Weissel, J., Peirce, J., Taylor, E., Alt, J., et al., 1991Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 121

21. GEOCHEMISTRY OF BROKEN RIDGE SEDIMENTS1

Robert M. Owen and Andrew R. B. Zimmerman

ABSTRACT

The present study involves the analysis and interpretation of geochemical data from a suite of sediment samplesrecovered at ODP Hole 752A. The samples encompass the time period that includes the lithospheric extension and upliftof Broken Ridge, and they record deposition below and above the mid-Eocene angular unconformity that denotes this uplift.A Q-mode factor analysis of the geochemical data indicates that the sediments in this section are composed of a mixtureof three geochemical end members that collectively account for 94.2% of the total variance in the data. An examination ofinterelement ratios for each of these end members suggests that they represent the following sedimentary components: (1)a biogenic component, (2) a volcanogenic component, and (3) a hydrothermal component. The flux of the biogeniccomponent decreases almost thirtyfold across the Eocene unconformity. This drastic reduction in the deposition of biogenicmaterials corresponds to the almost complete disappearance of chert layers, diatoms, and siliceous microfossils and iscoincident with the uplift of Broken Ridge. The volcanogenic component is similar in composition to Santonian ashrecovered at Hole 755A on Broken Ridge and is the apparent source of the Fe-stained sediment that immediately overliesthe angular unconformity. This finding suggests that significant amounts of Santonian ash were subaerially exposed,weathered, and redeposited and is consistent with data that suggest that the vertical uplift of Broken Ridge was both rapidand extensive. The greatest flux of hydrothermal materials is recorded in the sediments immediately below the angularunconformity. This implies that the uplift of Broken Ridge was preceded by a significant amount of rifting, during whichfaulting and fracturing of the lithosphere led to enhanced hydrothermal circulation. This time sequence of events isconsistent with (but not necessarily diagnostic of) the passive model of lithospheric extension and uplift.

INTRODUCTION

Marine sediments typically are mixtures of sedimentary mate-rial derived from different generic sources (e.g., biogenic, hy-drogenous, hydrothermal, volcanogenic). During the past decade,many workers have been involved in the development and refine-ment of mathematical techniques that are capable of deconvolut-ing the bulk composition of a sedimentary mixture into itsconstituent geochemical end members and of estimating both thecomposition and relative amount of each end member present(Dymond, 1981; Dymond et al., 1984; Leinen and Pisias, 1984;Chen and Owen, 1989; Olivarez and Owen, 1989). The applica-tion of this approach to a downcore sequence of sediment samplesresults in a detailed "chemical stratigraphy," which reflects boththe nature and relative intensity of the various processes that havecontributed to the development of the sediment column.

The present study involves the application of this approach toa column of Eocene sediments at Broken Ridge and is aimed atdeveloping a record of sediment compositional change duringpre-, syn-, and post-uplift deposition. Weissel and Karner (1989)have summarized various lines of evidence (see the followingdiscussion) that suggest that the uplift of Broken Ridge resultedfrom a "passive" rifting mechanism; that is, it was the result ofmechanical unloading of the lithosphere followed by isostaticrebound, as opposed to an "active" uplift in response to thermalor dynamic processes and magmatic thickening of the crust.Sengör and Burke (1978) have suggested that active rifting fol-lows a volcanism-uplift-rifting time sequence of tectonic events,whereas the passive rifting time sequence involves rifting-(possi-bly volcanism)-uplift. A detailed chemical stratigraphy of BrokenRidge sediments should enable us to determine if compositional

1 Weissel, J., Peirce, J., Taylor, E., Alt, J., et al., 1991. Proc. ODP, Sci. Results,121: College Station, TX (Ocean Drilling Program).

2 Department of Geological Sciences, The University of Michigan, Ann Arbor,MI 48109-1063, U.S.A.

variations within the sediment column are consistent with theproposed time sequence for a passive rifting mechanism. Weanticipate that tectonic events involving volcanism and riftingshould be manifested in the sediment column at Broken Ridge assignificant increases (relative to background levels) in the flux ofvolcanogenic and hydrothermal materials, respectively. More-over, the uplift of Broken Ridge is clearly denoted by the presenceof a major angular unconformity that was penetrated by drillingat Ocean Drilling Program (ODP) Hole 752A. Thus, the sequenceof tectonic events at Broken Ridge should be recorded in thecomposition of the geochemical end members that are present andin their position relative to each other and relative to the angularunconformity.

The upper Eocene sediments immediately overlying the angu-lar unconformity are dark brown in color as a result of staining byiron oxide. A specific objective of this study is to employ thechemical stratigraphy methodology to determine the cause of thisincreased iron flux. One possible explanation is that it resultedfrom the weathering of volcanic ash-rich Cretaceous sedimentsthat were exposed following the uplift of Broken Ridge. A resultin support of this explanation would be consistent with estimatesthat Broken Ridge experienced more than 2 km of vertical uplift(Weissel and Karner, 1989; Rea et al., 1990). Alternatively, in-creased iron fluxes have been linked to episodes of suddenlyenhanced hydrothermal activity at times of ridge jumps and riftinitiation (Owen and Rea, 1985; Lyle et al., 1987). If this can bedemonstrated at Broken Ridge, it would imply a significant degreeof temporal overlap between the rifting and uplift stages during apassive rifting event.

LITHOLOGIC SETTINGSeafloor basalt, formed at anomaly 18 time (42 Ma; Houtz et

al., 1977; Mutter and Cande, 1983), and a pronounced angularunconformity across Broken Ridge denote the separation of Bro-ken Ridge from the Kerguelen Plateau and the uplift of BrokenRidge, respectively. Hole 752A (located at approximately

437

R. M. OWEN, A.R.B. ZIMMERMAN

30.89°S, 93.58°E) was drilled north of the crest of Broken Ridgeand contains an extensive record of the sediment deposited beforeand after the mid-Eocene uplift of the ridge. Map views of Hole752A and a cross section of the seismic stratigraphy at BrokenRidge are shown in Peirce, Weissel, et al. (1989) and are repeatedhere as Figures 1 and 2, respectively, for the convenience of thereader. The uppermost portion of the sedimentary section, the"pelagic cap," consists of a nannofossil ooze with foraminifers.A disconformity at about 94 m below seafloor (mbsf) separatesooze of Pleistocene to late Oligocene age (oldest is 30 Ma) fromthe darker brown latest Eocene ooze (youngest is 38 Ma). TheOligocene hiatus on Broken Ridge has been interpreted as result-ing from a relative fall in sea level at this time (Rea et al., 1990),as opposed to an intensification of ocean circulation (e.g., seediscussion of Oligocene hiatuses in Kennett, 1982), because ofthe presence of a pebble layer at the disconformity. An angularunconformity (104-113 mbsf at Hole 752A) marks the period inwhich Broken Ridge was uplifted. The timing of the uplift ofBroken Ridge is well constrained: the oldest sediment overlyingthe angular unconformity is about 38 Ma and the youngest sedi-ment recovered from below it is about 45 Ma (Peirce, Weissel, etal., 1989). The middle Eocene to upper Maestrichtian nannofossilcalcareous (micritic) chalks and limestone underlying this uncon-formity dip to the north. The uppermost chalk unit (113-210 mbsfat Hole 752A) is greenish gray in color and contains volcanic ashlayers and minor occurrences of chert. Both of the unconformitiesare overlain by sand and limestone and chert pebble layers that

contain Fe-stained quartz grains and shell hash reworked from theunderlying sediment. Much of the shallow marine deposits over-lying the unconformable upper Eocene erosion surface, however,were unrecovered as a result of their sandy texture. At Site 755,farther south on Broken Ridge (Figs. 1 and 2), the Eocene uncon-formity is underlain by volcanic tuff and Fe-stained ashy lime-stone of Santonian age. Shipboard analyses of sediment in Cores121-755A-5R to 121-755A-12R report ash concentrations ofgreater than 50% in these sediments.

METHODSGeochemical analyses were performed on a suite of 46 samples

from Hole 752A that ranged in age from 19 to 57 Ma (79.7 to162.7 mbsf). The samples were taken from Sections 121-752A-9H-3 through 121-752A-18X-1 at intervals of, at most, 1.5 m;more closely spaced samples were analyzed from within thesediment overlying the Eocene unconformity (Tables 1 and 2).For comparison purposes, identical analyses were also performedon a suite of 10 samples of Santonian ash-rich sediments recov-ered from Cores 121-755A-5R to 121-755A-9R. All samples,consisting of 1- to 2-cm sections of core, were weighed, freeze-dried, and reweighed to calculate the dry-bulk density. The sam-ples were then ground with mortar and pestle and homogenized.Carbonate content was determined using a conventional pressurebomb technique. All elemental abundances were determined byinstrumental neutron activation analysis (INAA) at the PhoenixMemorial Laboratory, The University of Michigan, using stand-

29°S

92 96°E

Figure 1. Bathymetric map showing location of ODP Site 752, DSDP Site 255, and other Leg 121 sites at Broken Ridge.Contours are shown at 100-m intervals, except along the south-facing escarpment where contours are omitted for clarity(from Peirce, Weissel, et al., 1989).

438

N

753 752 754

GEOCHEMISTRY OF BROKEN RIDGE SEDIMENTS

755

Figure 2. Seismic stratigraphy showing Leg 121 sites at Broken Ridge (from Peirce, Weissel, et al., 1989).

ard INAA procedures (Gordon et al., 1968; Dams and Robbins,1970). Replicate analyses of NBS-1633A standard as well as"Standard Pacific Sediment" (Owen and Ruhlin, 1986) were per-formed to evaluate analytical precision (see Owen and Zimmer-man, this volume).

The statistical methods employed in this study are designed todiscriminate between different types of sediment in the samplesand to determine the relative contribution of each type to anygiven sediment sample. Both the number and the chemical com-position of the major geochemical end members in our samplesare determined by a quantitative sediment partitioning procedure,which involves a mathematical manipulation of the geochemicaldata set using multivariant Q-mode factor analysis and vectorrotation according to the technique of Leinen and Pisias (1984).The chemical significance (i.e., the nature of the source material)of each end member identified by this procedure is determined bycomparing key interelement ratios within each end member withanalogous ratios reported in the literature and known to be char-acteristic of pure geochemical source materials (Dymond, 1981;Graybeal and Heath, 1984; Leinen and Pisias, 1984). The relativeamount of each end member in each sample is determined by alinear programming procedure that provides an optimum solution(minimum error) for each sample. The computer programs neededto implement this technique were kindly provided to us by Drs.M. Leinen and N. Pisias. Finally, the total sediment mass-accu-mulation rate (MAR) is calculated as the product of the dry-bulkdensity and the linear sedimentation rate. The MAR, or flux ofeach end-member component to the sediment at any given time,is the product of the whole-sediment MAR and the fraction of asample's composition that each end member represents. TheMARs of each sediment end member identified by the factoranalysis were calculated using the linear sedimentation rates andage-depth correlations reported in Table 3, the dry-bulk densities

measured in this study, and the relative abundance of each endmember in each sample as determined by linear programming.

RESULTS AND DISCUSSION

Identification of Geochemical End MembersThe concentrations of 15 elements and the whole-sediment

MARs determined for the samples analyzed in this study areshown in Tables 1 and 2. The factor analysis identified threefactors that collectively account for 94.2% of the data; the esti-mated composition of each of these factors, or geochemical endmembers, is shown in Table 4.

The first end member, factor 1, is interpreted as a biogenic (i.e.,biogenic opal and refractory organic matter) end member. Severalinterelemental ratios of this end member, such as Mn/Ba, Al/Ba,Co/Ba, and Mn/Al, are in close agreement with analogous ratiosreported in the literature and considered to be characteristic ofsedimentary biogenic material (Table 5). All of the Ba and thehighest loading of Mn are associated with factor 1 (Table 4); Mnis enriched in phytoplankton (Riley and Roth, 1971; Martin andKnauer, 1973) and Ba is strongly associated with biogenic sedi-ments (Gurvich et al., 1978; Dehairs et al., 1980) and is generallyregarded as a proxy indicator of biological productivity (Boströmet al., 1973; Schmitz, 1987). The MAR of factor 1 (Fig. 3A)reveals a sharp decrease in the rate of biogenic sedimentationacross the middle Eocene unconformity. This is consistent withshipboard observations that note an almost complete absence ofchert layers, diatoms, and other siliceous microfossils in sedimentyounger than early Eocene in age (Peirce, Weissel, et al., 1989).This drastic change in biogenic sedimentation, from accumulationrates of 1.1 to 0.04 g/cm2/1000 yr (Fig. 3A), may be partlyexplained by the movement of the ridge away from productivehigh-latitude regions or by a reduction in bathymetrically induced

439

R. M. OWEN, A.R.B. ZIMMERMAN

Table 1. Major and trace element composition and whole-sediment mass-accumulation rates of BrokenRidge sediments.

Core, section,interval (cm)

121-752A-

9H-3, 125-1309H-4, 110-1159H-5, 125-13010H-1, 130-13510H-2, 130-13510H-3, 130-13510H-4, 130-13510H-5, 130-13510H-6, 130-13510H-7, 20-2510H-7, 55-6011H-1, 60-6511H-1, 130-13511H-2, 60-6511H-2, 130-13511H-3, 60-6511H-3, 130-13511H-4, 60-6511H-4, 120-12511H-5, 60-6511H-5, 130-13511H-6, 60-6512X-1, 10-1211H-6, 125-13012X-1, 20-2213X-1, 135-14013X-2, 130-13513X-3, 100-10513X-4, 52-5714X-1, 130-13514X-2, 130-13514X-3, 130-13514X-4, 130-13515X-1, 130-13515X-2, 130-13515X-3, 130-13515X-4, 130-13515X-5, 110-115

16X-1, 130-13516X-2, 130-13516X-3, 105-11016X-4, 130-135

17X-1, 130-13517X-2, 130-13517X-3, 125-13018X-1, 130-135

Depth(mbsf)

79.7081.0082.7086.3087.8089.3090.8092.3093.8094.2094.5595.3096.0096.8097.5098.3099.0099.80

100.40101.30102.00102.80103.40103.45103.50114.25115.70116.90117.92123.90125.40126.90128.40133.60135.10136.60138.10139.39143.30144.80146.05147.80153.00154.50155.95162.70

Al

4.553.263.567.214.623.851.922.982.613.394.704.293.943.102.582.001.821.261.401.361.741.661.531.243.795.472.275.643.383.053.333.642.192.473.265.481.283.652.181.282.222.592.573.614.143.84

Mn

0.2120.4100.2750.5580.6920.7140.3050.3060.1300.1130.2320.1050.1610.1450.108

0.07120.04160.02590.02430.02200.03080.02800.05550.0189

0.1490.7420.8360.6180.3150.1140.1610.1810.1660.1800.215

0.0650<0.0031

0.1030.2110.1140.1700.1760.1210.1760.1040.195

Ba

<o.oβo<0.142

0.1370.2280.3560.3200.1830.192

0.0533<0.040<0.0300.03250.03720.03030.0103<0.017<0.015<0.014<0.0180.02370.0333<0.0270.0317<0.019<0.046

6.164.84

0.4563.00

0.7051.051.111.02

0.9001.09

<0.0420.5180.740

1.391.121.351.22

0.9011.09

0.6510.932

Fe

1.842.122.454.253.924.012.865.1913.016.116.313.011.310.89.386.643.702.942.491.812.091.952.181.705.833.652.574.192.703.443.852.241.912.663.512.111.463.391.641.631.991.622.783.974.583.76

Co(ppm)

35.342.043.8

101125152

66.7102

13.417.150.123.520.221.118.213.08.406.726.464.745.124.6766.13.7139.320.013.625.319.319.223.212.910.716.718.611.97.1425.112.19.9810.39.4315.525.429.427.0

Se(ppm)

6.8113.611.519.922.020.011.513.816.619.124.120.618.116.013.811.27.496.727.255.457.426.637.465.2715.331.425.433.617.716.417.811.410.811.517.55.076.8012.712.18.6010.910.814.122.620.720.0

Zn(ppm)

<79.5300291437459341169153407396467328260228206160

86.459.150.467.971.740.082.848.5

191211232502296106184110

43.898.1183109

52.392.6174114101120138

87.3214117

Whole-sediment

Mass-accumulation rat

(g/cm2/1000yr)

0.3340.3400.3300.2860.2810.2740.2740.0420.0800.1250.1700.2370.3240.4200.4920.6390.7320.8910.8451.041.111.231.371.671.482.712.922.872.652.672.712.792.953.343.292.742.812.873.002.933.243.532.973.362.913.15

upwelling coincident with subsidence of the ridge (Rea et al.,1990).

Factor 2 is extremely Fe rich and represents about 80% of thecomposition of the sediment immediately overlying the middleEocene unconformity (Fig. 3B). The composition of this endmember estimated from factor analysis (Table 4) is consideredrepresentative of the Fe-stained Eocene sediments described pre-viously. A clear interpretation of the source of the Fe in thesesediments is especially important to understanding the riftingmechanism at Broken Ridge. For example, if the source of this Feis hydrothermal, then the presence of abundant hydrothermalmaterials in the post-uplift sediments would suggest that there isa considerable degree of temporal overlap between the rifting anduplift stages of a passive rifting event.

The Mn/Fe ratio of factor 2 (0.00521) is more than 2 orders ofmagnitude lower than that characteristic of common Fe-rich sedi-mentary components, such as hydrogenous material (Mn/Fe =

1.0-1.5; Dymond et al., 1984; Palmer and Elderfield, 1986),hydrothermal material (Mn/Fe = 0.29-0.43; Dymond, 1981;Greybeal and Heath, 1984; Ruhlin and Owen, 1986a, 1986b), orferromanganous oxyhydroxide products of oxic or suboxic diage-nesis (Mn/Fe = 5-70; Dymond et al., 1984; Chen and Owen,1989). The only other likely source of iron at Broken Ridge isderivation from subaerial or shallow submarine weathering ofunderlying Cretaceous sediments that were uplifted. Volcanic ashis ubiquitous throughout the pre-rift section, although in dimin-ishing abundance upsection. The youngest sediments that are bothrich in Fe and underlie the angular unconformity are ashy lime-stones of Santonian age that were recovered at Site 755. Toexplore this possibility, we have analyzed 10 samples of Santo-nian ash recovered from Cores 121-755A-5R through 121-755A-9R, calculated various interelement ratios from the mean valuesof these analyses, and compared these to analogous ratios calcu-lated from factor 2. The results of this comparison (Table 6)

440

GEOCHEMISTRY OF BROKEN RIDGE SEDIMENTS

Table 2. Rare earth element composition of Broken Ridge sediments.

Core, section,interval (cm)

121-752A-

9H-3, 125-1309H-4, 110-1159H-5, 125-13010H-1, 130-13510H-2, 130-13510H-3, 130-13510H-4, 130-13510H-5, 130-13510H-6, 130-13510H-7, 20-2510H-7, 55-6011H-1, 60-6511H-1, 130-13511H-2, 60-6511H-2, 130-13511H-3, 60-6511H-3, 130-13511H-4, 60-6511H-4, 120-12511H-5, 60-6511H-5, 130-13511H-6, 60-6512X-1, 10-1211H-6, 125-13012X-1. 20-2213X-1, 135-14013X-2, 130-13513X-3, 100-10513X-4, 52-5714X-1, 130-13514X-2, 130-13514X-3, 130-13514X-4, 130-13515X-1, 130-13515X-2, 130-13515X-3, 130-13515X-4, 130-13515X-5, 110-115

16X-1, 130-13516X-2, 130-13516X-3, 105-11016X-4, 130-135

17X-1, 130-13517X-2, 130-13517X-3, 125-13018X-1, 130-135

Depth(mbsf)

79.7081.0082.7086.3087.8089.3090.8092.3093.8094.2094.5595.3096.0096.8097.5098.3099.0099.80100.40101.30102.00102.80103.40103.45103.50114.25115.70116.90117.92123.90125.40126.90128.40133.60135.10136.60138.10139.39143.30144.80146.05147.80153.00154.50155.95162.70

La(ppm)

107166161220278247138151240244

99.971.567.574.767.245.825.722.621.123.228.228.048.124.670.6420388313189

40.351.059.052.549.159.461.725.441.460.540.455.253.958.261.534.363.2

Ce(ppm)

10511917225219622162.399.369.768.789.474.465.161.350.120.822.220.925.67.7412.512.922.125.335.5167100

90.259.447.441.741.029.431.437.713013.331.917.121.026.430.737.936.930.649.6

Nd(ppm)

151172241654518697202190195212100

77.666.785.250.142.523.832.858.217.730.211.729.639.281.830457225326241.3

<48.671.0

<53.5<60.5<62.852.242.556.3115

<47.8118

83.475.664.870.9131

Sm(ppm)

16.126.423.635.244.640.322.423.333.736.918.012.511.212.010.57.394.163.693.193.414.104.125.553.6710.058.950.642.127.06.488.328.877.966.929.5515.63.936.958.666.318.277.718.909.916.219.92

Eu(ppm)

3.645.904.435.929.907.783.457.797.437.843.992.572.492.342.321.48

0.9070.8230.7810.8070.9580.730

1.120.722

1.808.697.988.694.611.852.182.212.241.862.691.941.081.512.292.052.232.272.392.992.152.30

Dy(ppm)

12.235.823.642.971.357.728.627.741.034.816.812.413.312.612.26.505.354.692.643.633.143.184.573.8112.751.351.625.916.34.948.819.768.388.5312.418.33.906.7214.0

<6.3114.25.9312.411.19.4514.7

Yb(ppm)

10.719.014.821.431.128.913.615.515.615.28.237.446.596.495.784.142.482.382.172.052.601.963.361.985.5634.934.826.315.84.787.515.926.606.467.7412.23.216.046.215.685.695.596.957.935.506.87

Lu(ppm)

1.272.562.162.924.273.741.831.602.421.951.18

0.8970.8570.8450.7210.4960.3150.3070.3360.3650.2570.2720.3620.2460.5204.763.653.552.05

0.6160.8050.9410.9390.799

1.021.50

0.3000.627

1.290.4600.8720.7890.947

1.070.663

1.15

Table 3. Age-depth correlation for Hole 752A.

Agea

(Ma)Depth(mbsf)

Linear sedimentation rate

(cm/1000 yr)

17.1-23.723.7-28.1

73.1-91.791.7-93.4

0.280.04

Disconformity

37.8 104.3

Unconformity

55.3-57. 115.6-171.3 2.23

a From Peirce, Weissel, et al., 1989

441

R. M. OWEN, A.R.B. ZIMMERMAN

Table 4. Geochemical end-member composi-tion of Broken Ridge sediments determinedby factor analysis.

Element

Al (%)Fe (%)Mn (%)Ba (%)

Zn (ppm)Se (ppm)Co (ppm)

La (ppm)Ce (ppm)Nd (ppm)Sm (ppm)Eu (ppm)Dy (ppm)Yb (ppm)Lu (ppm)

% of totalVariance

Factor 1

15.90.001.3514.5

53194.119.1

36895

51151.415.154.048.36.35

35.4

% cummulative variance

Factor 2

17.841.8

0.2180.00

94884.568.1

227177156

37.28.7836.220.12.53

36.5

= 94.2

Factor 3

5.21.46

0.6700.00

35212.3

139

339257558

51.79.7461.632.74.32

22.3

indicate a strong similarity between the compositions of factor 2and the Santonian ash samples. This similarity is also evident ina comparison of the rare earth element (REE) patterns of factor 2and the Santonian ash samples (Fig. 4). A minor difference be-tween these REE patterns is that, relative to Santonian ash, factor2 exhibits a greater degree of La enrichment and Ce depletion.These differences, however, are consistent with the type of REEpattern expected for volcanogenic sediment that has been alteredby exposure to seawater and/or mixed with Ce-poor biogeniccalcite or chert (Fleet et al., 1976; Shimizu and Misuda, 1977;Ludden and Thompson, 1978; Desprairies and Bonnot-Courtois,1980).

The interpretation that factor 2 represents volcanogenic mate-rial is also supported by the fact that there is a close correspon-dence between the largest spikes in the MAR record (Fig. 3B) offactor 2 (e.g., at 124-126 and 138 mbsf) and the location of ashlayers in the sediment section below the Eocene unconformitywhich were visually identified during the shipboard examinationof sediments recovered at Site 752 (Peirce, Weissel, et al., 1989).

Factor 3, which accounts for 22.3% of the variability in thedata set, is interpreted as a hydrothermal end member on the basisof the comparison of interelement ratios summarized in Table 7.The interelement ratios shown in Table 7 emphasize Fe, Mn, andthe REEs, whereas we have deliberately avoided using ratios

Table 5. Comparison of elemental ratios of factor 1 (Table 4) vs. literature values forbiogenic sediments.

Ratio Factor 1 Biogenic sedimenta

Mn/BaAl/BaCo/BaMn/Al

0.09311.100.0001320.0849

0.0633-0.09781.00-2.260.0007100.0433

Values from Leinen and Pisias (1984), Graybeal and Heath

(1984), and references therein.

B(g/cm2/1000yr)

0.0 0.5 1.0 1.5 2.0 2.5 3.0

(g/cm2/1000yr)0.0 0.5 1.0 1.5 2.0 2.5 3.0

(mb

sf)

De

pth

75 -

85 •

95

105-

115-

125-

135-

1 4 5 ;

155-

IDisconformity

L^ - > v ^ ^ Unconformity

~•—‰

"^―pp

(mb

sf)

De

pth

75 -

8 5 -

9 5 -

105-

115 -j

125 •;

1 3 5 -

1 4 5 ;

155 -

I

Disconformity

~f3 Unconformity

(g/cm2/1000yr)

0.0 0.5 1.0 1.5 2.0 2.5 3.0

(mb

sf)

De

pth

75

85 -

95 1

105-

115-

125-

1 3 5 -

1 4 5 ;

155-^

J Disconformity

^^^~° Unconformity

V > - _ .

Figure 3. Mass-accumulation rates of geochemical end members in Broken Ridge sediments. MARs for each end member are calculated as the product of thewhole-sediment MAR (Table 1) times the relative amount of the end member in each sample. A. Factor 1 (biogenic component). B. Factor 2 (volcanogeniccomponent). C. Factor 3 (hydrothermal component).

442

GEOCHEMISTRY OF BROKEN RIDGE SEDIMENTS

Table 6. Comparison of elemental ratios of factor 2 (Table 4) vs. Santonian ash values.

Ratio Factor 2 Santonian asha

Al/FeMn/FeZn/FeCo/FeCo/MnZn/MnMn/AlZn/AlCo/Al

0.4260.005210.002270.0001630.03120.4350.0120.005300.000382

0.6190.01430.001670.0003810.02650.1170.02390.002700.000616

Mean values of 10 analyses of Santonian ash from Cores121-755A-5R through 121-755A-9R.

10

CΛ

IΦ

<

I0.1

Santonian ash

57La

59Ce Nd

67 69Yb

71Lu

61 63 65Sm Eu Dy

Atomic numberFigure 4. Comparison of REE patterns of factor 2 and Santonian ash. REEpatterns are normalized to North American Shale Composite (Haskin et al.,1968). Santonian ash is the average of 10 samples from Cores 121-755A-5Rthrough 121-755A-9R.

involving other elements (e.g., Al, Ba, Zn, Co, and Se) that weremeasured. Several authors (Dymond, 1981; Greybeal and Heath,1984; Leinen and Pisias, 1984) have summarized characteristicinterelement ratios for different geochemical end members foundin marine sediments. These ratios typically involve those ele-ments that are most commonly measured in sediment analyses(e.g., Fe, Mn, Co, Ni, Zn, Ba, and Al). Except for Mn and Fe,however, there is actually a considerable degree of overlap be-tween the "characteristic" interelement ratios involving Co, Ni,Zn, Ba, and Al for hydrothermal and hydrogenous material. Forthis reason, the data shown in Table 7 also include typical valuesfor hydrogenous sediments.

The Mn/Fe ratio is the most diagnostic geochemical ratio forhydrothermal sediments because (1) it reflects the composition ofthe two major metallic elements in the ferromanganese oxyhy-droxide hydrothermal precipitates that form from very Fe-richhydrothermal fluids, (2) the analytical error associated with themeasurement of Fe and Mn is usually less than for trace elements,and (3) it is significantly lower than corresponding Mn/Fe ratiosobserved for other common ferromanganous oxyhydroxide sedi-mentary phases (e.g., hydrogenous material or products of oxicand suboxic diagenesis). The Mn/Fe ratio of factor 3 (0.459) isvery close to the range of values (0.29-0.43) considered typicalfor hydrothermal sediments. Comparisons involving REE ratiosare shown in Table 7 because recent studies have found thathydrothermal precipitates are highly efficient scavengers of REEsfrom both hydrothermal vent fluids and from the ambient seawaterthat mixes with hydrothermal fluids (Ruhlin and Owen, 1986a,1986b; Owen and Olivarez, 1988; Olivarez and Owen, 1989). Thiswork has also shown that hydrothermal sediments exhibit highlycharacteristic interelement ratios among the REEs, and these ratios

Table 7. Comparison of elemental ratios of factor 3 (Table 4) vs. literature values forhydrothermal and hydrogenous sediments.

Ratio

Mn/FeEu/YbEu/LuCe/LaSm/Nd

Factor 3

0.4590.2992.250.7580.093

Hydrothermala

sediments

0.29-0.430.288-0.3701.80-2.450.12-1.080.13-0.22

Hydrogenous

b

sediments1.0-1.470.804.449.730.27 5

Values from Fleet (1983), Graybeal and Heath (1984), Ruhlinand Owen (1986b) , Owen and Olivarez (1988) , and Olivarez andOwen (1989b).

b Values from Fleet (1983), Dymond et al. (1984), and Palmerand Elderfield (1986).

443

R. M. OWEN, A.R.B. ZIMMERMAN

are significantly different than the analogous REE ratios charac-teristic of hydrogenous material. The comparison of REE ratiosfor factor 3 vs. hydrothermal sediments and hydrogenous sedi-ments (Table 7) illustrates this point and supports the interpreta-tion that factor 3 is representative of hydrothermal material.

Sediment Composition and Rifting MechanismsThe strongest evidence in support of a passive rifting mecha-

nism for the uplift of Broken Ridge is the fact that paleodepthestimates derived from the examination of microfossils of benthicfauna suggest that the seafloor was subsiding prior to uplift(Peirce, Weissel, et al., 1989). In other words, there is no indica-tion of pre-rift arching or shoaling as would be expected in thecase of active rifting. A comparison of the ages of sedimentsbelow (mid-Eocene) and above (late Eocene) the angular uncon-formity at Broken Ridge indicates that rifting was very rapid.Based on sediment loading and basement subsidence considera-tions, Rea et al. (1990) estimated that Broken Ridge was upliftedabout 2500 m within, at most, 2-3 m.y. Moreover, there is rela-tively low present-day heat flow at Broken Ridge (average = 1.04HFU) and no systematic thermal overprinting of magnetic data(Peirce, Weissel, et al., 1989). Collectively, these results areconsidered tentative because of the poor preservation of certainfauna used as paleodepth indicators and a limited number ofreliable geophysical measurements, but they are consistent withthe passive model for the uplift of Broken Ridge.

An intriguing compositional aspect of the sediment column atHole 752A is the presence of Fe-stained sediments immediatelyoverlying the angular unconformity. Given the tentative nature ofthe evidence in support of a passive rifting mechanism, it becameespecially important to determine whether or not the iron wasderived from a hydrothermal source. Within the context of apassive rifting mechanism, the presence of a significant amountof hydrothermal materials in the post-uplift sediments wouldsuggest that rifting and uplift occurred as a continuum at BrokenRidge, rather than as separate and distinct tectonic events. Alter-natively, a hydrothermal origin for the iron could be interpretedas reflecting a time sequence of distinct tectonic events in whichuplift preceded rifting; that is, a sequence thought to be charac-teristic of an active rifting mechanism (Sengör and Burke, 1978).Although our chemical stratigraphy results suggest that hydro-thermal materials are a major component of the sediment columnat Broken Ridge, they also suggest that the Fe-stained sedimentsoverlying the angular unconformity are derived from the weath-ering of ashes and are not related to a hydrothermal source.

There is at least a gross similarity between the sediment com-position record at Broken Ridge and those representative of activerifting situations. The flux of the hydrothermal end member (fac-tor 3) is greatest prior to the uplift of Broken Ridge and attains amaximum value below the Eocene unconformity (Fig. 3C). Thispattern is suggestive of pulsations in hydrothermal activity thathave been recorded elsewhere and have been related to tectonicreorganization events (Owen and Rea, 1985; Lyle et al., 1987).Reconstruction of the tectonic histories of many locations andexamination of the sediment composition of these locations haverevealed a clear link between active rifting events and periods ofincreased deposition of hydrothermal materials (some at ore-grade levels) that preceded and/or accompanied these events(Rona and Richardson, 1978; Rona et al., 1983; Lyle et al., 1987;Olivarez and Owen, 1989). For example, Lyle et al. (1987) exam-ined the Neogene history of hydrothermal deposition in associa-tion with tectonic reorganizations along the East Pacific Rise bymodeling the deposition of hydrothermal Mn in ridge flank sedi-ments as a function of distance from the ridge axis. Their resultsindicate the occurrence of episodes during which the depositionof hydrothermal materials was as much as 20 times higher than

equivalent Holocene deposition. Our results indicate that themaximum flux of hydrothermal materials at Broken Ridge isabout 3 times greater than early Eocene background levels; how-ever, this peak occurs at about 55.4 Ma, or about 10 m.y. prior tothe estimated time of 45 Ma for ridge uplift (Peirce, Weissel, etal., 1989). The youngest sample below the unconformity analyzedin this study corresponds to an age of about 55.3 Ma. Because ofthis 10-m.y. gap in sample coverage, it is not possible to determineif the peak at 55.4 Ma implies that hydrothermal circulationdiminished some 10 m.y. before uplift or if there were other,perhaps more intense, hydrothermal pulsations closer in time toridge uplift. We plan to address this problem in future studies byexamining sediments recovered at Hole 753A which were depos-ited between 47.0 and 48.8 Ma (Peirce, Weissel, et al., 1989)

SUMMARYSediments deposited below and above the angular unconform-

ity that denotes the uplift of Broken Ridge are composed of amixture of three major geochemical end members. An examina-tion of key interelement ratios for each of these end memberssuggests that they represent the following sedimentary compo-nents: (1) a biogenic component, (2) a volcanogenic componentchemically similar to Santonian ash, and (3) a hydrothermalcomponent. The MAR of the biogenic component is consistentwith shipboard observations that note an almost complete absenceof chert layers, diatoms, and other siliceous microfossils in sedi-ments younger than early Eocene in age (Peirce, Weissel, et al.,1989). We estimate that the rifting and uplift of Broken Ridgeresulted in a nearly thirtyfold reduction in the deposition ofnoncarbonate biogenic material. This drastic change in biogenicsedimentation, from accumulation rates of 1.1 to 0.04 g/cm2/1000yr (Fig. 3A), may be partly explained by the movement of theridge away from productive high-latitude regions or by a reduc-tion in bathymetrically induced upwelling coincident with subsi-dence of the ridge (Rea et al., 1990).

The Fe-stained sediment overlying the middle Eocene angularunconformity is primarily composed of sediment derived from thesubaerial or shallow marine weathering, erosion, and redepositionof Santonian ash and/or other chemically similar volcanogenicmaterial. The REE pattern of this Fe-rich material suggests thatthe ash has been altered to some extent via interaction withseawater. The exposure of large amounts of Santonian ash duringthe Eocene uplift event is consistent with estimates of 2500 m ofvertical uplift calculated using passive uplift mechanism assump-tions (Weissel and Karner, 1989; Rea et al., 1990).

The greatest flux of hydrothermal materials is recorded in thesediments below the angular unconformity at Broken Ridge. Boththe existence of a hydrothermal compositional end member inthese sediments and the relative flux pattern of this componentsuggest that the uplift of Broken Ridge was preceded by a signifi-cant amount of rifting during which faulting and fracturing of thelithosphere led to enhanced hydrothermal circulation. This timesequence of events is consistent with (but not necessarily diag-nostic of) the passive model for lithospheric rifting and uplift.

REFERENCES

Boström, K., Joensuu, D., Moore, C , Boström, B., Dalziel, M., andHorowitz, A., 1973. Geochemistry of barium in pelagic sediments.Lithos, 6:159-174.

Chen, J. C , and Owen, R. M., 1989. The hydrothermal component inferromanganese nodules from the southeast Pacific Ocean. Geochim.Cosmochim. Acta, 53:1299-1305.

Dams, R., and Robbins, J. A., 1970. Nondestructive activation analysisof environmental samples. Univ. Mich. Great Lakes Res. Div. Tech.Rept., 48:1-92.

444

GEOCHEMISTRY OF BROKEN RIDGE SEDIMENTS

Dehairs, R., Chesselet, R., and Jedwab, J., 1980. Discrete suspendedparticles of barite and the barium cycle in the open ocean. EarthPlanet. Sci. Lett., 49:528-550.

Desprairies, A., and Bonnot-Courtois, C, 1980. Relation entre la compo-sition des smectites d'alteration sous-marine et leur cortege de terresrares. Earth Planet. Sci. Lett., 48:124-130.

Dymond, J., 1981. The geochemistry of Nazca plate sediments: an evalu-ation of hydrothermal, biogenic, detrital, and hydrogenous sources. InKulm, L. D., Dymond, J., Dasch, D. M., and Hussong, D. M. (Eds.),Nazca Plate: Crustal Formation and AndeanConvergence: Geol. Soc.Am. Mem., 154:133-174.

Dymond, J., Lyle, M., Finney, B., Piper, D. Z., Murphy, K., Conard, R.,and Pisias, N., 1984. Ferromanganese nodules from MANOP Sites H,S, and R—control of mineralogical and chemical composition bymultiple accretionary processes. Geochim. Cosmochim. Acta,48:931-950.

Fleet, A. J., 1983. Hydrothermal and hydrogenous ferromanganese depos-its: do they form a continuum? The rare earth element evidence. InRona, P. A., Boström, K., Laubier, L., and Smith, K. L., Jr. (Eds.),Hydrothermal Processes at Seafloor Spreading Centers: New York(Plenum), 535-555.

Fleet, A. J., Henderson, P., and Kempe, D.R.C., 1976. Rare earth elementand related chemistry of some drilled southern Indian Ocean basaltsand volcanogenic sediments. / . Geophys. Res., 81:4257-4268.

Gordon, C. E., Randle, K., Goles, G. D., Corliss, J. B., Beesen, M. H.,and Oxley, S. S., 1968. Instrumental analysis of standard rocks withhigh resolution X-ray detectors. Geochim. Cosmochim. Acta,32:364-396.

Graybeal, A. L., and Heath, G. R., 1984. Remobilization of transitionmetals in surficial pelagic sediments from the eastern Pacific. Geo-chim. Cosmochim. Acta, 48:965-975.

Gurvich, Y. G., Bogdanov, Y. A., and Lisitzin, A. P., 1978. Behavior ofbarium in recent sedimentation in the Pacific. Geokhimiya, 3:359-374[translation by Scripta Publ.].

Haskin, L. A., Wildeman, T. R., and Haskin, M. A., 1968. An accurateprocedure for the determination of the rare earths by neutron activa-tion./. Radioanal. Chern., 1:337-348.

Houtz, R. E., Hayes, D. E., and Markl, R. G., 1977. Kerguelen Plateaubathymetry, sediment distribution and crustal structure. Mar. Geol.,25:95-130.

Kennett, J., 1982. Marine Geology: Englewood Cliffs, NJ (Prentice Hall).Leinen, M., and Pisias, N., 1984. An objective technique for determining

end-member compositions and for partitioning sediments accordingto their sources. Geochim. Cosmochim. Acta, 48:47-62.

Ludden, J. N., and Thompson, G., 1978. Behavior of rare earth elementsduring submarine weathering of tholeiitic basalt. Nature, 274:147-149.

Lyle, M., Leinen, M., Owen, R. M., and Rea, D. K., 1987. Late Tertiaryhistory of hydrothermal deposition at the East Pacific Rise: correla-tion to volcano-tectonic events. Geophys. Res. Lett., 14:595-598.

Martin, J. H., and Knauer, G. A., 1973. The elemental composition ofplankton. Geochim. Cosmochim. Acta, 37:1639-1653.

Mutter, J. C, and Cande, S. C, 1983. The early opening between BrokenRidge and Kerguelen Plateau. Earth Planet. Sci. Lett., 65:369-376.

Olivarez, A. M., and Owen, R. M., 1989. REE/Fe variations in hydrother-mal sediments: implications for the REE content of seawater. Geo-chim. Cosmochim. Acta, 53:757—762.

Owen, R. M., and Olivarez, A. M., 1988. Geochemistry of rare earthelements in Pacific hydrothermal sediments. Mar. Chern., 25:183—196.

Owen, R. M., and Rea, D. K., 1985. Sea-floor hydrothermal activity linksclimate to tectonics: the Eocene CO2 greenhouse. Science, 227:166-169.

Owen, R. M., and Ruhlin, D. E., 1986. Interlaboratory comparison of Leg92 standard sediment sample analyses. In Leinen, M., Rea, D. K., etal., Init. Repts. DSDP, 92: Washington (U.S. Govt. Printing Office),353-354.

Palmer, M. R., and Elderfield, H., 1986. Rare earth elements and neodym-ium isotopes in ferromanganese oxide coatings of Cenozoic foramin-ifera from the Atlantic Ocean. Geochim. Cosmochim. Acta,50:409-417.

Peirce, J., Weissel, J., et al., 1989. Proc. ODP, Init. Repts., 121: CollegeStation, TX (Ocean Drilling Program).

Rea, D. K., Dehn, J., Driscoll, N., Farrell, J., Janecek, T., Owen, R. M.,Pospichal, J. L., Resiwati, P., and the ODP Leg 121 Scientific Party,1990. Paleoceanography of the eastern Indian Ocean from ODP Leg121 drilling on Broken Ridge. Geol. Soc. Am. Bull., 102:679-690.

Riley, J. P., and Roth, I., 1971. The distribution of trace elements in somespecies of phytoplankton grown in culture. / . Mar. Biol. Assoc. U.K.,51:63-72.

Rona, P. A., Boström, K., Laubier, L., and Smith, K. L., Jr. (Eds.), 1983.Hydrothermal Processes at Seafloor Spreading Centers: New York(Plenum).

Rona, P. A., and Richardson, E. S., 1978. Early Cenozoic global platereorganization. Earth Planet. Sci. Lett., 40:1-11.

Ruhlin, D. E., and Owen, R. M., 1986a. Factors influencing the rare earthelement composition of hydrothermal precipitates, East Pacific Rise.In Leinen, M., Rea, D. K., et al., Init. Repts. DSDP, 92: Washington(U.S. Govt. Printing Office), 383-389.

, 1986b. The rare earth element geochemistry of hydrothermalsediments from the East Pacific Rise: examination of a seawaterscavenging mechanism. Geochim. Cosmochim. Acta, 50:393^Φ00.

Schmitz, B., 1987. Barium, equatorial high productivity, and the north-ward wandering of the Indian continent. Paleoceanography, 2:63-77.

Sengör, A.H.C., and Burke, K., 1978. Relative timing of rifting andvolcanism on Earth and its tectonic implications. Geophys. Res. Lett.,5:419^21.

Shimizu, H., and Masuda, A., 1977. Cerium in chert as an indication ofmarine environment of its formation. Nature, 266:346—348.

Weissel, J. K., and Karner, G. D., 1985. Drilling on Broken Ridge toevaluate theoretical models for rifting. JOIDES Drilling Prop., 135B.

, 1989. Flexural uplift of ridge flanks due to the mechanicalunloading of the lithosphere during extension. / . Geophys. Res.,94:13919-13950.

Date of initial receipt: 1 March 1990Date of acceptance: 9 November 1990Ms 121B-131

445