Lake_Youngs-Regional_Seismicity_Review_1985.pdf

I

I I

I I I

I I I I

I I

FROM EB~5 CO 8ELLE UU E ,W ~ 1 £1 . 6 . 199 1 1 1: 21 P. :2

August 15, 1985

Ebasco Services, Inc.

400 112th Avenue Northeast Bellevue, Washington 98004

Converse Consultants a.oltc:hnlc::el engln .. rlng .nd Applied Belene ••

84-5119-11

Attention: Mr. Don K. Smith, Project Manager

Subject: TECHNICAL MEMORANDUM Review of Regional Seismicity Lake youngs Project for Seattle Water Department

This memorandum summarizes the results of our review of regional seis micity for the Lake Youngs project. This submittal addresses item 1 of Attachment A to our purchase contract No. ES 7065. Probabilistic analy ses for thi s study were conducted by International Engineering Systems Corporation of Vancouver, British Columbia, under subcontract to Con verse Consultants.

Our firm previously completed an earthquake motions study for the Seattl e Water Department as a part of contract W79-33 to eval uate the "safety and seismic stability of 12 open " reservoir~. The report for this previous study was titled "Report on Earthquake Motions for Reservoir Seismic Stabil ity Analyses" and was dated October 15, 1980. The scope of the current study incl uded revi ew and updating of our previous work so that current design motions could be used 1n design of remedial con struction at Lake Youngs dams and 'more detail ed eval uat ion of Green Lake, Maple Leaf, and Lincoln Reservoirs.

ConW1H Conlullant., Inc.

300 Elliott Avenue West Suite 35U Se.ttle, W •• h;ngton 98119 Telephone 205 285-5200

FROM EBRSCO BELL EVU E, WR 113 . 8.1991 11 : 22 P . 3

-2-

F.arthguak€ Rec~r~

A review of seismic events recorded by the University of Washington seismograph network and information from the Canadian Earth Physics

Branch indicated that the time period 1979-1984 did not include a major earthquake in the Seattl e area. Numerous events have been recorded in association with volcanic activity at Mt. St. Helens, but these are typically small and have not affected structures in the Seattle area. A list of recorded earthquakes with magnitude 3.5 or larger during this period is presented in Table I.

Approach

Our understanding of the neotectonic framework for western Washington is similar to that described in our 1980 report. For this review, a prob abilistic analysis of the historical seismicity was conducted to allow greater flexibility in the shape of seismic source zones and to reflect the quality and completeness of the data.

The 1949 and 1965 earthquakes which caused Significant damage in western

Washington were both relatively deep, with fOci at about 70 and 58 kilo meters respect ivel y, whereas many small er earthquakes recorded 1 n the region are thought to have been relatively shallow. Detailed seis mographic data recorded in western Washington between 1970-1978 confirm a Significant division of earthquakes into shallow and deep suites, based on spatial separation of foci (hypocenters), energy release, magnitude-frequency rel ations, and focal mechanisms (Crosson, 1983).

Only about 10 percent of the earthquakes recorded during this period occurred at depths below 35 kilometers, but the total energy released by this deeper suite was slightly greater than by the more numerous earth quakes of the Shallow suite. The probabilistic analysis utilized in this review permitted incl usion of the fact that a significantly greater proportion of larger earthquakes occur below depth s of 35 kilometers as compared ~o the shallow depth data suite. , ..

Converse Consultants, tnc.

I

I -

~ROM EBASCO 8 ELLE UUE,W A 10 . 8 . 1991 1 1:23 P . 5

-3-

Seismotectonic Provinces and Recurrence Relationships

For purposes of this review, four seismotecton ic provinces were deline ated (Figure i), based on the distribution of historic seismicity and its relation to neotectonics. The rate of activity for each province is described by a magnitude recurrence relation of the form (Gutenberg and Richter, 1966):

N (~ M) = No exp( -6M)

N (2. M) = 0 for M< M - x for M > Mx where N (2. M) is the number of earthquakes of magnitude greater than or equal to M occurring in the province per year, No and 6 are recurrence p~rameters for the province, and Mx is the maximum earthquake magnitude believed possible within the province. The provinces and their computed recurrence relations are as follows:

SEISMOTECTONIC PROVINCES

Province Name

Puget Sound Shallow 25,800 7.80 2.24 7.5

Puget Sound Deep 25,800 2.70 1.63 7.5

North Cascades 44,500 1.51 1.84 7.5

Mt. St. Helens-Mt. Rainier 8,880 3.80 2.18 6.0

(N

JJ

is the number of events per year of M > 3 for the province as a wh e)

Recurrence parameters for the seismic source zones were derived from an analysis of historical earthquakes contained within each province. The data are from the Canadian Earth Physics Branch files, which include earthquake~" recorded by NOAA and the University of Washington sei smic network. To ensure an accurate count of the rate for each magni tude

'1 (§ g

0 :[ i " ;; !

t it n r > m

-0 -< m e

C lZ

C

O rn r n

It> rn

:;I o

II>

rl

P

:IE

.1

:x l ~

0 Z

1> r- I' ';

=J f' ,U lf

IC

I

~f e il

2.

1)

H r '" ffi is

C l

C l

, '0 o o

I I I I I I I

FROM E8ASCO BELL EVU E,W R 10.8.199 1 1 1: 23 P . 6

-4-level, the data set is restricted to events within the foll owing time magnitude windows, for which the catalogue is considered reasonably com pI ete:

2 < M < 3 1970 - present

3 < M < 4 1965 - present

4 < M < 5 1951 present - 5 '- M < 6 1915 - present

6 < M < 7 1900 - present

7 <M 1860 - present

Some modi fications to these windows were made for areas outside the Puget Sound regi on, for whi ch the coverage of small events is less com pl ete.

For the Puget Sound zone the relations were defined such that the sum of the deep and shallow rates would gi ve the corresponding total historical rate for the zone as a whole. The relative proportion of deep and shal low events was assigned based on data compiled by Crosson (1983).

puget Sound seismi city is reI ated to subduction of the Juan de Fuca pl ate beneath the North American pl ate as di scussed extensivel yin the

SCientific literature (for a recent review, see Taber and Smith, 1985).

Deep events (greater than 35 kilometers) are associated with gravita tional sinking of the subducting slab, while shallow events (less than

35 kilometers) are attributed to crustal subsidence in combination with regional compressive stresses. Within the Puget Sound province, seis micity was mOdeled by two overlying zones , one containing the shallow, crustal events and one contai ni ng the deep, subcrustal earthquakes. To date, large ea rthquakes within the zone, notably the M7.1 and M6.5 earthquakes near Seattle in 1949 and 1965, respectively, have had deep

(>50 ' kllometers) foc i . It is believed that future large earthquakes would likely be deep, although the possibility of large shalloYi earth-

)quake cannot be discounted. This view is supported by the data of

Crosson (1983), which show that larger (M>5) events are more often deep

I'

I

I I I

FROM EBRSCO BEL LE UUE.WR 1~ . 8 . 199 1 11:24 p , 7

-5-than shallow, whereas small (M2 to M3) events (Ire more often shall ow than deep. Accordingly, it is assumed that large events, up to M7.5, could be either deep or shallow, but a lower rate is assigned to large shallow events.

Seismicity in the North Cascades province probably arises from inter actions of the oceanic and continental pl ates. The eastern boundary is assumed to be near the east side of the high Cascades, ref1 ectlng the belief that a significant change in physiography signifies a significant change In the underlying stress regime. The zone experiences scattered seismicity, but a1 so contains one very significant historical earth quake, namely the 1872 event of approximately M7.4 (Malone and Bor, 1979). That earthquake may have occurred along the Straight Creek fault lone, or a northern extension, although a definitive correlation of the event with this or any other fau1 t has not been possibl e. For this analysis, a large earthquake, up to M7.5, is assumed possible anywhere in the North Cascades province, although this is considered a rare occurrence.

The volcanic zone to the south of the site encompasses seismicity geo graphically associated with Mt. St. Helens and Mt. Rainier. It has no real impact on the reservoir sites, due to the proximity of the sites to the more significant puget Sound and North Cascades source zones.

The boundary between the Puget Sound and North Cascades provinces cannot be precisely located considering existing data. The province boundary selected for this analysis is based on the landward- limit of the deep, subduction-related earthquakes, as defined by seismicity and tectonic considerations. This boundary al so 1 ies generally near the eastern limit of the Puget trough geologic structure; however, instrumentally recorded crustal seismicity extends diffusely east of the ass1gned pro vince boundary, especially north of the Devils Mountain fault. In· elusion of this more diffuse crustal seismiCity in the North Cascades province is expedient for this level of study as a means of -apportioning pre-instrumental sei smicity between the deep and shall ow Puget Sound provinces and In assigning a magnitude-recurrence relation to the North

Cascades source zone.

Converso Consultants, Inc,

• I , I I l I I

. F'RO M E BASCO BE LL E VU E , WA le. 8 .19 9 1 1 1 : 25 p, S

-6-

Probabilistic Seismic Risk Study

In areas such as the Puget Sound area, where active faults have not been identified and geologic structure is obscured by deep overburden, sta tistical analyses of the historical earthquake record provide useful estimates of seismic risk to engineering structures. In the case of relatively stiff structures, peak horizontal ground acceleration is an indicator of the effects of ground shaking in engineering practice, The ,probabl1 isti, method for this prel iminary seismic hazard determina tion was that of Cornell (1968) util izing the computer program by

McGuire (1976). The ~ethod assumes that seismicity of an area can be modeled by temporally and spatially homogeneous seismogenic source zones, Four source zones were defined on the basis of geologic and tectonic interpretation as described in the previous section. The magnitude recurrence reI ationship ,and upper magnitude bound for each zone were then establ ished based on observed seismicity and geologic evidence. These recurrence relationships, combined with a statistical model of motion attenuation and earthquake occurrence, were then used to develop a relationship between the expected level of strong ground motion (peak ground accel eration) and its annual probabil ity of being exceeded.

Reg i onal attenuat i on of shall ow earthquakes was taken as that descr,i bed by the empirical (log-rorma1) reI ation of Joyner and Boore (1981). For the shallow' suite of Puget Sound events, 'motions were assumed similar to those in Cal ifornia, with an effective depth parameter of 7.3 kilo meters. The largest hori zontal component of hor1 zontal accel eration obtained from the Joyner and Boore expression was reduced by 12 percent to represent mean peak ground acceleration (Campbell, 1981), which is an appropriate parameter for simpl1fied analyses of embankment stabil ity and I iquefaction potential.

For deep"Puget Sound events. the Joyner and Boore attenuation reI ation ships were modified to be consistent with recorded accelerations from the 1949 and 1965 events by including an average effective depth of 35

, F,ROM E B~SCO BELL E UU E ,W ~ l e.8 .1 991 11:25 P. 9

~7~ kilometers. The proportion of deep vs. shallow events was taken as that indicated by instrumental data since 1970 with known deep events, such as the 1949 and 1965 earthquakes, included in the appropriate data set

. by inspection.

Statistical scatter about the mean attenuation relations is char<lcter ized by the standard deviation of the natural IO\larithm of the residu als, or uncertainty factor. For purposes of basic dam stability analy ses, such as pseudostatic or slope deformation calculations, it is con sidered appropriate to include an uncerta inty factor in the estimated degree of ground shaking. For the low probability risk of interest 1n this preliminary eval uation, major risk contributions are from large, near~field events. An uncertainty factor of 0.4 was included in this analysis as appropriate for large, near~field events (Campbell, 1981).

Results of the Cornell analysis are summarized on Figure 2 for exceed~ ance probabilities in the range of 0.1 to 0.0001 per annum. Correspond~ ing return periods, defined as the inverse of annual probabil ity, are also shown.

Recommended Design Ground Motions

For critical structures, such as water-I mpounding embankments which could cause loss of life and significant property damage as a resul"tof failure, a low probability seismic event is cons idered appropriate for analyses of overall embankment stability. The level of shaking with an

~nnual risk of 1O~4, or return period of 10,000 years, often has been taken as the "maximum credibl e earthquake" for these types of projects.

As shown on Figure 2. a peak ground accelerat i on of 39 percent of grav ity is indicated by these statistical analyses as having an average return period of approximately 10,000 years. Input of coordinates for the various reservoir sites did not significantly change these results.

Therefore, the same seismic design criteria are recommended for all of the sites currentl y being studied.

Con ..... , Consultants. Inc.

-I , ,I I

I I I I I

I I I I I I , I I

, FROM EBA5CO BEL LEVUE , WR 10. 8 .19 91 11=26

e = C C

S O) Co

0)

U C

'" -0 0)

OJ

U

" Lo.J o

Expected Average Horizontal Peak Ground Acce leration (9)

PROBABILI TY OF HORIZONTAL PEAK GROUND ACCELERATION

LAKE YOUNGS SEISMICITY REVIEW

for EBASCO Services, Inc.

P . 1 (1

Proj6cI No.

84 -5119-11

Figure No.

I I I I I I I

I I I I I I I I

I

FROM EBASCO BELLE VUE.WR 10 . 8 .1 99 1 11 :2 7 P . 1 1

-8-

It is our understanding that a smaller, more frequent event is of inter est in eval uating the performance of Water Department facil Hies. As shown on Figure 2, an earthquake with an average recurrence interval of about 100 years is expected to result in average peak horizontal ground motions at the subject sites of approximately 12 percent of gravity.

Our analyses indicate the most significant contr1but1on to these low levels of risk is from large, shallow events 1n the Puget Sound region.

Estimated earthquake parameters for the two ris k level s of interest are listed below:

Level Earthquake Parameter Cl ass I

Richter Magnitude 7.0 to 7.1

Average Peak Horl zontal Ground Acceleration (9) 0.39

Annual Probability of Exceedance Expected Return Period (years) Epicentral Distance (km) Depth (km)

Dean E. Ryden PrinCipal Engineer

DER/ jm

INC.

10-4

10,000

< 15

5-20

ConverSo ConSUltants, Inc.

of Risk Cl ass I I

5.6 t o 6.0

0.12 10-2

< 30 5~20

I I I I I

FROM EB~S CO BELL EUU E.WA 10 . 8 .19 9 1 11 : 27 P . 12

-9-

REFERENCES

Campbell, K.W., 1981, Near-source attenuation of peak horizontal accel eration: .Bulletin of the Seismological Society of America, v.

71, n. 6, p. 2039-2070.

Cornell, C.A., 1968, Engineering seismic risk analysis: Bulletin of the Seismological Society of America, v. 58, n. 5, p. 1583-1606.

Crosson, R.S., 1983, Review of seismicity in the Puget Sound Region from 1970 through 1978, in Proceed ings of Workshop XIV - Ear t hquake Hazards of the Puge~Sound Region, Washington: U.S. Geological Survey, Open-File Report 83-19, P. 6-18.

Gutenberg, B., and Richter, C.F., 1966, Seismicity of the Earth: Stechert Hafner, New York.

Joyner, W.B., and Boore, O.M., 1981, Peak horizontal acceleration and velo . city from strong-mot ion records 1 ncl udi ng records from the 1979

Imperial Valley , California, earthquake: Bulletin of the Seis mological Society of America, v. 71 , n. 6, p. 2011-2038 .

Malone, S.D., and Bar, S.S., 1979, Attenuation patterns in the Pacific Northwest based on intensity data and the location of the 1872 North Cascades earthquake: Bulletin of the Seismological Society Of America,v. 69, n. 2, p. 531-546.

McGuire, R.K., 1976, Fortran computer program for seismic risk analysis:

U.S. Department of the Interior, Geolog1cal Survey, Open-File Report 76-67.

Taber, J.J. and Smith, 5.101., 1985, Seismicity and focal mechanisms associ ated with the subduction of the Juan de Fuca plate beneath the Olympic Peninsula, Washington: Bulletin of the Seismological Society of America, v. 75, n. I, pp. 237-249. -

Conve ... e Consult.nlB, Inc.