tracking ﬂuvial response to climate change in the paciﬁc...

ARTICLE IN PRESS

0277-3791/$ - se

doi:10.1016/j.qu

�CorrespondCollege, Univer

UK. Tel.: +01

E-mail addr

s.vanlaningham

rduncan@coas.

[email protected]

dgraham@coas

Please cite th

approach usin

Quaternary Science Reviews ] (]]]]) ]]]–]]]

Tracking fluvial response to climate change in the Pacific Northwest:a combined provenance approach using Ar and Nd isotopic systems on

fine-grained sediments

Sam VanLaningham�, Robert A. Duncan, Nicklas G. Pisias, David W. Graham

College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA

Received 12 August 2007; accepted 30 October 2007

Abstract

Traditional provenance techniques (Nd isotopes and clay mineralogy) are combined with recently developed bulk sediment 40Ar–39Ar

radiometric methods to determine how the terrestrial sources of sediment to the Oregon continental margin have changed over the last

25,000 years. Both Pacific Northwest river-borne detritus, and sediment from piston coring site EW9504-17PC (2671m water depth)

offshore southern Oregon have been analyzed. Nd isotopic analyses of river silts show a range of 10 units in eNd. North of the core site,

the Columbia River has eNd ¼ –7.6, while the Coos River has a value of eNd ¼ �10.8. Rivers proximal to the core site have more

radiogenic values from north to south, of eNd ¼ �5.0 (Umpqua River), eNd ¼ �1.3 (Rogue River), eNd ¼ �0.6 (Klamath River) and

eNd ¼ �3.0 (Eel River). Measured eNd in core sediments show subtle downcore changes, between eNd ¼ �0.9 and �2.5. The bulk

sediment 40Ar–39Ar plateau ages show more notable downcore variation between 25 and 14 ka, ranging from 113.5 to 124.0Ma, but are

still within the range of bulk ages previously measured on river mouth sediments. The Nd isotopic analyses are combined with bulk

sediment 40Ar–39Ar plateau ages into a ternary mixing model to quantitatively assess the sources of terrigenous material. Mixtures are

best described by three sources proximal to the core site (the Umpqua, Rogue+Klamath and Eel Rivers) from �14 ka to Present.

Sediment deposited in the interval from 22 to 25 ka is not adequately described by the present-day rivers and requires an additional

source. This additional source is best explained by an enhanced contribution from the interior Cascade volcanic arc, probably due to

glaciation in the Cascade Range and the presence of pluvial Lake Modoc in the Upper Klamath Basin at this time. From 22 to 14 ka, the

influence of Cascade Range sediment at the core site was overprinted by contemporaneous glaciation and sediment production in the

Klamath Mountains, and possibly addition of sediment from the Eel River region as well. Thus, differential erosion appears to play a

primary role in the provenance changes seen at the core site, and is more significant than sediment transport changes due to ocean

circulation during this time period. Overall, the combined Ar–Nd isotopic technique provides insight into the coupling of land surface

processes and ocean circulation related to climate change and will be a useful provenance tool in a variety of geologic/climatic settings.

r 2007 Elsevier Ltd. All rights reserved.

1. Introduction

Marine sedimentary records at continental marginscapture the integrated signal of oceanic, atmospheric and

e front matter r 2007 Elsevier Ltd. All rights reserved.

ascirev.2007.10.018

ing author. Present address: School of Geosciences, Kings

sity of Aberdeen, Meston Building, Aberdeen AB10 6ET,

224 273437.

esses: [email protected],

@abdn.ac.uk (S. VanLaningham),

oregonstate.edu (R.A. Duncan),

gonstate.edu (N.G. Pisias),

.oregonstate.edu (D.W. Graham).

is article as: VanLaningham, S., et al., Tracking fluvial respon

g Ar and Nd isotopic systems on fine-grained.... Quaternary S

terrestrial processes (e.g., Lamy et al., 2001; Pisias et al.,2001; Clift, 2006). Rivers communicate the terrestrialcomponent of change through the erosion, entrainmentand delivery of material to the continental margin (e.g., Cliftand Blusztajn, 2005). Changes over climatic timescales (o105

years) on the continental surface are mostly in response toatmospheric processes and because of this, combining marineand terrestrial studies at ocean margin core sites casts a widespotlight on the entire climate system. A blessing and curse inthe deep marine realm is that it preserves a thorough andcontinuous history of earth surface processes unaffected bytransgressions and regressions of the ocean. Yet, with that

se to climate change in the Pacific Northwest: a combined provenance

cience Reviews (2008), doi:10.1016/j.quascirev.2007.10.018

dx.doi.org/10.1016/j.quascirev.2007.10.018

mailto:[email protected],

mailto:[email protected]





ARTICLE IN PRESSS. VanLaningham et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]]2

comes the fact that material transported beyond thecontinental shelf and slope to the deep ocean is usually finegrained and difficult to characterize. Thus, a need remainsfor studies that can extract the important source informationcaptured in fine-grained silicates. This paper focuses on newisotopic approaches to characterize silt-sized river-borne,terrigenous sediment and documents the source of terrestrialmaterial to the northeast Pacific margin offshore Oregonover the Last Glacial–Interglacial period. We focus on thesilt-sized component because it contains rock-formingminerals and yet these grains are small enough to betransported long distances.

The study region along western North America showsevidence for a tight coupling between terrestrial–ocean–at-mospheric systems and lies in a transition zone between theNorth Pacific and Alaskan Gyres (Fig. 1A and inset).Radiolaria abundances that are typified by an easternboundary current species and pollen assemblages that aredominated by redwood co-vary on glacial–interglacialtimescales (Pisias et al., 2001). A hypothesis is that, whenthere is a strong eastern boundary current due to enhancedwind-stress curl along the margin there is increased coastalfog due to upwelling of cold, deeper waters, inevitablyleading to conditions favorable to redwood growth.However, what if this apparent coupling between terrestrialand ocean systems recorded in marine sediments is due toocean circulation changes? One way to test this is byexamining if the carrier of the pollen (terrigenous sediment)is changing contemporaneously with the pollen assem-blages on appreciable temporal and/or spatial scales.

With this motivation we combine Nd isotopic analyseswith a newly established 40Ar–39Ar incremental heatingtechnique (VanLaningham et al., 2006) applied to fine-grained sedimentary material to determine the terrestrialsources of sediment arriving at a core site located offshoresouthern Oregon. Bulk 40Ar–39Ar ages on Pacific North-west river sediments are published previously (VanLaning-ham et al., 2006), while the new Nd isotopic data from riversilts shown here provide additional means to preciselycharacterize the river-borne sediment in this region. Wethen present new Ar and Nd isotopic analyses on deepmarine sediments to address how sediment compositionchanges over the last 25,000 years and evaluate to whatextent the pollen record preserved at the core site is arecord of terrestrial climate change as opposed to changingtransport pathways of this pollen (and sediment) fromocean circulation or other surface process mechanisms. We

Fig. 1. Location map showing study rivers, core sites and bedrock geology o

Pacific Ocean and physiography (B), sediment loads (� 109 kg/year ) of major

rivers (D). Inset shows the larger circulation patterns of the North Pacific and

sediment load derived for the Columbia R. (fromWolf et al., 1999; Syvitski et a

modeled age used later in this paper. Legend for the geologic map: OLY ¼ Olym

SRTV ¼ Siletz River-Tillamook Volcanics; TYEE ¼ Tyee Formation Turbid

ciscan Melange; VA ¼ Valley Alluvium; SNB ¼ Sierra Nevada Batholith; CV

Canadian Cordillera; IBB ¼ Idaho-Bitterroot Batholith; VIR, CM and FP ¼

contribute to Columbia River). See text for further explanation of lithologies.

Please cite this article as: VanLaningham, S., et al., Tracking fluvial respon

approach using Ar and Nd isotopic systems on fine-grained.... Quaternary S

demonstrate that sediment provenance has indeed changed atthe core site, but only from proximal sources and thus, wasnot strongly driven by circulation changes. The change wasmostly due to differential erosion in river basins that result intransport of material from different parts of the onshorelandscape that are affected by climate differently. We thendiscuss our findings in terms of the continental PacificNorthwest–northeast Pacific oceanic climate system andprovide insights about terrestrial–ocean climate linkages.

2. Regional setting

This study examines a sediment core from the northeastPacific Ocean (Fig. 1) and the fluvial sources of thosesediments. All potential river-borne sediment sources to thecore site were examined. We sampled 14 major rivers fromas far north as the Olympic Peninsula, Washington to asfar south as San Francisco Bay, California to determinefine-grained sediment compositions.

2.1. Ocean circulation and climate

Present-day ocean circulation in the northeast PacificOcean is dominated by the California Current system. Itshows strong seasonality driven by changes in the locationsand intensities of the North Pacific and Alaskan Gyres,which reflect wind-stress conditions over the Pacific. Themajor components of ocean circulation are the CaliforniaCurrent, the Davidson Current and the California Under-current (Fig. 1A and inset), although smaller-scale flowexists (Hickey, 1979; Werner and Hickey, 1983; Hickey andRoyer, 2001; Hickey and Banas, 2003). In the spring andsummer, the California Current flows south along the westcoast as a �1000 km wide swath from the surface to 500mdepth (Hickey and Banas, 2003). The narrow (10–40 km)California Undercurrent flows northward beneath thesurface along the continental slope during this time.In fall and winter, coastal circulation changes consider-

ably. Surface flow along the margin switches to northwardfrom Point Conception, California past Oregon andWashington (Hickey and Banas, 2003) and is known asthe Davidson Current. It spreads �100 km wide across thedistal continental shelf and slope. Inner and mid-shelfcirculation generally follows trends of the California andDavidson Current, whereby flow is southward duringspring and summer and northward in fall and winter

f the western US. (A), present-day upper ocean currents in the northeast

rivers (C) and bulk sediment 40Ar–39Ar ages and eNd values of the major

locations of southernmost rivers in the study area. *Pre-dam estimate of

l., 2005). **The 40Ar–39Ar age of 109Ma for the Klamath River refers to a

pic Mountain Accretionary Complex; WCR ¼Washington Coast Ranges;

ites; KLA ¼ Klamath Mountain Accretionary Complex; FRAN ¼ Fran-

¼ Cascade Volcanics; CRB ¼ Columbia River Basalts; WWC ¼Western

Vancouver Island Ranges, Coast Mountains and Fraser Plateau (do not




ARTICLE IN PRESSS. VanLaningham et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]] 3

Please cite this article as: VanLaningham, S., et al., Tracking fluvial response to climate change in the Pacific Northwest: a combined provenance

approach using Ar and Nd isotopic systems on fine-grained.... Quaternary Science Reviews (2008), doi:10.1016/j.quascirev.2007.10.018



(Strub et al., 1987). The southward-flowing CaliforniaCurrent drives upwelling along the margin of the PacificNorthwest during spring and summer, although itsstrength and duration increase from north to south.Upwelling offshore Oregon occurs from about March toSeptember, while it generally shuts down during fall andwinter (Strub et al., 1987).

Sedimentation along the margin follows shelf and slopecirculation. Clay mineralogy of surface sediments (Karlin,1980) indicates that sedimentation offshore northernCalifornia, Oregon and Washington occurs dominantly inwinter and is northward, when the Davidson Current isactive and when the coastal region experiences extensiveprecipitation. There is some input of sediment related tospring runoff from alpine, interior regions in the Columbia,Umpqua, Rogue and Klamath River basins. However, claymineralogy data suggest that presently only the ColumbiaRiver has a north to south sediment delivery during thespring, when circulation offshore Washington and Oregontransitions to southward surface flow and also because thelarge Columbia freshwater plume disturbs circulation(Hickey and Banas, 2003).

2.2. Onshore climate setting

Presently, the Pacific Northwest consists of temperatecoastal, alpine and steppe provinces north of southernOregon/northern California. In California, the climate ismuch more moderate and Mediterranean-like and influ-enced by the Great Basin high-pressure system. Annualprecipitation in Washington coastal areas is approximately230–250 cm/year while the Olympic Mountains accrue asmuch as 650 cm/year (http://www.ncgc.nrcs.usda.gov/). Inwestern Oregon, 200–230 cm/year of precipitation occursalong the coast, between 250 and 400 cm/year accumulatesin the coastal mountains to the east (the Oregon CoastRanges and Klamath Mountains), while farther inland theWillamette Valley receives around 125 cm each year.

Whereas the coastal regions of Washington and Oregonshow substantial west–east orographic gradients in pre-cipitation, coastal California and the coastal mountainssouth of the Klamath Mountain region show very littleorographic control on precipitation patterns, although anotable latitudinal gradient is observed. Annual rainfalldecreases from about 175 cm around Eureka, CA (�411N)to around 100–150 cm in the San Francisco Bay area(381N).

Farther inland, the Cascade Range receives a diminishedamount of rainfall relative to the coastal mountains (about150–250 cm/year), with higher values generally in the northand at higher elevations. The Cascade Range creates aconsiderable rain shadow effect and precipitation reducesconsiderably over the Columbia Plateau and high desert ofeastern Oregon (25–50 cm/year). The headwaters of theColumbia River in the Rocky Mountains of BritishColumbia and Montana experience about 150 cm/yearwhile lower topography in these regions amass around



50–60 cm/year. Precipitation rates in the Snake Riverregion, the southern arm of the Columbia River, arearound 25–40 cm/year in the lower plains and 75–125 cm/year in the surrounding mountains (http://www.ncgc.nrcs.usda.gov/).

2.3. Present-day river discharge and sediment loads

River discharge and sediment loads from the PacificNorthwest vary widely (Karlin, 1980; Sherwood et al.,1990; Wolf et al., 1999; Wheatcroft and Sommerfield, 2005)and are presented in Table 1 (sediment loads from themajor rivers are represented schematically in Fig. 1B also).At 216 km3/year, the Columbia River releases over 20 timesmore freshwater to the northeast Pacific Ocean than thesecond largest contributor in the region (Rogue River,10.1 km3/year) and about 75% of the total freshwateroutput from all of the major Pacific Northwest riverscombined. In terms of total sediment loads, however, theEel River delivers the most sediment presently at about18� 109 kg of sediment per year, whereas estimates for theColumbia River suggest that it contributes only about5� 109 kg (Table 1). Notably however, dams may havereduced the Columbia River sediment load by at least 50%(Sherwood et al., 1990) and possibly as great as 75% (Wolfet al., 1999; Vorosmarty et al., 2003; Syvitski et al., 2005)during the last few hundred years, which would place itspre-dam sediment load in the range of 10–20� 109 kg/year,comparable to the present-day Eel River. There has been aless severe reduction in clay- and silt-sized sediments(�33%; Sherwood et al., 1990), which are the size fractionswe have analyzed. The Klamath (and Trinity Rivertributary), Rogue, Mad and Umpqua Rivers are the othermajor sediment sources to the margin and lie north andsouth of the core site.Sediment yields (a drainage area-normalized measure of

sediment flux) are extremely high in the central andnorthern California coastal rivers such as the Eel, Mad,Mattole, Trinity Rivers and Redwood Creek. This isrelated to the combined effects of erodable FranciscanMelange rocks (McLaughlin et al., 1994) and their degreeof shearing related to high deformation rates along the SanAndreas Fault and surrounding structures in the Mendo-cino Triple Junction (Merritts and Vincent, 1989; Snyder etal., 2000).

2.4. Geology of source terranes

Western North America has a diverse set of geologicprovinces that contribute to sediments accumulating in thenortheast Pacific Ocean (Fig. 1C). Sediment sources tothe margin in Washington are mostly from the rocks of theOlympic Mountains, Washington Coast Ranges andColumbia River. The Olympic Mountains are comprisedof a suite of sandstones, mudstones and volcanic lithologiesthat have Eocene to Miocene depositional ages (Brandonand Vance, 1992). In the southwestern coastal ranges of



http://www.ncgc.nrcs.usda.gov/




ARTICLE IN PRESS

Table 1

River statistics

River Basin size

(km2)

Avg. discharge (km3/

year)

Avg. annual load

(� 109 kg)

Sed. yield (� 103 kg/km2/

year)

Refs. Period of

record

N

Quinalt 684 2.6 0.1 125* (1)

Grays Harbor 5776 8.1 0.7 114.5+ (1)

Willapa Bay 1046 1.9 0.1 125*+ (1)

Columbia 661,211 216

Present-day 5.0 7.6 (3) 1964–1969

Pre-Dams 20.0 30.2 (4),

(5)

Tillamook Bay 1400 2.7 0.2 125*+ (1)

Siletz 523 1.4 0.1 125*+ (1)

Yaquina 665 1.0 0.1 128.8 (1)

Alsea 865 1.5 0.1 75 (2) 1939–P 188

Siuslaw 1523 1.8 0.1 62 (2) 1967–1994 826

Umpqua 9534 6.7 1.4 147 (2) 1905–P 139

Coosa 1567 2.7 0.2 150 (2)

Coquillea 1960 2.2 0.3 150 (2)

Rogueb 13,394 10.1 2.3 170 (2)

Chetcoc 702 2.3 0.2 250 (2)

Smith 1590 3.3 0.4 252 (2) 1931–P 264

Klamath 21,950 7.3 3.3 150 (2) 1927–P 1696

Trinity(Klamath

Trib.)

7390 4.7 6.8 920 (2) 1931–P 1876

29,340 12.0 10.1 344

Redwood Creek 720 0.9 1.3 1805 (2) 1953–P 2034

Mad 1256 1.3 2.6 2070 (2) 1950–P 1495

Eel 8063 6.6 18.0 2232 (2) 1913–P 1817

Mattoled 635 1.1 1.3 2000 (2)

Navarroe 785 0.5 0.2 300 (2)

Gualalae 901 0.4 0.3 300 (2)

Russian 3452 2.4 1.1 318 (2) 1939–P 1510

302

N is the number of days used to determine sediment yields. Yields from basins with letter subscripts were estimated based on the measured yields from

adjacent basins, aUmpqua, cSmith, dEel, eRussian. The Rogue (b) is a composite of the Klamath and Smith (refer to Wheatcroft and Sommerfield, 2005 for

details). Refs.: (1) Karlin (1980), (2) Wheatcroft and Sommerfield (2005), (3) Sherwood et al. (1990), (4) Wolf et al. (1999), (5) Syvitski et al. (2005).

S. VanLaningham et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]] 5

Washington and northwestern Oregon, source material ismade of Eocene sedimentary and volcanic rocks (Walkerand MacLeod, 1991). The Columbia River drains animmense area of western North America and a largevariety of rock types and geologic provinces. Lithologies inthe headwaters of the Columbia River (including theKootenay and Okanogan River tributaries) in BritishColumbia are Mesozoic accreted terranes of sedimentary,volcanic and plutonic origins as well as many stocks thatintruded into the accreted units (Monger et al., 1982;Ghosh, 1995). The Snake River, a large tributary of theColumbia River, erodes mostly Cretaceous Idaho-Bitter-root Batholith granites in the high-relief regions and thenflows across young Tertiary volcanic rocks in the SnakeRiver Plain. The headwaters of the Pend Oreille-ClarkFork and Clearwater tributaries drain the 55–62Magranitic batholiths of the Bitterroot Mountains and meta-sedimentary rocks with Mesozoic to Paleozoic depositionalages from the Rocky Mountains of Montana. Over muchof the Columbia Basin, the Columbia River and tributarieserode and flow through the vast area of Miocene ColumbiaRiver Basalts (�50% of the basin area) and the Tertiary to



Recent arc volcanic rocks of the Cascade Range (�8% ofthe Columbia River basin area). Even though all of thesegeologic provinces could be contributing to the sedimentcomposition of the fluvial material traveling through theColumbia River out to sea, it has been shown previouslyusing Ar isotopes that the present-day silt-sized ColumbiaRiver material is dominantly from the British ColumbiaCordillera (VanLaningham et al., 2006).Coastal sediment sources in central Oregon erode

turbidite sequences and oceanic basalts with Eocenedepositional ages. The largest rivers in southern Oregonand northern California begin in Cenozoic basaltic andandesitic rocks of the central and southern OregonCascades and either traverse the turbidites of the OregonCoast Ranges (Umpqua River) or the Mesozoic KlamathMountains (Rogue and Klamath Rivers) en route to thePacific Ocean. The Klamath accretionary complex iscomposed of meta-sedimentary, meta-volcanic, graniticand gabbroic rocks and ophiolitic sequences.The coastal range of northern and central California is

dominated by the Franciscan Melange, a Cretaceous tolower Tertiary sequence of sandstones and mudstones, with





large blocks (up to several kilometers in size) of serpenti-nite, blueschist, eclogite greenstone, chert and limestone(Blake and Jones, 1981; McLaughlin et al., 1994), althoughother less extensive volcanic lithologies are present in theregion as well. The Eel River, western North America’slargest present-day sediment producer, erodes these rocks,as do parts of the Russian River. Central Californiansediment sources (Sacramento and San Joaquin Rivers)drain the granitic rocks of the Mesozoic Sierra NevadaMountains, Great Valley sediments and lesser contribu-tions from the Klamath and Franciscan rocks.

2.5. Marine setting at the core site

Piston coring site EW9504-17PC (42.241N, 125.891W) islocated about 120 km west of the Oregon–Californiacoastline at a water depth of 2671m and lies in thetransition zone between the Alaskan and North PacificGyres (Fig. 1 and inset). The core was collected on a localbathymetric high on the east flank of Gorda Rise to avoidturbidites and to capture a continuous record of hemi-pelagic sedimentation. Sedimentation along the Oregon–California margin is driven by terrigenous sources and thusCorg is relatively high while CaCO3 abundances are low(5% on average; Lyle et al., 2000). The core has remarkableconsistency in both color and grain size (clayey silt)throughout the length of the �15m core. Only one visibleturbidite is present, a 2 cm thick sandy bed at 313 cm(�17.7 ka).

3. Materials and methods

3.1. Sampling and preparation

Sediments at core site EW9504-17PC are made up ofclay- to silt-sized particles. Hence, it is important tocharacterize similar size fractions from the river sedimentsthat contribute to margin sedimentation. Detailed descrip-tions of sampling and preparation procedures are providedin VanLaningham et al. (2006) and in an Oregon StateUniversity College of Oceanography reference manual(Robbins et al., 1984). Briefly, we sampled rivers neartheir mouths and usually above tidewater to obtain themost representative material that is transported out to sea.We treated the sieved 0–63 mm material with hydrogenperoxide to remove organic material while calciumcarbonate was removed with buffered acetic acid. Oxy-hydroxides were removed using 1.0N hydroxylaminehydrochloride buffered with sodium citrate. About75–200mg of silt-sized (20–63 mm) material was extractedfrom 100 to 200 g of bulk fluvial sediment by sieving andsettling, while centrifugation was used to extract 2–20 and0–2 mm size fractions.

In core EW9504-17PC, the terrigenous componentdominates the sedimentology of the core (calcium carbo-nate content varies from 0% to 13% while organic carbonvaries between 1% and 2%; Lyle et al., 2000). We collected



10 cm3, sampled every 2 ka from �25 ka to Present andapplied the same preparations to the core samples as theriver samples.

3.2. Nd isotopic procedures for river and downcore samples

We determined the Nd isotopic composition of bulk, silt-sized (20–63 mm) river and downcore samples. We firstpulverized about 50–100mg of sample and heated it in afurnace to 800 1C (we emphasize that only the sample splitused for Nd analyses were heated) to remove morerefractory organic compounds not dissolved in the H2O2

step. Samples were digested in Teflon Savillexs beakerswith 5ml of concentrated HF and 1ml of 8N HNO3. Thesamples dissolved within 24–48 h using a combination ofsonication and heating (at �110–140 1C). We then con-verted solutions to chlorides by drying down in 6N HCland then taking samples back up in 2.5N HCl for columnchemistry. Each sample was first passed through a singlecation exchange column of Dowexs AG50� 8 to separateREEs from all other elements, using 2.5N HCl followed bycollection of the REE in 6.0N HCl. A second chromato-graphy column using ln-spec resin (Eichroms industries)was used to separate Nd from other REEs, using 0.25NHCl as the elutant.Analyses were performed at the Oregon State University

Keck Laboratory using a Nus Instruments multi-collectorinductively coupled plasma mass spectrometer. Samplesentered the instrument in dilute HNO3 and were analyzedusing a multi-dynamic analysis setup. Nd isotopic ratios inthe samples and standard were mass fractionationcorrected to 146Nd/144Nd ¼ 0.7219 (Wasserburg et al.,1981). Measured values for the J-Ndi standard wereconsistently offset by 0.8 eNd units relative to the acceptedvalue (143Nd/144Nd ¼ 0.51211577; Tanaka et al., 2000).A correction was applied to samples and to BCR-1 rockpowders analyzed as unknowns (processed using the samechemical procedures as samples). Reproducibility of theJ-Ndi standard was 0.000024 (2s, n ¼ 47). The mean valuefor BCR-1, adjusted using the offset difference fromthe J-Ndi standard was 143Nd/144Nd ¼ 0.512632723(2s, n ¼ 10). Each sample was bracketed by a J-Ndi orBCR-1 during the analytical campaign. Typical ion beamintensities for samples during runs were �1.0–3.0V, whileintensities of around 2.0V were consistently acquired forJ-Ndi. A small correction was applied for isobaricinterferences of Sm on the 144Nd peak, when present, bymonitoring 147Sm during analysis. This correction wastypically less than the analytical uncertainty of themeasured 143Nd/144Nd.

3.3. 40Ar–39Ar incremental heating procedures for core

samples

A full description of the procedures used here for40Ar–39Ar analyses is provided in a previous paper(VanLaningham et al., 2006). Briefly, samples were





irradiated with the FCT-3 monitor standard(age ¼ 28.0270.28Ma; Renne et al., 1998) in the 1MWTRIGA reactor at Oregon State University to induce thereaction 39K (n, p) 39Ar. The irradiated samples wereanalyzed using the MAP 215-50 mass spectrometer in theAr geochronology laboratory at Oregon State University.They were heated incrementally with a defocused 10WCO2 laser programmed to traverse the sample during eachheating step (for approximately five minutes). Sampleswere degassed with 13–15 temperature steps, from 200 to300 1C to fusion at around 1400 1C. Zr–Al getters removedactive gases from the extracted Ar before measurement inthe mass spectrometer. Isotopic masses for 40Ar, 39Ar,38Ar, 37Ar and 36Ar were measured so that appropriatecorrections could be made for reactor-produced interfer-ences (McDougall and Harrison, 1999). The most im-portant of these corrections is an atmospheric correction of40Ar via the 40Ar/36Ar ratio and reactor-induced 39Ar and36Ar interferences from Ca that are corrected via 37Ar.

The plateau and integrated (total fusion) ages werecalculated from corrected 40Ar/39Ar ratios using ArArCalc(Koppers, 2002). A plateau age is a consecutive sequence ofconcordant heating step ages comprising more than 50% ofits total 39Ar released in the sample (McDougall andHarrison, 1999; Koppers, 2002). A mean square ofweighted deviations calculation was made for every plateauage to assess the goodness of fit. When this value is muchgreater than 1.0, it suggests that geological factors such asalteration, multiple age populations in a given mineralphase (VanLaningham et al., 2006) or reactor effects suchas Ar recoil, could be contributing to the measured agesand the weighted plateau ages might be underestimated(McDougall and Harrison, 1999; Wendt and Carl, 1991;Koppers, 2002). Potassium–calcium ratios (K/Ca) werealso calculated from measured concentrations of 39Arderived from K and 37Ar derived from Ca.

3.4. Other methods

For the mixing model discussed later in this paper, weuse Nd concentrations determined from sample dissolu-

Table 2

Age control points (ACP) for downcore age model

EW9504-17PC

depth (cm)

W8709A-13PC equivalent

depth (cm)

EW9504-17PC

depth between CPs

0.0 – –

135.4 69.1 120.3

255.7 189.9 25.5

281.2 220.0 39.3

320.5 257.8 15.4

335.9 268.9 20.5

356.4 289.0 19.4

375.8 313.0 70.3

446.1 396.5 –

aBased on reservoir corrected 14C dates (using CALIB4.1) from Mix et al. (1

for each core independently validates our choice of ACPs.



tions separate from those used for isotopic analyses,measured using a VG Instruments EXCELLs ICP-MS.Procedures used are summarized in Walczak (2006). 40Ar*(radiogenic 40Ar) concentrations for each river sample usedin the mixing model (see below) were estimated from theportion of the irradiated material used in the agedeterminations. Because the general trend in these esti-mates follow measurements of the K concentrations (usingICP-AES; VanLaningham et al., 2006), we are confidentthat the 40Ar* values are consistent, albeit crude estimatesof the real concentrations.We introduce clay mineralogic analyses made on the

river and core samples using X-ray diffraction in orderto supplement the other provenance techniques. Weincorporate standard clay mineral analyses (Moore andReynolds, 1989) using a Scintag Pad V diffractometer onglycolated smear slides of the 0–2 mm material from 21 to341 2y using Cu radiation. The data were analyzed usingMacDiff software (http://www.geologie.uni-frankfurt.de/Staff/Homepages/Petschick/RainerE.html). Mineral abun-dances were calculated using an internal talc standard(10% by weight) following the procedures of Heath andPisias (1979) so as to have a more quantitative assessmentof smectite, illite and chlorite+kaolinite abundances. Wecalculate ratios of these abundances to further avoid anyvariations related to changes in sediment flux.

3.5. Age model for core site EW9504-17PC

We improve the age model for EW9504-17PC over thelast 25 ka using the latest published reservoir-corrected 14Cdates (CALIB4) from foraminifera in nearby coreW8709A-13PC (Mix et al., 1999), which is �15 km awayfrom EW9504-17PC. Similar features in the CaCO3 recordsfrom each site were connected using tie points from the twocores (Table 2). A simple consistency test of the chosen tiepoints is done by comparing the thicknesses of sedimentbetween each tie point in both cores and suggests that thecontrol points are indeed reasonable choices because thethicknesses are comparable from each core over eachinterval (Table 2). The new age model adjusts depositional

W8709A-13PC

depth between CPs

Depositional age

(ka)aSedimentation rate

(cm/ka)

– 0.0 15.6

120.8 8.7 23.1

30.1 13.9 13.4

37.8 15.8 17.1

11.1 18.1 77.0

20.1 18.3 51.3

24.0 18.7 12.1

83.5 20.3 13.5

– 25.5 13.7

999). The consistency in thicknesses in the ‘‘Depth between ACPs’’ values



http://www.geologie.uni-frankfurt.de/Staff/Homepages/Petschick/RainerE.html

http://www.geologie.uni-frankfurt.de/Staff/Homepages/Petschick/RainerE.html



ages by 1–2.5 ka relative to previous age models (Lyle et al.,2000; Pisias et al., 2001) and has the largest impact aroundthe LGM.

4. Results

4.1. Nd isotopic analyses of river sediment sources

Nd isotopic analyses of Pacific Northwest Rivers arepresented in Table 3. The general features in this datasetare a range in eNd from �10.8 to �0.1 (analytical errors aretypically 0.2–0.4 eNd units). Triplicate analyses of KLA-2,processed separately through column chemistry, show arange of one eNd unit (eNd ¼ �0.670.7). The river sourcesnear the core site show a range of values betweeneNd ¼ �5.0 and �0.1. For the Umpqua River averageeNd ¼ �5.0, whereas the Rogue River has eNd ¼ �1.3 andthe Klamath River values are eNd ¼ �0.9. The largestsediment producer on the western margin of NorthAmerica, the Eel River, has average eNd ¼ �3.0.

Different size fractions were also measured for a fewsamples (COL-5, UMP-1B, KLA-2 and EEL-1B). TheColumbia River shows a notable shift to more radiogenicNd isotopic values in the finer silt (2–20 mm) and clay(0–2 mm), as does the Umpqua River in the 0–2 mmfraction. The Eel and Klamath Rivers do not show anysignificant shift in Nd isotopic values for different sizefractions.

Table 3

Nd isotopic results for Pacific Northwest rivers

Sample River # (Fig. 1C) Lat. (1N) Long. (1W) Siz

QUI-1 1 47.35 124.29 20–

COL-3 4 46.25 123.49 20–

COL-5 4 46.24 123.62 20–

UMP-1A 6 43.70 124.07 20–

UMP-1B 6 43.70 124.07 20–

COO-1C 7 43.38 124.18 20–

ROG-1 8 42.44 124.40 20–

ROG-5 8 42.44 124.38 20–

KLA-2 9 41.52 124.02 20–

KLA-2.2 9 20–

KLA-2.3 9 20–

EEL-1A 10 40.64 124.28 20–

EEL-2 10 40.63 124.28 20–

MAT-1 11 40.31 124.28 20–

RUS-2 12 38.43 123.10 20–

SCN-1 13/14 38.07 121.86 20–

COL-5 4 0–

COL-5 4 2–

UMP-1B 6 0–

KLA-2 9 0–

KLA-2 9 2–

EEL-1B 10 0–

eNd ¼ [143Nd/144Nd(sample)/143Nd/144Nd(CHUR)�1]� 104, where CHUR ¼ p

Analytical uncertainties are in the last reported significant figures.



Previous analyses of Columbia River Nd isotopiccompositions are generally comparable (Goldstein et al.,1984; Goldstein and Jacobsen, 1988) although slightlydifferent size fractions were analyzed. Their findings forbulk 0–63 mm sediment from the Columbia River wereeNd ¼ �5.5 (Goldstein et al., 1984), whereas a measure-ment on mostly clay-sized suspended material was eNd-

�4.5 (Goldstein and Jacobsen, 1988). Our results fromColumbia River were eNd ¼ �6.7 for 20–63 mm material,eNd ¼ �5.1 for 2–20 mm fine silt and eNd ¼ �4.1 for 0–2 mmclays, and generally consistent with these earlier studies.The trend to more radiogenic Nd values with finer materialis consistent with the fact that fine-grained rocks such asbasalts in the Columbia Basin are more radiogenic,whereas the coarser-grained granitic sources (BritishColumbia cordillera, for example) have a less radiogenicNd signature (Ghosh, 1995).

4.2. Nd isotopic analyses of downcore sediments in

EW9504-17PC

We analyzed fourteen samples from the core site for Ndisotopic composition (Table 4 and Fig. 2). Emphasis wasplaced on hemipelagic sediment samples from 25 ka to thePresent. However, the only notable coarse interval(turbidite?) was also analyzed in hopes of determining ifits source differed considerably from the hemipelagicmaterial. Generally, the samples have a small rangedowncore (eNd ¼ �2.5 to �0.9). The largest difference in

e (mm) 143Nd/144Nd Error, 1-s eNd Error, 1-s

63 0.512282 4 �7.0 0.1

63 0.512303 6 �6.5 0.1

63 0.512192 13 �8.7 0.2

63 0.512379 10 �5.0 0.2

63 0.512425 12 �4.1 0.2

63 0.512084 13 �10.8 0.2

63 0.512605 7 �0.7 0.1

63 0.512538 11 �1.9 0.2

63 0.512631 20 �0.1 0.4

63 0.512569 6 �1.4 0.1

63 0.512578 7 �0.2 0.1

63 0.512496 7 �2.8 0.1

63 0.512477 14 �3.2 0.3

63 0.512471 9 �3.3 0.2

63 0.512478 7 �3.1 0.1

63 0.512408 10 �4.5 0.2

2 0.512430 4 �4.1 0.1

20 0.512374 5 �5.1 0.1

2 0.512561 16 �1.5 0.3

2 0.512553 13 �1.7 0.3

20 0.512553 12 �1.7 0.2

2 0.512480 11 �3.1 0.2

resent-day chondritic uniform reservoir value of 143Nd/144Nd ¼ 0.512638.




ARTICLE IN PRESS

Table 4

Downcore bulk Nd isotopic results

Core site EW9504-

17PC, sample depth

(cm)

Age (ka) Size (mm) Total volts, Nd 143Nd/144Nd Error, 1s eNd Error, 1s

1 4 0.3 20–63 1.3 0.512511 10 �2.5 0.2

2 32 2.1 20–63 1.2 0.512531 10 �2.1 0.2

3 50 3.2 20–63 1.3 0.512534 7 �2.0 0.1

4 75 4.8 20–63 1.2 0.512529 9 �2.1 0.2

5 110 7.1 20–63 0.7 0.512545 12 �1.8 0.2

6 140 8.9 20–63 1.5 0.512544 9 �1.8 0.2

7 180 10.6 20–63 1.2 0.512550 11 �1.7 0.2

8 225 12.6 20–63 1.7 0.512545 10 �1.8 0.2

9 260 14.2 20–63 1.8 0.512513 9 �2.4 0.2

10 316 17.8 20–63 1.7 0.512540 8 �1.9 0.2

360 19.0 20–63 – – – – –

11 385 21.0 20–63 1.0 0.512577 13 �1.2 0.2

403 22.3 20–63 – – – – –

12 420 23.6 20–63 1.4 0.512594 10 �0.9 0.2

13 445 25.4 20–63 1.3 0.512541 10 �1.9 0.2

313-Turbidite 17.7 20–63 1.8 0.512521 11 �2.3 0.2

[143Nd/144Nd(sample)/143Nd/144Nd(CHUR)�1]� 104, where CHUR ¼ present-day chondritic uniform reservoir value of 143Nd/144Nd ¼ 0.512638.

Analytical uncertainties are in the last reported significant figures.

40Ar-39Ar Bulk Sediment

Plateau Ages (Ma)

De

po

sitio

na

l Ag

e (C

ale

nd

ar

ag

e,

ka

)

0

2

4

6

8

10

12

14

16

18

20

22

24

26

110 120 130 140

EW9504-17PC

EW9504-17PC

-4.0 -3.0 -2.0 -1.0 0.0

εNd

Fig. 2. Downcore eNd and bulk sediment 40Ar–39Ar ages for core site

EW9504-17PC. Dark gray diamonds are interpolated bulk sediment40Ar–39Ar ages, which are used in the mixing model discussion.


eNd from present-day values occurs over the depth intervalof 385–445 cm, at �22–25 ka, where eNd ¼ �0.9. There isalso a decreasing trend in eNd from 22 to 14 ka (Fig. 2).



4.3. 40Ar–39Ar plateau ages and K/Ca of downcore

sediments in EW9504-17PC

The bulk sediment 40Ar–39Ar and related K/Ca ratiosare presented in Table 5. The downcore 40Ar–39Ar agespectra consistently show plateau ages (Fig. 3), which areoften comparable in their degassing pattern to the spectrafrom nearby river sediments (published previously; seeVanLaningham et al., 2006). The core sediment sampleshave a range in plateau ages from 113.5 to 125.6Ma (notincluding the turbidite) and standard errors of around1.5Ma (2s). Similar to the downcore eNd values, the40Ar–39Ar ages show the largest change over a depthinterval of 403–445 cm, between 22 and 25 ka ago (Fig. 2).The K/Ca ratios, which are a crude indicator of miner-alogy, have similar spectra from sample to sample (Fig. 3).Generally, K/Ca values are around 1.5–2.0 (most likely amixture of kaolinite, K-feldspar and mica) at the lowerends of the spectra and decrease step-wise to values ofaround 0.1–0.3 (mostly plagioclase; see VanLaninghamet al., 2006) at the highest degassing temperatures. TheK/Ca values over age plateau-defining steps have a muchsmaller range of K/Ca ¼ 0.2–0.5 (Table 5).

4.4. Clay mineralogy

Talc-normalized clay mineral abundances are presentedin Table 6 for both river and core samples. Focusing on theclay mineral ratios in the rivers, the Columbia, Umpquaand Coos Rivers north of the core site generally showhigh smectite/illite ratios (range ¼ 2.3–3.1) while fluvialsediment sources at the same latitude and south of thecore site such as the Rogue, Klamath and Eel Rivers havelower smectite/illite ratios ranging between 0.3 and 0.8.




ARTICLE IN PRESSTable

5

Downcore

bulk

sedim

ent40Ar–

39Arages

Core

site

EW9504-17PC,

sample

depth

(cm)

Age

(ka)

Size

(mm)

40Ar–

39Arage

fusion

40Ar–

39Arage

plateau

Error,

2s

Steps

%Gason

plateau

MSWD

K/C

a

plateau

40Ar (r)/

39Ar (K)

fusion

Error,

2s

40Ar (r)/

39Ar (K)

fusion

Error,

2s

14

0.3

20–63

114.8

125.6

3.8

441

0.2

0.4

34.4

1.0

31.4

0.7

232

2.1

20–63

–124.2

Lin

ear

inte

rpola

tion

––

––

––

350

3.2

20–63

–123.2

Lin

ear

inte

rpola

tion

––

––

––

475

4.8

20–63

118.8

122.0

1.9

879

1.5

0.3

39.5

0.6

38.5

0.4

5110

7.1

20–63

116.1

121.7

2.0

764

0.1

0.4

39.7

0.6

37.8

0.5

6140

8.9

20–63

118.3

123.4

2.0

763

1.5

0.3

34.2

0.5

32.7

0.4

7180

10.6

20–63

126.6

130.1

1.7

667

0.4

0.4

36.2

0.4

35.2

0.4

8225

12.6

20–63

120.5

124.4

1.1

557

1.9

0.5

40.2

0.3

38.9

0.1

9260

14.2

20–63

120.9

124.0

1.0

458

0.5

0.4

39.4

0.2

38.3

0.1

10

316

17.8

20–63

118.6

120.8

0.9

877

0.3

0.3

38.0

0.2

37.3

0.2

360

19.0

20–63

114.3

119.0

1.1

556

1.5

0.3

37.7

0.2

31.4

0.1

11

385

21.0

20–63

–116.3

Lin

ear

inte

rpola

tion

––

––

––

403

22.3

20–63

109.8

113.5

1.3

555

6.5

0.2

36.3

0.4

35.0

0.1

12

420

23.6

20–63

–113.6

Lin

ear

inte

rpola

tion

––

––

––

13

445

25.4

20–63

110.0

113.8

0.9

770

2.6

0.3

36.2

0.2

35.0

0.1

313-Turbidite

17.7

20–63

125.7

128.9

0.9

570

0.7

0.2

40.8

0.2

39.8

0.1

S. VanLaningham et al. / Quaternary Science Reviews ] (]]]]) ]]]–]]]10



Kaolinite+chlorite/illite ratios are generally between 1.3and 2.0 with the exception of the Columbia and QuinaltRivers, which show distinctly lower kaolinite+chlorite/illite ratios (0.8 and 1.0, respectively).

5. Discussion

Now we discuss the possible sources of the terrigenouscomponent of marine sediment offshore southern Oregonand northern California, and do so by casting the temporaltrend of provenance changes in terms of the regionalclimate oscillations on glacial–interglacial to millennialtimescales. Fig. 4 shows the relationship of samples fromcore site EW9504-17PC relative to the fluvial sedimentsources in a diagram of eNd vs. 40Ar–39Ar bulk sedimentplateau age. The majority of the core samples plot betweenthe fields defined by the Eel, Rogue and Klamath Rivers,which are large coastal rivers (Table 1 and Fig. 4) proximalto the core site. Three samples, however, suggest anothersource not depicted in Fig. 4. We derive a mixing modelusing the Eel River (eroding the Franciscan Melange),Rogue and Klamath Rivers (eroding the Klamath Accre-tionary Complex and Cascade Volcanic Arc) and theUmpqua River (which drains mostly the Eocene turbiditesof the Tyee Formation and the Cascade Volcanic Arc) asend-members to examine whether mixtures of these fluvialsources describe the downcore eNd–

40Ar–39Ar composi-tional variations. These are the four largest rivers (bydrainage area) within a 400 km swath north and south ofthe core site and, fortunately, drain distinctly differentsource rocks.The mixing model can be used to examine whether

changes in sediment sources are due to changing influencesof these nearby rivers. Four samples analyzed for their Ndisotopic composition did not have 40Ar–39Ar age measure-ments made at the exact same depths and their 40Ar–39Arbulk ages have been interpolated to neighboring points(Fig. 2). Although this produces a slight (and probablyartificial) trend in those data, the 40Ar–39Ar ages fromnearby depths (Fig. 2 and Table 5) are also consistent withthe observed trend, supporting our contention that theseinterpolations reflect the natural variation.

5.1. The mixing model with three sediment sources

We construct a ternary mixing model using the standardbinary mixing equation from Langmuir et al. (1978; seeFaure, 1986 for a summary) described by

RM ¼ ½RAXAf þ RBXBð1� f Þ�=½XAf þ XBð1� f Þ�, (1)

where RM is the isotopic ratio of 40Ar/39Ar or eNd in amixture of two sources (A and B), XA and XB are theconcentrations of 40Ar* or Nd in the end members A andB, while f is the proportion of end member A (f ¼ A/A+B)and (1�f) is the proportion of end member B (following thenomenclature of Faure, 1986). We use the measured40Ar/39Ar ratios (proportional to 40Ar–39Ar plateau ages)




ARTICLE IN PRESS

5.0

4.0

3.0

2.0

1.0

0

40A

r-39A

r A

ge

(M

a)

113.8 ± 0.9 Ma

EW9504-17PC, 445 cm

Depositional Age: 25.4 Ka

EW9504-17PC, 360 cm

Depositional Age: 19.0 ka

119.0 ± 1.1 Ma

120.8 ± 0.9 Ma

EW9504-17PC, 316 cm


124.0 ± 1.0 Ma

EW9504-17PC, 260 cm


EW9504-17PC, 225 cm

124.4 ± 1.1 Ma


EW9504-17PC, 180 cm


130.1 ± 1.7 Ma

121.7 ± 2.0 Ma

EW9504-17PC,110 cm


122.0 ± 1.9 Ma

EW9504-17PC, 75 cmDepositional Age: 4.8 ka

EW9504-17PC, 4 cm


125.6 ± 3.8 Ma

0

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

EW9504-17PC,140 cm


123.4 ± 2.0 Ma

113.5 ± 1.3 Ma

EW9504-17PC, 403 cmDepositional Age: 22.3 Ka

128.9 ± 0.9 Ma

EW9504-17PC, 313 cm Turbidite


%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released

40A

r-39A

r A

ge

(M

a)

%39Ar Released40A

r-39A

r A

ge

(M

a)

%39Ar Released

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100K

/Ca

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

5.0

4.0

3.0

2.0

1.0

00

30

60

90

120

150

180

210

240

0 10 20 30 40 50 60 70 80 90 100

K/C

a

Fig. 3. 40Ar–39Ar incremental heating age and K/Ca spectra for silt-sized (20–63mm) sediments from piston core site EW9504-17PC, offshore southern

Oregon. Age spectra are black while the K/Ca spectra are medium gray. Analytical uncertainties (2-sigma) in age and K/Ca are depicted by the vertical

scaling of each step-age box. The only coarse-grained unit (turbidite?) within the core was also analyzed. It is shown with a light gray background to

distinguish it from the rest of the suite of samples deposited by hemipelagic sedimentation.


and normalize each sample to its unique J factor, which is ameasure of how much K (through 39K) has been convertedto 39Ar during irradiation. We usually discuss the40Ar–39Ar provenance information in terms of age and sothe 40Ar–39Ar ages are shown on a separate axis, alongwith the J-normalized 40Ar/39Ar ratios, recognizing thatthe 40Ar–39Ar ages approximate the true age in this x–y

mixing space. Eq. (1) is used to define a mixing field(through pseudo-binary mixing curves) with three distinctend-members, using the Eel River and the Klamath+Rogue Rivers for one component (right-side line; Fig. 5),the Eel River and the Umpqua River for the secondcomponent (lower line; Fig. 5) and the Umpqua River and



the Klamath+Rogue Rivers for the third component(upper line; Fig. 5). The mixing model parameters areshown in Table 7.The mixing model in Fig. 5 illustrates that indeed the

majority of samples can be adequately described as acombination of the three major sediment sources proximalto the core site. The Klamath Mountains source con-tributed the most sediment to the site except in theyoungest sample (�350 years ago) and at around 14 ka.At those times the Eel River contributed the most sedimentto the mixture (Table 8). The near-surface sample is amixture of 55% coastal California rivers, 34% KlamathMountain rivers and 11% Umpqua River. Present-day




ARTICLE IN PRESS

Table 6

Talc-normalized clay mineral abundances for rivers in the Pacific Northwest and Core Site EW9504-17PC.

Rivers name Lat. (1N) Long. (1W) %sm %ill %k+chl sm/ill k+chl/ill sm/k+chl

Quinalt River, n ¼ 2 47.35 124.29 2 35 35 0.06 1.00 0.06

Columbia River, n ¼ 5 46.25 123.49 19 9 7 2.28 0.77 2.78

Umpqua River, n ¼ 3 43.70 124.07 20 7 15 3.06 2.25 1.35

Coos River, n ¼ 3 42.44 124.40 19 7 13 2.97 1.92 1.59

Rogue River, n ¼ 3 42.44 124.38 14 18 32 0.82 1.82 0.44

Klamath River, n ¼ 4 41.52 124.02 6 20 35 0.33 1.73 0.18

Eel River, n ¼ 3 40.64 124.28 10 22 37 0.43 1.64 0.26

Mattole River, n ¼ 3 40.31 124.28 5 14 26 0.35 1.92 0.18

Russian River, n ¼ 2 38.43 123.10 20 16 22 1.30 1.33 0.95

Sacramento system, n ¼ 6 38.07 121.86 10 14 27 0.75 1.87 0.36

Core depth (cm) Age (ka)

1 4 0.3 6 17 25 0.38 1.52 0.25

2 32 2.1 7 15 22 0.44 1.51 0.29

3 50 3.2 7 20 30 0.36 1.53 0.24

4 75 4.8 4 13 20 0.33 1.54 0.22

5 110 7.1 17 31 52 0.56 1.69 0.33

6 140 8.9 5 16 34 0.32 2.07 0.15

7 180 10.6 8 18 32 0.44 1.83 0.24

9 260 14.2 8 15 30 0.54 1.93 0.28

10 316 17.8 8 15 26 0.51 1.67 0.30

10b 360 19.0 12 16 28 0.75 1.70 0.44

11 385 21.0 15 18 26 0.82 1.47 0.56

11b 403 22.3 14 16 22 0.88 1.38 0.64

12 420 23.6 11 15 23 0.74 1.52 0.49

13 445 25.4 13 18 27 0.71 1.52 0.47

313-Turbidite 17.7 8 12 22 0.61 1.77 0.34


sediment loads are inconsistent with these values (Karlin,1980; Wheatcroft and Sommerfield, 2005; Table 1). Intotal, 78% of the total present-day sediment load is fromcoastal California (Eel, Mad, Redwood Cr., Trinity R.),17% from rivers eroding the Klamath Mountains (Kla-math, Rogue, Smith) and 6% from the Oregon CoastRange Rivers (Umpqua, Coos-Coquille-Sixes-Elk).

Three samples (two have interpolated 40Ar/39Ar ratios)from around the Last Glacial interval do not fall within theternary mixing field nor do they trend toward any fluvialsource shown in Fig. 4. The 40Ar–39Ar ages are youngerbut with eNd values that are higher. The dominant geologicprovinces along the margin that could contribute younger,more radiogenic source rocks are the Columbia RiverBasalts and volcanic rocks in the Cascade Range. TheColumbia River could be a considerable source of sedimentderived from both of these geologic provinces, especiallywhen the Late Pleistocene Missoula floods sculpted theColumbia Basin (Bretz, 1969; Benito and O’Conner, 2003)and sediment from these events was being transportedacross the northeast Pacific Ocean (Zuffa et al., 1997;Wolf et al., 1999). However, the Cordilleran rocks fromthe continental interior (British Columbia, Montanaand Idaho) dominate the sediment signature of theMissoula floods in offshore sediments from Astoria Fan(Duncan et al., 1970; Kulm et al., 1973; Prytulak et al.,2006), with Nd isotopic values of around eNd ¼ �8 to �12(Prytulak et al., 2006). Therefore, if the Missoulafloods were responsible for the change in provenance at



EW9504-17PC around the time of the Last Glacial, thesediments would be expected to reflect older 40Ar–39Arages and less radiogenic Nd isotopic values, which is notthe case.

5.2. Sediments derived from the Cascade Volcanic Arc?

We further examine the mixing of sediment sources bydetermining whether the samples trending away from thedefined mixing field are consistent with mixing withthe Cascade Volcanic Arc 40Ar–39Ar ages and eNd values.The non-Cascade Range lower end member is simplydefined as the average of the core samples that are not fromthe Last Glacial (points 1–9; Fig. 5). With reasonable eNd

and Nd concentration parameters assigned to the Cascadesusing the GEOROC database (http://georoc.mpch-mainz.gwdg.de/georoc/), and 40Ar–39Ar ages and 40Ar*estimates calculated from the weighted average of mappedunits in the Umpqua Basin (VanLaningham et al., 2006),the 22–25 ka core samples fall directly on a mixing linetowards the Cascade Range and suggests an increase inCascades erosion relative to the coastal mountain ranges.

5.3. An anthropogenic bias?

A mismatch between the present-day 40Ar–39Ar agesmeasured from sediments at the mouth of the KlamathRiver and the predicted bulk sediment 40Ar–39Ar age(VanLaningham et al., 2006) offers us another avenue to



http://georoc.mpch-mainz.gwdg.de/georoc/

http://georoc.mpch-mainz.gwdg.de/georoc/


ARTICLE IN PRESS

40 60 80 100 120 140 160 180

-12.0

-10.0

-8.0

-6.0

-4.0

-2.0

0.0

2.0

40Ar-39Ar Bulk Sediment Plateau Age (Ma)

Klamath Accretionary Complex+

Cascade Headwaters

Melange

Tyee+

Cascade

headwaters

B.C. Cordillera/

Idaho Batholith/

Cascades/

CRB's

Tyee Fm.

Umpqua R.

Klamath R.

Rogue R.

Sacramento R.

Columbia R.

Quinalt R.

Coos R.

EW9504-17PC

Mattole R.

Olympic Mtns

Accretionary

Complex

ε Nd

Eel R.

Sierra Nevada/

Great Valley

Franciscan

Russian R.

Fig. 4. River and core samples in eNd–40Ar–39Ar bulk sediment plateau age space. Samples are plotted with 2-sigma analytical uncertainties on the Ar

plateau age weighted mean and epsilon Nd mean (The Columbia River Ar–Ar ages are total fusion ages since consistent age plateaus did not develop in

these samples). Samples with white-filled symbols denote rivers south of the core site while dark-filled symbols indicate rivers north of the core site. Ovals

around groups of fluvial samples indicate sediment sources with similar source rock combinations. For example, the Klamath and Rogue Rivers both

drain Klamath Accretionary Complex and Cascade Mountain lithologies, although in differing proportions. The core samples are also plotted (black

diamonds) and their location on this diagram suggest that sediments at this core site derive from the proximal sources of the Eel, Klamath and Rogue

Rivers, generally.


explore within the framework of our mixing modelapproach. Is the lack of present-day sediment contribu-tions from the eastern Cascades in the Klamath Basinrelated to the low relief of the Upper Klamath Basin or is ita reflection of land-use practices in the Klamath Moun-tains (Lower Klamath Basin)? It has been suggested thatgold mining and timber harvesting beginning in the 1800 sdrastically affected sediment production in the KlamathMountains (Nolan and Janda, 1995; Sommerfield andWheatcroft, in press). Of all the rivers in the study area thatmight reflect an anthropogenic bias in its isotopiccomposition of sediments, the Klamath River is one ofthe best candidates because about half of its basin lies eastof the Cascades in a fairly low-relief region while the otherhalf is in the high-relief, coastal Klamath Mountains(VanLaningham et al., 2006). This is in contrast to theRogue and Umpqua Rivers, for example, that erodetopography that has more similar relief structures in thediffering geologic provinces of the western Cascades andthe coastal Klamath Mountains and Oregon Coast Ranges.Moreover, because the Eel River erodes Franciscan



Melange along its entire length, erosion related to land-use might change its sediment flux but is unlikely to changeits provenance signature. If recent land-use practices arethe explanation, then the end members may not beproperly defined.

5.4. An improved mixing model

Because of this potential for an anthropogenic bias in thepresent-day samples and the fact that we cannot reconcilethe percent contributions calculated for the near-surfacesample in the Ar–Nd values with the river sediment loads,we recast the ternary Ar–Nd isotopic mixing model byincorporating the 40Ar–39Ar model predicted age of109Ma (from VanLaningham et al., 2006) rather thanthe measured 40Ar–39Ar values, and derive a new, drainagearea-weighted average age for the Rogue+Klamath Riversof 112Ma (instead of 140Ma). Even though the Ndisotopic signatures should also change by adding moreCascades material to the Klamath River pre-land use, wedo not make an attempt to correct for this for a couple of




ARTICLE IN PRESS

0.05 0.055 0.06 0.065 0.07 0.075 0.08 0.085-6

-5

-4

-3

-2

-1

0

40Ar/39Ar Ratios, J normalized (using 40Ar*)

ε Nd

Nd

εεNd = -3.2

εNd = -4.6

19

43

2

105 6 8

7

11

12

13

A

Approximate 40Ar-39Ar Ages

Oregon Coast Range Rivers

Coastal California Rivers40Ar-39Ar Age = 129 Ma

Klamath Mtns. Rivers40Ar-39Ar Age = 140 Ma

εNd = -0.9

80% Umpqua

60%Umpqua

40% Umpqua

20% Umpqua

20% Eel

80%

Um

pqua

20%

Kla

mat

h

60%

Um

pqua

40%

Kla

mat

h

40% U

mpqua

60% K

lamath

20% Umpqua

80% Klamath

40% Eel

60% Eel

80% Klamath

20% Eel

60% Klamath40% Eel

40% Klamath60% Eel

20% Klamath80% Eel

80% Eel

Turbidite

0.01 0.07

-4

0

+4C

asc

ades

Vo l

canic

Arc

B

CascadesVolcanic Arc

40Ar-39Ar Ages

40Ar-39Ar Age = 92 Ma

90 95 100 105 110 115 120 125 130 135 140 145

20 80 140

40Ar/39Ar

Fig. 5. Ternary mixing model for terrigenous sediments offshore Oregon and California. Dotted lines represent 20% increments of end-member

contributions assuming no differential erosion within basins of each end-member: (A) Sediment source end members to the core site are defined in40Ar/39Ar–eNd space. 40Ar–39Ar ages corresponding to the J-normalized 40Ar/39Ar ratios are shown at the upper horizontal-axis. The end members are

defined by: The Umpqua and Coos Rivers (‘‘Oregon Coast Range Rivers’’, north of the core site), the Klamath and Rogue Rivers (‘‘Klamath Mtns.

Rivers’’, �latitude of core site) and the Eel River (‘‘Coastal California Rivers’’, south of the core site). Core samples from EW9504-17PC are depicted as

medium gray-filled circles. Core samples that are interpolated in their 40Ar/39Ar ratios are white-filled circles. The numbers correspond to increasing depth

(or depositional age) with ‘‘1’’ being the shallowest (or youngest) and ‘‘13’’ being the deepest (oldest) in the core. A mixture of these three fluvial end-

members adequately describes the majority of core samples. However, samples 11–13 are not easily described in terms of these fluvial sources only. A

fourth component is needed (B) and a plausible source is the Cascade Volcanic Arc.

Table 7

End member compositions for Ar–Nd mixing model

End members Concentrations Isotope ratios

40Ar* Nd 40Ar/39Ara eNd

Present-day

Umpqua River 9.3E�13 19.0 0.0502 �4.6

Rogue+Klamathb 5.1E�13 24.5 0.0792 �0.9

Eel River 1.1E�12 19.0 0.0741 �3.2

LGM

Cascadesc 2.5E�13 15.0 0.0117 +4.1

Average of non-LGM samples from 17PC 6.1E�13 19.0 0.0711 �2.0

aThe 40Ar/39Ar ratios were multiplied by the J factor (in essence the dosimeter for irradiated material) for each sample.bAssuming that the modeled Ar age in VanLaningham et al. (2006) is the pre-land-use value.cThe 40Ar* was estimated from K concentrations in the GEOROC databadse.


reasons. We have not developed a weighted-averageapproach with the Nd system because of the paucity ofNd isotopic values in the host rocks. Because the range of



Nd isotopic variability is much smaller than the Ar ages inboth river and core sediments from this study, the effects ofland-use practices would have a significantly reduced




ARTICLE IN PRESS

Table 8

Percent contributions from river end-members, present-day values

EW9504-17PC sample

depth (cm)

Age

(ka)

Eel River

(%)

Klam+Rogue

(%)

Umpqua

(%)

1 4 0.3 55 34 11

2 32a 2.1 27 58 15

3 50a 3.2 19 63 18

4 75 4.8 20 60 20

5 110 7.1 5 73 22

6 140 8.9 10 72 18

7 180 10.6 30 65 5

8 225 12.6 14 71 15

9 260 14.2 43 42 15

10 316 17.8 7 69 24

13 445 25.4 0 67 33

313-Turbidite 17.7 60 37 3

LGM % Increase in cascade-derived

sediment

11 385a 21.0 16

12 420a 23.6 22

aInterpolated 40Ar/39Ar ratio (see text).


impact on the Nd component of the mixing field. Fig. 6shows that the downcore samples are generally oscillatingbetween the Eel and Rogue+Klamath Rivers throughtime. The samples numbered 11, 12 and 13, whichpreviously plotted outside of the ternary mixing field,now lie within the new mixing field and yet remainconsistent with the idea that there was an increase inCascade-derived material (Fig. 6B) because they are alsoon a mixing line between the Cascades and coastalCalifornia sediment sources. The mixing model predictionsof 74%, 22% and 4% from the Coastal California,Klamath Mountain and Umpqua end-members, respec-tively, in the near-surface sample are also more consistent(Table 9) with the present-day sediment load values(Wheatcroft and Sommerfield, 2005) of 78% from coastalCalifornia (Eel, Mad, Redwood Cr., Trinity R.), 17% fromrivers eroding the Klamath Mountains (Klamath, Rogue,Smith) and 6% from the Oregon Coast Range Rivers(Umpqua, Coos-Coquille-Sixes-Elk). Therefore, this newmixing model is more consistent with observations.

The change in sediment source to more KlamathMountains-derived sediment is likely related to differentialerosion in the Klamath Basin. However, it could be arguedthat the change may be from a significant shift in oceancirculation. Yet, this would require a �90% reduction insediment from the Eel River (as deduced from the mixingmodel). The commensurate increase in the contributionfrom Klamath Mountain rivers around 22–25 ka requireseither a complete shutdown of erosion in the Eel Riverbasin (highly unlikely), a reversal in the sediment transportdirection offshore, or an extreme change in the dispersivityof sediment along the margin.

If a change occurred in sediment transport directionfrom present-day south-to-north transport, then we mightexpect a change to more negative eNd values, because the



Oregon Coast Ranges and Columbia River are much lessradiogenic at eNd ¼ �7 to �11 (Fig. 4) and all sourcesnorth of the core site are less than eNd ¼ �5. The downcorevalues trend in the opposite direction (eNd ¼ �1.0 at theLGM), leading to the best explanation that an increase inthe supply of Cascades material through the Klamath,Rogue and Umpqua River systems.

5.5. Provenance linkages to glacial erosion

The downcore provenance records (40Ar–39Ar ages andeNd) can be compared with other climate proxies in coreEW9504-17PC and nearby marine and terrestrial recordsto investigate how the larger climate system might havedriven the provenance change. The previous sectionillustrated that Cascades-derived material more heavilyinfluenced the EW9504-17PC sediment signature justbefore the LGM and that a transition in provenance tomore present-day type sediment occurred from 22 to 14 ka.A comparison of these findings with a record from UpperKlamath Lake supports this interpretation. This 14C-datedrecord shows a large flux of glacial flour to Upper KlamathLake, which lies at the base of the eastern flanks of theOregon Cascades (Fig. 1) and feeds into the Klamath River(Rosenbaum and Reynolds, 2004). This record suggeststhat the flux of Cascades-derived material increased by upto a factor of four (relative to the pre-LGM values) duringglacial advance and retreat 22–14 ka (Fig. 7). At the sametime, core EW9504-17PC shows a notable linear change inthe 40Ar–39Ar ages, eNd, and clay mineralogical provenancerecords. Moreover, there is a nearly threefold increase(spike) in sedimentation rate from a background value of�15 cm/ka to over 75 cm/ka, coincident with the maximumglacial flour flux from Upper Klamath Lake. Although aturbidite in the core coincides with the spike in thesedimentation rate curve, it is only 2 cm thick and thus,accounts for very little of the total increase observedaround this time. We also recognize that a range of naturalprocesses complicate the interpretation of oceanic accu-mulation rates, not to mention additional uncertaintiesprovided by age model errors.If a large flux of 10–20Myr old, radiogenic (higher eNd

values) Cascades-derived material was being deliveredthrough the Klamath Basin at this time, then why do the40Ar–39Ar ages increase and the eNd values decrease at themarine core site during the LGM and deglaciation? It ispossible that, at the onset of glaciation in the CascadeRange (22–18 ka; Rosenbaum and Reynolds, 2004) therewas a concurrent increase in glacial flour flux from theTrinity Alps and other high topographic regions in theKlamath Mountains. The oldest plutons (�400Ma) inthe Klamath Mountains are exposed in the Trinity Alps(Irwin and Wooden, 1999) but they cover only a smallpercentage of the total basin (o5%). The surroundingglaciated rocks have an average 40Ar–39Ar age of around150Ma (VanLaningham et al., 2006). A large influx of thisolder material into the Klamath River could compensate




ARTICLE IN PRESS

0.05 0.055 0.06 0.065 0.07 0.075-6

-5

-4

-3

-2

-1

0

40Ar/39Ar Ratios, J normalized (using 40Ar*)

90 95 100 105 110 115 120 125 130

40Ar-39Ar Ages

0.01 0.08

-4

0

4

20 14040Ar-39Ar Ages

40Ar/39Ar

ε Nd

ε Nd

19

43

2

105 6 8 7

11

12

13

A

80% Umpqua

60% Umpqua

40% Umpqua

20% Umpqua

20% Eel

80%

Um

pqua

20%

Kla

math

60%

Um

pqua

40%

Kla

math

40%

Um

pqua

60%

Kla

mat

h

40% Eel

60% Eel

80%

Kla

math

20%

Eel

60%

Kla

math

40%

Eel

40%

Kla

mat

h

60%

Eel

80% Eel

Turbidite

Cascades

Volc

anic

Arc

B

εNd = -3.2

εNd = -4.6

εNd = -0.9

Oregon Coast Range Rivers

Coastal California Rivers

Klamath Mtns. Rivers(Modeled Klamath R.)40Ar-39Ar Age = 112 Ma



Fig. 6. Ternary mixing model as in Fig. 5 except with a modeled 40Ar–39Ar age for the Klamath River, which we use as a pre-European settlement value:

(A) sediment source end members to the core site are defined in 40Ar/39Ar–eNd space. Corresponding 40Ar–39Ar ages are shown at the upper horizontal-

axis. The end members are defined by: the Umpqua and Coos Rivers (‘‘Oregon Coast Range Rivers’’, north of the core site), the Klamath and Rogue

Rivers (‘‘Klamath Mtns. Rivers’’, �latitude of core site) and the Eel River (‘‘Coastal California Rivers’’, south of the core site). Core samples from

EW9504-17PC are depicted as medium gray-filled circles. Core samples that are interpolated in their 40Ar/39Ar ratios are white-filled circles. The numbers

correspond to increasing depth (or depositional age) with ‘‘1’’ being the shallowest (or youngest) and ‘‘13’’ being the deepest (oldest) in the core. This

illustrates that the downcore samples can be described as a mixture between rivers of the Klamath Mountains and those from coastal California such as

the Eel River. The fourth, Cascades component (B) is no longer needed to explain the samples from 22 to 25 ka. But most samples do lie on a mixing line

toward the coastal California river signature, which remains consistent with a differential erosion component contributing to the provenance evolution.

Table 9

Percent contributions of river end-members, modified Klamath Basin

EW9504-17PC sample

depth (cm)

Age

(ka)

Eel

River

(%)

Klam+Rogue

(%)

Umpqua

(%)

1 4 0.3 74 22 4

2 32a 2.1 59 39 2

3 50a 3.2 55 42 3

4 75 4.8 52 42 6

5 110 7.1 44 53 3

6 140 8.9 49 51 0

7 180 10.6 – – –

8 225 12.6 54 46 0

9 260 14.2 65 30 5

10 316 17.8 43 52 5

11 385a 21.0 17 81 2

12 420a 23.6 7 93 0

13 445 25.4 25 67 17

313-Turbidite 17.7 – – –

aInterpolated 40Ar/39Ar ratio (see text).




for the young ages contributed by the Cascades and serveto offset any trend toward younger bulk sediment ages.Records of Klamath Mountain glaciation are sparse and

timing is not well constrained (Sharp, 1960; Woods, 1976;Porter et al., 1983; Bevis, 1995). However, cirque-basedequilibrium line altitude (ELA) estimates suggest thatglaciers were at �1500m, while ELA estimates for thesouthern Oregon Cascades suggest it was �2000m (Porteret al., 1983). This is likely to have led to a major increase insediments derived from the older source rocks and to thetemporal trend observed downcore in the bulk sediment40Ar–39Ar ages and eNd. Moreover, downcore (chlorite+kaolinite)/illite ratios from the marine site increase linearlyfrom 22 to 14 ka and are also consistent with an increase insediments from the chlorite-rich Klamath Mountains inthat period (Fig. 7C).Other mechanisms that would change the type and

amount of material from the terrestrial source region arepossible. The Ar–Nd mixing model suggests an increase in




ARTICLE IN PRESS

% Full Glacial Average

20 40 60 80 1000

Glacial Flour Flux

kg/m /yr

0.0 0.1 0.2 0.3 0.4 0.530

0 0.2 0.4 0.6

EW9504-17PC, Radiolaria“California Current”

Factor Score

-0.1 0.1 0.3 0.5 0.7

Radiolaria “Bering Sea”Factor Score

A B

E

JH IGF

Ar- Ar Bulk SedimentPlateau Ages (Ma)

Depositio

nal A

ge (

Cale

ndar

age, ka)

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

110 120 130

EW9504-17PCEW9504-17PC

EW9504-17PC

EW9504-17PC

-4.0 -3.0 -2.0 -1.0 0.0

C D0 5025 75

Sedimentation Rate (cm/kyr)

Upper Klamath Lake

200 40

Little Lake, Oregon

Coast Ranges

Upper

Klamath Lake

50

Spruce Pollen, %Spruce Pollen, %

Spruce Pollen, %

0 10 20

5%

Lin

e

5%

Lin

e

5%

Lin

e

EW9504-17PC

1.00 1.50 2.00

Chlorite+Kaolinite/Illite

EW9504-17PC

εNd

Fig. 7. Northeast Pacific–Pacific Northwest climate and provenance records. A notable change occurs in EW9504-17PC 40Ar–39Ar plateau ages (A)

during the interval from 22ka until 14 ka (transparent gray band). A more subtle change throughout the same time period is seen in the eNd values

downcore. Although less diagnostic throughout the entire record, chlorite+kaolinite/illite ratios (C) show an increase over the same interval. A peak in

sedimentation rate (D) from EW9504-17PC coincides precisely with the peak in glacial flour flux from Upper Klamath Lake (E), which drains into the

upper Klamath River and is carried down to the northeast Pacific margin at 41.51N. Spruce pollen from Upper Klamath Lake (G) in the southern Oregon

Cascades (Hakala and Adam, 2004) and Little Lake (H) in the Oregon Coast Ranges (Worona and Whitlock, 1995) show that the increase in spruce pollen

downcore (F) is likely a combined affect of increased transport of pollen from coastal regions as well as pollen succession. Radiolaria factor scores (see

Pisias et al., 2001) represent the oceanic response to climate in the northeast Pacific. The peak and reduction in the ‘‘Bering Sea’’ radiolaria assemblages (I)

corresponds with the LGM/deglacial, while strengthening of the California Current (J) occurs after 14 ka. From these records, it appears that there is a

tight coupling between terrestrial and ocean climate.


Eel River material at younger downcore bulk 40Ar–39Arages (Table 8). Because the Eel River is the dominantsediment producer presently, and because precipitation wasenhanced around the Eel River region during the LastGlacial (Adam and West, 1983), it is possible that theprovenance change between 22 and 14 ka is also related toan increase in Eel River sediment to the margin; thiscomplicates our interpretation that a significant influencewas imparted from glacial erosion. The competing influ-ences of different sediment sources to the Oregon–Cali-fornia margin over this time interval are currently beingtested (VanLaningham, N.G. Pisias, S. Hostetler, R.A.Duncan, Exploring climate-driven erosion through a40Ar–39Ar detrital mixture model: A sensitivity test of

Pacific Northwest rivers to glacial–interglacial hydrologic

changes, in preparation).Secondary processes may also have been influential on

the Ar–Nd record including storage of material previouslyeroded from the Klamath Accretionary Complex, andremobilization of that material during the period of highglacial flour flux (and water?) from the Cascades. It is alsopossible that the higher concentration of suspendedCascades material led to enhanced erosion in the highergradient bed of the lower stretches of the Klamath Riverthrough an improved ability to abrade the river channel(Sklar and Dietrich, 1998)?



We emphasize that the relative changes between the Eel,Rogue+Klamath and Umpqua basins presented in Tables 8and 9 are only useful in a general way because the change inprovenance may be largely due to differential contributionsfrom different parts of these drainage basins. The Ar–Ndisotopic mixing model does not quantitatively address thispotential for differential erosion since it only accounts forchanges in the proportions of the three major sources.

5.6. A provenance change at 10 ka: changing ocean

circulation or sediment flux?

In the downcore record of 40Ar–39Ar plateau ages, anotable shift also occurs at 10.6 ka (Fig. 7a) in addition tothe change during the time of the Last Glacial. Althoughonly one sample has a significantly older 40Ar–39Ar bulksediment age than its neighbors, there is a trend towardolder ages on each side of the sample at 10.6 ka, suggestingthat a real provenance change may have occurred. TheAr–Nd isotopic data suggest an increase in the incorpora-tion of present-day Klamath+Rogue River (Fig. 6)material at this time. A variety of processes, both in theterrestrial and marine realms could have led to thisprovenance change.In the terrestrial environment, one explanation might

relate to the aggradational event recorded in river terraces





of the Oregon Coast Ranges. Several coastal Oregon riverfloodplains archive anywhere from 2 to 11m of sedimentabove a regionally expansive strath terrace (Personiuset al., 1993). The strath-forming event has been dated atapproximately 10.4 ka (average of nine calendar-calibratedradiocarbon ages from four rivers in Personius et al., 1993),contemporaneous with the 10.6 ka peak in 40Ar–39Arplateau ages. Because this increase in sediment productionwas regional (spanning across at least the central andsouthern Oregon Coast Ranges), it is thought that themechanism was climate driven (Personius et al., 1993). Anincrease in precipitation may have led to increaseddenudation of the coastal mountain landscape (Personiuset al., 1993). Alternatively, drying (drought?), whichcould have led to extreme events like forest fires wouldalso encourage denudation of the coastal mountain land-scape by reducing the amount of hillslope-stabilizingvegetation.

The terrestrial pollen records of Little Lake, Oregon(Worona and Whitlock, 1995), in Upper KlamathLake and in nearby Tulelake (Hakala and Adam,2004), during 12–9 ka support the hypothesis that drierconditions prevailed and may have led to a suddenlandscape denudation response. Alder pollen, an indicatorof a land surface disturbance such as fires or landslides(Pisias et al., 2001), shows a subdued maximum in both thedowncore EW9504-17PC abundance record as well as inthe Little Lake pollen history (Worona and Whitlock,1995).

Alternatively, a shift in ocean circulation, wherein areduction in advection of southerly water (and sediment)northward, could have also produced the observed changein Ar–Nd isotopic data at 10.6 ka. It would have led to anincreased proportion of more Klamath and Rogue Riversediment (older bulk sediment ages and higher eNd) relativeto the Eel River sediments. A fivefold increase in theabundances of subarctic diatom species (Lopes, 2006)suggests that a change in ocean circulation may haveoccurred around this time yet Radiolaria species (Pisiaset al., 2001) do not show any notable changes around10 ka. Was there advection of northern waters, whichwould have resulted in a reduction in the influence of theEel River and an increase in more Klamath-derivedsediment? This, however, might have led to less evapora-tion of ocean waters (reduced air–sea temperature gradi-ent), which in turn might have reduced or produced littlechange in precipitation. Although denser sampling at thecore site around this time interval is needed to resolvewhether the marine realm was perturbed by, and/orcaptures the terrestrial record of a 10 ka ‘‘event’’, thetentative relationships between terrestrial and oceanicsystems seen here suggest that a climatically driven pulseof precipitation (or extreme drought?), erosion andpossibly advection of northerly waters related to anaberration in the climate system may have occurred aroundthe Pleistocene–Holocene transition in the Pacific North-west–northeast Pacific region.



5.7. Provenance and the pollen record

Pollen records preserved in marine sediments record notonly vegetation succession related to climate change, butalso surface processes related to the transport of pollen tomarine sedimentary basins. In this final section, we brieflyexplore the meaning of downcore spruce pollen abundancechanges seen in core site EW9504-17PC (from Pisias et al.,2001) and nearby terrestrial lake cores. As we have alreadynoted, provenance signatures indicate that rivers trans-ported a greater proportion of material from the CascadeRange to the core site between 22 and 25 ka relative totoday. Then, from 22 to 14 ka, glaciers advanced andretreated, leading to a large influx of both eroded Cascadesmaterial and Klamath Mountain sediment (and possiblydetritus from the Eel River region), judging from thesimultaneous increase in 40Ar–39Ar ages and decrease ineNd values. The spruce pollen abundances from core siteEW9504-17PC shows a broad peak around the LGM(Fig. 7f), which could be interpreted as either a change intemperature/rainfall or a change in the flux of continentalmaterial (sediment and pollen) from the Cascade Rangeand coastal mountains.Spruce pollen abundances in Upper Klamath Lake

(Hakala and Adam, 2004) in the Klamath River Basinare generally low (0–5%), have a broad increase at theLGM and also have considerably lower abundances fromthose seen at the core site (Fig. 7). The marine sprucepollen record in EW9504-17PC has a more notable increasestarting at �23 ka to greater than 15% spruce and adecrease to 0–5% by 14 ka. Where does this spruce pollencome from? In the absence of a pollen record from theKlamath Mountains, we look to a lacustrine record from�250 km north of the core site in the central Oregon CoastRanges (Fig. 1A). This record suggests that the higherspruce pollen abundances in EW9504-17PC may besourced from the coastal areas rather than from theCascades. If so, then these records, in combination withwhat we have learned from the provenance records, suggestthat the spruce pollen in EW9504-17PC track thedifferential erosion patterns in the region and thus recorda complex signal of climate change and erosion/transportchanges in the terrestrial environment.

6. Conclusions

We have invoked a combination of 40Ar–39Ar plateauages and Nd isotopic compositions to characterize bulk,silt-sized fluvial and marine sediment in the Pacific North-west of North America. This has been done to investigatesediment transport changes related to differential erosionand ocean circulation and what this means in termsof terrestrial–ocean climate linkages. Bulk sediment40Ar–39Ar ages and eNd from rivers in the Pacific North-west have resolvably different characteristics that willbe useful to future studies along this continental margin.The coupled Ar–Nd isotopic technique provides robust





information about the terrestrial source of sediments onthe continental margin and provides a multi-tracerfingerprinting technique that is resistant to alterationprocesses.

At core site EW9504-17PC, sediment can be described asa mixture from the Klamath and Rogue Rivers (equi-latitudinal with the core site), the Eel River (south of thecore site) and the Umpqua River (north of the core site).Land-use practices in the Klamath River Basin appear tohave amplified the potential for problems in defining fluvialfingerprints from present-day river samples in that riversystem. Specifically, this amplification of land use relates tochanges in sediment production from the high-reliefKlamath Mountains relative to the low-relief easternCascades/high desert part of the Klamath River. With thisin mind, we conclude that the downcore sedimentsrepresent an oscillating mixture between the Klama-th+Rogue Rivers and the Eel River. The Umpqua andother coastal rivers north of 421N have had only a smallinfluence on sedimentation at the core site over the last25 ka. We also emphasize that there is no evidence that anysignificant amount of Columbia River material wasdeposited at the core (even during the Missoula Floods),based on the Ar–Nd isotopic compositions as well as claymineralogy.

Cascade Range-derived material from the headwaters ofthe proximal rivers to the core site (Klamath and RogueRivers) had a more significant sediment contributionbefore 22 ka relative to today, implying reduced precipita-tion in the coastal mountains or an increased ability ofCascade Mountain glaciers to erode the landscape. More-over, the presence of a pluvial lake in the headwater regionof the Klamath River was also potentially influential to theprovenance changes seen downcore. During the LGM andsubsequent deglaciation (22–14 ka) a commensurate in-crease in glacial erosion in the Klamath Mountainsintroduced more material from the Klamath AccretionaryComplex, leading to a net increase in bulk sediment40Ar–39Ar ages. An increase in material from othersediment sources such as the Eel River may have alsocontributed to older bulk sediment ages observed at thecore site.

The comparison of pollen records, in light of theinformation offered by the provenance data, illuminatesthat spruce pollen abundances in this core site reflect theintegrated signal of terrestrial vegetation succession (whichresponds to precipitation+temperature) as well as varia-tions in erosional flux (which also responds to precipita-tion). Because other pollen species are similarly affected,downcore pollen abundances reflect a complex integrationof precipitation-related processes.

Future studies can exploit the fact that the rivers in thisregion carry information related to erosion in the interiorCascades volcanic arc as well as the coastal mountains.Because the Cascades were glaciated during MarineIsotope Stages (MIS) 2, 4 and 6 and the KlamathMountains were only glaciated during MIS 2 and 6 (Bevis,



1995), future studies can assess the landscape response tothresholds in the climate system, glacial erosion, pollensuccession, sediment transport and oceanic changes.Furthermore, future studies linking provenance and pollenrecords in a suite of core sites to the north and south willyield a more complete understanding of terrestrial, oceanand atmospheric changes where the North Pacific andAlaskan Gyres intersect along western North America.

Acknowledgments

The authors would like to thank the reviewers of thismanuscript, as their comments greatly improved the finalpaper. We thank Andy Ungerer, Mysti Weber, AdamKent, Frank Tepley, Chris Russo, and John Huard fortechnical assistance. We are thankful to Alan Mix andSteve Hostetler for conversations related to PNW-NEPacific climate. Chris Goldfinger offered us insight aboutsediment transport processes in Cascadia. Although notdirectly cited, we are grateful for the body of work fromLarry Krissek, Ken Scheidegger, Verne Kulm and PaulKomar. They have paved the way for sedimentologicalstudies in the NE Pacific. Mitch Lyle and Walt Dean werehelpful in supplying data from their published works,which invigorated our interpretations of the data presentedhere. We thank Claire Schuft and Moya Duncan for helpwith sample preparation. This work was made possible byNSF Grant ATM0135294.

References

Adam, D.P., West, G.J., 1983. Temperature and precipitation estimates

through the last glacial cycle from Clear Lake, California, Pollen Data.

Science 219, 168–171.

Benito, G., O’Conner, J.E., 2003. Number and size of last-glacial Missoula

floods in the Columbia River valley between the Pasco Basin,

Washington, and Portland. Oregon Geological Society of America

Bulletin 115 (5), 624–638.

Bevis, K.A., 1995. Reconstruction of Late Pleistocene paleoclimatic

characteristics in the Great Basin and adjacent areas. Ph.D. Thesis,

Oregon State University, 277pp.

Blake, M.C., Jones, D.L., 1981. The Franciscan melange in northern

California: a reinterpretation. In: Ernst, W.G. (Ed.), The Geotectonic

Evolution of California. Prentice-Hall, Englewood Cliffs, NJ,

pp. 306–328.

Brandon, M.T., Vance, J.A., 1992. Tectonic evolution of the Cenozoic

Olympic subduction complex, Washington state, as deduced from

fission track ages for detrital zircons. American Journal of Science 292,

565–636.

Bretz, J.H., 1969. The Lake Missoula floods and the Channeled Scabland.

Journal of Geology 77, 505–543.

Clift, P.D., 2006. Controls on the erosion of Cenozoic Asia and the flux of

clastic sediment to the ocean. Earth and Planetary Science Letters 241,

571–580.

Clift, P.D., Blusztajn, J., 2005. Reorganization of the western Himalaya

river system after five million years ago. Nature 438, 1001–1003

doi:10.1038.

Duncan, J.R., Kulm, L.D., Griggs, G.B., 1970. Clay mineral composi-

tion of the Late Pleistocene and Holocene sediments of the

Cascadia basin, northeastern Pacific Ocean. Journal of Geology 78,

213–221.

Faure, G., 1986. Principles of Isotope Geology. Wiley, New York, 589pp.





Ghosh, D., 1995. Nd–Sr isotopic constraints on the interactions of the

intermontane Superterrane with the western edge of North America in

the southern Canadian Cordillera. Canadian Journal of Earth Sciences

32, 1740–1758.

Goldstein, S.J., Jacobsen, S.B., 1988. Nd and Sr isotopic systematics of

river water suspended material: implications for crustal evolution.

Earth and Planetary Science Letters 87, 249–265.

Goldstein, S.L., O’Nions, R.K., Hamilton, P.J., 1984. A Sm–Nd isotopic

study of atmospheric dusts and particulates from major river systems.

Earth and Planetary Science Letters 70, 221–236.

Hakala, K.J., Adam, D.P., 2004. Late Pleistocene vegetation and climate

in the southern Cascade Range and the Modoc Plateau region. Journal

of Paleolimnology (31).

Heath, G.R., Pisias, N.G., 1979. A method for the quantitative estimation

of clay minerals in North Pacific deep-sea sediments. Clays and Clay

Minerals 27 (3), 175–184.

Hickey, B.M., 1979. The California current system—hypotheses and facts.

Progress in Oceanography 8, 191–279.

Hickey, B.M., Banas, N.S., 2003. Oceanography of the US Pacific

Northwest Coastal Ocean and Estuaries with application to coastal

ecology. Estuaries 26 (4B), 1010–1031.

Hickey, B.M., Royer, T., 2001. California and Alaskan Currents. In:

Steele, J.H., Thorpe, S.A., Turekian, K.A. (Eds.), Encyclopedia of

Ocean Sciences. Academic Press, San Diego, CA, pp. 368–379.

Irwin, W.P., Wooden, J.L., 1999. Plutons and Accretionary Episodes of

the Klamath Mountains, California and Oregon.

Karlin, R., 1980. Sediment sources and clay mineral distributions

off the Oregon Coast. Journal of Sedimentary Petrology 50 (2),

543–560.

Koppers, A.P., 2002. ArArCalc-software for 40Ar/39Ar age calculations.

Computers and Geosciences 28, 605–619.

Kulm, L.D., Prince, R.A., Snavely Jr., P.D., 1973. Site Survey of the

northern Oregon continental margin and Astoria Fan. Initial Reports

of the Deep Sea Drilling Project XVIII (Appendix I, Part A),

979–987.

Lamy, F., Hebbeln, D., Rohl, U., Wefer, G., 2001. Holocene rainfall

variability in southern Chile: a marine record of latitudinal shifts for

the Southern Westerlies. Earth and Planetary Science Letters 185,

369–382.

Langmuir, C.H., Vocke, R.D., Hanson, G.N., Hart, S.R., 1978. A general

mixing equation with applications to Icelandic basalts. Earth and

Planetary Science Letters 37, 380–382.

Lopes, C., 2006. Late Quaternary Paleoceanography of the northeast

Pacific and Atlantic Oceans based on diatom transfer functions.

Oregon State University, Ph.D. Thesis, 250pp.

Lyle, M., Mix, A.M., Ravelo, C., Andreasen, D., Heusser, L.,

Olivarez, A., 2000. Millennial-scale CaCO3 and corg events along

the northern and central California margins: stratigraphy and origins.

In: Lyle, M., Koizumi, I., Richter, C., Moore, Jr., T.C. (Eds.),

Proceedings of the Ocean Drilling Program. Scientific Results, College

Station, TX.

McDougall, I., Harrison, T.M., 1999. Geochronology and Thermochro-

nology by the 40Ar/39Ar method. Oxford University Press, New York,

269pp.

McLaughlin, R.J., Sliter, W.V., Fredriksen, N.O., Harbert, W.P.,

McCulloch, D.S., 1994. Plate Motions Recorded in Tectonostrati-

graphic Terranes of the Franciscan Complex and Evolution of the

Mendocino Triple Junction, Northwestern California.

Merritts, D., Vincent, K.R., 1989. Geomorphic response of coastal

streams to low, intermediate, and high rates of uplift, Mendocino triple

junction region, California. Geological Society of America Bulletin 101

(11), 1373–1388.

Mix, A.C., Lund, D.C., Pisias, N., Boden, P., Bornmalm, L., Lyle, M.,

Pike, J., 1999. Rapic climate oscillations in the northeast Pacific during

the last deglaciation reflect northern and southern hemisphere sources.

In: Clark, P.U., Webb, R.S., Keigwin, L.D. (Eds.), Mechanisms of

Global Climate Change at Millennial Time Scales. American

Geophysical Union, Washington, DC.



Monger, J.W., Price, R.A., Templeman-Kluit, D.J., 1982. Tectonic

accretion and the origin of the two major metamorphic and plutonic

belts in the Canadian Cordillera. Geology 10, 70–75.

Moore, D.M., Reynolds, R.C., 1989. X-ray Diffraction and the

Identification and Analysis of Clay Minerals. Oxford University Press,

New York, 332pp.

Nolan, K.M., Janda, R.J., 1995. Impacts of logging on stream sediment

discharge in the Redwood Creek Basin. US Geological Survey

Professional Paper, vol. 1454.

Personius, S., Kelsey, H.M., Grabau, P.C., 1993. Evidence for

regional stream aggradation in the central Oregon Coast Range

during the Pleistocene–Holocene transition. Quaternary Research 40,

297–308.

Pisias, N.G., Mix, A.C., Heusser, L., 2001. Millennial scale climate

variability of the northeast Pacific Ocean and northwest North

America based on radiolaria and pollen. Quaternary Science Reviews

20, 1561–1576.

Porter, S.C., Pierce, K.L., Hamilton, T.D., 1983. Late Wisconsin

glaciation in the Western United States. In: Porter, S.C. (Ed.), Late-

Quaternary Environments of the United States. University of

Minnesota Press, Minneapolis, pp. 71–111.

Prytulak, J., Vervoort, J.D., Plank, T., Yu, C., 2006. Astoria

Fan sediments, DSDP site 174, Cascadia Basin: Hf–Nd–Pb constraints

on provenance and outburst flooding. Chemical Geology 233,

276–292.

Renne, P.R., Swisher, C.C., Deino, A.L., Karner, D.B., Owens, T.L.,

DePaolo, D.J., 1998. Intercalibration of standards, absolute

ages and uncertainties in 40Ar/39Ar dating. Chemical Geology 145,

117–152.

Robbins, J.M., Lyle, M., Heath, G.R., 1984. A sequential extraction

procedure for partitioning elements among co-existing phases in

marine sediments. Oregon State University College of Oceanography

Reference Manual, Reference 84-3, 45pp.

Rosenbaum, J.G., Reynolds, R.L., 2004. Record of Late Pleistocene

glaciation and deglaciation in the southern Cascade Range. II. Flux of

glacial flour in a sediment core from Upper Klamath Lake, Oregon.

Journal of Paleolimnology 31, 235–252.

Sharp, R.P., 1960. Pleistocene glaciation in the Trinity Alps of northern

California. American Journal of Science 258, 305–340.

Sherwood, C.R., Jay, D.A., Harvey, R.B., Hamilton, P., Simenstad, C.A.,

1990. Historical changes in the Columbia River estuary. Progress in

Oceanography 25, 299–352.

Sklar, L.S., Dietrich, W.E., 1998. River longitudinal profiles and bedrock

incision models: stream power and the influence of sediment supply.

In: Tinkler, K.J., Wohl, E.E. (Eds.), Rivers Over Rock: Fluvial

Processes in Bedrock Channels. American Geophysical Union,

Washington, pp. 237–259.

Snyder, N.P., Whipple, K.X., Tucker, G.E., Merritts, D.J., 2000.

Landscape response to tectonic forcing: digital elevation model

analysis of stream profiles in the Mendocino triple junction region,

northern California. GSA Bulletin 112 (8), 1250–1263.

Sommerfield, C.K., Wheatcroft, R.A., in press. Late Holocene sediment

accumulation on the northern California shelf: Oceanic, fluvial and

anthropogenic influences. Geological Society of American Bulletin.

Strub, P.T., Allen, J.S., Huyer, A., Smith, R.L., 1987. Large-scale

structure of the spring transition in the coastal ocean off western

North America. Journal of Geophysical Research 92, 1527–1544.

Syvitski, J.P.M., Vorosmarty, C.J., Kettner, A.J., Green, P., 2005. Impact

of humans on the flux of terrestrial sediment to the global coastal

ocean. Science 308, 376–380.

Tanaka, T., Shigeko, Togashib, Hikari, Kamiokab, Hiroshi, Amakawac,

Hiroo, Kagamid, Takuji, Hamamotod, Masaki, Yuharad, Yuji,

Orihashie, Shigekazu, Yonedaf, Hiroshi, Shimizug, Takanori, Kuni-

marug, Kazuya, Takahashih, Takeru, Yanagii, Takanori, Nakanoj,

Hirokazu, Fujimakik, Ryuichi, Shinjol, Yoshihiro, Asaharaa, Tani-

mizua, M., Dragusanua, C., 2000. JNdi-1: a neodymium isotopic

reference in consistency with LaJolla neodymium. Chemical Geology

168 (3–4), 279–281.





VanLaningham, S., Duncan, R.A., Pisias, N.G., 2006. Erosion by rivers

and transport pathways in the ocean: a provenance tool using40Ar–39Ar incremental heating on fine-grained sediment. Journal of

Geophysical Research 111 (F04014).

Vorosmarty, C.J., Meybeck, M., Fekete, B., Sharma, K., Green, P., Syvitski,

J.M., 2003. Anthropogenic sediment retention: major global impact from

registered river impoundments. Global and Planetary Change 39, 169–190.

Walczak, P.S., 2006. Submarine Plateau Volcanism and Cretaceous Ocean

Anoxic Event 1a: Geochemical Evidence from Aptian Sedimentary

Sections. Oregon State University, 172pp.

Walker, G.W., MacLeod, N.S., 1991. Geologic Map of Oregon. US

Geological Survey, scale 1:500,000.

Wasserburg, G.J., Jacobsen, S.B., DePaolo, D.J., McCulloch, M.T., Wen,

T., 1981. Precise determinations of Sm/Nd ratios, Sm and Nd isotopic

abundances in standard solutions. Geochimica et Cosmochimica Acta

45, 2311–2323.

Wendt, I., Carl, C., 1991. The statistical distribution of the mean standard

weighted deviation. Chemical Geology (Isot. Geosci. Sect.) 86, 275–285.



Werner, F., Hickey, B.M., 1983. The role of an alongshore pressure

gradient in Pacific Northwest coastal dynamics. Journal of Physical

Oceanography 13, 395–410.

Wheatcroft, R.A., Sommerfield, C.K., 2005. River sediment and shelf

sediment accumulation rates on the Pacific Northwest margin.

Continental Shelf Research 25, 311–332.

Wolf, S.C., Nelson, H., Hamer, M.R., Dunhill, G., Phillips, R.L., 1999.

The Washington and Oregon Mid-shelf Silt Deposit and its Relation to

the Late Holocene Columbia River Sediment Budget.

Woods, M.C., 1976. Pleistocene Glaciation in the Canyon Creek area,

Trinity Alps, California. California Geology, 109–113.

Worona, M.A., Whitlock, C., 1995. Late Quaternary vegetation and

climate history near Little Lake, central Coast Range, Oregon. GSA

Bulletin 107 (7), 867–876.

Zuffa, G.G., de Rosa, R., Normark, W.R., 1997. Shifting sources

and transport paths for the later Quaternary Escanaba Trough

sediment fill (northeast Pacific). Giornale di Geologia 59 (1/2),

35–53.




tracking ﬂuvial response to climate change in the paciﬁc...

Documents