ML20147D392
| ML20147D392 | |
| Person / Time | |
|---|---|
| Site: | Satsop |
| Issue date: | 02/29/1988 |
| From: | GEOMATRIX CONSULTANTS, INC. |
| To: | |
| Shared Package | |
| ML20147D389 | List: |
| References | |
| NUDOCS 8803040009 | |
| Download: ML20147D392 (394) | |
Text
{{#Wiki_filter:. . GEOMATAIX FINAL REPORT Seismic Hazards Assessment for WNP-3, Satsop, Washington Contract No. C-20453 Submitted to WASHINGTON PUBLIC POWER SUPPLY SYSTEM 3000 George Washington Way Richland, Washington 99352-0968 Geomatrix Consultants 8803040009 880229 DR ADOCK 0500 8
1 GEOMATAIX l 1 1 FINAL REPORT 1 l l l Seismic Hazards Assessment for WNP-3, Satsop, Washington Contract No. C-20453 Submitted to WASHINGTON FUBLIC POWER SUPPLY SYSTEM 3000 George Washington Way Richland, Washington 99352-0968 i Geomatrix Consultants
y Ona Market Plaza Spear Street Tower. Suite 71'. , (N5l057S 97 7 GEOMATAIX February 29, 1988
!~ Project 1133A D.W. Coleman Washington Public Power Supply System WNP-3 Site Workman Creek Road Warehouse #1 Elma, Washington 98541
Subject:
Final Report Seismic Hazards Assessment for WNe-3, Satsop, Washington Contract No. C-20453
Dear Mr. Coleman:
Enclosed please find five (5) copies of the subject final report incorporating all of your comments. If you have any questions, please do not hesitate to call. Sincerely, Ke n J. Coppersmith Project Manager dla Enclosures l l l l Geometrix Consultants, Inc. j Consulteng Eno<neers and Earth Scientrsts l
GEOMATRIX SATS0P SEISMIC HAl.ARD ANALYSIS TABLE OF CONTFETS Page
1.0 INTRODUCTION
1-1 2.0 APPROACH 2-1 2.1 Probabilistic Hazard Model 2-1 2.1.1 Formulation 2-1 2.1.2 Basic Components of Seismic Hazard Model 2-2 2.2 Use of Expert Opinion 2-7 2.2.1 Panel Selection 2-8 2.2.2 Expert Interview Process 2-10 23 Aggregation of Expert Opinion 2-14 5 0 SATSOP SEISMIC HAZARD MODEL 3-1 31 Subduction Zone Hazard Model 3-1 3 1.1 seismic source Model 3-1 3 1.2 Ground Motion Attenuation Model 3-7 32 Shallow Crustal Hazard Model 3-8 4.0 ANALYSIS RESULTS 4-1 4.1 Hazard Computation 4-1 4.2 Exceedance Frequency for Peak Ground Acceleration 4-2 4.2.1 Total Hazard 4-2 4.2.2 Subduction Zone Sources 4-2 4.2 3 Shallow Crustal Sources 4-6 4.3 Exceedance Frequency for Spectral Acceleration 4-7 REFERENCES APPENDIX A - DOCUMENTATION OF EXPERT INTERVIEWS APPENDIX B - CHARACTERIZATION OF SHALLOW CRUSTAL SEISMIC SOURCES APPENDIX C - CATALOG EVALUATION AND RECURRENCE RATES APPENDIX D - ATTENUATION RELATIONSHIPS FOR SUBDUCTION ZONE SOURCES
4
,1 GEOMATRlX TABLE.0F CONTENTS (cont'd)
L]ST OF TABLES JTable 2-1. Methodology for Conducting Satsop Seismic Hazard Analysis Table 3-1. . Summary of Expert Interviews LIST OF FIGURES
' Figure 1-1. Site Location Map Figure 2-1. Subduction Zone Hazard Model Figure 3-1. Subduction zone Eazard Model Figure 3-2. Aggregate distributions of 14 experts for ocean slab geotetry Figure 3-3 Aggregate distribution for probability of activity Figure 3-4. Aggregate distribucions for location of rupture Figure 3-5 Plate boundaries and regional seismicity Figure 3-6. Aggregate distrioutions for maximum magnitude Figure 3-7 Aggregate distribution of 6 experts for return period of large interface earthquakes based on paleoseismic data ,
Figure 3-8. Aggregate distributions for moment rate parameters Figure 3-9 Magnitude fregt.ency distributions used in the hazard analysis Figure 3-10. Complete subduction zone sources hazard model Figure 3-11. Shallow crustal sources hazard model Figure 4-1. Percentile haza.N1 curves for peak horizontal acceleration showing contribations of shallow crustal and subduction zone sources to total hazard. Figure 4-2. Contributions to total hazard shown in terms of magnitude, distance, and source contributions. Figure 4-3 Comparison of median hazard curves of individual experts. i Figure 4-4. 15th, Sot h , and 85th percentile hazard curves for individual experts. l
e GEOMATAIX TABLE OF CONTENTS (cont'd) i: LIST OF FIGURES (continued) Figure 4-5 15t h . 50t h , and 85th percentile Curves for interface (solid curves) and intraslab . (dashed curves) sources for individual experts. Figure 4-6. Comparison of experts' median hazard curves with aggregate hazard. ' Figure 4-7 Comparison of aggregation procedures for total hazard from subduction zone sources. Figure 4-8. Contribution of expert-to-expert uncertainty to total uncertainty. Figure 4-9 Contribution of uncertainty in subduction zone attenuation relationships to total uncertainty. Figure 4-10. Contributions of uncertainty in subducting plate geometry to total uncertainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lines) and the 15th and 85th percentiles considering only uncertainty in plate geometry (dashed line) . Figure 4-11. Contributions of uncertainty in source activity to total uncertainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lines) and the 15th and 85th percentiles considering only uncertainty in source activity (dashed line). Figure 4-12. Contributions of uncertainty in source segmentation to total uncertainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lines) and the 15th and 85th percentiles considering only uncertainty in source segmentation (dashed line) . Figure 4-13 Contributions of uncertainty in maximum extent of rupture to total uncertainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lines) and the 15th ggd 85th percentiles considering only uncertainty in maximum extent of rupture. Figure 4-14. Contributions of uncertainty in maximum magnitude to total uncertainty. Shown are the 15th and 85th percentiles con-l sidering all uncertainties (solid lines) and the 15th and 85th percentiles considering only uncertainty in mas:imum magnitude (dashed line). l l
GEOMATRIX TABLE OF CONTENTS (cont'd) LIST OF FIGURES (continued) Figure 4-15 Contributions of uncertainty in recurrence method (moment rate vs paleoseismicity) to total uncertainty. Shown are the 15th and 85th percentiles Considering all uncertainties (solid lines) and the 15 h and 85th percentiles considering only t uncertainty in recurrence method (moment rate vs paleoseismicity) (dashed line). Figure 4-16. Contributions of uncertainty in convergence rate /paleoseismic rate to total uncertainty. Shown are the 15th and 85th per-centiles considering all uncertainties (solid lines) and the 15th and 85 h percentiles considering only uncertainty in t convergence rate /paleoseismic rate (dashed line). Figure 4-17 Contributions of uncertainty in seismic coupling to total uncertainty. Shown are the 15th and 85*h percentiles considering all uncertainties (solid lines) and the 15th and 85th percentiles Considering only uncertainty in seismic coupling (dashed line). Figure 4-18. Contributions of uncertainty in magnitude distribution to total uncertainty. Shown are the 15*h and 85th percentiles considering all uncertainties (solid lines) and the 15th gnd 85th perCO3 tiles Considering only uncertainty in magnitude distribution (dashed line) . Figure 4-19 Peak acceleration hazard from shallow crustal sources. On the left, the 15th 50th and 85th percentile hazard curves for crustal and subduction zone sources are compared. Shown on the right are the median hazard curves for the various shallow crustal sources. Figure 4-20. Contributions to uncertainty in hazard for shallow crustal sources. The solid curves in each plot are the 15th and 85th percentile curves considering all uncertainties. The left plot shows the median hazard curves obtained using single attenuation relationship. In the center and right plots, the dashed curves are the 15th and 85th percentile hazard curves considering only uncertainty in maximum magnitude and recurrence rate, respectively. Figure 4-21. Hazard curves for 5% damped spectral velocity at periods of 0.15, 0.8 and 2 seconds. Shown are total hazard and hazard from subduction and shallow crustal sources. Figure 4-22. 15th. 50th and 85th percentile hazard curves of individual j experts for 5% damped spectral velocity at 0.15-second period. l
i l GEOMATRIX TABLE OF CONTENTS (cont'd) LIST OF FIGURES (continued) Figure 4-23 15'h. 50th and 85th percentile hazard curves of individual experts for 5% damped spectral velocity at 0.8-second period.- Figure 4-24. 15' h . Both and 85*h percentile hazard curves of individual experts for f.% damped spectral velocity at 2-second period.
- Figure 4-25 Comparison of experts median hazard curves with aggregate hazard for 5% damped spectral velocity at 0.15, 0.8 and 2 seconds.
Figure 4-26. Comparison of aggregation procedures for total hazard from subduction zone sources for 5% damped spectral velocity at 0.15. 0.8, and 2 seconds. I m
,,-.,_.m " - -
t GEOMATAIX SATSOP SEISMIC HAZARD ANALYSIS
~
1.0 INTRODUCTION
This report documents a probabilistic seismic hazard analysis carried out for the WNP-3 plant site at Satsop, Washington (Figure 1-1) . The major components of the study consist of the following: e Solicitation of expert scientific opinion regarding seismic sources and particularly subduction zone sources that may affect the site. e Explicit incorporation into the hazard analysis of this scientific understanding and the associated uncertainties. e Inclusion of the present state of knowledge and uncertainty regard-ing ground motion attenuation for both crustal and subduction sources. o Presentation of the hazard results showing relative contributions and sensitivities of the results to the inputs. Probabilistic seismic hazard analysis (SHA) involves assessments of the probability of earthquake locations, sizes, timing, and associated ground motions, which are coupled with a number of uncertainties. Major components of uncertainty at the Satsop site are due to inaccessibility of the subcuc-tion-related seismic sources (precluding conventional fault-specific paleo-
-ceismic studies), the relatively short historical observation period of about 130-200 yr (Heaton and Snavely, 1985), scientific uncertainties in the earth-quake behavior of the Cascadia subduction zone, and uncertainty in the atten-Vation of seismic wave energy generated from subduction sources to a site.
Therefore, in order to produce a SHA that will withstand intensive scienti-fic and regulatory ' _aw, the analysis must capture the present scientific uncertainty in several key tectonic issues such as the seismic capability of the interface between the Jurin de Fuca and North American plates. A complete SHA for the Satsop site must incorporate all known and potential c:ismic sources that may affect ground motions at the site. This includes potential sources related to subduction. The perception and understanding of subduction in the Pacific Northwest has evolved in the past several l
/ l GEOMATAIX y 1-2 l years. Past studies of historical seismicity have led to the conclusion , that the interface between the Juan de Fuca and North American plate is , either no longer undergoing differential slip (i.e. , subduction has ceased) or subduction is occurring aseismically. Improved instrumental seismicity ' coverage in recent years as well as re-examination of older historical earthquakes has confirmed a virtual lack of thrust-type earthquakes that would be celated to interplate displacement. At the same time, studies of high resolution seismic reflection data have shown clear evidence of into Quaternary and Holocene deformation in the young, water-saturated sediments of the outer accretionary wedge offshore, suggesting that plate convergence is still continuing. Offshore and onshore geophysical studies, including , the Lithoprobe project through Vancouver Island, have demonstrated that extremely high sedimentation rates have served to bury the Juan de Fuca , plate, and, because the plate is relatively young and buried essentially all the way to the Juan de Fuca ridge, the sediments are probably serving to . thermally insulate the plate and are themselves heated up. , 1 s The confirmation of historical quiescence of the plate interface and . increased understanding of other aspects of the Cascadia subduction zone have led to a variety of interpretations of the seismic behavior of the e plate interface. The extremes of these interpretations indicate that: * -
- 1) the historical record is characteristic of the long-term behavior of the ,
zone and slippage occurs aseismically, or 2) the historical record repre- -( sents an interseismic period between the occurrence of large interplate thrust earthquakes. A variety of behaviors between these two extremes have also been proposed. At present, the Cascadia subduction zone appears to be a unique zone in its
- combination of the youthfulness of the oceanic plate, its relatively low rate of convergence, and lack of interplate historical seismicity. An impor-tant issue is whether this behavior is merely a function of cur short period of observation or due to true differences with other subduction zones. ,
To carry out the Satsop SHA, a basic philosophy has been adopted. Those elements of the SHA that are not amenable to resolution in the near-term and
- with limited available resources should be defined by the current spectrum ,
GEOMATAIX , 1-3 , 1 of scientific thinking, as represented by expert opinion. Examples of these r' elements are the seismic activity of the plate interface, the amount of coupling oetween the plates, and the likely locations of earthquake rup-tures. Those elements of the SHA whose unc ertainty can be reduced signifi-cantly by data collection and analysis efforts are treated differently. In these cases, the Supply System has has carried out its own studies and the results supersede previous, more limited, studies. Important examples here are the empirical and numerical studies carried o"- ' stimate ground motions. It became c: car in the pro 6t as that existir.g published attenuation relationships do sat include important recent earthquake recordings (e.g. , Mexico and Chile), are not appropriate to the Satsop rock site conditions and are nt' 9ppropriate to the source-to-site distances of interest. As a result, net .elationships were developed for use in the hazard analysis (Appendix D). The seismic hazard methodology presented herein was developed and implemented after a careful consideration of recent major SHA's involving expert opinion (e.g. , those for the eastern U.S. by Lawrence Livermore National Laboratory and the Electric Power Research Institute) We have attempted to build on the strengths of those previous studies in dealing with experts and utilizing .* their opinions. In addition, we have added our own features to address the differences in the situations between this study and the previous studies. , A number of innovations were developed and implemented during this study such as: conducting one-to-one interviews with the experts; documenting the basis for the experts' responses; and dividing the tazard model into several compo-nents to allow for more complete representation of uncertainty. A detailed discussion of the methodology is given in the Section 2 of this report. Because of the large uncertainties associated with a hazard astessment of this kind, extensive sensitivity analyses were carried out to identify those elemerts of the hazard model that are contributing most to the hazard results and to the total uncertainty in the results. These are useful for identify-ing the relative significance of various components of the hazard model and determining the effect of uncertainty 1. these components on the final results. -
oo~- GEOMATmx 1-4 In sum, we believe we have captured the present scientific and tectonic understanding of the seiraic environment in the Satsop site region. The results provide a complete expression of the hazard at the site and the associated uncertainties. As such, they provide a solid basis for evaluating the seismic hazard at the WNP-3 site. 1
7 l 1 & GEOMATAfX l g/ v v g 'r u SAN JUAN ISLANDS '
/ ,
GBELLINGHAM NORTH s (Q
@%g-ob CAPE FLATTERY 4/7s VICTORIA or j N 04N o rFUCA CASCADES h PORT ANGELES k k EVEREU
+ + OLYMPIC ~ ' eSN0HOMISH g adEB MOUNTAINS PuGET sound 4 OBOTHELL o##gs[ 4 4 gb ,
SEATTLE X SN00 HOOD CANAL GAUBURN v KITSpP PENINSULA WISHKAH R.N ' g HOQUI AM . SITE OLY M PI A GRAYS HARBOR W ABE RDEEN ifCHEHAllS HIVER WILLAPA CENTR ALI A E
*gst R/vg c
% HILLS Id
' s Z tY 0 50 Miles a f I Figure 1-1 SITE LOCATION MAP (from WPPSS, 1982)
L d I
;'; s
$ 2 =l=
anomarna
~2 iO DAPPROACH'
\'
' 2 .' l . Probabilistic Hazard Model
~2.2.1 Formulation. Seismic hazard is expressed as L the probability that a ground motion parameter, Z (such as peak ground acceleration, spectral -
.valocity.. etc.) will exceed a specified level, z, during a specified time
{' period..t. 'The probability of exceeding a ground motion level at a site can be estimated from the inequality: P(Z)zlt)sv(z)t (2-1) where v(z) is the frequency or rate at which the level'of ground motion
'p;rameter Z exceeds z at the site. When dealing.with the low probability
' levels of interest in this application, v(z) t provides a good, conserva-tive estimate of the hazard. The parameter v(z) is obtained from the general expression:
.m" ,a V(z) = a(m ,)0 f(m) f(r) P(Z)zlm,r) dr im (2-2)
O'
- where a(m ) is the frequency of earthquakes above a minimum magnitude of engineering significance, m ; f(m) is the probability density function for evznt size between a and a maximum event size, m"; f(r) is the probability d:nsity function for distance to earthquake rupture; and P(Z>zlm,r) is the probability that, given a magnitude a earthquake at a distance r from' the site, the ground motion exceeds level z.
~
The probability functions contained in Equations 2-1 and 2-2 represent the i
. rtndomness inherent in the natural phenomena of earthquake generation and saismic wave propagation. For the Cascadia subduction zone one is faced i 'with considerable uncertainty in selecting the appropriate models and model
- ptrameters required to apply Equation 2 arising from limited data and/or citernative interpretations of the available data. The approach used in this study explicitly incorporates these uncertainties into the analysis to
- c ress their impact on the estimate of seismic hazard.
l l
GEOMATAIX 2-2' I 1he uncertainty in modeling the natural phenomena was incorporated into the hazard analysis through the .use of logic trees. The logic tree formulation for seismic hazard analysis (Power et al, 1981; Kulkarni et al, 1984; Youngs et al,1985; EPRI,1986; Coppersmith and Youngs,1986) involves specifying discrete alternatives for states of nature or parameter values and specifying - the relative likelihood that each discrete alternative is the correct value er state of the input parameter. The parameter values and their relative likelihoods are usually based on subjective judgment because the available data are too limited to allow a deterministic assessment or a formal statistical analysis. Figure 2-1 displays the general logic tree format used to represent the geismic hazard model alternatives. The logic tree is laid out to provide a progression from general aspects / hypotheses regarding the characteristics of coismicity and seismic wave propagation in the region to specific input parameters for individual faults and fault segments. The motivation for development of the various levels of the logic tree are discussed below. 2.1.2 Basic Components of Seismic Hazard Model. The seismic hazard model is divided into a number of components, most of which relate to the tectonic characterization of the potential seismic sources. The fourteen experts were responsible for characterizing these compor.onts (see Section 3). To hslp in understanding the hazard model, each component is discussed below. Crustal Geometry Each expert was asked to provide his interpretation of the three-dimensional geometry of the Cascadia subduction zone. Each expert provided a cross-sectional sketch of one or more possible geometries showing the location of the Juan de Fuca slab and the North American plate. Along-strike variations in geometry (such as changes in slab dip) are also specified. The most common basis for estimating the possibic position of the oceanic slab was the distribution of hypocenters of the deeper seismicity beneath Puget Sound, coupled with worldwide analogies to other subduction zones (e.g. ,
q
.- M GEOMATAlX 2-3 expected dapth required for magma generation may mark the depth to the slab beneath the Cascades).
Potential Seismic Sources The experts were asked to identify all potential seismic sources that could exist in the western Washington region. It was stated to the experts that shallot crustal potential seismic sources would be considered elsewhere in the study, but they were asked to identify those potential sources that might be present in the shallow crust (upper 20 km) but might not have a surface expression, Potential sources were not necessarily limited to those tectonic features that have been associated with seismicity during the historical period. Areal source zones as well as tectonic feature-specific sources could be identified. In general, all of the experts identified two potential seismic sources: 1) an intra-slab source whereby earthquakes are generated within the subducted oceanic slab, and 2) an interface source c whereby earthquakes are generated at the interface between the Juan de Fuca and North American plates. Probability of Activity Each seismic source is associated with an expression of the probability that it is active or seismogenic. Activity is used here to mean that the source is capable of generating tectonically-significant earthquakes. In general, for the subduction-related seismic sources, significant tectonic earthquakes were judged by the experts to be larger'than about magnitude 6. (Note that this is not the lower bound magnitude for integration of the hazard calcula-tion, which is discussed in Section 4.) The probability of activity is assessed to be a function of the tectonic role played by a potential source in the present stress regime, and unless that role is expected to change, the probability of activity is independent of timo. Thus, "activity" is a binary state (i.e., either "yes" or "no"), and the probability of activity i is an expression of the likelihood that the potential source lies in an active state or not. Not included here is the likelihood of earthquake recurrence during any specified time period. This is a function of the l recurrence rate, which is a separate component of the seismic hazard model. l
i f GEC)MAMX 2-4 L . The experts considered the probability of activity of the subduction sources (i.e., the intra-slab and interface) to be independent, based largely on observations of subduction zones worldwide. Location of Ruptures To model the seismic sources for the hazard analysis estimates are made of the three-dimensional location of ruptures for each seismic source. This is an assessment of the geometry of the surface over which future ruptures will occur. For example, an intra-slab seismic source might have the following rupture location characteristics: 1) in cross-section, earthquakes will occur in the upper ten kilometers (brittle portion) of the oceanic slab, 2) the downdip extent will be to depths of about 70 km and updip to the first bend in the slab offshore, 3) the earthquakes larger than magnitude 7 will occur at depths of 50 to 70 km, 4) in map view, the intra-slab seismicity will follow roughly the coastline to accommodate the "corner" near the Canada / U.S. boundary and will end at the Nootka fault on the north and the Blanco fracture zone on the south, 5) in map view, the relative frequency of earthquakes in the intra-elab source will spatially match that observed in historical seismicity (i.e. , higher concentration beneath the Puget , Sound / Georgia Strait region than to south or north). Another aspect of rupture locations that may be specified is that of seg-mentation of the source. This assessment allows for the possibility that future maximum earthquakes may not rupture the entire maximum dimensions of the source. Possible rupture segment boundaries may be identified and the probability that they will serve as rupture boundaries can also be assessed (o.g., the tear fault in the downgoing slab at a specified location has a 40% chance of serving to stop rupture coming from either direction on the plate interface). As with all components of the seismic hazard model, alternative hypotheses for the location of ruptures may be given and each can be associated with a relative weight or credibility.
GEOMATRIX 2-5 Maximum Earthquake Magnitude l' Each seismic source is associated with a maximum earthquake magnitude that ccrves as the upper-bound constraint on the recurrence relationship for that cource. Maximum magnitudes were often directly assessed by the experts based on the largest historically observed magnitudes or by analogy to- other cubduction zones. For the plate interface source, many experts indicated that the rupture dimensions, specified previously as part of locations of rupture, provided a reasonable basis for estimating maximum magnitudes. In these cases, the magnitudes were calculated by the elicitation team using the experts' rapture dimensions and the relationship between magnitude and rupture area by Wyss (1979). In general, the magnitudes determined in this canner ranged from 8-3/4 to 9-1/4. In a few instances, the experts specified d;veloping a maximum magnitude estimate from the relationships between plate age, convergence rate, and observed magnitude (Ruff and Kanamori, 1980), resulting in magnitude estimates of about 8.3 Uncertainty in the maximum magnitude estimate is expressed by the experts in t:rms of a range of values, a preferred value with associated bounds, or discrete values each associated with a relative weight. Convergence Rate Convergence between the Juan de Fuca and North American plates is considered p;rallel to the relative plate motion direction. The convergence rate is the relative rate between the two plates, derived in most cases from the cbsolute plate velocities of each plate. Various investigators have shown th:t the convergence rate at the Cascadia subduction zone has been decreasing ov:r the past few million years (e.g. , Riddihough,1984) . Because we are mo2t interested in the present behavior of the plate boundary, the experts were asked to give convergence rate estimates that are representative of contemporary rates. Of ten, a broad range of estimates was given, reflecting considerable uncertainty in the rate.
\
GEOMATRIX 2-6 Seismic Coupling Seismic coupling (a) is defined for this study as the percentage of the total convergence rate that is expected to be released as seismic energy (i.e., ratio between seismic moment rate and convergence rate). Coupling can be estimated from the historical record or from an assumed model. For example, the historical record in the Pacific Northwest shows that virtually no thrust earthquakes lcrger than magnitude 5 have occurred on the plate interface (i.e. , the historical seismic moment rate is very low) . If this behavior is judged to be representative of the longer term behavior then
.ccismic coupling would be very low. However, the historical quiescence may be interpreted by some to be the result of a short observation period and actually representative of interseismic quiescence. In this case, the i c:ismic coupling might be assessed to be high (i.e., a close to 1.0). A wide variety of approaches might be considered in arriving at a seismic coupling estimates ranging from detailed studies of the mechanical / thermal properties of subducted sediment to analogies to similar subduction zones worldwide.
Earthquake Recurrence Earthquake recurrence or the frequency of occurrence of various magnitude errthquakes is assessed for each seismic source. The experts were asked to cpecify the preferred method (s) for estimating recurrence including: the use of historical seismicity record, use of seismic moment rate, or geologic dtta regarding recurrence intervals. To use the historical seismicity record, the three-dimensional area for gathering recurrence statistics is cpecified as well as the area over which these recurrence rates are assessed to apply. For example, the deep seismicity zone () 30 km) beneath Puget t Sound may be specified to define a recurrence rate per square kilometer. This rate may in turn be said to also be appropriate for the source at this d:pth north and south of the seismicity zone. Th3 seismic moment approach to recurrence was used in many cases to define j tha recurrence for the plate interface. The convergence rato is multiplied l by the seismic coupling (a) to arrive at a seismic slip rate. To arrive at I I I
- , ~~
c, GEO M ATA D< 2-7 p l a seismic moment rate, the slip rate is multiplied by the total area of the seismic source (defined by the assessed source geometry and location of rup-tures) and an assumed rigidity (3 x 10 dyno/cm ), The use of seismic moment rate to define recurrence has become standard practice for crustal faults (e.g. , Anderson and Luco,1983; Youngs and Coppersmith,1985a) and appears to be supported by observations of seismic moment release observed for several subduction zones (Peterson and Seno,1984). To use the result-ing seismic moment rate, a recurrence distribution model must be specified that indicates the relative frequency of earthquakes of various magnitudes. The models considered by the experts included: 1) an truncated exponential magnitude distribution based on the familiar form log N = a-bM, 2) a charac-teristic carthquake model of the foru given by Youngs and Coppersmith (1985b) and 3) a maximum moment model as described by wesnousky, et al. (1983). Some ecperts used geologic evidence for the recurrence intervals between large earthquakes. Typically this type of data does not provide strong constraints on the size of the earthquakes giving rise to the geologic effect. For this study, we assume that the recurrence intervals apply to magnitudes within one-half magnitude of the maximum. The recurrence dis-tribution model then defines the recurrence rates for smaller magnitudes. 2.2 Use of Expert Opinion Several key tectonic issues (such as the seismogenic capability of the plate interface, the degree of seismic coupling between the plates, earthquake recurrence rates, and the like) are not amenable to resolution within the time frame of this study. Therefore, a decision was made to capture the present understanding and opinions regarding these issues through the use of experts most familiar with them. In deciding on an appropriate methodology for eliciting expert opinion, careful consideration was given to the strengths and weaknesses of recent large SHA's involving expert opinion (EPRI, 1986: and LLNL, 1985) because
- the level of uncertainty and the potential for short-term resolution of the issues is comparable.
l l f
r f-GEOMATRIX 2-8 By considering the attributes of these previous studies as well as the cpecific requirements for a hazard assessment at the Satsop site, a method-clogy was developed for utilizing expert opinion. The key attributes of the methodology and the purpose for each are given in Table 2-1, and further discussion is given below. 2.2.1 Panel Selection. The panel selected for the Satsop SHA was intended to span fields of expertise that cover in aggregate the entf.re range of the htzard model components (e.g., crustal geometry, seismic capability, conver-gence rate, maximum earthquakes, and earthquake recurrence rate). In addi-tion, it was felt to be desirable to attain a balance of disciplines per-tcining to the topic of subduction tectonics and seismicity (e.g. , geolo-gists, geophysicists, seismologists, laboratory experimentalists, empirical analys ts , etc. ) . The above considerations required that the total number of cxperts be relatively large (14) for studies of this kind. A primary consideration in the selection of experts was that they must have had some experience with the Cascadia subduction zone or allied experience with other subduction zones having similar characteristics. For example, a cuitable expert might be a geologist who is carrying-out analytical studies cf the seismic behavior of subduction zones that are subducting large amounts of sediments and who is familiar with the accretionary wedge characteristics of the Cascadia zone. Because a large part of the uncertainty associated with the Cascadia zone stems from determining its "uniqueneas" relative to other subduction zones, it is important that the experts be familiar with this zone in order to provide as site-specific a hazard model as possible. Finally, some of the experts have published opinions regarding the seismic behavior of the Cascadia subduction zone. To the extent possible, we cchieved a balanced cross-section of opinion in selecting the panel members. l l l 1 i 1
emm - - - OEOMATRIX 2-9 4,
, TABLE 2-1 METHODOLOGY FOR CONDUCTING SATSOP SEISMIC HAZARD ANALYSIS Attribute of Methodology Purpose o Large number of experts (14) e Spectrum of scientific opinion captured o Experts represent wide variety of e Incorporate full range.of disciplines perspectives and data sets i
c No single expert required to address e Avoid encouraging expert to go all aspects of hazard model beyond area of expertise o Experts provided with background e Encourage a uniform minimal information and topical reference level data base; provide a list- focus on key issues to SHA o Experts interviewed individually e Allow for free expression of and opinion not associated with opinion; highly focused expert by name discussion o Basis for decisions given and e Allows for a technical evalua-documented tion of the responses in terms of the scientific issues driving i thinking o Inter".ew senaaries provided to e Ensure accuracy and provide ; each expert for review opportunity to change opinion i upon reflection o Unzard model developed as components e Model is clearer to experts, allows for sensitivity studies o Full inclusion of uncertainty e Leads to more complete expres-expressed by experts sion of hazard;
r GEOMATAfX 2-10 t k' Using the above criteria, a balanced list of experts plus alternates was crrived at. The individuals were then contacted and all agreed to parti-cipate, with one exception (who felt that he had little to contribute to this study). The 14 experts were the following: Expert Affiliation John Adams Canadian Geological Survey Mark Cloos University of Texas, Austin Ron Clowes- University of British Columbia Daryl Cowan University of Washington Robert Crosson University of Washington Greg Davis University of Southern Califorria Thomas Heaton U.S. Geological Survey Thomas Hirde Texas A & M University Hiro Kanamori California Institute of Technology Vern Kula Oregon State University William McCann University of Puerto Rico Thomas Owens University of Missouri Robin Riddihough Canadian Geological Survey Garry Rogers Canadian Geological Survey In the course of discussions with the experts regarding participation in the project, two individuals requested that their responses not be attributed cpecifically to them by name, lest their response be construed as repre-centing an official position of their institutions. To guard the privacy of all the experts, it was dec_ded to not associate particular opinions with cxperts by name. Therefore, henceforth in this report, the experts are indicated by number only (e.g. , expert #1) . 2.2.2 Expert Interview Process. The elicitation of expert opinion for the Satsop hazard analysis occurred through a two-part interview process. The first interview (described in detail below) occurred in the summer of 1986. A second interview, conducted via telephone, was carried out in the fall of
GEOMATF41X 2-11 1987 The second interview wts a follow-up to the first and was held in light of feedback from the first interview (see Appendix A) and to gauge any change in opinion over the ene year time period. Several weeks prior to the interviews, each expert was sent an introductory letter (included in Appendix A) that contained the following: 1) a summary of the purpose of the study,
- 2) a review of the methods to be used to elicit expert opinion in the inter-view, 3) a list or questions that are likely to be asked, and 4) a biblio-graphy arranged topically. All of the cited references were made available to the experts at their request. The purpose here was to be sure that all of the experts were made aware of any references that they might not otherwise have come across.
During June and July,1986, each of the experts was interviewed in his office. Each interview lasted about five hours and was attended by an elici-tation team consisting of Dr. Kevin Coppersmith (geologist) Dr. Robert Youngs (hazard analyst), and Dr. Ted Habermann (seismologist) . Dr. Robert Winkler (expert opinion elicitation expert) attended the first interview and provided guidance throughout the interview process. Drs. Coppersmith and Youngs were heavily involved in the development and implementation of the EPRI hazard study (EPRI,1986) . The interviews opened with a short introduction to remind the expert of the purpose of the meeting and to set guidelines for the interview process. Examples of these introductory comments are given below: e We are using expert judgements because there exist limited or only indirectly relevant "hard" data, e We are using probabilities to quantify your judgements and uncertain-ties, but no high-powered statistical treatment is necessary, e The seismic hazard problem has been broken into component parts to help you think about the problem. e You may feel more comfortable abcat some parts of the hazard model than others - that's ok. You may defer on some aspects if you desire.
- f. . .'
GEOMATAIX
.l. .l' "r , 4 : , *.- l 2-12 .
- -=.
4 i M..6.w.t g N .w e What we are attempting to clicit are: 1) one or more interpretations . n i.' ?..
- regarding each component of the hazard model, 2) the probabilities >e !CE 4 associated with each interpretation, and 3) the qualitative basic for S .W ,: : f ..
f7 - your interpretations. ((*. N. .,<n.:-
, .ys.. ..q_..
.. (
e Try to consider all possible theories and hypotheses, not just "most P'? < . . 4 J- likely" scenarios. - FMeY h A* $ffh e Keep the uncertainties in mind in your responses. _ _ . s_ _ i e Do not mix values and judgement - we just want your scientific judge-j
- g ment, not your opinion regarding consequences or risk.
k gghp -p ( g e You will have an opportunity to review and revise our summary of this interview. (.h.o.h . {.{. G .?l{.j{ v= W ..N :. g . s.9 - [ Included in the intr Muctory comments were descriptions of the major elements ak. 4 VUMW' 8- .4 of the seismic hazard model and planned methods for incorporating uncertain-ty, such as the use of logic trees, p -.
- -- T.uhi!b: h. m .'
,; During the interview, individual components of the hazard model were dis- :( ..1 l-cussed separacoly (e.g., geometry of sources, convergence rate, seismic ,'yy.. 3 @ coupling, etc.). j.' 5 ; jy.-
3 M e % ." D ) ti The experts were encouraged to bring and use any data, maps, reprints, and murmur?mma j fy >.y T the like, that they felt were appropriate for explaining their reasoning.
- r irhs;U,7
. q. J i .
j Members of the elicitation team recorded in writing the basis for the inter- - (;.Q .f wg- ..: 3 ; pretations (e.g., "the dip of the oceanic slab is 11, based on seismic 7" ? a refraction data, instrumental seismicity, etc."). To define the uncertainty l i- in their estimates, the experts were asked for a range of possible values ts - r N ;N :W; j;- (e.g., "my preferred value is 11,, but the dip could be as high as 20, or as i*- # o . .a y low as 9,"). Some experts preferred to develop a probability distribution ty.F.g .
, v... ., .
[ ". . . e that expressed the uncertainty in their responses (e.g., a normal distribu-E_ tion centered on 11*). Capturing this uncertainty is important to fully k.. u .. . characterize the uncertainty in the hazard assessment. J-
.;~ ..
Asking the experts to provide the basis for their assessments proved to be ' :- - '-[ extremely valuable to understanding the reasoning process involved in arriv-ing at a judgement. It is clear that different types of data are viewed as g .] _ .[ . 9
'C
. .'.. - f Mip4
- .9 . M
- .f;3 t uk
GEO M AT AIX 2-13 1 being more diagnostic than others depending on the expert. For example, geologic evidence for large earthquakes and their recurrence intervals based on submarine turbidites was believed to be highly conclusive by some experts and merely suggestive or inconclusive by other experts. We believe that asking for their basis helped the experts to sort out the various data sets and to arrive at conclusions. By carefully treating the uncertainty, they were able to qualify the credibility of various interpretations. In going through the elements of the seismic hazard model on a component-by-component basis, experts were ellowed to decline to specify those components that they felt uncomfortable with. It is not surprising that several experts exercised this option given the wide range of fields of expertise involved in a hazard model of this type. In general, however, most experts were able to characterize nearly all components of the model and half provided complete hazard models (see Section 3) and to quantify their uncertainty and conri-dence in each component. Overall, the expert interviews were a highly successful part of the project. The experts were well-prepared, open, and willing to express freely their opinions.and uncertainties. Following the interviews, written notes were distilled, typed, and sent back to the experts for their review and revision. In general, the changes made were very minor. The interview summaries are given in Appendix A. As a follow-up to the 1986 interviews, a second set of interviews were car-ried out in the fall of 1987 via telephone. (The hiatus was due to a change in the focus of the WNP-3 program toward ground motion estimation methodolo-gies.) The second interview had two purposes: 1) to assess whether opinions would change given feedback (discussed below) regarding maximum magnitudes and recurrence, and 2) to allow the experts an opportunity to update their assessments. The feedback consisted of the following: during the course of the interviews, many of the experts had provided source characteristics related to maximum magnitude (e.g. , segmentation, interface width) and to earthquake recurrence maximum magnitude or recurrence values. Therefore, in
GEOMATAIX 2-14 - order for each expert to see the implications of their assessments, the calculated maximum magnitudes and recurrence relationships were provided to each expert prior to the second interview (Appendix A contains this second information package). The interview stepped through the various components of the model as was done the first time, allowing the expert to make any changes that he felt were appropriate. The results were then documented and checked by the experts for accuracy and are given in Appendix A. It is this updated set of expert opinions that are used to calculate the hazard at the Satop site and are discussed in detail in Section 3 e 23 Aggregation of Expert Opinion The product of the expert interview process was a set of interpretations of the present scientific knowledge concerning the seismic potential of the Cascadia subduction zone. Incorporation of this information into the eva'lua-tion of the seismic hazard required aggregation of multiple interpretations into a hazard model that, ideally, represents the combined information of . the experts and reflects the current level of uncertainty. The approach used in this study was one of "mechanical" aggregation in which the experts' probability distributions are mathematically combined to arrive at a single - distribution representing the uncertainty in the seismic hazard. As the assessment of expert opinion was structured to provide information on , multiple components of the hazard model the aggregation of these assessments could be done either at the component level or at the hazard level. Compo- i nent level aggregation would provide a single composite hazard model rather than a set of 14 hazard models. However, because of the sequential nature of the interviewing process, the experts' responses for some questions are conditional on their earlier responses. For example, there was some varia-tion in the seismic sources considered, and later questions are conditional on these sources. When two experts gave different tesponses on certain - questions, aggregating for some later questions would amount to aggregating judgements that were conditional on different sets of assumptions. Therefore, the primary approach to aggregation used in the study was to develop hazard
[ 2-15 GEOMATAIX g
'c. ras
.' g? g'g ')
j . models for each expert and aggregate the rest.lting distributions for seismic hazard. -M (* <-",;. . " ?n .~ F Aggregation at the hazard level raised the problem of experts providing [ incomplete hazard models because they felt uncomfortable making assessments
} "[-
b -
@O $l of particular aspects of the problem. It was judged that the most appro- '
h @( priate way to fill these "gaps" in an individual expert's hazard model was sb j g g .,, to use an aggregate of the assessments of the other experts. Thus the $ %gl( hazard models of some experts represent, to some extent, a composite hazard '
'I 'h g model.
Accordingly, a secondary analysis was conducted to extend the "gap y filling" procedure to full component level aggregation to see how the final hazard estimates differ and to learn more about the state of information y regarding each component.
~p fi .h Ei The combining rule for mechanical aggregation used in this study was a
{ simple average with equal weights, both at the component level and at the
- _ final hazard level. The simple average was used because there is no compel-b ling reason to assume that one expert is a "better assessor" than another without attempts to calibrate the experts. The simple average also tends to .
1 preserve the full distribution of interpretations made by the experts. k y j y ,; x q } p' . ' n - i- *
;.3 -
1 T
- [ - Sh w ?$4 1 r
m. y ' .".yw Ni-- ii a E 7 t s[.. , m
,, da ji. ; L{,*
h ,.
~, e, 1.l l4 W. 'I e'
9 V',' _f-
. .e
$ ,".~.'M.
3 ,. d un%x w'u
l r r 6 c t n d 5 c t n hEaaX id e e i e e deE t t i n m n r o m u6 e n Wc o e n t c o t o o W o a W n g p r o x p r a x x a x a iN W o E h C W
\
E [Cf h M p -. c 2 3 4 A W 0 0 0 / a r N e S
\
r r r r y y r r y c /y /y /y ig 0 0 0 o m m m 0 0 0
. de t m m m 3 4 5 oo / / /
GR 0 2 3 0 0 4 1 1 1 r o \ l [f - e e t e d c n lg M o rg n e r e m o do o d cg u e W G r R a z a H m 5 u 5 7 e im 8 8 9 n x o o Z Wy s n o i t e r w c t u e o r nt d u e p u i w r a d t n b ExRf u o S
/ f d
e d 1 n t n e
- o e t
n 6 e 2 6 m g m e g E s e R n S U U G I F e e f oty i t v w eM t t A c A, ac : S A t b
/T e s a c e S a r c f e r - r h u o e t o
o t r t n S I n I A B y e l l t r e e d o d o m Moe M M G
F GEOMATAIX 30 SATSOP SEISMIC HAZARD MODEL This section of the report summarizes the components of the seismic hazard model for the Satsop site. Included here are the "inputs" to the seismic hazard calculations, the results of which are given in Section 4. The potential future earthquake sources of significance to ground motions at the site can be divided into two groups: those associated with the Cascadia subduction zone and those located in the shallow crustel portion of the North American plate. Seismic hazard models were developed separately for each group of sources as discussed in Section 31 for subduction zone sources and 3 2 for shallow crustal sources. 31 Subduction zone nazard Model The various seismic source characteristics that were defined for potential subduction zone earthquake sources is given in logic tree format developed to model the subduction zone sources is shown in Figure 3-1. The logic tree progresses from an assessment of the geometry of the subducting slab to assessments of specific analysis parameters for individual sources. The assignment of parameter values and their relative likelihoods for the subduc-tion zone sources was based on the inputs from 14 experts (Section 2.2.2). The individual assessments of each of the experts are documented in Appendix A and are summarized in Table 3-1. As part of this study, site-specific ground motien attenuation relationships were developed based on an analysis of strong motion data including near-field data recorded during several recent earthquakes; this analysis is presented in Appendix D. 3 1.1 Seismic Source Model. As discussed in Section 2 3, the hazard model could be developed by either combining the assessments of all the experts for each parameter (termed "component level" aggregation) to arrive at an aggregated assessment over all experts for each parameter, or by developing O hazard model for each expert based on his individual assessments and then aggregating the computed hazard from the 14 models (termed "hazard level" aggregation). Because many of the assessments of the components were made conditional on other responses (e.g., an assessment of the maximum extent of
y-, GEOMATFHX 3-2 interface rupture may be conditional on the assessments made for the geometry of the oceanic slab) the hazard level aggregation approach was judged to be more appropriate. However, as indicated by the blank spaces in Table 3-1, not all of the experts provided a complete set of assessments for all compo- j nents. Where an individual expert declined to assess a particular compcnent. l a distribution of parameter values based on the assessments of the other I experts was substituted to complete the hazard model. A supplemental hazard analysis was performed using a composite hazard model constructed from the aggregated distributions for each component presented below. The component-level distributions for individual parameters w9re developed on a marginal basis. For example, if an expert has two altc uative oceanic slab geometries and for each, he assessed a distribution for maximum magni-tude, then his marginal distribution for maximum magnitude would be a combi-nation of the two distributions, each weighted by the probability that the particular slab geometry on which it is based is the correct geometry. The marginal distributions of the experts that made assessments of maximum magnitude could then be aggregated to form a single marginal distribution that could be used to fill gaps in the hazard models of those experts that did not assess maximum magnitude. The assessments made for each component of the hazard model are summarized below. Included here is a summary of individual experts' assessments for each component as well as the distributions of assessments across all experts for each component. The component-level distributions were used to complete the individual logic trees where necessary and to develop the composite hazard model. Crustal Geometry All of the experts provided an assessment of the cross sectional geometry of the subducting Juan de Fuca plate. Most of the experts provided only a single assessment consisting of the plate dipping at arproximately 11* and extending through the zone of deeper earthquakes lying at depths of 30 km or more beneath the site. Two experts provided a slight modification of the
O
/)OO==
OEOMATAIX 3-3 10' ' dip consisting of a flat lying slab with a double bend (see cross section for expert 6 in Appendix A as an example). Many of the experts preferred the model recently: proposed by Crosson and Owens (1987) that has an arch in the slab along strike. Figure 3-2 presents the aggregate distributions for slab geometry. Seismic Sources All of the experts identified the Juan de Fuca - North American plate inter-face and the subducting Juan de Fuca plate as potential sources of thrust and intraslab normal events, respectively. Some experts also identified potential sources in the overlying North American plate. Evaluation of the hazard from these crustal sources was included in the shallow crustal source model described in Section 3 2. Probability of Source Activity All of the experts made an assessment of the probability that the plate interface and the subducting slab are active or seismogenic (see Section 2.1 for discussion of "activity"). Figure 3-3 shows the distribution of assess-ments of activity for the intraslab and interface sources. The assessments for the intraslab source are generally at or near unity based on the past record of seismicity. The assessments for the interface range from near zero to near 1.0 with an average of 0.54. The assessments cluster near zero, near 0 5. and near 1.0. It should be noted that the that an adjustment was made to the assessments of experts 4 and 13 As indicated in Table 3-1, column 5. these two exports have probabilities of 0.9 and 0.85, that the maxikum magnitudes for the interface is M, 5 or less. All other experts Cade the assessment of activity in terms of the probability of the interface being able to generate tectonically significant events (K,>5). To put the assessments of experts 4 and 13 on a consistent basis they were adjusted to values of 0.075 and 0.0075, respectively, and their max 2 mum magnitude distri-butions renormalized to include only magnitudes larger than M, 5 These adjustments were discussed with the experts and they were .1 agreement.
w theOMATmeX . 3-4 Locations of Ruptures ne experts provided assessments on the limits of earthquake ruptures, both along the' length of the subduction zone as well as the up dip and down dip extent. Figure 3-4 provides histograms summarizing the responses obtained.
'Most experts considered the maximum limits of coherent rupture along the in-terface to be the boundary with the Explorer plate at the Nootka fault zone on the north and the Blanco fracture zone on-the south (see Figure 3-5). ~Several experts considered further segmentation of the interface to have some credibility, with a segment boundary generally in the vicinity of 46*N or segment boundaries on the northern or southern margins of the arch in the slab proposed by Crosson and Owens (1987). The assessments of the minimum depth of rupture along the interface ranged from 5 to 25 km and the maximum depth of rupture ranged from 35 to 60 km. The distributions for minimum and maximum depth of interface rupture shown in Figure 3-4 were used in develop-ing an expert's hazard model if he did not make an assessment.
A majority of the experts stated that they expect the future distribution of intraslab events to follow the observed pattern of historical seismicity
-with the majority of events occurring generally beneath Puget Sound. Alter-natives considered included completely uniform seismicity within the down-going slao or a concentration of larger events at deeper depths. Figu'e 3-4 shows the aggregtte distribution for seismicity distribution. The pattern of historical seiskicity generally inferred to lie within the Juan de Fuca plate is shown in A,>pendic C Figure C-2.
Maximus Magnitude n e experts that assessed maximum magnitudes for the interface either made a direct assessment or specified that it be calculated from the maximum rup-ture dimensions assessed above using the relationship between rupture area and magnitude proposed by Abe (1975) and Kanamori (1977). neir relationship can be writted as M = logg (A) + 3 99. Regression of published values of M, and Area for recent earthquakes holding the slope equal to unity yielded the same relationship between magnituds and rupture area. Twelve experts provided an assessment of maximum magnitude for the interfaces seven (58%)
~
r GEOMATRIX 3-5 specified the use of maximum rupture dimensions and five (42%) gave a direct 4 assessnont of the maximum magnitude on the basis of analogy with other sub-lJ [ duction zones or other techniques for magnitude estimation. The aggregate distribution shown in Figure 3-6 is for those five experts who made a direct assessment, and is thus conditional on the direct assessment procedure being the correct procedure. In general, the maximum rupture dimensions specified by the experts resulted in maximum magnitudes of about 9 If an expert did not assess interface maximum magnitude, then the marginal distribution used to represent the aggregated opinion of the other experts consists of 0 58 weight assigned to the magnitude value obtained from the experts assessment of maximum rupture dimensions and 0.42 weight assigned to the conditional ' distribution based on direct assessment. The distribution shown at the top of Figure 3-6 has a large probability of y 0 38 assigned to a maximum magnitude or 6. As this represents the judge-ments of two of the experts based on specific reasoning, it is an appropri-ate distribution for use in component level aggregation. However, it was judged that this assessment is significantly lower than would be obtained ! from a general population of scientists familiar with subduction zone earth-quakes and those experts who did not make any assessment of maximum magni-tude for the interface would, nevertheless, be likely to assign a much lower probability to a maxirum magnitude of 6. Accordingly, the conditional dis-tribution used for those experts who did not assess maximum magnitude (i.e. , the distribution for use in "gap-filling") was modified from that shown at the top of the Figure 3-6 by removing the assessments for very low magnitudes and renormalizing. The resulting distribution is shown in the middle of Figure 3-6. The maximum magnitude for the intraslab source was assessed by 11 experts on the basis of historical seismicity and analogy with other subduction zones. The aggregated distribution is shown at the bottom of Figure 3-6.
s f om=
===e.urrux 3-6 Earthquake Recurrence Method All experts who made an assessment of earthquake recurrence preferred to use historical seismicity data to define the recurrence parameters for intraslab events. Appendix C presents recurrence parweters for intraslab events l based on an analysis of the seismicity data. These parameters were used for all experts. Recurrence estimates for the plate interface were assessed either on the basis of a moment rate approach or on the basis of geologic evidence for the frequency of large events. In aggregate, the experts favor the moment rate approach slightly more than the use of the geologic data by the ratio 0 54 to 0.46. If an expert did not make an assessment of earth-quake recurrence for the interface, then both methods were used with the given weights.
Geologic Recurrence Rate Six of the experts chose to base the recurrence estimates for interface events solely or partially on geologic evidence for possible paleoseismic events, primarfly the data from coasta? subsidence and offshore turbidites. Figure 3-7 prosents the aggregated distributions for mturn period of large interface events. The distributions ara centered about an average recurrence interval of about 500 years. Con >argence Rate All of the experts made an assessment of convergence rate with most basing the assessment on the rate estimates published by Riddihough (1984), Nishimura and others (1984) and verplanck and Duncan (1987). Those experts that made a direct assessment generally gave a wide distribution of values with a mean value comewhat lower than the published estimates. Figure 3-8 thows the aggregate distribution for convergence rate estimates. Seismic Coupling Figure 3-8 shows the aggregate distribution- P the amount of seismic coup-ling between the Juan de Fuca and North American plates. Most of the experts gave a wide distribution for the amount of coupling with expert 1 giving a zero/one bimodal distribution. The bases for estimates of coupling were
r
* ** x i
3-7 quite varied and ranged from analogies with other subduction zones to thermal-mechanical modeling of the plate interface. The product of the plate interface area, the convergence rate and the amount of seismic coupling provide the rate of release of seismic moment. For an interface length of 800 km, a width of 100 km, a convecgence rate of 4 cm/yr and an aggregate mean of 0.4 for seismic coupling gives a moment rate of 3.84 x 1026 dyne-em/yr. Assuming all of the moment is released in magnitude 81 events, a moment rate estimate of approximately 200 years would be obtained for the return period of these events. Recurrence Model Three recurrence models for the form of the magnitude distribution were used for earthquake sources in the analysis: the truncated exponential distribu-tion, the characteristic magnitude distribution, and the maximum moient dis-tribution. Figure 3-9 illustrates the cumulative form of these three distri-butions and compares how they would estimate the frequency of smaller earth-quakes when the absolute level of seismicity is fixed by the frequency of the largest events. Based primarily on the historical absence of small- and moderate-magnitude events, most exports preferred the maximum moment or characteristic models. The aggregated distributions of the experts yielded weights of 0.52, 0 38, and 0.1 for the maximum moment, characteristic, and exponential models, respectively. 3 1.2 Ground Motion Attenuation. Appendix D presents an analysis of strong motion data from subduction zone earthquakes, including data form the recent earthquakes in Chile and Mexico. Two attenuation models were developed rep-resenting the uncertainty extrapolation of the empirically based attenuation relationship to magnitudes greater than M, 8. The two models are designated "S-Cubed" and "Joyner" indicating scaling laws based on the results of ground motion simulations (S-Cubed, 1988) and on theoretical source spectra and ran-dom vibration theory (Joyner, 1984). As indicated in Figure 3-9, the S-Cubed model is given greater weight (0.67 vs. 0 33) because it is based on simula-tion done specifically for ground motions at the WNP-3 site.
I GEOMATRIX 3-8 32 Shallow Crustal Hazard Model The logic tree format developed to model the shallow crustal sources in the North American plate $s shown in Figure 3-11. Eleven seismic sources were representea in the hazard analysis, consisting of five sources related to geologic / geophysical features and six distributed area source zones. The maximum magnitude and earthquake recurrence estimates for these crustal sources is presented in Appendix B. hree recently developed attenuation relationships for ground motions from shallow crustal earthquakes were censidered applicable for' estimating ground motions at the site from the identified crustal sources. The peak accelera-tion relatinnships published by Joyner and Fumal (1985) and campbell (1987) are based on ground motion data available through 1980 and are considered by their authors to be applicable to both soil and rock sites. The third rela-tionship (Ccomatrix Consultants, 1987) is a modified form of that published by Sadigh et al (1986) reflecting analysis of data recorded on rock sites post 1980. _ n e three relationships were given equal weight in the analysis. o i s b I P
mm TABLE 3-1e
SUMMARY
CF plosaBILlif op activity earreve_eacettype Litial octanic stat steester toirt rottattat seismic soveces
#1 for of deep seleeletty intre.eieb le) 1.0 le) 7 7$ (so.25) (a) J#
Intertsee [t] 9 35 to.25.e.5) [t) steenstems (t) 3 766 at teep etoplelte Intes*elab la) 0.8 (ej D6
- 7 (t) 2' dJ 3i 0.. (e.05.e..) it) <
..ter,.ee it)
- 4 t masse.er part et al 4<
#3 ter er deep assenten ts satre-eats f.) 1.0 tel
!aterface (t) e.9 [tl 16 (deeper port of al s (e II (DI 3.0 (a) 7 16 (ej 8(
d4 fep of deep sensetetty latre. east (e] 2< 0.075 (61 3 (e.3) I 3' Intertees (t) 4 (e.3) ) 4< 5 (e.3) )th) 6 to.e9) l 7 (e.es; I top et seep setestetty (0.85-0.s3 satre-enen tal .0 (s) 7 (al 3
#5 se* day (e.ne-o.15) Interface It) . e.5 (oo.5) (61 Tep et deep esteetetty, stagne Dead 40.)) antra-slet le) 1,e [sj 6-3/4 = 7 t (e) ' #6 l' Interface (t) 8.6$ (so.2) (t) steensione (b) 2<
31 Top of deep selseletty. deutte bene (e.7) 4i C e.) (so.2) (b.* Dimenaies: (t) ~I'
#7 for of deep esteeletty (e.7) Istre*elen (a) plate interfeee (t) latre-elet (a) s.e (a) $$sessione (6) J
#8 Top et deep senseletty
!aterface [t] e.5 (e.25
- 4 75) (ts for of deep eetestetty 8stra-eist to Some seeth (a) 0 9 (a) 6..4 (e.3) )
#9 I.e (0.25) I [a] 4 Intes.stan 50 - 75=e (t) 1.0 (6) 7.$ (o.55) I f.4 (0.8) I laterface (e) e.95 e e.05 (e) 7.e to.1) I strike-ente taunts to apper state (e) 75(0.8) ) (t) 7.8 (0.8) )
Aterellenary wedge feelte (e) 8.e (4) es.enesome (e)
- 7. re se e e-setes stat Irl s.e [e) 2.e (t) 74 (a) 7 5 to.a) ) te) 4.e (0.r) 3 8.( (a) 74 (e (,' (a) 3
#: 9 top of deep esteenetty 1stre-easb (a)
Interfoee (b) 0 7 to.6 = e.9) lb) steenstee. [t] 4.e [e) 7 = 76 (ej
$11 for of deep seisatetty letra-enab (a) a seterras. (e) e., (o.s - 3.0) It) steenesene (t) see, erset.: seeres tel n.0 te) 75 tel v.P er eeep essentestr s.tre..i.m t.) e.95 - t.0 to) 74 tel ~~h est
. tert. e 16) e.9 (b) 9 161 inte.-.nen t.) s.e 1 1 4 5 co.45) )
sa) t.,.t see, ease.s si, 7.0 (0.4) l teterrese it) e.e875 lb] 7.5 to.84) l (el 4.0 (0.01) ) l 4 10 5) 1 5 to.35) l (tl 6 (e.35) i 7 25 - 7 5 tel "i se, or e.ep eeseessity intve eset tel .o 141 sie s.ierrees it) s.35 te e.25) til 4.5 to 0 5n (6) ] aeseets...er = esse Isl 0 7 to 0 88 tel 7 to 0.251 tel
- iesee i Gian tal e.es te o.est 141 s tal
l _
\
)
I EXPERT IFTERVIEWS co.vt.ci.cs siis.ic stotocic .scussi.cs
.Ait (5B/1 1 D t ti'.# t .C t .t t.0. COUPLi.4 f.) p, Cit.t.Ct .0Dtt t...t t..?MQU.lfS (TS1 l
Ltei:) ...e..... .. ..seier i.i ** to.5 . . 66) is, ..si i i.i l r.... i to.,. .... . .. .i... . i t. l .. e i.i 33
]
l t s.is se.a.sie t. 4> i.3 - - 43o i : >-3$ ( 2.8) ' ! '(s.it -s u ,.t r.ie (e.1) st! , l i t. i,3 . .i..... .. .... iy i.i . 66 s.p....si.i i.3 Soo (400 i w. ) i.: I se...ste .. . i.i es...eisei. e i. i~ t..es .i.e.es... .. ... sir i.3 e.e5 to . ..isi c ... ....sie ini - l (3 5)
- p. (. 43 ...e.. .... toi O.es, I (3% . 4.) .l....se. .....iesty in) - 3.p ..ti.i i.:
> e, si (s.04) ...t...... .es. sei y i.) . 6 t. e.i5) 8.P.... isa ini Sve . 60. t asi) (.i > 20 ' O.e4) > 3e to.4) se...sie ..t. (0 7$) i.i s.p ..ti.& to.2) i.i > . 6e O.i) > . 43 - c.4) .. ,.ie to.25) i.i c....ei.>i iis to.8) i.i l - Se o.. ) t s. ie) ...... .. ee.....iir i.: i ti..) s. ..ii. i : - c.......sie i to.5) t.:
....e.. to.5) s.
1 (.ie *ie) - . 5 to.a . . 7) - - l
..? . ... (. 543 a.p ..si.2 is.......ri See t. ass) ioi l
...i.rie.,i .. ..sei y T1
, i. ci i........ i ..
.e .e,,i..=5
..e .....e ..... . <e.i> t.:
. .... .... it'
.: m... t e. i> i. t\PERTURE CARD 1 Also Avahble Ori l Apertirro Caril l i
5 3. i. i,, . . . . . . . ..... . , i.: . 3 t. . ..s> :. .e. ..i i...i m.. ...ii i e i. ci .i. ..ie.i .. .. .ii,i...i - ... ..ii. i... so. 35. i. l
.i..ie .. . i.i _..i........i .ise i.i i . 3, iis . .. . . .. e. .. .. . , i.i .. t...s . . 53 .. .. . i.i s . i.:
- ..i .... i.i ... ... . ..s i.i o t. i,, .41... e.i ..... . i, i.i ...,i. ...s> i.....i.i -
. . , . uu, .t........ .....ieis, i....ei ... .... .. ... . i.....i
.. .., t....
..3 .
i ,.,3
. 4i m ..i .. . i.i .. . (..ei ....... ..... i.:
>
- s. i3.3) ..is
.g c o..i5 s i se ts.sg 89030 con 9-*I 4 _ ___m_..____._____ __
i! m S0 ,x n o t o t d G e n e G e n e e se n O. t i e n n m r e m st e e e t
- o b u n o x M o W eb r
e n a ; r gr u as t L h h E Q h uiD t
/
- }\ f icgn 2 3 4 A o / mE e p 0 0 N . sC ea l e r r
"g\ r r r d
m o y y y hm /y /p 0 0 0 d r
" R t
e a 0 m 0 m m 0 m m 0 3
/
0 4
/
0 5
/
a z a 2 3 4 1 1 1 h s
}\ c e
c r e t n c e i g u nd eo m loo o rh r t o e s -. u e u G cM e e M n . o z _. e n md uu 5 o 9 mt myu a 2 i t oa c uw u w d b u r e S
#t4 R u
r mo "
" 1 f -
o 3
/
f e r a t d e u ,
- tn g a i
-.. n P y
e
. S U
f oy hc m t t eM t b a ac SA t
/}\ s e
s o o
- f _
r o e k i h - y A s _ l : . t r e . e d _ o d. m o u u e G
/r e
,.: ; i ) i - <
1: a j ., ihz )l - 9' , . y-g mammantnoo 6. Percent
. Slab 0 -20 40 60 - 80 -100 Geometry +----+----+----+----+--.-+----+--.-+----,----+--..+
Double Bend **** 11* ******************************* Arch Model ******************
.+....+....+....+....+....+....+....+....+....+....+
Distribution of 14 experts L s Figure 3-2.' Aggregate distribution of 14 experts for oceanic slab geometry 1 i t i
r GEOMATRIX Probability Percent of 0 5 10 15 20 25 Activity +----+----+----+----+----+----+----+----+----+----+
<0.025 ***************
10.05
- 0.10 ***************
f 0.15
- 0.20
- 0.25 *
{-- 0.30 *************** 0.35 ******************* 0.40 ************************** 0.45
- 0.50 ******************************
0.55
- 0.60
- 0.65 ***************
0.70 *************** 0.75
- 0.80
- 0.85
- 0.90 ********************************************
0.95 ***************
+----+----+----+----+----+----+----+----+----+----+
Aggregate distribution of 14 experts for interface activity Probability Percent of 0 20 40 60 80 100 Activity +----+--,-+----+----+----+----+----+----+----+----+ 0.800 ***** 0.850
- 0,900
- 0.950
- 1.000 ***********************************************
+----+----+----+----+----+----+----+----+----+----+
Aggregate distribution of 14 experts for intraslab activity Pigure 3-3. Aggregate distribution for probability of activity
( GCOMATRIX Percent 0 20 40 60 80 100 Segmentation +----+----+----+----+----+----+----+----+----+----+ f ****** none Nootka-Blanco ********************************* Segment @ 46*N ******** (- Segments ****** N & S of arch +----+----+----+----+----+----+----+----+----+----+ Aggregate distribution of 14 experts for interface segmentation Percent Depth 0 10 20 30 40 50 km +----+----+----+----+----+----+----+----+----+----+ 5 *************** 10 *************** 15 ************************** 20 ******************************************** 25
- 30 *****
+....+ __.+____+....+....+....+_...+....+....+-...+
Aggregate distribution of 14 experts for minimum depth of rupture on interface Percent Depth 0 10 20 30 40 50 km +----+----+----+----+----+----+----+----+----+----+ 35 ********* 40 ************************************************* 45 **************** 50 ***************************** 55
- 60 **
+ .-.+.__.+ ...+....+....+._-.+....+....+-...+....+
Aggregate distribution of 12 experts for maximum depth of rupture on interface Percent Seismicity 0 20 40 60 80 100 Distribution +----+----+----+----+----+----+----+----+----+----+ uniform
- observed **************************************
uniform N/S ***** variable N/S **********
+----+----+----+----+----+----+----+----+----+----+
Aggregate distribution of 14 experts for intraslab seismicity diatribution Figure 3-4. Aggregate distributions for location of rupture
w - f s
* =.. M ,
s t % .e
+ 4
'.q,./.,,...s . . 7,y . . . . . -. EMPLANATsO86
-J .
e * .
'. 4 ***
e e 2 74 u m 4.* en-
. 4D - e s usr.sume l
- e
. ,,-,,u 1 ..* : 0,' .; ,
. s ~..
s
~% >..3,.
. as -e.s u==im.d.
6 .. es, . . . .* \ .h ...
' k'\ ,.
"*
- 8 . R*'**"**d*******'
- ll- W e's.'.ed'asW'i8P
. 4. O ,. h $ *
..e,..e c~
.=~
x:..-
\ -
!. 'i~ .
.1t
, A. .ea.ye y sev usi s .** * ~.- ~~. 's m
=dv.Aa.nad
,,,y *
*9 = - Appe 4m.w b.e. .f C.ata.ac.I S m \
*. ; .? . - -
4 e . .gw ;. *.p , - *
\ 1 l.T.. - ; .\ i s_l ,i . . I ,Ns ,.
.=
' tw . \ .
y.
. 7-
, .k/ . huri p.4yL ="*"? . - *
. '. ; :s - * ,
, NOTE: te.e.ae p6.ei.d e..'id of Is77
- l. a - , .. ..
. . '*1- .
* . . m .':::
1\ o- 1, n.cf; u
- ~.~ no ,
. t:v
. ... .; . t .
en oem t l ,. - . L w 8 s. ,,. . ,~ l
. . v.4. ;e- :. .; ,
- ~ " , c p .. ..,5 m g.
I -. 4 ,. f -. .. ,; . f
, .- . 1
- e. .
e'a dr
-, . . i,-. Tar...;,, reE.- /*
y-T! * ** . . .u.t.c .. -t N.c s..e Pr.reces 3 & 5 in ri-J **., * * * .b
...,,*.8.
, . ..}, . at saner, a=atvsis auro ='
l ,
\
g A g, SttstescrTv CF TME PaCittC ason,
. . *k
- ST
.- .= w -- n- .., .., . -- -r
,- 4. d** . <. *' '
r.,,,. 23-2s Q v i Figure 3-5 Plate Boundaries and Regional Seismicity X
"_\
i onomarm
. Percent Maxianna 0 10 20 30 40 50
- f. Mannitude +----+----+----+----+----+----+----+----+----+----+
6.00- ***************************************
'6.25
- 6.50 *
[- 6.75
- 7.00 ***'
7.25 * * [ F 7.50 -
- 7.75 ~*
8.00 ***** 8.25 '***** 8.50 '********* 8.75 ********* 9.00 ***************************** 9.25 ***** 9.50 *****'
+....+....+....+....+....+....+....+....+....+....+
Aggregate distribution of 5 experts for directly assessed interface maximum magnitude Percent Maximum 0 10 20 30 40 50 Maanitude +----+----+----+----+----+----+----+----+----+----+ 8.00 ******** 8.25 ******** 8.50 ************** 8.75 ************** 9.00 ************************************************ 9.25 ********
'9.50 ********
+----+----+----+----+----+----+----+----+----+----+
Conditional distribution of directly assessed interface maximum magnitude used for "gap filling" Percent - Maximum 0 10 20 30 40 50 Magnitude +----+----+----+----+----+----+----+----+----+----+ 6.50 ***** , 6.75 **** , 7.00 ************************************ 7.25 *********************** 7.50 ************************************ f 7.75
- t 8.04 ** l
+....+....+....+....+....+....+....+....+....+....+ ;
Aggregate distribution of 11 experts for intraslab maximum magnitude r Figure 3-6. Aggregate distributions for maximum magnitude r [ r f
F GEOMArAIX 1 Return Percent Period 0 10 20 30 40 50 years +----+----+----+----+----+----+----+----+----+----+ 250
- 300 **
350 *** 400 ******** 450 ***************** 500 ****************************** 550 ******************** 600 *************** 650 ***** 700 ***** 750
- 800 ***
850
- 900 **
950
- 1000 *
+----+----+----+----+----+----+----+----+----+----+
Figure 3-7. Aggregate distribution of 6 experts for return period of large interface earthquakes based on palcoseismic data
- p. - - - . - - - - -
f.y - , amounmx - _ l. Percent
. Convergence 0.. 5. 10 15 20- 25 fs 'calvt .+----+----+----+----+----+--~~+----+-~~-+----+----+; >
L 0.5 *n ' i' ' 1.0. ** . , 1.5 ** fi 2.0 ************* 2.5 L**************** 13.0 ************************************************ :
- t. , .
- 1) 3.5 **********************************************
4.0' ******************************************** :; 44.5 -*************** 5.0 ;******************** ! 5.5
- 6.0 -*** 3
+....+....+....+....+....+....+....+....+....+....+
Aggregate' distribution of 14' experts.for convergence rate ' i Ptreent 0 5 10 "' 15 20 25 Alpha- +----+----+----+----+- --+----+----+----+----+----+ (0.025 ********************************************* 0.05 ********************* 0.10 ****** 0.15 **** 0.20 ***** 0.25 ****** 0.30 ******** 0.35 ******** 0.40 ******* 0.45 ******** 0.50 **********' O.55 ********** 0.60 ********** 0.65 ********* 0.70 ******** 0.75 ******* 0.80 ***** 0.85 '***** 0.90 ***** 0.95 ************ 1.00 '***********************
+....+....+....+....+....+....+....+....+....+....+
Aggregate distribution of 10 experts for alpha i Figure 3-8. Aggregate distributions for moment rate parameters
m,-
, ll CB50MATHEM r
L r 3 5 . 3 l 3 l 3 l l
.. . . . . . mm
. . . . Wedmwn Woment 10-1 !a.( :
- N. :
. s gs .
. 'N. . .
g . g .
' s 10-2 : _ N _
Ns :
. g .
s s . s
. s .
s . s
%g' 10-3 2*************************# :
. g
. t .
l
' ' ' I- '
104 ' ' ' ' ' 5 6 7 8 9 10 Magnitude Figure 3-9 Magnitude frequency distributions used in the hazard analysis
R1,X s t _ c t d se n d _ em n e a e e n l e n n t r u m n w t e e t o a e-ih s M n c M n s u e e g y w
, i a L et 0 w e M O M _ O H
/ \
f. o, 2 3 4 A d .r. 0 0 L / i. r ( N e . e o Sc
\
r ry r r y r l hn y y y e Vn Vrr 0 0 0 d r n m n 0 3 0 4 0 5 o
,, 0 e
0 0 r
/ / / m 2 3 4 1 1 1 d
k \ r a j"3 j : s e a a t n ' n n e s A a-m s s a m e e c r u - o - s m- 5 s 7 e m :- 9 d r 8 a n a o M- g z n ~ o i e . t e d i s r e n d b c u - u s [ ]d t e d e e n t t a n n a e y n l _ e y p - s e m n S -
, ~
U co - f o 0 1 S t t em a _ 3
/ m .
s u g e M e i f F
~
_ - r o r t e hi h A a . d d e s _ e = M > . 7 _ n e w(D E - e t m m-n m u s D W(4 0 - A
/ \ -
/\\ -
{- M GEOMATRIX (
-._._7 Attenuation Maximum Recurrence Recunrence Relationship Magnitude Method or Srp Rote compben s-1/2 Puget [ _ 0.031) 0.013/yr (5333)
I winamette
' f (5-3/4 /
/
0.b0.51118/yr (0.25)
! Local 6 Salamicity O. 2 h et R (0.438) '
-\6-1/4 (1.0)
}\0.\
yr
~
Olympic P
- I _ , , (0.25) \ 0. yr Geomatrix Offshore 6-1/2 (0.18)
(0.334) (0.031) Mt. Q Helens 0.1 mm/yr Hood C. [ (0.125) 0.2 mm/yr Nisqually ./ (0.25) Olympio , 7-1/2 , Srp Rote /,/ / 0.3 mm/yr (1,0) s Shelton -- (1.0)
\\-Q4 (0.25) mm/yr Joyner-Fumal (035)
(0.333) 0.5 mm/yr _ (0.125) Figure 3-11. Shallow crustal sources hazard model.
GEOMATAIX s - 4.0 ANALYSIS RESULTS 4.1 Hazard Computation r. Seismic hazard computations were made for each of the hazard model logic L trees developed in Section 3 utilizing the formulation given in Section 2. The hazard was computed considering the contributions of earthquakes of ( magnitude 4.0 and greater for the local shallow crustal sources and magni-tude 5 0 and greater for all other sources. The probability density func-tions for distance to earthquake rupture were developed by modeling earth-quake ruptures as rectangular rupture areas distributed over a fault plane. The plate interface and the individual shallow crustal features were modeled as single fault planes with earthquake ruptures distributed over the fault surface. Distributed area sources, including the intraslab source, were modeled as a series of parallel fault planes occupying the volume specified for the source. Spatially variable seismicity rate was modeled by specifying the fraction of the total seismicity that occurs on each fault plane. The mean rupture area of an event was specified by the fellowing relation-ships: rupture area = exp(2.28M, - 8.92) (4-1) for interface events, and rupture area = exp(2.0M,- 7 12) (4-1) for all other events. These relationships were developed by linear regres-sion of In(rupture area) on magnitude using the data presented by Abe (1975) and Wyss (1979). Hazard computations were made for each end branch of the logic trees devel-oped to model the uncertainty ir. input parameters. The resulting discrete distributions for annual frequency of exceedance of a range of ground motion levels were used to define 15'". 50th (median) and 85 percentile hazard
GEOMATAIX 4-2 curves to represent the uncertainty in the exceedance frequency. The compu-tations were made for peak ground acceleration and spectral velocity at periods of 0.15 and 0.8, 2.0 seconds. The periods of 0.15 and 0.8 seconds represent the periods of maximum amplification of spectral acceleration and l 1 spectral velocity, respectively, in the response spectrum developed for subduction zone earthquakes (see Appendix D). I 4.2 Exceedance Frequency for peak Ground Acceleration 4.2.1 Total Hazard. Figure 4-1 presents the computed hazard for peak j th D th horizonta? acceleration. Shown are are the 15 , 50 , and 85 percentile hazard cu.ves for the shallow crustal sources, the subduction zone sources, and the c>mbined total hazard. ne hazard curves for the subduction zone sources are from an equally weighted aggregate distribution of the 14 expert assessmenta. As can be seen, the hazard is dominated by potential subduction zone events. The 15th and 85th perf.entile Curves for the total hazard differ by approxiasately a factor of 2 at low acceleration levels to a factor of about 10 at high accesleration levels. Figure 4-2 presents the average contributions to the total hazard from events in vara.ous distance and magnitude increments and from various sources for peak accelerations of 0.1, 0 3, and 0 5 g. At low acceleration levels, the hazard results primarily from contributions from the smaller, more frequertt 9 events. As the neceleration level increases, the larger magnitude events increasingly dominate the hazard. There is also a major contribution from intraslab events in the magnitude 6 to 7 5 range. ne major contribution to
} the hazard is from events in the distance range of 25 to 65 km corresponding to the closer portions of the plate interface and subducting slab.
The hazard results for the two sets of seismic sources are discussed below. 4.2.2 Subduction Zone Sources. The seismic hazard from the subduction zone sources was computed using the input parameters defined by the 14 experts. The primary approach used in the analysis was to develop a set of input k____ . _ _ . _ _ _ . _ . _ _ _ _ _ _ _ _ . - . . _ _ - _ . _ _ . - _ . - _ _ - _ . _ _ _ _ -
GEOMATRIX 4-3 parameters for each individual expert and then average the resulting 14
~
hazard distributions to develop an aggregate hazard distribution. Results for Individual Experts. Hazard models were developed for each of the 14 9xperts using their individual parameters when they provided a response for a given component and aggregate marginal distributions of the assessments of the other experts when they did not respond. The complexity of the logic f trees for the individual experts varied dramatically from a minimum of 15 branches to a maximum of 1000 branches. In general, those experts with large. final logic trees did not provide a complete hazard model; the com-plexity reflects rather the use of the aggregate of the opinions of the other experts for missing components. Hazard computations were performed for each expert's model using the two attenuation models developed in Appendix D. Figure 4-3 presents the median hazard curves obtained from the logic trees of each expert. The range of median hazard curves spans about one end one-half orders of magnitude. Figure 4-4 shows the 15 *. 50 , and 85
- percen-tile bazard curves for each expert. As can be seen, there is generally nn
( order of magnitude difference between each expert's 15 th and 85
- percentile hazard curves, which is comparable to the variation between experts shown in Figure 4-3 Figure 4-5 shows the 15 , 50th, and 85 th hazard curves for the two subduc-tion zone sources treated separately for each expert. The solid curves are
! for the interface source and the dashed curves for the intraslab sources.
In many cases, 15 * , or 15th and 50 th hazard curves are not shown for the interface source, reflecting the assigned probability of the interface being inactive greater than the missing percentile hazard curves. For example, Expert 1 assigned a probability of 0 35 that the interface is active, giving a probability of 0.65 that it is inactive. Thus the 15 and 50* percentile hazard curves for the interface are zero. Experts 4 and 13 have assigned probabilities less than 0.1 that the interface is active at a magnitude level greater than 5 and thus all three percentile hazard curves are zero.
GEOMATAIX 4-4
~
As can be seen, th we is greater variability between experts in the assess-ment of the hazard from the interface source than the hazard from the intraslab sources. Aggregated Hazard. As indicated above, the primary approach used to aggre-gate the assessments of the 14 experts was to form an equal weighted average of the hazard distributions obtained for each expert. Figure 4-6 presents the resulting distribution for exceedance frequenc/. Shown are the 15 0 500, and 85* percentile hazard curves.for the aggregated distribution (solid curves) as well as the median hazard curves for the 14 experts. The th th 15 and 85 percentile hazard curves of the aggregated distributions encompass the median hazard curves of 10 out of 14 experts and an equal number of individual expert medians lie above and below the aggregate median. , Figure 4-7 presents the results of the alternate aggregation approach dis-cussed in Section 2 3 In this approach, the aggregated margina?. distribu-tions for each component of the hazard nodel were used to construct a single composite hazard model logic tree. As shown in Figure 4-7, the composite s model hazard curves are very similar to the aggragate hazard curves. Contributions to Uncertainty. The contributions, of uncertainty in various components of the hazard models to the uncertainty in the computed hazard are illustrated in Figures 4-8 through 4-18. Figure 4-8 compares the 15 0 and 85* hazard curves considering expert-to-expert variability (dashed t lines) with the 15 and 85 percentile hazard curves representing the th and 85* percentile total uncertainty. The difference between the 15 ranked experts is nearly equal to the difference between the 15 th and 85 th percentile curves including all uncertainties. Taking the relative differ-th th ence in the spread between the 15 and'85 percentile curves as a measure of the relative difference in the square root of the variance in hazard, the export-to-expert uncertainty contributes approximately two-thirds of the total variance in hazard. i
[ (, GEOMATAIX / 4-5 L , Figure 4-9 shows the contributions to uncertainty in hazard resulting from uncertainty in modeling subduction. zone earthquake ground motions. 'Ihe plot on the left compares the hazard curves obtained considering both interface and intraslab events and the plot on the right- shows the hazard curves considering only the hazard from interface earthquakes. As can be seen, there is only a minimal difference in the two hazard curves reflect.ag the large contrirbutions to hazard from events below magnitude 8 (Figure 4-2) for which there is no difference in attenuation relationships. Figures 4-10 through 4-18 present the contributions to the uncertainty in hazard due to the uncertainty 1n various components of the individual experts
. hazard models. In each figure the solid lines are the 15 h and 85* percen-f tile hazard curves resulting from the total uncertainty from all components of the hazard model and the dashed curves represent the 15 0 and 85
- percen-tile conditional mean hazard curves considering uncertainty only in the component identified in the figure title. (Fractiles of means are shown rather than fractiles of medians because they are more efficient to compute, although they result in a shift away from the medics hszard coward t,he higher percontiles of the distribud oa). The contributions to the total uncertainty vary from couponent to component and from expert to export. For
- any one component, the width of the distribution shown in Figures 4-10 through 4-18 reflects both the amount of uncertainty in ti.e assessment of the parameter and the sensitivity of the computed hazard to variability in-the parameter. For example, the results presented in Figure 4-10 show that the effect of uncertainty in the geometry of the subducting slab on the hazard is small. For most of the experts this results because they selected l
only a single model for the slab geometry. However, even for those experts who considered alternative geometries, such as experts 1 and 14, the impact on the hazard is relatively small. Alternatively, the results presented in Figure 4-11 indicate that uncertainty in source activity has a significant impact on the uncertainty in hazard. The reason for this large effect was , shown previously in Figure 4-5 As indicated in that figure, the hazard from the interface is generally comparable to or higher than the hazard from
GEOMATB11X 4-6 the intraslab events, thus the hazard is significantly altered depending on whether or not the interface is seismogenic. Examination of Figures 4-10 through 4-18 indicate that the major contri-bution to "within expert" uncertainty is from uncertainty in source activity. The contribution of "within expert" uncertainty to the total uncertainty can be estimated from the results presented in Figures 4-8 and 4-9 The total i variance in the hazard is the sum of the expert-to-expert variance, the variance due to uncertainty in modeling the attenuation characteristics, and the average within expert variance. As the expert-to-expert variance was estimated above to be approximately two-thirds of the total variance and the uncertainty in attenuation contributed little to the uncertainty in hazard, l the average "within expert" variance is approximately one-third of the total variance. 4.2 3 Shallow Crustal Sources. Figure 4-19 presents the hazard computed for 0 the shallow crustal scurces. The plot on the left compares the 15 , 50 0 and 85 percentile hazard curves for the shallow crustal sources with those k for the subduction zone sources. As can be seen, the median hazard from shallow sources is approximately one and one-half order of magnitude lower than the median for subduction zone sources and the range between the 15 0 and 85 percentile curves is greater. Shown on the right in Figure 4-19 are the median hazard curves for the various shallow crustal sources. As indicated by the comparisons in the figure, the hazard from shallow crustal earthquakes is dominated by the local random seismicity source. Figure 4-20 presents the contributions to the uncertainty in the hazard from shallow crustal earthquakes from uncertainty in selecting the appropriate attenuation relationship, uncertainty in maximum magnitude, and uncertainty in recurrence rate. The uncertainty in hazard is largely due to uncertainty in modeling attenuation with a small contribution from uncertainty in esti-mating earthquake recurrence rates. The large uncertainty in maximum magni-tude for the local random events has little impact. l l
m GEOMATAIX 4-7 4.3~ Exceedance Frequency for Spectral Velocity Figure 4-21 presents the computed hazard for 5-percent damped spectral velo-city at periods of 0.15, 0.8 and 2 seconds. Comparison of these results with the hazard curves for peak acceleration shown in Figure 4-1 indicates that the shallow crustal sources have a similar level of contribution to hazard for spectral accelerations as their contribution to hazard for peak acceleration. Figures 4-22 through 4-24 present the 15 th , 500 , and 85 0 percentile hazard curves for the individual experts, for periods of 0.15, 0.8, and 2 seconds, respectively. As can be seen, the uncertainty in hazard increases somewhat for longer period motions, reflecting greater impact of the uncertainty in the potential for very large events on the plate interface. The individual expert median hazard curves are compared with the 15 , 50 0 , and 85' percentile hazard curves for the aggregated distribution in Figure 4-24. As was the case for peak acceleration, the 15* and 85th percentile ? hazard curves for the aggregated hazard encompass the median curves for 10 to 11 of the 14 experts. The relative position of the median curves for I individdal experte is similar to that shown for peak acceleration in Figure 4-3 The results also indicate that the expert-to-expert varir.bility con-
*ributes approximately the same proportion of the total uncerttinty as was observed for peak acceleration.
Figure 4-26 compares the results for the two approaches used for aggregation. As was the case for peak acceleration, component level aggregation results in similar hazard curves to those obtained using the aggregate of the 14 experts' hazard distribution.
- p. 1 t c7
~' * >
- }
8, e y. Su '
, i h,.'
fdW, ,.. 3
~ '
\ '
l ,
< casoMATRex -
c . -; , , 1 3 _ g
- b. k' f f g .
., y . ' ,7 i +
- p. -- .,
i104 U. .. . i ,, .
- i. .. .
i i . _i_ .. i -i .. .
- ~ '
, Total Hozord Shollow Sources -
_s
- - -~- Subduction zone sources - , s
[-' 6 g s s s s s 10-2' %s s
.s s
r -
- g s s *
. , s -
t . ; g s N,. s .- s s s s _ s s s . _* s s x u s s
.s C- -
- s J. -
0- . s s 7I p <
. . - s s
{ .
+
's '
8Sg tt 10-3 r N, 's -
.. . . s -
.o
- s 3
C'
.s 'N..s Q.
N' s- -
~
s 4 .
's4
,s - - -
. . s
.<: . =, =.
. _ . . . 4 it0" --
'kd' " . ,
N, 7
. . s -
. . s s r .
q s, -
-\ \-
. . . s
. - . s
. s -
. . s
. . s
. s
. . a. s
. . . s .
's s
. s
' '
- i ' ' ' I' ' '
' 10-5
.1 .2 .3 .4 .5 .6 .7
; ,3 Peok Accelerotion (g)
Figure 11 -1. . Percentile hazard curves for peak horizontal acceleration showing
. contributions of shallow crustal and subduction zone sources to
' total hazard, s
s i L'._s
GEOMATAIX 15 . i i i . i i . i . i . i 0.1 g . . 0.1 g _ c E 0-25 km O interfoce
.9 j o -
j Q 25-65 km 8 Intraslob c - ~ c 0 >s5 km - E crustot O -l -
- -.e g 5 -
i.-
- e l ..
i::
; ;2#ii
*$- i '
O 15 . , i i i . , i i i 0.3 g _ _ 0.3 g . c 5 0-25 km O interfoce
.9 j o _ _
5 o _ O 25-65 km _ G tritroslob
'C 5
g_ g -
%h ,
- m
-:- . . . e :- . 1
~
O 15 . . , , . . . i . i - i i .-
, i -- . i i 0.5 g 0.5 g '
c E 0-25 km b O Interface l-
.9 j o _
E _8 25-65 km . _ 8 intraslob -
- O m: -
~
t ~O >65 km ~ E crustoi
-m -
M5
== ..:
j:; .:e
~
~ ~
~
f!!!!!,!
# #....#;: r[
..... ... ...... 7:.:' ' *" ' " ...
o 4 5 6 7 8 9 10 4 5 6 7 8 9 10 Magnitude Magnitude Figure l-2. l Contributions to total hazard shown in terms of' magnitude, distance, and source contributions. l
; r- ;7
.. - - . . Q, k ?t /.'
U
, osoMAtsetx tv :. 9, r-sy- )
S, , i G' , [
>~ ~
10 ,; ., . . hg . i
- i. i.
s . - i: , - 4
\. .
i 10-2 .
's. . : t l x
; 4 x.
-8
.e
.)
3 10 ' g - g -
--t- t
~ E ~ 10 ~3 -
11 c - - ' E {. s f 3 1 g 0 :10-4 - 14 g
- 5 2
a _ 5 12 13 7 m 10-5 ' ' ' ' ' ' ' ' ' ' '
.1 ,2 .3 .4 .5 .6 .7 Peck Accelerotion (g)
Figure 1-3 4 Comparison or median hazard curves or individual experts.
l! ; g~ ;. .d - 7
.g
<Ei y :,.
[ -
, g' ,%% ? , . i
~
w,',7>s,.y,-
;g . .
, y . .j< .
7 r.Q'-;
, 3- EE _X
'.~
g'.g' 2 j. c.Q3l ~... i ff_ Wfq. (- f: .
. ' _- 7 fx;". ~ , Q; - )
- : .:: q. . 3 ^ .. :3:
g
-33 ;. : ' .
ga
>. ,f; .--
. . ks :
~
- 4
( n 3- o 3 ;;
. ,A:'e ' ~ 1 . .
1 5 it e'
.- o m ._$:
. t 6 t r
;71 '
,^;, r e- r
,- - p \ le v e x
Q;yC , .;, . A E E-A c. c k s yf ' g.,.' t ' c t
< , - e r l- -
P
... \ .
_. : ;h(. r. ' r: : i
; . .. : ; 1 e
p Q 'r_ y, x 7 )
~ - ::- . ; : . -
..: l ,- : ,:: . g:: g5 . . . g e
7' ( n y(: - - 2 l 1 - o 1 1
\ ' 5 it a t ' , t o u y; -
i r
\.
r r
~ e e p te d p .
x e i x E E c v
-Q 6
' . ' 3 c A i d
[. , - ' k a e i n
. P
- ,. .: . . c. .l r: ! 1:: :" 35 .3 : . . 1 r
'o
- 3 . .
7 )
;- . ;- : 2- .
1 3 g f
--Q .
(
-s - .
o n o s
,c
. ,9 t 5 it e
,. I - t '
. . t r a
r v r e e e r
,s p p dc u
. , x E E x
3 c c A 6 k d a r e a P z T*y, .
- p. : ; .. .: . . 1 a
7 ) h 1 . g ( e
, n l 8 o i
- 7 ' 5 6 t 4yt ' , t o r
e r e r e n p p x le e E
=
E . c c 3 c r ?. - I A k e p - c _ e
..:" . .: 1 P h c* r. .. : ,: . :
t
- 7 )
. 5
- 1. : : g5 .. 3: . ~
g 8 5
.N ' .
6 t ' 5 ;d.r ( n o d n a 4- t r r e e p le , p . x e x c h E E 4
' . i J cA t 0
k c 5 e ,.
. . ; ~ .- 1 P h
t 7 ) 5
,. .~ g 1 ( 1 n
3 4 5 o it i t r
' . t r o r
e e p le . p e x x c 4 E E -
' 3 cA t ' . 4 k
o e e r P
.- \,: . r* . ,: ~ ; .- 1 u
g
. ,~ ; -
7 ) i
. . : g
,. 1 ( F n
1 2 o i t ' . t ' 5 st o r r r e e p e p le s x E E c
' . ' 3 cA 6
k c e 1 P E:- < .. . r. : , ;: E : . :~ ;5 .- : . .
, 3
- 5 ' 2 3 * '
- _ - - - - - - ~
0 1 0 1 0 1 0 1 0w 1 0 1 0 1 0 1 0 1 xEy,3 xy { C l il!lj ; ! i Y
~
R ,
- i, d
s 10-- i 6 4 6 i i . A i '
- i. 4
. .4 J . .J. i Expert 11 :
- Expert 1 . Expert 3 - Expert 5 : Exoert 7; Expert 9 . - Expert .13
. r - ,
h., 1s W " b,. j h..
- 5',, :. E.w
; w.
w "A :- u,, w -- . l- . , ,
- ww, -
s% - w w, . sw -
.- - " - ' .w
. {- o. ', ";, -
s 6 1 -3 .
- , r- . .. , , -
, , . s -
r s 7
- ..w ws *
. s o,, .
; ,,, : s :-
*::, s - -
-l:,
- - .s-
,,, .s - -
,o s ,, . . .
- s ,, - -
s.. ' D
**o. ,, -
10 -' :- N'g..3
- '! , e '.'2 ' -
- ,?, , , :- s, , , r
;,, , ; m. . . - . ,,..
,s- ,, . .
- ,,, w. ...,.
,5
- , w, . -
- .-u,,.,.
. %. s. s
.s - - .
ss .,. . -
' ' ' ' I ' ' i e i . i t r- t' 10-5 m-' s . . . . .
. . . . . . . : .~ . 2 L
- Empert 2 Expert 4
~ ~
Expert 6 : : Exoert 8 Expert 10 : : E.mert 12 : Expert 14 . l ( - , . . . r . . y-10-2 7
, g , g ,
4. p, . s
, gg, k'
3 :w w,, -
- r. w, . u, .
s.o. w .
. .sw .
- w. - .
. sw ,
.s., . ws. . . ,,u,, ,
. .o .
- . w. . . .
*s
.w ws .w
- . .,n%s . .
% .g %%% 4%g t
gg 444 g % %% %s % g %% %o.% r " 10'y :- u -
."o , --
w
, :- w's ,
sw = :- ,o,, u 's';, ,
; '!, , ':s,
,, . w 2
,s,
?
w, w,
- s. ,ss . ss g
gg
. gg. , _
g%. g ,% ,b, % %.% . . . ,, ,, = . ,,, . s.s w.
,- - g .. ,,
'w'*
10-* :-
' ,' , , :- '?!*
,s,ss. :
- - '20.
,u,
'2.*'.s
,ss.
'?,.','-- , . .
'2 s : , e s.
u - s ': : sss : , * *:
,. 9 '.,.,'s: ; ,., ': -
sss. .
, ss os =
, - g gg.
- t g4 -
ggg g q g v, s o*. , s.i s -
- g4% g
% g g t t
' - 1 i t t .'. . t I t . t
- t t . t t "
79.s
.1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 '.1 2- .5 .7 .1 .3 .5 .7 .1 .3 : .5 .7 .1 .3 .5 - .7 Feck Accelerotion (g) Peck Acceleration (g) Peok Acceteretion (g) Peck AccelercSon (g) Peok AcceleroGon (g) Peck Accelerotion (g) - Peok Accelerot;on (g)
S Figure 4-5 15'". 50**. and 85*' percentile curves for. interrace (solid curves) and 0 intraslab (dashed curves) sources for individual experts. 4 E x ,
v Lt - G GEOMATAIX f 10 -' . i i i . i N Aggregate over 14 experts f - ss ---- Median for individuct experts - s s e s _ k g s I\g s I% s g s
-%% s s s
%% s s s 10-2 _, ggs ,s s
\
- ', Vb's 's :
_ ss ts y s 's, _
, s eps's N -
_s s s s g.s ss
's s 's s _
-s s s
s g ss scs 's,
' s s -
>s 's
\ g%g'Os N 's 8 -
s s 10s 's 's - e s s s sN 's n 3 rr s g ,tt 'is sss s 's s 's s 8 10'3 L.
\ 's g
*t e'IN os s
's s s
s s
's 4
9 o,0 s s s ~,, s
's s Q Y -12s
's 's
's s 's
~
c - J s s
's s. Q@c es s '12 s
s
't ss ss,?s ' s' s s, s ,' s _
s s s - s s s q s, 'e s 's"~s,; s ' #s' s s Ns 'g YQs 's N"'Ec s s 8t4' # -'ss s, q 's %N '
'g 4 Y s 's N's
%, 44 ::,Rs ,
1
* * 's '
10 ~' 's *
*g N' ' -
_- \ + 4 4 /g's , 4 ,ws as- s s _ 's s
- 4
.g ' sss s_-
_ s t g s s 4, %
- s t e s t =
s s ** - s
+,k+
'% +
10-5 -
.1 .2 .3 .4 .5 .6 .7 Peak Acceleration (g)
Figure 4-6, Comparison of' experts' median hazard curves with aggregate hazard.
g. T: 4
+-
GEOMATAIX i 1 i i i
.10-1 _
i Aggregote over 14 experts
----- Composite Hozord Model f
)
-s s
> s
- - \ 's s
s r 10-2 - s. s s
- . s .
N \ _
\
s s, _ s % s 's _
- h. s %
o C _ s s
's s _
O s s a s ' cr s s
's s o
u'- 10-3
's s ' '
s _
- 's s o
3 , c
~
s ' o ~
- s , stg _
s
- s .
s
% g %
- s s
'n 's
/g,4
's ' 's 10 -' --
's s -
- 's :
N _ s, _ s s, _ s s, _
' ' ' I ' I ' ' ' l '
10-5 -
.1 .2 .3 .4 .5 .6 .7 Peak Acceleration (g)
Figure 4-7 Comparison of aggregation procedures for %tal hazard from subduction zone sources.
i GEOMATRIX k, I 10 -' ' ' ' ' ' ' i i i l ( l -
~
Total Uncertainty -
---- "Expert" Uncertainty -
s s 10-2 _
's, __
_s s . _ s s _ _ \s s, _ _ s _ s g s . s, s
- s -
s h s s g -
\ s -
e a s
's s C' g '
[10-3 -
's, s
's 's -
_o . s s _ a _ s 's .
- s % *
< _ s .
's s s 4- (4 s -
% ~
- 'y s,
's Gfg 's '
'l{tg 's '
10~'
- s s
s, s, . s . y . _ s _ s s y _ s s s, _ I I . I i i n' . 10-5 i i i i
.1 .2 .3 .4 .5 .6 .7 Peak Accelerotion (g)
Figure 1-8. 4 Contribution of export-to-expert uncertainty to total uncertainty. b
_q s,.:
- ]3 e
4,
. y 10 -' ' "
- i '
i > 1 I > i a
- ; I I I e. i .:
^ '
- Subduction Hozord _ Interface Hozord .
Total Uncertointy Toto! Uncertointi " - -
~
- - - S-Cubed Model , - - S-Cubed Model
---- Joyner Model - -
- - - - Joyner Modef : -
l s 10-2 #s _ ,_ ._ _s : c
- 's : t : - - 1
\s _ _ _
q _ . . _ g
>s _ % _ _
o % g _
\s _ _ _
o % CT
$10-3 r 'N s - -
r - ; 15 : : : : a : 'N 44 : : 44 : } 4 -
'N - - -
. g . _g - - ,
e . . i t . 4, sN _ _ t t, Ns
't.
10 " : - \ ~
\(N s -
's s . _
s s _
\_ _
- 's : g'%. :
_ _ _ 't. .
'4.
s.
-$ e I e l t l r l s l t t ! a ! t ! t ! % *! t
.1 .2 .3 .4 .5 .6 .7 .1 .2 .3 .4 .5 .6 .7 l Peck Accelerotion (g) Peck Accelerotion (g) 5 Figure 4-9 contribution'or uncertainty in subduction zone . attenuation !
relationships to total uncertainty.~. % X
4 r
~
(< . _
- GEOMATRIX 4 .
t - U; l1gc 44s s i . d l ' ' l. l: .i 8 l '
.[
i '
,. f
. .1
. Expert 1 -
Expert 4 Expert 7- Expert 10 : ' Expert 13 : J
\-
'10-2 A - ,
;, . s, - ., s .,
- w. .
N ,
.', s
~'s, . i,
= - .
'N {
- gn ',
's, 's ,
Ec 10-3 r Ns , :- Ns , r g , --
.N , :, ;
, s s 13 's ', . '. -
N, , .
.g N ,
N gs s, g , 10-* :- N,," :- 'N , :-
'q :- :- -
's ,
\ ,- 'g a .
j s, 10-5 10-' , , , i . i . . i . L.
- Expert 2 Expert 5 - Expert 8 . . Expert 11 Expert 14 ~
7 l' N, r 10-8 > s i 7 i 's, .
's, s,
, ;s
's, N, ', 's,
{ c 10-3 , :- -N, , r \, , r
's , :-
'N , :- ,
I; N, N, : N, N '. . r- .
- s g . .
.% N, N, -
'N -
'N, ',- -
10** :- N :- N' :- N: r N ,. t :
. s .
' ' 2- ' ' ' ' ' ' ' -' ' ' #
10-5 10-' . i . . , i , - i.
. i . . , . .
i . i .
. Expect 3 . Expert 6 ! Expert 9 - -
t xpert 12 i Combined ; 10-2 '
-{-
's, ,
', , , . 's ,
3 N N g s g' , ,
, 'N g c 10-3 =- N , r N , :-
N
's, s
~
j N, ,
's N, , 'N s {
c s , . s . s , , ,
.' 10-* r 'A :-
\q :- , e :-
'c;
'N ; .
10'S ' ' ' ' ' ' ' ' ' ' t f
'.1 .3 .5 .7 .1 .3 5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 Peok Accelerotion (g) Peck Accelerotion (g) Peak Accelerotion (g) Peak A:celerotion (g) Peok kccelerotion (g)
Figure 1-10. 1 Contributions of uncertainty in subducting plate geometry to total uncertainty. Shown are the 15th and 85'" percentiles considering all uncertainties (solid lines) and the 15*h and 85th percentiles considering only uncertainty in plate geometry (dashed line).
- b $
. CBEOMATRIX -
~
$l # -
t 10-8 i .
, i. .i e i _
i , s i. i
~
. Expert 1 '- Expert 4 ' 3 Expert 7 Expert 10 : - Expert 13 i
- s. _ _ _ .
s, '
.x 10-2 ,
s k
-{ ,
. g
's, 's, L
- c ' iO g. .
~
, r ,
, e , :,
, q l
'(
g , . , = , ,
\', - '
s ;
. ic-4 '
, - ', 3 ,.
7
',2
's: .
s 10-5 i ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
~ 10-' .
- i. , i i i _
i . ; . i Expert 2 ' Erpert 5 : Expert 8 - - Expert 11 - - Expert 14 : 10-2 .h ,
. $' ?.
.s
. 5
- i
- g. - - -
g, i , a 10-3 r \ ~ s
, 9 r ', 9 :- ', !
, i
\ '
}~
',', s, s
.4 -
10-* r 'N- - r ','
's ,
'Q .
's, g, \ . . ', , . .
' '\ ' ' ' ' ' ' ' ' ' ' L l- '
10-5 10-' i i . i i . . i .
- p. i .
i . . Erpert 3 Expert 6 - Expert 9 Expert 12 . Combined 5
, \ '
10-2 ' x s, , (, - 8 ',
', u g s,
, gs, g '
s ' C 10-3 :- ' 's m :- r N, , e 's :- N's %
~, ~, 's ~,'.
-p s, ,
s s, -
,s,
% w g . % , - % 4
-4 s, ,
N*,
% g % %
10 -' :- - 2: -
- - 's, s :- ', S
- u<
10-5 ' ' ' ' ' ' ' ' ' ' ' ' ' '
.1 .3 .5 .7 .1 .3 .'5 .7 1 .3 .5 .7 .1 .3 .5 .7 .1 .'3 .5 .7 Peak Accelerotion (g) Peck Accelerotion (g) Peak Accelerobon (g) Peok Acceleroth,n (g) -ecx Accelerobon (g)
Figure 14-11. Contributions of uncertainty in source activity to total uncer-tainty. Shown are the 13th and 85th Percentiles considering all uncertainties (solid lines) and the 15th and 35'h percentiles considering i.,nly uncertainty in source activity (dashed line) .
..4
- n. .
,g. - _ -
Ol
$~ *e-
!' , c3EOMATRIX - r - g 10M' ,. , i i, i i_ . . i. i , , Expert ~ 1. - Expert 4 - Expert 7 . Expert 10 : Expert 13 :
.\.
! 10*2 \ ,
, -\ , :- =
we- Ns N \, 's, _ ', . s , s , a 10-3 -- N -- N , -- N , r N , -, :
-c
}' N, N
's, s,
- , N, s,
's.
\,
s,
-4 g ,
g - , 10-* :: Ng -- Ns --
'q :- : --
N : g ;
'N '
- 't I . I - I - 1 - I - . f - t . - 1 -
3c-s - 10-' i i . i _ i i , , i i
. Expert 2 Expert 5 Expert 8 Expert 11 - Expert 14 {
10-8 m - m , , o'Y,
= '
- :s : , ,
7 :g :
\
- ; . :s .
I \ \ \- 's - $ g s c .10-3 r 's , r \, r - : -
, :- N j
'g 'N . . :
, '\
's 10** e-
..+@# e, r 'sh r
'g N e
\'
\ '
':. s- % ,
ss , . s 10-S 10-' .
, , r, , , , ,
i i ;
. Empert 3 Expert 6 - Expert 9 Expert 12 . CArbined 3
.10-2 , 'g , 'N ,
g- , . '\ ,
's, , , i :
s 3 , , , g ,+, s , s, - N - C 10-3 :- , e ' e - , e 's , - 's
+c#c ,'e \ ' \,
~
' ~
[ g ,, ; .
, j
.{ +g
'+e, . '.-
10-* --
'A :- 's' c
'J
+:
- - 'c!
' +
- 1 10'S ' ' ' ' ' ' ' ' ' ' I ' ' ' ' '
.1 .3 5 .7 .1 .3 .5 .7 .1 .3 .5 ./ .1 .3 .5 .7 .1 .3 .5 .7
- Peak Accelerotion (g) Peok Accelerotion (g) Peak Acceteroton (g) Peak Acce)votion (g) Peak Iccelerotion (g) i l
Figure f-12. 4 Contributions of uncertainty in source segmentation to total l uncertainty. Shown a.e the Irth j and 85th percentiles considering i all uncertainties (solid lires) and the 15th and 85'" percentiles ! considering only uncertainty in source segmentation (dashed line). I
r-GEOMATRIX I 10-' i ., i i , i i , i , i i i . Empert 1 - Expert i Expert 7 Expert 10 : Expert 13 - f . ., r 10-2 ,
. s, _, _ ,
- s x ,
y ', ', : i 's n,
. -
- N g ,
's,
+
g 's N, , e, c 10-3 :- N : - '
, :- +
,,~
, :- :- - m 0 % s :
- s, .
g , ,
~O s ,, ',,
4 \ 's, *- N. s
, , -, . \
10-* :- '. , --
'N : --
's ,
' .; N N ;
~ ' ~
\
(' . .
' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' \'
10-5 p 10-' , , , i i , , , i i Expert 2 Expert 5 ' Expert 8 Expert 11 Espert I t ; 10-2 i. f;, , g
" \ \ 's g '. ', -
N, s, . j c 10-2 N, ', ', ', - '
. ,\,
- - , , , r , , - - ,
iC , a
- s. '.
E ' " '
'., 'Ny\ N 10 --
., =- ,
,- r
.; r , r s,
*4-
\' \\ :
N \; -
' ' I ' ' *-
10-5 10-' , , , , i i i i i r- , , . Expert 3 Empert 6 Expert 9 Expert 12 . Cornbined i 10-2 's _ s, .; . , h , , - , , s y s, s, N, N N ' l g '
, N
~
N 's N C 10-3 ' N , r N , l y r , s,
, r s,
, r
=-
N, , N s N '
, % s -
's, N, 's, '., 'N h ,
~. . . ~.
10-4 :- A --
. N:
d N : 10-6 ' ' ' ' ' ' ' ' ' ' '
.1 .3 .5 .7 .1 3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 7 Peak Acceleration (g) Peok Aceleration (g) Peak Acelwot;on (g) Peok Accelerotion (g) Peckdcceteration (g)
Figure 4-13 contributions or uncertainty in maximum extent or rupture to total uncertainty. Shown are the 15th and 85*h percentiles considering all uncertainties (solid lines) and the l'jth and
, 85th percentiles considering only uncertainty in ma<imum extent or rupture.
~ -- - -
i..
- GEOMATRIX e
+
-n Ib . - I .i 3 i ' s i . a i - i -
. Expert l ' " Expert 4' ; Expert 7 ' g ., Expert 10 i -
Expert 13 ' 5 . i
'10-2
.., . , s, , e, ,.
.x i s,
- t -
8- ', : ',
. i' , . .
3 ,,
~~
.a - ':, .
,e a 10'3
'~
'J- , , -
'0, ,
s,e* ,
-3:
sc, .*,*, - :
-- 's,, ,
' 0,7, - *.6
\'q
~
t o-* r :- , - :- , :. 's, , N . 10** ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' .' ' h 10-' , i .
. . , a i.
~3 ,
Expert 2 Expert 5 ~ . Expert 8 . Expert 11 : ' Expert 14 : r s t - i ' 8 , , I ,? f l .h' '
- .E {'
L' -, 's : : . l,, e
't,,, -',- -
- ,, s. ,
c .10*8 :- .':, , :- 's , - 't, , ;- -, :- , s ,, . N, 't,, , l' y' ',,
?,, , ', -
\ _
10** r ' ' ,' ' , , r
* - ,g
- r . r c,+< -
4 J . ' g.s ic i i e i . . . i e i . o_ . . i 's 10 , .i . i r-- i . g , i . i . Expert 3 Expert 6 ~ Expert 9 Expert 12 : Combined i 10-8 .
~t
. h- lg -
', 't g - 4 g ',
C 10-3 r :- 4 4 7 r e,+,e 7 3 g
%g ' ', 0, -
s, , I . . g , ' O,, , N g*t g
'a
'0, 0 10-* -- \- -
- hq -
3 10-5 ' ' ' ' ' ' ' ' ' ' ' ' ' '
.1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 ! Peok Accelerotion (g) Peak Aceterotion (g) Peak Acceteroton (g) Peok Acelerotion (g) Peok Aceicrotion (g)
Figure 4-14. Contributionn of uncertainty in maximum magnitude to total uncer-tainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lir'O and the 15th ar'd 85t h percentiles considering only uncertaim , in maximum magnitude (dashed line).
w , r., .; r. . - ;--- ( * ',3.-: , t v
,:~
J..- U y E is -
+> - gagmrpngu .
cti
~
! 10~3 ,
i .
- Expert 1 Expert 4-Expert 7 Expert to ! -Expert 13 i-4
.10-2 ', f, m .
,. > s, g "
;f
, . ', 's , . 's, ,
7 N' g . F
~d-10-3 :-
- - 'N ' , :-
'N - :- ',
-.g O s s N % s N , 5, ?, s,
> ." J - .
- ', , N s,
's .
s
~10** r 's , --
' e: :- '
N-N,: - - - , N. !
's N.
e 10-5 i ' ' ' ' ' ' ' - ' ' ' ' ' ' ' ' ' ' [ 10 , , , , , Expert 2 bpert 5 - Dpert 8 Expert 11 - : Expert 14 4 . . .
'30-2
, , ? , 4- , ,
', E\,
, w : . , ! !
L -y_ :4 e, s
;. N, '
4',
- g. ,
+
. ', s , , .
a 10-3 # .,*e , ,
- - 0,
.se 3 -
s e
.3 r-
's, ,
r
'N s , c
's s :
3
, - . - 0, -
N ', :
- =s
. f. e,,5 N. ,
' .t 'ce, 'e,,g . ,N . 's, .
N .
. t o-*
~
sq :-
\ch.
'Nh , #' , r. 'N ,
7 ,N: s ! sc Q
': N- Ns :
, - N. .
' 4-
-h . A F A k __a A 10*' , .
Expert 3 : Expert 6. ~ Expert 9 - Expert 12 i : Cornbined i
. o g ,
'10-2 ,
3
~
g , ', , .
\,
N'
's, N, .e ,, . Ns .
s s h , s e 10'3 :- N, , r '00, :-
'N -, :- 'g :- \, =
N 'O, s
', 's s, e, -
, g g
. N 'O, . % Ns 's N 'O, NN 10'4 r A --
'Je
')
r , r N: :- 'o!
- 'u, 10*S..1 '.3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 Peck Acceleroteon (g) Peak Acelerotion (g) Peak Aceleration (g) Peck Acceleration (g) peak 5ccelerotion (9)
J Figure 4-15 Contributions of' uncertainty in recurrence methM (moment rate vs palcoscismicity) to total uncertainty. Shown are the 15th and 85th percentiles considering all uncertainties (solid lines) and -l the 15th and 85th percentiles considering only uncertainty in 'l
. recurrence method (moment rata vs palcoscismicity) (dashed line). '
;. I
, . g w'
- . +
, CBEOMATRIX e
I T 10-f , ,. , Expert l' Expert 4 : Expert 7 , Expert 10 : - Expert 13 : 10-2 , ._ , p ,z , ,. , x -4 : 1,
- g. _ : - 't t y . -
z%,1 a, . ,,
'00, st 10 3 :-
g\ 3 - ,, j \'O,i, . h '. '
- . '?, % .
, 'e, 10** :-
's, :- r , :- :- -
'}
=
10-S k ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' 10~( m , , , , , , , , , Expert 2 Expert 5 ' Expert 8 Expert 11 Expert 14 i ; { { 10'3 , * ,
- 1- - r x 3 : (
g.' :. ' ,
~ ~
'e,, '
.: 10-3 :. si, , :- - ',j , , :- ',,
NN . :- , g3
'e, :
,, . g :
= ; ,
'e, ; 'O, "
e%+
~ .
~n0 ,
, 'O, ,
10-* :- - *<, 'Or
'Od -
's. -
4 . . gt b , 5 : . . g w = l 10'S ' ' ' ' ' ' ' ' ' ' ' s' ' ' l.
+ 10-' ,
3 Expert 3 : Expert 6 - Expert 9 . Expert 12 . - CorrNned i
- l. L
[ _- 10*2 g . _. , , _ n l-g e,, : e, e,# C 10'3 . - - O , :- , :- , - 'O, ,
'O,
' 0,0,' '
'O,
'J, 'O,
'Or -
'?, 'O,
'Or 10 * * :- \'i :-
N',7 :- : -- ',
'1'3 10-5 ' ' ' ' ' ' ' ' ' '
.1'- .3 - .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 3 .5 .7 Peok Accelerotion (g) Peok Acelerotion (g) Peck Accelerotion (g) Peak Acelerotion (g) Peak Acelerotion (g)
Figure 1-16. 1 Contributions of uncertainty in convergence rate /palcoseismic rate to total uncertainty. Shown are the 15th and 85th percen-tiles considering all uncertainties (solid lines) and the 15"' and 85"' percentiles considering only uncertainty in convergence rate /paleeseismic rate (dashed line). I
W " :p9c.
~' '
. - ~_
+
. l
.13EOMATRIX -
Y Z
-10*'- . . .
i. i i , , J Expert 1 Erpert 4 : . Expert 7 Expert 10 ' ' Expert 13 i n-- - 110-8 y . x.
's 's Lg 8 ',,
'g g, ',
JQ 's',, \ 's
- d 10-3 r ,
, :- N , :- , ,
', . N N, 's .
g % , g g
.E . N . N . 's, 's ', .
10** r \ :-
'N --
\q - .
e- 'N ]l
',: . \ a
\s 10-5 ' ' ' ' ' ' ' '
.104 i ,
i i, i i i , i i , 3 _ , Expert 2 . Expert 5 : Expert 8 : Expert II : Expert 14 - V 3 i
- y. . ,
. .=
g ,
, h \
g-[ c '10-2
- 3- ,.
i , t .
.. N.,
3 y : irrs ,- 3
., e :- e, , h e \,; r \, ,
- s. \., z,
'$ \ :
. , i
" +
10-5 -- 10 r -- i i . . i , Expert 3 Expert 6 . Expert 9 Expert 12 . Comt>ined i 10 * * : : ', : ,
\ 't .
Ns .
'g -
's : '*4 s
- s s g
' - ' s
' s s
$ \ 9 ,
N
~
fp '
\ 's '
[ e N \ ,' C'10-3 r- N m r O',
, e-N m :- ,
~
r-N' s N
.. 'O, N, N ',
g sp, . g .
's 'Os
- 10**
'A ;-
\',',q r , -- - ,
*q 10** ' ' '
.1 .3 .5 .7 .1 3 .5 7 .1 .'3 .5 7 .1 3 .5 .7 .1 3 .5 .7 -
Peck Acelerotion (g) Peak Acelerot>on (g) Peak Aceteration (g) Peok kcelerotion (g) Peak Acelerotion (g) Figure f4-17 Contributions of uncertainty in seismic coupling to total uncer-tainty. Shown are the 15th and 85'h percentiles considering all uncertainties (solid lines) and the 15th and 85th percentiles considering only uncertainty in seismic coupling (dashed line). t
, . - . ,, _ . = - - - -
3 ._
>i_ ;. ' [. ! y...
GEOMATRIX I: 10 " ,, , , , i i f;}
- i. i.
Expert 1'- . Expert 4. - Expert 7 Expert to : Expat 13 :
~\
10-2 , s, , l, ., ., , ,
.g\
N
's, .
..3 ', .
g g,'s h ',
-t 10-3 i
I ? . ' '
, g '
- ', ; .g -
3 ; j Ng s i,
. I e
10-* :- - ',,: -
- - s, s, -
'q -
N g : I: ': : ', .
-$ a l t l a a l A l t a ! 1 l a a ! I l e a j n t a 10 -' , 3 - , , , , ,
Expert 2 Expert 5 Expert 8 Expert 11 . Expert 14 { (10-2 , , , k , y ,
- g. L 'i, .
'i, (* j p\ -
2. g s,
.\,
.4 n U
-- %,4,g .
i , , it 10'3
. s\ , :- \' , --
, , r \ 7 j s ',
< s, : ;
. a s,,s s,s ss r , ,
10-* r t'% :- N, e 'N; r ', r 'e h{ -
'h ,"
- 1. - s .
' ' ' ' ' ' ' ' ' ' -' ' ' ' W' 5
10 ~.' , , r- r, , , i i, , , i , ,
~
Expert 3 Expert 6 i Expert 9 Expert 12 Corr 61r e4 i t 10-2 , y _; , 3 _; {
,0 , * '
3 ,
't e
4 it 10'3 :-
, , r-
#c0 , :- , -
% , r- \* ::
s, -
+,c,e,
'O, ; 't, : %4 ;
s 'O -
' O ,,
's, 't, 10 :- A -- 'e s :- : :- +e#e, : :- ' r:
f ., 'O, !
- s IO'S ' '
,1 .3 .5 .7 .1 .'3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 .1 .3 .5 .7 Peak kcelerotion (g) Peok Mcelerotion (g) Peok Acelerotion (g) Peak Aceleration (g) Peak' Acceleration (c)
Figure 1-18. 4 Contributions or uncertainty in magnitude distribution to total uncertainty. Shown are the 15th and 85th Percentiles considering all uncertainties (solid lines) and the 155h and 85th percentiles considering only uncertainty in magnitude distribution (dashed line).
7.
.. y 3.. m_
4 10~' :
- I '
t I t I ' t
- l. -' I I. .:
~
- : : Medion hozord for: :
Sho!!ow Crustal Sources
-- Totot Shallow Sources
~ - ~' ~
---- Subduction Zone Sources -
1.ocol Random Wedge -
's 's -- Puget / Willomette 10-2 's's crustot rectures A, ', 's
, s ,
- , , s - , -
- , , s -
s s s, _- . _
% s g
>% . s s s - - -
u s s s S ', 's s N 4 . a CT s ' 's s o ' s s ' C 10-3 - ' N - - -
's s 's s, 's, : :
f i o _ . s. . 'l 3 - s s 'g - - C s s s .- '- - .
'944 l Q - . .
t _ , , s s _ _ \,' _
's, 'sN, ~. . .
s , t . s
' s -
\ .
's s, 10 " -
. s,
-\
s s , . . - o, p s s x _ s, . g\ . s, .
\
, i -
e+, s l
. s g \ .
l s, . . .
-$ t ! e t ! e n i % 9 e ! # ! t ! a b ! t
.1 .2 .3 .4 .5 .6 .7 .1 .2 .3 .4 .5 .6 .7 l
l Peak Accelerotion (g) Peck Accelerotion (g) Figure 4-19 Peak acceleration hazard from shallow crustal sources. On the left,.the 15'". Sot" _and 85'" g
.s
~
percentile hazard curves for crustal and subduction zone sources are compared. Shown on the right are the median hazard curves for the various shallow crustal sources. 4 9 x
w w 10 -' _ ,
~ "
_ Attenuation Relationship : : Maximum Mcgnitude : - Recurrence Rote .: Campbell (1987) _ ---- Joyner cnd Boore (1982) , _ _ _ ,
- - Geometrix (1987) 10-2 _
= __
= _
=
g _ _ g . a _ _ o- i ,, c> C 10-3 \ - -6 i
,: s's -
- \
_8 : : : :y \ 's, : c \ . '. E g .
. J _
,\ _
- - - e s
- - ~
10 " p
\ +
#s \\
g g s, g 7 i N g
- \s y\
- ' \s \ s :
s ++e _
# \
N 0, .
\
Q 'sN \ g s gs, . _
\s _
s
'N ' '
s s 10 ~5
.1 .3 .'5 .7 .1 .3 .5 .7 .1 .3 .5 .7 Peck Acceleration (g) Peck Accelerotion (g) Peak Accelerotion (g)
Figure 4-20. Co'tributions n to uncertainty in hazard for shallow crustal sources. The solid curves in each plot are the 15th and 85*h percentile curves considering all uncertainties. The left plot shows @ the median hazard curves obtained using single attenuation relationship. In the center and right 0 plots, the dashtJ curves are the 15th and 85th percentile hazard curves considering only uncertainty in maximum magnitude and recurrence rate, respectively. f, 3 R
a' 10-' :- ' i ' i ' i '
'l '
l ' I t. i '
.4
- l T = 0.15 sec : :'. T = 0.80 sec ': : ' T = 2.00 sec .:
't
. Total - - Toto! - - Total. - , 1
- - - - - Subduction Zone Sources- ----- Subduc%n Zoce Soerces . .---- Subduction Zones Sources g
- - - . ' Shallow Sources , *
- Sho'3w Sources Sho!!ow Sources
,s s
10-2 , 5 5 :' y $ g - m - s gg
- s g
"g g g g - s -
% r g g g f g
* = - .-
- g g \ g
' ' ' g -
x , . . s
- O .
s t C . ' s g
\ "
i T g
"[* % %
Q g g . 3 g - % g Q' **. g' % g .
% g
$10-3 h/; i N ,
2, 's ', - _s
's g
-- .** % = - -
g g O g * - - . . g s
- .1 , *g 'g .-
- ,"- s -
.- s -
*,.',g g .-
\ - " *'
g c - , l. g s . s . *.' g - g -
., s - -
.
- g g g
. ... %. g g
. . . , , . . .g ,
,, , \ s ,. g . . ,. ,g
\
#4 10 -' --
\
t., N, f,
".,'s, e --
. . - g s . s . .
,-.sg. s .
,l-s s . . _
s s Ns .. .
,.g.
,s N,
/,q ',
.r- N, .
.--.' ' .s '
.'. 1sth s
s -. ... ,
- .r s ,
' ~ ~ ' ' ' ' ~
Y(g.' , ' .
\ '#; , ,
\ '# ,# , pg '
, o -s . i .
- i. .t . s , i . .r .. i, . . i : .1 , .s 5 15 25 35 45 10 30 50 70 90 10- 30 50 70 90 Spectrci Velocity (cm/sec) Spectrol Velocity (cm/sec) Spectral Velocity (cm/sec)
Figure 4-21. Hazard curves for 5% damped spectral velocity at periods or 0.15, 0.8 and 2 seconds. Shown are .g total hazard and hazard frca subduction and shallow crustal sources. s B X
- y -w %,.- e w
'N 10 " , , , ,
3
, g i g ,
~
- : . : 1 Expert 1 Espert 3 Empert 5 Expert 7 : Expert 9 . : Expert 11 : Expert 13 :
10-2 "- [ -
., q , , q :- - :
'h $
~ ~
3
~
, [ [ [ .
cir r , r , , , - , c. , , , . ,
* ~
. ~
10-* :- , r , r 7
- - N, i-p r
. [ I I
10-5 10-' , , , . , g , g , , 5 ., 1 g
. Empert 2 Empert 4 Expert 6 Expert 8 E= pert 10 : Expert 12 : Expert 14 .
\
10-2 & , , , , , g . 1 l (g i -
- : : . . E .s
-s a 10-3 r
\ . .
r ,
- r. , ,
~ ~ ~ ' -
5 . 5 5 : 10-* --
- r .
, :- , r ,
r :- - , r , - 10-5 5 25 45 5 25 45 5 25 45 5 25 45 5 25 45 5
*~ '
25 N'.: 45 5
' 25 45 SV (cm/sec) SV (cm/sec) SV (cm/sec) SV (cm/sec) SV (cm/sec) SV (cm/sec) . SV (cm/sec)' -
R Figure f-22. 4 15'". 50th and 85'" percentile hazard curves of individual experts for 5% damped spectral - velocity at 0.15-second period. x
m m. c, . rw lai w 10-' , , , , , , , 3 h-
. - . : . . : : . . c ,
Empert 1 : Expert 3 . Expert 5 : - Empert 7 . Expert 9 : Expert 11. : : . Empert 13 . 10-2 , , , y Y ., N
~~
h
- : : 3
-. . - - - - . . e 10-3 r .
- - r -
, r. , ,
- 5 : E i !
10 -' r , r , r , r , r , r , W , _ . : : q_ - 3 i A f f f I
$ps f f i i 3 4 4 ] ! l .! I d 4
Espert 2 Empert 4 Empert 6 : Empert 8 . Expert 10 - . Expert 12 : - Expert 14
' ~i 10-2 y , , -. - , ,
3, , l7 . 5 a 10-3 -- :- -
- - , r
- r. :
) : : 4 : : : : . 10-* r - r -
- - r r , r 10-5 to 50 90 10 50 90 10 50 90 10 50 90 10 50 90 10 50 90 10 - 50 90 sv (cm/sec) sv (cm/see) sv (cm/sec) sv (cm/sec) sv (cm/sec) st (cm/sec) sv (cm/sec)
\
l 1 Figure 4-23 15'". Sot" and 85'" Percentile hazard curves of individual experts' for 5% damped spectral . g velocity at 0.8-seccad period. .
.X
ll R EX g y , . ;- : ; 0
- ,. : , 1 -
l 1 - 9 r 3 4 ) 1 1 ' c e t t r s r / e e l p x i p x 0 5 m c a m E E. ( v s r t c e p
. 7. ' :( .- ; - r 0 1
s
- _.~
3 ,.. . ,:: . l} .
]i . ;5
- 0 9
d e p' 1 - 2 ) c m 1 1 e a t t r s d r e e / p x I ,
. p x ' 0 me 5 %
5 E E ( v s r o f p:. -;f
,:: . 7 . c:.. :: .
- ci. . - 0 1 s t
0 r 3
.. , 1 .
l . 1 -
-; 9 e p
o '
)
c x-9 t t t r e s e r / e e l p x t , p m
' 0 5 m c a E E ( u v d
' s i v
i c r - .' c : r 0 d n 1 0 i
,. .~ , 1 l ,
i . 9 f 7 8
)
c o e t t s s r e r e / e p x p. x 0 mc 5 v r E E ( v u s c d
. c- { r o r t a 0 z
. . .~ , . ; . . - : . .
l; .
]E ' .
9 a h 5 6
)
c e e t t s l r e r e / i . p x p m ' 0 mc 5 td no E I ( ei v cr
' s re ep P
_, c . . , . ;' . 0 d 1 0
" n o
- 3. .: , . .. : .
l ;. 7 9 '5 c 8e 3 4
)
c s e d - t r t r s n2 e e / a p m s t , p m 0 5 mc t E s E ( h a t v s 0y 5t _( c r . : 0 .ic o 0 1
"l
'5 e l -
7 . . 9 1 v
)
1 2 c - t t e s . r r 4 e e / p t , p x ' 0 5 m 2 s c - E E ( 4 v s e r r; c- . - [ 0 u g
. . : - 1 1
' 2 > ' s ' '
- S i 0
F 0 0 0, 0 00 1 0 0 1 1 1 1 1 1 1 y{e 8fC iI l , 1l l ll w
, - m.
s 10-' : '
,1 '
I 1
-l
- 1 I
.1
- 6.
- : - :T = 2.00 sec
- T = 0.15 sec : T = O.80 sec _ - .' :
L Aggregote over 14 experts _ L Aggregate over 14 experts _ _ Aggregate over 14 expertsi _ . s s
---- - Median for individuct experts _ s,---- Median for individual experts . , ---- Mecion for individuct erperts _
_ s, _ g
- s s %
i, s , s s s s 10-2 s, \ _ . . \g __ - h
=_ =_
s, . s s, _ . - ,. _- _ s _ 4 '
- 's, 'N s
- _kiA s.N, ,
s
- k'#p ss 'c ,
,?, \ N g 's
- Is wi' t 't,w , ,
4 0\ \ ',,
's.s c _t 'y 's , ,, _ 2 wt$ u ,. s, _
s
- s'p ,'\* ss o 'N, _
g s s, % s%' , . % sQ % %' . o . y* s s 6 s+s,. , . ,. ',%,', s oN - s 's.
$10-3
%,,Y '?f, , ' ' ', ,#.+', s' \
\, \ IN s,
's \,
s\ s
--', n _- s - - ; -
'i, :,- .-
__o s s s s w s , . .
, % + -
,s, ,
l
,ss s ,-* sgg -
% s ,
a _. i. 3 s s _ _ s ssgg '
,-w. s
, g, ss ,g + ; g.s , - ,s ,- ,
\, ,N ss i
C . g g g _ _ t 3_ g g
, 4 s q
- s, g
ws, sws s - s 1s - s J _ i w 3 s _ _ i _ _. i
. . -2 ,_._
4
, ,ts., %s ss ,s +, %
i s s N, ,- , w %w ss s _ _i % , s 4 wwssss s s ss s ;e , s sss s s
- i. ss %,Ut
.ss -
i s 'g s 's s t g
', w" NNs;, *g '
- . s i , 't. g ++s ,.s N s t%
s s s g ssw t , t g
, O@O -
s g g s s s s s
-A s "s ssgw g g sg 4
-g g s%
g sg A~1
\ s',
10-4 - g s s',\ \s N _- ', "sng't, g
+,,;
s,s;, N,- 's -, s; s w s s,_ i. - s ss' .wo . -2 ss s s -, w ss s, s s _ , e,
, ss s s. _ s w sss s :Os : . ,-
s
'1 s, tsw.
? ,_
s, go s g _ s go s, - - ss% - -
, 4 ,
s s%% s go s _ _ .s - - -
\ , s.
gu g g ss ss g ws\osss-..s;&; s gu ,, _ s _ -
, s N :,- ,
os , - s gw , , s % .. s s, ' s _ tw s tw
. _ s 's s s _ _ .- s's s s ..s- g gw s s g s s s s
.s s,s
- se s . 1s
, ,s w s i Ns 10-5 ' ' ' ' ' ' #1 '
5 15 25 35 45 10 , 30 50 -70 -90 10: 30 50- 70 901 Spectro! Velocity (cm/sec) Spectrol Velocity (cm/sec) Spectrol Velocity (cm/sec), Figure 4-25 Comparison of experts aedian hazard curves with aggregate hazard for 5% daaped spectral velocity. - _ E at 0.15. 0.8 and 2 seconds. X
_~.
.=
~
F 10-'_ . , , i. i g
- T = 0.15 sec : : T = 0.80 sec : : T = 2.00 sec :
_ Aggregate over 14 emperL' _ _ Aggregote over 14 experts . _ Aggregote over 14 experts' .
---- Composite Hozord Model ,
Composite Hozord Model ,
---- Composite Mczord Modet ,
s, 10-2 ,
\ _ . _
\ ,
~: _s :
_* s s ,
\ _ _ l 3,
_ i s _\ u ' s , c $ ', s s s-y _ s g ', ', ', N, tt 10-3 -
\ ,s - -
's
's
- : : , 's
_o s : s : - _ -: o s s s s s N _
, s Ns -
c s, s s ,
\s _ _ s, _ _
g , _
, s
_ s . _ s _ .- , s -
's _ _ ', 's '
' 's , #44 ..
N i s
\s .
'34, s
's #
3,4,_ g % s g s % s s % 10-4 - s ' - - ' - - ' - i 's #s~j, e i v?Q' i i \ i s s s_ f \ \, s _ _ s s_ s
\ _
g . s
'v s . - _
s -_
\ _ _ \ _
'd e, 'f $, 4 _ _
's , _ _
's,, '
^
'4 N .
s , s . s s
-$ l e l n e% a ! e ! t ! \1 t l s e ! ' A 5 15 25 35 45 10 30 50 70 90 10 :30 50. 70 90 Spectro! Velocity (cm/sec) Spectrol Velocity (cm/sec) Spectrol Velocity (cm/sec) 4 Figure 4-26. 0 Comparison of aggregation procedures for total hazard from subduction -zone sources for 5% damped a spectral velocity at 0.15. 0.8, and 2 seconds. -
x
GCOM AT A D< REFERENCES Abe, K. ,1975, Reliable estimation of the seismic moment of large earth-quakes: Journal Physics of the Earth, v. 23, pp. 386-390. Adams,1984 Active deformation of the Pacific Northwest continental margin: Tectonics, v. 3, no. 4, pp. 449-472. Anderson, J.G., and Luco, J.E., 1983, Consequences of dip rate constraints on earthquake recurrence relationships: Bulletin of the Seismological Society of f.merica, v. 73, pp. 471-496. Campbell, K.W. ,1987, Predicting strong ground motion in Utah, an evaluation of urban and regional earthquake hazards and risk in Utah: U.S. Geological Survey Professional Paper (in press). Coppersmith, K.J. , and Youngs, R.R. ,1986, Capturing uncertainty in probabi-listic seismic hazard assessmentw within intraplate tectonic environ-ments: Proceeding, Third U.S. National Conference on earthquake Engineering, v. 1, pp. 301-312. Crosson, R.S., and owens, T.J., 1987, Slab geometry of the Cascadia subduc-tion zone beneath Washington from earthquake hypocenters and teleseis-mic converted waves: Geophysical Research Letters, v. 14, pp. 824-827 Electric Power Research Institute (EERI), 1986, Seismic hazard methodology for the central and eastern United States, Volume I: Methodology: EERI Document NP-4726, vol. I. Geomatrix Consultants,1987, Empirical ground motions investigations for Pacific Gas and Electric Company, Diablo Canyon Power Plant LTSP: Report for Pacific Gas and Electric Company, in preparation. Heaton, T.H. , and Snavely, P.D. ,1985. Possible tsunami along the north-western coast of the United States inferred from Indian traditions: Bulletin of the Seismological Society of America, v. 75, pp. 1455-1460. Joyner, W.B.,1984 A scaling law for the spectra of large earthquakes: Bulletin of the Seismological Society of America, v. 74, pp.1455-1460. Joyner, W.B. , and Pumal, T.E. ,1985, Predictive mapping of ground motion, in evaluating earthquake hazards in the Los Angeles region: U.S. Geological Survey Professional Paper 1360. Kanamori, H., 1977, The energy release in great earthquakes: Journal of Geophysical Research, v. 82, pp. 2981-2987 Kulkarni, R.B. , Youngs, R.R. , and Coppersmith, K.J. ,1984 Assessment of confidence intervals for results of seismic hazard analysis: Proceedings of the Eighth World Conference on Earthquake Engineering,
- v. 1, pp. 263-270.
e - ._ 1 GEOMATRIX R-2 l Lawrence Livermore National Laboratory (LLNL), 1985 Seismic hazard charac-terization of the eastern United States: Vols. I-III, April. Nishimura, C. , Wilson, D.S. , and Hey, R.N. ,1984 Pole of rotation analysis of analysis of present-day Juan de Fuca plate motion: Journal of Geophysical Research, v. 89, pp. 10,283-10,290. Peterson, E.T. , and Seno, T. ,1984 Factors affecting seismic moment release rates in subduction zones: Journal of Geophysical Research, v. 89, pp. 10,233-10,248. Power, M.S., Coppersmith, K.J., Youngs. R.R., Schwartz, D.P., and Swan, F.H., III, 1981, Seismic exposure analysis for the WNP-2 and WNP-1/4 Site: Appendix 2 5K to Amendment No. 18 Final Safety Analysis Report WNP-2, for Washington Public Power Supply System. Richland, Washington, September. Riddihough, R.R. ,1984, Recent movements of the Juan de Puca plate system: Journal of Geophysical Research, v. 89, pp. 6980-6994. Ruff, L. , and Kanamori, H. ,1980, Seismicity and the subduction processes: Phys. Earth Planet. Int. , v. 23, pp. 240-252. Sadigh, K. , Egan, J. A. , and Youngs, R.R. ,1986, Specification of ground motion for seismic design of long period structures: Earthquake Notes,
- v. 57, no. 1, p. 13 S-Cubed,1988, around motion simulations for thrust earthquakes beneath western Washington: Report prepared for Washington Public Power Supply System. July.
Verplanck, E.P. , and Duncan, R. A. ,1987, Temporal variations in plate convergence and eruption rates in the western Cascades, Oregon: Tectonics, v. 6, pp. 197-209 Wesnousky, S. , Scholz, C.H. , Shinazaki, K. , and Matsuda. T. , 1983. Earth-quake frequency distribution and mechanics of faulting: Journal of Geophysical Research, v. 88, pp. 9331-9340. Wyss, M. ,1979. Estimating maximum expectable magnitude of earthquakes from fault dimensions: Geology, v. 7, no. 7, pp. 335-340. Youngs, R.R. , and Coppersmith, K.J. ,1985a, Implications of fault slip rates and earthquake recurrence models to probabilistic seismic hazard estimates: Bulletin of the Seismological Society of America, v. 75, pp. 939-964. l l
W-- - - - - - - - _ _ _ _ _ _ _ _ _ _ _ _ i GEOMATAm f R-3 k Youngs, R.R. , and Coppersmith, K.J. ,1985b, Development' of a fault-specific ' earthquake recurrence model (abs.): Earthquake Notes Seismological l Society of America, v. 55, p. 16. h Youngs, R.R., Coppersmith, K.J., Power, M.S., and Swan, F.H., III, 1985. Seismic hazard assessment of the Hanford region, eastern Washington State: i_n Proceedings of the DOE Natural Phenomena Hazards Mitigation Conference, Las Vegas, Nevada, October 7-11, p.169-176. 1 l l l l t-
1 GEOMATRIX EXPIANATION TO ACCOMPANY TABLE 3-1 Table 3-1 summarizes the responses given by the fourteen experts, which are , further detailed in Appendix A. A more complete discussion of the components of the seismic hazard model is given in Section 2.1. Each of the columns in Table 3-1 is explained below. Note that blank columns or apparent omissions in the table are the result of the expert declining to characterize these aspects. Oceanic Slab Geometry Each of the experts developed a cross-sectional sketch of the geometry of the oceanic slab beneath western Washington. These sketches are included in Appendix A and described verbally in Table 3-1. Alternative models are given along with the relative weight assigned to each, expressed as probabilities summing to unity. Potential Seismic Sources The subduction-related potential sources of earthquakes are identified and each is assigned a letter, which is shown in brackets (e.g., "[a]"). These letters are used in subsequent columns to specify which seismic source is being described. Probability of Activity Probabilities of activity are given for each potential seismic source, spec-ified by a letter in brackets. Where expressed by the experts, ranges of estimates are given in parentheses. "Activity" is used here to signify capable of generating tectonically significant earthquakes (see Section 2.1) . Maximum Magnitude Direct assessments of the maximum earthquake magnitude are given for the sources specified in brackets. In some cases, a range of values is given, or a best estimate and uncertainty bounds, or discrete values with relative weights assigned to each value. Where the word "Dimensions" appears, the expert indicated that the rupture dimensions that he specified be used to
?
i' f' oso MAvsam f EXPLANATION TO ACCOMPANY TABLE 3-1 (cont'd) I calculat a magnitude (i.e., he did not provide a maximum magnitude estisate i H i L directly). See Section 2.1 regarding "location of rupture" to see how the i f i rupture dimensions were estimated.
. Convergence Rate The relative rate of convergence measured parallel to the convergence direc-tion between the North American and Juan de Puca plates is given in milli-l-
meters per year. In some cases, ranges are given or discrete values are : i given with associated relative weights. : Recurrence Method
. The manner in which the experts desired to have the earthquake recurrence rate specified is given in this column. Examples include recurrence based ;
on the historical seismicity record, geologic data for recurrence intervals, or seismic moment rate. The seismic moment rate approach (described in Section 2.1) utilizes the estimates of convergence rate and seismic coupling. Seismic Cctipling (a) Seismic coupling is the percentage of the total convergence rate that is expressed seismically. Therefore, if the coupling is very high (a = 1.0), then all of the ccnvergence rate will be expressed as earthquakes (i.e., the seismic moment rate from seismicity will be equal to that based on conver-gence rate). An a = 0 neans that convergence is occurring aseismically (i.e., thero is no seismic coupling). Recurrence Model The recurrence distribution function is specified in this column. Models requested by the experts include an exponential magnitude distribution (f.e., log N = a-bM); a characteristic magnitude distribution (Youngs and Coppersmith, 1985); and a maximum moment model (Wesnousky, 1983), t Geologic Recurrence for Large Earthquakes For. those cases where geologie data provide a basis for estimating recur-rence, an estimate of recurrence intervals for large earthquakes is given. L These recurrence intervals were generally judged appropriate for magnitudes at or near the maximum.
GEOMATRIX Appendices FINAL REPORT Seismic Hazards Assessment for WNP-3, Satsop, Washington Contract No. C-20453 Submitted to WASHINGTON PUBLIC POWER SUPPLY SYSTEM 3000 George Washington Way Richland, Washington 99352-0968 Geomatrix Consultants
GEOMATAIX Appendices FINAL REPORT ' Seismic Hazards Assessment for ; WNP-3, Satsop, Washington Contract No. C-20453 i 1 Submitted to { WASHINGTON PUBLIC POWER SUPPLY SYSTEM 3000 George Washington Way Richland, Washington 99352-0968 l i i Geomatrix Consultants l
7 i GEOMATAIX APPENDIX A DOCLMENTATION OF EXPERT INTERVIEWS g This appendix provides documentation of the expert interviews, summarized in Section 3 of the main report. As discussed in detail in Section 2.2.2, the expert interviews occurred in two parts. Phase I consisted of in-person interviews held at the experts' offices during the summer of 1986. The Phase II follow-up interviews were given by telephone in the fall of 1987 In the first case, an Information Package was cent to each expert to explain the objectives of the study and to describe the format. Prior to the follow- ! up interviews, materials were sent to each expert that summarized their pre-vious assessments and summarized the assessments made by all of the experts. l The summaries of their previous assessments included the calculated results l (i.e., maximum magnitude and recurrence relationships) derived from their , assessments. The documentation for the expert interviews given here include: l I e Information Package sent to the experts prior to the Phase I interviews. e Phase I responses of individual experts. e Informational materials sent to the experts prior to the Phase II follow-up interviews. This material includes: Example letter Attachment 1 - Summary of assessments for each expert including calculated results Attachment 2 - Phase I responses of individual expert
. (presented previously)
Attachment 3 - Summary of aggregate expert assessments Attachment 4 - Recent references , Attachment 5 - Updated seismicity cross-sections , o Phase II responses of individual experts. The interview summaries included here are based on written notes taken by members of the elicitation team and are focused on interpretations, uncer-tainties in each interpretation, and the basis for the responses given. We are not attempting in these summaries to provide a full "defense" of the expert opinions given because most responses are based largely on judgement. We are, however, trying to provide a third party with enough inroamation to understand the key data and interpretations that are driving the experts' opinions. Also, documentation is required because the experts relied,to ! some extent on new unpublished data or work in progress. Note that each expert reviewed his summary for accuracy and the accepted version is given here.
. w ; .-
weg-,z -~,y w m- , ~-, : ----4, -
,--~r-r- --s-__,~-~~-,mwo-y INFORMATION PACKAGE-
?> , (Sent Prior to Phase I Interview)
- k. .._ e T
- )
o <- a e 4 e C 1 5
'k 1 -
?
I s A . a 1 ]' a-7
, s 4
f J-s II i 4
-l t ?
l 4 t , J. 1 i. 1 e '.f 'P P 4 k-+,&,1 4%e&.m e ,
7 1
& . s :-
DEOMATRfX k' INFORMATION PACKAGE TO SEND TO SATSOP EXPERTS
' Introduction
.This information package'is-intended to provide you with the backgound and
' purpose-of-your involvement in the Satsop Seismic' Hazard Analysis.
-The problem that we are' addressing is: What is the probability of the
. occurrence of ground motions that woald exceed the seismic design basis at the Sacsop site? Answering this question requires a probabilistic seismic hazard analysis. We are incorporating the uncertainties that exist rcgarding the seismic sources that might affect-the Satsop site in the cotree models using probabilistic techniques. We cannot expect to solve t*,se questions regarding earthquake potential in the Pacific Northwest.
~
Therefore, we are making a ' snap-shot
- of the present stace of knowledge u11ng the opinions of experts onarding the nature and seismogenic potential cf the Cascadia subduction zone.
The object of this study is to solicit your opinion on the nature of the
~ Ca:cadia subduction zone thus providing a basis for constructing a source model for the probabilistic seismic hazard analysis. Thi_ is a probabilis-tic study, so ranges of values or weighted values may be given to particu-Icr parameters. It is likely that any given expert is not intimately fcciliar with all aspects of the problem at hand (e.g., one may have knowledge of geophysical constraints on the slab geometry but not of the details of the instrumental seismicity data set). Therefore, we will cllow each expert to qualify the respoases given relative to his perception of his expertise in the subject area.
The focus of the scientific inquiry is the Cascadia subduction zone--its features, behavior, and seismogenic potential. Inasmuch as analogies' to cther convergent zones shed light on this margin, the experto may wish to davelop such analogies. Please bear in mind, however, that the engineering
. result of this analysis is a probabilistic ground motion estimate at the Satsop site.
r 1 osemm .
;q
-i Basic Elements of a Probabilistic' Seismic Hazard Model
, The seismic hazard at a eite depends on:' the location of potential future.
certNuakes relative to the site; the rate of occurrence of future earth-quakes 'of various sizes; and the attenuation of ground motions with i l. !~ - distance. A probabilistic model of the seismic hazard requires 1) charac- ; L , i: torization ~of potential earthquake sources in terms of their location and l
'giometry relative to the site; 2)' the rate of ' seismic activity on each source Land the relative frequency of various size events; and 3) charac-tarization of the amplitude of ground motions as a function of ,
I' sturce-to-site distance and earthquake size. The latter of these (ground } notion-attenuation) is not the subject of our concern here. 'le are focused here on the source model (1 and 2 above), which defines the location and r occurrence of seicnicity. Further, we are concerned with j subduction-related seismic sources only; shallow crustal faults will be 'l modeled separately. l
- Seismic ' source modeling techniques for hazard analysis have become increas- -
ingly sophisticated in recent years. For example, sources ars usually modeled as three-dimensional surfaces, rupture size is constrained by mag- , nitude, the focal depth distribution (rupture nucleation locations) can be specified,-' fault segments can be modeled, and a variety of recurrence models can be incorporated including renewal or real-time models. In other words, the ' hazard analysis is capable of ef fectively modeling virtually any type c3 eartMuske behavior that is believed to be appropriate, and that ecn be characterized. To account for the uncertainties in' the source l models, simple probabilistic techniques such as logic trees have been d2veloped that . allow for a range of parameter values for any particular characteristic. Each value can be subjectively weighted as to its credibility or likelihood of being the correct value. Simply put, probabilistic approaches do not require that you make a "yes/no" decision;
-'only that you express your expert opinion and the uncertainties associated with it. The probabilistic methods that we will utilize as part of this . L project are discussed ja-more detail below.
i I I
n D g ' F4r this hazard analysis, we are attempting to model any potential seismic ' scurces that are believed by the experts to be associated with the Cascadia cubduction zone (e.g., plate interface, intra-slab, accretionary wedge, etc.). Any or all of these features may have some probability of being saismogenic. If. there is some finite probability of activity (however cmall), then the further characterization of the source (its geometry, ote.) can be carried out conditional on the source being seismogenic. Of course, if an element of the subduction zone has no probability of being saismogenic, further characterization is not required. Sitsop Site Hazard Analysis Uncertainties in the seismic hazard of the Cascadia subduction zone stem from the fact that no earthquakes largar than about magnitude 5 have been - unequivocally associated with the plate interface. No clear definition of sich geometry can be easily discerned from seismicity data alone (unlike most other convergent margins). Therefore, several essential hazard source characteristics cannot be di.ectly assessed, such as: Is subduction occurring beneath western Washington? What is the geometry of the inter-f:ce and the slab? If subduction is occurring, what is the degree of ( s2ismic coupling between the plates? Why have there been no observed i i interplate events? How is the maximum earttquake to be evaluated? How is
- carthquake recurrence to be evaluated? There are no clear answers to these questions but various lines of evidence from geologic, seismologic, and gnphysical data can be instrumental in providing constraints that can be i included in :he hazard model. For example, strain rate (usually fault slip j rete) can provide an important constraint on eartinuake recurrence (e.g.,
Anderson,1979; Anderson and Luco,1983; Youngs and Coppersmith,1985). To be useful, one must assess the component of the slip rate that is potential saismic strain energy (i.e., subtract aseismic slip from the total slip). In order to use the relative slip rate (convergence rate) at a subduction zone, it is necessary to estimate tim percent seismic coupling. Several studies ccnparing seismic moment rate (from historical seismicity) with picte convergence rate demonstrate a broad variation in tte percent coupling for worldwide subduction zones. Therefore, if the percent seismic
T GEOMAT5MX i coupling can be ec;;inated for th3 Cascadia subduction zone , in3an ingf ul constraints may be placed on earttquake recurrence. As another exampic, because this subduction zone is not well expressed from seismicity, we must rely on other data (e.g., refraction studies, broadband data, etc.) to estimate the slab geometry. Probabilistic Approaches to Eliciting Expert Opinion Many important decision-making problems involve a considerable degree of uncertainty and serious risks associated with that uncertainty. Therefore, it is important to obtain as much information as possible in order to understand and accurately represent the degree of uncertainty concerning events or variables of interest. Of ten, problems with serious risks are characterized by a lack of directly relevant experimental evidence. The "hard" empirical evidence may be only indirectly relevant (for health risks, consider experiments with new drugs on animals but not on humans) or may be too limited (for seismic risks, consider a limited history of reli-able records of seismic activity). As a result, most of the information available is subjective in nature, involving the judgments of experts who presumably will attempt to take into account any "hard" evidence, direct or indirect, that may be available. This is the situation that we face in the Satsop Seismic Hazard Analysis. To understand snd accurately represent the degree of uncertainty concerning events or variables of interest, we must utilize expert judgments and express them in a form useful for communicating and measuring uncertainty. This has been recognized increasingly in recent years and has led to use of experts' probability assessments as important inputs in decision and risk analysis problems. Examples include probability forecasts of rain and other meteorological events, risk assessments of health ef fects of specific air pollutants, and the recent study of seismic hazards in the Eastern United States by Electric Power Research Institute. People of ten think in terms of how likely certain events are but generally do not actually quantify their judgments in tet,ns of probabilitie s. Yet everyene is exposed to probability statements such as "The probability of
agg33,7,gg rain tomorrow is 30 percent," "There is a 10 percent chance that the patient will not survive," "The odds are 4-to-1 against the horse winning the race," "The probability of discovering oil if we drill in this location 10 5 percent," or "That team has only one chance in one hundred of winning the championship." We are going to ask you to make similar statements cbout the features, behavior, and seismogenic potential of the Cascadia cubduction zone. For example, we will be interested in your probability that the plate interface is seismogenic and your probabilities for different rates of plate convergence. It is important to emphasize that no background in probability theory is nooded, and you will not be asked to perform any fancy manipulation of probabilities. The problem will be broken down logically into small parts c3 that each question is clear, understandable, and easy to think about. We will use a number of methods to help you translate your judrents into probabilities. Your probabilities will be elicited in an interview c:ssion, and the role of the interviewers / analysts is to assist you in thinking about relevant qualitative issues and in representing your know-ledge in quantitative terms. The ultimate intent is to wind up with a set cf pecbabilities that accurately reflect your knowledge and uncertainties. It is also useful to obtain an idea of how confident you feel about these probabilities. When you give a probability of, say, 30 percent, we will c:k whether you are quite certain about that figura or whether it repre-c:nts an estimate but you feel that the probability might be lower or higher (for example, it might be as low as 25 percent or as high as 35 percent; or it might be as low as 15 percent or as high as 40 percent). We rccognize that it is often difficult to come up with just a single number, cnd giving a range in addition to the single number provides useful infor-mation. Finally, it is important for us to understand your reasoning process and the rationale for the probabilities that are given. Of parti-cular interest are any underlying assumptions or theories that you are i considering. Thus, the final outcome of the assessment process should be a set of probabilities, an indication (through ranges of values) of how vag'ue or confident you feel about these probabilities, and the qualitative reasoning behind the probabilities, k
l
, )
-62 f/V% ==
l C3mOMATRfX
. Some people tend to feel-more comfortable with what they view as "hard"
- data,' such as - a set ' of empirical observations , than with experts ' judg-
. ' menta1 ' probabilities which' are viewed by . some as representing "sof ter"
-data. Detailed evaluations, however, have shown that experts' probabili-ties can 'be very stable andz reliable. Weather forecasters' l probability forecasts provide an example in which the subjective probabilities tend to
. do at 'least as well as, and usually better than, probabilistic weather forecasts generated from combinations of statistical and physical models.
If the elicitation process is carefully designed, the stability and relia-bility of the results can be very high. Careful attention to all of an expert's uncertainties is. important, just as a full consideration of lpossible variability related to dif ferent sources of sampling or experimental error is vital-in empirical research. To help you represent your judgments most effectively and accurately, we will briefly discuss some factors that you should keep in mind when going through the probability elicitation task. The following four paragraphs provide some suggestions along these lines. It is Laportant that you consider relevant evidence in a systematic and effective manner. Your probabilities should be based on whatever information is available ebout the Cascadia subduction zone and what you know about subduction zones elsewhere in the world. It is important to try to consider all of the information and all of the- possibilities in terms of features, behavior, cnd seismogenic potential associated with the Cascadia subduction zone. Think about all scenarios that could possibly be consistent with the -infor-mation that is available. Do not just focus on a single, "most likely" ecenario or on a scenario that stands out in your mind for some reason. Think also of extreme scenarios, even if they are less likely . Consider information that might be inconsistent with a specific theory as well as information that might be consistent with the theory. Try to keep an open mind.
- . _ - - . . _ - . - - - - - _ - - a
n L 1
~7 GEOMAmm .j i
I You should also resist any tendency to place greater weight on pieces of information seen first or last. Early information can influence tlm way you think about a problem and the way you interpret and react to later ,
- in f o rmation. On the other hand, the most recent information seen is most -cccessible in your memory and may have undue influence on your judgments fer that reason. Review all of the available information, again trying to ksep an open mind.
Be sure to keep in mind the uncertainties associated with' data and othe r information.' The reliability and accuracy of information vary consider-chly, and you should be careful not to overestimate reliability or accuracy and to ignore uncertainties. 1 Finally, do not confuse values with judgments. We are interested in your scientific appraisal of probabilities concerning possible seismic sources casociated with the Cascadia subduction zone. The costs associated with 031smicity in view of the location of the Satsop site are a separate issue. In giving your probabilities, you should just be considering the likelihood of certain evente, not the potential consequences associated with those ovents. L i' As noted above, your probabilities will be elicit.ed in an interview cossion, which will last about one-half day. 7.te session will begin with a j brief introdoction to uncertainty and probability to familiarize you with i the basic notions of quantifying judgments in terms of probabilities. ; N3xt, we will discuss the Cascadia subduction zone, reviewing available ; ovidence and obtaining some of your qualitative judgments concerning the [ features, behavior, and seismogenic potential of the Cascadia subduction zone. Then we will ask you to quantify your judgments and to provide certain probabilities in a systematic manner. The interviewers /analys ts will assist you in this process of assessing probabilities. As the interview proceeds, you may think of aspects of the problem you had not recalled earlier. At any point, you can reconsider and change earlier probabilities. After tim probability assessment process is coupleted , we I r
GCOM AT A X will review both the qualitative judgments and the quantitative probabili-tie s . Af ter the interview, we will prepare a summary of the infonnation cbtained and send it to you so that you can see if it accurately reficcts your judgments. At that time, you can ma'..a any modifications that seem cppropriate. As we have mentioned, the ultimate intent is to wind up with a set of probabilities, together with an indication of the degree of confidence in the probabilities and a qualitative discussion of relevant factors, to accurately reflect your knowledge and uncertainties. In documenting our study, interpretations from all the experts will be cggregated for the analysis and particular interpretations will not be cttributed to individual experts. Some Likely Questions The following is a list of some likely questions to give you a feel for 'he type of infonnation that we will ask about during the interview.
- 1. What does the geometry of the plate margin beneath Western Washington look like? We will give you a graph and ask you to sketch possible models for this geometry. Then we would like you to assign probabilities to these inodels.
- 2. Consider the rate of plate convergence nonnal to the North American/ Juan de Fuca plate boundary. What is the probability that this rate of convergence is less than 10 rmn/ year? What is the probability that it is between 10 and 20 m/yr? Between 20 and 30 m/yr? Between 30 and 40 m/yr? Between 40 and 50 m/yr? Greater than 50 m/yr?
- 3. What are the possible seismic sources associated with subduction?
- 4. For the Cascadia subduction zone, what is the probability that each possible seismic source is seimogenic (active)?
- 5. If the plate interface has some probability of being seismogenic, what are the updip and downdip constraints (minimum and maximum depths) on the seismically coupled part of the in te r f ace ? We will ask for values and probabilities regarding the updip and downdip constraints.
- 6. What is the probability that the plate boundary is laterally seg-mented? If it is segmented, whete do you think the segment '
boundaries are? What i s t he probability that fault rupture will start /stop at these segment boundarie s?
9 GmOMATRIX , l
- 7. What is the longest rupture that may occur along the plate interface? We will ask about the probabilities of ruptures of various lengths . . >
- 8. If the plate interface ha's soine chance of being a-capable seismic -
source, what is your best estimate of the percent - seismic coupling between the plates? (Seismic coupling is defined here as the percent of. the convergence rate that is released as seismic energy; coupling may be a function of the historical observations or of an assumed model). What is the probability of less than 10 percent seismic coupling? What is the probability of between 10 and 30 percent seismic co.pling? Between 30 and 50 percent? Creater than 50 parcent?
- 9. Do you have direct estimate of earthquake recurrence for the plate interface? If se, express this as a recurrence interval for particular magnitude events or as a recurrence relationship of the form log N = a - be.
- 10. We will ask questions like those in 5, 6,. 7, and 9 with reference to intraplate seismicity, accretionary prism seismicity, and any other possible seismic sources.
These questions.are just intended to give you soise idea of the type of information that is of interest. During the interview session, terms will be defined precisely and clarification will be provided as needed.
e oca+4Aram j BI BLIOG RAPilY The- following references provide a representative sampling of the published dcta and interpretations portinent to this study. 1 Geometry of Cascadia Subduction Zone Atwater, T., 1970, Implications of plate tectonics for the Cenozoic tectonic evolution of western North America, Bull . Geol . Soc. Am. , 81, 3513-3536. Crosson, R.S., 1976, Crustal structure modeling oi earthquake data 2. Velocity structure of the Puget Sound region, Washington, J. Geophys. Res., 81, 3047-3054, 1976. Crosson, R.S., 1986 Where is the subducting slab beneath the Pacific Northwest, (abs .), Earthquake Notes , 1986 SSA Meeting, Charleston, South Carolina. Dickenson, W.R., 1970, Relations of andesites, granites, and derivative sandstones to arc-trench tectonics, Rev. Geophys., 8, 813-860. Ellis, R.M., et al.,1983. The Vancouver Island Seismic Project: A C0 CRUST onshore-of fshore study of a convergent margin: Canadian Journal of Earth Sciences, v. 20, pp. 719-741. Henderson, M., Crosson, P,S., and T.J. owens, 1985, seismic structures frau inversion of teleseismic wave forms on the Washington continental
- margin. EOS, Trans. Am. Geophy., Un. , 66, 987.
Hendrickson, M. A., 1986 The determination of seismic structure from tele-seismic P waveforms on the Washington Continental Margin: unpub. Masters Thesis, U. of Washington, 74 p. Kanasewich E.R., Clowes, R.M. , Green, A.G., Spencer, C., and C.J. Yorath, 1984, Lithoprobe I: Results of deep crustal vibrosis reflection program on Vancouver Island, EOS, 65, 989 (abs.). Langston, C.A., 1977, Corvallis, Oregon, crustal and upper mantle receiver structure from teleseismic P and S waves, Bull. Seism. Soc. Am. 67(3), 713-724. Lrngs ton, C. A. , and D.F. Blum, 1977, The April 29, 1965 Puget Sound earth 1uake ard the crustal and upper mantle structure of western Washington, Bull. Seism. Soc. Am., 67, 693-711. Lrngs ton , C. A. 1979 Structure under Mt. Rainier. Washington, inferred from teleseismic body waves, J. Geophys. Res. 84( H9), 4949-4762.
- osouArnm Langston, C..A., 1981, Evidence for the subducting lithosphere under southern Vancouver Island and western Oregon from teloseismic P wave conversions , J. Geophys. Re s. , 86 ( BS), 3857-3866.
McKenz ie , D. , and B. Julian , 1971. The Puge t Sound, Washington, earthquake and the mantle structure beneath the northwestern United States, Bull. Geol. Soc. Am., 82, 3519-3524. McHe c han , G . A . , an d G . D . S penc e , 1983, P-wave velocity structure of the earth's crust beneath Vancouver Island: Canadian Journal of Earth Science, v. 20, pp. 742-752. Michaelson, C. A. , 1983, Three-Dimensional velocity structure of the crust and upper mantle of Washington and Oregon, University of Washington Geology Dept., M.S. Thesis. Michaelson, C. A., and C.S. Weaver, 1986, Upper mantle structure from tele-seismic P wave arrivals in Washington and northern Oregon: J. Geophys. Res., v. 91, no. 52, p. 2077-2094. Owens, T.J., Hendrickson, M. , and R.S. Crosson,1986, Constraints on the subduction geometry beneath western Washington using broadband teleseismic P-waveform modeling (abs.), Earthquake Notes, 57, 1986 SSA Meeting, Charleston, South Carolina. Riddihough, R.P.,1979. Gravity and structure of an active margin: British Columbia and Washington: Canadian Journal of Earth Sciences, v. 16, pp. 350-362. Rohay, A.C., 1982 Crust and mantle structure of the North Cascad>3 Range Washington, Univ. of Wash., Seattle, Washington, Ph.D. Thesis. Spence, G.D., Clowes, R.M. , and R.M. Ellis,1985, Seismic structure across the active subduction zone of western Canada: Journal of Geophysical Research, v. 90, pp. 6754-6772. Taber , J.J. , 1983, Crustal Structure and Seiraicity of the Washington Continental Margin Ph.D. Thesis O iversity of Washington, Seattle, WA. Weaver, C.S. , and C.A. Michaelson, 1983, Segmentation of the Juan de Fuca plate and volcanism an the Cascade range: (abs . ) , EOS, v. 64, p. 886. Yora t h, C .J . , G reen . A.G . , Clowe s , R.M. , Brown , A.S . , Brandon , J .T. , Kanasewich, E.R., Hyndman, R.D., and C. Spencer, 1985. Lithoprobe, southe rn Vancouver Island: seismic reflection sees through Wrange11ia to the Juan de Fuca plate: Geology, v. 13, p. 759-762. Zervas C.E. , and R.S. Crosson, 1986, Pn observation and interpretation in Washington: Bull. Seis. Soc. Am., v. 76, no. 2, p. 521-546.
GEOMATRIX Scismicity of the Pacific Northwest Crosson..R.S., 1972, Small cartb l uakes, structure, and tactonics of Puget Sound area, Bull. Seismol. Soc. Amer. , 62, 1133-1171. Crosson, R.S., 1983, Review of seismicity in the Puget Sound region from 1970-1978: a brief summary: in J.C. Yount and R.S. Crosson (eds.) Earthquake Hazards of the Puget Sound Region, U.S. Geological Survey Open File Report 83-15, pp. 6-18. Heaton, T.H., and P.D. Snavely, Jr.,1985, Possible tsunami along the northwestern coast of the United States inferred from Indian traditions: Bulletin of the Seismological Society of America, v. '75, pp. 1455-1460. Rogers, G.C., 1983, Seismotectonics of British Columbia: unpublished Ph.D. dissertation, University of British Columbia, 247 p. Rogers , G .C. , 1985, Seismicity and seismic potential of Canada's western margin (abs.): The 23rd General Assembly of International Association of Seismology and Physics of the Earth's Interior, v.1, p. 53. Taber, J.J. and S.W. Smith,1985, Seismicity and focal mechanisms associated with the subduction of the Juan de Fuca plate beneath the Olumpic Peninsula, Washington: Bulletin of the Seismological Society of America, v. 75, pp~. 237-249. Washington Public Power Supply System (WPPSS),1982, Final safety analysis report - Supply System nuclear project no. 3, volume 3. Weaver, C.S., an? S.W. Smith,1983, R*;1onal tectonic and earthquake hazard l implications of a crustal fault zone in southwestern Washington: l Journal of Geophysical Research, v. 88, pp.10.371-10.384. l l Yellin, T.S., 1982, The Seattle earthquake swarms and Puget Basin focal mechanisms and their tectonic impilcations; unpublished M.Sc. Thesis , University of Washington, 96 p. l Plate Convergence in the Pacific Northwest l l Chase , R.L. , Tif fin , D.L. , and J.W. Hurray, 1975. The western Canadian i continental margin, in Canada's Continental Margins and Offshore , Petroleum Exploration, Canadian Society of Petroleum Geologists in association with the Geological Association of Canada, Calgary, Alberta, Canada, pp. 701-721. Nishimura, C., Wilson, D.S., end R.N. Hey, 1984, Pole of rotation analysis of analysis of present-da, Juan de Fuca plate motion: Journal of Geophysical Research, v. 89, pp. 10.283-10,290.
v Rid dihough , R.P. , 197 7 , A model for recent plate interactions of f Canada's west coast, Can. J. Earth Sci.. .Vol. ' 14, pp. 384-396. Riddihough, R.P. , 1981, Absolute tootions of the Juan de Fuca plate system: resistance to subduction?: (abs . ) , EOS, v. 62, p. 1035. Riddihough, R.P. , 1984, Recent movements of the Juan de Fuca plate system J. Geophys. Res., 89, 6980-6994. General References Regarding Seismic Coupling Adams, J., 1984. Active deformation of the Pacific Northwest continental raargin, Tectonics , 3(4), 449-472. Ando, M., and E.I. Balass,1979, Geodetic evidence of aseismic subduction of the Juan de Fuca plate: Journal of Geophysical Research, v. 84, ppe 3023-3027. Christensen D., and L. Ruf f,1983, Outer rise eartlquakes and seismic coupling: Geophysical Research Letters, v. 10, pp. 697-700. Davis ,' C .A. , Coombs , M. A. , Crosson, R.S. , Kelleher, J . A. . - Tillson, D.D. , and W.A. Keil,'1984, Juan de Fuca/ North American Plate convergence seismic or aseismic subductiont - Report prepared for Washington Public Power Supply System, Nuclear Project No. 3, August, 1984',' 40 p. Heaton, T.H., and H. Kanamori,1984, Seismic potential associated with subduction in the northwestern United States: Bulletin of the Seismological Society of America, v. 79, pp. 933-941. Hyndman, R.D., and D.H. Weichert, '1983, Seismicity acd Rates of Relative Motion on the Plate Boundaries of Western North America, Geophys. J. R. Astr. Soc., Vol. 72, pp. 59-82 Kanamori H., 1977 Seismic and aseismic slip along subduction zones and their tectonic implications: in M. Talwani and W.C. Pitman III (eds.), Island Arcs. Deep Sea Trenches and Back-Are Basins, Maurice Ewing Series I, pp. 173-174 AGU, Washington, D.C. LeFevre. L.V., and K.C. McNally,1985, Stress distribution and subduction of aseismic ridges in the raiddle America subduction zone: Journal of Geophysical Research, v. 90, pp. 4495-4510. McCann , W.R. , and C.R. Syke s , 1984, Subduction of aseismic ridges beneath the Caribbean plate: implications for the tectonics and seismic potential of the northeastern Caribbean; J. Geophys. Res., v. 89, -
- p. 4493-4519.
Peterson, E.T., and T. Seno. 1984, Factors affecting seismic moment release rates in subduction zones: Journal of Geophysical Research, v. 89, pp. 10,233-10,248.
GEOMATmX Reilinger, R., and J< Adams, 1982. Geodetic evidence for active landward tilting of the Oregon and Washington coastal ranges: Geophys. Res. Letters , v. 9, no. 4, p. 401-403. Ruf f, L. , and H. Kanamori, 1980, Seismicity and the subduction processes: Phys. Earth Planet. Int., v. 23, pp. 240-252. Savage , J .C. , Lisowsk i, M. , and W.H. Prescott, 1981, Geodetic strain measurements in Washington: Journal of Geophysical Research, v.
- 86. pp. 4929-4940.
Stein , S. , Wiens , D. A. , Engeln , J.F. , and K. Fiyita, 1986, Comment on "Subduction of aseismic ridges beneath the Caribbean plate: implications for the tectonic and seismic potential of the northeastern Caribbean: J . Geophys. Re s. , v. 91, no. B1, p. 784-786. Woodward-Clyde Consultants (WCC) 1984, Juan de Fuca Plate Comparison, report prepared for Washington Public Power Supply System WPPSS, Richland, WA. General References Regarding Subduction Pertinent to the Cascadia Zone Cande. S.C., and R.B. Leslie,1986, Late Cenozoic tectonics of the southern Chile trench: J. Geophys. Res.3 v. 91, no. B1, p. 471-496. Kanamori, H., 1981, The nature of seismicity patterns before major earttquake s: in D.W. Simpson and P.G. Richards (eds.), Earthquake Prediction, an International Review, Maurice Ewing Series IV, pp. 1-19 AGU, Washington, D.C. Ke lle he r , J . , and W. McCann , 1976, Buoyant zones, great earttquakes, and unstable boundaries of subduction: Journal of Geophysical Research, v. 81, pp. 4885-4896. Kelle he r , J . , and J . Savino , 1975. Distribution of the seismicity before large strike slip and thrust-type earttquakes: Journal of Geophysical Research, v. 80, pp. 260-271. Lay, T., Kananori, H., and L. Ruf f.1982 The asperity model and the nature of large subduction zone earttquakes: Earttquake Prediction Research,
- v. 1, pp. 3-71.
Mogi, K., 1979, Two kinds of seismic gaps: Pure Applied Geophysics,
- v. 117, pp. 1172-1186.
Sacks, I.S., 1983, The subduction of young lithosphere: Journal of Geophysical Research, v. 14, pp. 3355-3366. Scholl, D.W., 1974, Sedimentary sequences in the north Pacific trenches: in C. Turk and L. Drake (eds.), The Geology of Continental Margins, Springe r-Ve rlag , New York . pp. 4 9 3-50's .
=
GEOMATRIX L Seno, T., Shimazaka , K. , Some rv ille , P. , . Sudo,' K. , and . T. Eguchi, ; 1980, Rupture process of the Miyagi-oki, Japan, Earthquake of June 12,
.1978. Physics of the Earth and Planetary Interiors, v. 23, pp. 39-61.
- Uyeda, . S. , and H. : Kanamori, 1979, Back-arc opening and - the mode of subduc tion: Journal of Geophysical Research, v. 84, pp. 1049-1061.
Von Huene , R. , 1974, Modern trench sediments: in C. Burk and L. Drake (eds.), The Geology of Continental Margins, Springer-Verir.g, New York, pp. 493-504. Yokoyura, T., 1981, On subduction dip angles: Tectonophysics, v. 77, pp. 63-77. Ceneral References Regarding Earthquake Recurrence and Maximum Earthquake Assessment
,'.nderson, J.E., 1979, Estimating the seismicity from geological structure:
Bull. Seis. Soc. Am., v. 69, p. 163-185. Anderson, J.G., and J.E. Luco, 1983, Consequences of slip rate constraints on earthquake recurrence relations: Bulletin of the Seismological Society of America, v. 73, pp. 471-496. Schwartz, D.P., and K.J. Coppersmith,1984, Fault behavior and characteristic earthquakes: examples from the Wasatch and San Andreas faults: Journal of Geophysical Research, v. 89, pp. 5681-5698. Schwartz, D.P. Coppersmith K.J., and F.H. Swan, III, 1984, Methods for assessing maximum earthquakes: Proceedings of the Eighth World Conference on Earthquake Engineering, San Francisco, California, Vol. 1, pp. 279-285.- Youngs, R.R., and K.J. Coppersmith,1985, Inplications of fault slip rates and earthquake recurrence models to probabilistic seismic hazard estimates : Bulletin of the Seismological Society of America, v. 75, pp. 939-964. Geology Related to Subduction in the Pacific Northwest Conna rd , G . , Couch, R.W. , Roy , J . , and S. Kulm, 1983, Heat flow: Atlas of the Ocean Margin Drilling Program, Western Washington-Oregon Continental Margin, and Adjacent Ocean Floor, Region V, Joint Oceanographic Institutions, Inc. , Marine Science International, Woods Hole , HA, 1 map sheet plus text. Kulm, L.D., 1983, Western Washington / Oregon Juan de Fuca Project: unpub-lished dra f t report for Washington Public Power Supply System, 35 p. Kulm, L.D., and R.W. Embly, 1983, Contrasting tectonic-sedimentologic styles along the convergent Juan de Fuca plate boundary: EOS, . 64,
- p. 828.
' f -
~
E a ; d pw enex I :
?
E !!- e 0 a fr .i y: !!
's g. t*~$
B* ? 18:a 1 5
^~
- " i' u m
n, a v 87
=f i=
ti-
= =i :
S- &Ea o a
$ g. e E
- j +f b '
- i& '
s g
. ,9r# s.
AB e a o
- j. g6& klj - - }. iIv.II..I '
},
e e ig 8JI 7 e a =t id fE: 3* c 6 *3 2 e a ; d ,[ o' .. gi-; e J a - . . f ; ~ a -
? s~
u ig a a y 5 22 71~j 0 u *-- 9 : g i<! j %
** j= 5 e $
1
# 5 s
e... ,. -~. s f \
' I a
5 i . J *iI
- 4.
- t L.. L. . L .
l s ( t . ,
...~..ew..,
,1 /.t t ,
n . t . . l_* s A.
. . .,,sp.s
- /
- ~t p.:
- f
;< .t.
* . e,s i
!q .c/. m '..
. t Ms,-
-~
s
, . I.. .
- .d
} .. .
t l p:'7 - 4 , -xt
. n .Lo* .
*1 . ...
.h 1
- t. N .. .
/ p .: " .b . . .* *.
; 1.1 1\ -
~ * * y. , , .%
t . . i(e . $ y. s
- 11* .
. .1,N*.
..s . ,.
- j. n' .
, ...\ .n. 1 t
j g .; ys . . . . .
'}. *
;I*.,*,.kN I
- 6,.- * -
# . .,.. ~;... .
. .s.
l' ,. 'f '/ /y' * *, Q = / s.:--. r'y' , * . r
.L *@ '
=c
. .;r,w,:
nQx -
;r..*,**y :Nd (:: ,
f,: . 'I! ,
., z.. e g
' *'r. , . g .. rf...
.. ~ . . ~ .4, '
qz.ko .c ry . . -4 .u . s.
.q- .%~ ' .,
.c...
G. % . ** . .
&sy *':.g*..
- !..'. . ~ t ;
+n?%..
. , : . ;. ~ :I
' pk[y.
a .: .'1 . .(t' .' I, * . I
~ . - !
q> \ t
! 6 1 .
t . /. .
%. % * .., , , f * :
- -w
.v..* . .::g:.:{ ::':::-
r y I 1 1
.k
, s. 1 g
?
e f
- \ ~~ ,
s
-s ,
j i ~.? .: I , Q l w
.t
.] F y . 5 - . W , .' ] #
.. , 'r
,,f i)
~.
GEOMATFalX E W 12G* 12 5* 124* 123* 122* 121* 120* 11 9 ' 127* i I I I g i I I ,f
'o O IOOkm -
- n. 8 u * [
n HORIZONTAL SCALE .
- $ gjg [,
x10 VERTICAL EXAG. $ [a I gy .* 5m a ma a - - a -am m za
=
OLYMPICS 3 2 }4 CA$ CADES
$ 5 g 2 km -
-2 km 8 g l.
_l NGET BASIN a _o _ _ _ _ _ _ _ _ _ _ _ _ CASCA01A 0 -2 km - _-2 km I f I i i t 1 1 127* 12 6* 125* 124* , . I
- 121* 120' 119 ' '
8 i
- " I c s
N O,
/ .
Tak,e f*0 5
-50 km -
O - g (er 8 194$
-loo km O 100 km
-150 km N7 VERTICAL EXAG.
I I l l l 1 I # ,
-M .* -
-]._
m 7 PHASE I RESPONSES OF INDIVIDUAL ~EXP N y . 9 e
PilASE I' GEOMAT5HX RESPONSES BY EXPERT (11 Geometry
-Crrdibilities associated with geometries sketched on cross-section prcvided:
10 dip with doubic bend 0.20 15 dip. 0.50 25 0.30 Tha basis for these are: e The refraction data of Taber are good near the hinge area; his 10 dip east of the hinge is constrained by only two stations and a dip of 15 appears to be reasonable using his data. e The deep seismicity does not define a dip of more than a couple degrees; the T-axes of focal mechanisms are not as systematically oriented as suggested by Taber and Smith. e The 25 dip is consistent with depth of high velocity layer of Langston from the Longmire station and the proper depths for magma generation. e The Pn data suggest that the 6 - 7.8 km/see transition has essentially no dip and the 8.1 km/sec velocity is not see, although it is well-determined to the west offshore. e The velocity inversion of broad band data suggest dips of greater than 10 nnd most likely about 15 . e- Recent wirk by Canadian investigators suggest 15 dip. Convergence Rate
# 30 mm/yr normal convergence (110 mm/yr) based on analyses by Riddihough and Nishimura et al. These studies show that the rate has been decreasing over
< the past several years. Seismic Sources and Activity l Potential seismic sources are: l Intra-slab Interface "deep events" above Juan de Fuca plate for 15 and 25 dip models (remnant plate?)
Probability that the intra-slab. source is seismogenic: Given a dip of 10 : 1.0 based on occurrence of 1965, 1949' events Given a dip of 15 - 25 : 1.0 for deep zone (which may be remnant
. slab); 0.10 - 0.15 for deep slab-because of lack of current seismicity down to
! magnitude 2 Probability that the interface is seismogenic: 40% (ranging from 25% to less than 50%) based on: e ' Complete absence of thrust earthquakes that would be associated with stress buildup on the interface e The unusual nature of the margin relative to other margins globally e Adams turbidite data suggest possibility of large earthquakes, as perhaps will Brian Atwater subsidence data e Jim Savage most recent strain data does not see strain accumulation across Puget Basin, but may be shear strain accumulation across Strait of Juan de Fuca
. Location of Rupture Intra-slab source:
10 dip model: Eastern limit at about 122 because of age, depth, and temperature of plate Nestern limit for most of seismicity at about 124 based on observed drop-off of seismicity and the effect of the accretionary wedge although could have mag 4 to 5 events all the way to the ridge 95% of the seismicity would be expected between 122 and 124 15 - 25 models: Expect the seismicity to be above about 50 km but have little basis for estimating The remnant slab source would not expect it to be restricted in lateral extent to the Puget Sound region For all models, the observed seismicity provides a reasonable basis for the relative frequency of earthquake occurrence along strike
-3 GEOMMM Interface source:
' Downdip. extent should be at about 40 to 50 km depth based on pressure and temperature Updip extent to within about 50 km of the slab hinge point (toe of the continental slope at 125 ); above this would be in weaker materials of the accretionary wedge
-Along-strike representation of the interface is poorly constrained, e Michaelson and Weaver inversion is subject to considerable uncertainties, no alternative models were tested e 20% likelihood that the M & W boundary segments the interface e The ends i.' the Juan de Fuca plate (Nootka fault zone and Blanco fracture Lone) should be segment boundaries.
Maximum Earthquake Magnitude Intra-slab source-(10 dip) or remnant slab source (15 - 25 dips): 7.25 (10.25 with 7.25 slightly preferred) based on largest historical events and constraints from thickness of brittle slab Intra-slab source (15 - 20 dips) 5 - 6 based on observed events offshore Interface: Maximum dimensions (given above) provide reasonable maximum magnitude constraint Ruff and Kanamori relationship is not very applicable to the Juan de l Fuca plate because off edge of distribution l Seismic Coupling and Earthquake Recurrence Uncertainty is seismic coupling represented by a bimodal distribution; which says that the interface is either nearly entirely aseismic or is nearly completely locked. The probability mass near a = 0 and a = 1 ranges from 0.5 - 0.5 to 0.66 - 0.33, respectively. l The basis for this assessment is the following:
- If the coupling were a = 0.5, one would expect to see small to moderate magnitude thrust carthquakes in the region surrounding the imminent rupture; no such events have been observed.
l l
m GEOMATAW e There appears to be no known analog subduction zone that is completely locked;'this would imply a maximum moment-recurrence model (i.e., essentially no other events besides the largest magnitudes) e The strain data will be very important to assessing whether a very low a (aseismic slip) is occurring; at present, uniform strain accumulation does not appear to be occurring. Paleoseismic indicators (e.g., turbedites, Atwater subsidence data) suggest
. longer recurrence intervals (500 - 1,000 yr); due to present uncertainties, these_ data should only be used as a basis for comparison.
For intraplate and remnant slab sources, use historical seismicity for recurrence estimation.
GEOMATAIX E W 12 5' 124* 123* 122* 121* 12 0* 119 ' 127* 126' l I I I g i i I ,I I O lOOkm - E -
- a z a c5 5 i
HORIZON T AL SCALE J - $ m alO VERTICAL EXAG. $ ', 9 gQ .y [g
- 5 maa a 2 z;
- a .G22 3 3 y O N ICS l CASCADES
-2 km 8 $ 2 km -
C ASC A01 A B A -2 I I I I f I 1 1 A' A* 127 I 126' I - 125' 124' I" 8
- _ IM* 121' I w 120*
I 119' p O. Tabor i*& 3 __ o
- 50 km W L3 6- 10 ' -
\ 19'f 7 o
/ ,g
- 800 km .
O 100 km
-150 km 25 -
SCALE NO VERTICAL EXAG. I I I I I f I I EXPERT #1 Three configurations: shallow 10 with double bend 15 25
PHASE I GEOMATAIX > RESPONSES FROM EXPERT //2 ! Geometry Bamis-for geometry sketched on cross-section provided: e Refraction results by McClain beneath the continental slope show a 5 - 10 dip to the slab e Seismicity data constrain the slab location beneath Puget Sound e A steepening beneath Puget Sound (as suggested by Michaelson and Weaver, 1986) gets the slab to proper depths for magma generation baneath the Cascade. This also agrees with work by Langston. Note that the boundary sketched is the oceanic crust / mantic boundary (oceanic Moho) with the top of plate located as indicated. Crnvergence Rate A distribution of values is provided: 15 m/yr 10% 20 - 25 mm/yr 80% 30 mm/yr 10% (N:te that these are orthogonal rates of convergence.) Based on estimates made by Duncan and Verplanek, Riddihough, Engebretson, .cnd Jurdy. Seismic Sources and Activity Potential seismic sources are: Intra-slab Interface Probability that the intra-slab source is seismogenic: 60-70% based on occurrence of historical seismicity and w artainty in location of 1949 and 1965 carthquakes -Probability that the interface is seismogenic: 80%-(+5%, -10%) based on: e -McClain's refraction results show 4 - 4.5 cm/sec crust against the oceanic crust.
GEOMATRIX e The-isotherms and heat flow suggest that well-cemented dewatered rocks should be present, e Low grade metamorphism and' reduction would be expected (based on McClain's work). Loc-tion of Rupture Intra-slab source: The upper 10-15 km of oceanic crust is expected to be the brittle (seismogenic) part of the slab. The updip extent should be at about 123 longitude (near the bend in the plate). The downdip extent is uncertain; but 80-90% of the large earthquakes would be expected near the deeper bend in the plate. Interface: The updip extent of rupture lies approximately beneath the coastline based on: e This is western extent of Oligocene-Miocene volcanic rocks. Interface between Eocene volcanics and sedimentary Oligocene-Miocene accretionary wedge. e The subducted sediments have undergone low grade metamorphism and have had the fluids squeezed out, low porosity, and zero permeability. The downdip extent of rupture cannot be assessed with confidence. Along-strike segmentation of the interface is difficult because morphologic evidence would be obscured by the thick sediments and no deep reflection lines have been run parallel to the margin to look for sub-sediment morphologic charges. Between 46 and 47 lat. lies a free-air and Beugher gravity low that may represent a segment s (following M model of Kelleher et al.). Confidence level in this segment is 40% (+10%, -30%). Earthquake Recurrenc3 30% coafidence is given to the Adams turbidito interpretation because: e The cause of the turbiditos could be sediment loading at the slop, storm activity, or scismic triggering l
, , _ = . .
i 4-
-y,y T .+ 'f e Recurrence intervals of hundreds of years seems reasonable for this margin due to sediment loading because the shelf is over 3'i km wide.
e More work is needed to verify that the thickness of the hemipelagic clays is consistent among turbidites to verify that recurrence , intervals are regular. ,
'l
(, 4 I b j i [ i- I i s
~' - - - - ~ - = *
=---w rv e- ,e- #--ci,, , , , ~ , - . - .-n,,-.- .m--w,w-,me--%n .y.~~re- +ww- 7e+e- v'-'v--M=w=.y *--e-4~+----+ r----t""
,t GEOMATAIX E
@ 127* 126*= 125' -'l24* 123* 122* 121* 12 0' l19' l- 1 I g i i I .,8 I
=
O IOO km - g a a z @ u = HORIZONTAL SCALE .J L { cio 5 , alO VER!aCAL EXAG. f {a I ; _8g g
.. . 5m a e' s, =
, za
~"' =
- OLYMPICS a
2 y zN l CASCADES 3 2 km .
-2 km - O 8 I PUGET DASIN 3.________
CASCAOIA 8A
.2km I I i i I I I I i
A' A* 127 126*
.i 125*
12C 3 --
.I 121*
1 U 120* i 119' f ~ ~ , , , O* , TMar li6S 0 10f'03 ' O
~
-50 km %
~
194f
-LOO km , , dm cI Es cA -
4 , s
"" \
O LOO km
' ~
-150 km
[\ SCALE NO VERTICAL EXAG. b o M +--]' q9
\
l l l 1 I l ! ! l I t ! EXPERT #2 F 4 I I
PHASE I ""ANX RESPONSES BY EXPERT $3
'Germetry B231s for the geometry. sketched on cross-section attached:
e Focal mechanisms-of the. deeper seismicity, including the 1965 event, are normal suggesting they lie within the slab. e Langston's analysis of converted phases above the 1965 events shows a low velocity zone that is inferred to be a layer of subducted sediments. e No opinion is given regarding the location of the slab beneath the Cascades. C nvergence Rates 40.mm/yr (+19 mm/yr) based on the analyses of Nishimura et al. This rate cppears to be compatible with rates on major structures to the north and south, cuch as the Queen Charlotte fault and the San Andreas fault. The 40 mm/yr is ba:ed on an average rate over 700,000 yr. No data exists to determine shor'.er-tarm rates. Shortening rates given by Adams (25 mm/yr) are probably not true crustal shortening rates or would expect to see large mountain ranges like the Trtnsverse Ranges. S,ismic Sources and Activity Potential seismic sources are the following: i e Intra-slab source e Plate interface e Faults in accretionary wedge (analogy is,made to seismogenic faults in the wedge such as that giving rise to the 1945 Mikawa earthquake, which l experienced surface rupture. It is assumed that this type of source
- will be modelled by the known crustal faults in the site area and/or by l a random crustal source).
Probability that the potential sources are seismogenic: l Intra-slab: 1.0 based on the historical occurrence of the 1965 and 1949 i events, which are inferred to be intra-slaS events. Interface: 0.6 based on the following: e If the interface were creeping, one would expect to see more smali magnitude (M 4 6) thrust events. , o It would be unusual on a global basis for the margin to be completely quiet seismically, especially along its entire length.
GEOMATRIX e Global. comparisons of plato age, convergence rate, etc., suggest that the Cascadia zone should be seismogenic. e The probability may be as high as-80% if geologic evidence for large events is further (-upported; or it may be less than 50% if other analogous margins can be shown to be aseismic. Loc 7tions of Rupture Intra-slau: (Distribution for likelihood of earthquake occurrence shown on cross-section.) Two most likely locations are the outer rise and where they have occurred historically. Although very few outer rise events have occurred historically, they would be expected by analogy to other margins. In map view, the relative likelihood of intra-slab earthquake occurrence can be modelled either by the pattern of observed seismicity or by an assumption of a uniform distribution. Both of these models are given an equal weight of 50%. Interface: (Distribution for likelihood of earthquake occurrence shown on cross-section.) Basis for this distribution is global analogies to where seismic radiation typically occurs. In map view, the Blanco fracture zone and the Nootka fault zone, which mark the ends of the Juan de Fuca plate probably serve as segment boundaries. No strong evidence exists for segmentation within the plate, although ! it is possible for the interface to rupture along a length that is shorter than the entire 700 km-long margin. Mtximum Earthquake Magnitudes Intra-slab: 7-1/2 in the deeper part of the slab. 8 in the shallow part; based on global analogy (Sumba, Rat Island) Interface: 9 (+1/2) based primarily on maximum rupture dimensions and analogy j to the 1960 Chile earthquake. l Recurrence-Related Parameters Intra-slab: The historical seismicity provides a reasonable basis for estimating recurrence. Interface: Recurrence intervals may range from 200 to 20,000 yrs with a preferred value of about 400 yr based on: e Recurrence f rom of f shore turbidites appears to be about 400 - 500 yr.
va-y .
? -
h5 GEOMATRIX g ,
- Preliminacy evidence for subsidence by Atwater suggests recur-rence intervals longer than 1,000 yr, as do recurrence data in Alaska.
e- Historical data suggest that at least 150 yr and probably-200 yre have elapsed since the last major earthquake. A characteristic earthquake recurrence model is considered to be appropriate to explain the absence of moderate size events. The range of the characteristic magnitude should be 8 - 9. Saismic coupling (see distribution) is given an expected value of 0.66 based on global analogies and Kanamori's age vs. coupling relationship. .The value is lacs than 1.0 because post-seismic creep probably accommodates a considerable
- c. mount cf total convergence. Earthquake recurrence on the interface should be ba2ed primarily on the recurrence interval date (given above) and the seismic coupling estimate used as a secondary check on these estimates.
t e
GEOMATAIX y .
- E 127* 126' 125* 124* 123' 122' 121' 12 0 ' 11 9 '
I i i I I I ,1 I f o O 100 km - L HORIZONTAL SCALE 8 F e-7. tEE
. fag o e slo VERTICAL EXAG. $ 10 I
, eg Q -3
$g G ma s a S zs E d
2 e
.G22 ONICS l CASCADES
- 2 km 0 2 km ~ u l PUGET BASIN _g__._-- , C A S C A D'I A D A 2 km - . 2 km 1 I I f f I l I __ A' A*127 126* 12 5' 124' 12 ' _ 12 2 ' 121' 120* 119 '
'~ . -50 km N' 4 r a( w. m - ~
(rw54 -OQ '7*
-lO0 km Obd -
O IOO km
' ~ -150 krn SCALE NO VERTICAL EXAG.
I I I I I 1 l l EXPERT #3 cross-Section 10-20% of events Locati n Distribution for
/ .
5% of nts Intraslab events i { l .0 - - Distribution for prob that section of interface will
'- rupture in event
< GEOMATRIX PilASE I RESPONSES BY EXPERT //4
' Geometry LBasis for geometry sketched on-cross-section provided:
e Should be 100-125 km beneath volcanics based on global analogs e Shallow dip is consistent with young hot plate and slow convergence e The refraction data by Taber offshore are good [ Note that we have reinterpreted your drawing to reflect the fact that the Taber (1983) line represents the top of oceanic Moho, not top of the ccerenic plate (see attached). ] Convergence Rate A distribution of rates was given: 1 cm/yr 0.05 2 cm/yr 0.50 3 cm/yr 0.40-4 cm/yr 0.05 The basis for this estimate is: o Comparison with other margins e Presence of the volcanic arc e Presenza and number of earthquakes e Global plate reconstructions It is noted that uncertainties in these rates may result from the breaking up of the Juan de Fuca plate. Seismic Sources and Activity Two potential seismic sources are identified: e plate interface e intra-slab Probability that the intra-slab source is seismogenic: 1.0 based on historical occurrence of 1965 and 1949 ovents, which'are inferred to be intraplate events
GEOMATAIX Prsbability that the' interface source is seismogenic: 0.75 (10.25) based on the evidence that the rocks are being stressed near the interface, as witnessed by the small-magnitude intra-slab events (Note that "seismogenic" here is interpreted to be applicable down to magnitudes as small as M3). Location of Rupture Intra-slab source: Seismicity within the sis.b would be expected to occur within the upper 20 km;-90% within 5-15 km. The upper 5 km is probably weaker and unable to support large earthquakes, but could be the location of aftershocks. The weakness is probably due to the intermixing of sediments with basalts in the upper part of the slab. Along strike, a highen concentration of seismicity might be expected where they have occurred historically (low confidence in this assessment). Interface: The downdip extent would be at about 40 km based on e Globally, maximum depths are typically about 60 km, but because this is a young plate and has a slow convergence rate, a shallower depth is expected i e The presence of intra-slab events at depths of 40 km suggests that
- the rocks can support earthquakes at these depths The updip extent lies at about 20 km depth based on
l l- e Analogies to other subduction zones e This lies at about the eastern edge of the accretionary prism Along strike, the boundary with the Explorer Plate (Nootka fault zone) and tha Gorda Plate (Blanco fracture zone) should be barriers to interface rupture with a confidence level of 50% (145%). Maximum Earthquake Magnitude Intra-clab source: Uniform distribution between magnitude 7 1/2.
~
l l 1
.8 :'
aemTmix
-l . Interface:. i 3 30%-
4 30% 5' 30% 6 9% 7 1%~ S?ismic' Coupling and Recurrence
' Intra-slab source:
Use of historical seismicity to define recurrence seems to be reasonable (low confidence in this assessment).
. Interface:
.a = 0.05 (+0.15, -0.05) with 80% weight at 0.05.
-based on global analysis to other subduction zones such as Psrbados and southernmost Chile.
k
v GEOMATAIX W 122' E ns*
~
1278 126' 12 5' 124' 123* 121* 12 0* a i i l I
.O 8
100km 4 2 5l 3 e a z E o = { l HORIZONTAL SCALE .2 E- { c ,2g$ , slo VERTICAL EXAG. $' {" I G
, ;g g g du msa z .; g a 3
-ci2 2 ~
2 i~" =j OLYMPICS 1 zN l CASCADES
- 2 km O 8 2 km -
-l PUGET BASIN
.o---- --- -
CASCA0lA B ASI
--2,hm I I f 8 f 1 1 I f. ' A*127 I
126' 8 125*
-+
12t l-v l
- __ 122* _ _ gate w e a 120*
g A' 189'
~-
Tgh, lig $
@ g h _m_ O
-50 km 'r 4
#N '
op o 5lc e l.S u d( C r.
-LOO *.m
. N O 100 km c -d -m.k, l.ew.). ,f
-150 km _
3 NO VERTICAL EXAG. I I i 1 i i l I EXPERT #4 O
1~.
^
4 J PHA". I amAmx
-FZSPONSES BY EXPERT IS /
W > ~ Ge "me t ry' ~ e Thi following weights are given to the-geometries sketched on the crors-sections provided: A Model A (gradually increasing slab dip): 75 - 80% (+'.1 .15) 20 --25% (+ .1 .15) ~i Model B (shallow dip): biscd on: e To reach magma generation depths, need to be at abotn 100 km beneath Mt. Ranier ~ e The hypocenters for.the deeper seismicity do not define a dip e Taber's refraction data allow slab dips above 11* e Waveform modeling by Owens suggests dips of about 15 - 10' are likely at about 123.5*
- Model A is consistent with Michaelson and Weaver's 45' dips to east and with appropriate magma depths l
C nvergence Rate , 40 (+0, -15) mm/yr based on: e Riddihough analyses of plate reconstructions e The internal deformation accurring in the 'dxplorce and Garda plates suggents that the Juan de Fuca plate is changing character from an actively subducting plate to one that to passive and less rigid. l l e The convergence rate.s lack resolution and are averaged over the past l million years; a d'acrease in convergence rate has been noted. Therefore, the average cate over this period should represent the maximum rate. !. ~ Swicmic Sources and Activity Potential seismic sources are the following: e Intra-slab- , e Plate interface I
t CEOMATAIX ! pr<bability that the sources are seismogenic: Intra-slab: 1.0 based on the historical occurence of the 1949 and 1965 earthquakes'which can be attributed to the. slab Interface: 0.5 + 0.5 based on: e Heaton and Hartzell make a good case for the possiblity of large earthquake occurrence e Convergence is taking place but there have been no historical thrust earthquakes e It is rare to have an aseismic young plate e All of the evidence is circumstantial rather than defini-tive; therefore, a large uncertainty is assigned e Juan de Fuca tending toward behavior similar to Explorer and Gorda plates Location of Rupture _ Intra-slab: In cross-section, the upper 6 - B km of the slab is expected to be seismogenic. Downdip seismicity is not expected east or Mt. Ranier. The updip extent is essentially at the coastline based on the seismicity data and the youthfulness of the slab. In map view, the historical seismicity record can provide a basis for the relative frequency of earthquake occurrence; with the slab corner model of Rogers, a possible explanation for the localization of puget Sound seismicity. Intctface: The downdip extent of the seismogenic interface should be at about 123.5' whe:e the continental Moho lies against the interface. Updip, the interface should extend to the eastern boundary of the accretionary prism essentially at the coastline. Along strike, the boundaries with the Explorer and Gorda plates should be segment boundarios based on the different behavior of these other plates. Segmentation within the Juan de Fuca plate cannot be assessed with any confidence. Maxinum Magnitudes and Earthquake Recurrence
, Intra-slab: The historical record and analogies to intra-slab earthquakes globally suggest that M7 is a reasonable maxinum magnitude. The historical seismicity record provides a reasonabic basis for assessing earthquake recurrence on this source.
l_ 3 M osOMAMW l' 1 l' Interface: The characteristicu~of the Juan de Fuca plate suggest that.-it is
- i. unlikely that it.would' behave coherently in a single earthquake rupture, but a maximum magnitude estimate can not be made with any
! confidence. Recurrence on the interface'cannot be made with any l confidence, i-I l c p l-L 1 I l l' l L l t i 1 i
I. GEOMATRIX W- E 127' 12G* 125* 124* 123* 122* 121* 120* 119 ' I a I i l g l. O I IOOkm 2 5l g i a z 9 . . HORIZONTAL SCALE .J L { ,.Tu c $ , 10 VERTICAL EXAG. f { I, y g g
- O mas a S
$w .- .:22 2 zJ $
~ 2 km O f
'8 ll l CASCADES 2 km .
l PUGET BASIN g _. _ ._ - ' cASCADIA BA
. 2 km _ _
l' f I i t I f i A' A' 127 126* 125* 124' I-
- 121*
I c 120* l 19
- I 4 -4 %
y 'W ..~ "-' O,
/
tar 1905 9-u co
~- ~
o , ,.
-50 km 8 /(" ~ ~
N $1 (er l?4f 4 . h A E u %dt
-100 km ,.,
O 100 km O Cre ..<, c r a5l //oeo g ,3 , ,1,.
-150 km NO VERTICAL EXAG.
I I I i 1 l l t EXPERT #5
PHASE I RESPONSES BY EXPERT 16
.Ge* metry.
Basis- for geometry sketched on cross-section provided:
- Taber's refraction data in the offshore region are good e The slab should be about 100 km beneath Mt. Ranier to reach proper magmatic depths e seismicity data show down-dip tensional mechanisms, which is typical of intra-slab seismicity. The plunge of the t-axis of the.1965 earthquake
-is from teleseismic data and, therefore, not as affected by local structure as the mechanisms from the local network data.
- Th2 relative weight given to the geometry models shown are: Model A 70% + .05 based on T-axis of focal teleseismic data Model B 30% + .05 based on seismicity distribution C nvergence Rates Minimum convergence rate is 13 mm/yr, which is Savage's compression rate based en geodetic data. This value is given a weight of 0.1. 40 mm/yr is given a weight of 754 based on global reconstructions. It is unlikely that rates are higher than this because there are no large forces (slab pull), the ridge has rccently realigned, and the pole shifts show that the rate is slowing down. F211owing distribution given: Rate Probability 10 - 20 mm/yr 0.04 20 - 30 mm/yr 0.04 30 - 40 mm/yr 0.4 40 - 43 mm/yr 0.1 43 - 50 mm/yr 0.4 50 - 50 mm/yr 0.02 S31smic sources and Activity Potential seismic sources are the following: e Intra-slab source e Plate interface
^
e Upper plate above Blanco fracture zone (because of distance to site, this source is not considered further)
N GEOMATRIX Prcbability that these potential sources are seismogenic: Intra-slab: 1001 based on historical seismicity data Intsrface: 65% (115%) based on the young age of the slab as compared to other subduction zones, the convergence cate is slowing down meaning that it is being resisted, and resistance is giving rise to a rotation of the ridge and change in pole location. Locetion of Rupture Intra-slab: Earthquakes should be confined to the upper 10-15 km of the slab. In cross-section, they may occur 75 km (125 km) west of the free-air low (tranch) down to a maximum depth of 75 km. 75% (115%) of the M16 earthquakes would be expected in the bend region and where they have.been observed historically. The relative rate should decrease to the west by the square root of the distance. Interface: The downdip extent of the seismogenic interface should be about
.45 km (15 km) depth. Updip extent is at a depth of about 20 km, which is essentially the end of the crust in the overlying plate.
Hypocenters for earthquakes below M7 may occur candonly over the interface, but above M7 they are expected to nucleate toward the base of the interface. Along strike, segment boundaries to interface rupture and their credibilitie, cro the following: Blanco fracture zone (50% i 25%), Savanco fracture zone (no probability given), major change in the free-air gravity at 46' (25% i 25%). Ths probability of a rupture breaking more than a single segment boundary is low (10%). M9ximum Earthquake Magnitudes Intra-slab: Uniform distribution between 6-3/4 and 7-1/4 based on historical record. Interface: Use rupture dimensions given. Rtcurrence-Related Parameters Coupling: a = 0.610.15 (unif orm) based on dip angle (low dip = high a ), young age, and rate of convergence. Interface recurrence: should weigh equally estimates based on moment rate with Adam's paleoseismic estimates l Intra-slab recurrence: use historical seismicity
a I GEOMATRIX i y C 127* 126* 12 5' 124' 123' 122' 121' 12 0 ' ~ 11 9 ' I I i i I I ,1 I
-g o 100km O. z E = a HORIZONTAL SCALE 8 i dg5 -
10 VERTIC AL EXAG. & z o I G
, ;g g g ma a s
$$ g : .:22 2 s
py g OLYMPICS 2 CASCADES 2 en a > 2 km 0 o 2 km - 9 PUGET BASIN CASCAdlA BAS 4 -2 km - p-2 km l I i t I f I l i A' A' 127 12 6' 125' 124' I
- _
IR*_ _ _ _ 121' 120* 119' I i i~ ~ -
"I c y ~ O, TMac f*e 3
-50 km
- (cr Ib f I
* ** b ""
- Im
~~ ' 00 *
- M.AlA def, d w J,F u lp l -
O 100 km
-850 km -
NO VERTICAL EXAG. I I l 1 ! I _ l l EXPERT #6 Note: Cross-section redrawn to reflect that the offshore line labeled Tabor 1983 is the crust-Moho boundary, not the top of slab. 1 I
" N"*
PHASE I RESPONSES BY EXPERT 57 Geometry Ba31s for geometry sketched in cross-section provided: e Refraction data in the offshore, gravity data, and depths required for-magma generation suggest the general geometry given. o "Knee-bend" beneath Puget Sound may be due to phase changes in the slab, changes in the absolute-velocities of the slab, or the bulldozer-effect as the slab goes under continental crust. Knee-bends are relatively common in downgoing slabs.
- EMSLAB data beneath Vancouver Irland also support this model C*nvergence Rate 42 mm/yr (+ 10 mm/yr) as published by Riddihough (JGR, 1984). The rate between tha Juan de Fuca and Pacific plates is very well known, but because the P;cific-North America rate is uncertain, so is the Juan de Fuca-North America r;te. Bear in mind that these convergence cates are averaged over geologic time periods-and that the instantaneous rate is not known.
Petential Seismic Sources and Activity Intra-slab source Plate interface Probability that potential sources are seismogenic: Intra-slab: (no estimate given) Interface: 0.3 (+0.2) (see discussion under Recurrence-Related Parameters for basis) Locations of Rupture Intra-slab: Seismogenic part of slab should be upper 20 km based on age of the plate. The locations of earthquakes within the slab should approximate the distribution observed from historic seismicity data. The observed concentration beneath the Puget trough region may be due to 1) phase changes brought about because the slab is going into the mantle f aster to the north, or 2) a ' corner" in the slab that is accommodated by phase changes into the slab.
casoMATmax Intcrfaces (low confidence in this assessment). .One would expect the contact zone to be to the west of the Coast Ranges and to be very narrow due to the underplating occurring along the margin. The true
' amount of compression occurring along the margin is a function of
.the position of the first slab bend, the trench roll-back
' velocity, and the absolute velocity of the upper plate. Along strike, the Nootka fault zone and the Blanco fracture zone are likely segment boundaries. Tears or other segment boundaries that could relate to plate interaction are not evident in the Juan de Fuca plate.
Maximum Earthquake Magnitudes Intra-slab: no estimate given. Intceface: (low confidence in this assessment). A rupture of the entire plate (Nootka to Blanco fracture zone) seems to be the maximum rupture possible but have little basis for suggesting this is the case.
' Earthquake Recurrence-Related Parameters Intra-slab: Historical seismicity provides a reasonable basis for estimating intra-slab recurrence.
Intcrface: The behavior of the Cascadia subduction zone suggests that the interface is totally locked (a = 1.0) or tctally lubricated (a = 0), but not in between. The relative weight given to these models are the following: 0 0.7 (+ 0.2) 1.0 0.3 (+ 0.2) Profetence for the aseismic condition is based on the lack of historical thrust c>orthquakes, the steady uplift observed from the leveling data, and the good
- Svidence for extensive sediment subduction and underplating seen in the Vincouver Island geophysical work. No independent estimates of recurrence are givsn.
I
f GP.OMATRIX E W 125* 124* 123* 122* 121* 12 0
- 189 '
1278 126* I I gi i 1 ,I i I o O IOOkm - t a n
~ s =
i HORIZONT AL SCALE : [ $ ,;ge [; s10 VERTIC AL EXAG. $ Z T
- G I
, ,p, { *
.N g ass a za .
3 a ass 3 y OLNICS l CASCADES 2 km - -2 km 8 8 l PUGET BASIN 3_- . 2 km CASCAdA BAS 2 km - 1 I t e t I a 1 127' 12 S' 125* 124* _ 12 1* _ 812*_ _ _ _ 121* 120* 189'
, , : - - - w -- e i n i f .
O' TMar l*<& $ N g - g s w *8
-50 km
- o l O uw ' IMS
-100 km C"" -**bo -
'c Dw! er.(
O 100km
-150 km g NO VERTICAL EXAG.
t I I f I t I I EXPERT #7
~ PHASE'I RESPONSES BY EXPERT 88 Geometry Basis for geometry sketched on cross-section provided e ' Slab geometries determined f rom seismic ref raction, seismic
~
ref1wetion, seismicity, and other data. e The same general model for slab beneath Vancouver I. should apply to Washington because have same age of plate, sediment supply, etc. e Velocity models used for earthquake locations in southwest Canada are based on new ref raction results (1983 on). Convergence Rate 40 mm/yr (+ 10-15%) based on the Riddihough analysis. Note that he finds a de:rease in rate through the last several million years. His results are pecbably best for the more northerly parts of the Juan De Fuca plate, which 10 most appropriato for the present study. Sei"mic Sources and Activity Potsntial seismic sources are identified: Intra-slab source ! Plate interface Prebability that the potential sources are seismogenic: Intra-slab: 1.0 based on historic seismicity , Interface: range of 0.25 to 0.75 with preference for 0.4 to 0.5 on l' the following basis: The probability should be greater l that 0 because we see evidence of a sole thrust in the l ref raction and reflection data. The "E", zone seen in the l .Lithoprobe results, appears to act as a zone of decoup-ling, which is highly affected by fluids and sediments. The latest results show that the top of the slab is below the E zone by 1 to 1.5 see and is, therefore, separated from the E zone. The relationship between E zone and current rubducting plate is not totally clear.
.- - -_ - - . , , _. , .~
GEOMATRIX Location of Rupture Intra-slab: . Earthquakes would be expected in the upper 8-10 Km of the slab. In cross-section, the eastern extent should be at about 70 Km depth based on the seismicity data, heat flow data, and magnetotel-laric data. To west, intra-slab earthquakes may occur to the toe of the slope (125' west) where the slab first bends. In map view, the relative frequency of earthquake occurrence should approximate that' observed historically (i.e. higher concentration in Puget Sound
' region).due to a probable corner of the plate in this area.
Interface: Seismogenic part of the interface should be limited updip at
-about 124.5*-125' where the high velocity unit (above E1 ) pinches out. Downdip the interface should extend to about 123' where conti-nental mantle would come in contact with the slab. Along strike, the Nootka f ault zone would be expected to act as a segment boundary since reflection and refraction data show that it is a major boundary and crustal earthquakes are associated with the fault zone.
Maximum Earthquake Magnitudes,(low confidence in this assessment) Intra-slab: .no estimate, but note that it is a young plate with critical geotherm (500*C) limiting vertical rupture dimension to
< 8 km.
Interface: Reasonable to use maximum rupture dimensions. Ricurrence Related Parameters Intra-slab: No estimate. Interface: Seismic coupling estimated at 25% (range 204 - 50% baced on evidence for imbricated deformation. Undarplating or l j suberetion is taking place including sediments and perhaps some l -- of the oceanic plate. As a result of imbrication, the thickness ! of the sediments actually increases down the interface. This type of deformation is expected to give rise to aseismic behavior or only small earthquakes. l
l GEOMATRIX E W 12b' 12 5* 124* 123* 122* 121* 12 0' 11 9
- 127*
l 3 3 yI i 4 ,1 1 IOOkm o O - 9 i a p m . { HORIZONTAL SCALE alO VERTICAL EXAG.
.J p
T
=
{
- c 05 g;[
cj, ma ., -, 3 8 5m
- a .a g
3 OLYMPICS 2 2 .422 2 { l CASCADES 2 km 8 { 2 km . PUGET BASIN _g _ _. _ CASCADIA p 2 km .
.2km I 1 I g g 1 l g 127* 126* 125* 12t . _
_ 821*__ _ _ _ 12t* 120* 119' I i 6 i- - , - -
~
U i
/ - N ., "O .i ,
Tde 196$ e o
-50 km O -
'%r e O, 1917
-100 km Yeg dughgE, ek i O IOO km
-150km 3
NO VERTICAL EXAG. t I I I I I I I l l EXPERT #8 i i l l t t
PilASE I GEOMATAIX RESPONSES BY EXPERT 99 Ge~ metry Bacis for the geometry sketched on the cross-section provided: e The hypocentral locations and focal mechanisms of the deeper seismicity beneath the Puget Sound region suggest that these events are occurring within the upper part of the downgoing slab e The depths required for magma generation provide a suggestion of bending of the slab e The cut-off of selsmicity downdip in the slab is due to a change in the physical properties of the slab C'nvergence Rate 42 + 10 mm/yr based on studies by Riddihough. This rate is consistent with ch:rtening rates of 25 mm/yr based on onshore and offshore deformation analyses by Adams. S^ismic Sources and Activity Potential seismic sources are the followirig: e Intra-slab source (S) A,B. e plate interface e Deep intra-slab source 'C' (not considered further because probable lack of significance to hazard at site) e Strike-slip faults within upper plate (such as the St. Helens zone) east of 123' e Accretionary wedge faults (a generic fault (s) west of 123') e Tears in the down-going slab Probability that these potential sources are seismogenic: Intra-slab A 0.9 Intra-slab B 1.0 Intra-slab: 101 in deeper portion below 600* geotherm (east of about 121-122') "C'.
..y . . - _ _ - -
dN osoMArsax - Interface: .The fo11owing. distribution is specified: P (seismogenic)- relative weight 90 - 100% 0.75 80 - 90% 0.24 0 0.01 This distribution is for the probability of generating large-earthquakes (M > 8) and is based on worldwide analogies suggesting that it would be unusual for this margin to be aseismic, the turbidites offshore suggest possible large earthquakes, the stress provinces suggest a large locked area, onshore deformation
~
confirms active convergence, and terrace, tilt and geodetic data. TOcts: if the slab is segmented by tear faults,100% likely they would be seismogenic above 600* geotherm. Accretionary wedge faults: 1004 that a fault in this region is seismogenic based on the M 5.3 1904 earthquake
- Striku-slip zone (s):
100% that a fault of this type is seismogenic based on the 1946 Vancouver Island event, which had a strike-slip mechanism. These zones may be the result of the obliqueness of the convergence vector. Locations of Rupture (~ I ' Intra-slabs the seismogenic part of the slab should be the upper 10 km. The pattern of historical seismicity provides a reasonable basis for estimating the relative frequency of earthquake occurrence along strike and downdip. l Interface: two models of the location of the seismogenic interface in cross-section l l 124.7* - 122.7* 80% likelihood 124.7* - 122* 20% likelihood Preference is given to the first model because 122.7* is the l , margin between the Puget Basin and the Olympics as well as close to the Weaver and Smith stress boundary. In map view, the l-interface may be segmented at the Blanco fracture zone and the Nootka f ault zone (80% likelihood). Additional possible segment at the Michaelson and Weaver segment boundary; which is also near a change in observed crustal stress (20%; +30s, -20% likelihood). The margin is narrower in southern Oregon. l l
asomarpux
; Accretionary wedge faults:
- seismogenic f aults _ should lie between about 123' and 124.5* above the interface. To the west of.
'124.5* deformation accurs aseismically in soft-sediments.- Analysis by Adams suggest that the rate of deformation decreases to the east within the accretionary' wedge..
TeOrs in slab: ' most likely: location 9ould be northern Oregon (see segmentation of interf ace dicussion above). - Strike slip faults: between the western edge of the Fuget Trough to the.cen-
. tral cascades; at the eastern boundary.of the stress ~
province of Weaver and Smith. Maximum Earthquake Magnitude Intra-slab (shallow part, "A"): _the following distribution is given: 6.5 0.1
.7. 0 0.25 7.5 'O.55 8.0 0.1 based on worldwide analogies and the age of the oceanic crust.
Intra-slab (deep part, "B"): 7.0 0.1 7.5 0.8 8.0 0.1
' based on the historical seismicity record and the thickness of the brittle part of the slab (10 km).
Interface: -Use the rupture lengths developed earlier to assess Mmax. Favor large rupture based on turbidite data, locked zone from focal mechanisms. Accretionary wedge: Use the following distribution
, 7.5 0.8 8.0 0.2 -
I based on dimensional arguments, i Strike-slip faults: % 7-1/2 based on Weavet and Smith's assessment for the St. l Helens zone.
s< 4 GEOMATRIX 4 L
- Earthquake Recurrence-Related Parameters
~
. Intra 7 slab *A*: use worldwide dat to determine seismicity of this age _ ocean floor.
Intra-slab *B": Use historical seismicity data to constrain recurrence for intra-slab sources. Intstface: . Primary basis for recurrence on the interface is the 430 yr (+25%) recurrence interval derived from the turbidite data. Seismic coupling may serve as a basis for comparison and should have the following distribution: 5%
< 95%
v I I I
'O- 0.5 0.7 1.0
- Coupling is believed to be high based on the evidence for deformation in the overlying plate in the onshore and offshore; and the tide gauge and terrace deformation suggest that their zero isobases have the same positioned relation as at the Shikoku I, Japan.
Te2rst Should already be included by intra-slab seismicity. - Accretionary wedge: Use historical seismicity data for recurrence. Strike slip faults: Use Weaver and Smith's earthquakes for the St. Helens zone and compare the total rate for the strike slip region with this rate. l l r
7 GEOMATAIX W E 1278 126* 125* 124* 123* 122* 121* 12 0* 119 ' I 4 I a 4 I I i g O IOOkm 2 5 i i y a u e { HORIZONTAL SCALE > s 3 ,cto $ , alO VERTICAL EXAG. $ { I
- y .2 a d m ., 8 5m 2052l 2 -f a z .J g .J sy 5= OLYMPICS 2 i JCA3 CADES zu
-2 km g ! 2 km .
PUGET BASIN
.o ---- - _ -_ .
C ASCA01A B ASIN
--2 km 2 km .
I I f I l l _l I A 1278
- A' 126* 125* 124,; _
f
- 121* 120* 119'
- U 4 I -4 l '" I l y N O, ,
T4r 1905
. go ,, "mk , %p c} h .n J< % < * *I b <- _
sL. , , . ..>
'l,y %N N9 -100 km w4,, s(al, g gi httf O'
100 km
" d* *' 18 * -150 km g NO VERTICAL EXAG.
I I I I I I l t EXPERT #9 4
PHASE I
^
RESPONSES BY EXPERT 010 Grometry
. Basis.for geometry sketched on cross-section provided:
e' Focal mechanisms for the 1949 and 1965 earthquakes, as well as those for the smaller-magnitude events are primarily extensional, thus they likely. lie within the upper part of the subducting slab. e Iacreasing dip to the east is required to attain magmatic generation depths and is in agreement with the T-axis-of the 1965 earthquake. e T-axes for the smaller magnitude seismicity are not very well-resolved from the available data. They do not have to coincide with slab dip as they are likely controlled by local structure. o Waveform modeling studies, such as those by Crosson and Owens, that suggest steeper slab dips are tentative at the present time. c*nvergence Rates 35 - 50 mm/yr (+19 mm/yr) based on global studies such as that by Nishimira et cl. Uncertainty is 954 confidence level of Hishimica. Note that these rates are cveraged'over the pact 1 my and any more recent changes in rate can not be determined from available data. Offshore and onshore deformation rates are comewhat lower than 35 - 50 mm/yr. Stismic Sources and Activity Potential seismic sources are the following: e Intra-slab source e Plate interface o Accretionary wedge: (Example of this type of source is the 1945 Mikawa earthquake, which was accompanied by surface rupture and to have recurrence intervals of about 10,000 years. It is assumed that this type of source will be modeled by the known crustal faults in the site region and/or as a random crustal source.) Probability that the potential sources are seismogenic: Intra-slab: 1.0 based on the occurrence of the 1965 and 1949 events which are inferred to be intra-slab events. , Interface: 0.7 (0.6 to NO.9) based on the following: e No other subduction zones show complete absence of thrust earthquakes.
I reEOMATRIX i e Comparison with other margins in terms of plate age, convurgence rate, etc., suggests large thrust events are possible. e Rivera plate (1932 event and aftershocks) and southern end of 1960 Chile rupture are examples of young plates that have had thrust events, o Southernmost Chile is as quiet as the Cascadia zone but it has recently cubducted a ridge - more analogous to coastal California. e Preseismic compression can be expressed by compressional outer rise events; very preliminary analysis of the 5.1 event.off Oregon suggests that it does not appear to be a compressional event. e There is some evidence that 1949 and 1965 type of events occur at subduction zones which are strongly (seismic) coupled. Locations of Rupture Intra-slab: Elastic part of slab should be the upper 10-20 km based on the age of the plate. The larger intra-slab events (M>6) should occur below depths of 30 km and as deep as 60 km, which is the approximate maximum depth of the historical events. In map view, the observed deeper seismicity has a nonuniform distribution but this may be due to differences in detection capability, particu-larly from Washington to Oregon. A 50% weight is given to modelling the intra-alab seismicity according to the historical distributions and a 50% weight is given to a uniform distribution. Interface: Downdip extent of seismogenic interface should be at depth of about 30-40 km based on global analogies. Updip extent should be to about the trench (long. 125') based on OBS studies in Japan. The Blanco fracture zone and Nootka fault zone should serve as segment boundaries to interplate rupture. The evidence for internal segmentation of the Juan de Fuca plate is weak. t M9ximum Earthquaks Magnitudes Intra-slab: 7-1/4 (+1/4) based on the young slab age and the size of the 1949 event. Interface: Use two approaches and assign a 50% weight to each:
- 1. Rupture dimensions as specified above.
- 2. Age vs. convergence cate vs. magnitude relationships of Ruf f and Kanamori. (Note that this relationship arrives at e,ssen-
[ tially a moment rate, not necessarily the maximum magnitude.) l t i
GE OMATF4tX R currence-Related Parameters Intra-slab: The historical seismicity data set provides a reasonable basis for assessing earthquake recurrence. Interface: Seismic coupling on the interface is estimated to be 30% (with a maximum range of greater than 0% to less than 50% ) based on the plate age vs. coupling relationship of Kanamori and Astiz. This relationship shows a decrease in coupling for progressively younger plates, that are younger than 20 my. Tsunami and turbidite data suggest that the recurrence interval for large interface events should be longer than 150 yr. Use the seismic coupling estimate as the primary constraint on recurrence, with the 150 yr estimate as a check.
r = .- - GEOMATRIX E 3 127* 126* 12 P' 124* 123* 122* 121* 12 0* 11 9 ' I i i gI I I ,1 i O 100km - 3 i HORIZON T AL SCALE 1
.8 7'
e-
't JE
{o Q. o sIO VERTIC AL EXAG. $ ',, 9 ,] 2 5 Z o G ma s u . S W5 . a s ass s 7 _ f.y OLYMPlCS l CASCADES 2n . 2km-
-2 km 8 3 l PUGET BASIN
_o _ _ _ _ _ _ _ _ _ _ _ - o .,
- 2 km C ASC AOI A D ASJN -2 km -
t i I f 8 I I I Al A' 127 126* a q 125* 124' i l' v - - _ 11 2' _ _ _ 121*
" i U 120*
i 119' a T,(,, sig $
' N c@, -
O~ ; ,
** ** J' ,i ' * * <* -53 km ' S f ( te - t fl
- IMS, ,
~ -sGO sm ,_ 0 c, ,
( , ,, 4 _ gg,Le b o .~) u ,! O IOO km ,
-IS0 km NO VERTICAL EXAG.
i i i t i I l I EXPERT #10 l l L
PilASE I GEOMATRfX RESPONSES BY EXPERT 511 Geometry C: sis for geometry sketched on cross-section provided e Pocal mechanisms for d,-ep events beneath Puget Sound and the Georgia Strait indicate normal faulting; essentially no thrusting. e These events should be near the upper part of the slab based on tempetature constraints e T-axes of the 1965 and 1976 events are about 30' and support the increase in dip to the east; the depth necessary for magma generation also support a steeper dip as well as theoretical considerations of buoyancy due to phase changes e The hinge in the slab near the coastline is supported by seismicity data along the Canadian margin e Lithoprobe results combined with seismicity data are similar to the geometry shown Convergence Rate 42 mm/yr (+ 10) Based on the analyses of Riddihough Note that the evidence suggests that the rete is decreasing and the most recent rates are averaged over the past 1/2 million years. Therefore the contemporary rate may be less than 42 mm/yr. Seismic Sources and Activity Potential seismic sources are identified: Intra-slab source Interface source Deep crustal source (that may not be identified at the surface) Probability that the potential sources are seismogenic: Intra-slab: 1.0 based on historic seismicity Interface: 0.9 (.8 to 1) based on analogy to other subduction zones such as southern Chile. If the margins were aseismic it would be anomalous implying special conditions and we don't see evidence for this.
GEOMATRIX De:p' crustal source: ,The source of 1946 Vancouver Island event is the type of event for this source. Although the 1946 event was coincident with the Beaufort fault zone, it apparently had no aftershocks (suggesting a deep' crustal source) and a focal mechanism that does not appear to be consistent with an intra-slab source. If such a source exists, its probability of activity is 1.0. Location of Rupture Intra-slab: within the upper 10 km of the slab. Updip and downdip extent
. based on high concentration of instrumental seismicity. Large events in depth range 40 to 70 km (east of 124'). Earthquakes up to about M5 might be expected to west of about 124' based on the youth of the plate, historical seismicity, and analogies. Along strike, use distribution of observed seismicity to define the location and relative rate of intra-slab seismicity.
Interface: In cross-section the seismogenic part of the interface is . expected to be west of the intersection of the continental Moho with the slab at about 123', and west about 100 km to beneath the top of the continental slope. Along strike the Hootka fault zone and Blanco fracture zone are expected to act as segment boundaries separating the Juan de Fuca plate from the Explorer and Gorda plates respectively. 1 The Michaelson and Weaver proposed segment boundary is very unlikely because there is no evidence for it in the seismicity data (i.e., 3 similar boundaries identified in other subduction zones are seismogenic). No other segment boundaries are evident. De:p crustal source: Above 30 km depth anywhere within the upper plate l along the entire margin. Maximum Earthquake Magnitudes , s Intra-slab: 7 1/2 (slight preference for 7-1/2) based on historical seismicity and thickness of seismogenic part of slab. Interface: Maximum magnitude should be based on dimensions of rupture (given above). Ruff and Kanamori relationship does not have direct applicability to this margin except to infer that large earthquakes are possible. Detp crustal source: 7-1/2 based on size of 1946 event (7.3) and the fact that a large event would likely result in surface rupture.
. I t
r
GeoMATmix
,3, Recurrence Related Parameters Interface ' Recurrence intervals for-large earthquakes (1 8-1/2) should be a minimum of 300 years based on historical observation of quiessence and Adams recurrence from turbidite data. Because of the essential
- l. ~ absence of thrust events, a characteristic earthquake or maximum moment recurrence model is appropriate (as opposed to an exponential magnitude distribution).
i Intra-slab: Use seismicity data to estimate earthquake recurrence. l-l Deep crustal source: From historical record, recurrence interval for M7 [ sho'lld be minimum'of 150 yr. 1 I l l 9 l 1 l
GEOMATRIX E } 1278
~
126* 12 5' 124* I?3' 122' 12 t* 120' lis' i I I ,1 i I I I g = 0 100km -
= E a HORIZON T AL $CALE J $ $ ccI g [3 .
Ig 2 ? slo VERTICAL EXAG. $. ! 9 La., 8 5, a a ' z .;
- a -22 2
-W 5 2 2 OLYMPICS CASCADES
$3 ;
- 2 km .
- 2 km - O l GET BASIN 3 --._ ----
C ASCADIA BASIN , , 2km I I f f I i 1 I
+
d# A D.,;-
~ -50 km Or,. ea .1 ..,.
No q cred/Ndo 6 L.y . f ,I Jm a, g,. N L pk _
-100 km O IOO km l -150 km SCALE NO VERTICAL EXAG.
i l i I I I I J l EXPERT #11 l
PHASE I GEOMATmiX RESPONSES BY EXPERT #12-Geometry Bacis for the~two geometries sketched on the cross-section provided:
- Taber's-refraction data and seismicity data provide principal basis e The presence and position of the volcanic arc argues for steepening and against flattening of the slab e Steeper slab models (e.g., Crosson) must be confirmed by positive velocity evidence for existence of the slab at greater depth R21ative confidence in the two models is the following:
Progressive steeping model 0.8 Flattening model 0.2 Convernence Rate 3 - 3.5 cm/yr (with a 60% likelihood); ranging from 2.5 to 4 cm/yr Bared on Riddihough analyses and Atwater vector solutions Seismic Sources and Activity Two potential seismic sources are identified: Intra-slab source Interface source Prebability that the intra-slab source is seismogenic: 95 - 100% based on historical seismicity (with some uncertainty die to the mechanism and location of these events). Probability that the interface is seismogenic: l 20% (with a range of 10% to 40%) based ont e No thrust events have been observed along the entire length of the l margin e Global analog subduction zones are seismogenic e 1949 or 1965 earthquakes may have occurred on the interface, although they were more likely intra-slab events
I f omoMAmm , Location of Rupture I
' Intra-slab source:
The historical' seismicity record provides a reasonable basis for the location and relative frequency of earthquake occurrence, both down-dip and along strike. .The seir.mogenic part of the slab is 6 - 10 km l thick. -0.05 probability that.' future locations are completely random, l Interface source: Downdip extent of rupture is at about 50 km depth where the slab comes into ' contact with cor.tinental mantle. Updip extent is probably
-limited by eastern extent of underplating; but this location is uncertain.
Segmentation of the interface along strike probably occurs at the Juan de Fuca-Explorer plate boundary and, to the south, with the Gorda i plate. There is a 30% likelihood that the change in orientation of the volcanic trend at southern Vancouver Island represents a segment boundary. Maximum Earthquake Maanitude ( Intra-slab source: I o
.The largest historical events (1949, 1965) probably provide a basis for the maximum events but there is a 20% chance that the maximum event could be larger.
Interface: l l No real basis for making assessment. Snismic Coupling and Earthquake Recurrence Ssismic coupling (a) estimates to be between 0.05 and 0.5 with a preferred
- l. value of 0.1; there is an 80% likelihood that a lies between 0.05 and 0.15.
The basis for the coupling estimate is a consideration of all types of data related to subduction including the historical seismicity reccrd, back-arc epreading, slab age and dip; sediment accretion of underplating; Chilean
- vs. Mariannas type behsvior, etc. The geologic evidence tends to favor low
! coupling along the Cascade zone. Coupling and convergence rates provide a l reasonable basis for estimating earthquake recurrence along the interface. ! .The historical seismicity record may be used to estimate recurrence on the j intra-slab source. ( ( I l
r l GEOMATAIX 'W . E 127* 12G* 125* 124* 123* 122' I ?_ l* 12 0* 189 ' I i i g i i 1 ,4 I O -lOOkm - a - = 3 e a z g HORIZONTAL SCALE J E I ,8 6 slo VERTICAL EXAG. $ [" I I if 2 f dg m a maa w 422 ' za g py g OLYMPICS 2 2
) CASCADES zo . -2 km 8 8 2 km -
NPUGET BASIN C ASCADIA BASIN , ,
- 2 ko I I t I t t I 's A' *
- A' 127 126* 125* 12 4* _ _l _ _ 121* 120* 189*
i I" " U I I . i-
. y*i
^90 * - - -
.{og o( slob ,
- 50 km % S k U f' * **$
~
*nb fro"tss^Ndd 1947 \
~' '* -
- 4. r .t 3I . \.
O' 100 km I t *'S
' * "3 -- -150 km -
NO VERTICAL EXAG. I I I t i f f I j, EXPERT #12
PHASE I hh GEOMATRIX RESPONSES OF EXPERT $13 Geometry Basis for geometry sketched on cross-section provided: o Deepor hypocenters are separated from shallower events e Depths of some events (e.g., 1965) are too deep for interface events o Normal focal mechanisms are typical of top of slab extension and graben formation seen within slabs worldwide e Magma source depths are typically about 100 km, but this is a thin, hot slab so depth contraint not strong e would not expect piece of young thin plate to break off and stay seismically active Note that we have reinterpreted your drawing to reflect the fact that the T;bor (1983) line represents the transition to oceanic Moho, not top of oceanic plate (see attached). C*nvergence Rate 43 1 mm/yr Based on analysis of Nichimura et al. (1984) In agreement with evidence of deformation and accretion offshore. Seismic Sources and Activity I Two potential seismic sources: e plate interface e intra-slab Probability that intra-slab source is seismogenic l 80s (1 about 10%): based on occurrence of deeper seismicity with j normal mechanisms and on the estimated age of this slab and the l convergence rate. l Probability the interface source is seismogenic 51 1 51: Based on fact that essentially all deep events down to , magnitudes less than 3 do not slow thrust mechanisms, but show down-dip tension consistent with intra-slab seisnacity. Historical record also precludes interface events.
+
'4 OEOM/dMIX Location of Rupture s
Intra-slab source: e 90% of the moderate-to-large events will occur between 50 and 80 lan depth due to age of the plate, maximum downdip tension, possible
' bending of plate e In map view, may expecte higher. concentrations where have had them historically (e.g., Puget Sound region); perhaps due to fabric >
resulting from extension of Savanco and other Explorer Plate fracture zones into this area Interface source: o Seismogenic portion will extend f rom east of accretionary wedge (15-20 km depth) to depths of_40-50 km: accretionary wedge not sufficiently strong to stor1 seismic energy; 40-50 km cut-off depth consistent with worldwide cases and the age /cate of Juan de Fuca plate e Segment boundaries generally at boundary with Explorer Plate and . with Gorda plate Maximum Magnitude g Intra-alab source: the following magnitudes and associated probabilities: l e Stated values 6-1/2 (80% j; 10%), 7 (70% j; 10%), 7-1/2 (25% j; 10%), 8 (14) translate to 6-1/2 (0.45), 7 (0.4), 7-1/2 (0.14), 8 (0.01) e Based on historical seismicity and age and rate of subducting slab I Interface: l e Stated values 4 (35%), 5 (254, 6 (10%) translate to: 4 (0.5), 5 (0.35), 6 (0.15) e Based on young hot plate, slow convergence and correlation to worldwide subduction zones and large amount of sediments that may be contained between plates. i coupling e Seismic coupling: 5% (+ St) e Based on the young age, low convergence ratos the large amount of sediment being subducted; even with accretion, a large amount of
- sediment is being subducted Recurrence e Intra-slab - use historical seismicity
-U 1 ( GEOMATRIX is W E 1278 126* , 12 5' ' 124' 123* 122* 121' 12 0* 189 ' , I i 4 a i i i i 100km j j O. ~ H . . HORIZONTAL SCALE 'J- I- $ .ja $ , s10 VERTICAL EXAG. $ [* Z f g 3 8 5m a ma s - za 3 a 2 2 azz y -
* = OLYMPICS
$$ j CA3 CADES
-2 km 2 km -
GET BASIN
--2 km - - - - - - ]- - -
CASCADIA B ASIN 2 km - t f f I t t i 1
*N Nef lI 1278 126' 125' [ I24' .
__ IM* _ _ _ 121* 120* 119*
, , i .- - . --
TMa/ - $
% al N i
'N O, ,
c 116 * % o
-50 km
- 4. s.l.~ do , ** o ,
d c4 e .'s w k3 W
-100 km ( vp g (,..g. s -
O 100km
-150 km g NO VERTICAL EXAG.
t I I t i I ! I EXPERT #13 k
?
PHASE I GEOMATRIX RESPONSES BY EXPERT #14 , Geometry
")
Basis for geometries given on cross-section provided: 4e Depth to magma generation should be 100-125 km and used locations beneath Mr. Ranier. e The deeper earthquakes beneath Puget Sound do not appear to be North American plate earthquakes; either they are within the Juan de_Fuca plate'or in a remnant slab. m Two models are given e ' Model A (shallow slab dip): based on assumption that deep earth-quakes are occurring within the upper part of the oceanic slab. This is compatible with the extensional focal mechanisms of the small events as well as the 1949 and 1965 earthquakes. e Model B (shallow and steeply-dipping interfaces): based on Crosson's and Owens' inversion results calling for a steeper dip to the Juan de Fuca slab. Note that this model does not assume that the deep seismicity beneath the Puget Sound is in the North American plate, but that it is occurring in a wedge between two interfaces. Relative preference is given to Model B, and it is given a weight of 65% (115%).
~
Convergence Rate On the basis of published estimates such as those by Riddihough and Nichimura et al., the following distribution is given:
> 50 nu/yr 0.05 40 - 50 mm/yr 0.30 30 - 40 mm/yr- 0.40 20 - 30 mm/yr 0.20 10 - 20 mm/yr 0.05 S^tsmic Sources and Activity The following potential seismic sources are identified, as a function of g20 metric model:
Model A e Intra-slab source e Interface source e Accretionary wedge faults e Tears in the down going slab
^
L t
,ju e' *2=
M GeOMATRtX The probability that each of the potential seismic sources are seismo~ genic e io the following: Model A e Intra-slab:- 1.0 due to occurrence of 1949 and 1965 events e Interface: 0.35 (10.15) based ons ,
- Seismic quiescence is marked and no thrust earthquakes have occurred
- The thermal history of the Juan de Fuca plate strongly suggests that the plate is very hot due to blanketing of the plate very near the ridge; the sediments prevent convective loss of heat as well as water circulation; therefore, the plate has a very young thermal age
- Relatively low rate of convergence
- Analogy to other subduction zones suggest that some possibility exists for interface earthquakes and very long recurrence intervals, e Accretionary wedge fault: 70% (110%)
- There are many known faults between the coast and the trench showing' young displacement, some may be seismogenic l
e Tear in slab: given that a tear exists, probability of it being l seismogenic is 0.05 (10.05) based on:
- To function as a later tear fault, different movement rates of the slab would be needed and there is no evidence for this. ,
I
- For dip-slip, would require large differences in slab age and this is not the case.
i l Model B , 1 e Shallow interface: 0.23 (10.15) l
- lower than Model A because would expect a lovet strain rate and less likelihood for stick-slip behavior i e Deep interface: 0.3 (10.1)
- lower than shallow interface in Model A because it would be deeper and hotter; r.bsence of observed seismicity e Deep intra-slab: 0.1 based on absence of observed seismicitv t
i I l l r l . I t l
1 M osoMArmix e Shallow intra-slab (Model A) would be remnant slab (Model B) with the probability of seismogenic the same (1.0)
-e Accretionary prism and tear fault the same as Model A.
Location of Ruptgre Model A
' Intra-slab: Seismogenic thickness of slab is about 10 km In cross-section, intra-slab events will occur between 124 longi-
.tude on the west to 70 km slab depth (about 122 on the east.
80-90% of the larger earthquakes (M15) would be expected in the slab bend area, where the 1943 and 1965 events occurred.
~
Along-strike distribution should reflect the relatively higher rates of occurrence where they have occurred historically, which is also the location.of the change in trend of the plate. Interface: In cross-section, the seismogenic part of the interface would be expected to be between the trench axis (125 ) on che west to a depth of about 50km on the east (at the bend). 50 km is the typical maximum depth of thrust events on subduction zones worldwide. l l Alongstrike,segmentboundarieswouldbeexpectedatthenorghern t and southern ends of the Juan de Fuca plate (about 44 and 49 latitude). A low probability (0.2 i 0.15) is given to the likelihood that the Juan de Fuca plate interface is internally segmented. Major ! differences lo slab geometry (such as proposed by Michaelson and l Weaver) would be expressed in the volcanic axis and they aren t.i If a segm9nt boundary exists, it would be coincident with the southern cutoff of seismicity south of Puget Sound and on strike with the northeasterly convergence direction. Accretionary wedge: potential seismogenic faults would be expected ; between the coastline and halfway down the continental slope, with ' the highest likelihood at the shelf / slope break. Model_B Shallow interface and shallow intra-slab (remnant slab) same location : in cross-section as Model A. Along strike, the shallow interface and remnant slab do not extend south of Puget Sound zone of seismicity. ! Deep Interface: In cross sections, extends from trench to depth of 50 km; same '.ateral constraints as Model A. i i i
, t, osoMArmx -
y ' Deep-intra-slab: Location of rupture is not known. Tear: TIf a tear exists, it would most likely be at the segment boundary described above. M"ximum Earthquakes Model A Interface: 8 (19 5) (0.66 weight between 7.5 and 8); dimensional arguments that give estimates of magnitude 9 are not applicable.to the Cascadia subduction zone due primarily to-different strain
-rat s.
Intra-slab: 7.25 - 7.5, which is slightly larger than historical observation (1949, 1965). Accretionary wedge: 7 (10.25). The occurrence of earthquakes in the [- wedge is localized by accretionary processes (fluid pressures, t 4 etc.) and the expected seismogenic area is limited. Model B Shallow interface: 7.25 (10.25) less than Model A because of slower strain rates resulting'in different rhelogy. Accretionary wedge and remnant slab (shallow intra-slab) same as. Model A. Deep interface: .75 for 7.5 to 8, .25 for 8 to 8.5 Deep intra-slab: No basis for estimate. Tear fault: 5. There is no need for either significant strike-slip or dip-slip displacement in the slab. Rrcurrence-Related Parameters Model A Intra-slab: The historical seismicity record provides a reasonable basis for estimating recurrence. Interface: (FaArly low confidence in estimating seismic coupling) e High heat flow, low convergence rate suggest that the plate should have a very young thermal age.. e Seismic coupling estimated at 10% (15%) based on Kanamori relation between slab age and coupling. L . l
4 oeonaarnix e For interface recurrence, use coupling and maximum carthquake estimates. Model B Shallow and deepLinterfaces: Coupling val'.e of Model A is appropriate but convergerace rate should be partiti;ned between the upper and lower interfaces with a ration of 1:1 to 1:6, respectively (preference at 3:1). Accretionary wedge: Adams finds about 25 mm/yr of shortening in the accreticeary wedge; an estimated 5% aay be expected to be released as seismic energy. , Assum g that the plate interface is seismogenic (by either Model A or Model B), a maximum moment recurrence model is appropriate.
GEOMATAIX E k 127' 126' 12 5' I24* 123' 122' 121' 12 0' l19' i I a i i l i I 0' IOOkm -2 5: e a z a u . g i HORIZONTAL SCALE .J E- -! ' -ju $ , slo VERTICAL EXAG. f {* I
- 8y g*
de O mas a g : 422 z .J
= OLYMPICS 2 s -
i' GE zu ~~- g CASCADES l 2 km O 8 2 km - l PtliET BASIN C ASCADIA BA 2 km -
.2km t i I f f I l i IA 127 ' A' 12 6' 125' 124' - 12 1' - 112*_ _ _ _ 121' 120' 189' i i
- .pw e i O
n g y - T 6,e l'JO S - O %S3 #'A Ek
-50 km Modc( A Llw (,o -
%,h#
o Top E p100 km gp pik M )d Q - T- Jok a. O 100 km k850 km g NO VERTICAL EXAG. 1 i t i I I I I EXPERT #14 l I I I
+ - - -
)
l v PliASE II INFORMATIONAL MATERIALS
. SFET TO EXPERTS PRIOR TO FOLLOW-UP INTERVIEW _
v im
000 Markat Plaza Spour Stroet Tower. Suite 717
$5}N$7 GEOMATAIX SAMPLE LETTER SENT TO EXPERTS September'18, 1987 Project 1133A-1.4 Prof.
Subject:
Satsop, Washington Seismic Hazard Aaalysis
Dear :
Many apologies for the long delay in getting back to you concerning the subject study. h e past several months ~can be summarized by the following. ~The tectonic assessments for the hazard analysis were completed at the end of November,1986. The results of this analysis essentially provide a quan-tification of the probability of occurrence of various magnitude earthquakes on various seismic sources. ne next step of the hazard analysis specifies the level of ground motion at the site given these earthquakes. For these predictions, grcund motion attenuation relationships are required that des-cribe the rate of ground motion amplitude variation with distance as a func-tion of magnitude. R ese relationships are readily available for shallow crustal earthquake sources. However, relationships for subduction zones, particularly for sites within 30 km, potentially directly above the plate interface, magnitudes 2,8, and on rock sites have not been previously developed. As a result, several months of the project were devoted to ground motion analysis. The ground motion analysis has included the development of appropriate empirical attenuation relationships. Considerable effort was focused on developing a data set of close-in, rock site recordings for both intraplate and interface earthquakes. Here, data from the 1985 Chile and Mexico earthquakes have proven invaluable. The preliminary results of this effort were presented at SSA in April, 1987 (Youngs et al, Seis. Res. Ltr., v. 58,
- p. 29, 1987). ne ground motion-related studies were completed in July of this year.
So at present, we are equipped to carry out a full hazard analysis for the Satsop site. We are, however, concerned that during the intervening time, the opinions of one or more of the experts may have changed in light of recent studies. In addition, we feel that it may be useful for each expert to see the calculated results of his assessments and to re-examine his assessments in light of those given by all of the experts. As a result, we are asking you to review the attached materials and to participate in a-telephone interview aimed at identifying any changes in your previous assessment. Geomatr Ix Consultants, Inc. consuit.no Eno.neecu and Earu, sc.entots
! l
-Prof.
September 14, 1987 " " ^ " "
'Page 2 hhe materials attached'to this letter are the-following:
Attachment l'.- A summary of your assessments-(referred to by your expert number). This summary includes not only the direct assessments given by you (e.g. , probability of activity), but also- the calculated results of your assessments.
~
For example, you may have specified that the maximum magnitude son the plate interface be calculated from the dimensions of rupture: Attach-ment 1~shows the magnitude values that result from these dimensions. As another example, you may have specified that earthquake recurrence be calcu-lated from the convergence rate, seismic coupling, and a particular recur-rence model. The calculated recurrence relationships based on these speci-fications are given in Attachment 1. Please review these results and-their c associated uncertainties. Bear in mind that it is possible to change either the parameters or the calculated results (e.g., recall that maximum magni- ! tude or recurrence intervals could be assessed directly). Attachment 1
-(along with the cross-secticas in Attachment 5) will provide the basis for the present reassessment.
Attachment 2. A detailed summary of the assessments given by all of the experts. These are the individual assessments (each checked for accuracy) given by the experts. These are important because they provide the scien-tific bases for the conclusions drawn. You will find yours according to your expert number. Please bear in mind that these summaries reflect opinions given in 1986 and are subject to modification. Attachment 3. Overall-summary and analysis of expert assessments. Ihis attachment summarizes the range of expert opinion given for the parameters ' of interest to the hazard analysis. This document may help provide a context for your responses, although no attempt has been made to call out individual' expert opinions. Again, please remember that these assessments. are subject to change. Attachment 4. Recent references related to the Cascadia subduction zone. The papers attached have either come into print or have been accepted for publication since our last meeting. Also included are the abstracts of papers that we are aware of that have been submitted for publication. Copies of these papers are not yet available, subject to acceptance for publication. Attachment 5 Seismicity cross-sections. As an aid in evaluating your assessments of crustal geometries, we are providing updated seismicity cross-sections. The uncertainty in hypor.entral location is given by the error bars. .The cross-section of particular importance for this analysis is Cross-section F, which passes through the site. 8 L. - . - -
Prof. September lli, 1987 """^mm Page 3 The procedure that we will follow for the reassessments will consist of the { following. Please review the attached materials, paying particular atten-tion to your previous assessments, their calculated results, and the assess-ments made by others. We will call you in the near future to set up an appointment for a telephone interview, which will occur in late September or early October. During the interview, we will step through your assess-ments and the essociated uncertainties and ask you for any changes. As was done previously, we will make out best attempt to record both your assess-ments and the basis for them. Our record will then be sent to you for
, review to ensure accuracy.
Once again, we apolegize for the long delay in following up on this hazard assessment. As far as we know, it is one of the largest seismic hazard analyses involving expert opinion that has yet been undertaken. Your participation is invaluable end greatly appreciated. We are pleased to offer a $250 honorarium as a small expression of thanks for your efforts on this phase. As promised, the complete Satsop Seismic Hazard Analysis will be provided to you upon completion. We look forward to speaking with you soon. Best regards, Kevin J. Coppersmith Project Manager dla l l Attachments l
. GEOMATAIX ATTACIDENT 1 Sununary of Phase I Assessments for Each Expert Including Calculated Results
Ji- [ N GEOMATRIX ., 1 ( Expert 1
SUMMARY
OF MODEL PARAMETERS
' Slab Geometry: '
s s Three dips: Model A ~10* (0.2) Model B - 15* (0.5) Model C - 25' (0.3)- ConverRence rate: 30_t10 mm/yr
,,i [
-The resulting distribution for convergence rate mm/yr is percent: 0 10 20 30 4 50 prob mm/yr +----+----+-- -+----+----+----+----+----+0----+----+
0.050 20.00 ****** 0.300 25.00 ******************************* 0.300 30.00 ******************************* 0.300 35.00 ******************************* / 0.050. 40.00 ****** *
.' prob value +----+----+----+----+----+----+----+----+----+----+
k Seurces and probability of activity: Model A - Interface 0.4 (0.25 - < 0.5) Intraslab 1.0
~
a Models B and C - Interface 0.4 (0.25 - < 0.5)
- Intraslab (Juan de Fuca) 0.1 - 0.15 "deep zone" beneath Puget Sound 1.0 '
MOximum extent of rupture on interface: Updip - toe of continental slope Downdip - depth of 40 to 50 km (equal weights) Along strike --Nootka to Blanco (0.8) Nootka to Blanco segmented at 46*N (0.2) Interface Maximum Magnitude: Use maximum rupture dimensions Model A - unsegmented area = 134400 to 160000 km2 M 9.25 segmented area = 67200 to 80000 km2 gv 9,o Model B - unsegmented area = 76000 to 92000 km2 g W 9.0 segmented area = 38000 to 46000 km2 M 8.75 Model C - unsegmented area = 28800 to 38400 km2 M" 8.5-8.75 segmented area 14400 to 19200 km2 M" y 8.25-8.5
=
4 +
1 1 4i ,y Q 'P - s GEOMATRIX - l
- s. -
-(
Interface Earthquake Recurrence:
- Use. moment rate approach with':
moment rate = convergence rate *a* interface area a = 0.05 (0.58) or 0.95.(0 g ) .( .e use "maximum moment" magnitude distribution. + :; Attached Figure 1 shows the resulting distribution of: recurrence 31 estimates for interface events. Maximum event magnitude is assumed-to. be uniformly distributed in the range of the expected maximum magnitude given above t 0.25 magnitude units. Repeat times for maximum events range from 300 to 20,000 years. Figure 2 shows the effect of choice of a onLrecurrence estimates for Model'A. The remaining variation in recurrence estimates shown.in Figure 2 reficct the effects of variations in convergence rate and maximum magnitude. i & Lo' cation of intraslab events: Model A - 95% between 122*W and 124*W Models B and C "dcep zone" between 122*W and 124*W intraplate shallower than 50 km Along strike - match observed relative frequency Intraslab Maximum Magni!, .'
- Model A and "deer r for Models B and C - 7.25 t0.25 resulting distri ,
percent: 0 10 20 30 .40 50 prob value +----+----+----+----+----+----+----+----+----+----+ g 0.300 7.000******************************* t 0.400 7.250***************************************** 0.300 7.500******************************* prob value +--- '----+----+----+----+----+----+----+----+----+ Models B and C - 5 to 6 uniform distribution Intraslab Earthquake Recurrence: Historical seismicity used to compute a- and b-values for exponential
-model. FortfnTraglaInevents in Models B and_C, the seismicity rate was estimated from offshore eventF~wiliiil nt.he Juan. de Fuca_ pjate away from the spreading centers and fractdri zohes. Figure 3 shows the
. recurrence relationship used for the intraslab events in Model A and the deep zone in Models B and C. Thiscurveisbase[s'onallrecorded events not inferred to lie within the North American plate. Figure 4 shows the recurrence relationship for the offshore Juan de Fuca' plate used to model the intraslab recurrence for Models B and C.
4 >m . . . . . . . ,. w > . +-. , u ,
-s.
... , w ;%< . 4 . a-1 gi.,c s . : 4. 3 j
' , , .,. . , - 3 s On, ' 4h :' M ,
e i.
'i 9le?s.
l>,u.c o~h p>- 7
' lr , ' ' h'; __ s h i._ , f . gy?)' '
l _i , L, s q e.-...; . p ;.. :. vyf. - y y [ .4
.j , . , '$p >
Jt _ gu -
. 1: _ - casomarpeex ;
y; , < - <
' > , t -
- y. , , d
, > ^ g jo .; - *-
.. , 4 , ! l .i .i i ..
* ,. d, . , ,. 3
, . .. .,. < ". - t m
,ps. ,
, y -
- 1
. N :' ,
t :s, * '
,s . .1 , ,
. l
, ';s . l )<j _.
'fmyt,1Modelk , .
- s'
. .M
, ' - - - ;.- Evpwt,1, Model 8 .
~
W '-)/JJ - ...,*E9mt i Mod.tl C i . 8 , f m - .
- i,
' . ' _O, c'4 .i Q"
c ,
, . 10 _. - -
w (> .f .' . / ,,. . u:. u.; c , e -o . 's
.a u
c 3 0_. 2 _.
- c. .
- l. 3' s 5 [
- g _. .. , 1 y
L
.a 9e..%...........
. _ _ _ _ _ _ _ _ _ _ _ . _c _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.
[. -~. n , %, lS, _ . L
~
vi - . v, . ., ( .- $ E, 10-3 .
's s
. s .
- . s .
l.s; . , o - . ..
.. . ... . . . . . .. ........s...............
~. oX : ;
. i L 10-4 :-
- . . . . . . . . . . . . . . . . 4.pe..............
F -
%. i l --
. N i :
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .._ seg
- . . . %.l .
r
!s 9
g . i , i , i S 6 7 8 9 10
- e. eAarnitude s
, a Figure ,1 Estimated recurrence distribution for Exoert 1 i
4
'f%
~ '.:
j:, a.- -
?
,, . -ya s . s :- , ,
e qQ 'i
, GEOMATRIX -
g: 1
- .u 10 ; . ,
, . . .i ,
, .i . _
.na .. -
, . .t Model d lockhd ' '
IJ00+l A urdo3+J ,. 1 _
~ -
10-1 -
~
... Er - m. s . u. e -
- <-, 3' u c .10_2
- c. _
m-5 e
. ,y . , -
.g . ' ' ,
- g - .-
,?" '
)'
< Cff __
l 5 ~~ L) ..
~
.3D*4 ,
s , 3 -- s- % g - - 4 1 !_ s . L , . ..
~____________________trX__..__________
e l 10-' - N , _____________________s_o_%_. -x :
- .
- ________________s '
s _
.ry,________________,s,, s :
, s. _
s ._ s i. F k yo_s . i . r- . i . i 5 6 7 8 9 10-Magnitude . Figure 2 Effect of Alpha for Expert 1 n
.,+
=
>l
, .1 _. _ _ _ . __ . _ - - _ _ . . . _ , . -. . . . -
i b GEOMATAIX FIGURE.3 8-value - Deep Zone
;s. 200 g. .
g . i .
- i. . .
j . 100 :-
% - 0.10 -
50 - b = 0.71 -
.(1 0 .
O 20 - 10 -- 5 - 2 - - G c
, 1 3
W
- l. * .
E *5 -. - o . . 3 j .2 - !~ ( .I _~
~
O 2
.05 ' - -
i . 6
.02 -
O ( l .01 . .
) 5
.005 -
.002 - -
' I ' I ' ' I ' ' ' ' ' '
.001 ' '
O 1 2 3 4 5 6 7 8 Magnitudo
GEOMATAIX , i >t B-value --
. FIGURE h 'Of f Shore
..200 ,
100 -- 50 - - b - 0.71 - 20 - - 10- r. - 5 - -
- 2. - -
D g 1 _ a cr a_ - O
- j. .5
~
l l ao . 6 : l C j g .2 - 0 0 0 0 -
.1 r
- O ( :
.05 - -
.02 - -
.01, --
.005 - -
.002 - -
l . .001 O 1 2 3 4 5 6 7 8. Magnitude (- i l -
4 5 GEOMATAIX
. Expert 2
SUMMARY
OF MODEL PARAMETERS Slab geometry: Approximately 10' dip through deep seismicity ;9, Convergence rate: 15 to 30 mm/yr with the following distribution-perciant: 0 10 30 40 50 prob mia/yr +----+----+----+-- 20 0.100000 15.00 ***********
-+----+----+----+----+----+----+
0.400000 20.00 ***************************************** 0.400000 25.00 ***************************************** 0.100000 30.00 *********** prob value +----+----+----+----+----+----+----+----+----+----+ Sources and probability of activity: Interface 0.8 (0.7-0.85) Intraslab 0.6 - 0.7 l L Maximum extent of rupture on interface: Updip - coastline Downdip - not assessed, use AGGREGATE distribution for maximum depth of rupture (km) percent: 0 10 20 30 50 i-prob. d (km)+- --+----+----+----+----+----+----+-- 40 -+----+----+ 0.083 30.00 ********* j 0.250 35.00 ************************** l 0.317 40.00 ********************************* l 0.125 45.00 ************** l 0.225 50.00 ***********************
+____+ ___+____+_.__+.___+ ___+ ___+.___+.___+____+
Along strike - Nootka to Blanco (0.6) Nootka to Blanec. segmented at 46*N (0.4) Interface maximum magnitude:
.Not assessed - use AGGREGATE distribution a) 0.55 weight assigned to estimate from maximum rupture area ur. segmented model - rupture area 64800 to 133600 - H 949e25 segmented model rupture area 32400 to 66800 - My8Y75-9
, a GEDMATRIX l
R
, b).0.45 weight assigned to following distribution-percent: 0 10 20 30 prob Mmax-50' O.073333 7.50
+----+----+----+----+----+----+----+--
40-+----+----+ ' 0.073333 7.75 ******** 0.073333 8.00-******** 0.390000 8.25 ****************************************
'O.098373 8.50 ***********
0.083333 8.75 ********* 0.083333- . 9.00 ********* 0.083333 9.25 ********* 0.041667 9.50 ***** prob
+----+----+----+----+----+----+----+----+----+----+
Combining a) and b) the distribution over all maximum rupture geometries is percent: 0 10 20 30 40 prob Hmax 0.033000 7.50
+----+----+-~--+----+----+----+----+----+----+--
50
-+
0.033000 7.75 $*** ' ' O.033000 8.00 **** 0.175500 8.25 ******************* 0.044250 8.50 ***** 0.180500 .8.75 ******************* 0.329000 9.00 ********************************** 0.153000 9.25 **************** prob
+----+----+----+----+--- +----+----+----+----+----+
Interface earthquake recurrence: 0.3 weight assigned to geologic estimate of 430 (t25%) yrs 0.7 weight assigned to moment rate approach moment rate = convergence rate
- a
- interface area Assessed distribution for convergence rate mm/yr is give above a'pha not assessed - use AGGREGATE distribution of experts
\
GEOMATRIX ' Y
. percent: 0 5 10 15 20
. prob a 0.032
+----+----+----+----+----+----+----+- --+----+-- 25 0.00 *******
-+
0.191 0.05 *************************************** 0.145 0.10 ******************************
.0.045 0.15 **********
0.036 0.20 ******** 0.050 0.25 *********** 0.046 0.30 **********
~0.037 0.35 ********
0.024 0.40 ******
"X- 0.020 0.45 *****
0.023 0.50 ****** 0.028 0.55 ******* 0.033 0.60 ********
;0.038 0.65 *********
0.041 0.70 ********* 0.030 0.75 ******* 0.019 0.80 ***** 0.015 0.85 ****
'O.011 0.90 ***
.0.043' O.95 **********
0.093 1.00 ******************** prob
+----+----+----+----+----+----+----+----+----+----+
Magnitude distribution model not assessed - use AGGREGATE e assessment of thn experts exponential. (0.23) ' characteristic (0.41) maximum moment (0.36)- . Figure i shows the resulting distribution of recurrence estimates. Figure 2 shows the effect of variation in acximum magnitudn on recurrence estimated using the moment rate approach, and Figure 3 shows the effect of choice of magnitude distribution model on recurrence estimates.
..Tha' earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows:
Return Period (yrs) for events of dagnitude 1 M Hagnitude M Sth percentile 50th percentile 95 percentile 5 7000 20 l' 0.4 6 7000 75 2 l 7 7000 275 6 8 - 630 60 . 9 - 10000 400 I
+- w ~ , ,,- r .
---e
GEOMATAIX Y 4-Location of intraslab events: Updip extent -123*W Downdip uncertain but 80 -'90 % near bend at 122*W. Along strike - variation not assessed, use aggregate opinion 0.9 variable matching observed seismicity pattern 0.1 uniform along strike Intraslab Maximum Magnitude: Not assessed - use following AGGREGATE distribution _ percent: 0 10 20 40 50 prob Mmax +----+----+----+----+----+-- 30 0.041 6.60 *****
-+----+----+----+----+
0.030 6.75 **** 0.355 7.00 ************************************ 0.219 - 7.25 *********************** 0.345 7.50 ************************************ 0.000 - 7.75
- 0.010 8.00 **
prob
+----+----+----+----+----+----+----+----+----+----+
Ir.i raslab earthquake recurrence:
-Not assessed - use historical suismicity. Figure 4 shows recurrence relationship for deep earthquakes assumed to be occurring within downgoing slab.
This curve is bases on all recorded events not inferred to lie within the North American plate. I 4 A _ , ,,r ~ r ' " ^ ' '
-7,,
s' ~ . i
~ '
< s. :7
> s ,
.b ' -
c
,- 4 GEOMATRIX :
.,t f
,.()=.
- y > -
'u
. s --
. g!-
, - i.. i i
,i ; i 4 i
- i. i. .
- y. 1 m
- f l,
.w -j _.
- - 97 :
s . 4 i 0 -
' .e .
ji. , u,
..o . g 19-2 __ *O
-I(E i 3o -
~
N
~
a ~ ~
--f.
, j o-3
., - . ~ .
\ -
3 ~ c 10_4 . N . 10-6 ' ' ' ' ' ' ' 5 6 7 8 9 10 f.4agnitude Fioure 1 Estimated recurrence distribution for Expert 2 r T *--r m W4-
-> 74. . > y }. ~.
',I, /
L , L s uaOMATRIX ' b J 1
+
=
' 10 _ ,
Chxacteris6c Mu from geomstri _
- c. uu = 7.5. moment rot + ~ ,
i q 7
. s, .
10-1 : . - N - - s% - : j :
. l .
e- e.
- u. . ' .
-3 r 1 a
c 10-2 : _ i _ j. E . 9 - . l - 0 - t . e i s , . i- i y i . i l1 10 : - l -
~
l e i, .
. (- .
l
. t .
, . 8
- i. - t
[- I l-10-' - e 7 ! . l . g
. I .
- t. .
l . i f 1 I I I i l 10'~0 5 6 7 8 9 . 10 Magratude Figure 2 Effect of Max Magnitude for Expert 2
GEOMATAIX 10 _ , EgonenSol
- - - - Choroctedett; a
* *
- IAowimum flornent 1 -
's 10-1 -
N, _- x 9
'N 's :
B a
's - -
r.r -
's E 's u- -
's s _
~O a
's s
's E 10-2 __
s, __ <- _.....................a . s _
- .s ,
.,y - . o -
.s .-
.~
. s, -
~
~
o - -
- h. s.
.s 10-3 --
\., .,
.s -
\, .
g _ 10 -' -
\ -
\ 3 g _
g g 3 g-s . 1 . r . t . t 5 6 7 8 9 10 tAagnRude Figure 3 Effect of mea gnitude distribution model for Expert 2
w-
- j_
~
GEOMATAIX
, '\
FIGURE . B-value - Deep Zone 200 . , . , . , ,
, . g .
i . . 100 r
%- 0.10 -
50 - - b = 0.71 - (1 O - G 20 - - 10 --
.5 -
2 - - i? C 1 c a cr - N *0 ' ~ o - - o C ' Q .2 - -
.1 - -
- O
.05 - -
1
.02 - -
0 (
.01 r
- ) :
.005 -
.002 - -
.001-
' ' ' ' ' I ' ' ' ' ' i ' f '
0' 1 2 3 4 5 6 7 8 Magnitudo
y . 050MATREM Expert 3
SUMMARY
OF MODEL PARAMETERS Slab Geometry: Approximately 10* through deep seismicity Convergence rate: 40:119 mm/yr The resulting distribution for convergence rate mm/yr is percent: 0 10 30 40 prob mm/yr +----+----+----+-- 20 50 0.050000 20.00 ****** -+----+----+----+----+----+----+ 0.300000 30.00 ******************************* 0.300000 40.00 ******************************* 0.300000 50.00 ******************************* 0.050000 60.00 ****** prob a-
+----+----+----+----+----+----+----+----+----+----+
Sources and probability of activity: Interface 0.6 Intraslab 1.0 Haximum extent of rupture on interface: l Updip - minimum depth 5 km l l Downdip - taximum depth 45 km (resulting width 200 km) Along strike - Nootka to Blanco I Interface Maximum Magnitude: directly assessed according to following distribution l:6 percent: 0 10 20 30 4 50 ! prob Hmax t 0.125000 8.50
+----+----+----+----+----+----+----+----+0
----+----+
l 0.250000 8.75 ************************** O.250000 9.00 **************************
- - 0.250000 9.25 **************************
0.125000 9.50 ************** prob
+----+----+----+----+----+----+----+----+----+----+
GEOMATAIX
-Interface Earthquake Recurrence:
Use geological assessment according to following distribution: percent: 0' 5 10 15 20 25 prob years +----+----+----+----+----+----+----+----+----+----+
.0.006940 200.00 ** -0.'055560 300.00 ************
0.103300 400.00 ********************** 0.104170 500.00 **********************
- 0.097220 600.00 ********************
0.090280 700.00 ******************* 0.083330 800.00 ******************
-0.076390 700.00 ****************
0.069440 1:;00.00 ********x****** 0.062500 1100.00 ************** 0.055560 1200.00 ************ 0.C';8610 1300.00 *********** 0.041670 1400.00 ********* 0.034720' 1500.00 ******** 0.027780 1600.00 ******* 0.020830 1700.00 ***** 0.013890 1800.00 **** 0.006940 1900.00 ** 0.000870 2000.00
- prob +----+----+----+----+----+----+----+----+----+----+
(moment rate approach used as check with : moment rate = convergence rate *o* interface area for a = 0.66 and Hmax = 9, return period = 333 yrs) use "characteristic" magnitude distribution Attached Figure 1 shows the resulting distribution of recurrence estimates for interface events. Figure 2 shows the effect of Mmax on recurrence estimates. The earthquake recurrence estimates shown in Figure 1 can be summarized in terms of return period for events of various sizes as follows: Return Period (yrs) for events of Magnitude 1 M Magnitude M Sth percentile 50th percentile 95 percentile 5 21 7 2 6 87 30 8 , 7 364 132 41 8 1050 430 166 9 - 1320 340
e-,: , o- y r GEOMATAIX ~ 18 3 Location of intraslab events:
.10; to 20 % near outer rise 5;%.at intermediate depths remainer'in deep zone of observed seismicity Along strike - match observed' relative frequency Intraslab Maximum Magnitude:
Mmax = 7.5 in deeper part of slab
= 8.0 in shallow part' Intraslab Earthquake Recurrence:
. model.
Historical seismicity used to compute a- and b-values for exponential Figure 3 shows the recurrence relationship used for the intraslab events. This curve.is bases on all recorded events not inferred to lie within the North American plate. J l l 1 l l l l 1 1
1 GEOMATRIX o V 10 - ' i ' i 4 1 3 s a b m i
~
L a 1 J M M d es e M 4 10-' -N .; r I w% : P:r - e eu . s Ia &f E 10-2 --
'N ._
\ ;
J 3 9N \ \sN - G . A, e
\, ,~~ h. s l 1
o - 10-3 _
\ \ _.
i N \
\
\ -
\ \j
'l 10-' _-
E. ] W N r [ 5 5
- 9
<- 10 Ha p % de r-- . ,.
) hl f f $ l( '
ll l {l l { k{ R 'p 64 f
( k - N' I 7 g pj.j' '_1.
, 2:7 4. - ~ fr - '#. s ..
t,y.: ,
- ~
~
8.. ^ ' pt- _
- c r , o.o r= -
71 ~ 'y;; q
~
I "t'(- h t ft
' < ,, y .. fM i T y ,5'
.s .
', J), .- >
}- !
i r, if, . .; .- ' - wr 1
. 830.
4 y. i, i i, i i
- i. ..
. c. -
a . [6: f.-
. v. , -fc -; Mo - 8.5 -
- uu . n .
l ;; 1- --- - E, .. , - i: ".
's,- _
t' 4 s,.g _ Yy -% , , .10-8 .- 'N, -
- s.s ,s
- _ s -
s ., r- . . v .- c .
< .. g. .
..= .- ,.
- u. _
s% - cC , 3 0-2 _ s,-g y .- , I .. -
- g .
. s .
1 . ~ e . .
~~, _
3 3 o -
.~.a.s s
- j. %
y w t
.30-s -
N _ . g . r ,z . _-.- ' s, e -
- e s
[ .c t -
- c. .
p r . , s, 10-' -
~
so-5 - - - - - 5 6 7 8 9 10 Magdhede Figure 2 Effect, of rnax magnitude for Expert 3
;s, GEOMATAM s .
' FIGURE 3 B-value -
Deep' Zone
-200 ,
100 r ;
; g- 0.10 :
50 - Q b - 0.71 e Q _ O 20 - 10 r - 5 - 2 - D g 1 r ; o -
.e : :
E .5 - E - o o C 4 .2 -
.i z
- O
~
.05 -
.02 -
0 (
.01 .
- ) :
.005 -
- \r -
.002 -
.001
' ' ' ' ' I ' ' ' ' ' ' ' ' '
O 1 2 3 4 5 6 7 8 Magnitudo
e v e i I. n
, GEOMATRIX .j i
l l l j Expert 4 j
SUMMARY
OF MODEL PARAMETERS i Slab neometryr. Approximatelyfl0*' dip through deep ~ seismicity Convernence rate: Following distribution assessed percent: 0 20 prob, mm/yr +----+----+----+-- 40' 60 80 100 0.05 10.00 **** -+----+----+----+----+----+----+ 0.50 ,20.00 ************************** 10.40 30.00 ********************* 0.05 40.00 **** prob
+----+----+----+----+----+----+----+----+----+----+
Sources and probability of activity: Interface Intraslab 0.75 1.0 (10.25) renormalized to 0.075 for events > M 5.0 Maximum extent'of rupture on interface: 1 Updip - 20 km depth Downdip - 40 km depth Along strike - Nootka to Blanco barrier to rupture (0.5) Interface maximum magnitude: Assessed distribution percent: 0 10 20 30 50 prob Mmax
+----+----+----+----+----+----+----+-- 40 -+----+----+
0.300000 0.300000 3.00 ******************************* 4.00.*******************************
~0.300000 5.00 *******************************
0.090000 6.00 ********** 0.010000 7.00 ** prob +----+-~~-+----+----+--~~+----+~~~~*~~~~+~~~~+~~~~+ Renormalized distribution for events > mag 5 percent: 0 20 40 60 80 300. prob Hmax +----+----+----+----+----+----+----+----+----+--- 4 0.90000G 6.00 **********************************************
-0.100000 7.00 ******
prob' +----+----+----+----+----+----+-~~~+----+~~* ' ' *
.m
- 3 3
~
f
' GEOMATRIX
- Interface earthquake recurrence:
use moment rate approach moment rate = convergence rate
- a
- interface area Assessed distribution for convergence rate'mm/yr is give above alpha assessed as follows percent: 0 20 40 60 80 100
, prob a +----+----+----+----+----+----+----+----+----+----+
0.050000 0.00 **** 0.800000 0.05 ***************************************** 0.050000 0.10 **** 0.050000 0.15 **** 0.050000 0.20 **** prob +----+----+----+----+----+----+----+----+----+----+ Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) l- characteristic (0.41) l maximum moment (0.36) I
' Figure 1 shows the resulting distribution of recurrence estimates, Figure 2
, shows the tifect of variation in magnitude distribution model on recurrence j' estimated us?.ng the moment rate approach. l The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: Return Period (yrs) for events of Magnitude 2 M Maanitude M Sth percentile 50th percentile 95 percentile 5 10 0.3 0.06 . 6 15 0.6 0.5 7 - - 51 1 l
_;; 8 GEOMATRIX i Location of intraslab events: Use observed seismicity pattern
-Intraslab Maximum Magnitude:
Assessed distribution as follows: percent: 0 10 20 30 40 50 prob- Maax_ +----+----+----+----+----+ ---+----+----+----+----+
-0.330000 7.00 **********************************
0.340000 7.25 ***********************************
-0.330000' 7.50 **********************************
prob '+----+----+----+----+----+----+----+----+----+----+ Intraslab earthquake recurrence: Use historical seismicity. Figure 3 shows recurrence relationship for deep earthquakes assumed to be occurring within downgoing slab. This curve is bases on all recorded events not inferred to lie within the North American plate. l l i.
GEOMATM' t 10 rg { ; i l 7
, j r 3
i p C N s,>'A . i 5%,% <W' \ \ 4 X'/ - i l l\ 1 V_ e l \ : se W M W W pe 10-1 N-x
\
- i. -
e LL - 3J E 10-2 _. y .
.y -
a - 5 c r 5 - 0 10-3 - m
~
i l 10-' : J se p e be
' i i i g 3-6 .
i l i 1 - 1 . . .J 5 6 3 9 IO M a ,3 nil.u d e PigUfe ! t.S ,t..trYtal.Fi t eOUfTeriC e O!9(TIOUllOri IOr t.
) D "n.
't
'I c3 ,
i< -
' GEOMATRIX FICURE j :-B-value -
Deep Zone 200- , . , , , l: . 100 .
. . - 0.10 .-
50 . - b = 0.71 -
>Q G .
O 20 - 10 :- : 5 - 2 . e D c - 1 -
$ 5 b w .
& 5 -
o . . 3
.2 -
.1 .
- O !
.05 - -
.02 -
0 (
.01 r :
. ) .
.005 -
.002 - -
.001 O 1 2 3 4 5 6 7 8 Magnitudo
GEOMATmf X - Expert 5
SUMMARY
OF MODEL PARAMETERS iSlab geometry: Model A gradually increasing dip approximately 17" below site (0.75-0.80) Model B - approximately 10' dip through deep seismicity (0.20-0.25)- Convergence rates. 25 to.40 mm/yr with the following distribution percent: 0 10 -20 30 40 prob mm/yr +----+----+----+----+----+----+----+----+----+----+ 50 0.028 25.00 **** 0.222 30.00 ***********************~ 0.444 35.00 *********************************************
'0.306 40.00 ********************************
prob +----+----+----+----+----+----+----+----+----+----+ Sources and probability of activity: Interface 0.5 (10.5) Intraslab 1.0 Maximum extent of rupture on interface: Updip - coastline Downdip - 123.5'W longitude Along strike - Nootka to Blanco i Interface maximum magnitude: Not assessed - use AGGREGATE distribution a) 0.55 weight assigned to estimate from maximum rupture area ; Model A - rupture area 43200 - M" 8.75 Model B - rupture area 64000 - M 9 b) 0.45 weight assigned to following Mistribution ; percent: 0 10 20 30 40 50 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0.073333 7.50 ******** 0.073333 7.75 ******** 0.073333 8.00 ******** 0.390000 8.25 **************************************** 0.098333 8.50 *********** 0.083333 8.75 ********* [ 0.083333 9.00 ********* ! 0.083333 9.25 ********* 0.041667 9.50 ***** prob +----+----+--- +....+....+....+....+....+....+...., j l i I
- i s 3
b OEOMATRIX s Combining a) and b)L the distribution over'all maximum rupture geometries is
, percent: 0_ 10 '20- 30 '40 50-
' prob Hmax +----+----+----+----+--'-+----+----+----+----+----+
0.033000 57.50 ****
. O.033000 7.75 **** 'J ' O.033000 8.00-****-
0.175500 ~ 8.25 ******************* 0.044250 8.50 ***** 0.463750 '8.75 ***********************************************' O.161250 9.00 ***************** 0.037500' .9.25 ***** prob- .+----+----+----+----+----+----+----+----+----+----+
. Interface earthquake recurrence:
Not assessed - use AGGREGATE distribution of.other experts as follows: 0.48 weight assigned to following AGGREGATE distribution of geological estimates of return period for~large events percent: 0 20 40 60 80 100 prob yrs -+----+----+----+----+----+----+----+----+----+----+ 0.001388 200.00 *
-0.025988 300.00 **
0.671900 400.00 *********************************** 0.154718 500.00 ********* 0.019444 '600.00 **
= 0.018056 700.00 **
0.016666 -800.00 **
- 0.015278 900.00 **
- 0.013888 1000.00 **
0.012500 1100.00 ** 0.011112 1200.00 ** 0.009722 1300.00
- 0.008334 1400.00
- 0.006944 1500.00
- 0.005556 1600.00
- 0.004166 1700.00
- 0.002778 1800.00
- 0.001388 1900.00
- 0.000174 2000.00
- prob +----+----+----+----+----+----+----+----+----+----+
l 0.51 weight aasigned to moment rate approach ' t.
' moment rate = convergence rate
- a
- Interface area j
Assessed distribution for convergence rate mm/yr is give above
GEOMArFttX alpha not assessed - use AGGREGATE distribution of experts percent: 0 5 10 15 20 25 prob a +----+----+----+----+----+----+----+----+----+----+ 0.032 0.00 ******* 0.191 0.05 *************************************** 0.145 0.10 ****************************** 0.045 0.15 ********** 0.036 0.20 ******** 0.050 0.25 *********** 0.046 0.30 ********** 0.037 0.35 ******** 0.024 0.40 ****** 0.020 0.45 ***** 0.023 0.50 ****** 0.028 0.55 ******* 0.033 0.60 ******** 0.038 0.65 ********* 0.041 0.70 ********* 0.030 0.75 ******* 0.019 0.80 ***** 0.015 0.85 **** 0.011 0.90 *** 0.043 0.95 ********** 0.093 1.00 ******************** prob +----+----+----+----+----+----+----+----+----+----+ Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) characteristic (0.41) maximum moment (0.36) Figure 1 shows the resulting distribution of recurrence estimates, Figure 2 shows the effect of variation in maximum magnitude on recurrence estimated using the moment rate approach, Figure 3 shows the effect of choice of magnitude distribution model on recurrence estimates, and Figure 4 shows the effect of slab geometry on the recurrence estimates. The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: Return Period (yrs) for events of Magnitude 1 M Magnitude M 5th perce.itile 50th percentile 95 percontile 5 4200 17 0.4 6 4200 63 2 7 5250 240 9 8 - 480 60 9 - 12000 540
att s/f.'/ ' 'e GEOMATRIX _4 - 4
-s- :di
- location of'intraslab events:
Updip extent ' coastline s
. Downdip extent - Mt. Ranier Along strike - match observed seismicity pattern Intraslab' Maximum Magnitude:
Magnitude 7 Intraslab earthquake recurrence: I
-Use historical seismicity. Figure'5 shows recurrence relationship for deep earthquakes assumed to be occurring within downgoing slab. This curve is bases on all recorded events not inferred to lie within the North American plate.
i l
.- - , - - , ---_.-, - .- y- r mw~ . + -----~ -- - - - - - - - - - ---
I e GEOMATAIX 10 , , ,
~ 3 1
- \
1 ,
- 9sy, -
jo-i _ x ; y - s a
~
- h. -
3 Sofa s gQ 10-2 - s\ -
,, : \
33 . s 3 p - 5 o -
%.y 10-3 --
s
\ -
\
s%
\ \
% s. \.\
30-4 _ i .
- ' ' ' I t o-6 - '
5 6 7 8 9 - IO I/QQENkude Figure i Eshmated recurrence dieltibution icr Exoert 5
f~ ' W GEOMATAIX to t ' i ' i i ' i~ ' F 'l - t C ct,xasuieu.. anax fm 9m u,
----. ,,m3 . u
-+
1 r - 1 _s,% g g . 10-1 _ s s s. N p : s.,
& _ N, .
e < m _ i 3= ', 8 10-2 , s - l . t$ - J 0 - a -
~
o E - l\. I, 10-3 : l
'\ -
\s
- l \ -
\
\
' \
~
l
, \.
10-' : ;
\ -
l
~
e
,i i . i i, i i 30-6 . . .
s a 7 a 9 . to
).isgnitude Figure z tfiect of Max e.iagnitude r.or typert :-)
W s GEOMATRIX 10 _ i , r- - i i i i .- i
-- Op I 5..vw is 103
- - - - - Dw:ttaie ti-: I i
~
. . . . >u .an.tn us. .ii s
[ [ s 10-1 -s ' s
- 's x -
's 8
e _- 's r
\ s e 's (L. -
's
~O a 's %
5 10-2 _ 's, ; s y _.....................,,,,,
= ,
o - .
.s s s -
S s E
's N
=
o _
'.s~,'..s
. '.s>\
10-3 :-
*\
'.\%
O. . _ \. - 4, 10 " _~ i *=.\. s
~ i t
t 10-6 ' ' ' ' ' ' ' ' \ I 5 6 7 8 9 10 Magnitude Figure 3 Effect of rnagnitude distribution model for Expert 5
V OEOMATRIX
, 10 ,;
i i , , . T- i J
- 1
- Charactuletla - Model A ovarztstette - Hose e 1 r
- ~
'N
's 10-' r -
's -
s
- s x - ss 8
v
's s
\ s -
w e s,
- u. -
's, -
o a
's %,
E 10-2 r -- 4 : 's s e y o s o - s - b - s G * , o 5 ~ 's s s g g-3
's
,' s s g.gs, .
s s
\q -
i t i 10 " r \ 'i -
; \* :
~ \'i
\t s
- 1 1 - 1 - 1 10-5 >
5 6 7 8 9 10 Magnitude I 9 9 V V Fia..ure 4 Effect or elab a.eometry for txper t o
GEOMATRIX FIGURE' I- B'value - Deep Zone ' 200 . j , j , 100 :- ;
- - 0.10 -
50 - b == 0.71 - (1 O' . 20 - 10 r ; 5 - ' 2 - D c - 1 e - 3 -
,o- - l . ,i u
tu
.5 -
o - o . C g .2 -
.1 -
- ()
~
.05 -
l' - t . t
.02 -
() (
.01 -
- ) -
.005 - -
l - . 1 f
.002 - --
,001 ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
O 1 2 3 4 5 6 7 8 Magnitude j t i i 6 6 d i h
, s , , _ - - - . ,
t.. . GEOMATRIX - Expert 6
SUMMARY
OF MODEL PARAMETERS Slab' Geometry: Model A - 10' through deep seismicity. (0.70 10.05) Model'B -~10* with double bend through deep seismicity '(0.30 10.05)
'Convernence rate:
10 to 60 mm/yr with the following distribution: percent: 0 10 20 30 40 50 prob mm/yr +----+----+----+----+----+----+----+----+----+----+ 0.010 10.00 ** 0.020 15.00 *** 0.020 20.00 *** 0.020 25.00 *** 0.110 30.00 ************ 0.200 35.00 ********************* 0.183 40.00 ******************* 0.274 45.00 **************************** 0.14'3 50.00 **************** 0.010 55.00 ** 0.005 60.00 ** prob +----+----+----+----+----+----+----+----+----+----+ Sources and probability of activity: Interface 0.65 (10.15) Intraslab 1.0 Maximum extent of rupture on interface: Updip - depth 20 km Downdip - depth of 45 km (15) l Along strike - Nootka to Blanco (0.5)
- Entire plate (0.5) l Interface Maximum Magnitude
Use maximum rupture dimensions Model A - unsegmented area = 157500 km2 M 9.25 i segmented area = 120000 km2 gW 9.25 ! Hodel B - unsegmented area = 210000 km2 M" 9.5 I segmented area = 160000 km2 M" 9.25
. Interface Earthquake Recurrence:
Use moment rate approach with weight (0.5) moment rate = convergence rate *o* interface area . convergence rate distribution given above t (
r GEOMATRIX a assessed'according to following distribution percent: 0 5 10 15 20 25 prob a +----+----+----+----+----+----+----+----+----+----+ 0.083 0.45 ****************** 0.167 0.50 ********************************** ~' 0.167 0.55 ********************************** 0.167 0.60 ********************************** 0.167 0.65 ********************************** 0.167 0.70 ********************************** 0.082 0.75 ***************** prob +----f----+----+----+----+----+----+----+----+----4 Usekeologicestimateof430i25%withweight(0.5) Magnitude distribution - exponential (0.6)
- characteristic (0.4)
Attached Figure 1 shows the resulting distribution of recurrence estimates for interface events. Maximum event magnitude is assumed to be uniformly distributed in the range of the expected maximum magnitude given above t 0.25 magnitude units. Figure 2 shows the effect of magnitude distribution model on recurrenca estimates, Figure 3 shows the effect of uncertainty in segmentation, and Figure 4 shows the effect of slab geometry. The aarthquake recurrence relationships shown in Figure 1 can be summarized in t,.rms of return periods for various sizes as follows: Return Period (yrs) for events of Magnitude t M Magnitude H 5th percentile 50th percentile 95 percentile 5 9 1 0.2 6 37 4 0.7 7 182 21 3 8 675 118 20 l 9 1860 525 200 l i l Location of intraslab events: ! l Updip - 75 km west of trench ! , Downdip - depth of 75 km 75% of mag >6 near bend j Along strike - match observed relative frequency l l l Intraslab Maximum Magnitude: Uniform distribution 6.75 to 7.25 Intraslab Earthquske Recurrence: liistorical seismicity used to compute a- and b-values for exponential model. Figure 5 shows the recurrence relationship used for the intraslab events in Model A and the deep zone in Models B and C. This curve is bases on all recorded events not inferred to lie within the North American plate. l
e D GEOMATAIX to _ , , , r-
'Nx N.
E' 1sy, _ Sbf' 10-' _
\
- ~
Q' : g .
? -
c _
\%
va T2 10 :
* : .N :
J : '\ .i j ~
\\ s s
~
5
\
~
'\,\
10-3 -
}
8 \
- s 10-4 -
4
\
I 1 - I i 10-6 . - i 5 6 7 8 9 10 Magnituda Figure 1 Eslimated recurrenco distrii>ution for Expert 6
h GEOMATRIX 10 _ i ; ,
; i-----
)
bp. miial
---- *tww,v relsit<
1
~
I 4 1 _- ;
-'s s, -
s 10-' -
's, ..
- 's s I
- 's e - ',
Pr .
's, _
e s,
- u. .
s s D g a s E 10-2 __ s, , ;
-t -
" s \ -
hs y -
's, '_g 2
- 's, -
s p *
~~ '
5 . O ,' s 10-3 ..
\
- s
. \, s i -
\\ I 1
\s 10 -' .L. '4
- i
\
- i
> \ .
\ ,
10 -f' ' ' ' ' ' ' ' ' \ ! 5 6 7 8 9 10 Magnitude Figure 2 Effect of maanitude distribution model for Expert 6
GEOMATRIX 10 _ i i , g - -- , Secoentea
's - - - - Uns-;miw
% s, <
1 - s
- s, -
s
- s -
s _- s s .
's -
s
= %
s 10-1 : -
's s, -
x - s s 8,
, _ 's, .
's s s
u'. - s, s 3a s s 5 10-2 __ v . s s 3c -
~ . ==
4 g -
\s 5
u -
's.
\
\
10-3 --
\
r r., _- \ i -
- \
s 1 i L i 1 10-* -
- I l
10-6 . . 1 - I - I _ 5 6 7 8 9 10 f.h2gnitode H-.avre a tfr.ect or. seamentation for t> pert o-
GEOMATRIX 10 , M
=- UOdei 4 r vue: S \
ss - 1 -- g -.-
- s
\
~
g .
. % =
~~, \
's s (s . %
~ %
e w - B .J
's
- E 10-2 _. ',
3
\ 4 j
A o , 5 ' o - \
\
\
\
10-3 p
\\\. .
~ \\
\ \
s
\s -
_ \\i _ i t j Q'd 7 'n - be be
! Y
. g 5 6 7 8 9 10 Hagnitude
. -r. - - -
tl0.Ure 9 t.ileCl Of S,Gb I G.eornOI I.V for t> pert h
- r.
GEOMATRIX z-
.j FIGURE t] B-value -
Deep Zonc . ? _e 200 . 100 :- -- 50
%- 0.10 .-
b = 0.71 - (1 O . G 20 - 10 :- 5 -
- 4
- l. 2- -
D c
, 1 3 .
.w -
\ -
- 8 .5 : .
~ '
- b. -
i V - o .
.2 -
i
.1 :-
O t
.05 -
t
.02 -
k - 0 ( i
.01 -
- ) ~
.005 -
l -
.002 -
.001 O 1 2 3 4 5 6 7 8 Magnitudo i
[
V GeOMATRIX Expert 7
SUMMARY
OF MODEl PARAMETERS Slab geometry: Approximately 10* dip through deep seismicity Convergence rates . 42 110 mm/yr with the following distribution - percent 0 10 20 30 40 50 prob am/yr +----+----+----+----+----+----+----+----+----+----+ 0.05 30.00 ****** 0.30. 35.00 ******************************* 0.30 40.00 ******************************* 0.30 45.00 ******************************* 0.05. 50.00 ****** l prob +----+----+----+----+----+----+----+---~+----+----+ l Sources and probability of activity: Interface 0.3 (10.2) Intraslab' not assessed - use AGGREGATE assessment of 0.96 Maximum extent of rupture on interface: Updip - depth of 15 to 20 km (equal weights) Downdip - coast ranges depth 30 km Along stt:iko - Nootka to Blanco Interface maximum magnitudes i Use maximum rupture area I rupture area 44000 to 68000 - My 8.75-9 Interface earthquake recurrence Use moment rate approach moment rate = convcrgence rate
- a
- interfaco area Assessed distribution for convergence rate mm/yr is give above alpha assessed as 1.0 (conditional on source active)
Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) characteristic (0.41) maximum moment (0.36) - Figure 1 shows the resulting distribution of recurrence estimates and Figure 2 shows the effect of choice of magnitude distribution model on recurrence estimates.
n - - _ . _ _ _ O t, , GFOMATRIX The earthnuake recurrence relationships shown in Figure 1 can be summarized ' in' terms of return periods for events of'various sizes as follows: Return Period (yrs) for events of Magnitude t M V MaRnitude M Sth percentile 50th percentile 95 percentile 5 630 6 0.4 6 630 22 2
- '. 7 630 100 9 8 630 310 60 1 9 13400 1400 533 Location of intraslab events:
Updip .. downdip, and along strike - match observed seismicity pattern Intraslab Maximum Magnitude: Not assessed - use following AGGREGATE distribution percent 0 10 20 30 40 50 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0.041 6.60 ***** 0.030 6.75 **** 0.355 7.00 *********************h************** 0.219 7.25 *********************** 0.345 7.50 ************************************ 0.000 7.75
- 0.010 8.00 **
prob +----+----+----+----+----+----+----+----+----+----+ Intraslab earthquake recurrence: Use historical seismicity. Figure 3 shows recurrence relationship for
- deep earthquakes assumed to be occurring within downgoing slab. This curve is bases on all recorded events not inferred to lie within the North American plate.
I I i
- _ - . - . =
,F GEOMATAIX 10 _ i i i . J I 1
~\
Wlo _
^
N 10-' -- - ke - d., - t -
~O a
E 10-2 -- - 5 b o - s% 10-3 - 3 ' 10 -' -- 1 4 I l 10-6 . . . I . . 5 6 7 8 9 10 Magnitude Figure 1 Estirnated recurrence distriisution for Expert 7
t*' - 4 GEOMATF11X 10 _ , i . , ,
; , -~ '
T'
- - - ~ ~ ~ ~
m AsNisi , ano nw%tt.:
\ .
. . . leo.Luu.n IJament 1
m 4 D V s s 10-' :-
's s -
g x
~
's h.,
= 's
. s
** s LA. -
s
'% N
-S ,
e - 10-2 2 s _ s
- s
-s -
2 -
's \ J 2 s
-...................................s,\,
o w
\\,
10-3 -- \N
't.
\
\us 1
10 " -- 8 8 4 1 1 so-s - ' - ' - ' - ' ' 5 6 7 8 9 10 MapRude Fiaure 2 Effect of rnoonitude detributiori inodel for Expei! -
,p % ,
r q -}..y. q.+m ' *
~ -
CBEOMATHtX I r;; , l; q3 3
, tg \ I y 4.
, _ , [~k ? 7IGURE 3 B-value - Deep Zone ; u..
-- J200' -
i - i
- . i - .
i . . . 100- r-.
. .e
~
- i o g= 0.10 '
- M - . l b - 0.71 - !
i O G ~ G 20 - .
.10 -- -
5 - . 2 - 1 . d l '. t T. .
~
.5
{ - . f - ,. j .2 - ..
.i -
() :
.05 - -
.02 -
3 O t
- .01 -- -
) .
L' .005' I ' l - .
.002 - .
.001 O 1 2 3 4 5 6 7 8 Magnitude l
I l l~,
m - ( ~ L > OsOMATalM ' i g Expert 8
SUMMARY
OL)iODEL PARAMETERS
. . ~
Slab Reometry: . Approximately 10' dip through deep seismicity Convernence rate: 42 110 mm/yr'with the following distribution c percent 0- _ _10- 20 30 40 :50 prob am/yr +----+----+----+----+----+----+----+----+----+--->+ 0.05 30.00 ******- 0.30 35.00 ******************************* 0.30 40.00 ******************************* 0.30 45.00 ******************************* - 0.05 50.00 ****** prob +----+----+----+----+----+----+----+----+----+----+ l Sources and probability of activity: ; Interface 0.45 (0.25-0.75) Intraslab 1.0 Maximum extent of rupture on interface: Updip - 124.5'W to 125'W (equal weights) Downdip - 123' W Along strike - Nootka to Blanco Interface maximum maanitudet Use maximum rupture area rup+.ure area 89600 to 119100 - yM 9-9.25 Interface earthquake recurrence: Use moment rate approach moment rate = convergence rate
- a
- interface area Assessed distribution for convergence rate mm/yr is give above alpha assessed according to followint di s tribution percent: 0 10 20 30 40 50 prob a +----+----+----+----+-- -4 --+----+----+----+----+
0.017860 0.15 *** 0.142860 0.20 *************** . 0.260710 0.25 ^*********************** ** 0.228570 0.30 **********************<e 0.171430 0.35 ****************** 0.114290 0.40 ************ 0.057140 0.45 ******* 0.007140 0.50 ** prob - +----+----+----+----+----+----+----+----+- --+----+
GEOMATRIX Magnitude distribution model not assessed - use AGGREGATE assessment "x - of the experts --- exponential (0.23) characteristic (0.41) maximum moment (0.36) Figure 1 shows the resulting distribution of recurrence estimates and Figure 2 shows the effect of choice of magnitude distribution model on reciterence estimates. The earthquake recurrence relationships shown in Figure 1 can be summarized ' in terms of return periods for events of various sizes as follows: Return Period (yrs) for avents of Magnitude h M Mannitude M 5th percentile 50th percentile 95 percentile 5 2500 14 0.3 6 2500 57 1 7 2500 270 6 8 2500 500 36 9 4400 1700 380 Location of intraslab events:
, Updip - 125'W Downdip - 70 km depth Along strike - match observed seismicity pattern Intraslab Maximum Mannitude:
Not assessed - use following AGGREGATE distribution percent 0 10 20 30 40 50 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0,041 6.60 ***** 0.030 6.75 **** 0.355 7.00 ************************************ 0.219 7.25 *********************** 0.345 7.50 *******************************t**** 0.000 7.75
- 0.010 8.00 **
prob +----+----+----+----+----+----+----+----+----+----+ - l Intrastab earthquake recurrence: Not assessed - use historical seisuicity. Figure 3 shows recurrence relationship for deep earthquakes assumed to be occurring within downgoing slab. This curve is bases on all recorded events not inferred to lie within the North American plate. i
- y. - .. - - . 4- ..
.. 4, 2
j.#. ,-
...gu, -. ;
.~
gg . ,
, m "' ,
- .7,, . < < ,
,ar ,
-.<- r 9.t 4 1*
'~'
i , , . t.,-
- , t- . s s
_. -S's - s
+-
- GEOMATRIX _
,) y
<., : ; ; -n .
.. ~~r .
t.'<
* - ,a\-:.t-- .:L .; ,i ., i
- ~
4 , ii 10<;.
,_ .c ,.
-, , , ,_ _ ;. j . _ , .
r s er - < , et ,
- a s
1- - r e
"- 2
. g..
- g - -
gg off _ i 10-1 :.
- r - n -
i - 9 _
- { .-
s
'u_. _-
Te e sj,. c . suUe -2 _ .Q~- \ _
- o. _
g
. . s
<s.
m 10-3 --
. r
.O L /o [
l _ - N s N, \ ~ 10-' :- .
\ .
i 10-6 ' ' ' ' ' 8 - I ~ . 5 6 .7 8 9 10 Magnitude
/F ,
Figure '1 Estimated recurrence distribution for Expert 3
,, f.
\ \
A em m, ,,s 4 mr,, E.-',. , .J.s,, , i - ~ -ye- + v
W7:. . ; 2.f ( s .p'
~~. '
Fl %';,,,. J - ':Jr ,'
' i < :- > - <
d[y I '.pN%Jf; i A ' .
< , t ,
it ffi, , GEOMATRIX - g,
- . a- .i . : ' '. .. 1 ' ' - ' -
s 2
- t -
, c .
1102 _ i
- j. . ; .
. 3- .. 7---, ; j=
, ! i.(- . _'. 4 y < _
--
- kXpGlenilai
- -- - - 97t:Wie'h a a *
- 84 smimum IJ Jme.il 1.-
s t s
.10-' -
N.s -
's s-
,s. .
g . 's
... _ .s, 8.-
_ 's s .
.w. \'
, : LA. - 'g -
* \
;3
\
s, e c ,10..: - s s
.s, s __
- ' =
~
. i_ s s
. P , '. o - %, ._
J. 5. -- N N
.5 p-s,'s s o- .
s,
. . . . . . ' ..............................h,
.=
10-3 - Ns
- \-
\
s
- \ _
\.
,i -
i i 1 410-' --
-l -
1
- 1
~
.4 l
L 104 ' ' ' ' ' ' I- - L 5 6 7 8 9 10 l -~ Magnitude \ Figure -2 ~ Effect of magnitude dietribution model for Expert 8 I g, 4
. k I' i
f
, i .,
GEOMATRIX -
.FICURE] B-value -
Deep Zone 200 ' l ' i '
- l. i j . . , ,
\ 3 --
100 -
. 7
; g= 0.10 50 b o.71 _
d a - O 20 - . 10 -- 7
- ~
5 : .
.2 - -
c e .1 ~. 3 ' tr - _- e g .5 - . 3 j .2 - .
.1 -- -
4
- O -
.05 - - .
' ~
.02 - .
0 (
.01 -- --
J :
.005 : .
.002 - .
' I . f , t , e , , , , , , ,
.001 I 2 3 4 5 6 7 a r
Magnitudo
q l D. GEOMATRIX - T
,y Expert 9
SUMMARY
OF MODEL PARAMETERS Slab-geometry:
-Approximately;10* dip through deep seismicity:
Convergence rate:
.42 110 mm/yr with'the following distribution percent: 0 _
10 12 0 30 40 50 prob- ~ mm/yr +----+----+----+----+----+----+----+----+----+----+ 0.05 30.00 ****** 0.30 3.5 00 ******************************* 10.30 40.00 ******************************* 0.30 . 45.00 ******************************* .. 0.05 50.00 ****** prob +----+----+----+----+----+----+-- -+----+----+----+ Sources and probability of activity: 3-Interface 0.92 (0.8-1.0). Intraslab shallow 0.9 Intraslab deep 1.0 Maximum extent of-rupture on interface: up/down dip extent - 124.7*W to 122.7*W (0.8) 124.7*W to 122.0*W (0.2) Along strike - Entire Zone (0.2) Nootka to Blanco (0.6) Nootka to Blanco segmented at 46*N (0.2) Interface maximum magnitude: Use maximum rupture area Entire zone - rupture area 178800 to 241200 - M y 9.5 Nootka - Blanco 119200 to 160800 - M 9.25 Nootka - Blanco segmt at 46*N 59600 to 80400 - M" 9.0 Interface earthquake recurrence:
~
Use geological estimate of 430 yrs (t257.) Use characteristic magnitude distribution model Figure 1 shows the resulting distribution of recurrence estimates. The carthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: 7
-- -~ y , y, , -v.-. ,---q
. ._ ~ - ! .l t X ' GEOMATRIX Return' Period (yrs) for events of Magnitude 1 M Magnitude M Sth percentile -50th percentile 95 percentile 5 4 3 2. 6- ~17 11 7 7 77 53 32 8 256 427 132 9 600 427 330 Location of intraslab events: Match observed seismicity pattern up/down dip and along strike Intraslab Maximum Magnitude: Shallow zone use following. distribution percent: 0 20 40 60 80 100 prob Hmax +----+----+----+----+----+----+----+----+----+----+ 0.10 6.50 ****** 0.25 7.00 ************** 0.55 7.50.***************************** 0.10 8.00 ****** prob +----+----+----+----+----+----+----+----+----+----+ Deep zone use following distribution percent: 0 20 40 60 80 100 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0.10 7.00 ****** 0.80 7.50 ***************************************** 0.10 8.00 ****** prob- Mmax +----+----+----+----+----+----+----+----+----+----+ Intraslab earthquake recurrence: Use historical seismicity. Figure 2 shows recurrence relationship for deep zone. This curve is bases on all recorded events not inferred to lie within the North American plate. Figure 3 shows the recurrence relationship for the shallow zone based on the offshore recordings within the Juan de Fuca plate. E
g, , - .. ., -
", t ,4
^ ' ' , ; c. .
=
=r I g g
61 ..;-,
. seOMATRW ::
4
-10: .,
7.y . f 1 _'
,w- j _.
;i ' ,
95*/, . 10-5 --
- eI- : 5*4,
.'.,?r . .
s
. iu _ .
o.
.3 E 10-2 g ..
o . - a - n
. g. %
, s.
o - 10-3 - 10 : l-- . l . 10-6 ' ' ' ' ' ' ' I -- f' 5 6 7 8 9 to Mt jnitude \
. Figure 1 Estimated recurrence dietribution for Expert 0 l'
y, , 1,, ,, . _ , p: ' ' P s I GEOMATRIX o, _ [" FIGURE R- B-value - u dI+ c[ic ~l h h4 f, kC ' Deep Zone l 200 , l ,
- i. .. . 100. r -
- g- 0.10 -
50 - b - 0.71 -
~
lt G 20 -- - 10 -- - 5 - 2 - .
' i?
-g 1 r 2 .
o-
$ .5 :
} -
o . - 2 C 4 .2 - .
.1 --
() :
.05 :
.02 - -
O b
.01 -- ._
? )
.005 - :
.002 - .
.001 'I ' I ' I ' I ' I ' l ' l '
l- 0 1 2 3 4 5 6 7 8 i l- Magnitudo L
- -J.. , .. .,, - - . . - . - - , ,.-..-, ,,
GEOMATAlX FIGURE' 3 B-value - Off Shore { 3(,lfg isd fa t 200 , 100 -- ' 50 - b - 0.71 - 20 - 10 r --
- r. -
5 -
- 2 -
g- 1 - i o cr a- O
$ '*0
- l o -
O 3 . C
- j. g .2 -
0 0 0 0 -
.1 :
' ~ i l
- O e ;
.05 -
1 ,. .02 - L '. .01 r -- l' .005 - i - 002 -
.001 ' ' ' ' ' ' ' ' ' ' ' ' '
O 1 2 3 4 5 6 7 8 Magnitudo 4 _ , y
ceoNsnTno< l 1 Expert 10
SUMMARY
OF MODEL PARAMETPRS Slab Reometry: Approximately 10' dip through deep seismicity Convergence rate: 50 mm/yr with the following distribution percent 0 10 20 30 40 50 prob mm/yr+----+----+----+----+----+----+----+----+----+----+ 0.25 35.00 ************************** 0.25 40.00 ************************** 0.25 45.00 ************************** 0.25 50.00 ************************** prob e +----+----+----+----+----+----+----+----+----+----+ S,ources and probability of activity: Interface 0.7 (0.6-0.9) Intraslab 1.0
. Maximum extent of rupture on interface:
up dip - 125'W down dip - 30 to 40 km depth Along strike - Nootka to Blanco Interface maximum magnitude: Maximum rupture area - weight 0.5 rupture crea 116000 - M M 1 Age versus convergence rate yweight 0.5 Mw 8.3 Interface earthquake recurrence: Use moment rate approach moment rate = convergence rate *a* interface area Convergence rate specified above a specified by following distribution i l l I
1 a GEOMATRIX
<<J
?
.g;. percent: 0 5 .
- 10. 15 -20 prob a 0.006250
+----+----+----+----+----+----+----+----+----+--
0.05 ** 25
-+
' 0.050000 0.10 ***********
? 0.100000 0.15 ********************* 0.150000~ 0.20-******************************* 0.200000 0.25 *****************************************.
- 0.233330 0.30.************************************************
' 0.166670 0.35 **********************************
O.083330 'O.40 ****************** 0.010420 0.~45 *** prob
.+----+----+----+----+----+----+----+----+----+----+
Use exponential magnitude distribution model Figure 1 shows_the resulting distribution of recurrence estimates and Figure 2~shows't.he effect of maximum magnitude on earthquake recurrence. The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: Magnitude M
. Return Period (yrs) for events of Magnitude h M Sth percentile 50th percentile 95 percentile 5 4 1 0.2 6 17 5 1
- 7. 88 27 6 8 500 170 61 9 -
1930 Location of intraslab events: Large events-occur in depth range 30 to 60 km Along strike - 0.5 weight to observed seismicity pattern 0.5 weight to uniform distribution Intraslab Maximum Magnitude: 7.25 10.25 Intraslab earthquake recurrence: Use historical seismicity. Figure 3 shows recurrence relationship for deep zone. .This curve is bases on all recorded events not inferred to lie within the North American plate. E
,r ,
.g-- ..
y ,
'c' w,- .- -
*av:
_ _ ./ r_.* ,
. , _ _ ,.c.
,l ','.':<9 ' '
- g. j t_ -- 'i , " C- c aammx -
2 .2 - _ ,
, + .
_4
. s x . , -.
j4.. og .
;10' . .
, ,. y.. ,--
g , _ 1- ; .- ., } '[:
'~~
* - 1
- 4 - s l:.
" R. .
1- 7 .. - w* w s e 1 .
- 10-1 r. --
; 1
.g,
. n 9
3, ,
. g .
e-
- u. -
- f. .
3
.a
.. E 10-2 _. ,_
m: < - o - 10-3 -
\
- - \
\ _
10 " r .
- \
\
l I' 10-8 ' ' ' ' ' ' ' - S 6 7 8 9 10 i- Magnitude 1 l Figure 1 Estirnated recurrence distribution for Expert 10 p . I b
,4
i r
,' m~; .
.uw .
v ., . 1 .., _I,.. -
- s. - '/>-.
- j.. t , , .., ,
- '. ~t .\
. Ri me
; x z
-r m
,j @
a a: Hto . -
# .. 3 :r. 1 .- - y. c- g - _-
c ., o _ - ..
-. a.
.,- - 663-8.2s c +QL unw- u-
' 'a e
i 1 r
~ . _
.s cs,. _
b i g
, g -
. 110-5 - 's,.s .
;A .:-. . s
.s 3 .
s sg s .. 1 _9
,3 . ,.
~
j . 6 %
'k .
g. J ,
. g 10-2 _ . s, c ,
s
'ss i _ ~W< .
',o-
> s-
-sp' *
- q- ._ f' ~
_O 3- .s.5 e r-o - s, r..- g s x 10-3 . . s y -
* .g
- , ..g
_ s
- s s, ,
~
ep 1 i, 1
- s.
i, . - i
. g-t
. i 4
. s: I e
, g g.4 . l . 1 . I' I !
5 6 7 8 ~ 9 , 10 Magnitude Figure 2 Effect of Maxirnurn rnagnitude for Expert 10 j' : at
s 1
- f Pt GEOMATRIX FIGURE.3. B-value'-
/
Deep Zone i '(
- 200 . , .' , ,
100 r -- 3 g- 0.10 50 - b = 0.71 - Cl O . O 20 - -
.10 :-
- S - -
2 . . I? 5 1 r a tr t{ .5 - - o . . 3 C 4 .2 - -
.1 . .
-O :.
.05 - -
-(
.02 - -
0 (
.01 - .
- ) ~
.005 - -
.002 - -
' I ' I ' I ' I ' I ' I ' I '
.001 O 1 2 3 4 5 .6 7 8 -
Magnitudo 4
h il GEOMATAIX g
- '4 4
l' - Expert 11
SUMMARY
OF MODEL PARAMETERS Slab geometry: g Approximately 10' dip through deep seismicity.
#1 Convergence rate:
42 110 mm/yr with the following distribution percent: 0 . 10 20 30 40 50 prob mm/yr +----+----+----4----+----+----+----+----+----+----+ 0.05 30.00 ****** . 0.30 ~35.00 ******************************* 0.30 '40.00 ******************************* R 0.30 '45.00 ******************************* ; 0.05 50.00 ****** prob +----+----+----+----+----+----+----+----+----+----+ Sources sid-probability of activity: Interface 0.9 (0.8-1.0) Intraslab 1.0
~
Maximum extent of rupture on interface: up dip extent - top of continential slope
.down dip extent - 123'W Along strike - Entire Zone (0.09)
Nootka to Blanco (0.91) Interface maximum magnitude: Use maximum rupture area Entire zone - rupture area 122400 - M 9.25 Nootka - Blanco rupture area 81600 - M" 9 Interface earthquake recurrence: Use geological estimate of 430 yrs (i257.) Use characteristic magnitude distribution model Figure i shows the resulting distribution of recurrence estimates. The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: Return Period (yrs) for events of Magnitude 1 M Magnitude M Sth percentile 50th percentile 95 percentile 5 5 4 3 6 19 15 11 7 85 68 51 8 284 227 177 9 670 533 400
'pt- .' : 'j I '
i- - g h GEOMATAIX it
- 'c0' '-
Location ~of intraslab events: i Hatch observed seismicity pattern up/down dip and alcag strike. -Large events in depth range 40 to 70 km. Intraslab Maximum Magnitude: Use following distribution
. percent: ,0 10 20 30 40 50
. prob Mmax +----+----+----+----+----+----+----+----+----+----+
0.30 7.00 ******************************* 0.33 7.25 ********************************** 0.37 7.50 ************************************** prob +----+----+----+----+----+----+----+----+----+----+ Intraslab earthquake recurrence: Use historical seismicity. Figure 2 shows recurrence relationship for deep zone. This curve is bases on all recorded events not inferred to lie within the North American plate. t 6
,~ - , ~ - - , - , - - - , , - -
,s. ., , . - - - - - _ - - - - _
It' g, - ' I
# * ' ' r; , . - +. , -
3..1 l,f.??. p, _ i
~
aeOMATS4m - rt
" i' .
'10.
.* ic >
-i.
i i .
._n.
- .i n.* q-
}
I < / 4 t-
- P d-2 . ;1: f ._.
4 *
~
I 10-8 - w -
.g -
ai _
- 5. _
. v-u.; _
g
.3
. E 10-2 _.
.] \ _
i : N-p u - 10 ~3 r _ s
- 4 10-' _-
r - _ (: 10-6
- I ' ' ' ' ' I I i-5 6 7 8 9 -
to Magnitude P P Fiaure 1 Estirnated recurrence distribution for Expert 1. l .\ L jA.
GEOMATMIX /
-FIGURE 1 B-value .
Deep Zone 3-20 200 ,
,100 r --
- g- 0.10 ;
50 - - b - 0.71 - C2 O . G
'20 - -
10 :- ;
- 5 . -
2 -
.y-g- 1 r 1
a, . .- u b.
.5 - -
o . C
.2 4
.1 . .
() :
.05 - -
.02 - -
0 (
.01
- ) :
.005 - -
~
.002 - -
.001 - ' ' '
' I ' ' ' I ' ' ' ' '
O 1 2 3 4 5 6 7 8 Magnitudo *
+:
j! GEOMATAIX a Expert 12
SUMMARY
OF MODEL PARAMETERS l' Slab geometry: Model A - approximately 10* dip through deep seismicity (0.80)- Model B - approximately 10' dip with reverse bend (0.20) 1
-Convergence rate:
25 to 40 mm/yr with the following distribution percent: 0 1C 20 30 40 50 prob mm/yr+----+----+----+----+----+----+----+----+----+----+ 0.200000 25.00 ********************* 0,300000 30.00 ******************************* 0.300000 35.00 ******************************* 0.200000 40.00 ********************* prob _+ ----+----+----+----+----+----+----+----+----+----+ l Sources and probability of activity: Interface 0.2 (0.1-0.4) Intraslab 0.95-1.0 Maximum extent of rupture on interface: Updip - eastern extent of underplating - 15 to 20 km depth (equal L weight) l Downdip - 50 km depth Along strike - Nootka to Blanco (0.7) Nootka to Blanco segmented at tip of Vancouver I (0.3) l l-t Interface maximum magnitude: Not assessed - use AGGREGATE distribution a) 0.55 weight assigned to estimate from maximum rupture area Model A - unsegmented area 106400 to 124800 - M 9.25 , segmented area 86450 to 101400 - H" 9 to 9.25 l Model B - unsegmented area 138400 to 161600 - H" 9.25 l segmented area 112450 to 131300 - H" 9.25 l b) 0.45 weight assigned to following distribution percent: 0 10 20 30 40 50 prob Hmax +----+----+----+----+----+----+----+----+----+----+ 0.073333 7.50 ******** 0.073333 7.75 ******** 0.073333 8.00 ******** 0.390000 8.25 **************************************** . 0.098333 8.50 *********** 0.083333 8.75 ********* 0.083333 9.00 ********* 0.083333 9.25 ********* 0.041667 9.50 *****
+----+----+----+----+----+----+----+----+----+----+
prob
GEOMATFHX Combining a) and b) the distribution over all maximum rupture geometries is percent: 0 20 40 60 80 100 prob Hmax +----+----+----+----+----+----+----+----+----+----+ 0.033000 7.50 *** 0.033000 7.75 *** 0.033000 8.00 *** 0.175500 8.25 ********** 0.044250 8.50 *** 0.037500 8.75 *** 0.103500 9.00 ****** 0.521500 9.25 *************************** prob Hmax +----+----+----+----+----+----+----+----+----+----+ Interface ear?.hquake recurrence: Use moment rate approach moment rate = convergence rate
- a
- interface area Assessed distribution for convergence rate mm/yr is give above assessed distribution for alpha percent: 0 20 40 60 80 100 prob o +----+----+----+----+----+----+----+----+----+----+
0.100000 0.05 ****** 0.600000 0.10 ******************************* 0.127550 0.15 ******* 0.048980 0.20 *** 0.040820 0.25 *** 0.032650 0.30 *** 0.024490 0.35 ** 0.016330 0.40 ** 0.008160 0.45
- 0.001020 0.50
- prob +----+----+----+----+----+----+----+----+----+----+
Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) characteristic (0.41) maximum moment (0.36) Figure 1 shows the resulting distribution of recurrence estimates, Fir;ure 2 shows the effect of variation in maximum magnitude on recurren'ce estimated using the cament rate approach, Figure 3 shows the effect of choice e,f magnitude distribution model on recurrence estimates, and Figure 4 shows the effect of slab geometry on the recurrence estimates. The carthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows:
l l GEOMATAIX l l Return Period (yrs) for events of Magnitude 1 M Magnitude M Sth percentile 50th percentile 95 percentile 5 10000 42 1 6 10000 148 5 7 10000 427 17 8 - 1950 170 9 - 10500 3160 Location of intraslab events: Updip, downdip extent, and along strike 0.95 weight - match observed seismicity pattern 0.05 weight - uniform Intraslab Maximum Magnitude: Not assessed - use AGGREGATE assessment of other experts percent: 0 10 20 30 40 50 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0.041 6.50 ***** 0.030 6.75 **** 0.355 7.00 **********************6************* 0.219 7.25 *********************** 0.345 7.50 ************************************ 0.000 7.75
- 0.010 8.00 **
prob +----+----+----+----+----+----+----+----+----+----+ Intraslab carthquake recurrence: Use historical seismicity. Figure 5 shows recurrence relationship for deep earthquakes assumed to be occurring within downgoing slab. This curve is bases on all recorded events not inferred to lie within the North American plate. l L_ _ _ -_ __-____- _____-__- _ - - - - - - - - - - - - - - - - - -
GUOMATRIX . 10 _ , , , r-- -
, T- -
1 q 10-i - s - 8,
.Er
\
u \
$ 10-2 _-
\ ,
)
's E to N g j a
c
~
\ -
10-3 i \ y i NN 10-' 7 5% \,') -
\
\
~
\
\
~
\
s g g-6 . , I l , l _ 5 6 7 8 9 10 Macnitude Figure 1 Estirnated recurrence dietribution for Expert 12
;;;,,:e> n> s -
. , n. . _ . '
; D ' ' ' - ~ ' '
,.(y;
, r -
13a ,.
1 9.... ,
M
' i '
- py
,3 s .-
Y_" osomamw s e-c . rg7 < - t- ,.
- ', ~
,_ ~
-_ . s
.s l- ,
'1 , -
.j o _
l _ . ,
. , ,, - y ,
, .q ,
*.. J --
g 1 c
. uu Ruptwo k.m -
- w
----+ n 7.5
-.g s.
? -
1' - -Q.g.
~
m
- s, s. -
s; . -
. s,,g
- y. .
.% g s.%'- .
\
's 10-1 -~ :
's -s .
s. gg ' '
=
. *. % =
l u.- - - ,e _ JO j
'. E 10-2 __
l _- _g . > I
.)O .
I-
.' 5
- g. -
l
.5 l o -
10-3 --
-i \
4. n s t
< 4s l, . -
g .
~ I t
t 10 " -- l N\ l s _ i
. l
. I
_ i t
\
t 1 %
. 1 I
i 10-6 ' ' ' ' ! ' ' ' - -
~5 6 7 8 9 10 Moonitude Figure 2 Effect of Max Magnitude for Expert 12
. s b, b -s.*
_ .m - -
I ( ;
'q, .
+ -
.g ,
asomamx - 10 .. 41
.I e 1 ,
g , t -
.. . { p g ]lg['
- - - - -
- paws.e4Uc
* *
- tJodmwn Houe.d :- .
.c 1- -
P 10-8 -- --
- 'i .
sg. . s, ~
. 5,
^
, . s- .
1 " .e s,%
- u. .. .s,.
- s. .
.3 s
's, 5 10-2 _.- _
- s. .
9, .
.g .
.s
- : . *s
*3g
.o -
5, .
.g 10-3 --
.s'..
A
'q\'
jo-4 -
. v
\.s i s b
\
- 4 4
104 ' ' ' ' l 8 - 5 6 7 8 9 to Magnitudh Figure 3 Effect of magnitude distribution model for Expert 1.?
-357 p; ,
,I.
, f,
- E..- -4 -
>- t s s > '.,x 1.
r'
,k-4
,4 t,
{.
- v. ;
i
~ '
,'. N '4
, . , g ,
V s r' -e,, . _ . ~ GELOMATRIX -
- e, _ .
( 3
. , c>
.,. :10 __. . .. j. , _ , ., ,
y , _,- -,. p.
.r? -.Y. , , -.. ..--, ,: -
n' ,J'. ,
- t.
8.
' n.
, i s i ,
, t - Llod9t A
, _ ; p - - - - ~ O Met n y
g
'1 -
i
- a s
~
)/ '%
, g s s
, 10-' -
's .
s,,
,' F : %
s s
- s v; - ',
8-e
- 's,. >
4- %
- u. -
s% - , r o J g
- s .
. c 10-2 :- - 's _.
a. 4 - - s
.g -
- s. .
.O -
h\ ,. r ; . - t- f , c, . \, s s 3 107 . : -
's
- s 2-g ,
- s 4
\ i-.
- s, .)
10 -
- 's, .
r s, is is
% s
\s, 3
1 04 ' ' ' ' ' ' ' l- - 5- 6 7 8 9 , to Magnhde Figure 4 Effect of skib geometry for Expert 12
s , 4 l N GEOMATRIX r .: : - FIGURE B-value - Deep Zone 200 , i . . 4 1 100 :
- ? .
. AS - 0.10 :
50 - ~ b = 0.71 - (2 0 - G 20 - ~ 10 r ~ 5 - ~ r 2 - - D - _ 1
- c. .
a er : - u.
- 2. .5 - ~
o . - 3 C
- .4 .2 - .
t-
.1 : ?
- 0 -
.05 - ~
.02 - - -
O G i .01 .- 1
) :
l \ .005 - ~ \^ - -
.002 - ~
l-
, i i , t . t . I e I I .
.001 i 0 1 2 3 4 5 6 7 B-Magnitudo 4
4 < grv w -,,w- e ,, .,4, w~e-e - ,~- - -wq ~
'A-
l I , i GEOMATRIX - o
. Expert 13
SUMMARY
0F MODEL PARAMETERS l- Slab geometry: Approximately 10* dip throug!. deep sofsmicity Convergence ~ rate: Following distribution assessed percent: 0 10 20 30 ~ 40 50 prob mm/yr +----+ ---+----+----+----+----+----+----+----+----+ 0.05 20.00 ****** 0.30 30.00 ***********.*f;**************** 0.30 40.00 ******tM'********************* 0.30 'o. . -****************************** 0 05 S3.00 ****** prob +----+--+-+----+----+----+----+----+----+----+----+
- Sources and probability of activity:
Interface 0.05 (10.05) renormalized to 0.0075 for events > M 5.0 Intraslab 0.8 (t0.1) Maximum extent of rupture on interface: Updip - 15 to 20 km depth Downdip - 40 to 50 km depth Along strike - Nootka to Blanco Interface maximum magnitude: Assessed distribution percent: 0 20 40 60 80 100 prob Mmax +----+----+----+----+----+----+----+----+---~+----+ 0.50 4.00 ************************** , 0.30 5.00 **************** 0.15 6.00 ********* prob +----+----+----+----+----+----+----+----+----+----+ Renormalized distribution for events > mag 5 t percent: 0 20 40 60 80 100 prob Hmax +----+----+----+----+----+----+----+----+----+----+ 1.0 6.00 *************************************************** prob +----+----+----+----+----+----+----+----+----+ ---+ i ! Interface earthquake recurrence: Not assessed - use moment rate approach as Mmax to sma.'1 to produce geological evidence , moment rate = convergence rate
- a
- interface area
i 3 5M GEOM ATRIX 2 Assessed distribution for convergence ra re mm/yr is giue above alpha assessed as follows percent: 0 10 20 30 40 50 prob a +----+----+----+----+----+----+----+----+----+----+ 0.30 0.001 ******************************* 0.40 0.050 ***************************************** 0.30 0.100 ******************************* prob +----+----+----+----+----+----4----+----+----+----+ Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) characteristic (0.41) maximum moment (0.36) Figure 1 shows-the resulting distribution of recurrence estimates, Figure 2 l shows the effect of variation in magnitude distribution model on recurrence estimated using the moment rate approach. The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for events of various sizes as follows: l-l Return Period (yrs) for events of Magnitude t M Magnitude M Sth percentile 50th percentile 95 percentile 5 1 0.1 0.02 l 6 3 0.4 0.1 l 7 - - l Location of intraslab events: Use observed seismicity pattern, 90% of large events in 50 to 90 km depth range l Intraslab Maximum Magnitude: Assessed distribution as follows: l percent: 0 10 20 30 40 50 prob Hmax +----+----+----+----+----+----+----+----+----+----+ 0.45 6.50 ********************************************** 0.40 7.00 ***************************************** 0.14 7.50 *************** 0.01 8.00 ** prob +----+----+----+----+----+----+----+----+----+----+ Intraslab earthquake recorrence: Uae historical seismicity. Figure 3 shows recurrence relationship for es assumed to be occurring within downgoing slab. This deepearthquap/onallrecordedeventsnotinferredtoliewithinthe carve is base North American plate.
?; ,
O GeOMATRIX Assessed distribution for convergence rate mm/yr.is give above alpha assessed as follows percent: 0 10 20 30 40 50 prob a +----+----+----+----+----+----+----+----+----+~---+ 0.30 0.001 ***********************k******* 0.40 0.050 ***********************k***************** 0.30 0 . 10 0 * * * * * * * * * * * * * * * * * *
- A * * * : t * * * * * *
- prob +----+----+----+----+----+----+----+----+----+----+
Magnitude distribution model not assessed - use AGGREGATE assessment of the experts exponential (0.23) characteristic (0.41) maximum moment (0.36) r. Figure i shows the resulting distribut. ion of recurrence estientes, Figure 2 shows the effect of variation in magnitude distribution model on recurrence estimated using the moment rate approa.ch. The earthquake recurrence relationships shown in Figure 1 can be summarized in terms of return periods for ovents of various sizes as follows: Return Period (yrs) for events of Magnitude t M Magnitude M Sth percentile 50th percentile 95 percentile 5 1 0.1 0.02 6 3 0.4 0.1 7 - - - Location of intraslab events: Use observed snismicity pat. tern, 90% of large events in 50 to 90 km depth range
-Intraslab Maximum Magnitude:
Assessed dis'ribution as follows: percent: 0 10 20 30 40 50 prob Mmax +----+----+----+----+--- F----+----+----& ---+----+ 0.45 6.50 ********************************************** 0.40 7.00 ***************************************** 0.14 7.50 *************** 0.01 8. 00 ** prob +----+----+----+----+----+----+----+--- ' - -+----+ Intraslab earthquake recurrence: Use historical seismicity. Figure 3 shows recurrence relationshi,p for es assumed to be occurring within downgoing slab. This deepcarthquap/onallrecordedeventsnotinferredtoliewithinthe curve is base North American plate.
._. _ . ~ , . - _ _
a l D GEOMATAIX too _ , 3 r- - nrX to r - 1 p 4 a e : N ( L. -
?.3 5 10- -
E a s : 4 s ~ 5 o - 19-2 _. 10-3 r ~ m-* ' ' ' *---- 5 6 7 8 9 30 Magnitude Figure 1 Estirnated recurrence distribution for Expert 13
e GEOMATAIX 100 ; i
; i
; i
- 1
--- - C93nentioi
_g ---- Chorotteitik ,
\ -
* *
- thimum Llane.it N
10 - g _
- 's ~
-_ . . ; 3 ' 0, , ,{ .
a f . _ l
) .
1 : x : 8., _ 1 s
? ~ ~0 3
E 10-1 -- g _ o - -
!5 -
o - 10-2 _. ,_ E_ i 10-' r -
' ' I L 10-4 ' ' -
.a.
5 6 7 8 9 - 10 Magnitude Figure & Effect of magnitude distribunon model for Expert 13 l
+
, GEEOMATRIX '
t FICURE 3 -' B-value -
-Deep Zone-200 . , ,
i
.i 100 r-g- 0.10 50 - -
b = 0.71 - Q O .. G 20 - -
~0 1 --
5 - - 2- - - c - - 1 t- e a :. o- . s 8 . -
- u. -
o .
.a .
C 4 .2 - - l ) , i j .1 --
- O
,oS - -
l i
.02 - l -
f. O (
.01 r ;
- ) -
, .005 - - l
.002 - -
.001 O 1 2 3 4 5 6 7 8 .
Magnitudo y- - - - - .- - + _ - , 't &
D g GEOMATRIX r Expert 14
SUMMARY
OF MODEL PARAMETERS
-Slab Geometry:
Two dips: Model A -19* to 10* (0.35)
.Model B - 20 to 25* (0.65)
'Convernence rates-The a .assed distribution for convergence rate mm/yr is percent: 0 5 10 15 20 25 prob mm/yr +----+----+----+----+----+----+----+----+----+----+
0.012500 10.00 **** 0.025000 15.00 ****** 0.062500. 20.00 ************** 0.100000 25.00 ********************* 0.150000 30.00 ******************************* 0.200000 35.00 ***************************************** O.175000 40.00 ************************************ 0.150000 45.00 ******************************* 0.100000 50.00 ********************* 0.025000 55.00 ****** prob +----+----+----+----+----+----+----+----+----+----+
' Sources and probr.bility of activity:
Model A - Interface 0.35 (10.15) ! Intraslab 1.0 31 Model B - Shallow Interface 0.25 (10.15) Deep Interface 0.3 (10.1) Intraslab (Juan de Fuca) 0.i "deep zone" beneath Puget Sound 1.0 Maximum extent of rypture on interface: , Model A - Updip at 125' Downdip at depth of 50 km Along strike - Nootka t) Blanco (0.8) Nootka to Blanco segmented at 47'N (0.2) Model B - Updip arc 125* both interfaces Downdip at depth of 50 km both interfaces Along strike - Nootka to Blanco (0.8) Nootka to Blanco segmented at 47'N (0.2) ' shallow interface exists only north of 47'N 6 Interface Maximum Magnitude: l Shallow interface - 8 (10.25) Model A, 7.25 (10.25) Model B resulting L in following distribution l i o.- -s
u
- GEOMATRIX f'
percent: 0 10 .20 30 40 50 prob Hmax +----+----+----+----+----+----+----+-- -+--- v----+ 0.214500 7 . 0 0 * * * * * * * * * * * * ': * * * * * * * *
- 0.221000' 7.25 ***********************
0.291500 7.50 ****************************** 0.077000 7.75 ********* 0.077000 8.00 ********* 0.059500 8.25 ******* 0.059500 8.50 ******* prob +----+----+----+----+----+----+----+----+----+----+ Deep interface - 8 (10.25) Model B resulting in following distribution percent: 0_ 5 10 15 20 25 prob Mmax +----+----+----+----+----+----+----+----+----+----+ 0.220000 7.50 ********************************************* 0.220000 7.75 ********************************************* 0.220000 8.00 ********************************************* 0.170000 8.25 *********************************** 0.170000 8.50 *********************************** prob +----+----+----+----+----+----+----+----+----+----+ I:+-AI4ce Earthquake Recurrence: Use moment rate approach with: moment rate = convergence rate *a* interface area a assessed according to following distribution percent: 0 20 40 60 80 100 prob a +----+----+----+----+----4----+----+--E---+----+ 0.200000 0.05 *********** 0.600000 0.10 ******************************* l 0.200000 0.15 t********** prob 4----+----+----+----+___+...+....+....+.__.t....+ i Use "maximum metent" magnitude distribution I Attached Figure 1 shows the resulting distribution of recurrence estimates for interface events. Maximum event magnitude is assumed to be uniformly distributed in the range of the expected maximum . magnitude given above 1 0.25 magnitude units. Figure 2 shows the effect of choice of maximum magnitude on recurrence estimates and Figure 3 shows the differences between Model A and Model B. The recurrence ratimates shown in Figure 1 can be summarized, in terms , of return period for various size events as follows Return Period (yrs) fot events of Magnitude 1 M Magnitude M Sth percentile 50th percentile 95 percentile 5 260 33 8 260 33 8 7 260 34 8 8 - 427 60 ,
.1 J
l>
~
GEOMATWlX 3 Location of intraslab events: Model A - between 122*W and 124*W-Model B "deep zone" between 122*W and 124*W Along strike - match observed relative frequency
-for Model B "deep zone" exists only north of 47'N
'Intraslab Maximum Hannitude:
Model A and "deep zone" for Model B - 7.25 to 7.5 Model B - not assessed Intraslab Earthquake Recurrence: Historical seismicity used to'sompute a- and b-values for exponential model. For intraslab events in Model B the seismicity rate was estimated from offshore events w' thin the Juan de Fuca plate away from the spreading centers and fracturt zones. Figure 4 shows the recurrence relationship used for the intraslab events in Model A and the deep zone in Model B. This cutve is bases on all recorded events not inferred to lie within the Norta American plate. Figure 5 shows the recurrence relationship for the offshore Juan de Fuca plate used to model the intraslab recurrence for Model B. i l I i l i
r y D GEOMATRfX 10 , i i- ; ,, i--, j i 1 : J
*) 5 %
X 10-1 _-- X \ 8 -
\ 4 g -
so % \ l o c \ 3 3s \, \ 5 10-2 _- v - 3o e e- ',o \ \
\,
3 - E \ \ 3 \ o -
\
10-3 :- 1 \ t \
- \
l 10-' r - 10 4 ' ' ' ' ' ' ' ' 5 6 7 8 9 10 Mapitude Figure 1 Eeurnated recurrence distribution for E peri 14
\
) C GEOMATFtlX to y, 3 r , , 1
- i
_ -- u :.s 1455 1 1 J 1 : t o-' - X E., N U tt _ r, a c g 30 2 ...................... ,
. _ s
_ s - 4, c _ s 4 3 's l c - s n s s._. ~, 10-3 - \ -- _ s
- \
~ s
_ s _ i t 8 l - I 10 -' : - 4 - _ g 1 l
- 0 3
- i t I g g-6 i f i l I i i l * }
s 6 7 a 9 . to MagnRude naure 2 u.fect of Max IAco.n.tua. e ior t.xpert . 4 i i
to , . . , - M
'osom^mx 1 . I r -
1 p mdt.; . t uxa e i I 1 - 10-5 _- m _ N p ___,.__~___,_____,___ _ 3 s % 7 -
\
\
37
.sm -
,-s x 3 v a i \ l E 10-2
\\
s \ q s . w
! : s -
sa - s s '\, .
- s g -
s 5 ' y s
\
10-3 -
- i s
_ 6 t 4
\
\
'. \.
. 10-' t- .
_ s _ s g _3 , , , i . i . i. ... 5 3 7 8 9 10 f,{ 33ni!U d3 Figure 3 Effect of elab geornetry for Expert 14
GEOMATRtX
. s-av,. .. . .
+
FIGURE 'B-value - ' j'f > Deep Zone 2001 , g .. .
~
.100 7 --
- g- 0.10 -
- 50 .- -
b ,= 0.71 -
'(1 .O .
G 20 -- - O, 10 : z 5 - nr 2 - I 1 _
-S :
{ . t l I u
- ta.
.5 -
l
. .g j 3
C { .2 - l .1 - l
-. : O :
I - ! .05 - l . l- - r i f>
.02 -
O (
.01 l-
) -
,005 : :
.002 - -
.001 ' ' ' ' '
O 1 2 3 4 5 6 7 8 Magnitudo
i.- : a' _ CU l . ,. , , osomarsesx-- L is yi FIGURE B-value '..
- - . Off Shore' 200- ,
100- :- - ' 50 -- - b - 0.71'- - n-20
- 4-10 --- ' -
5 - 2 . c 1 cr (i. O
~
f 5L ,- l i . O C ' 4 .2 - O O O O - i
.1 r
}
- O'( ~
.05 -
.02 --
t i
.01 --
.005 - - -
t
.002 - -
t
.001
' ' ' I ' ' ' ' ' ' ' I ' I '
O 1 2 3 4 5 6 7 8 Magnitudo ' a
, -_ M
- o. ,
Q , GEOMATRIX : ,. r I t ,
. gs
,1P r o
, .:p i V
. AITACl9 TENT 2 pl L- ,
Assessments Made by Individual Experts' , 4 (These are the Phase I assessments included previously in this > a'ppendix.) s
.E, i I
. t
. 4 E
i 4 l t b L
?
r b f f
?
l p W i I L
, -, - .. _ .-._m._,_,,..__ _ _ ..m., ._,,;._,...__, , _ . _ . . . , . - . . . . . , , , _ _ . _ , , , _ __ _ . . . - - , _ _ _ - - .
T GEOMATAIX ATTACIDtDir 3 S m ry of Aggregate Expert Assessments a
,er----. -r-- , ,,-.
DEOMATRIX
SUMMARY
OF EXPERT ASSESSMENTS FOR SUBDUCTION ZONE SCISMIC SOURCES The logic tree format developed to model the subduction zone sources is shown in Figure 3-1. The logic tree progresses from an assessment of the geometry of the subducting slab to assessment of specific analysis para-meters'for individual sources. The assignment of parameter values and th61r relative likelihoods for the subduction zone sources was based on the inputs from 14 experts. The individual assessments of each of the experts are documented in Appendix A and are outlined in Table 3-1. The assessments-made for each component of the hazard model are summarized below. Crustal Geometry All of the experts provided an assessment of the cross sectional geometry of the subducting Juan de Fuca plate. Most of the experts provided only a single assessment consisting of the plate dipping at approximately 10* and extending through the zone of deeper earthquakes lying at depths of 30 km or more beneath the site. 'Ihree experts provides alternative geometeries consisting of a more steeply dipping interface with a possible remnant slab or secondary interface it. the vicinity of the observed deep seismicity. Two experts provided a slight modification of the 10* dip consisting of a flat lying slab with a double bend (see cross section for expert 6 in Appendix A as an example). Figure 3-2 presents the aggregate distributions for slab geometry. Seismic Sources All of the experts identified the Juan de Fuca - North American plate inter-face and the subducting Juan de Fuca plate as potential sources of thrust and intraslab normal eventa, respectively. Several experts also identified potential sources in the overlying North American plate. Evaluacion of the hazard from these crustal sources was included in the shallow crustal source model.
GEOMATRIX l Probchility of Source Activity All of the experts made an assessment of the probability that the plate interface and the subducting slab are seismogenic. Figure 3-3 shows the distribution of assessments of activity for the intraslab and interface sources. The assessments for the intraslab source are generally at or near unity based on the past record of seismicity. The assessments for the interface range from near zero to near 1.0 with a nearly uniform distribu-tion and an equal weighted average of 0.49 It should be noted that tha that an adjustment was made to the assessments of experts L. and 13 As indicated in Table 3-1, column 5, these two experts have probabilities of 0 9 and 0.85 that the maximum magnitudes for the interface is M, 5 or less. All other experts made the assessment of activity in terms of the probabil-ity of the interface being able to generate tectonically significant events (M,)S). To put the assessments of activity of experts 4 and 13 on a consistent basis they were adjusted to values of 0.075 and 0.0075. respectively, and their maximum magnitude distributions renormalized to include only magnitudes larger than M 5 All but two of the experts assigned a value of unity to the probability that the intraslab source, as represented by the deep zone of seismicity is active. Location of Ruptures The experts provided assessments on the limits of eartnquake ruptures, both along the length of the subduction zone as well as the up dip and down dip extent. Figure 3-4 provides histograms summarizing the responses obtained. Most experts considered the maximum limits of coherent rupture along the interface to be the boundary with the Explorer plate at the Nootka fault zone on the north and the boundary with the Gorda plate at the Blanco frac-ture zone on the south (see Figure 3-5) . Several experts considered further
~
segmentation of the interface to have some credibility, with a segment boundary generally in the vicinity of 46*N. The assessed minimum depth of rupture along the interface ranged from 5 to 25 km and the maximum depth of rupture ranged from 35 to 50 km.
r - , occ:: ,swn A majority of the' experts stated that they expect the future distribution of intraslab events to follow the observed pattern of historical seismicity
- with.the majority of events occurring generally beneath Puget Sound. Alter-natives considered included completely uniform seicnicity within the down-going slab or a concentration of larger events at deeper depths. Figure 3-4
-shows the aggregate distribution for seismicity distribution.
Maximum Magnitude The experts that assessed maximum magnitudes for the interface either made a direct assessment or specified that it be calculated from the maximum rup-ture area assessed above using the relationship proposed by Wyss (1979). The ten experts that provided an assessment of maximum magnitude for the interface were nearly er )nly split (0 55. 0.45: one expert using both methods) between the use of maximum rupture dimensions and a direct assess-ment of the maximum magnitude on the basis of analogy with other subduction zones or other techniques for magnitude estimation. The aggregate distri-bution shown in Figure 3-6 is for those who made a direct assessment, and is thus conditional on the direct assessment procedure being the correct procedure. If an expert did not assess interface maximum magnitude, then the marginal distribution used to represent the aggregated opinion of the other experts consists of 0 55 weight assigned to the magnitude value obtained from the experts assessment of maximum rupture dimensions and 0.45 weight assigned to the conditional distribution bases on direct assessment. The distribution shown at the top of Figure 3-6 has a large probability of 0 38 assigned to a maximum magnitude of 6. As this represents, the judge-ments of two of the experts based on specific reasoning, it is an appropri-ate distribution for use in component level aggregation. However, it was judged that this assessment is significantly lower than would be obtained from a general population of scientists familiar with subduction zone earth-quakes ar.d those experts who did not make any assessment of maximum magni-tude for the interface would, nevertheless, be likely to assign a much lower probability to a maximum magnitude of 6. Accordingly, the conditional distribution used for those experts who did not assess maximum magnitude was modified from that shown at the top of the Figure 3-6 by removing the
I vn GEOMATRIX
-4
' assessments ' for very low magnitudes and renormalizing. The resulting distribution is shown in the middle of Figure' 3-6.
The maximum magnitu6e for the intraslab source was assessed by 11 experts on the basis of historical seismicity and analogy with other subduction zones. The aggregated distribution is shown at ti.e bottom of Figure 3-6. Earthquake Recurrence Method All experts who made an assessment of earthquake recurrence preferred to use historical seismicity data to define the recurrence parameters for intraslab events. These parameters were used for all experts. Recurrence estimates for the plate interface were assessed either on the basis of a moment rate approach or on the basis of geologic evidence for the frequency of large events. In aggregate, the experts favor the moment rate approach slightly more than the presently available geologic data by the ratio 0 58 to 0.42. If an expert did not make an assessment of earthquake recurronce for the interface, then both methods were used with the given weights. Geologic Recurrence Rate Five of the e.perts chose to base the recurrence estimates for interface events solely or partially on geologic evidence for possible paleoseismic events, primarily the data on turbidites. Figure 3-7 presents the aggre-gated distributions for return period of large interface events. The distributions are tightly clustered about the estimate of 430 years given by Adams (1985). Convergence Rate All of the experts made an assessment of convergence rate with cost basing the assessment on the rate estimates published by Riddihough (1984) and Nishimura and others (1984). Those experts that made a direct assessment generally gave a wide distribution of values with a mean value somewhat ' l lower than the published estimates. Figure 3-8 shows the aggregate distribution for convergence rate estimates. i
GEOMATAtX Seismic Coupling Figure 3-8 shows the aggregate distributions of the amount of seismic coup-ling between the Juan de Fuca' and North American plates. Most of the experts gave a wide distribution for the amount of coupling with expert 1 giving a zero/one bimodal distribution.
- The combination of the plate interface area, the convergence rate and the amount of seismic coupling provide the rate of release of seismic moment.
For an interface length of 800 km, an average width of 150 kna. a conver-gence rate of cm/yr and an aggregate mean of 0.4 for seismic coupling gives a moment rate of 5 76 x 1026 dyne-cm/yr. Assuming all of the moment is , released in magnitude 9 events, a moment rate estimate of approximately 700 years would be obtained for the return period of these events. Recurrence Model . Three recurrence models for the form of the magnitude distribution were used ( for interface events in the analysis: the truncated exponential distribu-tion, the characteristic magnitude distribution, and the maximum moment distribution. Figure 3-11 illustrates the cumulative form of these three distributions and compares how they would estimate the frequency of smaller earthquakes when the absolute level of seismicity is fixed by the frequency ; l of the largest events. The aggregated distributions of the experts yielded probabilities of 0.23, 0.41, and 0 36 for the exponential, characteristic, t and maximum moment models, respectively. i l l
tu*
- M Oceanie Saab Ge_ . , #Dep) Posentief probahmety of Seesang Soestees Men Aeelv6ty
#t Top of deep seismicity (0.2) gg
' intra-stet (el la 11 15' eip (o si 7 75 (1025) intwtue (6) o.4 (o 25 - 0 5) fel 25' eip to 3) 06m nsuns i
*oeep events * (vor 15* a 75' eips)Ici - t o (ci 7.25 (to.25) I Deep slet (for 15' ai 25' dips) (41 0.1 - c.15 Idl 5 - 6 141
#2 Top of deep seism: city
' Intre-slet le) 0 6 - 0.5 (el -
intersue 161 ea(0.7- oest(61
#3 Top of deep sessawcety natre-sien tel 1411 e lanes., pe intertue tel o.s (o.5 - o.as tel 7-in 14.ep.
#4 s( in) Int Top of deep see edcity intra-slet (el SA lel 7 - 7 1/t h natweece tbl c15 ( 0 25) lol 3 m3) ?
4 (o.3) 5 m3) s (o.es , 7 (OAll
#5 4
Top of deep seismietty (0.2 - c25) entre-ssee tal tA lol 7 tel 17' - 20' eip (t?5 - c.4) intertu a It) e.5 (10.5) 161 84 Top of deep : MMr. single bend (0.7) betre-seat tel 1A fel 6-3/4-71/4 Interface it) 4.65 (10.151 tel Top of deep seismicity, soutee bend (0.3) Dimensas.s 16: e
#7 Top of deep se6senicJty g0.7) entra-stet fel S5(102)th) Dimensieres It poete saterface tel
#$ Top of deep seismacity entre-sleb (el 1A lol Dernenetens it; interface it! O.4 - O S (0.25 - 010 (b)
#9 Top of seen seismicity entre-sset to Sota septe. lel eJ (el s.5 (e.t l intra-sset to - 75em Inl int ,ue ic]
1A (61 jd,Q t,
.. - 14 75) . . . 1) .
Strhe-sep faults in upper plate (el g"0 8 *} t8 7A M Accretionary wedge faults (el 1.0 14] I II Tews in 1 . ; --; seen til 1A tel oimensient (c) 7-t/2141 75 mal g, 40 52)
#to Yes et seep seismicity intvowat (s? t.o 1 1 7-t/4 (a t/4](el aniertue tol o.7 to s - e s)In) oimenssens (0.51 8.3 (o.sl
#1i Top of deep seism 6 city intre-ste6 (el 1 A [e]
7 t/2 (el antertue (b) 0.s to s - 1.0) It) Dimensaens 161 Deep crustet sourte Icl 14 lc) 7-1/2(c)
# 12 Top of eeep se6sm6 city, single bend to8) Intre-steb fel CAS - 1.0 lel Top or seen sensadca ry, souble bend (0.7) v7-1/4[el Interface tal 02(0.1-odiin) ela Top or seep seismicity intse-sien tel e s (to.1) lel s 5 (o.45) '
anentece (b) 0 05 (10 05) Itil , le)
.N. $.'3.3
)
4 to si I 5 (0.35n tal 6 $15) e14 Top et eeep seismac4ey n35) entre-stem (snenew e6e moden let iA tal 7.25 - 7.5 tel steeper ein (o.65) interface (shonow eie enouen in) s.35 (scits) tal a tro.5) tal Shenow Interface (steep die moeel)Icj 0 25 (to 85)Icl 7 25 (t(L25) ici Deep enserface (steep dip enodet) 141 33 (to t)(di 7.5 - 8 to 75a Remnant sieb (steep esp modet) le] 18 (el 8-85 M , oeep intre-stem (sies, e., modes)It] t e it! 7.25 - 74 tel Acceeuenery wedge Igl o 7 (to 11(91 7 tto 25)(si Teees in simb th) 005 (t o os) (n) 5 tal
.m
OF WIPERT IffTERVIEWS -- .. muss e: . _ ___ : -- powde _ . neee launtres sessenc securreno necue'ence haeehed _c.upsene top ce.sesse mecurrence _ is.4ee ta se tertheuates M el 38 (* tel - Ma'*' ices sensur ta ci <o to s - e es) tapene eies (e el g - woment rete 101 il <s tas . 043) . Mn moment given e = 8 (b) 15 (41) Coelogic (0.3 c: ca,e)(el .. to - 25 so.a) <430 (b) 30 tal) e et d as tt199 Meterical sessa**v (el e.as peonentw to) poet C3 al 400 (200 - 2.0001 lo) O*oto08c data (bl Cherectoristic [bl 19 10461 Wstorical '-tery (a) 20 10.5) e.0C 10 = E15) ** - 30 10 4) gg de (8 06) 40 (25 - col Historical seismicity (al == f.apenential(al - lol 10 = 20 to.ed) testorncei seismicity (a)
- 0.0 (1415l Empcnencellal 430 (61
. gg Coologic date (0.5) (b)
Decnontiel 10 %) (b) de - 43 te.1) neomont rose tas)It) 43 - 50 10 4) Cherectertess sta) In! 50 - 80 to 02) 42 (t10) testor6 cal solas. deity tal 1(l.Op Empos entsoi (ej - 40 (110 = 15%) ~~ 825 (42 . ES) -- -- I 42 (tiel Mstorical sets. M, Id.te.f] 4.710.5 - IA) taponential Id.de.f] i) 430 (125%) (c) Chare:teristic [c) Coolog6c data (cl 3 r a 35 - 50 (t19) Motoreces sets M, tal eJ to - e.5) Emponential is.bl -- ggg Moment rete Ibl 42 (110) pastor 6 cal sets;.M r (a.cl == Esponeettet !a cl 1300lb) Geo4ogic dote (bl Characteristic (b) [150 (cl 30 = 35 (25 - 40) lestorical se!s.ev (el El 10.05
- 0.5) Esponentief (al -
omen. e i.i o (tin eesio,ica. seis , iai
. 0i na0si !
s.-owei i.: .. TI i APERTURE l
- i. . ) Ms.o,tca, s ce, is. ,i ai o0 0s> E - w(wi CARD L
$'N 4* - 0 mi h uoment voie (e.cAel Maa moment !bAtel p
id
>50 to 05) Also A, ,wardI on j Apertine Card g?O bO'tOD M Ok =
4
h GEOMATRIX EXPl.ANATION TO ACCOMPANY TABLE 3-1 Table 3-1 summarizes the responses given by the fourteen experts, 9hich are further detailed in Appendix A. A more complete discussion of the components of the seismic hazard model is given in Section 2.1. Each of the columns in Table 3-1 is explained below. Note that blank columns or cpparent omissions in the table are the result of the expert declining to characterize these aspects. Oceanic Slab Geometry Each of the experts developed a cross-sectional sketch of the geometry of the oceanic slab beneath western Washington. These sketches are included in Appendix A and described verbally in Table 3-1. Alternative models are given along with the relative weight assigned to each, expressed as probabilities summing to unity. Potential Seismic Sources The subduction-related potential sources of earthquakes are identified and cach is assigned a letter, which is'shown in brackets (e.g., "[a]"). These letters are used in subsequent columns to specify which seismic source is being described. Probability of Activity Probabilities of activity are given for each potential seismic source, cpecified by a letter in brackets. Where expressed by the experts, ranges of estimates are given in parentheses. Activity" is used here to signify capable of generating tectonically significant earthquakes (see Section 2.1). 3 Maximum Magnitude Direct assessments of the maximum earthquake magnitude are given for the cources specified in brackets. In some cases, a range of values is given, or a best estimate and uncertainty bounds, or discrete values with relative weights assigned to each value. Where the word "Dimensions" appears, the
- i<
'""^""
EXPLANATION TO ACCOMPANY TABLE 3-1 (cont'd) cxpert indicates that the rupture dimensions th:t he specified.be used to eniculate a magnitude .(i.e., he did not provide a maximum magnitude esti-mate directly). See Section 2.1 regarding "location of rupture" to see how the rupture dimensions were estimated. Convergence Rate The relative rate of convergence measured parallel to the convergence direction between the North American and Juan de Fuca plates is given in tillimeters per year. In some cases, ranges are given or discrete values cre given with associated relative weights. Recurrence Method The manner in which the experts desired to have the earthquake recurrence rate specified is given in this column. Examples include recurrence based on the historical seismicity record, geologic data for recurrence inter-vals, or seismic moment rate. The seismic moment rate approach (described in Section 2.1) utilizes the estimates of convergence rate and' seismic coupling. Seismic Coupling (a) Seismic coupling is the percentage of the total convergence rate that is cxpressed seismically. Th2refore, if the coupling is very high (a = 1.0), then all of the convergence rate will be expressed as earthquaker (i.e., tihe seismic moment rate from seismicity will be equal to that based on convergence rate). An a = 0 means that convergence is occurring cseismically (i.e., there is no seismic coupling). Recurrence hodel The recurrence distribution function is specified in this column. Models requested by the experts include an exponential magnitude 41% ribution (i.e., log N = a-bM) a characteristic magnitude distributien (Youngs and Coppersmith, 1985); and a maximum moment model (Wesnousky, 1983).
p,.,...,.i . . . . . - . _ . . .. - IT
- A EXPLANATION TO ACCOMPANY TABLE 3-1 (cont'd)
Geologic Recurrence for Large Earthquakes
' For those cases where geologic data provide a basis for estimating recurrence,:an. estimate of recurrence intervals for large earthquakes is given. 'These recurrence intervals were generally judged appropriate for magnitudes at or near the maximum.
\
e t c d 6 o n d 6 t n a0inX 6 i e ! e e 4 e n r e m t n i r m dug e t o e o o t n o c a W n t c W in p r o p o g x a x x a x o E h a a W E h Wg C M
/
f [Cf c 2 3 4 A i
/
im's e 0 0 0 N S c [ r
/y
/ /g r
/y g r y r y
g 0 0 0
#i o lot e o e
m m m m m m 0 3
/
0
/
4 0 5
/
GR 0 2 0 3 0 4 r 1 1 1 o e
/ t
[f c c n n e i g e o r m lo w o e c W G l e e R d o M n d 5 5 r s. 7 9 T 8 8 a z M a W H s e e c r w r t u e o u nt p e id r r o t u w a S ExR n f o e n o Z n d n o n t e o 6 e n i t c m e t n g e m g c e s u e m g n S d e U b S u S _ 1 e - f e 3 o iv _ iv t e t c c e - o t a A n r u I - t S g b
/T e i
F s la c e S a r c r f e _ - r h - u o e t o tr t o S n in I y A B
- t r l e
l e b e d d ta m o o S o M M e G
- > - nygsy~- v -
+
;e '. -
4 -= t e (11EOMATmtX e,
.i a . ,
r Percent Oceanic Slab 0 20 ~40 60 80 Geometry-
+----+----+----+--.-+.---+--.-+....+....+....+..100- .+
10' double bend *** 10 ***************************************** - 15* ****** 25' ****
+-...+....+....+....+....+....+....+....+....+....+
Distribution of 14 experts h t i r
.i I
Figure 3-2 Aggregate Distribution for Slab Geometry .' r F P f b r I i 4 , i I I i r I ! k i [ t I
- l. . . _ . _ . - ,, .. _ . _ = - _ - - . . - - .. - - - - - - - - - - - - -- ---- - = -- ~'
up , - g.
. > t,- _
r.
,,ei -
,3
's v
(BEOMAT5tlX g Percent' ! 0. 5 10' 15 20 25 Activity +---~+----+----+----+----+----+----+----+----+ ---+' 0.000 *************** " 0.050 4- t ,u 0.100 ***************- ' O.150 *-
.O,200 ***************- "
-0.250 **********
0.300 *************** '
+- 0.350 ******
0.400 ***************- ' O.450 *************** 0.500 ***************
.7
.0.550 * .
O.600 *************** 0.650 *.************** 0.700 *************** 0.750 * ' O.800 *************** 0.850 * - ' I 0.900 ****************************** '
+----+----+----+----+----+----+----+----+----+----+
Distribution of 14 experts for interface activity , Percent I O 20 40 60 80 100 $ Activity +----+----+----+----+----+----+- -+----+----+----+- ' t 0.650 ***** ' 0.700 *
-0.750
- 0.800 *****
0.850
- e O.900
- 0.950
- i' 1.000 ********************************************
i
+....+....+....+.--.+....+.-..+....+....+....+....+ '
Distribution of 14 experts for intraslab/ deep zone activity i
+
[ r Figure 3-3 Aggregate Distributions for Probability of Activity f I
}
I i
DeOMATRIX Percent 0 20 40 60 80 Segmentation
+----+----+----+----+.---+--.-+----+.--.+-...+..100 .+
none ***** Nootka-Blanco ****************************************** 6'N **** So1 Vancouver I'**
+....+....+....+....+....+....+....+....+....+....+
Distribution ~of 14 experts for interface segner.tation-4 Percent Minimum- 0 10 20 30 40 - Depth (km) +----+----+----+----+----+----+----+----+----+-- 50 -+ 5 ************************** 10 **** 15 ************************ 20 *********************************************** 25 ****
+....+....+....+....+....+....+....+....+....+....+ r Distribution of 12 Experts for Minimum Depth of Rupture on Interface l
Percent Maximum- 0 10 20 30 40 Depth (km) +----+----+----+----+----+----+----+----+----+-- 50 -+ 30 ********* i 35 ************************** l 40 ********************************* 45 ************** l 50 *********************** *
+....+....+....+....+....+....+....+....+....+....+ i.
l l Distribution of 12 Experts for Maximum Depth of Rupture on Interface I l l Percent Seismicity 0 20 40 60 80 100 Pattern +----+----+----+----+----+----+----+----+----+----+ Uniform
- Observed **************************************
Uniform N/S ***** Variable N/S **********
+....+....+....+....+....+....+....+....+....+....+
Distribution of 14 for Intraslab Seismicity Pattern i l ~ Figure 3-4 Aggregate Distrioution For Location of Rupture , l l r t I ( 1 (
*w
'I' *** " ve we ""
or t
. e ,u, ur
-(##
- 3. J'.~;' . '
. ~.
+ .* .
' ?/ .C , . q'" u == w~
..l . , *\ , a 1, ~
s.
.* . \. * . . <
. f g. .
*
- eB - 8 9 u.v'* *.
..=. c*n;.
%\ .
. u N.
. $ 0 - S 9 W.ywmede
. . . . . . . ~
...'..h.g 1 -
_r 9-.. s O *p
. .- 4.
,, * = s. e <
- i i - i ~ ~ - ~,
g
- y. }
. 4. ,/
e Asie
- - ,.4,neged
- . -y.H.y .nen.1,.
-- Ape m.fl.a
- g .
~ ~ . . . , ,,
t .,
\ , 1 -- Aac w.# cea* w s
.j s *
- K
. .. . * ~ ~ %. . . g. t. ;. * - *
\ ,
i
; .. e .
- 1. . d.c . . ..v. .=. .
J
. ,,.c.
.s .e
- 1\ s .
.. . . j. .. .
\v >
,g
. . h. ...- .m.- Ig *. *
. .m . *
,' .~.' O. *. g
' ' " * .q/ * * *
,Y ' . - ,. . . . .
.i d
wott: te.c.ai-o w.w e..ae.e sen
*
- ef .4 t
,c [ , Sou CE: a grader Pasda ,e
=.. . . .e L.;: . t .
. tm .,
. - . e .
> 6.a eu.si. .wes e osis n.m
,. . . \
.4 f . l
. 4 4" l ..-"
*. w-
~ . 9 40 80 t
- 12,0 100 20,0 Md.e
*. i A.;.['@j%.%g.i.;,;g:.. ., . .
;n% *
,s .a
....--~..C.-...-,......
N,
.. . ;. _.._5 .
~. .-,. m
. , . . , = *. . *:
. .. . .. e . . , , ,, a .a s ,,,. . . m. ,
s* W "
, i. ..
Ig. ** * . I* * *** l
.., ......, r - h .~ W *-
5tr$MICITV CF TWE PActf sC N 37 ri 2s.2s Figure 3-5 Plate Boundaries and Regional Seismicity f
ze.
. .e c
' _3'
,3
~
p > 05 s GEOMATRIX r Percent Maximum 0 11 0 20 30 40- 50 Masnitude +----+----v----+----+----+----+----+----+----+----+
, 6l.00 ***************************************
6.25 *-
-6.50 * .
'.75 * .
6 7.00 ***
'7.25'*
7.50 *****-
'7.75 *****
8.00:t**** 8.25 ************************ 8.50 ******* 8.75 ****** 9.00 "*****
.9.25 "*****
9.50 n*** 4.----+----+---.+----+----+----+----+----+----+----+ Conditi.onal distribution of 5 experts for interface maximum magnitude used for component level aggregation Percent , Maximum 0 10 20 30 40 50 Magnitude i----+----+----+----+----+----+----+----+----+----+ 7.50 ******** 7.75 $******* 8.00 ******** 8.25 **************************************** 8.50 *********** 8.75 ********* 9.00 ********* 9.25 ********* 9.50 *****
+----+----+----+----+----+----+----+----+----+----+
Conditional distribution for interface maximum magnitude used for "gap filling" Percent Maximua 0 10 20 30 40 50 MaRnitude +-- 6----+----+----+----+- --+----+----+----+----+ 6.50 ***** ; 6.75 **** 7.00 *$;**************************"******* 7.25 *******~*************** 7.50 ******* **************************** 7.75
- 8.00 **
+ ....+----+----+----+----+----+----+----+----+----+
Distribution of 11 experts for int.ruslab maximum magnitude Figure 3-6 Aggregate Distributions For Maximum Magnitude
- .. - . .~.
,g.
, t q-7> ['q ;Q 1 t GEOMATRIX
- p p ,
1 Percent Return Period-0 -21 '40 60 80 100-
!(years)~. . ----+----+----+----+----+----+----+----+----+----+-
_ ,200 * - 300 **'
- 400 ********************e**************
500 *********
~600 **
-700 **
800 **
'900 **
r .. 1000 **:
^'
1100 ** 3 1200 **
-1300 **
1400 ** 1 1500 ** ' 1600 **
'1700 **
;1800 **
- 1900 **
- 2000'** ~
+-___+.___+____+.___+ ___+____+,___+____+__._+.___+
Distribution of 5 experts . P
'd
- Figure 3-7 Aggregate Distribution Return farlod of Large Magnitude Events Based on Geological Data n
y i k
g}"i
-- frary- <
n ,' ;d c [,,4-.,'i L 9:%' ' ~, , , . j t Ie -
\
i[y
. Mosemaarnix
-l Percent a
. Convergence 40 15 101 .15 25 "a ; Ra t e ( 'e m / y r ) ' + - - - - + - - - -+ - - - - +, - - - -+ -'- - -+ - - --+----+----+-
-+- - - -+- - 2 0 .
> 1. 0 : ** . <
11.5'*** f . t o.******************--
,2.5~****************
3.0:***************************************
'3.5 **************************************
4.0'******************************************** 4.5 *********************** 5.0 *********************** 5.5'** 6.0 ***
+ ..+..__+_.__+... +....+_...+ ___+__..+____+....+-
. Distribution of 14 experts foriconvergence rate Percent Seismic 0 5 10 15 20-Coupling +----+----+----+----+----+----+----+----+----+-- 25 -+
-0.00 *******
0.05 *************************************** 0.10 ****************************** 0.15 ********** ' 0.20 ******** 0.25 *********** 4' o,30 ********** 0.35 ******** O.40 ****** 0.45 ***** O.50 ****** 0.55 *******
. 0.60,********'
4 0.65 ********* 0,73 ********* ir 0.73 ******* 0.80 *****- 0.85 ****' o- < 0.90 *** 0.95 ********** 1.00 ********************
.g..__+....+____+.___+ ___+_ ..+ ...+_.__+____+....+
[/ 0 ' Distribution of 11 experts for seismic coupling l 1. Figure-3-8 Aggregatt Distribution For Moment Rate Recurrence Parameters it h I l
, :lu a , -- , , . . , . . . . _ _ _ _ . . _ . ., . . . . , . , _ . .
- Y\{,.. ,
Jh :,gy , 3, s 1- s
- GEOMATRIX :
l;i & i
, 1, 1 t s li'
,t
, % ., y' g ;
-' ^
1' . , . i ,
- s. .
hk
.4
~
~
Exponential -
- _ _ _ . Chaoctaris6c -
. Maxirnt.m Moment
- 1 o-1 ' A- _
..s g
~
g . g . u s . C %
- e. s -
E e
-}}