 7 -

. O . g o.g #O.c.8 0  !

O 2. , o To 0 - e ,8'  : *O O

,0 0

o  : g  !

• O O

,O

• 3 3

Oo O . .

@ j 3

O - - 00 (p$ -

l i

• . O I O o3 )00 -

g

~ -

O c o e a o O 80 - at 89 o -

g O N x:

U Sd 3 oas i O55 Oss O E O + 1 O +

i i  ! . > l i I , . I . 1 i .  ! , I i o N m m o N n N - o T 7 -

7 7 lonrisay paz!iowsta ionp; sad paz!iowsta Figure 30. Weighted normalized residuals for rock and soll data plotted against magnitude and distance. Residuals are based on fit of Equations 5 and 6.

1 _ . . ......, . . ._ _ .

. . .....i . . ._ _ . .......; . .

7:  :  :  :  :  : .$

ci 5  :  :  :

c _ _ _ _ _ _

e _ _ _ _ _ _

u p g .2 - - -

o - -

o E

e a

S c

o j _

o +

,e, s o, So .05 - - + - -

m-s u

'4s _ _

s

• O . _

00 '

• +'

y - '

,s, sO -

m

@ g .02 - '* - -

o 's, - -

3 e s O -

O ' s g Q-

,oj _ Mw 5.3-97 ', _ __ Mw 5.8-6.2 g ', __ _ Mw 6.3-6.7 o '

6 Introslob ',  :  : 4 ntroslob O 's  :  : * 'at' 5' b O ':

{---

interface O '

3 .005  : O laterfoce ', :  : O O  : O interface ',:

g ---- Introsfob - Rock ',- - -- Introsfob - Rock 'Q

---

Intrastob - Rock -

F Interfoce - Rock Interfoce - Rock Interfoce - Rock f

2 .002 * ' '' ' ' '- ' ' ''"' ' ' ' ''' '

o g 1 _ . . . . ...., . . . . . . . ....., . . ._  ;

EE  :  :  :  :  :  :

13 5 -

e m . . _ _ _ _

g _ - _ _ _ ~,0 O _

_, -,. '2 - - - -o' , - -

Oy G _

N '

s3 O O 8 ,O -

O'O s x c .j _  % 's _ _

O,s OO 'g m- .o _

\e s - -

g  : ,  : O s s  :

e m

se .05

's O

'O ,

g' o.

o s, o s, O

~

m u -

O ,

J, -

m o

4 O , , s LD s ,e

_x '02 - s - -

,0-a o s Oo ,

E S ,__ __ Mw 7.3-7.7 O 's Mw 7.8-8.2 O 01 _ Mw 6.8-7.2 o O t

ITI  : e introstob '- -

4 Introslob  :  : 4 Intrastob _

O

$ .005  : O Interfoce  :  : O 8aterfoce _  : O interface  :

R_ ---- Introslob - Rock O - -- introstob - Rock g -

---

Introsfab - Rock -

Interfoce - Rock O Interfoce - Rock Interfoce - Rock 3

.002 E 10 20 50 100 200 50010 20 50 100 200 50010 20 50 100 200 500 Distance (km) Distance (km) Distance (km)

1 . . . . ._ . . ._ -_ . . . ,_

.....3 _

D  :  :  :  :  :  :

e b

- O -

,C - _ _

O _ _-

m - ~ ~

u

.M g .2 -

O (p -

' - 'g,, _

a h

n. m

[

O,Ns, s * ' l _

[

- pO s, E

E

,,, O

's ,

i_

- g .C5 -

M '

O s, s

' 4 s u ' '

m #

s

@ x .02 -

s' - -

s 8 ', s a ',

' __ Mw 5.8-6.2 \_ ._ Mw 6.3-6.7 U r M w 5.30 5.7 E .01 ,  ; _ ,

4 Int'oslob (:;:  : 4 Introslob \[ [ 4 Introslob

• h a .005  : O t.iterface  : O interface  :  : O interface g ---- Slob - Soil -

---

Slob - Soil O

---

Siob - soi -

Interfoce - Soil m ?nterface - So?. Interfoce - Soil I - - - - ' - - - ' ' -

@ 002 ' '

3, 1 _ . . . ..... . . ._ r . .......; . . . _ _ . . . . . . , . . ,_

=r  :  :  :  :  :  :

13-OO _

• 5 - - - -

-O' _

e ,, - -

-O -

o Og - -

rn .2 *'s - -

n

@ Op o_. 3 -

h, c Q 's '

1. o .1 r O -

s , _ _

i  : , s,  :  : O O,  :  : O ', :

l,ii m _ _ _ _ _

• C5 - -

O cr m

Tu o

O n

e _ _ O - ,

m o - _ v _ _

e

o. .x .02 -

• \- - - -

o 8 -

_ Mw 7.8-8.2

~01 - Mw 6.8-7.2 ._ Mw 7.3-7.7 _ _

E E , introsiod _ E 4 Introsicb O:

CE E e intrastab i E .005  : O interface  :  : O Interfoce  : O interfoce  :

C - - - - Slob - Soil -

- - - - Slob - Soil O - -

- - - - Slob - Soit

$ Interfoce - So.1 Interfoce - Soit Interfoce - Soi

.002 -

10 20 50 100 200 500 10 20 50 100 200 50010 20 50 100 200 500 Distance (km) Distance (km) Distonce (km)

i l

j l

l 50e . . . . ,,,

, i i i.i., i i, (a) 200 - -

3 / ~~,

)100 -

[_, '

$g  :

/ ,'

j 3 50 -

a' -

l g

,r' ~

l E

,' ~

O r i

} 20 -

m ,  ;

,# R < 150 km 10 .- ,' -----

R < 150 km 7

~ ' ~

i,in iiil , i iiiiil i ii 5

i

.02 .05 .1 .2 .5 1 2 5 Period (sec) i 3 i i4

....; i i i

,i.iig i ii

- ~

(b) 2.5 - -

i , 's ,

2 -

1 y -

< s 51.5 - ,' -

o -

VI ,

1 A

---

st w I speevoi she '

.5 - _

Smoothed SpecVol Shope s l

l

' ' ' d ' ' ' ' ' '*

0

.02 .05 .1 .2 .5 1 2 5 Period (sec)

Figure 33. Median spectral shapes (5% damping) for Mw 8 subduction zone earthquake. Top plot shows median shapes for distance <150 km and

> distance >150 km. Bottom of plot shows statistical shape for <150 km smoothed spectrum in terms of spectral acceleration amplification.

1 io .

i i

~

] T = 2 sec  !

S F_ T = 0.2 sec . -

O

~

l

\ . . - - l 2 - - - O -

E OO j O

82 e

ix 1  : -

-;>W = = =- A Q W -

- ,g..

- hg- O 5-l o O O  % - -

0 7. - ' o

- 1 1

5 .5 o O O O ,o[' 8 8 W

3 - . _

7 ,,< . o .

O

/o ,, o OO s O 7 /> 0 i s .2 - - -

J y O /.- O O 1 o O 4 E -

zg .1  :

?

O  :

i 1

.o5 - - -

0 - -

O Subduction dato O Subduction dato Sod;gh et ol.1986 - -

. Sod;gh et. of.1986 -

, - - - - Joyner end Boore 1932 .

, - - - - Joyner end Boore 1982 ,

- ~ RVr -- RVT l Subduction - TNs Stucy Subduction - TNs Study

.o1 5 6 7 8 9 5 6 7 8 9 i

Mognitude Mognitude i

Figure 34. Relative spectral amplification (5% damping) as a function of mag-nitude. The data points represent relative response spectral ordinates compared to the spectral shape for a Mw 8 event shown in Figure 33. The curves present relationsh!p developed previously for crustal events and the relationship developed in this study for subduction zone earthquakes.

1

, , ,,,,i , ,

, ,,,,,i _

.5 - - -

.. s s,

, as n

.2 -

's ~ - -

'. s

.s

~

o .s v

- 's .s c .1 7 ,' s --

h.s

.9

_ .s

.s

.s, e

. .s - _

S .05 -

's,

's , -

e . .

l 8 ' . '. u 3 02 -

o . 's -

W .s Mw 8 Mw 7

'01 - -

---- siod - so? ." :---- stod - Sea In'.erface - Soa - -

Interfoce - Sou -

.005 -

Bufeiizo (1985) - biondo Bufoiiro (1985) - biondc .

002 1 --- , , , , ,

i i ,,iii , , ,.

, , ,,,,i ,

.5 - - -

.2 - - -

n w ~' ,

~,

s 8 .1 r ', t -

.' s I z  : 's  :  : .' s  : )

e .

.s _ _ .s _

1 so .05 -

.' s - - -

,' s - - .'.s Q

fo .Q2 - - - ' -

s a- 1 M* 7

'- M* O i

'01 - -

---

intreslob - Rock ,','

}---- Introslob - Rock ,

Interf ace - Rock k -

Interfoce - Rock -

.005 -

eurakzo (1985) - crme Bufoiiro (1985) - t,rme  : I

.002

' ' '' ' ' ' ' ' ' l ' ' ' l 10 20 50 100 200 50010 20 50 100 200 500 Distence (km) Distance (km)

Figure 35. Comparison of median relationships developed in this study with those developed by Bufal!za (1984).

1 ,

iii>ig i iii . . .. l

. i 1 i

.; . i iig

.5 - - -

_ ' ' ~ ~ . .. -

s

% g

.2 - '

's -

3 's -

s v

's . s 8_

.1 .- , 's s  : r -

, 's -

s g .

s s

s s

m . ,

o

.05 -

's

• \-'

v s .l -

N -

,s

.s -

i .02 N-o /

Q.

Mw 7 Mw 8

'01 -

---- Sieb - Soil 2---- Slob - Soil Interfoce - Soil - -

Interface - Soit -

• 005 -

l twosaki (1980) soit

. - tweseki (1980) soil 22 ' ' ' ' ' ' ' ' '' ' ' '

1 .

, ,....i . , ,. _

.5 - - - -

N m

.2 -

's s -

ss s, 's s c 'j s - -

's s

.9

's  :

. . . s .

O .

\ . .

., s .

2o .05 -

,s s

U '& g V . - -

< 's. ,

s

'y*-

f .02 - -

o

• s Q.

Mw 7 's - -

Mw 8

'Q 1

}- -- IntresIOb - Rock y ---- Introslob - Rock ,

interfoce - Rock k -

Interfoce - Rock - 1 005 -

1 I.csoki (1980) rock

. 1.osoki (1980) rock

.002

' ' 'I ' ' ' ' ' '' ' ' '

10 20 50 100 200 50010 20 50 100 200 500 Distance (km) Distance (km)

Figure 36. Comparison of median relationships developed in this study with those developed by !wasaki(1980).

1 l

l l

1 _ , , - , , ,,,; , , ,_ _

, , ,,,ii , , ,.

~

.5 _ -

l l

l .2 - - - -+ .

l 7., ,

N ',',' '

l C s -

• ] -

N 'l 's

.9  :  :

~

'N  :

} .- N ' ,' s ,  :  : N,N  :

• .05 .s - -

N*.N -

3 x ,q - -

s ,3" y -

N '. - -

g . -

% .02 -

T.s. - -

N(s . -

e s

' \ '

• Mw 6  %'

Ma 5 -

q

.01 -

---- Introslob - Soil

%,g -.";  :---- Introslob - Soil

. s,--

- Interfoce - So;l s N- -

Interfoce - Soil -

'005 _- -

Crouse et oi (1988) d=70km 's - -

Crouse et ol (1988) d=70km s.

~,,vuse ;,i et (1983) d=30km s. _ - - Crouse et oi (1988) d=30km .

.002 1 _

.,,i,ij . . ..

.5

, , ' g- ',,

~, .

s

,2 . N ; ,, ' n . - N

'\ _

^

m . 's N Ns v N '

N N 8 .1 r N N 7 r N \ 7 gm  : N \  :  : N ,

',s  :

N .s - -

N

@ 05 - -

N ' * ,s v

_ . . N 's v _

V \ *\

4

s' , N

% .02 -

Ns - - -

)

e

c. ( .

Mw 7 -

Mw 8 --

~01 ~

---- Introslob - So:I  :---- Introsiob - So,1 )

1 Interfoce - Soil - -

Interfoce - Soil -

.005 ~

Crouse et ol (1988) d=70km _ Crouse et ol (1935) d=70km  : j

_ - - Crouse et at (1953) d=30km - - Crouse et ol (1938) c=30ker. . !

l 002 l 10 20 50 100 200 500 10 20 50 100 200 500 l Distance (km) Distance (km)

Figure 37. Comparison of median relationships developed in this study with those developed by Crouse and others (1980).

21 to C 4 . , ,

, ,,iig , ,, , , ,

,,,,,g , , , , , ,

,,,,,g , ,,

p _ _ - _ _ .

-S 8 2 - - - - -

w +

2 O

• O

@ $ 1 E  : : * '+ 0 i E , ,

7 5 Y { O o

• O { { #o,cP h { { t

o. .e 5

_ og _ _

+g .. _

o ,

<: 8 + e 0 e I O _ _ _ _

d o Yo .2 - - - - - -

E >

g S .1 -- Mw 5.3-5.7 m _ Mw 5 8-6.2

-- Mw 6.3-6.7 y - - - ~

y 4 Rxk 0 Rd 7 { 4 Rxk }

.C5 - O sod -

- O soa _

_ O sed _

O .g3 * ' ' ' 'a'al ' ' ''*I

• u a > ' 'a nI ' in E

s 4 , ,

gt _ _ - _ _

3e 8 2 - - -

= 0 8 O *

{ 1 r i r +

• l' o ge

_ , e' *  :

If 0 b O OC+% [ [ , ' O,*

, 6* _ _

k .e

'O

O, *O  :  :

O -

+

• 0 .0 0

E b _

,0 0 g_ _ _ _

Q t* , _

S m

by *2 _ _ _

O _ _

g >

S .1 -- Mw 6.8-7.2 _ _ Mw 7.3-7.7 - Mw 7.8-8.2 y  :

~ ~ ~ ~

O  : 6 Rock 4 Rock 4 ' Rock -

n _ _

_O

_O x .05 - O soa _ _ soa - - soa _

m -

.03 ' ' '' ' ' ''' ' ' ' ' ' '"' '

8

. 10 20 50 100 200 50010 20 50 100 200 50010 20 50 100 200 500 o

= 0. stance (km) Distance (km) Distance (km) c.

E

.2 ,

, , , ,,,,i , , , , , ,

,,,,ri , ,,

n di .1 r , r -- * ,

C, n, - - .-

~~ , _-

-'s, o

, ~~,o a _c - - -

.C5

,0 ca c 0 .

e .g .,

8 's O O

,'s 9 - o cr g :

o 02 - ' - -

e - -

m e v

' ', *

• O6 0, s

's s

$ 0: # ' '

O 's O g

.01 -

s - _ __ _ _

am o  :  :  :

s ,

o o o s - -

s - -

'-u -

~ ~

3m .005 -

s s

s - -

OO ',--

s -

s

[Et g > _ Mw 5.3-5.7 ', _ _ Mw 7- 6.2 '

- Mw 6.3-6.7 'c

~

Intrasfob - rock O '

C C m

o

.00.9 4 '

'- e 'oslob - rock 4 Intrasfob - rock  ;

'O

. . _r -. o- O Interfoce - rock  ; o .nterface - rock .

O interface - rock O

3 .001 -

__-- Intrastab -

s

---

Intrasfob -

Intraslob j a -,  : Interfoce  :  : Interface 3  : Inte-foce

- .0006

~

~

' ' 'I ' ' '

' ~

m g- .4 , ,

, , ,,,,i , ,, ,

,,,,,,i , ,,

" O 2 - -

Z is

. ' O~~ .

m ' g '. 's n j ._

, _ _ O -_

5 'q  !  ! dQ ,

O O D

< .9 .C5 -

O',

s - - -

m g -

s, - -

O 's -

o' '

O 2_ y

- - O s, - - s _- - -

I; @ 6 O_

g <

u

.02 -

O 's s

s

. O s

s O  :

j Cl O '

E Q~

i F O ',-~  :

o r, s

7 m > .005 -

O c -

~

O Mw 6.8-7.2 Mw 7.3-7.7 [ Mw 7.8-8.2 g O I

-, o,. 4 Introsiob - rock 4 Introsicb - rock 4 Intreticb - rock o 002 - -

o - -

o O laterfoce - rock O interfoce - rock o interface - rock X --- Introslob --- Introslob ---- Introslab e .001 r On r r

_- - Interfoce - -

Interfoce - -

Interface -

C '

~ ~ ' - '

@ .0006 ' ' '

10 20 50 100 200 50010 20 50 100 200 50010 20 50 100 200 500 DistOnce (km) DistonCe (km) DistCnce (km)

500 ,......i , , ,

(a) .

200 ,o' 3 ,s' '

, \

N y 100  : ,'

sg - o 3 50 -

l x , I l

g -

~

}

o 20 -

Ln ,

, R < 100 km 10 r ,

.___ g , too km 7 5

.02 .05 .1 .2 .5 1 2 5 Period (sec) 3 ,,

,,,,,, , ,r-  ;

~ ~

(b) 2.5 ,

l, 's, ,

l 6

2 - _

1 l

g - , - i g

s o 1.b - -

x '

, s S -

a -

l 1 -

.5 -

~~~~

S'*I'8" ' SP ' ' Sh P' -

Smoothed Spectrol Shope l

l 0~

.02 .05 .1 .2 .5 1 2 5 Period (sec)

Figure 40. Median vertical spectral chapes (5% damping) for Mw8 subduction zone earthquake. Top plot shows median shapes for distance <100 km and distance >100 km. Bottom plot shows <100 km statistical shape and smoothed spectrum in terms of spectral acceleration amplification.

s'J. O scn ,g,if?cs n

' obc $', ',k b 'kb

' R)w,'O,',t>

se ss ,

e Ob O O A' O e O

' ' Front tg 4 4

3flckO,,

c,0 9cgo,,

- "6

'fGjl?gg$#

G eos hi 2 s

'6 G,i eo ee-'0 e i s

, ,0 s '

s'

C;'O '140,, O ,

Gr', + H /

6 - 's a) Schematic representation of rupture propogation.

l S oo -- ,

/

l {r-

/

l t 3

G i f

/

/

0 '

O Tp Time b) SchematR: representatbn of time history of slip for a single subelement on fault.

Figure 41. Schematic representation of source rupture model used in simulations

T T y T I I I g g g g g q y y y I g y y g g g y y ( [

Coleto de Ccmpos NS - Mexico Lo Union NS - Mexico 100 --

necor eed r Recoreed

_ ---- S,muta t.on

: ---- Simula tion y 50 -

s,' ' - -

a ',

,j%

q -

i i G .

, ', 's s . - *%

y f ,

v <

> 20 -

,.' s

, '>~,' i s

- , i

, i e t,,

$ 10 --

, '/\

._? -

y u ,

o 5 -

G I w . . -

2 -

,,! , , ,  ! , , , ,,iil i i i , i i i , I , i i

)

200 ,

,, . ' i . . .

...i . . > > > . . .

i La Vithto NS - Mexico Volporciso UFSM N70E - Chile 100  : Dewces -~

Recorded S.mviction  :  : ---- Simulction g 50 - - -

',',i

',,'s

{ -

i E -

-- i ' i ,' ' - -

,, i, i,

v ,

%s g 20 -

,,'l,,

I i

}

i .

3, ' s i'

$ 10 --

.-2 - - - _

y - -

c 5 -

O -

2 - - - -

,,,,1 , , ,,,,,! , , , ,,,,1 , , , , , , , , 1 , , ,

3 05 .1 .2 .5 1 2 5.05 .1 .2 .5 1 2 5 Period (sec) Period (sec)

Figure 42 Comparison of response spectra (5% damping) for simulated motions and recorded motion for the 1985 Mw 8 Mexico and Chile earthquakes,

100 _

i

,,1 1 i

, . ,i,i .

~

50 -

./..-

1. T 7- d'N ~ .

/, d" \\

U s

s 2o - /

./,

\q\,-

[ -

N 10 - ' -

.N M -

f' p.-

f-y 5 -

), ', -

u k'E -

R = 29 km -

& '- M

• 8' '

W 2 - -

. .. M.83

---

Mn85 1 --

-- Ma86 --

M . 8.9  :

~

.5 100 .

,...i . .

.....i . ..

50 - -

re.--- N

- gr4e s, s .

q /:-

8 20 -

d' .~. . ' , )* -

} ./ \ t e' f' /

\

3 E

[

, ' -' /

\1 z \

-O 5  :

/( /

3 -

R = 38 km -

o

& 2 -

/',/. . l.

"" 7~7 -

W /

/ M., 8.1

/ Mw f,.3 1 7 ,

M. 8 3 -

/ - - M e 8.6  :

~

.5 '

.02 .05 .1 .2 .5 1 2 5 Period (sec)

Figure 43 Smooth spectra for simulated motions from f arge interface events beneath WNP-3 site.

l l

l 4 .

i T = 0.2 sec. R = 29 km T = 2.0 sec, R = 29 km Equobons 6 and to - ~

Equot.ons 6 ond to -

C5=o ----

C5-0

- - C5=-C2/C3 - - C5= -C2/C3

$ Smuiotion $ Simulation g2 - ... Fit to simulations - -

Fit to simutobons -

i _t. -

& ,.+-

s .-

& , . 4- '%

,+ ,> ..-' ,

3 _

' , .,,/' -

,.,6-

/. .

- 7,,..-s -

.6 4 ,

i i i T = 0.2 sec, R = 38 km T = 2.0 sec, R = 38 km Ewations 6 ond 10

- ~

Equations 6 ond to -

C5=0 ---- C5=o

- - C5 = -C2/ C3 -- C5 = -C2/C3 ,.

$ Smuiction $ Simulation .

g2 - . . Fit to s.mutot;ons - - -

Fit to simu lations -

g

/ .

O; f'..-

4 +-

w .-

,f - -

g' ,s 1

/. .

. / l

, J. .

4' <.

l .6 '

I 7.5 8 8.5 9 7.5 8 8.5 9 Magnitude, Mw Magnitude. Mw Figurc44. Comparison of near field magnitude scaling relationships based on simulated motions with magnitude scaling based on empirical attenuation relationships. Solid curve is empirical attenuation relationship developed from data for Mw 8. Short and long dashed curves are limiting constraints on parameter C5 . Dotted line is linear fit to simulated data constralned to pass through 1.0 at Mw 8.

4 .

i i

i i

T = 0.2 seC, R = 30 km _ _

T = 2.0 sec R = 30 km ,

Ecu1.ons 5 cnc 10 Ecuot.ons 5 ond 10

.. . "S-Cubef' Model ----

"s-Cutei' Modet

.. Joyner Joyner 3 -

m '

II ,

2 , ,'

• i '

s' ,

D , '

) ,

D ,

/

.6 4 .

i i

i i i i T = 0.2 sec, R = 40 km , ,

T = 2.0 sec. R = 40 km -

Ecuctions 5 cnc 10 -- Ecuot ons 5 ond 10 s ,

"S-Cubes' Mocel ---- "s-Cube r vocel Jeyner - Joyne<

g 2 -

II ,

at ,

• 1$,

u ,-

N ,

V) ,

) _ f<,e:7 _ _

, e l'4

.6 7.5 8 8.5 9 7.5 S 8.5 9 Magnitude, Mw Magnitude Mw Figure 45. Near field magnitude scaling for attenuation relationships developed in this study.

l l

l 6 _ _ _ . _ _ _

l T T T .

. I .Y Y Ig 3 y 5 .I E U y y y 5 yy to E 100 - Mw 7.25 = r Mw 7.9 =

cm -  :  :  :

4 - - -

a m 50 - - -

. v - - -

N O. g O O

E 20 -

-o 2, '

<m3 o

60=V 3 to

_g y3m  :  : -

u-  ; 5 - -

a a-O"Y m - - -

CD 3 --$

- ~ ~

xoo e U 2 - - -

z<-

ge3 a

LA

) _

Heaton end Hortzen 1986 _

Heaton ond Hortrett 1986 _

o o

TNs study - S-Cu>ed  :  : TNs study - S-Cut >ed  :

O G. - - -

g ] g- -

TNS s'udy - Joyne - -

TNs study - Joyner

$ f

• Js O. 3 -

3 ,

3 ,_,.....t i , . .....I . . . .,.....I . . . . . .! ,

a3e 2gg ID -

200 . . . . , , , .

., , ,m ,

S$o m -* $

' im r Mw 8.5 r Mw 9

=

r Mw 9.5 =-

C - -

o.m 50 -

axg o - - - , . - - . . .

co - - - -

._, g . 20 - - - - -

y :r m o ,-

-, m v - e.

e ci x 10 -

ya gt -

g - - - - -

oa

• 5 -

3 m > - -

3 e g i

\

m~ v y3 e ~

9 . - -

,- - - a 3 Q G O =-

y _ Neoton ond Hortreu 1986 _ ,_

- Hectoen ond HortzeH 1986 _ ,,

- - Heaton ond HortreN 1986 _

ins study - S-Cut >ed  : TNs study - S-Cubed  :  :  :

m' e ins studf - S-Cut >ed a

- .5

- --- TNs study - Joyner  ;  ; - --- TNs study - Joyner ,

---- TNs study - Joyner  ;

r> - - - - -

g 3

, . .,f n . ....t . . . , n . . . .st . . a n . . . .I . , . . anal . . 1 1 annat -

Q- 02 .05 .1 .2 .5 2 5 .02 .05 .2 .5 2 5 .02 .05 1 .1 1 .1 .2 .5 1 2 5 e

x Period (sec) Period (sec) Period (sec)

127.00 120.00 S9.00 , ,

99,00 V v g, 3 - --

x .

. ,\ . .+

(

~s x \V ..:er.

• Mg ,fM,Q . ,

. ' ,,. l . .* 3 * .

l *. f

, . /t *

+ + +. .*'. , . 2+ +

• .. ~

T. m

. s' ,

• . . =,4, (

• . , * ['p .

I l

<>. t .

, . I. . **

s . i .-  ;,*',

+

, ,  ?.* . ,. ;

.ij8i' ),

3

w. .

.. s. , .

,t

• 4 + -

yy.q., . .

+

• 9 =

f.

f . .

45.00 ( *' 45.00 127.00 320.00 ci i i i i i i i i i izx y b

% T 2 - G T 4 ~

au

~

1  ::

= -* - .."  :

~ R as 2

. = .. .

2 7 3 00 G 0

", 1 , . , , , , , ,

SE o e P) ate .

.. *?.,-

g50

, .. . ne ,%* **

1. ) ,, . ,

7,,,,fgg,.ture i

o 60 - *(*.*.

70 -

80 -

90 -

l 1:1 Vertical saggeration Figure 47. Down dip cross section through subduction zone in site vicinity.

Minimum distances to subduction zone sources are shown.

k 12 ,

n > >n new "r va B 25 et 33 km ---- ~+oa v. 9 :5 et 33

• n -

e4 m s4ts 1 - - 5#1 $

- - $!E -

/ \ /

/ / \ 1

/ \ /

/ \ \

3 3 - - -

/

7 / \ / \

.?

o / \ / \

1 / \ / \

$ 6 -

/ \ - -

/ \ -

?o -

/

i

\ / \

f ,,

/

', - \

[

{4- / . . -

x

-/ ,

s,

\

N

/

s

\

' ~

. r, /

3 ,.- ~, \

2

, x / ,'

, \

s

, y- -

s Hori2cntal '.'.

Vertico!

O  !

02 .05 .1 2 5 1 2 4.02 05 1 2 5 1 2 4 Fe'iod (se:) c .:: (3e:)

e Figure 48. Comparison of 5% damped horizontal and vertical response spectra for a Mw81/4 interface earthquake at a distance of 33 km with the SSE spectra.

1 i

l l

s 12  ; ,  ;

v. 9 5 1 95

---

u. 9 med o, et 33 km ---- v9 ~e: cm et 33 km

w. e 3 u. e s i _ _ _ s2 3

_ - - su 3 J

/ /

/ \ / \

/ /

/ \ / \

/ *

\ / '

~

p g _ _

s

/. .\ / \

g /,/ .\ T

/ \

, /. - ,- , .\ l- / \

$ 6 -

/; ,-

'-s, *

.x

^

/ -,' s t ,_., ,.y

, y 2 . < ,

~

g / ,,' i, ' \ . /

,' 7 ', \

• c 4 _ ./ , '

/ ,'

',\

\

/

/. , /

N\.,' , \ N

,-. ./,, i, x _../. '

\, x r

, x . . y,. N ,;. x (

' , . N

,. s-3

' \-

xs-

'S-Cubec' Scaling 'Joyner' 5 :. ng -

0 0: 05 .1 .2 .5 1 2 4 .02 C5 1 2 5 1 2 4 Period (se:) Pe .:: (se:)

Figure 49. Median 5% damped horizontal response spectra for interface events in the magnitude range of Mw 81/2 to 91/2 for two forms of large magnitude peak acceleration scaling.

1 1

I

1.2 ., ,, .; .;

w. e 75 u. S 7*

---

W.85 84th per:eatJe et 33 km ---- u.85 64tn percentde et 33 h.m

v. 8 25 v. 8 25 i _ -_ - tsc -

- - ssc 3 3

'/ /

/ '\ -

< \

y- . / .g

'. ./..

_  ;/,v

' \. ,, _ _ .7, , ' . ' . , ' \,

) '\ .

1: ' ,\

a  :) 'v.

' L' ~,\

5 '

'?- 'N.  ;{ '\

b s

~Y

,); 'q' . - -

.; a -

,Q . '/

s hy ,-l '\.' ,'

'/ ' .\

- ,7

'$ y '..\

M" -

j .\ _

~ ~

/y'

';. \

',;. N .

', x

. \ ', N 2 _

'.\ ,

s.

N_

',N

'. N_ ,

, , 'q.

'S-Cubed' Scoring 'Joyner* Scoling 4 g 1! .i .! .i 02 05 .1 2 .5 1 2 4.02 C5 .1 2 5 1 2 4 Peiod(sec) Pe,o (se-)

Figure 50. 84th percentile 5% damped horizontal response spectra for interface events in the magnitude range of M w 81/4 to 8 3/4 for two forms of large magnitude peak acceleration scaling.

100 ,;

50 -

,'n..-, .- # IhN ~ .

d," N\

s 20 ..h' 'k:

b 10 -

,s' o f, - .

~

\#

y -

2 o />,,' -

3 5 ,, ',

? j;}' - S-Cubed (1988)

{

g 2 -

/;', v. 81

u. 8 3

,e . . . . u. 8 5

, -- u. 8 6 1

? -

u. 8 9

.... Sg

.5 200 ,

1, i 1% -

~,(..::: :..".'

~ ' '{ -

50 .

n

,'/,',

,'/,'

,'/, ~%

l i N '/,'

e h 20 ,' / a',

b ,

o / *'

e' Q 10

,' /,

>/,,'..

, / ,.'-

$ , /,'?

Heaton cnd Hortrell (1936) b l 0 <'

$ ,' v. 7.9 2 ,' u. 6 5

.... u.,

-- u.

, f .... sm

.J.

.5

.02 .C 5 .1 .2 .5 1 2 5 Period (sec)

Figure 51. Comparison of ho:izontal SSE spectrum (5% damping) with response spec ra from simulated motions for large interface earthquakes.

l l

1.2 ,

i ,, .i ,i meson men

- - - - me an Ma 7.5 ct 70 km - ---- mesa va 7 5 et 70 km 84th 84 th 1 - - sst ,3

-- sst -

/ \ /

/

\

/ \

/

- / \ - -

/

3c. .e _

S / \ / \

$ / \ / \

~t .6 -

/ g

/ \ -

3 / \ / \

~

S / \ / \

(,

c' -

/

,,,,,' , \ / \

/ x / \

, 7

~, \ / \

N / \

, y ,e , , -- - , s 4 - .- s ,

, s -

', ,- N

's , 's, N Hon,zontc! ',, Vertical '.,

0

L '

.02 .05 .1 2 5 1 2 4.02 05 .1 .2 5 1 2 4 Period (sec) Per,od (se )

Figure 52. Comparison of 5% damped horizontal and vertical response spectra for a Mw71/2 intrastab earthquake at a distance of 70 km with the SSE spectra.

e a

w STRONG MOTION DATA BASE l

m D

6 1

STRONG MOTION DATA BASE Date Earthquake Lat Long FD RT mb Ms Mw Station C HD RD Comp Anax ALASKA 1964.06.05 Alaska 60.35 145.87 16 th? 5.2 Cordova R 27 N286 0.0306 N196 0.0349 1968.12.17 Alaska 60.15 152.82 82 ss 6.3 Seward R 205 N090 0.0224 N000 0.0380 Seldovia R 130 N090 0.0269 N000 0.0413 1971.05.04 Adak 51.42 177.21 38 th? 6.8 Adak R 77 N180 0.1168 N090 0.2076 1974.04.06 Shumagin !$ 54.87 160.29 37 th 5.8 5.6 Sand Pt R 65 65 N120 0.n768 N030 0.0911 1974.04.06 Shumagin is 54.90 160.29 40 th 6.0 5.8 Sand Pt R 61 64 N120 0.1002 NO30 0.1201 1974.08.13 Adak 51.49 178.11 47 th? 6.1 Adak R 123 W180 0.0223 N090 0.0298 1974.11.11 Adak 51.59 178.08 69 th? 6.1 Adak R 128 W180 0.0310 N090 0.0466 1975.07.25 Alaska 55.04 160.41 38 n 5.8 5.6 Sa-d Pt R 51 N120 0.0098 -

NO30 0.0130 1976.02.22 Aleuti ons 51.57 176.81 61 n? 4.7 Adak R 72 N180 0.0282 N090 0.0670 1979.01.27 Alaska 54.79 160.64 53 th 6.2 Sand Pt R 82 N197 0.00 77 W107 0.0094 i O 1979.02.13 Alaska 55.17 156.94 47 th 5.8 6.5 Sand Pt R 231 N197 0.0228 .

N107 0.0422 ,

1979.02.28 St Elias 60.64 141.59 13 th 7.2 7.5 Yakutat S 167 101 N009 0.0829 N279 0.0620 fcy Bay $ 76 43 N180 0.1747 WO90 0.0982 Munday Creek R 73 50 NS 0.0640 EV 0.0416 1981.12.28 Alaska 54.67 160.41 33 th? 3.8 Sand Pt R 82 N070 0.0188 W340 0.0246 19a3.02.14 Alaska 54.74 158.88 25 th 6.3 Sand Pt R 125 N250 0.0077 N160 0.0068 Simeonof RL 41 N070 0.0305 k340 0.0567 Chernabura R 52 N070 0.0478 W340 0.0413 Pirate Shake RL 87 N072 0.0121 W342 0.0250 1983.02.14 alaska 54.85 156.84 25 th 6.0 Sand Pt R 121 N250 0.0058 pp>

N160 0.0040 Simeonof RL 38 k070 0.0284 h340 0.0413 Chernabura R 54 N070 0.0170 N340 0.0206 Pirate Shake RL 81 N072 0.0151 N142 0.0139 1986.05.07 Andreanof is 51.41 174.83 16 th 6.8 7.7 8.0 tdak Hanger S 151 60 Long 0.2500 Tran 0.2000 t

\

STRONG MOTION DATA 8ASE Date Earthquake let Long FD R1 Ft) Ms Mw Station C HD RD Corrp Aman CASCADIA 1949.04.13 Puget Sound 47.10 122.75 54 n 7.1 olppia 5 56 N176 0.1743 N266 0.3272 Seattle S 80 N182 0.0601 N272 0.0765 1965.04.29 Seattle 47.40 122.40 57 n 6.5 6.7 etnia S 80 N176 0.1641 N266 0.2008 Seattle SEF S 61 N238 0.0836 .

N148 0.0591 ,

tacoma S 60 H090 0.0754 N180 0.0459 CHILE -

1945.09.13 central Chile 100 n? 7.1 Santiago E de ! RL 106 Long 0.1310 fran 0.0670 1952.04.29 central Chile 10 th? 6.0 Santiago E de ! RL 172 Long 0.0070 Tran 0.0060 - .

1953.v9.04 ce tral Chile 50 n? 6.4 Ssntiago E de ! RL 146 Long 0.0150 Tran 0.0170 1958.09.04 central Chile 15 th? 6.8 Santiago E d 96 Long 0.0300 ,

Tran 0.0520 , .1 1965.03.28 La Ligua 61 n? Long 0.1870 6.4 7.2 7.4 Santiago E )

Tran 0.1710 '

1967.09.26 central Chile 84 n? 5.6 Santiago E c. Long 0.0280 Tran 0.0250 1971.07.08 Valparaiso 40 th 6.6 7.7 7.8 Santiago E de i RL 175 101 Long 0.1340 Tran 0.1650 1973.10.05 central Chile 33 th? 6.7 Santiago E de i RL 130 Long 0.0110 Tran 0.0100 1974.11.12 central Chile 90 n? 6.2 Santiago E de ! RL 95 Long 0.0330 Tran 0.0440 Cerro St Lucia R 94 Long 0.0310 1978.12.21 central Chile 46 n? 5.3 Chillan SA 66 Long 0.0510 iran 0.0500 Talca SA 106 Long 0.0260 Tran 0.0310 1979.07.05 central Chile 56 n? 5.8 La Ligua A 63 Long 0.2020 Paptdo R 68 Long 0.2000 vina del Mar A 112 Long 0.0340 ,

Tran 0.0250 ValparaisouTFSN R 117 Lone 0.0110 .

1981.11.07 central Chile 32.20 71.34 65 n? 6.8 Papudo R 75 N50E 0.3790 S40E 0.6050 3

La Ligua A 76 N70W 0.3660

$20W 0.4720 <

Llclieo A 178 N10E 0.0730 S80E 0.1970 San Felipe A 106 520E 0.3760 N70E 0.3710 Peldehue A 137 EW 0.2900 A

e

STRONG Moil 0N DATA BASE Date Earthgaake Let Long F0 RT rb Ms Mw Station C H0 RD Coop Amax Santiago E de ! RL 166 NS 0.0770 EW 0.0o00 1985.03.03 San Antonio 33.24 71.85 33 th 6.9 7.8 8.0 Illapel SA 194 142 N20W 0.1200 S70W 0.1000 Los Vilos R 15/ 107 NS 0.0300 EW 0.0400 La Ligua A 114 67 N70W 0.1900 S20W 0.1330 Papudo R 98 54 N50E 0.1300 S40E 0.4700 Zapallar R 94 51 NS 0.3200 EW 0.3300 San Felipe A 127 94 S10E 0.3500 N80E 0.4700 Lloyllay SA 103 73 N80W 0.3400 S10W 0.4900 Vina del Mar A 53 42 N70W 0.2280 S20W 0.3560 Valparaiso E.A. A 46 39 W50E 0.2930 S40E 0.1630 ValparaisoLTFSM R 46 39 S20E 0.1640 N70E 0.1790 l Peldehue A 123 99 EV 0.6400 Quintay A 39 37 WS 0.2000 EW C.1800 Santiago E de ! RL 125 102 NS 0.1100 EW 0.1100 Llolleo A 58 42 S80E 0.4260 N10E 0.6690 Melipilla S 88 58 FV 0.6000 NS 0.6700 Papel A 97 42 NS 0.3100 EW 0.1400 Pichttecu R 145 42 NS 0.2700 EW 0.1800 San Fernando A 168 102 WS 0.2300 EV 0.3400 Iloca A 188 70 NS 0.2200 EV 0.2800 Hualane A 188 77 WS 0.1700 EV 0.1400 Constitucion R 239 119 NS 0.1400 EW 0.0800 Talca SA 274 122 N80W 0.1600 N10E 0.1700 Cauquenes A 308 181 NS 0.0900 EW 0.1200 Chillan Viejo A 371 246 N80E 0.0600 N10W 0.0700

STRONG Moll 0N DATA GASE -

Date Earthquake Let Long F0 RT mb Ms Mw Station C NO R0 Corp Amax o

JAPAN a, l

1956.02.14 Chiba Pref 35.70 139.90 45 n 6.0 tkO24 A 50 NS 0.0771 EW 0.0589 1962.04.23 Japan 42.23 143.92 60 n 7.0 hk005 A 96 NS 0.2870 EW 0.5275 1962.04.30 Miyagi Pref 38. 73 141.13 35 ss 6.5 th001 A 69 54 NS 0.0731 EW 0.0527 1963.05.08 lbaragi 36.40 141.18 40 th? .5.1 kt001 A 63 AS 0.0301 -

g EW 0.0314 it003 A 63 NS 0.0620 EW 0.0662 1963.D8.04 Chiba 35.43 140.35 39 *.57 5.1 kt014 A 52 NS 0.0944 ,

EW 0.0807 ,

1964 02.05 lbaragi 36.40 141.07 54 th 5.6 6.0 kt001 A 65 NS 0.0574 EW 0.0405 NS 0.2068 j EW 0.1393 NS 0.0422 -

EW 0.0323 1964.06.16 Niigata 38.35 139.18 40 r 6.1 7.4 7.6 1 SA 71 57 WS 3.1314 EW 0.1742 1964.11.14 lbaraki 36.47 140.63 69 n 4.9 5.1 kt001 A 69 NS 0.2655 EW 0.2429 1965.04.20 shirucka 14.88 138.30 40 ss 5.6 6.1 cb002 SA 45 NS 0.1220

• EW 0.0674 ebOOS SA 50 NS 0.1107 , .,

EW 0.1558 1965.10.26 Kunashirl is 43.73 145.52 159 n 6.2 7.1 hk004 A 227 NS 0.1076 ,

EW 0.0681 1967.11.19 lbaragt 36.43 141.22 48 th? 5.6 6.0 kt001 A 69 NS 0.4733 EW 0.3570 1968.04.01 Hyuganada 32.28 132.53 37 th 6.2 7.6 7.5 kkO14 SA 322 207 WS 0.1381 EW 0.1695 o ks002 A 75 50 NS 0.2941 EW 0.3581 ks003 A 135 114 NS 0.0335 EW 0.0330 sk005 SA 167 127 NS 0.0720 ,

EW 0.1093 sko06 A 88 59 WS 0.1967 EW 0.2282 1968.04.01 Myuganada As 32.24 132.21 40 th 5.8 6.3 sk006 A 110 Long 0.0560 Tran 0.0640 1968.05.16 Tokachi 0ki 40.73 143.58 20 th 6.0 8.2 8.2 hk003 A 291 120 NS 0.2258 -

EW 0.1583 hk009 SA 320 157 NS 0.1855 EW 0.1972 *~

hkO13 A 160 50 NS 0.1276 -

EW 0.0772 e th014 A 188 81 NS 0.1681 EW 0.1607 i

1

STRONG N0fl0N DATA 8ASE Date E arthquake Lat Long FD RT mb Ms Hw Station C Hn RD Comp Amax Tokachi Oki th020 4 244 124 NS 0.2317 (c ont 'd) EW 0.2003 th029 A 168 84 WS 0.3177 EW 0.2102 1968.05.16 Tokachi Okt AS 41.42 142.58 26 n 6.3 7.3 hk003 A 237 hs 0.1160 EW 0.0924 hkO13 A 79 43 NS 0.1091 EW 0.1159 th014 A 227 138 hs 0.1578 EW 0.1268 1968.05.18 fakachi Oki AS 40.33 143.40 20 th 4.9 5.1 th005 A 203 NS 0.0870 EV 0.0194 1968.05.23 Tokachi Oki At 40.25 142.57 30 th 5.4 6.3 th014 A 92 Nt 0.1353 EW 0.1179 1968.07.01 Saitama 35.08 139.43 68 n? 5.9 5.6 6.1 tkOS6 A 91 S03E 0.0819 W87E 0.1299 1968.07.05 Miyagi 38.43 142.22 44 th 6.0 6.4 th005 A 76 535E 0.0421 W55E 0.0690 1968.08.06 W. Shikoku 33.30 132.38 48 n 6.3 6.5 6.8 cg005 SA 126 hs 0.0840 EW 0.06&3 1968.08.07 Hokkaido 42.97 144.97 68 n 5.6 5.7 hk004 A 75 63 NS 0.0499 EW 0.08 77 1968.10.08 Chiba 35.52 140.15 73 th 5.2 5.3 kt004 A 82 $33W 0.0743 SS TE 0.0341 1968.11.14 Iwate 40.15 142.78 40 th? 5.5 6.0 tbO14 A 92 NS 0.1114 EW 0.0780 1969.04.21 Hyuganada 32.15 132.12 39 th 6.1 6.5 ks002 A 66 $30W 0.0874

$60E 0.1225 1970.01.21 Hokkaido 42.38 143.13 25 th? 6.3 6.7 hkO13 A 51 Long 0.1519 Tran 0.2192 1970.04.01 Iwate 39.75 142.05 75 n 5.8 6.0 th014 A 77 NS 0.1934 EW 0.1649 1970.07.26 Myuganada 32.07 132.03 47 th 6.1 7.0 7.0 ks002 A 51 S30W 0.1393 560E u 1424 ks003 A A4 55 Long 0.0366 Tran 0.0351 1970.07.26 Hyuganada As 32.12 132.10 47 th 6.1 6.0 ks002 A 51 S30W 0.0682 S60E 0.0720 1971.01.05 Aichi 34.43 137.17 44 n 5.6 5.7 kkO26 A 78 NS 0.0950 EW 0.1062 1971.06.13 lbaraki ' 36 .23 140.97 55 th 5.5 5.3 kt001 A 65 hs 0.2529 EW 0.1638 1971.08.02 Erimomisaki 41.23 143.70 45 n 6.6 7.3 hko04 A 201 159 $15W 0.0914 ST5E 0.0776 1971.10.11 Chiba 35.90 140.55 40 th 5.2 5.2 kt050 A 42 $29W 0.0485 561E 0.1726 1972.02.29 Hachijojima 33.18 141.27 50 th 6.) 7.2 kt004 A 2 60 $33W 0.0813

$57E 0.0595 1972.05.11 Kushi ro 42.60 144.93 63 n? 5.5 5,8 hk004 A 71 $15W 0.1456 S T5E 0.0817 l 1973.06.17 Nemuro oki 52.97 145.95 41 th 6.5 7.7 7.8 hk004 A 134 $15W 0.2048

$75E 0.1293

STRONG MOTION CATA BASE Date Earthquake Lat long FD RT mb Ms Mw Station C HD RD Corp Amax e

1973.11.19 Miyagi 18.88 142.15 56 th 6.1 6.5 th033 A 121 NS 0.0538 EW 0.0648 1974.03.03 Chiba 35.57 140.88 56 th 5.6 6.1 kt036 A 63 NS 0.0364 EW 0.1138 1974.07.08 lbaragi 16.42 141.20 45 th 6.0 6.1 kt036 A 87 NS ^498 EW 0.0527 1974.09.04 Iwate 40.18 141.93 52 ss 5.3 5.6 th029 A 67 NS 0.0890 EW 0.0860 1974.11.09 Tomskomal 42.48 141.78 125 n? 6.0 6.5 hk016 SA 126 SOSE 0.0899 N92E 0.0854 1974.11.16 Chiba 35.75 141,25 44 th 5.8 5.6 kt016 A 58 NS 0.0704 EV -0.0932 1978.06.12 Miyagi Ken Oki 38.15 142.17 40 th 6.8 7.5 7.6 th033 A 108 66 NS 0.3204 EW 0.2938 kt014 A 442 390 NS 0.0326 EW 0.0442 th028 A 108 66 NS 0.3252 EW 0.5415 th]14 A 168 120 NS 0.1896 EV 0.1468 th029 A 2 75 229 NS 0.0908 EV 0.0513 th020 A 321 279 NS 0.0397 EW 0.0397 th013 A 179 156 NS 0.0710 EW 0.0734 MEXICO 1962.05.11 Mexico 17.25 99.53 40 th? 7.0 Alameda Cen DF SA 249 N11W 0.0480 N 79E 0.0420 1962.05.19 Mexico 17.12 99.57 33 th? 6.7 Alameda Cen DF SA 262 N11W 0.0390 N71E 0.0310 1962.11.30 Mexico 17.30 99.43 57 n? 5.8 Alameda Cen DF SA 245 N11V 0.0070 N71k 0.0050 1964.07.06 Wexico 18.03 100.77 100 n? 7.4 Ciudad Univ R 245 NS 0.0200 EW 0.0150 Wonoate) HS OF S 256 EW 0.0480 Nonoalco AS OF S 256 NS 0.0310 EV 0.0200 Nonoalco MOS OF S 256 NS 0.0310 EW 0.0310 Ncacelco HP OF $ 256 NS 0.0390 EW 0.0450 1965.06.24 Mexico 17.00 99.60 51 n? 4.6 Acapulco Pel R 64 NS 0.0870 EW 0.0980 1965.08.23 Mexico 16.30 95.80 16 th 7.6 7.4 Ciud2d Univ R 499 447 NS 0.0042 EW 0.0029 Nonoalco AS OF S 505 456 NS 0.0210 EW 0.0095 1965.11.01 Mexico 17.00 99.70 58 n? 4.4 Acapulco Pet R 65 NS 0.0800 EW 0.0570

STRONG M0i!ON DATA BASE Date E ar thquake let Long F0 RT mb Ms Mw Station C HD RD Corp Amax 1965.12.09 Mexico 17.30 100.00 57 n? 6.0 6.3 Acapulco Pel R 77 WS 0.2350 EW 0.1330 Nonoalco AS OF S 262 NS 0.0071 EW 0.0097 1966.04.11 Mexico 17.98 102.75 30 th? 5.5 Infiernillo Por R 102 S68W 0.0230 N22W 0.0230 1966.09.25 Mexico 18.30 100.80 79 n? 5.5 5.7 Infiernillo P R 143 S68W 0.0170 N22W 0.0260 1967.04.20 Mexico 16.86 99.50 76 n? 4.3 Acapulco Pet R 88 NS 0.0480 EW 0.0540 1967.06.07 Mexico 17.10 99.90 47 th? 4.4 Acapulco Pet R 55 NS 0.0650 EW 0.0490 1968.02.03 Mexico 16.70 99.40 9 th? 5.7 5.7 Acapulco Pel R 59 NS 0.0210 EW 0.0320 1968.07.02 Mexico 17.64 100.27 41 th? 5,9 6.5 Acapulco Pel R 105 NS 0.0900 EW 0.0570 Nonoalco AS OF S 238 NS 0.0130 EV 0.0160 1968.08.02 Mexico 16.59 97.70 16 th 6.3 7.1 7.2 Acapulco Pet A 247 154 NS 0.0084 EW ,0,0120 Ciudad Univ R 345 330 NS 0.0150 EW 0.0120 Nonoalco AS OF S 355 341 WS 0.0260 EW 0.0420 Nonoalco NP OF S 355 341 NS 0.0320 EW 0. 04 70 1971.09.05 Mexico 17.09 99.81 50 n? 5.2 5.0 Acapulco sop S 57 N00E 0.1700 N90W 0.2350 1973.08.28 Mexico 18.27 96.60 S4 n? 6.8 7.1 Oaxaca F de Med S 156 N00E 0.2030 N90W 0.1670 Mina ti tlan S 241 N00E 0.0170 W90W 0.0180 Pajaritos S 255 N00E 0.0600 W90W 0.0570 Palacio dl Dep S 302 N00E 0.0180 N90W 0.0170 1974.11.17 Mexico 17.00 100.10 33 th? 4.7 Acapulco SOP S 43 NS 0.1300 EW 0.1160 1975.03.14 Mexico 16,60 93.40 155 n? 5.5 Tuxtla, Gutier. S 159 NS 0.0860 EW 0.0840 1975.12.04 Mexico 16.59 99.50 89 r? 5.0 Oaxaca F de Med S 264 N00E 0.0270 k90E 0.0170 1976.04.27 Mexico 16.43 99.68 33 th? 4.9 Acapulco SOP S 62 N00E 0.0420 N90W 0.0470 1976.06.07 Mexico 17.40 100,64 45 th? 6.1 6.4 Acapulco SOP S 111 N00E 0.0560 W90W 0.0500 1978.03.19 Mexico 17.03 99. 7t- 36 th? 6.4 6.4 Acapulco SOP S 44 N00E 0.3910 N90W 0.8500

STRONG M0!!ON DATA BASE Date Earthquake Lat Long FD RT mb Ms Mw Station C HD RD Comp Amax 1978.11.29 Daxaca 15.77 96.80 18 th 6.8 7.8 7.6 Caxaca F de Med S 122 121 N00E 0.2200 W90W 0.1460 Minatitlan S 322 296 N00E 0.0230 N90W 0.0310 j

Pajaritos S 344 317 N00E 0.0i60 l N90W 0.0120 Puebla S 376 348 N00E 0.0130 N90W 0.0200 Ciudad Univ R 4J8 415 N00E 0.0180 N90W 0.0180 Hospital ABC $ 466 423 N00E 0.0051 W90W 0.0031 Nonoalco HP DF S 466 424 N00E 0.0190 N90W 0.0260 1978.11.29 Daxaca 1st AS 16.16 96.75 33 th 5.3 Danaca F de Med S 1 08 N00E 0.0580 N90W 0.0390 Puebla S 358 N00E 0.0020 N90E 0.0020

( Ciudad Univ R 441 N00E 0.0020 N90W 0.0020 1978.11.29 caxaca 2nd AS 16.18 96.63 22 th 5.7 Caxace F de Med S 102 N00E 0.1000 N90E 0.0910 Puebla S 361 N00E 0.0040 N90E 0.0040 Ciudad Univ R 446 N00E 0.0051 N90W 0.0041 1979.03.14 Cuerrero 17.46 101.46 20 th 6.5 7.6 7.5 Sicartsa CM S 97 71 N00E 0.2600 N90E 0.2990 WOCE 0.2690 N90W 0.3130 Infiernillo CM R 103 73 NOCE 0.1220 N90W 0.1070 Ciudad Altamita S 133 116 N00E 0.1610 N90E 0.1250 Acapulco Pet R 183 142 N00E 0.0280 N90W 0.0350 Acapulco SOP S 184 143 N00E 0.0430 N90W 0.0420 Apat ti ngan S 205 174 N00E 0.0510 N90W 0.0630 Ciudad Univ R 323 295 N00E 0.0170 N90W 0.0140 Alberca Olimp. S 326 299 N00E 0.0310 N90W 0.0380 SAHOP S 330 N00E 0.0340 N90W 0.0310 Loteria Nat Sot S 332 305 N00E 0.0390 N90W 0.0330 Nonoa(co AS OF $ 334 308 N00E 0.0420 N90W 0.0340 Texcoco Chiral. S 349 320 WOOE 0.0330 W90W 0.0230

STR0hG MOTION DATA BASE Date Eaathquake Lat Long F0 RT mb k3 Mw Station C HD RD Corp Amax Guerrero Texcoco Cen. La S 349 321 N00E 0.0420 (cont'd) N90W 0.0490 Texcoco Sosa S '4 327 N00E 0.0560 N90W 0.0530 Puebte S 394 356 N00E 0.0150 N90W 0.0130 1981.09.17 Mexico 16.16 +99.83 17 th? 5.4 San Mar;os S 25 N00E 0.3600 N90W 0.1550 Acapulco SOP S 58 N00E 0.2100 W90W 0.1610 1981.10.25 Playa Azul 17.75 102.25 20 th 6.2 7.3 7.3 Sicartsa CM S 29 20 NS 0.2540 EW 0.2380 Sicartsa CT $ 29 20 NS 0.2120 EW 0.2490 Infiernitto Por R 72 43 $68W 0.1330 Ciudad Altami. S 187 154 NS 0.0450 EW 0.04 00 Acapulco Pet R 2 75 258 NS 0.0090 Chilpancingo S 301 274 N35W 0.0270 NS$E 0.0310 Ciudad Univ R 376 338 N00E 0.c i20 N90W 0.0140 Hospital ABC S 378 340 WOOE 0.0049 N90W C.0084 viveros S 379 341 N00E 0.0160 N90W 0.0160 Alberca Olimp. S 380 142 NS 0.0270 EW 0.0270 Loteria Nat Sot 4 385 347 NS 0.0220 EW 0.0170 Patacto Dr,orte S 397 359 NS 0.0180 fV 0.0220 Ter",0co Chirut. S 4 04 365 is 0.0160 EW 0.0100 Texcoco Cen Lag S 402 364 NS 0.0160 EW 0.0220 Texcoco Sosa S 405 367 NS 0.0310 EW 0.0290 Puebla S 461 422 FPJE 0.0100 N90W 0.0067 Apattingan S 149 127 NS 0.0680 EW 0.0760 Nonoalco As DF S 386 348 NS 0.0130 EW 0.0140 Iv85.09.19 Michoacan 18.18 102.57 16 th 7.0 8.1 8.0 Caleta de Campo R 28 15 N00E 0.1410 N90E 0.1410 La villita R 46 19 NS 0.1230 EV 0.1230 La Unico R S6 23 WS 0.1690 EW 0.1500 Zihuatenejo R 136 28 NS 0.1050 EW 0.1640

STRONG N0fl0N DATA 8ASE Date Earthquake Lat Long F0 Ri mb Ms M.# Station C HD to com Amax Michoacan Papsm a R 189 79 NS 0.1650 (cont'd) EW 0.1190 E. S'<ill R 231 126 NS 0.1040 EW 0.0830 At,yee R 252 148 NS 0.0540 EW 0.0601 El Cayaco S 275 166 NS 0.0418 EV 0.C489 Coyuca R 294 187 NS 0.0428 EV 0.0357 La Venta R 324 216 NS 0.0183 EW 0.0214 Cerro de Piedra R 349 241 NS 0.0275 EW C.0153 Las Mesas R 355 246 NS 0.0224 EW 0.0183 YeltIanguls R 325 217 NS 0.0255 EW 0.0183 El ocotito R 340 232 WS 0.0499 EV 0.0550 Teacalco R 333 251 NS 0.0499 EW 0.0245 CitsLd Univ R 379 290 NS 0.0285 EW 0.0347 Ide ! Patio R 379 290 NS 0.0326 EW 0.0357 Mesa vibradora R 379 290 NS 0.0377 EV 0.0398 Sismex Puebla S 469 380 NS 0.0306

)

EW 0.0336 Tacubaya S 380 291 NS 0.0347 EW 0.0336

${smex viveros S 381 292 NS 0.0449 EW 0.0428 C de Aba. Frig. SA 389 300 NS 0.0826 EW 0.0968 C de Aba. Ofiel SA 389 300 NS 0.0703 EW 0.0215 S. de Ccn y Tra SA 385 296 NS 0.0999 EV 0.1710 flahuac sect >as SA 394 305 NS 0.1390 EW 0.1090 t

flahuac Deport. SA 392 303 NS 0.1200 EW 0.1140 Zacatula RL 122 20 500E 0.2764 I N90W 0.1859 Apat zingan S 104 103 $00E 0.0693 N90W 0.0826 l

l l

l

STRONG MT!CW DATA BASE Date Earthquake Lat Long F0 RT tt Ms Mw Station C HD RD Corp Amax 1985.09.21 Michoacan AS 17.62 -101.81 20 th 7.5 7.6 Zihuatenejo R 39 25 $00E 0.1650 W90W 0.1430 Papanoa R 85 34 500E 0.2570 N90W 0.2260 El suchil R 135 77 500E 0.0887 N90W 0.0744 Coyuca R 199 138 S00E 0.0428 N90W 0.0479 Cerro de Piedra R 250 190 S00E 0.0133 N90W 0.0102 Teacalco R 278 246 NOOE 0.0312 N90E 0.0226 La Union R 40 40 S00E 0.0497 N90W 0.0781 La VitLita R 77 59 S00E 0.0341 N90W 0.0410 Zacatute RL 77 57 $00E 0.0741 N90W 0.0732 Atoyac R 144 98 N90W 0.0761 500E 0.0809 El Cayaco S 169 114 N90W 0.0439

$00E 0.0615 Coyuca R 199 138 N90W 0.0488 S00E 0.0429 xaltianguls R 214 173 N90W 0.0166

$00E 0.0175 La Venta R 216 168 N90W 0.0195

$00E 0.0137 1985.10.29 Michoacan AS 17.58 102.64 20 th 5.6 5.4 5.9 Caleta de cacpo R 59 N90W 0.0390 e00E 0.0303 1986.01.24 Michoacan AS 17.24 101.44 20 th 4.5 Papanoa R 25 N90W 0.0166 500E 0.0214 La Llave R 43 N90W 0.02d3

$00E 0.0147 1986.02.01 Guerrero 16,95 100.14 36 4.1 Ocotillio R 47 N90W 0.0147

$00E 0.0088 xattIanguis R 60 N90W 0.0117

$00E 0.0166 El ocotito R 83 N90W 0.0107

$00E 0.0147 1986.02.07 Michoacan AS 17.65 101.45 20 th 4.9 Zihuatenejo R 20 N90W 0.0137

$00E 0.0058 Papanoa R 60 N90W 0.0088 S00E 0.0107 1986.04.30 Michoacan 18.02 103.06 20 th 6.2 7.0 6.9 Catete de Carpo R 18 N90W 0.0995 S00E 0.0790 Arteaga R 90 N90W 0.0224 500E 0.0283 Fito de Cabatio R 343 N90W 0.0039 500E 0.0039 xaltianguis R 369 N90W 0.0019 S00E 0.0010 y

STRONG Moil 0N DATA 8ASE Date Earthq2ake Let Long F0 RT st Ms Mw Station C HD RD Corp Amax 1986.05.05 Michoacan As 17.77 102.80 20 5.6 5.5 5.9 Caleta de Canpo R 40 N90W 0.0341 500E 0.0517 1986.05. N Guerrero 16.85 98.93 36 5.2 4.2 5.2 Las vigas R 49 N90W 0.0664 l

S00E 0.0809 Las Mesas R 68 N90W 0.0449

$00E 0.0322 Cerro de Piedra R 83 N90W 0.0147 500E 0.0117 El ocotito R 84 N90W 0.0273

$00E 0.0497 Xattfanguis R 95 N90W 0.0088

$00E 0.0C98 1956.06.11 Guerrero 17.86 100.34 50 5.1 La comunidad R 61 N90W 0.0498

$00E 0.0517 La Llave R 89 N90W 0.0244 500E 0.0205 1986.06.16 Guerrero 17.08 99.62 34 4.5 Xattianguis

• 36 N90W 0.0595

$00E 0.1688 Las ' ass R 39 N90W 0.0702

$00E 0.0341 Cerro de Piedra R 48 490W 0.0205 iO0E 0.0195 Coyuca P 61 N90W 0.0117 500E 0.0088 Las vigas R 64 N90W 0.0126

$00E 0.0126 l

1986.11.04 Michoacan 17.79 102.02 15 4.8 La Union R 35 N90W 0.0312

$00E 0.0263 Zihuatersejo R 65 N90W 0.0126 500E 0.0058 PERU 1947.11.01 Peru 30 th? 7.3 7.7 Lima I G RL 260 Long 0.0063 Tran 0.0061 1951.01.31 Peru 50 n? 6.0 Lima i G RL 116 Long 0.0620 Tran 0.0810 1952.08.03 peru 50 n? 5.3 Lima I C RL 125 Long 0.0270 Tran 0.0270 1957.01.24 peru 50 n? 6.2 Lima I C RL 120 Long 0.0100 Tran 0.0090 1957.02.18 Peau 1 >0 n 6.5 Lima I G RL 152 Long 0.0400 fran 0.0340 1966.10.17 Peru 10.92 78.79 24 th 6.3 7.8 8.1 Lima I G RL 206 167 tong 0.4040 Tran 0.2740 1970.05.31 Pwru 9.36 78.87 56 n 6.6 7.7 7.9 Lima I G RL 3 74 255 Long 0.1290 fran 0.1320 1971.11.29 Peru $4 n 5.3 Lima I G RL 138 Long 0.0600 fran 0.0900 1974.01.05 Peru 12.40 76.31 98 n 6.6 Lima I G RL 123 Long 0.0900 fran 0.1100 Zarate RL 122 Long 0.1570

[

1. 77 1ran 0.1720

STRONG MOTION DATA 8ASE Date Earthquake Lat Long F0 RT nb Ms Mw Station C HD RD Corp Amax 1974.10.03 Peru 12.39 77.66 27 th 6.6 7.2 8.1 Lima I G RL 87 70 Long 0.2330 Tran 0.2100 Casa or Huaco RL 92 75 Long 0.2000 Tran 0.2500 1974.11.09 Peru 12.44 77.46 30 th 7.2 Lima I G RL 68 Long 0.0500 Tran 0.0700 La Motina SA 103 Long 0.1200 iran 0.1000 SOLOMONS

• 967.11.14 Long Island 5.46 147.05 194 n? 5.8 Yonki U Ramu SA 243 Long 0.04 73 Tran 0.0471 1968.04.29 Long is 5.39 146.14 31 th? 5.6 Yonki U Ram SA 101 Long 0.0217 Tran 0.0268 1968.06.03 Long is 5.46 146.91 182 n? 5.5 Yonki U Ram SA 226 Long 0.0266 Tran 0.0337 1968.06.17 N. Huon 6.25 146.56 106 n? 5.3 Yonki U Ram SA 124 Long 0.0306 Tran 0.0389 1968.09.16 New Britain 6.08 148.77 49 n? 5.9 Yonki U Ram SA 313 Long 0.0071 Tran 0.0060 1969.01.07 Arona 6.20 146.44 111 n? 5.2 Yonki U Ram SA 122 Long 0.0125 Tran 0.0124 1968.03.10 Unbol Is 5.60 147.29 194 n? 5.7 Lee Base A 232 Long 0.0270 Tran 0.0207 Yonki U Ram SA 253 Long 0.0396 iran 0.0309 1969.06.24 untol Is 5.85 146.79 117 n? 5.3 Lae Base A 153 Long 0.0197 Tran 0.0224 Yonki U Ram SA 154 Long 0.0222 Tran 0.0251 1969.08.02 Lee 6.52 146.92 33 th? 5.2 Lee Base A 40 Long 0.0264 Tran 0.0312 1969.08.03 Danfu 4.25 153.06 59 n? 5.4 Rabout A 113 Long 0.0290 Tran 0.0192 1969.08.22 Sotcren is 7.60 156.00 80 n? 5.1 Lee Base A 999 Long 0.0074 Tran 0.0126 1969.09.07 Taki 6.61 155.74 174 n? 5.2 Panguna R7 179 Long 0.0594 Tran 0.0435 1970.03.28 sougainville Is 6.26 154.6' 63 n? 5.9 Pangma R7 114 Long 0.0828 Tran 0.1221 1970.05.13 Untol Is 5.90 146.79 116 n? 5.0 Yonki U Ram SA 152 Long 0.0260 iean 0.0309 1970.10.31 Utingen 4.93 145.47 42 th 6.0 7.0 Yonki U Ram SA 162 Long 0.0864 iean 0.0734 1971.02.12 Wasu 6.28 146.50 123 n? 5.6 Lee Base A 109 Long 0.0406 Tran o 0340 Yonki U Ram SA 136 Long ..o04 Tran 0.1659 1971.02.13 Vssu 6.06 146.25 114 n? 5.4 Lee Base A 158 Long 0.0103 Tran 0.0090 Yonki U R yu SA 120 Long 0.0797 Tran 0.04 74

STRONG moi!ON DATA BASE Date Earthquake Lat Long FD RT mb Ms Mw Station C HD RD Comp Amax 1971.03.13 Madang 5 . 75 145.39 114 n? 6.2 Yonki V Ramu SA 142 Long 0.0165 Tran 0.0134 1971.07.14 New Britain is 5.52 153.86 43 th 6.0 7.8 8.0 Panguna R7 205 153 Long 0.08 75 Tran 0.1245 1971.07.19 Annanberg 4.90 144.52 75 n? 5.6 Yonki U Ramu SA 232 Long 0.0176 Tran 0.0150 1971.07.26 New Ireland is 4.93 153.18 43 th 6.6 7.9 8.1 Panguna R7 301 251 Long 0.0370 fran 0.0596 1971.08.07 New Ireland Is 3.87 152.04 24 th? 5.1 Rabaul A 49 Long 0.0534 Tran 0.0356 1971.09.14 New 9ritain is 6.46 151.55 22 th? 6.1 6.3 Rabaul A 259 Long 0.0135 Tran 0.0114 1971.09.25 Lee 6.54 146.64 111 n? 6.3 7.0 Lee Base A 119 Long 0.1323 fran 0.1144 1971.10.14 Kokopo 4.58 152.40 25 th? 5.5 Rabaut A IS Long 0.0126 Tran 0.0111 1971.10.28 Buka is 5.57 153.99 107 n? 5.8 6.5 Rabaul A 271 Long 0.0624 fran 0.0695 1972.11.05 Long Is 5.40 146.70 229 n? 5.4 Lee com DW A 274 Long 0.0229 Tran 0.0227 1973.03.22 Wasu 6.16 146.93 102 n? 5.1 Lee Civil Aviat A 120 Long 0.0408 Tran 0.0297 1973.08.13 Marienberg 4.50 144.10 109 n? 5.9 Yonki U Ramu SA 304 Long 0.0263

• Tran 0.0275 Lee CivD Aviat A 419 Long 0.0259 Tran 0.0209 1973.11.25 Medang 5.89 145.53 101 n? 5.0 Yonki U Ramu SA 119 Long 0.00 70 fran 0.0081 1974.03.04 Urboi is 5.88 147.11 67 n? 4.9 Yonki U Ramu SA 148 Long 0.0158 Tran 0.0127 1974.03.25 Saldor 6.03 146.08 110 n? 5.4 Yonki U Ramu SA 113 Long 0.0633 Tran 0.0463 1974.09.20 Saidor 6.20 146.10 105 n? 5.8 Intake U Ramu R 106 Long 0.0155 Tran 0.0323 Yonki U Ramu SA 106 Long 0.1554 Tran 0.2299 1981.12.13 Solomon is 6.39 154.93 50 n? 5.9 6.0 460 B Penguna M R 80 Long 0.0745 Tran 0.09 72 1981.12.13 Solonen is 6.54 154.92 48 n? 5.5 5.7 460 3 Panguna M R 80 Long 0.0262 Tran 0.0636 1983.03.18 Solomon Is 4.&3 153.58 89 n? 7.9 Arawa Town R 281 Long 0.0221 Tran 0.0263 Bato Bridge R 285 Long 0.0346 Tran 0.0329 BvE 80 Pangsna S 285 Long 0.2890 fran 0.2750

SfaowG Mott0N CATA 8ASE LEGEN0 F0 = focal depth (km)

RT e rupture type: th shallow thrust n normal r reverse ss strike slip C = site classification: R rock NL rocklike A alluviin S soit SA - sof t allwiun HD = hypocentral distance (km)

RD e distance to rupture surf ace (km) l l

l l

_ - _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ _ _ - _ - - _ _ _ _