Deep-Water Turbidites as Holocene San Andres Fault

7/31/2019 Deep-Water Turbidites as Holocene San Andres Fault

1/26

Key words paleoseismology earthquake submarine recurrence patterns submarine landslides turbid flows

1. Introduction

The study of plate boundary fault systemshas been revolutionized by two relatively recentsub-disciplines: paleoseismology and crustalmotion as measured by the Global PositioningSystem. GPS technology now makes it possibleto measure crustal strain accumulation at plateboundaries with a high degree of certainty in

1169

ANNALS OF GEOPHYSICS, VOL. 46, N. 5, October 2003

Deep-water turbidites as Holoceneearthquake proxies: the Cascadia

subduction zone and NorthernSan Andreas Fault systems

Chris Goldfinger (1), C. Hans Nelson (2) (*),Joel E. Johnson (1) and the Shipboard Scientific Party(1) College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon, U.S.A.

(2) Department of Oceanography, Texas A & M University, College Station, Texas, U.S.A.

AbstractNew stratigraphic evidence from the Cascadia margin demonstrates that 13 earthquakes ruptured the margin fromVancouver Island to at least the California border following the catastrophic eruption of Mount Mazama. These 13 eventshave occurred with an average repeat time of 600 years since the first post-Mazama event 7500 years ago. The youngestevent 300 years ago probably coincides with widespread evidence of coastal subsidence and tsunami inundation in buriedmarshes along the Cascadia coast. We can extend the Holocene record to at least 9850 years, during which 18 events cor-relate along the same region. The pattern of repeat times is consistent with the pattern observed at most (but not all) local-ities onshore, strengthening the contention that both were produced by plate-wide earthquakes. We also observe that thesequence of Holocene events in Cascadia may contain a repeating pattern, a tantalizing look at what may be the long-termbehavior of a major fault system. Over the last 7500 years, the pattern appears to have repeated at least three times, withthe most recent A.D. 1700 event being the third of three events following a long interval of 845 years between events T4and T5. This long interval is one that is also recognized in many of the coastal records, and may serve as an anchor pointbetween the offshore and onshore records. Similar stratigraphic records are found in two piston cores and one box corefrom Noyo Channel, adjacent to the Northern San Andreas Fault, which show a cyclic record of turbidite beds, with thir-ty-one turbidite beds above a Holocene/.Pleistocene faunal datum. Thus far, we have determined ages for 20 eventsincluding the uppermost 5 events from these cores. The uppermost event returns a modern age, which we interpret islikely the 1906 San Andreas earthquake. The penultimate event returns an intercept age of A.D. 1664 (2 range 1505-1822). The third event and fourth event are lumped together, as there is no hemipelagic sediment between them. The ageof this event is A.D. 1524 (1445-1664), though we are not certain whether this event represents one event or two. The fifthevent age is A.D. 1204 (1057-1319), and the sixth event age is A.D. 1049 (981-1188). These results are in relatively goodagreement with the onshore work to date,which indicates an age for the penultimate event in the mid-1600s, the most like-ly age for the third event of 1500-1600, and a fourth event 1300. We presently do not have the spatial sampling need-ed to test for synchroneity of events along the Northern San Andreas, and thus cannot determine with confidence that theobserved turbidite record is earthquake generated. However, the good agreement in number of events between the onshoreand offshore records suggests that, as in Cascadia, turbidite triggers other than earthquakes appear not to have added sig-nificantly to the turbidite record along the northernmost San Andreas margin during the last 2000 years.

Mailing address: Dr. Chris Goldfinger, College of Oceanicand Atmospheric Sciences, Oregon State University, Marine Ge-ology Active Tectonics Group, Ocean Admin. Bldg. 104, Corval-lis, Oregon 97331, U.S.A.; e-mail: [email protected]

(*)Now at : Instituto Andaluz de Ciencias de la Tierra,CSIC, Universidad de Granada, Campus de Fuente Nuevas/n, Granada 18071, Spain.


2/26

only a few years. However, real-time strainmeasurements can only represent a fraction ofone strain cycle. Fundamental questions such asthe utility of the seismic gap hypothesis, cluster-ing, and the applicability of slip-predictable ortime-predictable models remain unansweredbecause we rarely have a long enough earth-quake record. Characteristic earthquake modelsassume that stress buildup is proportional to thetime since the last earthquake. The seismic gaphypothesis follows directly from this assump-tion, and is the basis for probabilistic predictions

of seismicity. Recently, characteristic earthquakemodels and their derivatives have been chal-lenged by new models of stress-triggering andfault interaction. In these models, strain rechargefollowing an earthquake is supplied only indi-rectly by the underlying motion of the plates,and the stress on each fault segment is controlledby the action and history of the surrounding seg-ments (e.g. Stein et al., 1992; Ward and Goes,1993). The applicability of these models maywell be system dependent, may differ betweenplate boundary versus crustal faults, or dependon magnitude. What is needed most to addressthese questions is data on spatial and temporal

earthquake recurrence for more fault systemsand over longer spans of time, so that meaning-ful statistical conclusions can be drawn.

Paleoseismology can address these questionson active continental margins. Two methodolo-gies used in recent years are coastal paleoseis-mology and investigation of the turbidite eventhistory. Unlike more typical paleoseismology onland, neither technique uses fault outcropsbecause the faults are entirely offshore, and bothtechniques must demonstrate that the events theyare investigating are uniquely generated byearthquakes and not some other natural phenom-enon. Nevertheless, these problems can be over-

come, and both techniques can be powerful toolsfor deciphering the earthquake history along anactive continental margin. The turbidite recordhas the potential to extend and earthquake histo-ry farther back in time than coastal records, 10.000 years, more than enough to encompassmany earthquake cycles. In recent years, tur-bidite paleoseismology has been attempted inCascadia (Adams, 1990; Goldfinger and Nelson,1999; Goldfingeret al., 2003; Nelson et al., 2003),

Puget Sound (Karlin and Abella, 1992), Japan(Inouchi et al., 1996), the Mediterranean(Kastens, 1984; Anastasakis and Piper, 1991;Nelson C.H. et al., 1995), the Dead Sea (Niemiand Ben-Avraham, 1994), Northern California(Field, 1984) and even the Arctic ocean (Grantzet al., 1996), and is a technique that is evolvingas a precise tool for seismotectonics.

We have been investigating the use of turbi-dites as paleo-earthquake proxies at the Cascadiasubduction zone and along the Northern SanAndreas Fault. A growing body of evidence from

both of these margin settings suggests that withfavorable physiography and sedimentologicalconditions, turbidites can be used as earthquakeproxies with careful spatial and temporal correla-tions.

New stratigraphic evidence from the Ca-scadia margin confirms that 13 turbidites weredeposited in major channel systems along themargin from Vancouver Island to at least theCalifornia border (Adams, 1990) following thecatastrophic eruption of Mount Mazama.Nelson et al. (2003) have interpreted these tur-bidites as having been triggered by great sub-duction earthquakes. These 13 events have

occurred with an average repeat time of 600years since the first post-Mazama event 7500years ago (fig. 1a,b), and the youngest event 300 years ago approximately coincides withwidespread evidence of coastal subsidence andtsunami inundation in buried marshes alongthe Cascadia coast. In this paper we furtherdevelop the earthquake hypothesis for the ori-gin of Cascadia turbidites, and extend the Ho-locene turbidite record to 9850 years, duringwhich 18 events may correlate along the sameregion. We also observe that the sequence ofHolocene events in Cascadia appears to containa repeating pattern of events, a tantalizing first

look at what may be the long-term behavior ofa major fault system.

On the Northern California margin, we col-lected two piston cores and one box core fromNoyo Channel, draining the Northern Califor-nia continental margin adjacent to the San An-dreas Fault. Like Cascadia, these cores show acyclic record of turbidite beds that may repre-sent Holocene earthquakes on the northern seg-ment of the San Andreas Fault. Here we present

1170

Chris Goldfinger, C. Hans Nelson, Joel E. Johnson and the Shipboard Scientific Party


3/26

1171

Deep-water turbidites as Holocene earthquake proxies: the Cascadia subduction zone and Northern San Andreas Fault systems

(?)

Nitin

atFa

n

Astoria Fan

-130 -129 -128 -127 -126 -125 -124 -123 -122 -121 -120 -119 -118

39

40

41

42

43

44

45

46

47

48

JUAN DEFUCAPLATE

PACIFIC PLATE

NORTHAMERICAN

PLATE

Barkley Canyon

Juan de Fucauan de FucaCanyon

Astoria Canyonoria CanyonWillapaCanyon

Rogue CanyonogueCanyon

Klamath Canyon

TrinidadCanyon

Eel Canyon

Grays Canyon

Quinault Canyon

BaseofSlopeChann

el

Astoria

Channel

13/18

13

13/18

13/18

(?)

31

GORDAPLATE

BlancoFractureZone

Gorda

Rid

ge

Tufts AbyssalPlain

)

Mendocino Escarpment

-200

-3100

-2700

Washington

VancouverIsland

(?)

(?)(?)

Rogue Channel

PC30PC31

TC30

TC31

KC55

Cascadia Channel

PC22

PC23

PC25

BC24

Juan De Fuca Canyon

PC11

PC12

TC11

TC12

CasadiaChannel

ColumbiaRiv

er

Smith Canyon

67

40mm/yrJDF-NAM

SanAndreasFault

Noyo Canyon

Mendocino Canyon

California

Oregon Astoria ChannelBC04

(Location of fig. 4)

SixesR

iver

Bradley Lk.

13?/18?

200

Submarine canyons and channels

N

0

Contour interval 100 m

km

1999 Core locations

13/18 Post-Mazama Holocene turbidite events

Juan

De

Fuca

Rid

ge

Fig. 1a,b. a) Cascadia margin turbidite canyons, channels and 1999 core locations. Major canyon systems andthe number of post-Mazama and/or Holocene turbidites are shown in red. Mazama ash was not present inBarkley Canyon cores or in the cores south of Rogue Canyon. Area of fig. 4 shown by box at Astoria Canyon.b) Synchronicity test at a channel confluence as applied where Washington channels merge into the CascadiaDeep Sea Channel. The number of events downstream should be the sum of events in the tributaries, unless theturbidity currents were triggered simultaneously, as by an earthquake.

a

b


4/26

new evidence that these turbidites may repre-sent an earthquake record of the Northern SanAndreas system. In total, thirty-one turbiditebeds are found above the Holocene/Pleistocene( 9 ka) boundary yielding an average recur-rence time of 300 years. The ages of theyoungest five Northern San Andreas events arein reasonably good agreement with onshorepaleoseismic event ages and repeat times. Whilewe presently cannot unequivocally demonstratethat these deposits are earthquake generated, theage similarities and similar number of onshore

and offshore events suggests this origin.

2. Cascadia paleoseismic history

The past occurrence of great earthquakes inCascadia is now well established in Cascadiacoastal marsh stratigraphy (e.g. Nelson A.R.et al., 1995; Atwater and Hemphill-Haley,1997; Clague, 1997; Kelsey et al., 2002), andtsunami records in Japan (Satake et al., 1996).More recently, the turbidite event record isbeing investigated as a long-term proxy forHolocene great earthquakes (Nelson et al.,

2003). Consequently, attention has turned tomagnitude, recurrence intervals, and segmenta-tion of the margin. Segmented and whole mar-gin ruptures should leave distinctly differentstratigraphic records in both the coastal marsh-es and the offshore turbidite record, howeverthe difficulties of precisely dating stratigraphicevidence of past earthquakes makes unravelingthe paleoearthquake record a challenging prob-lem.

3. The coastal record

Paleoseismic evidence on land is found in theform of subsided marsh surfaces and tsunamirunup or washover deposits of thin marine sandlayers with diatoms that are interbedded withinestuarine or lake muds (Atwater, 1987, 1992;Darienzo and Peterson, 1990; Atwater et al.,1995; Hemphill-Haley, 1995; Nelson et al., 1995;Hutchinson et al., 2000; Kelsey et al., 2002;Nelson et al., 2000). The tsunami deposits arefound several kilometers inland from the coast,

up river estuaries, or in low-lying freshwaterlakes near sea level, but above the reach ofstorm surges. A 3500-year record of such tsuna-mi events is encountered in Willapa Bay, Wa-shington (Atwater and Hemphill-Haley, 1997).A 7300-year record is under investigation inBradley Lake, Oregon (Nelson et al., 2000; Ol-lerhead et al., 2001). A 5500-year record is alsobeing studied in Sixes River estuary (Kelsey etal., 1998, 2002). Time variance for recurrenceintervals of the Willapa Bay tsunami events is asmuch as 990 years or as little as 140 years

between events (Atwater and Hemphill-Haley,1997). A total of 17 tsunami events have beenrecognized in the Bradley Lake sediments, wherethe maximum (900 years) and minimum (100years) time variance are similar to that found inWillapa Bay (Hemphill-Haley et al., 2000). TheBradley Lake tsunami events are presumed to belocal and not distant tsunami on the basis of theheight requirement to wash over the spit that sep-arates the lake from the sea. The Sixes Riverrecord of 11 events in 5500 years is found 110km south of Bradley Lake, and also exhibits awide variance in time between subduction zoneearthquake events from 70 years to 900 years

(Kelsey et al., 1998, 2002). In addition to thesimilar range in time variance between recur-rence of events ( 100 to 1000 years) in theselong-term coastal records, the average recurrenceinterval is similar for the Willapa and Sixes Riverestuaries (500-540 years) and slightly less forBradley Lake (440 years) (Kelsey et al., 1998,2002; Nelson et al., 2000). Kelsey et al. (1998,2002) find that the Willapa and Sixes estuarypaleoseismic events are comparable for 2400 to3500 years BP, differ between 2400 to 800 yearsBP, and both include the 1700 A.D. event (Jacobyet al., 1997; Yamaguchi et al., 1997). The recordsof repeated subsidence and tsunami inundation

followed by emergence are most likely paleoseis-mic events because multiple soils buried by estu-ary muds show evidence of rapid subsidence,followed immediately by incursion of tsunamisands with marine diatoms over the wetland soilsurface. Some are associated with liquefactionfeatures (Atwater and Hemphill-Haley, 1997;Kelsey et al., 1998, 2002). The Bradley Lakerecord is based only on tsunami deposited sands,exhibits a relatively greater number of events per

1172



5/26

unit time than the estuary records, and has 12events requiring a tsunami inundation of> 5.5 m.

The most abundant high precision age dataare available for the most recent subsidenceevent, which probably occurred within a fewdecades of 1700 A.D. ~ 300 years ago. Dendro-chronology of western red cedar in Washingtonand other Northern Oregon estuaries show treedeath occurred between August, 1699 and May1700 (Jacoby et al., 1997; Yamaguchi et al.,1997), and was most likely due to salt waterincursion due to subsidence. The age of this

event is supported by evidence of a far-fieldtsunami in Japan on January 26, 1700 A.D.,which has been attributed to a subduction earth-quake on the Cascadia subduction zone (Satakeet al., 1996). The ~ 300 year event is wide-spread, with evidence found from NorthernCalifornia to Vancouver Island (Clague, 1997).For older events, error bars for numerical agesare significantly larger and the difficulty inidentifying anomalous local subsidence eventsincreases. The land record of submergence, treedeath, and tsunami, as well as the turbiditerecord offshore all require correlation alongstrike to determine whether events represent

whole margin or segmented rupture. In the caseof the land record, earthquake origin has beenquite well established, but correlation remainselusive. Conversely, with the offshore record,the consistency of the records and several keytests of synchronicity demonstrates a probableconsistent correlation, while demonstratingearthquake origin is more difficult.

4. Turbidite correlation and dating methods

The use of turbidite deposits as paleoearth-quake proxies is, like other forms of paleoseis-

mology, dependant on as complete as possiblean understanding of the depositional environ-ment, physiography, and dynamics of emplace-ment of the stratigraphic record. In this section,we outline some of the methods we and othersare developing to test the viability of this proxyrecord.

Channel pathway analysis Turbidity cur-rents result from channel wall slumps, mostly

of unconsolidated material, in the upper reach-es of canyons. Large slumps of older blockymaterial into a channel may attenuate or deflectturbidity currents that might otherwise travel tocore sites on the abyssal plain. We analyze thethalweg profiles of the channel systems, alongwith the swath bathymetry along the channelwalls to search for such slumps that may haveblocked or diverted turbid flows, thus biasingthe stratigraphic record downstream (fig. 2).Another possible influence is folding and fault-ing. Griggs and Kulm (1970) showed that major

turbidity currents are up to 100 m high and 17km wide in Cascadia Basin. Thus faulting andfolding at reasonable slip rates, and even mod-erate slumping are perhaps unlikely sources ofdisruption of channel flow, though large slidesand faults with high slip rates might be capableof biasing the downstream record.

Core siting In our Cascadia work, wefound that surface morphology in the channelswas a critical factor in choosing exact sample

1173


Distance (m)

0 20 000 100 000 120000

Depth

(m)

0

500

1000

1500

2000

3000

Axial gradient profile for Astoria Canyon and Main Channel

2500

40000 80 00060000 140 000

Slump

Fault

Distance (m)

Depth(m)

0 40000 80000

2000

2200

2400

2600

2800

3200

Axial gradient profile for base of Slope Channel

3000

20 000 60 000 100000

Mazamaturbiditepresent

Mazamaturbiditeabsent

Fig. 2. Example of channel blocking by a large sub-marine slide along one part of the Astoria fan system.This channel lies along the base of the continental slope.


6/26

location. In a gross sense, distance from thesource and distance from the channel thalwegcontrols grain size. In detail, we found that theturbidite record was very robust and best pre-served in thalwegs, on point bars, in the lee ofpoint bars, and on low terraces above the thal-weg. In general, we try to sample each systemboth down channel, and across major channelsin order to capture a complete event record andtest for sensitivity in both directions. The qual-ity of the record, and our ability to capture theHolocene interval was, however, sensitive to

local conditions. After some trial and error we

found that sampling in the lee of point bars wasa good strategy in proximal locations (fig. 3).These locations included the complete record,did not include very coarse material, and tend-ed to have a somewhat expanded record, whichincreases dating precision by reducing the sam-ple interval needed to get enough forams for14Cdating. In more distal localities, we sampledlocal sediment pools to find an expanded recordin areas that otherwise had lower than optimalsedimentation rates. In any given setting sometrial and error is required to choose sites that

offer optimal conditions, an expanded record,

1174


Vizcaino Channel

Noyo Channel

AbandonedChannelHeads

ActiveChannel

Heads

Mendocino Triple Junction

N

San Andreas Fault

Fig. 3. GLORIA sidescan image overlaid on SeaBeam bathymetry showing the truncation of channel heads ofNoyo and Vizcaino channels (dashed line) and the active channels of the Delgada Fan, Northern California mar-gin. The 90 bend in the head of Noyo Canyon at the bottom of the image (.just above the N on the north arrow)has been controlled by the SAF. Abandoned channels appear to result from dextral SAF motion. Location shownon fig. 9. Fault location from Castillo and Ellsworth (1993) and interpretation in this study.


7/26

and a Holocene section that can be cored withexisting gear. Backscatter data, where available,

were used together with bathymetric data toreveal depositional patterns, relative grain sizedistribution, and to determine the most activechannels (fig. 4).

Age Control AMS 14C ages are determin-ed from planktonic forams deposited in thehemipelagic sediments that underlie each tur-bidite sequence. Benthic forams could be re-worked, as could previously deposited plank-tonic forams entrained in the turbidite deposit,thus care is taken to select only hemipelagicplanktonic forams. Benthic forams may be un-reliable and give erratic ages, possibly due to

the unknown deep-water reservoir correction(M. Kashgarian, Lawrence Livermore Lab.,2000). Our foram assemblages were dominatedby Turborotalia pachyderma (sinistral and dex-tral) and Globigerina bulloides, but used all(~ 20) planktonic species in most cases (D.Boettcher, MicroPaleo Associates, written re-port, 2003). Previous work has shown that in midto distal parts of the Cascadia Channel systems,there is little evidence of erosive turbidite em-

placement, thus sampling in the hemipelagicinterval immediately below the coarse tur-

bidite base is used to date events. Our ration-ale is to collect samples from the youngesthemipelagic interval between events to datethe overlying event. Though there may besome erosion of this interval, thus biasingevents to slightly older ages, we use more dis-tal cores, which are less likely to have basalerosion for primary dating, and inspect thebases visually for evidence of erosion.Although bioturbation blurs the age of the sur-face 10 cm or so, there is some evidence thatlarge grains such as forams are not moved ver-tically within the sediment as much as the sed-iment is bioturbated, reducing this problem

somewhat (Thomson and Weaver, 1994). Thecores are examined for organic matter andmica, which is characteristic of turbidite tailsin this area, and samples are taken once theupper and lower limits of the turbidites arefound. The hemipelagic fraction is also distin-guished by greater bioturbation (e.g. Nelson,1968; Griggs et al., 1969).

Because planktonic forams can be reworkedwhen they are entrained in the turbidite deposit,

1175


Fig. 4. 100 m resolution shaded swath bathymetry of a portion of Astoria Canyon, Northern Oregon. The sit-ing of turbidite cores at sea is facilitated by using 3D visualization of the canyon/channel morphology. To obtainthe most complete turbidite record we avoid erosive channel thalwegs and coarse grained frontal point bars. Wefound non-erosive complete records in the lee of point bars, such as this site for core M9907-04BC.


8/26

we sieve the > 0.062 mm sand fraction andcarefully hand-pick planktonic forams onlyfrom the minimum interval needed to acquireenough forams (1-4 cm) of hemipelagic sedi-ment below the base of the turbidite and aboveany underlying turbidite tail. Raw AMS radio-carbon ages are reservoir corrected and con-verted to calendar years (cal years) by themethod of Stuvier and Braziunas (1993).Though there may be a time-varying surfacereservoir age (Southon et al., 1990), this has notyet been demonstrated. In Cascadia Channel,

our AMS intercept age for the youngest event is304 years BP, for the known 1700 Cascadiaevent. Additionally, the first post-Mazama tur-bidite event also is within 150 years of themost current Mt. Mazama eruption age (7627 150; Zdanowicz et al., 1999). Both these agesindicate the reservoir corrections are approxi-mately correct.

Event correlation and recurrence intervalsWe address the mean interval between turbiditeevents and variability of this mean by two semi-independent methods: 1) directly dating theevents, and 2) detailed analysis of the hemi-

pelagic sediment interval between turbidite e-vents in selected new cores. This requires de-tailed analysis of coarse fraction constituents toseparate the turbidite tail muds from the incep-tion of the hemipelagic sediment. We analyzedall cores using the OSU Geotech MST system,collecting gamma (GRAPE) density, P-wavevelocity and magnetic susceptibility series foreach core. In order to establish a completechronology at a given site, correlation betweenthe piston and trigger cores, box cores, and some-times multiple cores at the same site is need-ed. This is because the tops of cores are some-times lost (most common with piston cores),

and separation of core segments may also oc-cur in the piston core. Correlation between mul-tiple cores aided by the MST data is essential inthese cases. We also utilize x-radiographs toimage details of the individual turbidites,search for multiple coarse pulses, or multipleevents, and identify the top of the turbidite tails.The x-radiographs are proving invaluable, par-ticularly in the San Andreas cores where multi-ple pulses are common.

5. The Cascadia turbidite record

Concurrent with the discovery of the firstburied marsh sequences on land, Adams (1985,1990) assessed the possibility that turbidites inchannels of Cascadia Basin contained a recordof great earthquakes along the Cascadia margin.He examined core logs for the Cascadia Basinchannels, and determined that many of themhad between 13 and 19 turbidites overlying theMazama ash datum. In particular, he found thatthree cores along the length of Cascadia chan-

nel contain 13 turbidites and argued that these13 turbidites correlate along the channel (as didGriggs and Kulm, 1970). Adams observed thatcores from Juan de Fuca Canyon, and belowthe confluence of Willapa, Grays, and Quinaultcanyons, contain 14-16 turbidites above theMazama ash. The correlative turbidites inCascadia Channel lie downstream of the con-fluence of these channels. If these events hadbeen independently triggered events with morethan a few hours separation in time, the chan-nels below the confluence should contain from26-31 turbidites, not 13 as observed. The im-portance of this simple observation is that it

demonstrates synchronous triggering of turbi-dite events in channel tributaries, the head-waters of which are separated by 50-150 km.Similar inferences about regionally triggeredsynchronous turbidites in separate channels arereported in Pilkey (1988). The extra turbiditesin the upstream channels may be the result ofsmaller events. This synchronicity test is apowerful relative dating technique that is com-pletely independent of radiocarbon dating,which rarely has the precision to correlateevents.

In July, 1999, we collected 44 (4 diam.) pi-ston cores, 44 companion trigger cores (also

4) and eight 30 cm box cores in every majorcanyon/channel system from the northern limitof the Cascadia subduction zone near theNootka Fault, to Cape Mendocino at its south-ern terminus (fig. 1a; Goldfinger and Nelson,1999; Nelson and Goldfinger, 1999; Goldfingeret al., 2003). Cores were run through the MSTscanner as whole rounds to collect density,velocity, and magnetic susceptibility data, thensplit, photographed and described on board.

1176



9/26

We find that thirteen post-Mazama eventsare found along 600 km of the margin in theCascadia, Willapa, Grays, Astoria, Juan deFuca, and Rogue Canyon/Channel systems(figs. 1 and 5 to 7; Nelson et al., 2003). Thoughthe Holocene-Pleistocene boundary is a bit lesscertain as a datum, these same channels alsohave 18 post-datum events, extending this rec-ord to ~ 10.000 years. In previously existingcores, we found only 3 post-Mazama events inmiddle and lower Astoria Channel, which ap-

peared to contradict Adams (1990) hypothesisfor 13 events. In 1999 cores, we find a progres-sive loss of turbidites from 13 to 10 to 7 to 6 to5 events at each successive downstream chan-nel splay in the distributary upper Astoria Fan.This down-channel loss of events resulting inonly 3 events in the mid-lower Astoria Channelexplains the previous contradiction, and showsthat the post-Mazama turbidite record is consis-tent along the margin.

6. Mechanisms for triggering of turbid flows

Adams (1990) made a convincing case forsynchroneity of Cascadia margin events. Butare these events all triggered by earthquakes?Adams suggested four plausible mechanismsfor turbid flow generation: 1) storm wave load-ing; 2) great earthquakes; 3) tsunamis, and 4)sediment loading. To these we add 5) nearbycrustal earthquakes, 6) in-slab earthquakes, 7)aseismic accretionary wedge slip, 8) hyperpyc-

nal flow, and 9) gas hydrate destabilization. Allof these mechanisms may trigger individualsubmarine slides and or turbid flow events, buthow can earthquake-triggered events be distin-guished from other events? Investigators haveattempted to distinguish seismically generatedturbidites from storm, tsunami, and otherdeposits. Nakajima and Kanai (2000) and Shikiet al. (2000) argue that seismo-turbidites can insome cases be distinguished sedimentolog-

1177


0

100

200

300 H

P

CORE 12PCJuan de Fuca

Channel

TurbiditeEvent No.Depth in

cm

4

12

5

7

89

1011

1314

15

1617

12

1920

3

18

21

6

400

500 486Missing

392(470-280)

4156(4316-4053)

2696(2743-2491)

1465(1555-1341)

607(652-522)

4865(5027-4812)

6353(6478-6264) 6638

(6375-6507)7282

(7403-7203)

972(1089-908)2310

(2360-2172)

3234(3343-3096)

4378(4491-4227)

7860(7955-7756)

8966(9300-8864)

9397/9220/9109(9755-9011)

9851(10290-9583)

10625(11154-10318)

11902/11835/11728(12292-11128)

TRANSITION ZONE

KEY TO CORES

Turbidite event number

Hemipelagic mud

Turbidite sand/silt bed

*607(652-522)

Cascadia MarginA.D. 1700 event255 BP (301-101)in Cascadia Channel

#

Mazama Ash

Fig. 5. Schematic core diagram for piston core M9907-12PC from Juan de Fuca Channel. Calibrated radiocarbonages are shown with 2 error ranges. AMS ages are obtained from planktonic forams in the hemipelagic mud beneath

events. Recurrence intervals between events are calculated and plotted in fig. 8. Location shown on fig. 1a.


10/26

ically. They observe that known seismicallyderived turbidites in the Japan Sea and LakeBiwa are distinguished by wide areal extent,multiple coarse fraction pulses and variable

provenance (from multiple or line sources), andgreater depositional volume than storm-gener-ated events. These investigators observe thatknown seismo-turbidites caused multiple slumpevents in many parts of a canyon system, gen-erating multiple pulses in an amalgamated tur-bidity current, some of which sampled differentlithologies that are separable in the turbiditedeposit. In the Japan Sea, the stacked depositsare deposited in order of travel time to their

lithologic sources, demonstrating synchronoustriggering of multiple parts of the canyon sys-tem. These turbidites are also complex, withreverse grading, cutouts, and multiple pulses.

Gorsline et al. (2000) make similar observa-tions regarding areal extent and volume for theSanta Monica and Alfonso Basins of theCalifornia borderland and Gulf of Californiarespectively. In general, these investigatorsobserve that known storm sediment surges arethinner, finer grained and have simple normallygraded Bouma sequences. We observe thatmany, but not all, of the turbidite sequences inCascadia are similar to the historically known

1178

Fig. 6. Summary core logs and digital photographs of all sections of Juan de Fuca core M9907-12PC. Turbiditesshown as T followed by numerals. See fig. 1a for location.



11/26

seismic turbidites reported by Nakajima andKanai (2000; their fig. 3).

While there may yet be applicable global,regional or local criteria to distinguish betweenturbidite triggers, these are at present poorlydeveloped. Thus far in Cascadia, we have notattempted to distinguish between triggeringmechanisms with sedimentological criteria, buthave used the spatial and temporal pattern ofevent correlations and Adams synchronicitytest at the confluence of Willapa, Juan de Fuca,and Cascadia channels to establish a regional

correlation that cannot be the result of triggersother than earthquakes. We confirm Adamsresults from the channel confluence of 13 post-Mazama events both above and below the con-fluence. Because turbidity currents deposit their

loads in a matter of hours, they are excellentrelative dating horizons, the relative age resolu-tion provided by this feature of turbidites is fargreater than any radiometric or other absolutetechnique. The synchronicity of event recordsestablished at the confluence effectively elimi-nates non-earthquake triggers because otherpossible mechanisms are extremely unlikely totrigger slides in separate canyons only a fewhours apart. This would have to have take place13 consecutive times to produce the core recordwe observe. The correlation is strengthened by

extending the record to 18 Holocene events atall sites between the Smith River and Juan deFuca Canyon. While this does not have thesame power as the confluence test, the similar-ity is striking, and very unlikely to have oc-

1179


Fig. 7. Summary core logs and digital photographs of all sections of Cascadia Channel core M9907-25PC. Seefig. 1a for location.


12/26

curred without a regional earthquake trigger.An alternative explanation could be a series ofearthquakes that took place within hours ofeach other along the length of the margin.While possible, the precision of the relative dat-ing provided by the time of settling of turbidflows restricts this to only a few hours time forCentral Washington. Additionally, whatever se-quence of events occurred, it would also have totake place in such a way that the stratigraphicrecord yields 18 events in all locations. Againthis is not impossible, but far less likely than

the simplest explanation, that of 18 plate wideevents. We are presently exploring other corre-lation techniques that may enable direct linkageof the Rogue and either Cascadia or Roguerecords, which do not have a confluence, andthus are not positively correlated.

We can make some additional observationsabout the applicability of other triggers thatmight be expected to generate turbid flows.Storm wave loading is a reasonable mechanismfor triggering of turbid flows, but is an unlike-ly trigger in Cascadia. On the Oregon andWashington margins, although deep-waterstorm waves are large, the canyon heads where

sediment accumulation occurs are at waterdepths of 150-400 m. These depths are at orbelow the maximum possible for disturbanceby storms with historical maximum signifi-cant wave heights of ~ 20 m, though rare megastorms cannot be ruled out. Tsunamis may alsoact as a regional trigger, however the 1964AlaskaMw 9.0 event did not trigger a turbiditeobserved in any of the cores, although it didserious damage along the Pacific coast (Adams,1990). Crustal or slab earthquakes could alsotrigger turbidites. To investigate this possibility,we resampled the location of a 1986 box core inMendocino Channel, where the uppermost event

is suspected to be the 1906 San Andreas event.Since 1986, the Mw 7.2 Petrolia earthquakeoccurred in 1992, either on the plate interface orlowermost accretionary wedge landward of thissite (Oppenheimeret al., 1993). Despite the epi-central distance of only a few kilometers fromthe canyon head, we were surprised to find nosurface sand in the 1999 box core, nor was itpresent at other Southern Cascadia Channel lo-cations. At least for this event, anMw 7.2 earth-

quake was not sufficient to trigger a turbid flow,or alternatively, insufficient sediment was pres-ent. Conversely, theMw = 6.9 Loma Prieta earth-quake apparently did trigger some type of turbidflow in Monterey Canyon at a much greater epi-central distance (Garfield et al., 1994).

The remarkable similarity of turbidite rec-ords in channels systems monitoring the north-ern 2/3 of the Cascadia margin suggests strong-ly that at least this part of Cascadia has experi-enced 13 post-Mazama events, and 18 Holo-cene events (fig. 1a,b). These events were suf-

ficiently large to both generate turbidites, andto correlate along the length of the margin.Further correlation is presently underwayusing radiocarbon dating, but will face difficul-ties with the accumulated errors and uncertain-ties. Nevertheless, the circumstantial case pre-sented by Adams (1990) is now considerablystronger.

South of the Rogue Canyon, the turbiditeevent frequency for the Northern Gorda plateregion may also contain 18 events in Smith andKlamath canyons. The Northern Californiachannels contain no Mazama ash, and have amore diffuse Holocene/.Pleistocene faunal

boundary, thus correlation will depend heavilyon radiocarbon dates. For the Southern Gordaarea the events are much more frequent. AMSages for events in the Southern Gorda regionindicate average turbidite recurrence intervalsof 133, 75, and 34 years respectively for theTrinidad, Eel, and Mendocino channels, in-creasing progressively toward the Mendocinotriple junction. The number of possible earth-quake sources for triggers also increases pro-gressively toward the triple junction, and in-cludes the Mendocino and Blanco faults, in-ternal Gorda plate faults, and perhaps theNorthern San Andreas. The potential for sedi-

mentological triggering is also higher, with theEel, Trinidad, and Mendocino Canyon headsvery close to the coast, increasing the likeli-hood of hyperpycnal flow input. Three earth-quakes of magnitude 6.9 to 7.4 have occurredin the past 21 years in the triple junction area.However, we estimate that the 10-14 cm ofhemipelagic sediment at the surface inMendocino Channel represents 50 to 70 ofdeposition, thus these three earthquakes have

1180



13/26

not triggered turbid flow events in MendocinoChannel.

Unlike the other Cascadia systems, both Eeland Mendocino canyon heads erode close to theshoreline where the canyon heads may intersectlittoral drift sediment that can be funneled down-canyon by storms. Given the probable mix ofinterplate, intraplate, and sedimentological (non-earthquake) events in the Southern Gorda region,we are as yet unable to make detailed inferencesabout the earthquake record in this area.

7. Turbidite recurrence interval data

We have plotted the turbidite event age ver-sus recurrence interval for the individual eventpairs available as of this writing (fig. 8). Theplot includes the best of our key cores thusfar, M9907-12PC from Juan De Fuca Canyon(figs. 5 and 6). This is the first of our cores tohave all Holocene turbidite events dated, withseveral latest Pleistocene events included, up toevent T21 (T19 is not included due to age rever-

1181


0

2000

4000

6000

8000

10000

12000

0 500 1000 1500 2000 2500

Recurrence (years)

Calibrated

14CA

ge

(BP

1950)

Juan de Fuca Canyon

Juan de Fuca Canyon

Cascadia Channel

Rogue Canyon

Rogue Canyon

Kelsey et al., (2002)Sixes River Estuary, OR

Atwater and Hemphill-Haley

(1997) Willapa Bay, WA

Garrison-Laneyet al. (2002)

Lagoon Creek, CA

Clarke and Carver (1992)

Humbolt Bay, CA

Nelson et al., (2000)

Bradley Lake, OR

Fig. 8. Summary plot of offshore and onshore calibrated 14C Ages and individual recurrence intervals. Partialresults from the turbidite record based on available radiocarbon ages. Multiple plots are shown where series inJuan de Fuca and Rogue channels is broken by a bad or missing date. Where three key sites are dated, note goodcorrelation between Juan de Fuca Canyon, Cascadia Channel, and Rogue Canyon at 4000-6000 years BP. Notegood agreement between Juan de Fuca Canyon and onshore records at 0-6000 years BP. Extra events inSouthern Oregon are indicated in the 3000-4000 BP range (also noted by Kelsey et al., 2002). For turbidite dates,intercept ages are plotted at the upper bound of each interval. Offshore turbidite data shown as solid lines, dashedlines show land data cited in legend. The pattern of events contains age offsets, both between land and offshoredata, and among the offshore cores. These are almost certainly due to spatial and temporal variations in watermass age during the Holocene, which is corrected here using the fixed values of Stuvier and Braziunas (1993).Resolution of time and space variant reservoir corrections is a work in progress.


14/26

sal for that event). The plot also includes partialrecords for the Cascadia and the Rogue sys-tems, for which complete AMS age results arenot yet available. The recurrence period versustime plot has some strengths, the principal onebeing that comparing cores uses the pattern ofrecurrence rather than the raw ages, which havenot been corrected for sedimentation rate, pos-sible basal erosion, and tuning of the 14C reser-voir correction. Each interval is arbitrarily plot-ted at the younger event age that delimits theinterval. Where an event is defined by multiple

ages due to multiple intercepts on the14

C cali-bration curve, we use the median age if there isno statistical preference.

The average post-Mazama recurrence timebased on the Juan de Fuca core is 581 years,with the shortest interval being 222 years (T8-T9), the longest 1488 years (T10-T11). Theaverage post-Mazama recurrence interval usingan age for T13 averaged from all channels is 600 years. The mean recurrence time basedon all available intervals is 564 years, whichis the Holocene average for 18 events terminat-ed by the A.D. 1700 event that occurred 250years before the 1950 reporting standard for cali-

brated radiocarbon ages (9841-250/17). Thisaverage is based on the average of three AMSages for event T18 from our key cores in Juan deFuca, Rogue, and Cascadia channels. These agesare 9849 (10287-9784), 9851 (10290-9583), and9824 (10274-9540) for Cascadia, Juan de Fuca,and Rogue channels respectively.

With a single complete, but uncorrected re-cord, the patterns may (or may not) be robust,but there does seem to be a repeating pattern ofa long interval ending in an earthquake, fol-lowed by a moderately long interval, then 1 or2 shorter intervals. Over the last 7500 years,the pattern appears to have repeated three times,

with the most recent A.D. 1700 event being thethird of three events following a long interval of845 years between events T4 and T5. This longinterval is one that is also recognized in manyof the coastal records, and may serve as ananchor point between the offshore and onshorerecords. The partial Rogue and Cascadia recordsshow a possible pattern correlation with thecomplete Juan De Fuca record for the overlap-ping events.

While it is temping to expound about earth-quake clustering and long-term fault behavior,we emphasize here that the analysis shown infig. 8 is incomplete, and confirmation of thispattern will require age data from the other keycores that is not yet available. We are encour-aged that despite occasional reversed ages andother problems inherent in paleoseismology,the extensive turbidite event record in Cas-cadia Basin will overcome these problems andusing pattern matching and correlation, willprovide a robust long-term paleoseismic histo-

ry.

8. Preliminary investigation of turbidite se-quences along the Northern Californiamargin

Following on the successful Cascadia effort,we are beginning an investigation of the Holo-cene rupture history of the Northern San An-dreas Fault using similar methods. The planincludes collection of a spatially extensive setof new piston cores in channel systems drainingthe adjacent continental margin from south of

San Francisco to the Mendocino Triple Junc-tion. This work will utilize a number of Keycores to develop the Holocene event record inall channel systems draining Northern Cali-fornia. Fortunately, we already have one of theseKey cores, M9907-49PC, collected in 1999as part of the Cascadia project as a control corein Noyo Channel (figs. 9 to12). We are pres-ently using this piston core and a companionbox core to develop an event history for thenorthern end of the San Andreas system.

The physiography of the margin is ideal forthis investigation. The Northern San Andreasfault has a single main fault strand (Castillo and

Ellsworth, 1993), and few other regional or localseismic triggers are available. The margin chan-nels are numerous, and offer excellent spatialsampling of the Holocene record along this im-portant fault system. The Northern San Andreasoffers an excellent opportunity to investigate thelong-term history of a relatively simple fault sys-tem, and relate that history to the current contro-versy over characteristic versus stress triggeringmodels of earthquake behavior.

1182



15/26

1183


Fig. 9. Map showing channel systems, onshore major faults, and location of 1999 core M9907-49PC on NoyoChannel. Summary of land paleoseismic data along the San Andreas is also shown. Inset shows age constraints forrecent events reported at land sites with turbidite event data (T-values and horizontal lines due to basal erosion, T3

cannot be dated). Horizontal dashed band in mid 1600s is the estimated age of the penultimate event from trenching(.T.) and dendrochronology (D) at Grizzly Flat. Grey vertical bars in inset show the 2 age ranges of radiocarbon datesused to constrain the age of onshore recent events. Up arrows indicate the event post-dates the age range, down arrowsindicate the event pre-dates the age range. Red vertical bars on turbidite events represent the one and two sigma rangefor Noyo Canyon ages. Numbers (with errors) and letters locate slip rates and event chronology sites. PA - PointArena; FR - Fort Ross; V - Vedanta; O - Olema; D - Dogtown; GG - Golden Gate; SF - San Francisco; SE - SealCove; LCS - Lower Crystal Springs; F - Filoli; AN - Ano Nuevo; SJ - San Jose; NA - New Almaden; GF - GrizzlyFlat; C - Corralitos; CC - Coward Creek; W - Watsonville; PG - Pajaro Gap; SJB - San Jaun Bautista; SC - Santa Cruz;M - Monterey; NC - North Coast segment, containing all segments. (Land data after Schwartz et al. (1998), modifiedfrom Niemi and Zhang (2000). Data added at Arano Flat from Fumal et al. (1999)). Offshore channels interpretedfrom GLORIA sidescan sonar data and offshore bathymetric compilation, this study.

Bodega C

hannel

PioneerChannel

NoyoChannel

Cordell Channel

Farallo

nCha

nnel

Guala

laCh

anne

l

VizcainoChannel

4000

3000

2000

PointArena

BS

F

PointDelgad a

HF

1400

1500

1800

1300

1200

1100

1000

900

1600

800

700

600

500

300

200

1000

1900

A.D. 1664

Pointarena

Vedanta

dogtown

GrizzlyFlatT

GrizzlyFlatD

CowardCreek

AranoFlat

SAR

SGF

SJ B

SC

M

SJ

SF

SE

PG

NA

W

O

SSCM

PA

FR

VD

LCS

F

GF

CC

NC

GG

PS

C

50 km

37 N

38 N

39 N

40 N

124 W 123 W 122 W

AN

1650-1812

post 1430-1630

post 1520-1776NoyoCanyon

modern

SAF

SAF

Calibratedage,

A.D.

400

49PC

A.D. 1204

25.52.5

171

243

174

73

1650-1812

Post 1430-1630

Post 1520-1776T:Pre 1640-1660

1430-1670

D:Post 1630-1776

A.D. 542

A.D. 1049

A.D. 1524

A.D. ???

T2

T4

T5

T6

T8

T7

T1

1700

49PC

LakeM

ountain

Fau

lt

GarbervilleFault

Maacam

aFault

BartlettSprings

Fault


16/26

9. Northern San Andreas seismotectonicsetting

The San Andreas Fault is probably the best-known transform system in the world. West ofthe Sierra Nevada block, three main fault sys-tems accommodate ~ 75% of the Pacific-NorthAmerica plate motion, distributed over a 100km wide zone (Argus and Gordon, 1991). Theremainder is carried by the Eastern Californiashear zone (Argus and Gordon, 1991; Sauber,1994). The Northern San Andreas is the main

system, accommodating about 25 mm/yr ofthe ~ 34 mm/yr distributed across Western Cali-fornia. Most of the remainder is taken up on theparallel Hayward-Rogers Creek system, and theslightly divergent Calaveras Fault System fur-ther to the east. South of San Francisco, thetransform system becomes more complex, andincludes the offshore San Gregorio Fault, which

joins the Northern San Andreas at Olema, justnorth of San Francisco (fig. 9). Between SanFrancisco and Cape Mendocino, the main strandof the San Andreas is a relatively simple systemwith most strain localized on the main strand.Seismicity offshore is virtually nil, with the

exception of the Mendocino triple junctionregion. Since the 1906 rupture, the main SanAndreas has been nearly aseismic, with only afew small events near Pt. Arena.

10. Northern San Andreas onshorepaleoseismicity

The San Andreas system has been intenselystudied on land, and has been divided into seg-ments based on its historical record of earth-quake behavior. The northern segment rupturedin the 1906 Mw 7.8 earthquake, and rupture ex-

tended from the at least San Francisco north toShelter Cove near Point Delgada (fig. 9). The pa-leoseismic history of the Northern San Andreassystem is presently under investigation usingtrenching and marsh coring. TheMw 7.8 earth-quake in 1906 clearly ruptured the surface a-long the San Francisco Peninsula to as far northas Point Arena (Lawson, 1908). Some debateexists regarding the full length of the 1906 rup-ture. Original investigations of surface rupture

are summarized in Lawson (1908), and includ-ed a description of surface rupture as far northas Shelter Cove. Much later, McLaughlin et al.(1979, 1983) questioned these reports on thebasis of bedrock mapping which indicated thatthe main fault strand had not displaced featuresdated at 13000 years BP. Seismological evi-dence does not require slip on the San Andreasnorth of Point Arena, though Thatcher et al.(1997) infers from geodetic data that 8.6 m ofslip occurred on the fault in the Shelter covearea. Brown (1995) re-examined surface mor-

phology and the original field reports by F.E.Mathes from 1906, and concluded that the orig-inal reports of surface rupture were correct, andthat many effects of the rupture are still observ-able today. Most recently, Prentice et al. (1999)also re-examined Mathes field notes and pho-tographs and trenched along the 1906 rupture.Like Brown (1995), they conclude that abun-dant evidence for 1906 rupture exists, and esti-mate a minimum slip-rate of 14 mm/yr for theNorthern San Andreas based on a 180 m offsetof colluvial deposits dated at 13 .180 170 calyears BP. The southern end of the ruptureextends as far south as the Santa Cruz moun-

tains (Schwartz et al., 1998), giving a minimumrupture length of 470 km.

The paleoseismology of the Northern SanAndreas has been investigated at Olema, 45 kmnorth of San Francisco, at Dogtown, close to theOlema site, at Point Arena, and at Grizzly Flatsin the Santa Cruz mountains. At the Vendantasite near Olema, Niemi and Hall (1992) foundthat offset stream channels showed that the faultruptured along a single main strand, and offsetstream deposits dated at 1800 78 years by 40-45 m. The maximum Late Holocene slip ratederived from these data is 24 3 mm/yr, in goodagreement with geodetic data. They estimate that

if the 4-5 m slip event recorded in 1906 werecharacteristic, the recurrence time for suchevents would be 221 40 years. At Point Arena,145 km to the northwest, Prentice (1989) rec-ognized four events that offset a Holocene allu-vial fan channel by 64 2 m. The maximum sliprate calculated at Point Arena is 25.5 mm/yr, inexcellent agreement with the Olema data. Theaverage slip per event at Point Arena implies arecurrence time of 200-400 years (Prentice,


1184


17/26

1989). Dated offset terrace deposits suggest thatthis rate has not changed by more than about20% since Pliocene time (Prentice, 1989). Thebest age derived for the penultimate event is themid 1600s (Schwartz et al., 1998), and themost likely ages for the previous three eventswere: #.3 1300 (post A.D. 1150, pre A.D.1650), and two events pre A.D. 1210 and postA.D. 1, totaling five events in 2000 years (Pre-ntice, 1989, 2000; Niemi and Zhang, 2000).Schwartz et al., 1998 also show an additionalevent at several sites in the early-mid 1500s.

A controversial aspect of Northern San An-dreas tectonics has been whether the fault is

segmented, with variable behavior for each seg-ment, or whether the 1906 rupture was charac-teristic. The consistent slip rates found north ofthe Golden Gate, slow to about 17 mm/yr southof the Golden Gate. This and a lower co-seis-mic slip south of the Golden Gate (Segall andLisowski, 1990; Prentice and Ponti, 1997;Thatcheret al., 1997) led investigators to con-clude that the fault is segmented near the Gold-en Gate. The Working Group on CaliforniaEarthquake Probabilities (1990) applied a uni-form slip rate to the fault, and concluded that

segments with lower-co-seismic slip in 1906should have more frequent events to fill the slip

1185


Fig. 10. Summary core logs and digital photographs of all sections of Noyo Channel core M9907-49PC. Seefig. 9 for location.

Silt

Sand

Vfs

Vfs/silt

Sand and silt

Hemipelagic clay

Turbidite clay

Burrows

510

520

530

540

0

10

20

70

80

90

100

110

120

130

140

150

16

180

190

200

210

220

250

2

130 280

300

310

340

350

380

430 550

560

570

580

590

600

610

620

630

64

650

66

67

680

700

690

Section 5depth in

cm

Section 4depth in

cm

Section 3depth in

cm

Section 2depth in

cm

Section 1depth in

cm

49PC

1-38 Turbiditeidentification


18/26

deficit. Schwartz et al. (1998) argues that thesegmentation is simply a reflection that the off-shore San Gregorio Fault (fig. 9) absorbs someof the slip, and the slip-rate on the main SanAndreas is correspondingly reduced. They ar-gue that the through-going rupture in 1906 wasnot segmented, and further, that the penultimateevent, which occurred in the mid-1600s, rup-tured approximately the same distance, and hada magnitude similar to the 1906 event.

11. New results: Northern San Andreasturbidite record

During our 1999 Cascadia cruise, we collect-ed two piston cores and one box core from NoyoChannel, 150 km south of the southern end ofthe Cascadia subduction zone. We did this bothto test the distance at which large ruptures wouldgenerate turbidites, and to investigate whetherthe Northern San Andreas had generated a tur-bidite record of its own. We found that the NoyoChannel cores near the offshore Northern SanAndreas Fault do show a good cyclic record ofturbidite beds (figs. 10 and 11a-d). In Core 49PC,

we find thirty-one turbidite beds above the Holo-cene./.Pleistocene faunal and lithologic datum asdefined by Duncan et al. (1970), which NoyoCanyon has a 14C age of 9000 years BP.

Thus far, we have determined ages for 20(of 38) events including the uppermost 5 eventsfrom cores 49PC/TC and adjacent box core50 B.C. using AMS methods. The uppermost

event returns a modern age, which weinterpret is likely the 1906 San Andreas earth-quake. The penultimate event returns an inter-cept age of A.D. 1664 (2 range 1505-1822).The third event and fourth event are lumpedtogether, as there is no hemipelagic sedimentbetween them. The age of this possible coupletevent is A.D. 1524 (1445-1664). The coupletcould represent two ruptures with little timebetween, an aftershock, or perhaps erosion atthe base of T3. The fifth event age is A.D. 1204(1057-1319), and the sixth event age is A.D.

1049 (981-1188). These results are in relativelygood agreement with the onshore work to datewhich indicates an age for the penultimateevent in the mid-1600s (figs. 9 and 11a-d), themost likely age for the third event of A.D. 1500-1600, and a fourth event A.D.1300. We presently do not have the spatial sam-pling needed to test for synchroneity of eventsalong the Northern San Andreas, and thus can-not determine with confidence that the ob-served turbidite record is entirely earthquakegenerated. However, the good agreement innumber of events between the onshore andoffshore records suggests that either turbidite

triggers other than earthquakes appear not tohave added significantly to the turbiditerecord along the northernmost San Andreasmargin during the last 2000 years, or thesimilarity of records is a coincidence.

With ages for 20 events, we can begin tomake some preliminary observations aboutthe turbidite event history. Figure 12b shows

1186

Fig. 11a-d. a) Top 93 cm of trigger core M9907-49TC and (b) the gamma-ray and magnetic susceptibility logs(uncorrected) for the entire length (700 cm) of piston core M9907-49PC from Noyo Canyon. The characteristiclow gamma and high magnetic susceptibility of the turbidite sands is used to identify and correlate turbiditeevents observed in the cores. c) A schematic core diagram of this piston core (49PC) is also shown. (Note cor-

relation is from the piston core to the trigger core, thus depths of events do not match exactly because the trig-ger core has more compaction than the piston core). Of the 38 total events seen in the piston core, 31 areHolocene age. The youngest 7 events seen in the trigger core (a) correlate with the youngest 7 events of the pis-ton core (b and c). The white buttons on the trigger core and the red horizontal lines on the log identify turbiditeevents. The recurrence times (c) of 216 and 234 years are calculated based on the age of T14 and the 1906 event.The Holocene-Pleistocene boundary is here defined at the 1 to 1 ratio of radiolarians to forams (Duncan et al.(1970); base of T31) is also apparent as a color change in the sediments, and is shown as the white curve nearthe bottom of (c). d) Sample core from the 1999 R/V Melville Cascadia Turbidite Paleoseismicity Cruise. Themiddle section of core shows the sharp color change from light grey to olive-green, characteristic of thePleistocene-Holocene boundary, and several thin, grey turbidite sands (labeled with Ts) separated by olive-green hemipelagic mud. Note dropstones in the Pleistocene section.



19/26

1187


c

d

a b


20/26

the age of the core with depth, and shows agood AMS age record, with no reversingintervals in those dated so far. Figure 12ashows the recurrence interval of events plot-ted with depth. It is apparent from both plotsthat both sedimentation rate and turbidite eventfrequency have not been constant through theHolocene. If the turbidite frequency is an earth-quake proxy, then the repeat time for eventsalong the northernmost segment near NoyoCanyon has decreased during the Holocene.This may be due to a number of possible caus-es: 1) the behavior of the fault has changed, for

example more slip shifting from other parts ofthe plate boundary system to the main SanAndreas in the Late Holocene; or 2) the recordincludes a non-earthquake climatic or sedimen-tation record that has changed through theHolocene. Climatic and sedimentation changeswould tend to favor a reduction in sedimenta-tion during the Holocene as sea level rose andseparated canyons from their river sources, andterrestrial erosion rates fell. We observe just the

opposite, with an increase in sedimentation rate(based on hemipelagic sediment between tur-bidites) that parallels the increase in turbiditefrequency. This observation tends to support achange in fault behavior, however this issuecannot be resolved with our single core.

12. Discussion

As the Cascadia turbidite project has ma-tured, we have learned a number of things aboutthe subtleties of using turbidites to precisely date

earthquakes. The most important is that regionaland temporal correlations are the strongest evi-dence of earthquake triggering, and radiocarbonevent ages are much less so. This is not surpris-ing, and is simply the result of the errors andunknowns inherent in the dating process. Someof these problems have been known to land pale-oseismologists for years; others are unique to thedeep-sea environment. Our goals in radiocarbondating have not been to use the ages to demon-

1188

Turbidite event recurrence versusdepthNoyo Canyon core 49PC

RECURRENCE (YEARS)

0 200 400 600 800 1000 1200 14000

100

200

300

400

500

600

700

DEPTH

(cm)

Turbidite Age (BP) versusdepthNoyo Canyon core 49PC

AMS C14 AGE (YEARS BP)

0 2000 4000 6000 8000 10000 12000 14000 160000

100

200

300

400

500

600

700

DEPTH

(cm)

Fig. 12a,b. a) Event recurrence versus core depth for Noyo Canyon core 49PC. b) Event age versus depth in core49PC. 2 errors for (b) shown by horizontal bars. Both plots show an increase in the interval between events downcore. We cannot determine the origin of these rate changes with our single Noyo core. The youngest interval in (b)represents an average recurrence time of 216-234 (events 3 and 4 could be either 1 or 2 events). The good corre-spondence between these events and the land record suggests that this interval is most probably a good proxy forthe Northern San Andreas earthquake record.


a b


21/26

strate earthquake origins, since the spatial andtemporal correlations do a much better job. Thepurpose of the ages is to attempt a correlationwith the land record, and to begin to integratethe two to cull more information about the earth-quakes themselves in terms of the spatial andtemporal patterns of strain release, about whichwe presently know very little.

Correlating paleoseismic events using thecalibrated ages has long been the Achilles heelof paleoseismology. Because the ages them-selves include many sources or error that must

be propagated, it is sometimes difficult to geteven known events to correlate this way. Landpaleoseismologists have worked to overcomethis problem by using large quantities of ages tobe able to do statistical analyses of the cloudof data associated with each event to make thecorrelation more robust. With offshore data, weare unable to use that tool at present because theexpense of collecting the original samples witha large ship has thus far precluded collectingenough sample volume to do more than one ortwo dates from each event. The limitation isimposed by the abundance of the planktonicforams that we are dating, of which there are just

enough in one or two 4 diameter cores beloweach event for one or two ages. What we do havein the offshore record is a longer record at mul-tiple locations. The tools that this provides offeranother way to arrive at a robust record. As men-tioned above, pattern matching offers a way toreveal a robust signal with less dependence onthe individual data points, which may be in errorfor a variety of reasons. With multiple plots ofrecurrence, turbidite thickness, hemipelagicthickness, velocity, magnetic susceptibility etc.,we can correlate events and do statistical analy-sis on the correlation of the curves. This can thenbe used in turn to identify outliers.

In fig. 8 we also show the land data fromWillapa Bay (Atwater and Hemphill-Haley,1997), a record from Lagoon Creek in NorthernCalifornia (near the Klamath River, Garrison-Laney et al., 2002), Humboldt Bay (Clarke andCarver, 1992), and Southern Oregon (Kelseyet al., 2002). These records show similarities tothe longer offshore record from Cascadia Basin,but it is too early to comment extensively. Ourfirst complete core shows differences with

events T2 and T3 of the onshore record, whichwe suspect may be due to the proximal envi-ronment of core 12PC. Previously we and oth-ers have noted a discrepancy between the aver-age recurrence time seen in the marsh record,and the average recurrence time from the tur-bidite record. With a more complete record, wethink that this apparent discrepancy may simplybe an artifact of the different time spans beingreported. Our average reported repeat time of 600 years is for the 13 post-Mazama events.Now, with a complete record for one of our key

cores, we see that if we calculate recurrence forthe same time span that Atwater (1987) do, weget a very similar result. At Willapa Bay,Atwater and Hemphill-Haley (1997) report anaverage repeat time of 533 years (7 events in3500 years); Kelsey et al. (2002) report 529years for the Sixes River in Southern Oregon (11events in 5500 years) and Garrison-Laney et al.(2002) report 608 years for Lagoon Creek. In ourrecord for the youngest Holocene interval, wecalculate an average recurrence of 490-550 yearsfor the last 3500-4000 years. Thus, it may be thatat least for the last 3000-4000 years, there is nodiscrepancy, and thus it will be possible to corre-

late events one for one. This adds strength to boththe land and offshore records in that input fromindependent upper plate earthquakes appear notto be significant at least for the data analyzedthus far for Washington, Oregon, and NorthernCalifornia as far south as Humboldt Bay.

The curves show a possible anomaly in the3000-4000 year range between the records inSouthern Oregon, the coastal records to thenorth and south, and the turbidite record. As didKelsey et al. (2002), we observe extra eventsin the onshore records from Bradley Lake andthe Sixes River areas relative to both the onshorerecord and our turbidite record. Our Rogue

Canyon cores directly sample the subductionzone events at this latitude, and the record of tur-bidites matches the other offshore records. Theextra events in Bradley Lake suggest thatSouthern Oregon lakes and estuaries experienceearthquakes or some other phenomenon thatleaves similar records in addition to the region-al great earthquake record. These could be up-per plate events, smaller interplate events, ormay possibly have some other origin.

1189



22/26

13. Future directions

If we can successfully calibrate our agessuch that they are equivalent to event ages in thecoastal marshes, we can find matches and mis-matches in the two records which will lead tobetter understanding of both records. For exam-ple, some of the turbidite events are clearlylarger than others, at least in terms of the thick-ness of the turbidite. Do these events correlateto larger onshore submergence, and or widerspatial recognition on land? Are they a result of

a longer interval preceding that event, and thusmore accumulated sediment was triggered? Orperhaps the longer interval ended in a largerearthquake as well. Two independent data setsallow a powerful analysis of these factors.

Moving toward a marine turbidite recordthat is both compatible with land records andcalibrated properly has some complications, butresolution of these complications is possible,and will increase the resolution and utility ofturbidites as earthquake proxies. Improvementsin dating fall into two categories: 1) arriving atthe best event age, and 2) calibrating the marineages to be compatible with the land record.

Arriving at the best event ages will involve re-moving or attempting to at least bracket some ofthe sources of error inherent in dating the eventsthus far. These include age bias induced by sam-pling below, not in each turbidite event. Sincewe do not date microfossils in the turbiditesthemselves because of potential reworking, wemust date them either above or below eachevent. We date below the events because thetransition where the fine-grained turbidite tailtransitions to hemipelagic sedimentation is verysubtle at times, and inconsistent, while the baseof the turbidite is very distinct. The downside ofthis is that there may be basal erosion. We

attempt to counter that by taking cores in non-erosive depositional environments such as dis-tal channel reaches and the lee sides of pointbars, but we are not always able to do that.Planktonic forams are thus collected from 2-3cm intervals of sediment below each turbidite.The AMS age represents the midpoint of thatinterval, which is older than the event, thus wemust correct for that. The correction however isalso not simple, since we need to determine the

local sedimentation rate (minus the turbiditeswhich are instantaneous) for each event, or atleast for an interval of time including that event.Final determination of hemipelagic sedimenta-tion rates and these corrections await a futurepaper.

We also would like to assess and, if possi-ble, make some correction for basal erosion, orat least identify events in cores that may havehad some unknown amount of erosion. We canevaluate the ages that have suspected basal ero-sion effects by comparing the maximum thick-

ness of hemipelagic sediment between each tur-bidite event. First, we can compare the multiplecores at each site to determine the maximumhemipelagic thickness found at each site foreach of the turbidite events. Then we can com-pare the maximum hemipelagic thicknessfound at any site for each turbidite event (i.e.T1, T2, etc.). If a sample for a radiocarbon agefor a specific turbidite event (i.e. T5) has beentaken at a site with less than the maximum hem-ipelagic thickness observed nearby, we willknow that there has probably been basal erosionby the turbidity current at that site for thatevent.

The second category of corrections to makethe marine cores compatible with land marshand tree-ring ages involves the corrections ap-plied to the radiocarbon ages themselves. Landages are corrected to a calibration curve that isgetting better with time as more data is used tocompute the curves. Marine ages have an addi-tional complication in that the exchange of airwith the ocean is not instantaneous, even in sur-face waters, and thus the water has an ageassociated with it that must be removed. This isdone using paired shell and wood dates frombays and estuaries that are incorporated into amarine database. This database has regional

adjustments (Delta R) that are made to the basiccorrection. In Cascadia, as in most regions,there are a number of calibration points alongthe coasts of Oregon, Washington, and Cali-fornia in the literature. The question is whetherthese values are valid for planktonic but pelag-ic (not estuarine) forams we use for AMS dat-ing. To a first order, we know that contempo-rary values are not far off, since we get ageswithin 50-100 years for the #1 event, which is

1190



23/26

known to be from the1700 A.D. earthquake(assuming acceptance of the premise of region-al spatial and temporal correlation).

We are very fortunate that we have the 1700earthquake because this known event offersthe possibility of calibrating the Delta R correc-tion. A related issue is that it is not known ifthe marine calibration corrections vary in time.We are again fortunate that we have not one,but two events of well determined age, the A.D.1700 earthquake, and the eruption of Mt. Ma-zama. There was some delay between the erup-

tion and the deposition of the turbidite, whichmakes this datum a bit less precise, however itstill offers a calibration point better than anyother available in the region. These two eventscover much of the Holocene, and spatiallyinclude nearly all of Cascadia, offering a meansto develop a calibration correction that accountsfor both temporal and spatial variability to atleast a first order for the Holocene.

14. Conclusions

Results from Cascadia thus far support

the conclusion that 18 large earthquakes mayhave ruptured at least the northern 2/3 of theplate boundary that we have examined. Synchro-nous triggering of turbid flows in several Was-hington channels has now been demonstratedfor 18 events spanning the entire Holocene.Radiocarbon ages cannot prove or disprovesynchronicity for the 18 events observed inother channels. While segmented rupture withclose temporal spacing cannot be ruled out, theidentical records found in the Rogue Channelwould have to be produced by a remarkablyconsistent long-term rupture pattern, with es-sentially no variability through 18 events over

9850 years. We are also encouraged by theclose agreement between preliminary data andthe onshore paleoseismic record along theNorthern San Andreas Fault. Preliminary re-sults from Noyo Channel indicate the potentialfor applying this method, utilizing additionalcores from other channels, to fully develop theHolocene event-history of the Northern SanAndreas Fault System. Most importantly, bycarefully correlating the turbidite and land

records, it should be possible to construct reli-able event records back 10.000 years forCascadia, the Northern San Andreas and otherfault systems. Doing so would allow investiga-tions of clustering, triggering and other recur-rence models, which are presently difficult totest due to the relatively short instrumental,historical and onshore-paleoseismic records.

Acknowledgements

We thank the officers and crew of the ScrippsVessel R.V. Melville with special thanks to BobWilson. We thank the members of the ScientificParty: Drew Erickson, Mike Winkler, Pete Kalk,Julia Pastor, Antonio Camarero, Clara Morri, GitaDunhill, Luis Ramos, Alex Raab, Nick PisiasJr., Mark Pourmanoutscheri, David Van Rooij,Lawrence Amy, and Churn-Chi Charles Liu.This research was supported by National ScienceFoundation grants EAR-0001023, EAR-017120and EAR-9803081 and U.S. Geological SurveyNational Earthquake Hazards Reduction Pro-gram award 01HQGR0051, and Cooperative A-greement 1434WR97AG00016. We also thank

John Adams and Phil Barnes for thoughtful re-views that much improved the paper.

REFERENCES

ADAMS, J. (1985): The potential for subduction zone earth-quakes in the Puget Sound, Washington area, Eos,Trans. Am. Geophys. Un., 66, 1071.

ADAMS, J. (1990): Paleoseismicity of the Cascadia subduc-tion zone: evidence from turbidites off the Oregon-Washington margin, Tectonics, 9, 569-583.

ANASTASAKIS, G.C. and D.J.W. PIPER (1991): The characterof seismo-turbidites in the S-1 sapropel, Zakinthosand Strofadhes basins, Greece, Sedimentology, 38,717-733.

ARGUS, D.F. and R.G. GORDON (1991): Current SierraNevada-North America motion from very long baselineinterferometry: implications for the kinematics of theWestern United States, Geology, 19,1085-1019.

ATWATER, B.F. (1987): Evidence for great Holocene earth-quakes along the outer coast of Washington State,Science, 236, 942-944.

ATWATER, B.F. (1992): Geologic evidence for earthquakesduring the past 2000 years along the Copalis River,southern coastal Washington, J. Geophys. Res., 97,1901-1919.

ATWATER, B.F. and E. HEMPHILL-HALEY (1997): Recurrence

1191



24/26

intervals for great earthquakes of the past 3500 years atNortheastern Willapa Bay, Washington, U.S. Geol.Surv. Prof. Pap. 1576, pp. 108.

ATWATER, B.F.,A. NELSON, J.J. CLAGUE, G.A. CARVER, D.K.YAMAGUCHI, P.T. BOBROWSKY, J. BOURGEOIS, M.E.DARIENZO, W.C. GRANT, E. HEMPHILL-HALEY, H.M.KELSEY, G.C. JACOBY, S.P. NISHENKO, S.P. PALMER,C.D. PETERSON and M.A REINHART (1995): Summaryof coastal geologic evidence for past great earthquakesat the Cascadia subduction zone, Earthquake Spectra,11, 1-18.

BROWN, R.D. (1995): 1906 Surface faulting on the SanAndreas Fault near Point Delgada, California, Bull.Seismol. Soc. Am., 85 (1), 100-110.

CASTILLO, D.A. and W.L. ELLSWORTH (1993): Seismote-ctonics of the San Andreas Fault System betweenPoint Arena and Cape Mendocino in NorthernCalifornia: implications for the development and evo-lution of a young transform, J. Geophys. Res., 98,6543-6560.

CLAGUE, J.J. (1997): Evidence for large earthquakes at theCascadia subduction zone,Rev. Geophys., 35, 439-460.

CLARKE, S.H. and G.A. CARVER (1992): Breadth of inter-plate coupling in the Southern Cascadia subductionzone: implication for earthquake magnitudes, Geol.Soc. Am. Abstr. Prog., 24, pp. 15.

DARIENZO, M.E. and C.D. PETERSON( 1990): Episodic tec-tonic subsidence of Late Holocene salt marshes,Northern Oregon Central Cascadia margin, Tectonics,9, 1-22.

DUNCAN, J.R., G.A. FOWLER and L.D. KULM (1970):Planktonic Foraminiferan-Radiolarian ratios andHolocene-Late Pleistocene deep-sea stratigraphy off

Oregon, Geol. Soc. Am. Bull., 81, 561-566.FIELD, M.E. (1984): The Submarine Landslide of 1980 off

Northern California, U.S. Geol. Surv. Circ., 938, 65-72.FUMAL T., G.F. HEINGARTNER and D.P. SCHWARTZ (1999):

Timing and slip of large earthquakes on the SanAndreas Fault, Santa Cruz Mountains, California,Geol. Soc. Am. Abstr. Prog., 31, A-56.

GARFIELD, N., T.A. RAGO, K.J. SCHNEBELE and C.A.COLLINS (1994): Evidence of a turbidity current inMonterey Submarine Canyon associated with the 1989Loma Prieta earthquake, Cont. Shelf. Res., 14 (6),673-686.

GARRISON-LANEY, C.E., H.F. ABRAMSON and G.A. CARVER(2002): Late Holocene tsunamis near the southern endof the Cascadia subduction zone, Seismol. Res. Lett.

GOLDFINGER, C. and C.H. NELSON (1999): Holocene recurrenceof Cascadia great earthquakes based on the Turbidite event

record,Eos, Trans. Am. Geophys. Un., 80, 1024.GOLDFINGER, C., C.H. NELSON and J. JOHNSON (2003): Ho-locene earthquake records from the Cascadia subduc-tion zone and Northern San Andreas Fault based onprecise dating of offshore turbidites, Annu. Rev.Geophys., 31, 555-577.

GORSLINE, D.S., T. DE DIEGO and E.H. NAVA-SANCHEZ(2000): Seismically triggered turbidites in small mar-gin basins: Alfonso Basin, Western Gulf of Californiaand Santa Monica Basin, California Borderland,Sediment. Geol., 135, 21-35.

GRANTZ, A., R.L. PHILLIPS, M.W. MULLEN, S.W. STARRATT,G.A. JONES, S.S. NAIDU and B.P. FINNEY (1996): Character,

paleoenvironment, rate of accumulation, and evidence forseismic triggering of Holocene turbidites, Canada abyssalplain,Arctic Ocean,Mar. Geol., 133, 51-73.

GRIGGS G.B. and L.D. KULM (1970): Sedimentation inCascadia. Deep-Sea Channel, Geol. Soc. Am. Bull., 81,1361-1384.

GRIGGS, G.B., A.G. CAREY and L.D. KULM (1969): Deep-sea sedimentation and sediment-fauna Interaction inCascadia Channel and on Cascadia Abyssal Plain,Deep-Sea Res., 16, 157-170.

HEMPHILL-HALEY, E. (1995): Diatom evidence for earth-quake-induced subsidence and tsunami 300 years agein southern coastal Washington, Geol. Soc. Am. Bull.,107, 367-378.

HEMPHILL-HALEY, E., A.R. NELSON, H.M. KELSEY and R.C.WITTER (2000): Displaced marine diatoms in a coastalfresh-water lake: microfossil evidence for Holocenetsunamis on the South-Central Oregon coast, inProceedings of the Geological Society of AmericaPenrose Conference on Great Cascadia EarthquakeTricentennial (Or. Dept. of Geol. Min. Ind.), edited byJ. CLAGUE, B. ATWATER, K. WANG, M.M. WAND andI. WONG, p. 53.

HUTCHINSON, I., J.J. CLAGUE, P.T. BOBROWSKY and H.F.L.WILLIAMS (2000): Investigations of Cascadia paleo-seismicity in southwestern B.C. and northernmostWashington State, in Proceedings of the GeologicalSociety of America Penrose Conference on GreatCascadia Earthquake Tricentennial (Or. Dept. of Geol.Min. Ind.), edited by J. CLAGUE, B. ATWATER, K. WANG,M.M. WAND and I. WONG, p. 61.

INOUCHI, Y., Y. KINUGASA, F. KUMON, S. NAKANO, S.YASUMATSU and T. SHIKI (1996): Turbidites as records

of intense palaeoearthquakes in Lake Biwa, Japan,Sediment. Geol., 104, 117-125.

JACOBY, G.C., D.E. BUNKER and B.E. BENSON (1997): Tree-ring evidence for an A.D. 1700 Cascadia earthquake inWashington and Northern Oregon, Geology, 25 (11),999-1002.

KARLIN, R.C. and S.E.B. ABELLA (1992): Paleoearthquakesin the Puget Sound region recorded in sediments fromLake Washington, U.S.A., Science, 258, 1617-1620.

KASTENS, K.A. (1984): Earthquakes as a triggering mecha-nism for debris flows and turbidites on the CalabrianRidge,Mar. Geol., 55, 13-33.

KELSEY, H.M., R.C. WITTER and E. HEMPHILL-HALEY(1998): Response of a small Oregon estuary to coseis-mic subsidence and postseismic uplift in the past 300years, Geology, 26.

KELSEY, H.M., R.C. WITTER and E. HEMPHILL-HALEY

(2002): Plate-boundary earthquakes and tsunamis ofthe past 5500 years, Sixes River estuary, SouthernOregon, Geol. Soc. Am. Bull., 114 (3), 298-314.

LAWSON, A.C. (1908): The California earthquake of April18, 1906,Rept. State Earthquake Invest. Comm. (Car-negie Institution, Washington D.C.), vols. I and II.

MCLAUGHLIN, R.J., D.H. SORG, J.L. MORTON, J.N. BAT-CHELDER, H.N. HEROPOULOS, H.N. OHLIN and M.B.NORMAN (1979): Timing of sulfide mineralization andelimination of the San Andreas Fault at Point Delgada,California, Eos, Trans. Am. Geophys. Union, 60,p. 883.

MCLAUGHLIN, R.J., K.R. LAJOIE, D.H. SORG, S.D.

1192



25/26

MORRISON and J.A. WOLFE (1983): Tectonic uplift of amiddle Wisconsin marine platform near the Mendocinotriple junction, California, Geology, 11, 35-39.

NAKAJIMA, T. andY. KANAI (2000): Sedimentary features ofseismoturbidites triggered by the 1983 and older his-torical earthquakes in the eastern margin of the JapanSea, Sediment. Geol., 135, 1-19.

NELSON, A.R., B.F. ATWATER, P.T. BROBOWSKI, L.A.BRADLEY, J.J. CLAGUE, G. CARVE, M.E. DARIENZO,W.C. GRANT, H.W. KRUEGER, R. SPARKS, T.W.STAFFORD and M. STUIVER (1995): Radiocarbon evi-dence for extensive plate-boundary rupture about 300years ago at the Cascadia subduction zone, Nature,378, 371-374.

NELSON, A.R., H.M. KELSEY, E. HEMPHILL-HALEY and R.C.WITTER (2000): Oxcal analyses and varve-based sedi-mentation rates constrain the times of C14 dated tsuna-mis in Southern Oregon, inProceedings of the Geolo-gical Society of America Penrose Conference on GreatCascadia Earthquake Tricentennial (Or. Dept. of Geol.Min. Ind.), edited by J. CLAGUE, B. ATWATER, K. WANG,M.M. WAND and I. WONG, 71-72.

NELSON, C.H. (1968): Marine geology of Astoria deep-seafan, Ph.D. Thesis, Oregon State University, Corvallis,p. 289.

NELSON, C.H. and C. GOLDFINGER (1999): Turbidite eventstratigraphy and implications for Cascadia basin pa-leoseismicity, Eos, Trans. Am. Geophys. Un., 80,733-734.

NELSON, C.H., B. SAVOYE, J.P. REHAULT and C. ESCUTIA(1995): Interfingering of Western Corsican marginaprons with the Var Fan lobe and an apparent Late Qua-ternary Corsican paleoseismic event, in International

Association of Sedimentologists 16th Regional Euro-pean Meeting, Aix-les-Bains, Savoie France, Abstractsvolume, p. 113.

NELSON, C.H., C. GOLDFINGER, J.E. JOHNSON and G.DUNHILL (2003): Paleoseismic history of the Cascadiasubduction zone derived from turbidite stratigraphy,U.S. Geol. Surv. Prof. Pap. (in review).

NIEMI, T.M. and N.T. HALL (1992): Late Holocene slip rateand recurrence of great earthquakes on the San AndreasFault in Northern California, Geology, 20, 195-198.

NIEMI, T.M. and Z. BEN-AVRAHAM (1994): Evidence forJericho earthquakes from slumped sediments ofthe Jordan River delta in the Dead Se

Documents

Deep-Water Turbidites as Holocene San Andres Fault