OREGON GEOLOGYtaylors/g473/seismic_hazards/... · 2009-04-18 · 40 OREGON GEOLOGY, VOLUME 63,...

OREGON GEOLOGY

IN THIS ISSUE :

The Portland Hills faultSeismic rehabilitation at the Oregon State Library in Salem

Results of a new method of Fourier grain-shape analysis

Oregon Department of Geology and Mineral Industries Volume 63, Number 2, Spring 2001

OREGON GEOLOGY, VOLUME 63, NUMBER 2, SPRING 2001 39

The Portland Hills fault: An earthquake generator or justanother old fault?by Ivan G. Wong, Seismic Hazards Group, URS Corporation, 500 12th Street, Suite 200, Oakland, CA 94607; MarkA. Hemphill-Haley, Seismic Hazards Group, URS Corporation and Department of Geological Sciences, University ofOregon, Eugene, Oregon 97403; Lee M. Liberty, Center for Geophysical Investigation of the Shallow Subsurface,Boise State University, 1910 University Drive, Boise, ID 83725; and Ian P. Madin, Oregon Department of Geology andMineral Industries, 800 NE Oregon Street, Suite 28, Portland, OR 97232

ABSTRACTSeveral lines of indirect evidence

and preliminary interpretations of re-cently collected seismic reflectiondata have led to the conclusion thatthe Portland Hills fault at the easternbase of the Portland Hills appears tobe capable of generating large-mag-nitude earthquakes. Although nohistorical earthquake can be associ-ated with the Portland Hills fault,small-magnitude seismicity in thepast 20 years in the vicinity of thePortland Hills fault zone, which in-cludes the Oatfield and East Bankfaults, suggests that one or all ofthese structures may be seismogenic.The Portland Hills fault may be40–60 km long, probably dips to thesouthwest beneath the PortlandHills, and may slip in a reverse-oblique sense. Limited observationssuggest that, on average, the inter-vals between large earthquakes area few thousand to more than10,000 years. Given its location inthe midst of the Portland metropoli-tan area, rupture of the PortlandHills fault resulting in a large earth-quake could be devastating. Futurestudies are required to characterizethe earthquake potential of the fault

in a more definitive manner and toprovide an improved basis for pre-dicting the hazards that would resultfrom such a large earthquake.

INTRODUCTION

Since the mid-1960s, it has beensuggested that a fault is located atthe eastern base of the northwest-trending, westward-sloping PortlandHills and within the Portland metro-politan area which has a populationof 1.4 million (Figure 1). Little con-sideration was given to the possibili-ty that the “Portland Hills” fault wasactive1 (or seismogenic, i.e., capableof generating earthquakes). This wasconsistent with the prevailing viewheld until the past decade that Ore-gon was not particularly seismicallyactive. However, as speculationturned to recognition in the PacificNorthwest that the Cascadia sub-duction zone megathrust was seis-mogenic, attention began to shiftalso to evaluating the earthquake po-tential of inland crustal faults, partic-ularly those located in or near urban

areas. For example, significant ef-forts have been focused on the Seat-tle fault since the suggestion that itruptured in a large earthquake(>MW [moment magnitude] 7)1,100 years ago (Bucknam and oth-ers, 1992). If urban crustal faultssuch as the Seattle and the PortlandHills faults are seismogenic, theycould generate large, disastrousearthquakes similar to the MW 6.9earthquake in Kobe, Japan, in 1995.

In the following paper, we reviewwhat is known about the geology ofthe Portland Hills fault, particularlyits characteristics, which might helpquantify its earthquake potential.Recent studies, which have focusedon this quantification, are discussed.Last year, Wong and others (2000a;2000b) released scenario earthquakeground shaking hazard maps for thePortland metropolitan area. They as-sumed a MW 6.8 earthquake ruptur-ing 30 km of the Portland Hills faultat the base of the Portland Hills (Fig-ure 2). The potential ground shakingfrom such an event would greatlyexceed the ground motions from aMW 9 Cascadia subduction zonemegathrust event in the Portlandarea. This paper elaborates on the

Editor’s NNoteDuring tthe ffinal sstages oof ppreparing tthis aarticle ffor ppublication, eevidence wwas ddiscovered tthat tthrows nnew

light oon tthe qquestion iin tthe ttitle, wwithout iinvalidating tthe eessence oof tthe ppaper. IIn aa ssidebar oon ppages 447–49,Ian PP. MMadin ggives aa ppreliminary aassessment oof tthe nnew-ffound eevidence ffor aactivity oon tthe PPortland HHills ffaultin ttimes mmore rrecent tthan hhad bbeen kknown bbefore. AA mmore eextensive ddescription aand ddiscussion wwill bbe ppub-lished bby MMadin iin Oregon GGeology at aa llater ddate.

1 We consider a fault in the Pacific Northwestto be “active” if it has moved (been dis-placed) at least once during late Quaternarytime (the past 780,000 years).

40 OREGON GEOLOGY, VOLUME 63, NUMBER 2, SPRING 2001

basis for selecting a MW 6.8 sce-nario, although the evidence at thattime was not compelling that thePortland Hills fault was seismogenic.

HISTORICAL SEISMICITY

The Portland area has exhibited alow to moderate level of historicalseismicity (Figure 2), compared toother areas in the Pacific Northwest.The area is not as seismically activeas the Puget Sound region to thenorth but may be the most activearea in Oregon for events of MW≥3.0 based on the historical record(Wong and Bott, 1995). Detaileddiscussions of the historical seismici-ty can be found in Bott and Wong(1993) and of instrumentally record-ed seismicity since 1982 in Yelin andPatton (1991) and Blakely and oth-ers (1995). Based on the historical

earthquake record, seven felt earth-quakes (Richter magnitude [ML]>3.5) have occurred in the vicinity ofPortland since 1850 (Bott andWong, 1993).

The first significant earthquake tostrike the Portland area occurred onOctober 12, 1877. It was felt intowns around Portland, but its maxi-mum intensity was in the city (Modi-fied Mercalli [MM] VII). This eventof approximate magnitude ML 5¼damaged chimneys and was feltover an area of 41,000 km2. OnFebruary 3, 1892, a “severe” earth-quake of estimated ML 5 causedbrick buildings to sway and windowsto rattle in Portland, terrifying its oc-cupants. A ML 4½ event on Decem-ber 29, 1941, shook an area ofabout 9,000 km2, including thetowns of Portland, Hillsboro, Sher-

wood, and Yamhill in northwestOregon and Vancouver and Wood-land in southwest Washington. Ef-fects included shattered windows,cracked plaster, and overturned ob-jects. Another ML 4½ earthquakeoccurred somewhere between Port-land and Vancouver on December15, 1953. Effects were similar tothose encountered in 1941. Anearthquake on November 6, 1961(ML 5) shook a large area (23,000km2) of northwest Oregon andsouthwest Washington. Damage in-cluded a fallen chimney, crackedplaster, broken interior lights,jammed doorframes, and groceriesthrown from shelves. The instrumen-tal location of this event is uncertain,although an aftershock felt mostly inPortland suggests the main shockalso occurred there. On January 27,

Figure 1. Oblique air photo of the Portland Hills toward the northwest and inferred location of the Portland Hills fault.The fault is mapped near the eastern base of the escarpment north of West Burnside Street. To the south, the inferred faultsteps toward the east, where it traverses the downtown area and crosses the Willamette River (Figure 2). Fault locationbased on Madin (1990). Air photo courtesy of Northern Light Studio.


1968, an earthquake estimated tobe ML 3.7 in size was felt in Port-land. The event is believed to haveoriginated in east Portland but at adepth of 20 to 24 km.

The most significant local earth-quake in Portland’s history occurredon November 5, 1962. Yelin andPatton (1991) calculated a magni-tude of MW 5.2 for the event andBott and Wong (1993) assigned aML 5.5 based on its felt area andother instrumental estimates. Theearthquake was felt widely over anarea of 70,000 km2 in northwestOregon and southwest Washington,

causing damage typical for an earth-quake of this size: broken windows,cracked and fallen chimneys, andcracked plaster. The main shock wasfollowed by numerous small but un-felt aftershocks. With the exceptionof the 1962 earthquake, which wasrelocated by Yelin and Patton (1991)to a location in Vancouver (Figure 2),the large location uncertainties of thehistorical earthquakes with ML >3.5in the Portland area prior to 1982prevent a determination of whetherany of these events were associatedwith the Portland Hills fault or adja-cent major structures.

PREVIOUS FAULT STUDIES

Numerous geologic and a fewgeophysical studies of the PortlandHills fault have been performedsince the 1960s. The early studiesstemmed from attempts to explainthe presence of the Portland Hills,which are underlain by uplifted Ter-tiary rocks. A major obstacle in char-acterizing the Portland Hills fault isthat nowhere has the fault beendocumented to be exposed at theground surface. Given the heavyvegetation, urbanization, and cata-strophic floods (i.e., the late Pleis-tocene Missoula floods ca. 12,000 to15,000 years ago), there may beareas where the fault extends to thesurface but has simply not been de-tected. The following summary istaken principally from Yelin and Pat-ton (1991), Unruh and others(1994), and Blakely and others(1995).

The Portland Hills comprise anorthwest-trending, asymmetric an-ticline that was uplifted principally inthe late Miocene and Pliocene (after6 Ma). Several models have beendeveloped to explain the origin ofthe northeastern margin of the anti-cline, which forms an abrupt escarp-ment. Both Schlicker and others(1964) and Balsillie and Benson(1971) suggested that the escarp-ment (Figure 1) is the surface ex-pression of a steeply northeast-dip-ping fault. Schlicker and others(1964) inferred that the fault was ofnormal slip. Beeson and others(1976) proposed that the fault waspart of a Portland Hills-ClackamasRiver structural zone and that thezone accommodated right-lateral,strike-slip motion. They also sug-gested that the Portland basin was apull-apart structure and that thewest-bounding Portland Hills faultwas a right-oblique fault. On1:24,000-scale quadrangles of thePortland area, Madin (1990) showedthe Portland Hills fault extending adistance of about 40 km along thePortland Hills (Figure 2). Yelin andPatton (1991) also inferred that the

Figure 2. Portland Hills fault zone and historical seismicity in the vicinity ofPortland, 1841 to 2000. Faults shown are those considered to be potentially ac-tive by Wong and others (2000b). Site NCP is also shown. Earthquake symbol sizeis a function of magnitude. The largest event shown is the 1962 ML 5.5 earth-quake. Smallest events are ML ≤1.0.


Portland basin was a pull-apart basinand that the Portland Hills fault andtheir postulated Frontal fault zone tothe east bound the basin (Figure 2).Their model was based in large parton a re-analysis of the 1962 Port-land earthquake (MW 5.2), whichappears to have resulted from nor-mal faulting within the Portlandbasin. In their model, the PortlandHills fault is a right-lateral, strike-slipfault. In contrast to previous models,Unruh and others (1994) interpretedthe Portland Hills anticline as a fault-bend or fault-propagation fold andbased this interpretation on surfacemapping and sparse drill-hole data.In their model, the Portland Hillsfault is a southwest-dipping, blindthrust fault.

In the most extensive study offaulting to date, Blakely and others(1995) performed a high-resolutionaeromagnetic survey of the Portlandarea. On the basis of interpretationsof the survey and an analysis of thecontemporary seismicity, they identi-fied two additional faults, the EastBank and Oatfield faults (Figure 2),which they include as part of thePortland Hills fault zone. Both faultswere first identified by Beeson andothers (1989, 1991) on the basis ofgeologic mapping. The Portland Hillsfault, also an element of the Port-land Hills fault zone, was not wellimaged by the aeromagnetic data(Blakely and others, 1995). The EastBank fault was previously knownthrough shallow well data that indi-cate ≤200 m of vertical displacementof the underlying volcanic basement.As mapped by Madin (1990) andBeeson and others (1991), the faultis overlain by Quaternary deposits.The Oatfield fault along the south-western slope of the Portland Hillswas also previously known from thestudies by Beeson and others (1989)and Madin (1990).

The aeromagnetic data suggestthat the Oatfield and East Bankfaults are steeply dipping structuresthat exhibit principally reverse slip(Blakely and others, 1995). This slip

would be consistent with earlier ob-servations of folding and thrusting(Beeson and others, 1985, 1989,1991; Unruh and others, 1994).

In an analysis of contemporaryseismicity, Blakely and others (1995)noted that two sequences from1982 to 1985 near Sauvie Island(ML ≤2.8) and 1991 at the northend of the Portland Hills (ML ≤3.5)may be associated with the PortlandHills fault zone, although they ap-pear to have occurred at mid-crustaldepths (15–20 km). Thus relatingthese events to mapped faults isproblematic, given the uncertaintiesof the subsurface geometry of thefaults and the hypocentral locations(Blakely and others, 1995). No otherevents appear to be associated withthe fault zone, particularly along itscentral and southern extent (Figure2). Focal mechanisms of the 1982 to1985 and 1991 sequences exhibitboth right-lateral and reverse slip(Blakely and others, 1995; Figure 3).A component of strike-slip motionalong the Portland Hills fault zone isconsistent with the pull-apart basinmodel.

No evidence for recent displace-ment on the fault had been ob-served prior to 1998. In a photogeo-logic investigation of the fault, Geo-matrix Consultants (1993) did notobserve evidence for Holocene orlate Pleistocene activity, but the fac-tors mentioned previously could ob-scure any surficial evidence (Figure1). Balsillie and Benson (1971) notedseveral geomorphic features indica-tive of Pleistocene uplift of the Port-land Hills (linear northeastern frontof the Portland Hills, aligned triangu-lar facets, and knickpoints in north-east-flowing streams), which may besuggestive of activity on the Port-land Hills fault. Unruh and others(1994) judged the fault to be “po-tentially active” on the basis of theirreview of available data and geo-morphic evidence for middle to lateQuaternary uplift of the PortlandHills and homoclinal westward tiltingof early to middle Pleistocene de-

posits in the Tualatin basin. The geo-morphic evidence included recentstream incision and downcutting,minor deformation of Tualatin basinfill on the southwest flank, andthrust faulting of the Tertiary Trout-dale Formation in the southern por-tion of the Portland Hills (Unruh andothers, 1994).

In a seismic source characteriza-tion of the Portland Hills fault for thedevelopment of statewide groundshaking maps, Geomatrix Consul-tants (1995) considered two models:a relatively steeply (70º) west-dippingreverse or reverse-oblique fault anda blind shallow-dipping thrust fault.The latter model suggested byUnruh and others (1994) was as-signed little weight. Although therewas no definitive surficial evidencefor the fault being seismogenic, Ge-omatrix Consultants (1995) adopteda relatively high probability of 0.70that the fault was active and basedthis on possible deformation of latePleistocene sediments and the topo-graphic expression of the PortlandHills. They also estimated that poten-tial rupture lengths for the PortlandHills fault range from 28 to 62 km(Figure 2). The latter value is basedon lineaments observed on air pho-tos to the north and south of themapped trace. Slip rates of 0.05 to0.2 mm/yr were estimated, consid-ering the tectonic setting and re-gional kinematics (Geomatrix Con-sultants, 1995).

RECENT INVESTIGATIONS

The Portland Hills fault as well asthe East Bank and Oatfield faultswould pose the greatest seismic haz-ard to Portland because of theirproximity and their potential to gen-erate large-magnitude earthquakes(MW ≥6.5) (Wong and Silva, 1998).Wong and others (2000b) character-ized the Portland Hills fault in theirdevelopment of scenario and proba-bilistic ground shaking maps for thePortland metropolitan area by revis-ing the assessments of GeomatrixConsultants (1995). They assumed a


higher probability (0.8) that the Port-land Hills as well as the East Bankand Oatfield faults were seismo-genic. The basis for this assessment,were the following observations:

1. Contemporary microseismicityhas been observed in the vicinity ofthe Portland Hills fault, with the ma-jority of events concentrated atdepths of 5–20 km (Yelin and Pat-ton, 1991; Blakely and others,1995). Although no earthquakes canbe definitively associated with thefault, the occurrence of events issuggestive that there are activestructures nearby. As observed bymany (e.g., Kafka and Levin, 2000),the presence of small earthquakes,more often than not, delineatesareas where larger earthquakes arelikely to occur. We believe the Port-land Hills fault zone or specificallythe Portland Hills fault is the likelysource for future large earthquakesin the Portland area.

2. The few earthquake focalmechanisms in the Portland area andsurrounding region indicate thatnorthwest-striking structures such asthe Portland Hills fault are favoredto be seismogenic in a north-southtectonic compressive stress field(Yelin and Patton, 1991; Blakely andothers, 1995; Wong, 1997) (Figure3). The similarly oriented northwest-striking Mount Angel fault to thesouth may have been the source ofthe ML 5.6 Scotts Mills earthquakeof 1993 (Madin and others, 1993;Thomas and others, 1996).

3. Pratt and others (2001) ac-quired and interpreted single-chan-nel marine seismic reflection dataacross the projections of the Port-land Hills and East Bank faults wherethey cross the Willamette River nearRoss Island and near the Multnomahchannel, respectively. Unfortunately,the Ross Island profiles across thePortland Hills fault are uninter-pretable. The data near the Multno-mah channel appear to indicate thatthe East Bank fault coincides with alarge paleochannel. The layered stra-ta, including a late Pleistocene un-conformity at the base of the paleo-

channel, may be cut by a nearly ver-tical fault. Thus the East Bank faultappears to have been active at leastin the late Pleistocene (Pratt andothers, 2001).

4. Finally, the apparent youthfulgeomorphic expression as evidencedby the abrupt escarpment on theeastern side of the Portland Hills issuggestive of recent activity on thefault. It is possible that this youthfulmorphology is the product of flood

sculpting and scour, but it may alsoindicate late Quaternary and possiblycurrent uplift of the Portland Hills.

Wong and others (2000b) basedan estimated range of slip rates forthe Portland Hills fault on (1) a qual-itative comparison of its geomorphicexpression with faults in other re-gions; (2) long-term slip rates forthe Gales Creek and Mount Angelfaults, which are located to thesouthwest and south of the Portland

Figure 3. Crustal earthquake focal mechanisms of northwestern Oregon andsouthwestern Washington. Most of the mechanisms exhibit strike-slip and oblique-reverse faulting in response to a N-S to NE-SW-directed maximum compressivestress. Figure is modified from and data sources described in Wong (1997).


basin, respectively; (3) historical seis-micity; and (4) estimated range ofdisplacements by Pratt and others(2001). From detailed mapping, R.E.Wells (personal communication, 1997)estimated a long-term slip rate forthe Gales Creek fault of about 0.24mm/yr since 50 Ma and 0.6 mm/yrsince 15 Ma. Wang and Madin (2001)have estimated a slip rate for theMount Angel fault of approximately0.23 mm/yr for the past 30 ka.

Bott and Wong (1993) calculatedearthquake recurrence for the Port-land region based on the historicalseismicity record, which dates backto about 1850. Assuming a Guten-berg-Richter relationship, we findthat the return period of earth-quakes of ML 6.5 and greater isabout 1,000 years. For an averagedisplacement of about 0.5 m, appro-priate for a MW 6.5 earthquakebased on the relationship of Wellsand Coppersmith (1994), the equiv-alent slip rate would be about 0.5mm/yr, distributed across faultswithin the Portland region. If thePortland Hills fault zone (includingthe Portland Hills, East Bank, andOatfield faults) and the Frontal faultzone are taking up most of this slip,the average slip per fault zone wouldbe about 0.25 mm/yr.

Pratt and others (2001) estimatethat the late Pleistocene (~15,000-yr-old) unconformity across the EastBank fault at the Multnomah chan-nel may be displaced vertically “sev-eral” meters. Assuming that this ap-parent vertical displacement is atleast 3 m, that would indicate a sliprate of about 0.2 mm/yr without ac-counting for any horizontal compo-nent of slip.

Wells and others (1998) suggest-ed that the Cascadia forearc is mi-grating northward along the coastand is breaking into large rotatingblocks. One of these blocks, theOregon coastal block, is accommo-dating part of the 7–9 mm/yr ofnorth-south shortening of the fore-arc. Wells and others (1998) alsosuggested that the shortening istaken up by deformation along block

margins. The northeast boundary ofthe Oregon block is located near thePortland Hills fault zone (Wells andothers, 1998).

To constrain maximum slip-ratevalues, we can use the convergencerate of 4–7 mm/yr proposed byWells and others (1998) for the Ore-gon forearc. Of this total rate, R.E.Wells indicates that northwesternOregon may be shortening at a rateof 2–3 mm/yr (personal communica-tion, 1999). Within northwesternOregon, the major fault systems ap-pear to be the Portland Hills, theGales Creek-Mount Angel, Mollala-Canby, and Frontal fault zones(Blakely and others, 1995, 2000;Wong and others, 2000b). If thesefour fault zones are taking up a sig-nificant portion of the 2–3 mm/yr ofregional shortening and if the activi-ty on the fault occurs at about thesame maximum rate, the slip ratesalong them could range up to amaximum of 0.5–0.75 mm/yr. Suchhigh rates seem unlikely except forpossibly the Portland Hills fault, be-cause they would result in much morepronounced geomorphic expressionof the faults than is observed. Further-more, these high slip rates assumethat there is no aseismic componentto the crustal deformation.

Based on the above observations, arange of slip rates of 0.05–0.4 mm/yrwas adopted for the Portland Hillsfault as well as the East Bank andOatfield faults (Wong and others,2000b). Using the relationship ofWells and Coppersmith (1994) tocalculate an average displacement of0.8 m for a MW 6.8 earthquake), wefind that the average recurrence in-terval would range from 2,000 to16,000 yrs.

The maximum earthquake thatthe Portland Hills fault appears to beable to generate is in the range ofMW 6.8–7.2, as calculated from thepotential rupture lengths of 28–62 kmestimated by Geomatrix Consultants(1995). This is a significant range inmagnitudes. Because the resultinghazards will also differ significantlyfrom the low end to the high end of

this range, it is critical that the ex-tent of the Portland Hills fault and itspossible segmentation and thus po-tential rupture lengths be better de-fined. In the modeling of the scenarioground motions for an event on thePortland Hills fault, Wong and others(2000a; 2000b) adopted a MW 6.8event rupturing 30 km of the fault,predominantly along its most geo-morphically pronounced portion atthe base of the Portland Hills.

The fault was assumed to rupturewith oblique slip (50 percent reverseand 50 percent right-lateral strike-slip), dip 70º to the southwest be-neath the hills, and extend to adepth of about 19 km (resulting in afault width of about 20 km). Al-though focal mechanisms in the re-gion exhibit predominantly strike-slipmotion (Yelin and Patton, 1991;Blakely and others, 1995; Wong,1997), with some showing reverseand oblique slip (Figure 3), the pres-ence of the Portland Hills indicatesthat reverse faulting is the primarymode of deformation if uplift is stillongoing.

Future studies of the fault shouldalso focus on constraining the dip ofthe fault. The assumed value of 70ºis extremely uncertain. Because the dipcontrols the downdip width of thefault, the potential area available forrupture may be significantly larger, ifthe fault has a shallow dip (<50º).

The dip of the fault will also shedsome light on the structural relationshipbetween the Portland Hills and Oatfieldfaults. Wong and others (2000b)considered two possible structuralgeometries for the Portland andOatfield faults due to the closenessof the faults (separated by 3–5 km;Figure 2) and their proposed respec-tive westward and eastward dips assuggested by Blakely and others(1995): (1) the two faults each dipat about 70° and merge at a depthof about 5 km to form a single zoneat greater depths Figure 4a), or (2)the two faults have steep dips(>80°) and are not structurally con-nected (Figure 4b). In the first case,i.e., if the two faults merge, they


could be part of a flower structureprovided they are predominantlystrike-slip faults. If they are reversefaults, the Oatfield fault may be aback-thrust to the main PortlandHills fault.

The implication of both structuralmodels is that in any given earth-quake either one or both faults mayrupture coseismically. If both faultsrupture together in a single event,large near-field ground motions willbe distributed over a wider areathan if only a single fault were torupture.

Furthermore, the proximity of theEast Bank fault (Figure 2) suggeststhat it, too, might be structurallyconnected with the above faults.That is unlikely, however, at least ac-cording to Blakely and others (1995),whose interpretations of aeromag-netic data show the East Bank faultdipping away to the east (Figure 4).

CURRENT INVESTIGATION

With funding support from theU.S. Geological Survey NationalEarthquake Hazards Reduction Pro-gram, we employed multiple geo-physical methods, including high-resolution seismic reflection, groundpenetrating radar (GPR), and mag-netic profiling to locate the PortlandHills fault at North Clackamas Park(NCP) south of Portland (Figure 2).We also incorporated data from near-by water wells to correlate the ob-served strata with horizons in the geo-physical data and to aid in laying outthe locations of the seismic lines.

Two high-resolution seismic pro-files at the NCP site provide detailedimages of the upper 100 m of thestratigraphic section, and enable usto locate significant offset in theMiocene-age Columbia River BasaltGroup (CRBG) rocks and overlyingsediments (Figure 5). Ground mag-netic profiles along our seismic tran-sects correlate with offset and orien-tation in the volcanic basement orCRBG rocks. The magnetic surveys,well logs, and ground reconnais-sance show that CRBG rocks cropout near the southwest portion of

Figure 4. Schematic block diagram of the Portland Hills fault zone. Shown aretwo possible interpretations of its subsurface geometry. A third model is that theEast Bank fault dips to the west and is structurally connected to the Portland Hillsand/or Oatfield faults.


the NCP site. Seismic, magnetic, andwell-log information indicate thatbasalts dip shallowly to the north-east and are observed at depthsgreater than 100 m to the north ofthe site. The seismic data show astrong-amplitude, steeply (>20°)dipping horizon that is likely the topof the CRBG sequence (Figure 5). Awell located less than 300 m to thenorthwest of the seismic line en-countered basalt at a depth of morethan 150 m. Reflections fromyounger sediments (possibly TertiaryTroutdale to uppermost PleistoceneMissoula flood deposits) also dipsteeply. To the east, near the south-central portion of NCP, reflectionsassociated with younger sedimentsappear flat lying, but are slightly fold-ed synclinally and faulted toward thewest. This major change in dip appearswithin the Pliocene to latest Pleisto-cene(?) sediments. The dipping stra-ta are imaged to within about 10 mbelow the land surface (Figure 5).

We interpret these data to repre-sent a major splay of the PortlandHills fault that offsets post-CRBGsediments. The change from flat-lying CRBG rocks and younger sedi-ments to steeply dipping strata pro-

vides an indication of the fault loca-tion. The data do not, at present,provide unequivocal evidence forthe style of deformation. However,the position of the offset CRBGmarker is consistent with a south-west-dipping reverse fault which dis-places the Portland Hills northeast-ward over the adjacent basin.

To date, evidence shows that post-CRBG sediments have been faulted. Ifthe youngest faulted sediments im-aged with high-resolution seismic-re-flection methods are Missoula-flood-related deposits, then there was atleast one episode of coseismic surfacerupture in the past 15,000 to 12,000years. The focus of our continuinginvestigation is to determine the ageand extent of these deposits, to as-sess the number of times they havebeen displaced, and to determinethe style of deformation.

CONCLUSIONS AND “WHAT IF?”

There are several lines of indirectevidence that suggest the PortlandHills fault is an active and seismogen-ic structure. Direct evidence includesthe seismic reflection results at NCP,although the interpretation is still pre-liminary. We hope to collect more

definitive evidence during field inves-tigations during the coming summer.If doubt of the earthquake potentialof the Portland Hills fault is removed,the current year’s planned and fu-ture investigations will focus on char-acterizing the maximum earthquakepotential of the fault, its subsurfacedimensions and geometry, rupturecharacteristics, and the frequency oflarge-magnitude events. Regardingthe latter, it will be critical to deci-pher the history of prehistoric earth-quakes on the fault in an effort toprovide some insight as to when thenext event may occur.

The consequences of a futurelarge earthquake on the PortlandHills fault could be severe. Wongand others (2000a) estimated thatground shaking, as characterized bypeak horizontal acceleration, couldexceed 1.0 g for a MW 6.8 event2.(The peak horizontal accelerationrecorded in the 1962 earthquake atthe only existing strong-motion sitewas 0.10 g, in downtown Portland.)

Figure 5. High-resolution seismic reflection profile at the North Clackamas Park.

2 Earthquake ground motions are often ex-pressed in terms of acceleration and the unit“g”, which is the gravitational acceleration atthe earth’s surface equal to 980 cm/sec2. Theonset of light damage is at about 0.1 g.


If the Portland Hills fault were torupture predominantly with reverseslip, an analogue for such an eventwould be the MW 6.7 earthquake of1994 in Northridge, California,which generated some of thestrongest ground shaking everrecorded (as high as 1.9 g) andcaused 58 deaths, and propertydamage of $20 billion. The source ofthe Northridge earthquake was ablind reverse fault which dips about40°–50° to the south beneath theSan Fernando Valley.

Unlike the Northridge fault, how-ever, the Portland Hills fault is situat-ed in the midst of a major metropol-itan area, where the majority of olderbuildings do not meet current build-ing code seismic design criteria, par-ticularly those pre-code unrein-forced masonry structures. Of spe-cial concern, as observed in theNorthridge earthquake, is the po-tential for rupture directivity, whichwill be directed toward downtownPortland. The directivity effects willbe greatest if the Portland Hills faultruptures with reverse slip. Since di-rectivity is a long-period effect (>1.0sec), high-rise buildings, which areconcentrated in downtown Portland,could be particularly susceptible tovery strong ground shaking (Theseeffects are not illustrated on the sce-nario maps of Wong and others[2000b] because they go out onlyto a period of 1.0 sec).

As future investigations proceed,it will be important to update theassumptions considered in the Port-land Hills fault scenario groundshaking maps, (e.g., magnitude, slip,etc.) in order to provide the mostrealistic estimates of ground shakingpossible so that adequate mitigationmeasures can be taken to preparethe citizens of the Portland metro-politan area for this potential disas-ter. Future studies of the East Bankand Oatfield faults and their rela-tionship, if any, to the Portland Hillsfault are also critically needed.

The Portland Hills fault at Rowe Middle School

by Ian, P. Madin, Oregon Department of Geology and Mineral Industries,and Mark A. Hemphill-Haley, Senior Geologist, URS Corporation

For decades, geologists have recognized that the Portland Hills fault is amajor geologic structure that runs right through the Portland Metro area.The fault is responsible for the unusually straight and sharp edge of thePortland Hills north of downtown Portland, but is less obvious southeast ofdowntown. Scattered outcrops and data from water well logs were used toapproximate the location of that southeast extension of the fault on recentgeologic maps. Geologists from the Oregon Department of Geology andMineral Industries (DOGAMI), URS Corporation (an engineering firm), andBoise State University began a research project in 2000 to study the Port-land Hills fault in more detail. The study, funded by the U.S. Geological Sur-vey through the National Earthquake Hazard Reduction Program (NEHRP),seeks to determine whether the fault is active, and therefore poses anearthquake threat to the region. In this case, “active” means that it hasmoved during the last 10,000 to 15,000 years.

The NEHRP study used two geophysical methods to image the fault andwas conducted in North Clackamas Park and on adjacent private property. Thesite was chosen because the inferred line of the fault passed through it, andthe area was relatively undeveloped and quiet, which aided in the collectionof good geophysical data. The resulting imagery confirmed the presence of asignificant fault, located the main fault to within ~100 ft, and showed fault-ing and folding to within about 50 ft of the surface. It did not definitively showevidence of activity, because the seismic and magnetic techniques could notimage the shallowest layers, and the age of those layers was unknown.

This May, a DOGAMI geologist was collecting more magnetic readingswith a handheld instrument to try to map the fault line around the NorthClackamas Park site. While collecting data at the site, he noticed sedimentlayers in a retaining-wall excavation, and these layers appeared to be fold-ed. The deposits at this site are sand and silt layers left during the ice ageby the Missoula floods and are 15,000 to 12,800 years old. Hence anyfaulting or folding of these sediments would point to an active fault.

A subsequent two-day examination and logging (by DOGAMI and URSgeologists) of the sediments exposed in the retaining-wall face showed thatthe entire sequence of sediment layers is folded. We believe that this fold-ing is evidence for an active fault beneath the site and that the fault is ei-ther the Portland Hills fault or a closely related fault. Preliminary interpre-tation of the log for the cut shows that the southwest end is about 5 fthigher than the northeast end, and the whole sediment sequence has beenshortened about 3 ft. This suggests a total of 6 ft of movement. This is con-sistent with the assumption that, during the past 12,800 years, two earth-quakes may have occurred on the fault.

We considered other geologic processes that might have caused the ap-pearance of folded layers but concluded that they could not produce thearrangement we observed. We intend to use more geophysical studies atthe site to confirm the presence of a fault beneath the site. An example ofthis approach is an earlier high-resolution magnetic survey across the fault,which shows an abrupt jump in the magnetic field strength coincident withthe folded sediments. We believe this occurs because the depth to highlymagnetic volcanic bedrock changes abruptly across the fault.

(See illustrations on next two pages)

(Continued on page 50)


The magnetic grid map shows the location of magnetic datapoints (black diamonds) with respect to the cut and includes cul-tural features. The postion of the sharp anomaly occurs at the same location as the sharp fold in the cut. Points not included inthe grid were strongly affected by piles of rebar deployed along the cut for construction. The cross section of the magnetic gridshows a sharp and relatively high-amplitude (~700 gamma) anomaly.

The preliminary log of the cut shows a series of fine sand layers deposited by the late Pleistocene Missoula Floods. Each layer, orrhythmite, represents a single flood event, and the rhythmites are typically separated by paleosols, which are marked by accumula-tions of clay, iron oxides, and root casts. The log shows a stack of at least seven rhythmites, all of which appear to be folded in thesame form. Some channeling is evident in the form of truncated paleosols but is generally restricted to one or two rhythmites.

NE

NE

SW

SW

The fault at Rowe Middle School

These two pages show, far left, a lo-cation map of the new discovery site;on this page, top, a magnetic grid mapand graph of the cut that revealed thenew information; and, below this text,a description of horizons for the pre-liminary log of the cut, split across bothpages at the very bottom.


ACKNOWLEDGMENTS

Our thanks to Silvio Pezzopane,Rick Blakely, Ray Wells, and TomPratt of the U.S. Geological Surveyfor their generous sharing of ideasand information, and to MarilynMackel, Maria Gallegos, and FumikoGoss for their assistance in thepreparation of this paper. The paperbenefited from reviews by RickBlakely, Tom Pratt, and Susan Olig.

REFERENCES CITEDBalsillie, J.J., and Benson, G.T., 1971, Evidence

for the Portland Hills fault: Oregon De-partment of Geology and Mineral Indus-tries, Ore Bin, v. 33, p. 109–118.

Beeson, M.H., Fecht, K.R., Reidel, S.P., andTolan, T.L., 1985, Regional correlationswithin the Frenchman Springs Member ofthe Columbia River Basalt Group: New in-sights into the middle Miocene tectonicsof northwestern Oregon: Oregon Geolo-gy, v. 47, no. 8, p. 87–96.

Beeson, M.H., Johnson, A.G., and Moran,M.R., 1975, Portland environmental geol-ogy—fault identification (Final technicalreport to U.S. Geological Survey, in coop-eration with Oregon Department of Geol-ogy and Mineral Industries): Portland,Oreg., Portland State University, Depart-ment of Earth Sciences, 107 p.

Beeson, M.H., Tolan, T.L., and Anderson, J.L.,1989, The Columbia River Basalt Group inwestern Oregon; geologic structures andother factors that controlled flow emplace-ment patterns, in Reidel, S.P., and Hooper,P.R., eds., Volcanism and tectonism in theColumbia River flood-basalt province: Ge-ological Society of America Special Paper239, p. 223–246.

Beeson, M.H., Tolan, T.L., and Madin, I.P.,1991, Geologic map of the Portland quad-rangle, Multnomah and WashingtonCounties, Oregon, and Clark County,Washington: Oregon Department of Geol-ogy and Mineral Industries GeologicalMap Series GMS–75, 1:24,000.

Blakely, R.J., Wells, R.E., Tolan, T.L., Beeson,M.H., Trehu, A.M., and Liberty, L.M.,2000, New aeromagnetic data reveal largestrike-slip(?) faults in the northernWillamette Valley, Oregon: Geological So-ciety of America Bulletin, v. 112, no. 8, p.1225–1233.

Blakely, R.J., Wells, R.E., Yelin, T.S., Madin, I.P.,and Beeson, M.H., 1995, Tectonic settingof the Portland-Vancouver area, Oregonand Washington: Constraints from low-al-

titude aeromagnetic data: Geological Soci-ety of America Bulletin, v. 107, no. 9, p.1051–1062.

Bott, J.D.J., and Wong, I.G., 1993, Historicalearthquakes in and around Portland, Ore-gon: Oregon Geology, v. 55, no. 5, p.116–122.

Bucknam, R.C., Hemphill-Haley, E., andLeopold, E.B., 1992, Abrupt uplift withinthe past 1,700 years at southern PugetSound, Washington: Science, v. 258, p.1611–1614.

Geomatrix Consultants, Inc., 1993, Seismicmargin earthquake for the Trojan site:Final report prepared for Portland GeneralElectric Trojan Nuclear Plant, Rainier, Ore-gon, Project no. 1348, var. pag.

Geomatrix Consultants, Inc., 1995, Seismicdesign mapping, State of Oregon: Final re-port to Oregon Department of Transporta-tion, Project no. 2442, var. pag.

Kafka, A.L., and Levin, S.Z., 2000, Does thespatial distribution of smaller earthquakesdelineate areas where larger earthquakesare likely to occur?: Seismological Societyof America Bulletin, v. 90, no. 3,p. 724–738.

Madin, I.P., 1990, Earthquake-hazard geologymaps of the Portland metropolitan area,Oregon: Oregon Department of Geologyand Mineral Industries Open-File ReportO–90–2, 21 p., 8 maps, 1:24,000.

Madin, I.P., Priest, G.R., Mabey, M.A., Mal-one, S., Yelin, T.S., and Meier, D., 1993,March 23, 1993, Scotts Mills earthquake-western Oregon's wake-up call: OregonGeology, v. 55, no. 3, p. 51–57.

Pratt, T.L. Odum, J., Stephenson, W., Williams,R., Dadisman, S., Holmes, M., and Haug,B., 2001, Late Pleistocene and Holocenetectonics of the Portland basin, Oregonand Washington, from high-resolutionseismic profiling: Seismological Society ofAmerica Bulletin (in press).

Schlicker, H.G., Deacon, R.J., and Twelker,N.H., 1964, Earthquake geology of thePortland area, Oregon: Oregon Depart-ment of Geology and Mineral Industries,Ore Bin, v. 26, no. 12, p. 209–230.

Thomas, G.C., Crosson, R.S., Carver, D.L., andYelin, T.S., 1996, The 25 March 1993Scotts Mills, Oregon, earthquake and af-tershock sequence: Spatial distribution,focal mechanisms, and the Mount Angelfault: Seismological Society of AmericaBulletin, v. 86, no. 4, p. 925–935.

Unruh, J.R., Wong, I.G., Bott, J.D., Silva, W.J.,and Lettis, W.R., 1994, Seismotectonicevaluation of Scoggins Dam, Tualatin Pro-ject, Northwestern Oregon: Unpublishedreport prepared for U.S. Bureau of Recla-mation, 206 p.

Wang, Z., and Madin, I.P., 2001, MountAngel fault characterization using SH-wave refraction and reflection methods(final technical report): Unpublished reportprepared for the U.S. Geological Survey,USGS Award No. 00HQGR0009.

Wells, D.L., and Coppersmith, K.J., 1994,New empirical relationships among magni-tude, rupture length, rupture width, rup-ture area, and surface displacement: Seis-mological Society of America Bulletin, v.84, no. 4, p. 974–1002.

Wells, R.E., Weaver, C.S., and Blakely, R.J.,1998, Forearc migration in Cascadia andits neotectonic significance: Geology, v.26, no. 8, p. 759–762.

Wong, I.G., 1997, The historical earthquakerecord in the Pacific Northwest: Applica-tions and implications to seismic hazardassessment, in Wang, Y., and Neuendorf,K.K.E., eds., Earthquakes—Converging atCascadia: Oregon Department of Geologyand Mineral Industries Special Paper28/Association of Engineering GeologistsSpecial Publication 10, p. 19–36.

Wong, I.G., and Bott, J.D.J, 1995, A look backat Oregon's earthquake history,1841–1994: Oregon Geology, v. 57, no.6, p. 125–139.

Wong, I.G. and Silva, W.J., 1998, Earthquakeground shaking hazards in the Portlandand Seattle metropolitan areas, in Dak-oulas,P., Yegian, M., and Holtz, R.D.,(eds.), Geotechnical earthquake engineer-ing and soil dynamics III: American Societyof Civil Engineers Geotechnical SpecialPublication 75, v. 1, p. 66–78.

Wong, I.G., Silva, W., Bott, J., Wright, D.,Thomas, P., Gregor, N., Li, S., Mabey, M.,Sojourner, A., and Wang, Y., 2000a, Earth-quake scenario ground shaking map forthe Portland, Oregon, metropolitan area.Portland Hills fault M 6.8 earthquake,peak horizontal acceleration (g) at theground surface: Oregon Department ofGeology and Mineral Industries Interpre-tive Map Series IMS–15, scale 1:62,500.

Wong, I.G., Silva, W., Bott, J., Wright, D.,Thomas, P., Gregor, N., Li, S., Mabey, M.,Sojourner, A., and Wang, Y., 2000b, Earth-quake scenario and probabilistic groundshaking maps for the Portland, Oregon,metropolitan area: Oregon Department ofGeology and Mineral Industries Interpre-tive Map Series IMS–16, 16 p., 11 sheets,scale 1:62,500.

Yelin, T.S., and Patton, H.J., 1991, Seismotec-tonics of the Portland, Oregon, region:Seismological Society of America Bulletin,v. 81, no. 1, p. 109–130.

(Continued from page 47)

OREGON GEOLOGYtaylors/g473/seismic_hazards/... · 2009-04-18 · 40 OREGON GEOLOGY, VOLUME 63,...

Documents

Transcript of OREGON GEOLOGYtaylors/g473/seismic_hazards/... · 2009-04-18 · 40 OREGON GEOLOGY, VOLUME 63,...