Major Element, Trace Element, and Isotopic Composition of Emperor Seamount Chain and Ontong Java Plateau Basalts

MAJOR ELEMENT, TRACE ELEMENT, AND ISOTOPIC COMPOSITION OF EMPEROR SEAMOUNT CHAIN AND ONTONG JAVA PLATEAU BASALTS

A Dissertation

Submitted to the Graduate School of the University of Notre Dame

in Partial Fulﬁllment of the Requirements for the Degree of

Doctor of Philosophy

John T. Shafer, B.S.

Clive R. Neal, Director

Graduate Program in Civil Engineering and Geological Sciences Notre Dame, Indiana

December 2006 This document is in the public domain. MAJOR ELEMENT, TRACE ELEMENT, AND ISOTOPIC COMPOSITION

OF EMPEROR SEAMOUNT CHAIN AND ONTONG JAVA PLATEAU BASALTS

Abstract by

John T. Shafer

The Ontong Java Plateau (OJP) and the Hawaiian Ridge-Emperor Seamount

Chain (HR-ESC) are two of the largest volcanic structures on Earth. As the OJP covers an area of more than 2 × 106 km2, the plateau represents the largest known magmatic event in Earth’s history. Equally impressive, the HR-ESC stretches for more than 6,000 km through the north-central Paciﬁc Ocean. Despite the magnitude of these events, the chemistry of signiﬁcant portions of the OJP and the HR-ESC are poorly known to completely unknown. A full understanding of the chemistry of the OJP and the HR-ESC is necessary to make realistic models of their formation. Major element, trace element, platinum group element (PGE), and isotopic ratio analyses have been conducted on samples collected from several locations on both the OJP and the HR-ESC using inductively coupled plasma-optical emission spectroscopy (ICP-OES), ICP-mass spectrometry (ICP-MS), and negative thermal ionization mass spectrometry (NTIMS).

Using a combination of the aforementioned analytical techniques, it was discovered the OJP has more geochemical variation than previously recognized, especially at the margins of the plateau. The discovery of a previously unknown John T. Shafer lava type from the volcaniclastic succession at Site 1184 on the eastern salient of the OJP and more primitive lava types from basaltic crustal xenoliths suggests that further sampling via deep sea drilling is necessary to better understand the complex origins of the worlds largest igneous province. ESC seamounts have been assumed to undergo similar stages of development as their younger Hawaiian counterparts despite the paucity of non-dredge samples. The investigation of a greater than 300 km thick sequence of basement from Nintoku Seamount conﬁrms that a typical ESC volcano develops in a similar manner to modern Hawaiian volcanoes. PGE and Re/Os isotopic work on three ESC seamounts suggests that both recycled oceanic crust and outer core material contribute to the overall PGE budget in the Hawaiian plume but that it requires unusual melting conditions for both sources to manifest themselves. DEDICATION

To my wife and family

ii CONTENTS

FIGURES...... vii

TABLES ...... xi

ACRONYMS ...... xiii

ACKNOWLEDGMENTS ...... xv

CHAPTER 1: INTRODUCTION TO LARGE IGNEOUS PROVINCES ANDTHEPROBLEMATHAND ...... 1 1.1 OverviewofProblem ...... 1 1.1.1 TheOntongJavaPlateau ...... 1 1.1.2 Hawaiian Ridge-Emperor Seamount Chain ...... 4 1.2 SolvingtheProblem ...... 6

CHAPTER 2: THE ONTONG JAVA PLATEAU: GEOLOGICAL SETTING, PETROLOGY, & GEOCHEMISTRY . . . 7 2.1 Overview...... 7 2.2 Geological Setting and Physical Geology ...... 7 2.3 ModelofFormation...... 8 2.4 Petrology&Geochemistry ...... 11 2.5 VolcaniclasticrocksandSite1184 ...... 15

CHAPTER 3: HAWAIIAN RIDGE-EMPEROR SEAMOUNT CHAIN . . 17 3.1 Overview...... 17 3.2 HistoryofMantlePlumes ...... 17 3.3 The Hawaiian Ridge-Emperor Seamount Chain ...... 24 3.3.1 PhysicalGeology ...... 24 3.3.2 Geochemistry of the Hawaiian Islands ...... 27 3.3.3 Geochemistry of the Emperor Seamount Chain (ESC) . . . 34 3.3.4 Movement of the Hawaiian Mantle Plume ...... 38 3.4 Summary ...... 39

iii CHAPTER 4: COMPOSITIONAL VARIABILITY IN LAVAS FROM OCEAN DRILLING PROGRAM SITE 1184, ONTONG JAVA PLATEAU . . . 41 4.1 Samples ...... 41 4.2 Results...... 43 4.2.1 Petrography...... 43 4.2.2 MajorElements...... 45 4.2.3 TraceElements ...... 51 4.2.4 Radiogenicisotoperatios...... 58 4.3 Discussion...... 61 4.3.1 Theinﬂuenceofalteration ...... 61 4.3.2 Petrogeneticinterpretations ...... 69 4.3.3 Sourceregion ...... 73 4.4 Conclusions ...... 74

CHAPTER 5: ONTONG JAVA PLATEAU XENOLITHS FROM THE IS- LANDOFMALAITA ...... 76 5.1 GeologyofMalaitaandtheSolomonIslands ...... 76 5.2 TheMalaitanAln¨oite...... 79 5.3 Previous Work on Aln¨oite Xenoliths ...... 80 5.4 The Xenolith Suite Studied for this Work ...... 82 5.5 AnalyticalMethods...... 82 5.6 Results...... 82 5.6.1 Petrography...... 82 5.6.1.1 BM/SI-7...... 83 5.6.1.2 PHN5802...... 83 5.6.1.3 PHN3939...... 84 5.6.1.4 PHN5765...... 85 5.6.1.5 PHN5792...... 86 5.6.1.6 PHN5793...... 87 5.6.1.7 PHN5843...... 88 5.6.1.8 PHN5853...... 89 5.6.1.9 PHN5854...... 89 5.6.1.10 ST-7...... 90 5.6.2 MajorElementData ...... 92 5.6.3 TraceElementData ...... 96 5.6.4 Sr-andNd-IsotopicData...... 103 5.6.5 ElectronMicroprobeData ...... 105 5.7 Discussion ...... 111 5.7.1 OriginofREEAnomalies ...... 111 5.7.2 Thermobarometry...... 115 5.7.3 Relationship of the Basalt/Gabbro Samples to the OJP . . 120

iv 5.8 Conclusions ...... 125

CHAPTER 6: HAWAIIAN PLUME SOURCE HYBRIDIZATION: EVI- DENCE FROM NINTOKU SEAMOUNT, EMPEROR SEAMOUNT CHAIN ...... 127 6.1 NintokuSeamount ...... 127 6.2 Samples ...... 129 6.3 Results...... 132 6.3.1 MajorElements...... 132 6.3.2 TraceElements ...... 137 6.3.3 SrandNdIsotopes ...... 148 6.4 Discussion ...... 152 6.4.1 Chemical and Isotopic Variations with Depth (Stratigraphic Age)intheSite1205DrillCore ...... 152 6.4.2 Inﬂuence of Alteration of Chemistry ...... 154 6.4.3 Role of Fractional Crystallization and Crystal Accumulation 157 6.4.4 Role of Melting Processes and Source Heterogeneity in Cre- atingVariationWithinNintoku Lavas ...... 162 6.4.5 Model for the Formation of the Nintoku Post-Shield Basalts 163 6.4.6 A Physical Model for Nintoku Post-Shield Magmatism . . 174 6.4.7 Constraints from Sr Isotope Variations with Time on Pet- rogenesis of Hawaiian-Emperor Post-Shield Lavas . . . . . 176 6.5 Conclusions ...... 179

CHAPTER 7: THE PLATINUM GROUP ELEMENT AND RE–OS ISO- TOPIC COMPOSITION OF THE EMPEROR SEAMOUNT CHAIN . 181 7.1 Introduction...... 181 7.1.1 DetroitSeamount, ODP Leg 197,Site 1203...... 182 7.1.2 Nintoku Seamount, ODPLeg 197, Site 1205 ...... 183 7.1.3 KokoSeamount, ODPLeg197,Site1206 ...... 185 7.1.4 PGEsandtheHawaiianPlume ...... 186 7.2 AnalyticalTechniques ...... 189 7.3 Results...... 189 7.3.1 Petrography...... 189 7.3.1.1 DetroitSeamount...... 189 7.3.1.2 NintokuSeamount ...... 190 7.3.1.3 KokoSeamount...... 190 7.3.2 MajorandTraceElements ...... 191 7.3.2.1 Detroit Seamount (∼ 81Ma) ...... 195 7.3.2.2 Nintoku Seamount (∼ 56Ma)...... 197 7.3.2.3 Koko Seamount (∼ 48Ma) ...... 199 7.3.3 Platinum Group Elements and Re–Os Isotopes ...... 200

v 7.4 Discussion ...... 204 7.4.1 Significance of Pd and Os Depletions ...... 205 7.4.2 PGEandRe–OsSystematics ...... 210 7.4.3 Effect of Sulfide Retention in the Source of the Hawaiian Plume ...... 220 7.5 Conclusions ...... 223

CHAPTER8:CONCLUSIONS ...... 226 8.1 TheOntongJavaPlateau ...... 226 8.2 Hawaiian Ridge-Emperor Seamount Chain ...... 228 8.3 Summary ...... 231 8.4 RelevanceandFutureWork ...... 232

CHAPTER 9: ANALYTICAL TECHNIQUES ...... 234 9.1 OverviewofTechniques...... 234 9.2 SamplePreparation...... 234 9.3 LossonIgnition...... 235 9.4 MajorElementAnalysis ...... 237 9.5 TraceElementAnalysis...... 240 9.6 Platinum Group Element Analysis ...... 246 9.7 IsotopeAnalyses ...... 248 9.7.1 Sr-, Nd-, & Pb-Isotope Analysis as part of Chapter 4 . . . 249 9.7.2 Sr- & Nd-Isotope Analysis as part of Chapter 6 ...... 250 9.7.3 OsIsotopeAnalyses ...... 252

BIBLIOGRAPHY ...... 253

vi FIGURES

1.1 Map of the Ontong Java Plateau and surrounding area ...... 2 1.2 Map of the Hawaiian Ridge-Emperor Seamount Chain ...... 5

2.1 Map of the Ontong Java Plateau, highlighting the island of Malaita 12 87 86 2.2 Graph of Sr/ Sr vs. ǫNd for OJP basalts and MORB basalts . . 13 2.3 Primitive mantle normalized proﬁles of basalts recovered from Site 803...... 14 2.4 Map of the Ontong Java Plateau, highlighting the eastern salient andSite1184 ...... 16

3.1 Schematic of a rising mantle plume ...... 20 3.2 Map of the Hawaiian Ridge-Emperor Seamount Chain ...... 25 3.3 Map of the Hawaiian Islands showing the Kea and Loa trends .. 28 3.4 208Pb/204Pb vs 176Hf/177Hf, 87Sr/86Sr, and 143Nd/144Nd isotopic ratios showing the distinct ﬁelds for Kea and Loa trend volcanoes . 29 3.5 MgO vs SiO2 forKilaueatholeiiticbasalts ...... 30 3.6 Incompatible element diagram of Kilauea shield stage tholeiites and EPRMORB...... 31

3.7 Map of the distribution of R/Ra and ǫNd on the big island of Hawaii 33 3.8 Origin of Alkalic Pre- and Post-Shield Basalts ...... 34 3.9 87Sr/86Sr vs. seamount age for the Hawaiian Ridge-Emperor Seamount Chain ...... 36

4.1 Stratigraphic section of core recovered from Hole 1184A ...... 42 4.2 Photomicrographs of Site 1184 basalt clasts ...... 44 4.3 (a) Loss on ignition (LOI) (wt.%) vs. CaO (wt.%) and (b) LOI (wt.%) vs. K2O(wt.%)...... 46

4.4 Classiﬁcation of Site 1184 basalt clasts. (a) SiO2 (wt.%) vs. Na2O 4 + K2O (wt.%) and (b) Nb/Y vs. Zr/(P2O5 × 10 ) ...... 47

vii 4.5 Primitive mantle normalized incompatible element plots of Groups 1,2,3,and4 ...... 57 87 86 206 204 4.6 (a) Sr/ Sr(t) vs. ǫNd(t); (b) Pb/ Pb(t) vs. ǫNd(t) for Site 1184 basaltsamples...... 61 4.7 Sr (unleached, ppm) vs. 87Sr/86Sr...... 63

4.8 (a) (La/Ta)PM vs. (Yb/Ta)PM; (b) (Nb/Ta)PM vs. (Th/Ta)PM; (c) (Zr/Ta)PM vs. (Hf/Ta)PM ...... 65

4.9 CaO (wt.%) vs. (Y/Ho)PM ...... 69 4.10 (a) Group 1 basalt clasts normalized to average Kwaimbaita-type basalt; (b) Group 2 basalt clasts normalized to average Kwaimbaita- type basalt; (c) Group 3 basalt clasts normalized to average Kroenke- type basalt; and (d) Group 4 basalt clasts normalized to average Wairahitobasalt ...... 72 4.11 Average Group 4 clast composition, average Group 4 volcaniclastic matrix composition, average high-Nb group of Fitton & Godard [78], and average Kwaimbaita-type basalt, all normalized to primi- tivemantle ...... 73

5.1 MapofSolomonIslandChain ...... 78 5.2 Map of north-central Malaita showing locations of aln¨oite pipes and sills...... 81 5.3 Photomicrograph of BM/SI-7 under crossed polars ...... 84 5.4 Photomicrograph of PHN 5802 under plane polarized light .... 85 5.5 Photomicrograph of PHN 3939 under crossed polars ...... 86 5.6 Photomicrograph of PHN 5765 under crossed polars ...... 87 5.7 Photomicrograph of PHN 5793 under crossed polars ...... 88 5.8 Photomicrograph of PHN 5853 under crossed polars ...... 90 5.9 Photomicrograph of PHN 5854 under crossed polars ...... 91

5.10 SiO2 and MgO (wt.%) vs LOI (wt.%) for the Malaitan xenolith suite 95

5.11 SiO2 (wt.%) vs Total Alkalis (Na2O+K2O) (wt.%) discrimination diagram ...... 97 5.12 Chondrite normalized REE profiles for the Malaitan Xenoliths and average Kroenke, Kwaimbaita, and Singgalo basalts ...... 98 5.13 Primitive mantle normalized incompatible element profiles for the Malaitan Xenoliths and average Kroenke, Kwaimbaita, and Singgalo basalts...... 103 87 86 5.14 Sr/ Srt vs ǫNd(t) for the Malaitan xenoliths and the host alnöite 106

viii 5.15 Pyroxene ternary classification diagram ...... 111 5.16 Ca-Mg-Fe* compositions of clinopyroxene, orthopyroxene, and gar- netinMalaitanxenoliths...... 112 5.17 Garnet proportions in eclogites from various locations...... 113 5.18 Chondrite normalized REE profiles of samples BM/SI-7, PHN 5802, PHN5843,PHN5853,andST-7...... 115 5.19 Crustal samples normalized to average Kroenke and Kwaimbaita basalt ...... 122 5.20 TiO2 (wt.%) vs MgO (wt.%) for crustal xenoliths ...... 124 6.1 Stratigraphic profile for Hole 1205A Nintoku Seamount ...... 128

6.2 Classiﬁcation of Site 1205 basalts, SiO2 (wt.%) versus total alkalis (Na2O+K2O)(wt.%) ...... 137 6.3 MgOvariationdiagrams ...... 138 6.4 Chondrite-normalized REE proﬁles for Site 1205 basalts ..... 148

6.5 ǫSr vs. ǫNd showing Hawaiian and Emperor post-shield basalts and theNintokupost-shieldbasalts ...... 152 87 86 6.6 MgO (wt.%) vs. Sr/ Sri and ǫNd ...... 153 6.7 DepthproﬁlesfromHole1205A ...... 154 6.8 Initial 87Sr/86Srvs.La/Yb,Nb/Zr ...... 155

6.9 P2O5 (wt.%) vs. K2O/P2O5 ...... 156 6.10Ce/Pbvs.Nb/U ...... 157 6.11 MgO variation diagrams showing only the conglomerate clasts and the alkalic basalts from the uppermost igneous basement . . . . . 159 6.12 Majorelementmodeltestﬁgure ...... 162 87 86 6.13 Sr/ Sri vs.La/Ce ...... 168 6.14 Graphical representation of the results of the three models pre- sentedinFigure6.13 ...... 174

6.15 Distance from Kilauea (km) vs. ǫSr and age of crust at time of seamount formation vs. ǫSr ...... 177 7.1 Bathymetry oftheDetroitSeamount region ...... 183 7.2 Bathymetry of the Nintoku Seamount region ...... 184 7.3 BathymetryoftheKokoSeamountregion ...... 185

7.4 SiO2 (wt.%) vs. total alkalis (Na2O+K2O) (wt.%) discrimination diagram ...... 196

ix 7.5 Incompatible trace element concentrations of Detroit, Nintoku, and Koko seamounts normalized to primitive mantle ...... 197 7.6 Harker diagrams showing major element trends of Detroit, Nintoku, andKokoSeamounts ...... 198 7.7 Th (ppm) vs Nb (ppm) (a) and La/Yb vs Nb (ppm) (b) . . . . . 199 7.8 Primitive mantle-normalized PGE proﬁles of the ESC basalts ana- lyzedinthisstudy ...... 202 7.9 187Os/188OsvsAge(Ma)...... 203 7.10 187Os/188OsvsOs(ppt) ...... 204 7.11 Pd(ppb)vsCr(ppb)...... 206 7.12 LOI (wt.%) vs Pt/Pd from ESC high-MgO and picritic samples . 207 7.13 Primitive mantle normalized plot of the ﬁve samples with Os concentrations...... 208 7.14 Os/IrvsCr(ppm)andCr/LafortheESClavas ...... 210 7.15 Graphical representation of the results of the 187Os/188Os mixing model ...... 218 7.16 Re/Yb and Pt/Yb vs Sm/Nd of ESC samples from this study and HawaiianIslandsamples ...... 225

9.1 Flowchart diagraming reduction of raw ICP-OES data ...... 241 9.2 Flowchart diagraming reduction of raw ICP-MS data ...... 245

x TABLES

4.1 MAJOR ELEMENT CONCENTRATIONS (WT.%), LOSS ON IGNITION, AND CORE INTERVALS FOR ODP SITE 1184 BASALTIC CLASTS ...... 48 4.2 TRACE ELEMENT CONCENTRATIONS (PPM) OF LEG 192, SITE 1184 BASALTIC CLASTS ...... 53 4.3 STRONTIUM, NEODYMIUM, AND LEAD ISOTOPIC DATA OF ODP SITE 1184 BASALTIC CLASTS ...... 59

5.1 MAJOR ELEMENT CONCENTRATIONS AND LOI VALUES (ALL VALUES IN WT.%) OF MALAITAN XENOLITHS . . . . 93 5.2 TRACE ELEMENT ABUNDANCES (ALL VALUES IN PPM) OF MALAITANXENOLITHS ...... 99 5.3 SR- AND ND-ISOTOPIC DATA (ELEMENTAL ABUNDANCES IN PPM) OF MALAITAN XENOLITHS ...... 104 5.4 ELECTRON MICROPROBE DATA ON CLINOPYROXENE, OR- THOPYROXENE, OLIVINE, AND GARNET CRYSTALS FROM MALAITANXENOLITHS ...... 108 5.5 TEMPERATURE ( ◦C) AND PRESSURE (KBAR) OF CLINOPY- ROXENE FROM MALAITA XENOLITHS ...... 116

6.1 SUMMARY OF UNIT NUMBER, RECOVERY DEPTH, AND ROCK TYPE OF SAMPLES RECOVERED FROM SITE 1205A 130 6.2 MAJOR ELEMENT ABUNDANCES FOR LEG 197, SITE 1205 BASALTS ...... 133 6.3 TRACE ELEMENT ABUNDANCES (PPM) OF LEG 197, SITE 1205BASALTS,PART1 ...... 140 6.3 TRACE ELEMENT ABUNDANCES (PPM) OF LEG 197, SITE 1205BASALTS,PART2 ...... 144 6.4 STRONTIUM AND NEODYMIUM ISOTOPE DATA FOR LEG 197,SITE1205BASALTS ...... 149 6.5 MAJOR ELEMENT MODEL PARAMETERS ...... 160 6.6 RESERVOIR COMPOSITIONS (PPM) AND PARTITION CO- EFFICIENTS ...... 166

xi 6.7 TRACE ELEMENT MODEL RESULTS ...... 170

7.1 MAJOR ELEMENT CONCENTRATIONS AND LOI VALUES (ALL VALUES IN WT.%) OF ESC SAMPLES ...... 192 7.2 TRACE ELEMENT ABUNDANCES (ALL VALUES ARE IN PPM) OFESCSAMPLES ...... 193 7.3 PLATINUM GROUP ELEMENT ABUNDANCES (ALL VALUES AREINPPB)OFESCSAMPLES ...... 201 7.4 RE/OS ISOTOPIC DATA (CONCENTRATION VALUES IN PPB) OFESCSAMPLES ...... 203 7.5 MIXING PARAMETERS FOR 187OS/188OSMODEL ...... 216 7.6 PARAMETERS FOR RESIDUAL SULFIDE MODEL ...... 221

9.1 PROTOCOL FOR ICP-OES ANALYSIS ...... 238 9.2 STANDARD AND SPIKE CONCENTRATIONS (PPB) USED IN TRACE ELEMENT ICP-MS ANALYSIS ...... 242 9.3 PROTOCOL FOR ICP-MS ANALYSIS ...... 244

xii ACRONYMS

LIP Large Igneous Province

OJP Ontong Java Plateau

OJPT Ontong Java Plateau Terrain

HR Hawaiian Ridge

ESC Emperor Seamount Chain

MOR Mid-ocean ridge MORB Mid-ocean ridge basalt

EP R East Paciﬁc Rise

DMM Depleted mantle material

ODP Ocean Drilling Program

DSDP Deep Sea Drilling Project

HSDP Hawaiian Scientiﬁc Drilling Project MBSL Meters below sea level

MBSF Meters below sea ﬂoor

ICP − MS Inductively coupled plasma-mass spectrometry

ICP − OES Inductively coupled plasma-optical emission spectroscopy

LA − ICP − MS Laser ablation-inductively coupled plasma-mass spectrometry

USN − ICP − MS Ultrasonic nebulization-inductively coupled plasma-mass spectrometry EMP Electron microprobe

xiii NTIMS Negative thermal ionization mass spectrometry LOI Loss on Ignition

ME Major elements

T E Trace elements

P GE Platinum group elements

IT E Incompatible trace elements

HFSE High ﬁeld strength elements REE Rare earth element

LREE Light rare earth element

HREE Heavy rare earth element

SRM Standard reference material

DCS Drift corrector solution

xiv ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Clive R. Neal, for his help, support, and guidance during my time at Notre Dame. I would also like to thank Dr. Jinesh Jain and Mr. Dennis Birdsell for help with ICP analyses. Finally, I would like to thank Dr. Jim Ely, Dr. Bill Chazey, Mr. Will Kinman, and Mrs. Georgianna Kramer for all of their useful discussions.

xv CHAPTER 1

INTRODUCTION TO LARGE IGNEOUS PROVINCES AND THE PROBLEM AT HAND

1.1 Overview of Problem

1.1.1 The Ontong Java Plateau

The Ontong Java Plateau (OJP) is the most voluminous of the world’s Large

Igneous Provinces (LIPs) and is by far the largest known magmatic event on Earth. Greater than 2 × 106 km2 of Paciﬁc Ocean seaﬂoor is covered by the main plateau (Fig. 1.1). Crustal thickness beneath the plateau may be as great as 30–35 km [90, 211], indicating that the volume of igneous rock comprising the plateau could be as high as 6 × 107 km3 [79]. The OJP formed rapidly at ∼ 122 Ma [34, 158, 241, 242], with a possible second minor plateau-building event at ∼ 90 Ma [158, 187, 241]. Despite the OJP’s great size, the geochemistry of the plateau is remarkably uniform. Tejada et al. [242] named two chemically and isotopically distinct stratigraphic units: the Kwaimbaita Formation and the Singgalo Formation [79] that are also present at Site 803 (Units C-G and Unit A, respectively) on the northern margin of the OJP and from Solomon Islands on the southern margin (Fig. 1.1). A third basalt type was found during Ocean Drilling Program (ODP) Leg 192 at Sites 1185 and

1187. This basalt type has lower MgO and trace element contents than either the

1 150 155 160 165 170 175 15 15 0 00 -6 -5 N 000 -400

0 -50 00 00 -50 -50 -4000 -6 00 00 00 -2000 802 0 -50 -3000 East Mariana 0 00 -50 00 00 00

-3 00 10 -30 -40 10 Basin 00 -3000

-40 -40 00 -50 -30 00 00 -30 -50 -40 00 00 462 00 -40 0 00 00 -4000 -3 -30

-30 5 00 5 807 -40 Nauru 00 803 Basin

1187 -4000 1000 kilometres -30 289 00 0 -4 -20 1185 0 000 00 Kroenke Canyon -2 1186 0 -4 1183 00 000Ellic Basin High

00 0 plateau -3000 -50 -30 00 00 -20 -2000 5 -20 5 00 0 1184 ient -70 -3000 00 sal 00 -1000 Stewart Basin -20 -10 Santa Isabel ern -40 00 00 -3000 0 East

-30 -1 00 00 Malaita 000 -400 0 -4 0 -500 -10 10 00 0 10 -30 -30 San Cristobal 00 00 -60 S -2000 00

150 E 155 160 165 170 175

Figure 1.1. Map of the Ontong Java Plateau and surrounding area

Kwaimbaita and Singgalo Formations and was termed Kroenke-type basalt [79]. Chapter 2 will provide a more in depth discussion of the geology, petrology, and geochemistry of the Ontong Java Plateau. The OJP has been sampled by deep sea drilling in eight locations, seven of which are located on the high plateau (Fig. 1.1). The other location is on the unnamed northern ridge of the eastern salient of the OJP (Fig. 1.1) that had not been drilled prior to ODP Leg 192. Because of this, the relationship of the eastern salient to the high plateau was unknown. Workers speculated that it may be contemporaneous with the high plateau, the product of a postulated plume tail after emplacement of the high plateau, or possibly the locus of the ∼ 90 Ma eruptions [132, 154, 210, 241]. It is impossible to understand the formation of the OJP without knowledge of the tectonic setting, igneous petrology, eruption setting, and stratigraphy of the eastern salient. Therefore, Chapter 4 will focus

2 on the petrology and geochemistry of basalt samples recovered from ODP Leg

192, Site 1184 on the eastern salient of the OJP. These samples suggest that the eastern salient was formed concurrently with the high plateau and that Site 1184 of the OJP was erupted subaerially [33, 243]. All three basalt types found on the high plateau are present at Site 1184, while there is one novel basalt type

[78, 223]. These ﬁndings suggest that there was more evolution of OJP magmas on the eastern salient than on the high plateau. While deep sea drilling has told us much about the composition and structure of the OJP, it’s impact is limited to the top few hundred meters. Field work in the Solomon Islands has, through sampling in eroded anticlines on Malaita, penetrated

∼ 4.5 km into the lava pile. New drilling ships could potentially recover upwards of several kilometers of basement although these expeditions have yet to be ﬁnalized. If the OJP is 33 km thick [90], this means that at least 85% of the plateau has not been sampled. This is a serious problem if we are to understand the origin of the known basalt types. However, due to the collision of the OJP with the old Solomon arc during the early Neogene as a part of southwest-directed subduction, obducted sections of OJP basement are present on the islands of Malaita, Santa Isabel, and San Cristobal (Fig. 1.1) [51, 54, 193]. A small number of deep crustal xenoliths were found within aln¨oite pipes that intrude the exposed OJP basement. These samples appear to be metamorphosed OJP basalts and may be from deep sections of the OJP that have not, and may not ever be, sampled by deep sea drilling. These samples provide a intriguing potential glimpse into the interior of the OJP and will be further described in Chapter 5.

In summary, as Tejada et al. [240] stated, our current sampling of the OJP can be likened to “pin-pricking an elephant”. The multitude of origination models

3 reﬂects the lack of detailed knowledge of OJP composition and age [240]. The new data presented in Chapters 4 and 5 will begin to help ﬁll out hitherto unknown parts of the story.

1.1.2 Hawaiian Ridge-Emperor Seamount Chain

The Hawaiian hotspot has produced the ∼ 6, 000 km long Hawaiian Ridge- Emperor Seamount Chain (HR-ESC) over the course of it’s > 86 Ma history.

Active volcanism occurs on the Hawaiian Islands at the eastern end of the Hawai- ian Ridge and the islands and seamounts become progressively older to the west and north (Fig. 3.2). The Hawaiian Islands are by far the most sampled and studied ocean island chain in the world. They are the basis of mantle plume theory and is the most convincing example of lithospheric plates changing direction

[44, 59, 153, 198, 252]. Chapter 3 discusses the geology, petrology, and geochemistry of the Hawaiian Ridge-Emperor Seamount Chain and discuss the history and literature on the development of mantle plume theory. Despite the great number of samples and studies done on the Hawaiian Islands, the ESC seamounts were virtually unknown until recently. Deep Sea Drilling Project (DSDP) Leg 55 carried out the ﬁrst organized drilling expeditions to the ESC, but only deeply sampled several seamounts. For example, Nintoku Seamount, situated roughly in the middle of the ESC (Fig. 3.2), had only 33 m of drill core recovered. ODP Leg 197 picked up where DSDP Leg 55 ended, and was able to more deeply drill several of the seamounts not sampled by DSDP Leg 55. In a problem similar to that facing the OJP, models developed on the basis of chemistry of basalts from the Hawaiian Islands need to be tested against the chemistry of the older products of the Hawaiian hotspot. Therefore, knowledge of the

4 Figure 1.2. Map of the Hawaiian Ridge-Emperor Seamount Chain and surrounding areas. Modiﬁed from Tarduno et al. [236]. Shaded areas are regions of sea ﬂoor extending above approximately 3500 meters below sea level (mbsl). chemistry of the ESC is imperative to test theories and to develop new ideas about the evolution of the Hawaiian hotspot magma chemistry. In fact, one of the stated goals of ODP Leg 197 was to “obtain geochemical data needed for understanding the compositional variability of volcanic products from the Hawaiian hotspot” [236]. In order to further this goal, Chapter 6 describes the chemistry of samples recovered from Nintoku Seamount during ODP Leg 197 and provides a model for their formation. These samples provide insights to ESC magma formation.

The compositional variability of the Hawaiian hotspot has been a much discussed topic [124, 201], generally limited to trace element and 87Sr/86Sr and 143Nd/144Nd and Pb isotopic variations. Very little is known about the evolution of the platinum group elements (PGEs) throughout the history of the Hawaiian hotspot. PGEs and their related isotope ratios are the centerpiece to a contro-

5 versy about the signiﬁcance of coupled enrichments of 187Os/188Os and 186Os/188Os as well as enriched PGE abundances in Hawaiian picrites and high-MgO basalts [20, 25, 26]. Theories based on PGE chemistry are exclusively limited to the Hawaiian Islands, as PGE data is generally nonexistent from the ESC. Chapter 7 describes data PGE abundance and 187Os/188Os isotopic ratio data collected from

ESC seamounts and the Hawaiian Islands.

1.2 Solving the Problem

The research described in this dissertation requires major- and trace-element data (ME and TE, respectively), PGE data, and isotopic data (187Os/188Os, 87Sr/86Sr, 143Nd/144Nd). The majority of the data collected in this dissertation has been collected by myself and where it was not, it will be pointed out in the text. Nevertheless, all data presented in this dissertation has been interpreted by me and I have had the responsibility of publishing the results. Data collection techniques will be described in detail in Chapter 9.

6 CHAPTER 2

THE ONTONG JAVA PLATEAU: GEOLOGICAL SETTING, PETROLOGY, & GEOCHEMISTRY

2.1 Overview

The purpose of this chapter is to summarize new research into the origin and evolution of the Ontong Java Plateau (OJP) and integrate it with previous studies. It will also serve as the location for background information in Chapters 4 and 5. Basic information (such as size, age, basalt types) was introduced in Chapter 1.

2.2 Geological Setting and Physical Geology

The Ontong Java Plateau (OJP) is the largest of the Earth’s large igneous provinces (LIPs) and covers an area of approximately 2 × 106 km2 (similar to the size of Alaska). With crustal thicknesses beneath the plateau of approximately 30–35 km [90], the total volume of igneous rock forming the plateau may be as great as 6 × 107 km3 [79]. The gross structure of the OJP can be divided into two main parts: the high plateau in the west and the eastern salient (Fig. 1.1). Much of the plateau stands at least 2 km above the surrounding ocean ﬂoor [194] under 2 to 3 km of seawater [172]. The structure is bounded to the north by the East Mariana Basin, to the northwest by the Lyra Basin, and by the Ellice Basin to the southeast. The southwest margin of the OJP impinges upon the the Solomon

7 Arc, which has resulted in the folding and uplifting of the southern margin of the

OJP over the last 6 Ma [78, 133] forming several of the Solomon Islands (Malaita, Santa Isabel, Ramos, and Ulawa).

2.3 Model of Formation

The high magma production rates necessary for the formation of the OJP require an anomalously hot mantle [79]. Because of this, the plume-head hypothesis has been suggested as a mechanism by which enormous volumes of basaltic magma could be erupted over an extremely short period of geologic time [30, 210]

(see Chapter 3, Section 3.2). Trace element and isotopic data are consistent with high degrees of melting (at least 30%) of a mildly depleted peridotite source [35, 78, 107]. Such high degrees of melting are likely to only be achievable by decompression melting under thin lithosphere with very high potential temperatures [79]. Melting conditions of this nature require that the mantle source be actively fed into the melting zone, which can be easily achieved with a mantle plume [79]. PGE abundances are elevated relative to primitive mantle estimates [35]. En- riched PGE abundances are consistent with a mantle plume originating at the core-mantle boundary [35]. However, efforts to locate the hotspot responsible for the OJP has largely been unsuccessful. Several authors have favored the Louisville hotspot as the source of the OJP [157, 210, 238], but new plate reconstruction models based on a fixed hotspot frame of reference place the location of the OJP at 43 ◦S at the time of OJP formation [133] which is 9 ◦ north of the current position of the Louisville hotspot. This indicates that the Louisville hotspot was not responsible for the formation of the OJP or that the hotspot has undergone significant drift since 120 Ma. This is not totally implausible, since new research

8 on the Hawaiian hotspot [235, 237] suggests that the Hawaiian hotspot has drifted southward during the formation of the Emperor Seamount Chain. However, the paleomagnetic paleolatitude of the OJP at the time it formed (∼ 25 ◦S) [212] further increases the discrepancy with the Louisville hotspot. Riisager et al. [212] interprets the discrepancy as evidence that the Louisville hotspot may have moved southwards over the last 120 Ma. Taking into account both hot-spot motion and true polar wander [11], formation of the OJP by the Louisville hotspot is possible but unlikely. In addition, there is no clear chain of ocean islands connecting the Louisville hotspot and the OJP as predicted by the plume head hypothesis. The deficiencies in the plume head hypothesis has led to several other formation mechanisms being proposed. The large size, high melting conditions, and limited variation in magma compositions have lead Ingle & Coffin [114] to suggest a meteorite impact origin for the OJP. Ingle & Coffin [114] cite the lack of a known hotspot or hotspot track that can be convincingly tied to the OJP, the lack of uplift during formation and subsidence since then, the anomalously high melting conditions, and isotopic similarity to some Indian Ocean MORB as evidence of an alternate formation method than by a impinging plume head or hotspot. In order to generate the quantity of magma necessary to form the

OJP, a sizable bolide (> 10 km diameter) would be required. This bolide would vaporize the approximately 4 km water column and probably most or all of the lithosphere [114]. The resulting crater would be ≥ 200 km in diameter and the near instantaneous generation of large quantities of melt would ﬁll the crater [114]. This model accounts for the limited variations in magma chemistry, the lack of uplift associated with massive volcanism, and elevated PGE abundances. However, a bolide of the size necessary to produce the OJP would have been

9 larger than the impactor responsible for the mass extinction at the Cretaceous-

Tertiary boundary. No mass extinctions occurred at the time of OJP formation [79]. In addition, the trace element geochemistry of the OJP is also inconsistent with formation by bolide impact. The enormous amounts of magma generated by the impact of the meteorite into Pacific Ocean lithosphere would result in magmas that are essentially identical to MORB. While the OJP does have some isotopic similarity to Indian Ocean MORB, the overall trace element geochemistry is not MORB-like. The elevated PGE abundances that are cited by Ingle & Coffin as being the result of distribution of elevated PGE abundances in the bolide throughout the generated melt can also be accounted for by origination of a mantle plume from the core-mantle boundary [35]. The sum of these short- comings is at least as severe, if not more so, than those of the mantle plume theory. Another theory is that there is a layer of unusually fertile, volatile-rich, near- solidus mantle material with an OIB-like composition in the upper asthenosphere (termed the “perisphere”) that was present beneath the current position of the OJP at 120 Ma [10]. Sudden rifting caused cataclysmic melting of the perisphere resulting in the formation of the OJP. As the perisphere source is tapped, MORB- like mantle underlying the perisphere should be increasingly incorporated into the melt. Difficulties in this model include the fact that OJP magmas are volatile poor [213] and, so far, no MORB-like magmas have been found. In fact, the most evolved and presumably latest forming OJP magma are even less MORB like than earlier magmas.

In the end, no one formation theory can account for all of the characteristics of the OJP. The plume-head hypothesis appears to ﬁt more of the observations than

10 the other hypotheses [79]. Still, the lack of uplift associated with the formation of the plume and the lack of an apparent hotspot track are issues that remain to be solved.

2.4 Petrology & Geochemistry

As mentioned in Section 2.2, the only subaerial exposures of OJP basalt are on the Solomon Islands, the thickest being on the island of Malaita (Fig. 2.1). The Malaitan Volcanic Group is composed of a succession of early Cretaceous tholeiitic pillow basalt, sheet ﬂows, and sills that are 3–4 km thick [79, 192, 242].

Pelagic chert and limestone is thinly interbedded at rare intervals, suggesting high eruption frequency and emplacement in deep water [79]. Crustal thickness of the OJP averages around 36 km [90] which translates into a volume of at least 5 × 107 km3 and could be as high as 6 × 107 km3 [46, 79, 154]. OJP related volcanism may be even greater, as lavas ﬁlling the adjacent Nauru, Mariana, and Pigafetta

Basins have been proposed to be closely related to the OJP [31, 32, 90]. Age dating of the OJP has been diﬃcult. The widely used 40Ar/39Ar method for dating igneous rocks is problematic as OJP basalts contain very little potassium. Published 40Ar/39Ar data point to a major episode at 122 Ma and a possible second minor event at ∼ 90 Ma [158, 241, 242]. However, Chambers et al. [34] suggest that the younger 40Ar/39Ar ages are the result of Ar recoil and therefore represent minimum ages. Biostratigraphic dating of overlying sediments suggests that magmatism extended into the late Aptian, which corresponds to age ranges of 122-112 Ma [225] supporting the 122 Ma 40Ar/39Ar age. Re-Os isotopic data from ODP Leg 192, Sites 1183, 1185, 1186, 1187 deﬁne an isochron with an age of 121.5 ± 1.7 Ma [188].

11 150 155 160 165 170 175 15 15 0 00 -60 -5 N 00 -4

000 -50 00 00 -50 -50 -4000 -6 00 00 00 -2000 802 0 -50 -300 0 East Mariana 0 00 -50 00 00

-300 00 10 -30 -40 10 Basin 00 -3000

-40 -40 00 -50 -30 00 00 -30 -50 -40 00 00 462 00 -40 0 00 00 -4000 -3 -30

-30 5 00 5 807 -40 Nauru 00 803 Basin

0 1187 -400 1000 kilometres -30 289 00 0 -4 -20 1185 0 00 00 0 Kroenke Canyon

-20 1186 -4 1183 00 000Ellic Basin High

00 0 plateau -3000 ient -50 -30 sal 00 -2000 -2000 5 5 -20 ern 00 0 1184 -70 East -3000 00 00 -1000 Stewart Basin -20 -10 Santa Isabel -40 00 00 00

0 -30

-30 -1 00 -4 00 0 Malaita 000 00 -4 0 -500 -10 10 00 0 10 -30 -30 San Cristobal 00 00 000 -60 S -2 00

150 E 155 160 165 170 175

Figure 2.1. Map of the Ontong Java Plateau, highlighting the island of Malaita

The Malaita Volcanic Group was subdivided by Tejada et al. [242] into two chemically and isotopically distinct formations: the Kwaimbaita Formation and the Singgalo Formation (Fig. 2.2). The basalt sequence exposed on Malaita re- sembles an expanded version of the core recovered from Site 803 on the northern ﬂank of the OJP (Figs. 1.1 and 2.1), where the Kwaimbaita Formation is compositionally identical to units C-G and the Singgalo Formation is identical to unit

A. A recently discovered third lava type was found at Sites 1185 and 1187 near the submarine Kroenke Canyon and was termed Kroenke-type basalt [155]. This lava type has higher MgO and lower concentrations of incompatible trace elements than either the Kwaimbaita- or Singgalo-type basalts but is isotopically identical to Kwaimbaita-type basalts.

12 12

MORB

Kwaimbaita, Kroenke 6 Wairahito (Units C-G) Epsilon Nd (t)

4 Singgalo (Unit A)

A) 2 0.702 0.7025 0.703 0.7035 0.704 0.7045 0.705 87Sr/86Sr (t)

87 86 Figure 2.2. Graph of Sr/ Sr vs. ǫNd for OJP basalts and MORB basalts. Data sources: [154, 156, 241, 242]

Kwaimbaita-type basalt is found at all but one drill site and therefore represents the dominant magma type while the Singgalo-type basalt is signiﬁcantly thinner (> 2.7 km thick vs. 750 m maximum thickness, respectively) and is apparently volumetrically minor [79]. The isotopic similarity of the Kroenke-type basalts to the Kwaimbaita-type, combined with their less evolved trace element composition suggests that it may be the parental magma for the bulk of the OJP

[240]. Tejada et al. [240] modeled the Sr-, Nd- and Hf-isotopic composition of the Kwaimbaita- and Kroenke-type basalts as primitive mantle that experienced a minor fractionation event about 3 Ga prior to OJP formation. In addition to

13 100 Units C-G eltn Unit A a M e viti

10 m ir P/elp m a S 1 Th Nb Ta La Ce Pr Sr Nd Zr Hf Sm Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu

Figure 2.3. Plot showing primitive mantle normalized proﬁles of basalts from Unit A and Units C-G from Site 803. Data source: [155]

minor isotopic heterogeneity, the OJP is remarkably homogeneous in trace element composition (Fig. 2.3). Incompatible element abundances in OJP magmas can be modeled by ∼ 30% melting of a peridotite primitive magma source from which about 1% by mass of average continental crust had been extracted [78]. This modeling is consistent with the isotopic modeling of Tejada et al. [240]. Thin lithosphere in conjunction with very hot mantle potential temperatures is required to produce a 30% melt of peridotite mantle. These conditions are consistent with the conclusion of Kroenke et al. [133] that the OJP formed near a recently abandoned spreading center. High degrees of melting are supported by peridotite phase equilibria consider- ations [107], where forward- and inverse-modeling approaches obtain values of 27 and 30% partial melting (fractional and equilibrium melting, respectively). PGE systematics of OJP basalts also conﬁrm high degrees of partial melting. High PGE concentrations in OJP basalts are consistent with ∼ 30% melting of a peridotite source from which sulﬁde phases are exhausted during melting [35]. Some basalt samples have PGE abundances too high to be accounted for by normal peridotite

14 mantle sources, so Chazey & Neal [35] speculated that minor amounts of outer core material were incorporated into the OJP source region. The isotopic similarities of the incompatible-element poor and MgO-rich Kroenke- type lavas found at Sites 1185 and 1187 to the ubiquitous Kwaimbaita-type magmas suggest that the former may be parental to the latter. Fractional crystallization models conﬁrm that this is indeed possible [78, 240] and melting experiments show that the phenocryst assemblage and major element characteristics can be adequately modeled by fractional crystallization in shallow (< 6 km) magma chambers [218].

2.5 Volcaniclastic rocks and Site 1184

ODP Leg 192, Site 1184 is located on the eastern salient of the OJP (Fig. 2.4), which prior to Leg 192 had not been sampled by deep sea drilling. The most significant finding from Site 1184 was the thick sequence of volcaniclastic rocks that unequivocally indicated that the succession was the result of large phreatomagmatic subaerial eruptions. Wood fragments were found at the boundaries of four of the six subunits (as defined by Thordarson [243]), and as lava flows on the main plateau were erupted under deep water [192, 213], the volcaniclastic succession at

Site 1184 provide the only evidence of subaerial eruptions from the OJP. However, basement lava ﬂows were not reached. The volcaniclastic succession at Site 1184 is interpreted to consist of density- current deposits formed by erosion and reworking of nearby ediﬁces over a period of about 3 Ma [155, 243]. However, the origin of the deposits remained unclear due to ambiguity about the relationship of the eastern salient and Site 1184 to the main plateau (see Subsection 1.1.1). In addition, fossil ages obtained from

15 150 155 160 165 170 175 15 15 0 00 -60 -5 N 00 -4 000 -50 00 00 -50 -50 -4000 -6 00 00 00 -2000 802 0 -50 -300 0 East Mariana 0 00 -50 0 00 00 -300 00 10 -30 -4 10 Basin 00 -3000

-40 -40 00 -50 -30 00 00 -30 -50 -40 00 00 462 00 -40 0 00 00 -4000 -3 -30

-30 5 00 5 807 -40 Nauru 00 803 Basin

0 1187 -400 1000 kilometres -30 289 00 0 -4 -20 1185 0 000 00 Kroenke Canyon -20 1186 0 1183 0 High

00 0 plateau -50 -30 00 -2000 5 -20 5 00 0 1184 ient -70 -3000 00 sal 00 -1000 Stewart Basin -20 -10 Santa I ern -40 00 East 00

-30 00

-500 10 0 10 0 San Cristobal 0 -60 S -20 00

150 E 155 160 165 170 175

Figure 2.4. Map of the Ontong Java Plateau, highlighting the eastern salient and Site 1184

nanofossil assemblages diﬀer from radiometric ages by as much as 75 Ma, again leading to uncertainty about the origin of the Site 1184 volcaniclastic deposits

(mid-Eocene vs. lower-Cretaceous, respectively) [33, 155]. Chemical and isotopic evidence suggest a strong aﬃnity to OJP basalt types, lending credence to the radiometric age of ∼ 123 Ma [33, 78, 223, 251]. In addition, detailed examination of the Site 1184 drill core indicates that nanofossils present in the core are likely to be a late addition to the succession [243]. A steep paleomagnetic inclination of the Site 1184 rocks also is consistent with a Cretaceous emplacement age [212].

16 CHAPTER 3

HAWAIIAN RIDGE-EMPEROR SEAMOUNT CHAIN

3.1 Overview

The purpose of this chapter to summarize new research into the origin and evolution of the Hawaiian Ridge-Emperor Seamount Chain and integrate it with previous studies. It will also serve as the location for background information in Chapters 6 and 7. Basic information (such as size, age, basalt types) was introduced in Chapter 1.

3.2 History of Mantle Plumes

Mantle plume theory begins with Hawaii. The Hawaiian Islands were first studied in detail in the mid to late part of the 19th century by James Dana, who determined the initial structure and history of the Hawaiian volcanoes. It wasn’t until the discovery of plate tectonics in the early 1960’s [246] that Wilson [252] postulated that the linear oceanic island chains in the Pacific Ocean could be caused by the Pacific Plate moving over a fixed magmatic point source termed a hotspot. This idea readily took hold and has been the preferred explanation for linear chains of oceanic islands and seamounts, although other alternative ideas have been proposed [5–9, 128]. In the 40 years since the hotspot theory has been proposed by Wilson, the exact nature and source of these hotspots has been

17 debated. First, what is the nature of the hotspot? Does the source of the magmas arise from the lower crust, shallow mantle, the 410 km or 660 km discontinuity, or perhaps the D” layer at the core-mantle boundary? How does the magma rise through the mantle? These questions lead to idea of mantle plumes, or thermally buoyant upwelling of less viscous mantle material. There are several lines of evidence that suggest that mantle plumes (or at least some of them) originate at the core-mantle boundary [30, 53, 62, 95, 167]:

1. Computer modeling and scale models of rising plumes indicate that only plume heads arising from the core-mantle boundary can achieve the size (> 1000 km diameter) required to form the large volumes of magma that

compose submarine plateaux like the Ontong Java Plateau and Kerguelen Plateau.

2. The approximately fixed position of hotspots relative to one another is consistent with a deep mantle source. The fixed position would be difficult to

maintain in the convecting upper mantle.

3. The amount of heat transferred to the base of the lithosphere by plumes is approximately 12 per cent of the Earth’s total heat ﬂux and is comparable to estimates of the amount of heat released by the core as it cools.

4. Periods of increased plume activity seem to correlate with periods of normal polarity and decreased pole reversals in plume-related basalts [139, 140]. This seems to imply that the beginning of heat transfer across the D” layer (the origins of a mantle plume) alters the convective ﬂow in the outer core

resulting in a change in the Earth’s magnetic ﬁeld.

18 5. The Earth’s core is expected to be high in Os due to the highly siderophile

nature of the platinum group elements and, by extension, to have highly

1 radiogenic levels of 186Os/188Os (190Pt → 186Os, T 2 = 450 Ga) and mod-

1 erately radiogenic 187Os/188Os (187Re → 187Os, T 2 = 41.6 Ga) compared to the lower mantle. Some plume-derived lavas have correlated elevations

in their 187Os/188Os and 186Os/188Os ratios which suggests contamination by outer core material. This theory will be explored in greater detail in Chapter 7.

6. Mantle tomography suggests that at least some mantle plumes extend to the base of the lower mantle [120]

The theory states that a disruption along the D” layer causes a local heat anomaly to cross from the outer core to the lower mantle. This causes the adjacent mantle to become thermally buoyant and begin to rise and, at the same time, begin to entrain nearby mantle. If the thermal anomaly is large enough (computer simulations suggest that the anomaly needs to be in excess of 200 ◦C), the budding plume head begins to ascend from the lower mantle entraining nearby mantle material and leaving the trailing plume tail (Fig. 3.1). Simulations suggest the the plume head would then impinge upon the lower lithosphere in 50 to 100 My

[53]. After the plume head reaches the lithosphere, it begins to ﬂatten and eventually intersects the mantle solidus and undergoes rapid decompression melting, producing large volumes of basaltic magma and creating large oceanic plateaux or continental ﬂood basalts (generically termed “large igneous provinces” or LIPs). As the overriding lithosphere begins to move away, the plume tail creates smaller oceanic islands and, over time, a linear oceanic island chain.

19 Figure 3.1. Schematic of a rising mantle plume. The ﬁgure shows the rising hot, buoyant material has entrained cooler surrounding mantle and is trailed by a hot tail of thermally buoyant material from the source of the mantle plume

However, despite multiple lines of evidence that hotspots are the result of mantle plumes, there are multiple arguments for diﬀerent mechanisms. For example, the Hawaiian Ridge-Emperor Seamount Chain, often times quoted as the prime example of a linear oceanic island chain produced by a mantle plume, has no corresponding oceanic plateau. Other oceanic plateaux and continental ﬂood basalts that are attributed to plume head magmatism have no corresponding oceanic island chain (for example, the Ontong Java Plateau). The geochemical arguments

20 about enriched levels of Os isotopes have been countered by invoking recycled, seawater-altered oceanic crust [142] or recycling of Fe–Mn crusts [219]. The relative ﬁxed positions of hotspots has come under criticism and, in fact, evidence has been shown to indicate the the Hawaii hotspot has moved southwards at rates of approximately 30 mm/year [235, 237]. Some workers believe that while mantle plumes exist, they do not necessarily have to form at the D” layer and have to be the result of thermal anomalies across the core-mantle boundary. Sinking slabs of oceanic crust are thought be able to maintain distinct geochemical and physical properties as they descend into the mantle and could be a potential mechanism for the production of mantle plumes. These arguments are generally based on isotopic evidence [142, 150] from Hawaiian plume lavas. Oceanic crust and sediments likely to have higher 87Sr/86Sr and lower 143Nd/144Nd than depleted mantle, which explains some of the isotopic characteristics of the depleted Koolau end member in Hawaiian magmas [142]. In addition, the recycling of oceanic crust and sediments is a viable way of producing O-isotope heterogeneities. The depleted Hawaiian Koolau end member shows a δ18O enrichment that can be reproduced by addition of approximately 30% recycled material to the Hawaiian plume. Oceanic crust is also quite high in Re due to the incompatible nature of Re during mantle melting and is correspondingly low in Os due to osmium’s compatible nature. Because of this, oceanic crust evolves highly radiogenic 187Os/188Os (due to the beta decay of 187Re to 187Os) over time. Addition of recycled material to plume material creates a hybrid source with 187Os/188Os values similar to that of modern Hawaiian basalts. The recycled oceanic crust theory generally requires that ∼ 30% recycled oceanic crust and pelagic sediment is mixed into the plume to generate the observed isotopic heterogeneities and compositions seen in Hawaiian plume material.

21 Not surprisingly, the proponents of a lower mantle/core-mantle boundary origin of mantle plumes counter the previous arguments by suggesting that the oceanic crust does not have enough Pt to be able to substantially alter the 186Os/188Os composition of Hawaiian lavas. Oceanic crust and sediments can have Pt/Os ratios up to 45 times greater than primitive mantle, however, this elevated ratio is due to the extremely low levels of Os in these reservoirs rather than elevated Pt. The binary mixing equation between two isotopic end members (see Chapter 7, Section 7.4.2, equation 1) indicates that if the abundance of component is extremely low it takes large amounts of that component to facilitate a change in the mixture’s composition. In this case, it takes between 70 and 95% of the recycled component to change the 186Os/188Os composition of the mixture to match the maximum found in the Hawaiian lavas [25]. This amount of recycled material required is much too great and would cause the resulting lavas generated from such a hybrid source to be tonalitic rather than tholeiitic, which are not found in Hawaiian lavas [25]. Brandon et al. [25, 26] suggested that the coupled enrichment in 186Os/188Os and 187Os/188Os values in Hawaiian lavas are the result of contamination of the Hawaiian plume with ∼ 2% outer core material, thereby showing that the plume originated from the core-mantle boundary. Examination of the Re–Os and Pt–Os systematics in other mantle plume basalts indicate that there may be a common highly radiogenic 187Os/188Os and 186Os/188Os component to all of the examined mantle plume basalts (Hawaii, Gorgona Island, Noril’sk komatiites, and SW Oregon Os-rich alloys). Brandon et al. [27] postulated that this common component is the outer core.

However, recent work [219] on the 182Hf-182W isotopic system from the same Hawaiian picrites containing the coupled radiogenic Os isotopes suggests that

22 there is no contribution of material from the outer core. Tungsten is a siderophile element that would have been partitioned into the Earth’s core during core formation while Hf is lithophile and would have stayed in the silicate mantle. Because

182 of this, the core will be highly depleted in W and have a strongly negative ǫW value. Any source that samples the outer core should inherit this depleted signature. The Hawaiian picrites that show the coupled enrichments in Os isotopes do not show a negative ǫW signature [219]. Schersten et al. [219] suggested that the coupled enrichments in Os isotopes and bulk silicate Earth ǫW values in the Hawaiian picrites can be caused by incorporation of < 2–4% of Mn-nodules into the source region. This hybrid source can reproduce the Os anomalies and yet yield no W-anomalies [219]. In addition, because the Mn-nodules are so high in Mn (∼ 25 wt.%), there would be a 5–9% increase in Mn concentrations in the source and the Fe/Mn ratio would decrease [219]. Changes in Mn concentrations would be hard to detect due to signiﬁcant variations in major element concentrations due to melting conditions. However, Humayan et al. [113] have conducted precise measurements of the Fe/Mn ratio of Hawaiian lavas and determined that that there is a 1 to 2% increase in the mole fraction of iron in the mantle beneath Hawaii, in contradiction to hypothesis put forth by Schersten et al. [219]. These data suggest that Mn-nodules cannot be the source of the Hawaiian Os isotope signal. At this point, the evidence for core-mantle interaction is inconclusive. How- ever, for the purpose of this dissertation, I believe the combination of physical geology, computer modeling, and geochemical evidence make the core-mantle boundary theory the most compelling. In any case, it seems clear the outer core must have some role to play in the origination of at least some large and long-

23 lived mantle plumes such as Hawaii. It also seems likely that recycled oceanic crust and perhaps a veneer of oceanic sediments can be subducted and remain a distinct component and play a role in some of the geochemical signatures that have been observed from oceanic islands around the world.

3.3 The Hawaiian Ridge-Emperor Seamount Chain

The intraplate Hawaiian Ridge-Emperor Seamount Chain (HR-ESC) is an approximately 6,000 km long, age-progressive chain of volcanic islands and seamounts (Figure 3.2). Active volcanism occurs on the Hawaiian Islands and the volcanoes become progressively older to the west and north. The oldest preserved volcanoes, Meiji and Detroit Seamounts, have an age of approximately 76–85 Ma [72, 123].

3.3.1 Physical Geology

The Hawaiian hotspot produces shield volcanoes, a type of structure that is broad and gently sloped resulting from the eruption of magma less viscous than lavas produced at subduction zones. This hotspot has produced the largest volcanoes on Earth, the largest of which rise nearly 9 km from the ocean ﬂoor [190]. Hawaiian volcanoes tend to go through a series of stages during formation (although not every volcano goes through every stage). The stages are [190]:

1. Eruption of submarine alkalic lavas.

2. Eruption of tholeiitic lava pillows. Subsidence begins to occur.

3. Eruption of shallow water hyaloclastics.

4. Eruption of main tholeiitic shield.

24 Figure 3.2. Map of the Hawaiian Ridge-Emperor Seamount Chain and surrounding areas. Modiﬁed from Tarduno et al. [236]. Shaded areas are regions of sea ﬂoor extending above approximately 3500 meters below sea level (mbsl).

5. Eruption of post-shield alkalic lavas.

6. Period of quiescence and erosion.

7. Eruption of rejuvenated stage lavas. Typically characterized by eruption of strongly alkalic lavas.

8. Atoll formation.

9. Erosion and subsidence resulting formation of a seamount.

By the end of the active life of the volcano, approximately half the erupted sequence has been subsided below the ocean ﬂoor [190]. As mentioned before, not all Hawaiian volcanoes have gone through each stage. In particular, the post-shield alkalic lavas and rejuvenated stage lavas are not present on all volcanoes. This

25 could be because these lavas tend to be volumetrically minor in comparison to the shield stage tholeiites and the long period between eruptions of post-shield alkalic and rejuvenated stage lavas means that erosion could could remove them entirely. The order of the stages is due to the progression of the Paciﬁc plate over the Hawaiian hotspot [190]. As the hotspot leaves a volcano and begins to create a new one over fresh lithosphere, the initial amounts of lithosphere melting and magma production are small. These low volume melts are alkalic in nature. As the magma pathways become more pronounced under the proto-volcano, the amount of primary magma from the plume increases and tholeiitic lavas are erupted in deep water. The high pressure from the water and low gas contents of the tholeiitic magma result in pillow lavas that characterize stage 2. As the volcano grows in size, eventually the water pressure lessens and some of the lavas are erupted as hyaloclastic deposits (meaning that they erupt somewhat explosively due to steam generated by the hot magma interacting with seawater). Once the volcano becomes subaerial, the volcano grows rapidly due to the high magma production rates and easy transport to the surface through the well developed plumbing system and magma chambers (stage 4). Eventually, the hotspot moves on and magma production and amounts of partial melting decrease. In stage 5, alkalic lavas re- place tholeiitic lavas. However, the production rate during stage 5 is lower relative to the stage 4 and there may be long periods of dormancy between post-shield alkalic lava eruptions. Eventually, all eruptions stop and erosion is dominant (stage 6). Somewhere between half a million and 3 million years later, some volcanoes undergo a rejuvenated stage of volcanism. The lavas erupted during this stage are generally highly alkalic. Some contain garnet xenoliths, indicating an origin at great depths within the mantle (> 70 km). All Hawaiian volcanoes appear to

26 go through these stages [190] and drill cores recovered from ESC seamounts show that some are capped by post-shield alkalic lavas and that the bulk of the shield is tholeiitic [236]. Bathymetry and geophysical measurements indicate that the structure of these seamounts are very similar to those from the Hawaiian islands [236].

3.3.2 Geochemistry of the Hawaiian Islands

The geochemical and petrologic structure of a mantle plume is assumed to be axisymmetric during its ascent to the base of the lower lithosphere, however, interaction with the moving plate should cause the plume to become deformed and that some aspects of the symmetry would be modiﬁed [67, 208, 209]. This should include the geochemical structure of the plume. In fact, studies of the Hawaiian plume show that there are two distinct geochemical trends present in the Hawaiian Islands [86]: the Kea and Loa trends (Fig. 3.3. The Kea trend volcanoes are oriented such that they traverse the northeastern part of the assumed plume top while the Loa trend volcanoes are aligned more closely to the plume axis (Fig. 3.3).

Kea trend lavas tend have lower 208Pb/204Pb at a given 206Pb/204Pb than Loa trend volcanoes [135, 141] and form different trends in plots of 208Pb/204Pb vs Hf, Sr, and Nd isotope ratios (Fig. 3.4). The differences in isotopic ratios between the two trends are presumed to reflect changes in the source region of the magma, i.e., the Loa and Kea trend volcanoes sample different parts of a geochemically zoned plume. Recent research [206] suggests however that the two (at least) mantle components necessary to account for the Kea and Loa trend geochemical characteristics may be present to at least some extent in all lavas and that the plume is likely not

27 Figure 3.3. Map of the Hawaiian Islands showing the Kea and Loa trends. Prominent volcanoes are as follows: KI, Kilauea; ML, Mauna Loa; MK, Mauna Kea; H, Hualalai; KO, Kohala. Reproduced from Ren et al. [206].

to be compositionally zoned. Instead, the chemical diﬀerences in the two trends is controlled by the thermal structure of the plume. Ren et al. [206] studied the major and trace element composition of melt inclusions in olivine in order to better observe the composition of Hawaiian magmas prior to shallow level magma mixing. Ren et al.’s [206] data suggest the random heterogeneities are more common than a structured zoned plume would suggest and envision streaks or ribbons on one component being deformed and stretched throughout a matrix of the other other component.

28 stpcrto hwn h ed o e n o rn volca trend Loa and Kea for ﬁelds the showing ratios isotopic rmGOO est:http://georoc.mpch-mainz.gwdg.de/ website: GEOROC from iue3.4. Figure 208 143Nd/144Nd 87Sr/86Sr 176Hf/177Hf Pb/ 204 bvs Pb 176 Hf/ 29 208 208 208 177 Pb/ Pb/ Pb/ 204 204 204 Hf, Pb Pb Pb 87 Sr/ 86 r and Sr, 143 Nd/ os Data noes. georoc/ 144 Nd hr rpinrtsaegetydmnse eaiet the to relative diminished greatly are rates eruption th and where stage alkalic post-shield the during common more much mugearites and benmoreites, trachytes, hawaiites, as of such formation the to leading chambers a magma clinopyroxene level higher of some fractionation frac Further during 3.5). olivine (Fig. of [258] removal the to attributed usually typic are in elements trace and major [153] both in Katsura trends & geochemical MacDonald and [198] Powers by conducted were rsalzto hudices h ereo niheti elements. enrichment ble addition of In degree the melting. increase partial should crystallization of degree lower suggestin a MORB, and/or to source relative elements trace incompatible in tholeiit Kilauea average The basalts. island oceanic other h rc lmn opsto fHwia hlitsi gen is tholeiites Hawaiian of composition element trace The Hawa of composition element trace and major the on work Early iue35 g sSiO vs MgO 3.5. Figure oto nSiO on control est:http://georoc.mpch-mainz.gwdg.de/georoc/ website:

SiO2 (wt.%) 2 n g ocnrtos aafo GEOROC from Data concentrations. MgO and 2 o iae hlitcbslssoigolivine showing basalts tholeiitic Kilauea for

Oli

vine

Co ntrol Line ntrol opsto senriched is composition e hs vle aa are lavas evolved These . h ihyincompati- highly the n dpaicaeocrin occur plagioclase nd hedbidn stage. building shield lHwia tholeiites Hawaiian al inlcrystallization tional ihyeovdlavas evolved highly eueae stage rejuvenated e ute fractional further , oeenriched more a g rlysmlrto similar erally h dominant The . inbasalts iian Figure 3.6. Incompatible element diagram of Kilauea shield stage tholeiites and EPR MORB. Note how the highly incompatible elements on the left hand side of the diagram are enriched in the Hawaiian tholeiite and depleted in the EPR MORB. Kilauea data from GEOROC website (http://georoc.mpch-mainz.gwdg.de/georoc/Entry.html). EPR MORB data from Regelous et al. [202].

Shield stage tholeiites from individual volcanoes screened to ∼ 7–8% MgO are generally quite uniform [147]. However, systematic differences in composition are seen from individual volcanoes that cannot be attributed to differentiation of these magmas from a single common source [147]. Because nearly 90% of the island of Hawaii is submarine and the earliest shield-building stage lavas are similarly buried and far below sea level, it is difficult to determine the long-term evolution of the shield-stage tholeiites. Recent drilling from the Hawaiian Scientific Drilling Project (HSDP) recovered greater than 3000 m of stratigraphic section encompassing basalts from both Mauna Loa and Mauna Kea, which allows for examination of long-term evolutionary trends in the shield-building stage [73, 77]. REE inverse modeling performed on these lavas indicate that at least there are

31 at least two distinct populations of Mauna Kea lavas; one of which samples a relatively fertile, possibly interior part of the plume and another more depleted source that would lie on the periphery of the plume [77]. The data of Feigenson et

87 86 al. support the idea of the chemically zoned plume. Isotopic data(ǫNd, Sr/ Sr, and 3He/4He) from Mauna Loa and Mauna Kea lavas from the HSDP pilot hole

[67] and from the literature were used to create and isotopic “map” of the Hawaiian plume (Fig. 3.7). The isotopic data suggest that the high 3He/4He values are conﬁned to the core of the plume and correspond to potential temperatures in excess of 1565 ◦C. The contours on the isotopic map (Fig. 3.7) are elongated in the general direction of plate motion. Epsilon-Nd contours extend past the big island of Hawaii and in some locations plume ǫNd exists with asthenospheric

3 4 87 86 He/ He. Low ǫNd and high Sr/ Sr data from Mauna Loa form a blob like

87 86 structure, which suggest that the low ǫNd-high Sr/ Sr component is not a ring of entrained lower mantle material. However, the isotopic HSDP data also suggest that the plume structure is chemically zoned and is being deformed by the passage of the Paciﬁc Plate. While Hawaiian volcanoes are predominantly composed of tholeiitic basalt, they also feature volumetrically minor late-stage alkalic lavas and occasional eruptions of high-MgO and picritic lavas. MacDonald & Katsura [153], after studying the major element composition of Hawaiian lavas, developed the dividing line between alkalic and tholeiitic lavas on the SiO2–Total Alkali (Na2O+K2O) discrimination diagram. The relationship of the post-shield alkalic lavas to the shield building tholeiitic lavas has not always been clear. Geochemical models used to explain their relationship have been proposed [36] (see Chapter 6) by arguing that as hot enriched plume material impinges upon the base of the lower lithosphere it

32 Figure 3.7. Map of the distribution of R/Ra (ratio of the 3He/4He ratio of the sample to that of the atmosphere) and ǫNd on the big island of Hawaii. Note the elongation of the R/Ra core plume signal and the low ǫNd blob oﬀset from the plume core. Reproduced from DePaolo et al. [67].

creates low volume, small percentage incipient melts of the isotopically and incompatible element depleted wall rocks. These incipient (low-degree of partial melting) melts are created in the beginning stages of volcano formation as the plume material just begins to create the magma plumbing pathways and during late stages as the plume is moving away and cooler plume material creates smaller degree alkalic melts. The incipient melt mixes with the plume material creating a hybrid source capable of producing the range of compositions seen in Hawaiian lavas (Fig. 3.8).

33 0.65

20% 0.6 0.8% 10% 5.0% 3.0% 1.0% 0.55 1.0%

0.5 1.5% e C /a L 0.45 2.3%

0.4

UD 0.35 Re-fertilized Oceanic Lithosphere

MORB Source 0.3 0.702 0.7025 0.703 0.7035 0.704 0.7045 0.705

87Sr/86Sr

Figure 3.8. Origin of Alkalic Pre- and Post-Shield Basalts. Plot shows how low degree incipient melts of lithospheric wall rock mixing with plume material can generate the trace and isotopic characteristics of alkalic post-shield basalts. These data are from Nintoku seamount. Full description of the model used to generate this ﬁgure can be found in Chapter 6 and in Chen & Frey [36].

3.3.3 Geochemistry of the Emperor Seamount Chain (ESC)

Early sampling of the ESC was conducted during the Deep Sea Drilling Project (DSDP) Legs 55 and 145, which recovered samples from Suiko, Nintoku, Ojin, and Detroit seamounts [18, 42, 130]. Prior to that expedition, the only available samples from the ESC were dredged fragments of uncertain origin. De- spite the advances made during DSDP Legs 55 and 145, only broad general- izations could be made, as the majority of the drilling only shallowly pierced the seamounts (drilling at Nintoku seamount on Leg 55 recovered only 32 m of

34 basement and at Site 884 at Detroit Seamount only 87 m was recovered). De- spite this, some of the gross features of the ESC became apparent and it was determined that seamount morphology and structure are similar to those of the modern Hawaiian islands [18, 42, 130]. Drilling at Site 432 (Nintoku Seamount) during DSDP Leg 55 recovered alkalic basalt and evolved fragments of hawaiite contained within a conglomerate unit overlying the volcanic basement suggesting that Nintoku may be capped by post-shield alkalic basalts [130]. The similarities in morphology and lava petrology suggests that the ESC volcanoes passed through a series of evolutionary stages similar to those of young Hawaiian volcanoes [149, 201].

The more recent Ocean Drilling Program (ODP) Leg 197 drilled Detroit, Nintoku, and Koko Seamounts [236]. In each case, signiﬁcantly deeper (several hundred meters vs 10’s of meters) penetration was achieved than during DSDP drilling. Combining the results for these drilling legs shows that the geochemistry of lavas from the ESC vary dramatically from one end to the other. Tholeiitic and alkalic lavas from Meiji and Detroit Seamounts are depleted in highly incompatible trace elements compared to their younger counterparts from the Hawaiian Islands [124, 201]. Initial 87Sr/86Sr ratios of tholeiitic and alkalic lavas from Detroit and Meiji Seamounts extend to unradiogenic (depleted) values relative to lavas from younger Hawaiian Ridge seamounts [112, 124, 201] (Fig. 3.9). In contrast, the available geochemistry data for lavas from younger Emperor Seamounts (between Suiko and Daikakuji, 65–42 Ma) show that these seamounts have major and trace element compositions similar to those of young

Hawaiian lavas [18, 41, 43, 59, 60, 124, 130, 201]. However, with the exception of Suiko Seamount, samples from the younger Emperor Seamounts are limited to

35 Distance from Kilauea (km)

1,000 2,000 3,000 4,000 5,000 M ( e (

87 86 c 0.7038 Sr/ Sr n e fid e fid rS

6 0.7034 100 g 8 / a ts a r Seamount-crustag ala idf ef e e S ren i c 7 a y s 8 a h urc-tn s o a elc k io o u hiN w olo o o K rte di k re k nij 0.7030 a 50 a o n yru M M P a P niP O K u D t o Y u L e

? m Suiko a

0.7026 e 0 20 40 60 80 S ref )ry Seamount age (Myr)

Figure 3.9. 87Sr/86Sr vs. seamount age for the Hawaiian Ridge-Emperor Seamount Chain. The 87Sr/86Sr values of the ESC seamounts decreases with increasing seamount age and with seamount-crust age diﬀerence. The 87Sr/86Sr composition of Detroit Seamount is nearly identical to that of MORB. Figure from Keller et al. [124].

a few dredge and shallow drill sites. Dredged samples in particular are generally highly altered, so the primary geochemistry of lavas from most of these sites is poorly known. Figure 3.9 shows that the 87Sr/86Sr values of the ESC decrease with time until they are approximately equivalent to MORB at Detroit and Meiji seamounts. Keller et al. [124] interpreted these data as being indicative of mixing between a “Kilauea-like” plume component and a depleted component with the isotopic characteristics of MORB. This idea is lent credence by the fact that some plate reconstructions place an active spreading center near the position of the Hawaiian hotspot at the time of Detroit Seamount formation [159]. The idea of magma mixing is supported by detailed laser ablation ICP-MS work on plagioclase crystals from Detroit Seamount that showed that the cores of the plagioclase crystals were enriched in Sr, Ba, and the LREE relative to the rim of the crystal [129]. Kinman

36 & Neal [129] interpreted these results to suggest that the crystal initially grew in a plume magma dominated magma chamber, but later grew in a more MORB rich magma chamber, possible due to the inﬂux of MORB magma into the magma chamber. Due to the Hawaiian plume being in close proximity to a mid-ocean ridge at ∼ 81 Myr, the thickness of the lithosphere during formation of Detroit and Meiji seamount was much thinner than present. Studies have shown the oceanic lithosphere thickness increases until the lithosphere age reaches ∼ 70 Ma, at which point it stays relatively constant [189, 201]. Regelous et al. [201] reports that some aspects of the Sr-Nd and Nd-Hf isotopic compositions of De- troit and Meiji lavas are inconsistent with magma mixing between MORB and plume lavas. In addition, the Pb isotope composition of Detroit lavas are different from EPR MORB. While the Detroit lavas overlap the MORB ﬁeld in

ǫSr–ǫNd space, they tend to have a higher ǫSr for a given ǫNd and lie on an extension of the Hawaiian array [201]. These factors led Regelous et al. [201] to propose that the systematic changes in isotopic and trace element composition along the ESC is a result of changing lithosphere thickness. In this theory, the Hawaiian plume is envisioned as a “plum-pudding” type mantle [178, 179, 196,

228, 233] with enriched heterogeneities distributed throughout a depleted refractory matrix. When the plume is situated under thin lithosphere (as would be the case at Meiji and Detroit seamounts), the increased melting column allows for a greater percentage of a depleted refractory matrix to contribute to the melt and lowers trace element abundances and depletes the isotopic signature. Under thicker lithosphere, the melting column is shorter (degree of partial melting is lower) and less of the depleted matrix can contribute to the melt, yielding en-

37 riched isotopic and trace element signatures like those from the younger ESC and

Hawaii.

3.3.4 Movement of the Hawaiian Mantle Plume

The bend in the HR-ESC is one its most prominent features and has been cited as a prime example of a fixed hotspot recording a change in direction of the overriding plate. The logical next step was to try to use hotspots as a fixed frame of reference to deduce plate motions [167]. Despite the allure of this feature, soon after this hotspot reference frame was attempted to be put into use discrepancies were reported [166]. Models of past plate motions failed to predict the bend [230] and tests of the fixed nature of hotspots predicted large relative motions of Hawaii to other hotspots [166]. At this point, it was unclear whether the bend in the HR-ESC was due to a change in plate motion, relative motion of the hotspot, or a combination of the two. More recently, Norton [186] suggested that the bend records the spot when the Hawaiian hotspot became fixed in the mantle. In this scenario, during formation of the ESC, the Hawaiian hotspot moved southward until 43 Ma, at which point the hotspot became fixed in the mantle and the more westwardly track of the Hawaiian Ridge was formed. Tarduno & Cottrell [235] and Tarduno et al. [237] tried to verify this hypothesis by collecting paleomagnetic and paleolatitude data from the ESC. Their early data (based on limited sampling before ODP Leg 197) suggested that the Hawaiian hotspot may have moved at rates comparable to lithospheric plate motions (> 30 mm/yr) during formation of the ESC. In addition, they determined that Atlantic hotspots moved at similar rates during the early to mid-Cretaceous. After ODP Leg 197 provided a greater number of samples from the principle seamounts along the ESC, their

38 initial ﬁndings were conﬁrmed [237]. Tarduno et al. [237] determined that the

Hawaiian hotspot had moved in a nearly southerly direction at a rate in excess of 40 mm/yr based on paleolatitude data from basalts from Koko, Nintoku, Suiko, and Detroit seamounts. These results suggest that plate motion changes may not be as fast or as abrupt as previously thought.

If the cessation of southerly movement of the Hawaiian plume at 43 Ma is the result of the plume becoming “ﬁxed” in the mantle, then perhaps there is geochemical indicator that could tie the ﬁxing of the plume to a physical event. Or is the movement of the plume related to large-scale mantle dynamics that are periodic or episodic? My research attempts to, at least in part, to address this.

3.4 Summary

The research presented in this dissertation will be used to help answer several questions about the source location and composition of the Hawaiian hotspot.

On the basis of previous research, it is assumed that the Hawaiian hotspot is a mantle plume. Chapter 6 will attempt to determine if a typical ESC seamount is chemically similar to the modern Hawaiian Islands. In addition, application of petrogenic models from the Hawaiian Islands will be applied to the Nintoku samples to determine if the same mechanism is viable during the formation of the ESC. These data will help determine if the geochemical dynamics found in the modern Hawaiian Islands were similar during the formation of the ESC. Chapter 7 examines the platinum group element and 187Os/188Os composition of three ESC seamounts. As expanded upon in Chapter 7, the PGE’s and their associated isotopic systems are very useful to examine the source characteristics of lavas. Chapter 7 attempts to add insight to the ongoing question of whether the Hawaiian

39 plume originates at the core-mantle boundary and if it is contaminated with a small amount of outer core material.

40 CHAPTER 4

COMPOSITIONAL VARIABILITY IN LAVAS FROM OCEAN DRILLING PROGRAM SITE 1184, ONTONG JAVA PLATEAU

Background information, geological setting, and introductory material covering the Ontong Java (OJP) and the Ocean Drilling Program (ODP) Leg 192, Site 1184 are presented in detail in Chapter 2. Detailed descriptions of analytical techniques can be found in Chapter 9.

4.1 Samples

The drilling at Site 1184 recovered an upper interval of calcareous ooze (Unit I) and more than 330 m of volcaniclastic rocks (Unit II) (Fig. 4.1). Unit II is divided into six subunits (A, B, C, D, E, and F) on the basis of changes in grain size, lithology, and sedimentary structures [243]. Common lithologies include tuﬀ, lapilli tuﬀ, and lapillistone, all of which contain occasional lithic clasts. The lithic clasts are primarily basalt, diabase, or re-worked volcaniclastic material. The near absence of large lapilli suggests that the volcaniclastic material was deposited some distance from the volcanic center [155]. For the most part, the rocks recovered from Site 1184 are moderately to highly altered, although some unaltered glass is present below Core 38R. Small oxidized red lapilli are present, especially in Subunit IIC, and wood fragments are present at the boundaries between subunits

41 200 Subunit IIA (lithic vitric tuff & lapilli tuff)

250 Subunit IIB x B1 (lithic vitric lapilli tuff to x B2 lapillistone) x x B3-B8 300 WOOD x x C1 Subunit IIC x C2-C3 (lithic vitric lapilli tuff to 350 lapillistone & trachylitic lapillistone)

WOOD

Depth (mbsf) 400 Subunit IID (lithic vitric lapilli tuff to lapillistone) x D1

450 WOOD

Subunit IIE (lithic vitric tuff & lapilli 500 tuff) x E1-E2 WOOD 525.1

Figure 4.1. Stratigraphic section of core recovered from Hole 1184A showing location of samples and of wood-bearing horizons. Modiﬁed from Mahoney et al. [155] and White et al. [251]

IIB-C, IIC-D, IID-E, and IIE-F. A thinly bedded, ﬁne-grained tuﬀ in Subunit IIC appears to be a primary ash-fall deposit [155]. These features indicate emergence of the OJP in this region. A total of 14 basaltic clasts extracted from Subunits IIB, IIC, IID, and IIE were selected for this study (Fig. 4.1). In hand sample the clasts were non-vesicular to sparsely vesicular and aphyric to sparsely plagioclase-phyric basalts. The clasts were generally small (1–4 cm) and were a grayish color. These samples are re- ferred to as Clasts B1–E2, with the letter referring to the subunit and the number referring to the position within the subunit. For example, Clast B2 is higher the core than Clast B7 (See Tables 4.1 and 4.2)

42 4.2 Results

4.2.1 Petrography

The generally small size of the clasts (occasionally < 1 cm after grinding) meant that thin sections could only be made for nine of the 14 samples, but at least one thin section was made from each Site 1184 subunit. Plagioclase and clinopryoxene are the most common groundmass phases in these samples, varying from extremely fine-grained (< 0.05 mm) to fine-grained (up to ∼ 0.3 mm) with a subophitic-intergranular texture. In one of the most altered samples (Clast C1, LOI = 10.38 wt.%), secondary zeolite is extremely abundant (identified as natrolite based on this sample having Na2O=5.73 wt.%). A trachytic texture is still evident in Clast C1 (Fig. 4.2a). Even in samples with relatively low LOI

(e.g. Clast B6, LOI = 4.88 wt.%), clinopyroxene is generally altered and replaced with by clays (Fig. 4.2b). In samples with larger grain sizes (e.g. Clast E1), clinopryoxene is generally partially altered and replaced with brown clays, but the plagioclase crystals are generally unaltered (Fig. 4.2c). In Clasts C1, B3, B8,

E1, and E2, zeolite minerals are present as vesicle ﬁll and veins in the surrounding matrix.

43 Figure 4.2. Photomicrographs of basalt clasts. (a) Group 4 clast C1 (sample 24R-2, 68–70 cm) in plane-polarized light showing clear zeolite (probably natrolite) and small plagioclase laths. (b) Group 1 clast B6 (sample 21R-6, 73–77 cm) in plane-polarized light showing clinopyroxene altered to a brown mineral; small plagioclase laths deﬁne a subtrachytic texture. (c) Group 1 clast E1 (sample 44R-2, 101–105 cm) under crossed polars. This sample has one of the lowest LOI values in the Site 1184 suite, yet clinopyroxene still shows signs of alteration. Mineral abbreviations are as follows: ZEO, zeolite; PLG, plagioclase; CPX, clinopryoxene; and AM, alteration mineral.

44 4.2.2 Major Elements

Major-element concentrations and LOI of the 14 basalt clasts analyzed by ICP-OES are presented in Table 4.1. These data have been recalculated on a volatile-free basis. The uncorrected major element totals vary from 97.0 to 106.8 wt.%. The large range of major-element totals is controlled by two samples: the consistent low totals of sample C2 (replicated multiple times on diﬀerent fusions) and the high total of sample D1, for which sample powder was extremely limited and thereby precluded multiple fusions. We suggest that these data be treated with caution. The altered nature of all Site 1184 clast samples is highlighted by the fact that the majority of samples have at least 5 wt.% LOI (the total range is 1.6–11 wt.%). Concentrations of CaO, SiO2, and MgO are negatively correlated with LOI (CaO shows the best correlation as seen in Figure 4.3a), and

K2O displays a poor positive correlation with LOI (Fig. 4.3b). In terms of SiO2 vs. total alkalis classiﬁcation diagram (Fig. 4.4a), these samples are transitional between alkalic and tholeiitic. However, as indicated in Figure 4.3, alteration has variably increased Na2O and K2O levels. Fresh glass clasts analyzed by White et al. [251] plot as low-K tholeiites. When elements considered to be immobile under low-temperature alteration conditions are used to discriminate between alkali and tholeiitic basalts, the basalt clasts plot strictly in the tholeiitic ﬁeld (Fig. 4.4b).

45 15 a)

10 CaO

0 0 2.557.5 10 12.5 10 b)

6 O 2 K 4

0 0 2.557.5 10 12.5

LOI

Figure 4.3. (a) Loss on ignition (LOI) (wt.%) vs. CaO (wt.%) and (b) LOI (wt.%) vs. K2O (wt.%) indicating that major elements can be mobilized and either lost (CaO) or gained (K2O) during sea-water alteration of basalts.

46 15 a) Site 1184 Basalt Clasts Phonolite Site 1184 Glass Clasts

Tephriphonolite Trachyte

10 Alkalic Phonotephrite Trachyandesite O 2 Tholeiitic Rhyolite

+ K Basaltic Tephrite

O Trachyandesite 2

Na Hawaiite 5 Andesite Dacite

Basaltic Basalt Andesite 0 40 50 60 70 80

SiO2

1.25 b)

Alkalic 1

Tholeiitic 0.75 Y Nb/ 0.5

0.25

0 0 0.05 0.1 0.15 0.2 0.25 4 Zr/(P2O5*10 )

Figure 4.4. Classiﬁcation of the Site 1184 basalt clasts. (a) SiO2 (wt.%) vs. Na2O+K2O (wt.%), where the dashed line separates Hawaiian tholeiitic and alkalic lavas [153]. Data for the Site 1184 glass clasts are from White et al. [251]. Rock type ﬁelds are from Le Bas et al. [145]. 4 (b) Nb/Y vs. Zr/(P2O5 × 10 ) after Winchester & Floyd [255]. The Site 1184 clasts plot as transitional between alkalic and tholeiitic on the SiO2 vs. total alkalis plot, yet strongly tholeiitic on the Nb/Y vs. 4 Zr/(P2O5 × 10 ) plot, suggesting that Na2O and K2O were increased during alteration of these samples.

47 TABLE 4.1 MAJOR ELEMENT CONCENTRATIONS (WT.%), LOSS ON IGNITION, AND CORE INTERVALS FOR ODP SITE 1184 BASALTIC CLASTS

1 2 Core, Interval Clast Group Depth SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total LOI

(cm) (mbsf)

18R-2, 102–104 B1 1 261.0 49.6 1.24 12.2 13.1 0.26 8.94 9.73 4.44 0.34 0.17 99.4 4.25 48 20R-4, 50–52 B2 2 282.9 51.6 1.01 14.5 7.81 0.14 7.09 13.3 3.90 0.61 0.08 99.2 6.22

21R-5, 40–47 B3 2 293.1 51.2 1.40 15.9 10.9 0.04 4.40 4.60 3.16 8.25 0.16 100.4 7.32

21R-5, 55–60 B4 2 293.2 47.2 1.26 18.6 18.7 0.10 10.6 0.85 1.75 0.81 0.14 99.2 7.85

21R-6, 28–30 B5 3 294.5 53.5 0.73 17.0 7.74 0.11 5.47 12.6 2.46 0.33 0.08 98.6 3.96

21R-6, 73–77 B6 1 294.9 49.9 1.02 14.6 11.4 0.30 8.24 9.93 4.06 0.56 0.03 100.0 4.88

22R-1, 24–27 B7 3 297.5 50.8 1.02 15.3 10.3 0.28 8.11 8.63 4.46 1.05 0.03 103.1 6.87

22R-4, 120–123 B8 2 301.8 52.4 1.01 13.8 7.16 0.16 7.75 13.4 3.37 0.75 0.21 97.3 4.01

24R-2, 68–70 C1 4 318.0 48.5 1.05 17.1 12.7 0.28 7.07 4.83 5.73 2.00 0.67 103.8 10.38

25R-5, 16–18 C2 4 330.9 49.9 2.12 13.0 18.0 0.39 6.44 5.26 3.38 1.10 0.37 97.0 7.22 TABLE 4.1 (CONTINUED)

1 2 Core, Interval Clast Group Depth SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total LOI

(cm) (mbsf)

25R-6, 115–117 C3 4 333.1 50.1 1.97 15.2 15.0 0.39 5.11 6.49 4.53 0.91 0.33 99.1 11.83

35R-8, 32–35 D1 2 432.3 49.5 1.18 12.8 13.0 0.18 7.60 9.88 5.222 0.56 0.07 106.8 4.87

44R-2, 101–105 E1 1 511.7 49.3 0.95 13.4 11.1 0.21 7.90 13.9 2.95 0.26 0.04 100.0 3.17 49

44R-2, 126–132 E2 1 512.4 48.9 0.97 14.0 11.5 0.19 7.89 13.3 2.99 0.27 0.03 100.3 1.68

BHVO-2 average 50.0 2.71 13.7 12.3 0.18 7.24 11.4 2.21 0.52 0.29

Recommended values3 49.9 2.73 13.5 12.3 0.13 7.23 11.4 2.22 0.52 0.27

S.D. (1σ) (n = 7) 1.5 0.07 0.5 0.2 0.01 0.09 0.3 0.12 0.07 0.04

BIR-1 average (n = 2) 50.2 1.02 16.2 12.1 0.22 9.89 14.3 2.34 bd 0.07

Recommended values4 47.96 0.96 15.5 11.3 0.18 9.70 13.3 1.82 0.03 0.021 TABLE 4.1 (CONTINUED)

1 2 Core, Interval Clast Group Depth SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total LOI

(cm) (mbsf)

BPL-1 average (n = 2) 50.8 2.49 16.8 14.1 0.24 8.43 10.6 2.85 0.30 0.72

Recommended values5 46.1 2.30 14.8 13.6 0.18 8.12 9.7 2.31 1.11 0.49 50 1All major- and trace-element data presented have been normalized to 100% on a volatile-free basis

2Un-normalized major-element total of each clasts + LOI

3BHVO-2 recommended values from USGS preliminary certiﬁcate of analysis

4BIR-1 values from Flanagan [80] and Gladney and Roelandts [91]

5BPL-1 recommended values from Scott Hughes, Idaho State University, personal communication 4.2.3 Trace Elements

Considerable variability in trace element abundances (Table 4.2, re-calculated to a volatile-free basis) and in element/element ratios, especially for ratios of nor- mally immobile elements (e.g. Nb/La), is displayed by the Site 1184 basalt clasts. We have subdivided our basalt clasts into four groups on the basis of incompatible- element proﬁles (Fig. 4.5a-d). Group 1 comprises four clasts (B1, B6, E1, and E2; Fig. 4.5a) that have primitive-mantle-normalized proﬁles similar to those of the Kwaimbaita Formation lavas. These clasts also possess among the lowest LOI values of all clasts analyzed in this study (1.68–4.88 wt.%; Table 4.2). Group 2 clast compositions (B2, B3, B4, B8, and D1; Fig. 4.5b) show greater variability, although the heavy REE (HREE) have similar normalized abundances to those of lavas from the Kwaimbaita Formation. These clasts exhibit higher LOI values

(4.01–7.85 wt.%). Clast D1 exhibits a deep Zr–Hf trough and light REE (LREE) enrichment (Fig. 4.5b). This could reflect the incomplete dissolution of a refractory oxide or other phase, possibly zircon, during sample preparation; however, the HREE are not depleted relative to the Kwaimbaita Formation lava compositions (Fig. 4.5b). The extremely limited amount of basalt clast from this sample precluded multiple dissolutions (i.e. flux fusion), so the veracity of this anomaly cannot be assessed and, thus, must be interpreted with caution. However, in other aspects the sample is similar to the Kwaimbaita Formation basalts (i.e. HREE, Nb, and Ta abundances). Group 3 clasts (B5 and B7; Fig. 4.5c) have normalized profiles similar to those of Kroenke-type basalts [78], although in Clast B5 there are depletions at Nb–Ta and Zr–Hf relative to the REE. These profiles have been duplicated by separate dissolution, including the sodium peroxide fusion method [148]. Clasts B5 and B7 also contain the highest Ni abundances of the 14 clasts

51 analyzed in this study (222–296 ppm; the range of Ni abundance reported from the Kroenke-type basalt is 170–230 ppm [78]). Clast B5 also has high TiO2 abundances comparable to those of Kroenke-type basalts (TiO2 =0.73 wt.% vs. 0.733 wt.% in average Kroenke-type basalt), but has lower overall Zr and Nb contents (24.2 and 1.36 ppm, respectively, vs. 40.6 and 2.18 ppm in average Kroenke- type basalt). Clast B7 contains higher Zr (56.8 ppm), Nb (2.95 ppm), TiO2 (1.02 wt.%) abundances, and appears to be intermediate between the Kroenke-type and Kwaimbaita-type compositions. Group 4 clast compositions (Clasts C1, C2, and C3) have more enriched incompatible-element concentrations (speciﬁcally, Nb, Ta, Th, and the HREE) than average Wairahito basalts (the most evolved rock type so far found on the OJP) described by Birkhold [24] from the island of Makira (Fig. 4.5d). Clasts C1 and C3 exhibit negative Sr and Eu anomalies that are consistent with plagioclase fractionation. In addition, the trace-element composition of the volcaniclastic matrix within which the three Group 4 clasts were contained also shows increased levels of Nb, Ta, and Th over the REE [78]. However, the Nb, Ta, Th, and REE abundances of the matrix are not as enriched as in the Group 4 clasts. Clasts B2, B5, B8, E1, and E2 all exhibit a positive Y anomaly of varying magnitude, which has been replicated twice using separate dissolutions, including sodium peroxide fusion. Such an anomaly is not seen in lavas typical of the Kwaimbaita-type lavas.

52 TABLE 4.2 TRACE ELEMENT CONCENTRATIONS (PPM) OF LEG 192, SITE 1184 BASALTIC CLASTS

Core 18R-2 20R-4 21R-5 21R-5 21R-6 21R-6 22R-1 22R-4 24R-2 25R-5 25R-6 35R-8 44R-2 44R-2 24R-2 25R-5 25R-6 BHVO-2 S.D. Recommended

Interval (cm) 102–104 50–52 40–47 55–60 28–30 73–77 24–27 120–123 68–70 16–18 115–117 32–35 101–105 126–132 68–70 16–18 115–117 n = 7

2 Clast B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 D1 E1 E2 MC1 MC2 MC3 53 Group 1 2 2 2 3 1 3 2 4 4 4 2 1 1

Li 9.09 3.69 7.06 10.5 bd 3.74 1.28 2.97 8.74 5.51 6.29 2.59 5.71 2.49 8.78 9.38 10.8 4.7 0.7

Be 0.16 0.15 0.08 0.11 0.23 0.26 1.16 0.12 0.20 0.23 0.28 0.31 0.20 0.25 0.49 0.52 0.54 1.05 0.1

Sc 66.6 39.3 41.9 56.1 74.6 52.8 221.9 48.0 17.7 37.8 42.8 48.9 46.2 45.4 49.7 51.8 53.6 33.1 3.0

V 568.4 192.7 522.5 771.1 1259 375.9 1785 410.3 179.8 183.2 192.6 482.1 296.6 285.5 217.0 235.8 238.5 336.3 14.9

Cr 646.8 298.5 177.7 243.5 337.3 378.1 946.4 210.8 13.7 15.4 33.4 273.9 325.6 362.7 177.1 194.8 192.6 284.7 48.8

Co 55.5 25.4 97.5 45.0 58.7 51.2 181.0 32.9 24.7 38.8 38.3 57.3 51.0 54.3 50.3 50.9 54.2 52.3 4.3

Ni 107.2 71.2 88.9 88.5 234.3 89.4 307.7 70.9 14.3 12.3 17.5 110.5 111.9 116.5 68.3 75.6 74.1 127.6 7.9

Cu 74.7 9.24 406.8 27949 232.9 169.5 804.6 51.8 49.3 61.9 69.2 49.7 153.1 162.0 138.1 144.9 146.4 139.0 2.2 TABLE 4.2 (CONTINUED)

Core 18R-2 20R-4 21R-5 21R-5 21R-6 21R-6 22R-1 22R-4 24R-2 25R-5 25R-6 35R-8 44R-2 44R-2 24R-2 25R-5 25R-6 BHVO-2 S.D. Recommended

Interval (cm) 102–104 50–52 40–47 55–60 28–30 73–77 24–27 120–123 68–70 16–18 115–117 32–35 101–105 126–132 68–70 16–18 115–117 n = 7

2 Clast B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 D1 E1 E2 MC1 MC2 MC3

Zn 211.3 24.0 17.9 45.6 23.3 139.5 632.2 43.3 123.4 192.0 198.7 73.2 140.8 93.6 151.3 148.1 167.0 111.0 5.4

Ga 18.8 16.2 15.1 26.9 23.1 20.4 81.1 16.6 32.1 24.7 18.9 19.2 17.9 18.8 35.1 24.0 34.4 23.3 1.4 54

Rb 4.57 2.03 16.5 12.9 1.76 1.85 5.06 2.40 5.50 12.0 7.87 0.21 0.33 10.8 9.36 11.2 9.7 1.5 9.8

Sr 82.7 105.1 49.9 31.7 369.1 96.0 92.8 123.4 148.7 89.8 222.7 88.2 102.0 96.3 145.5 117.7 189.2 399.0 22.8

Y 21.8 44.3 30.7 28.0 21.0 22.3 16.1 55.5 45.5 56.6 54.3 23.5 30.5 46.5 39.3 38.0 36.7 26.0 1.9

Zr 66.3 42.9 57.9 85.5 24.2 56.8 56.8 59.9 212.2 181.1 187.6 28.1 52.7 53.7 90.1 89.2 86.1 172.1 15.4

Nb 3.50 2.24 3.66 4.00 1.36 3.21 2.95 6.56 17.5 15.1 14.7 3.51 2.60 2.96 8.78 7.40 8.41 19.8 1.0

Ba 7.65 28.9 157.7 54.1 14.2 26.4 39.7 40.7 19.1 46.5 47.5 8.17 9.29 9.55 32.1 36.4 127.7 138.9 3.3

La 3.22 1.67 1.98 8.4 1.92 3.12 2.89 3.83 7.59 7.11 7.39 4.01 2.78 2.85 7.35 7.88 7.40 16.3 0.4

Ce 9.03 6.88 5.25 17.5 5.96 9.12 7.48 12.1 24.2 21.1 21.2 13.6 7.82 8.37 18.3 18.4 17.7 40.4 1.7

Pr 1.51 1.30 0.86 2.3 1.05 1.45 1.14 2.1 3.68 3.38 3.28 2.07 1.29 1.34 2.70 2.70 2.59 5.9 0.1 TABLE 4.2 (CONTINUED)

Core 18R-2 20R-4 21R-5 21R-5 21R-6 21R-6 22R-1 22R-4 24R-2 25R-5 25R-6 35R-8 44R-2 44R-2 24R-2 25R-5 25R-6 BHVO-2 S.D. Recommended

Interval (cm) 102–104 50–52 40–47 55–60 28–30 73–77 24–27 120–123 68–70 16–18 115–117 32–35 101–105 126–132 68–70 16–18 115–117 n = 7

2 Clast B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 D1 E1 E2 MC1 MC2 MC3

Nd 7.16 7.09 4.21 10.6 5.05 6.91 5.75 10.1 16.8 16.3 15.3 8.98 6.46 6.30 11.7 11.6 11.1 25.1 0.5

Sm 2.66 2.70 1.76 3.22 1.67 2.50 1.90 3.08 5.87 5.72 5.56 2.92 2.15 2.42 3.55 3.46 3.37 6.6 0.5 55

Eu 0.80 0.81 0.67 0.74 0.59 0.87 0.69 1.01 1.54 1.59 1.55 1.08 0.82 0.84 1.39 1.29 1.28 2.19 0.1

Gd 3.57 3.63 3.23 4.13 2.46 3.32 2.69 4.32 7.41 8.27 7.67 3.63 3.09 3.24 5.24 4.98 4.79 6.75 0.1

Tb 0.62 0.63 0.67 0.72 0.42 0.61 0.45 0.72 1.31 1.48 1.42 0.64 0.55 0.56 0.88 0.89 0.83 1.01 0.05

Dy 4.13 4.17 4.71 4.83 2.54 4.21 2.97 4.49 8.44 10.3 9.64 4.27 3.59 3.84 6.17 6.01 5.63 5.68 0.26

Ho 0.85 0.88 1.10 1.05 0.55 0.92 0.59 1.00 1.77 2.21 2.14 0.88 0.80 0.80 1.38 1.34 1.27 1.04 0.04

Er 2.45 2.58 3.45 3.31 1.65 2.66 1.69 2.92 4.93 6.81 6.44 2.50 2.28 2.36 4.22 4.11 3.87 2.77 0.12

Tm 0.37 0.38 0.56 0.50 0.25 0.39 0.25 0.43 0.75 0.99 1.00 0.37 0.35 0.36 0.64 0.63 0.58 0.38 0.03

Yb 2.57 2.55 3.44 3.38 1.40 2.62 1.62 2.65 5.29 7.53 6.74 2.25 2.25 2.45 4.06 4.04 3.80 2.15 0.1

Lu 0.33 0.34 0.53 0.51 0.22 0.39 0.21 0.40 0.75 1.06 0.97 0.31 0.32 0.33 0.60 0.59 0.29 0.02 0.28 TABLE 4.2 (CONTINUED)

Core 18R-2 20R-4 21R-5 21R-5 21R-6 21R-6 22R-1 22R-4 24R-2 25R-5 25R-6 35R-8 44R-2 44R-2 24R-2 25R-5 25R-6 BHVO-2 S.D. Recommended

Interval (cm) 102–104 50–52 40–47 55–60 28–30 73–77 24–27 120–123 68–70 16–18 115–117 32–35 101–105 126–132 68–70 16–18 115–117 n = 7

2 Clast B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 D1 E1 E2 MC1 MC2 MC3

Hf 1.79 1.54 1.80 2.45 0.75 1.65 1.56 1.36 5.31 5.01 4.65 0.99 1.55 1.51 2.41 2.39 2.25 4.48 0.12

Ta 0.21 0.25 0.25 0.23 0.08 0.21 0.16 0.20 1.13 0.89 0.91 0.22 0.16 0.18 0.46 0.42 0.42 1.30 0.1 56

Pb 0.38 0.14 0.06 0.04 bd bd bd bd 0.64 0.32 0.31 0.06 0.08 0.05 0.76 0.90 0.71 1.54 0.07

Th 0.27 0.22 0.27 0.33 0.07 0.24 0.14 0.60 1.78 1.61 1.50 0.18 0.22 0.24 0.65 0.60 0.57 1.30 0.1

U 0.05 0.01 0.09 0.13 0.05 0.08 0.04 0.01 0.06 0.11 0.34 0.10 0.09 0.05 0.11 0.12 0.13 0.47 0.04

1BHVO-2 recommended values from USGS preliminary certiﬁcate of analysis [254]. Recommended values in italics are for BHVO-1 [92]

2Concentrations of samples MC1–MC3 are of the volcaniclastic matrix surrounding clasts C1–C3, respectively. Clast names are composed of the subunit from which the clast was extracted and relative position within the core (see section 4.1) 100 a) Avg Leg 192 primitive basalts Avg Kwaimbaita

Avg Wairahito Avg Singgalo 10

1 Clast B1 Clast E1 Clast B6 Clast E2 0.1 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 100 b)

Clast B2 Clast B8 1 Clast B3 Clast D1

Clast B4 0.1 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 100 c)

Samples/Primitive Mantle Samples/Primitive 1 Clast B5 Clast B7 0.1 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 100 d) Clast C1 Clast C2 Clast C3

1 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu

Figure 4.5. Primitive mantle normalized incompatible element plots of (a) Group 1—Clasts B1, B6, E1, and E2; (b) Group 2—Clasts B2, B3, B4, B8, and D1; (c) Group 3—Clasts B5 and B7; and (d) Group 4—Clasts C1, C2, and C3. The gray profiles represent Kwaimbaita-type basalt (circle), Singgalo-type basalt (triangle), Kroenke-type basalt (square) and the Wairahito Formation basalts (diamond). Group 1 clasts show similar profiles to Kwaimbaita-type basalt, Group 3 clasts show affinity to Kroenke-type basalt and Group 4 clasts are distinct from the main basalt types. Group 2 clasts exhibit more compositional variability, which is probably due to more intense alteration, although they have HREE profiles similar to Kwaimbaita-type basalt. See text for details. Data from which average OJP compositions were taken are: [24, 155, 172, 241, 242].

57 Fractionation of the LREE from high ﬁeld strength elements (HFSE), such as

Nb, is seen in Clasts B2, B3, B4, B8, C1, C2 and C3 (Fig. 4.5b, d). Clast B8 is Kwaimbaita-like over most of the proﬁle and in certain incompatible-element ratios (La/Lu = 9.6 vs. 10.5 for average Kwaimbaita-type basalt, La/Ce = 0.32 vs. 0.34), yet has a (Nb/La)PM ratio of 1.7 (vs. 1.09 for average Kwaimbaita-type basalt). As with the positive Y anomaly, this type of fractionation if not seen in typical Kwaimbaita Formation basalts [172, 241, 242].

4.2.4 Radiogenic isotope ratios

Four samples (Clasts B4, B7, C3 and E1) were analyzed for Sr-, Nd-, and Pb- isotope ratios, and Clast B8 for Sr- and Nd-isotope ratios (Table 4.3). The ﬁve clasts analyzed represent at least one sample from each of the groups deﬁned on the basis of incompatible elements. All isotope data were age-corrected to 120 Ma [33] using high-resolution ICP-MS parent-daughter data obtained from splits of the solutions analyzed for isotopes, except for U and Th values of Clast B4, which were determined by high-resolution ICP-MS on leached powders. Age-corrected

87 86 ǫNd values (ǫNd(t)) form a narrow range (+6.0–+6.5), whereas initial Sr/ Sr

87 86 206 204 206 204 ( Sr/ SrI) and age-corrected Pb/ Pb ( Pb/ Pb(t)) exhibit relatively more variation (0.70288–0.70480 and 18.139–18.458, respectively; Table 4.3). The range of ǫNd(t) is within that for the Kwaimbaita basalts (Fig. 4.6a, b). Although the wide range of Sr- and Pb-isotope compositions overlaps with the range for the Kwaimbaita and the Singgalo basalts [242] (Fig. 4.6a, b), we interpret the range in our data to be a result of secondary alteration eﬀects (see subsection 4.3.1 below). Therefore, we shall not use the Sr- and Pb-isotopic results in the following petrogenetic interpretations.

58 TABLE 4.3 STRONTIUM, NEODYMIUM, AND LEAD ISOTOPIC DATA OF ODP SITE 1184 BASALTIC CLASTS

Core 21R-5 22R-1 22R-4 25R-6 25R-6 Dupl. 44R-2 44R-2 Dupl.

Interval (cm) 55–60 24–27 120–123 115–117 115–117 101–105 101–105

Clast B4 B7 B8 C3 C3 E1 E1

Group 2 3 2 4 4 1 1

Rb1 18.2 5.86 1.07 2.20 0.09

Sr 43.3 113.7 110.2 18.5 64.4

87Sr/86Sr 0.706850 0.703944 0.704852 0.703467 0.703654

87 86 Sr/ Sr(t) 0.704779 0.703690 0.704804 0.702881 0.703647

Sm 0.9 0.48 3.68 3.13 1.27

Nd 3.51 1.18 11.1 6.03 2.74

143Nd/144Nd 0.512924 0.513002 0.512976 0.513042 0.513012

143 144 Nd/ Nd(t) 0.512803 0.512810 0.512819 0.512797 0.512793

ǫNd(t) 6.2 6.3 6.5 6.1 6.0

Th 0.18 0.31 1.20 0.30

U 0.35 0.06 0.23 0.06

206Pb/204Pb 19.101 ± 8 18.831 ± 7 19.053 ± 3 19.037 ± 6 18.567 ± 2 18.566 ± 6

207Pb/204Pb 15.541 ± 7 15.543 ± 6 15.577 ± 2 15.556 ± 5 15.519 ± 1 15.516 ± 6

208Pb/204Pb 38.763 ± 17 38.614 ± 15 39.000 ± 6 38.941 ± 13 38.380 ± 4 38.366 ± 14

59 TABLE 4.3 (CONTINUED)

Core 21R-5 22R-1 22R-4 25R-6 25R-6 Dupl. 44R-2 44R-2 Dupl.

Interval (cm) 55–60 24–27 120–123 115–117 115–117 101–105 101–105

Clast B4 B7 B8 C3 C3 E1 E1

Group 2 3 2 4 4 1 1

206 204 Pb/ Pb(t) 18.144 18.458 18.139 18.124 18.371 18.370

207 204 Pb/ Pb(t) 15.499 15.525 15.532 15.511 15.509 15.506

208 204 Pb/ Pb(t) 38.059 38.059 37.437 37.380 38.060 38.046

1Rb, Sr, Sm, Nd, Th, U, and Pb concentrations (ppm) from analysis of leached powder. Age (t) is taken as 120 Ma. Duplicate analyses were made on different dissolutions of the same unleached powder. Analytical uncertainty for 87Sr/86Sr measurements is ±0.000018 but in-run precisions were better than ±0.000012. Sr-isotopic ratios were measured by dynamic multi-collection, fractionation-corrected to 87Sr/86Sr = 0.1194 and normalized to 87Sr/86Sr = 0.71025 for NBS 987. Analytical uncertainty for 143Nd/144Nd measurements is 0.000014 (0.3 ǫ units) but in-run precisions were better than 0.000010. Nd-isotopic ratios were measured in oxide form by dynamic multi-collection, fractionation-corrected to 146NdO/144NdO = 0.72225 (146Nd/144Nd=0.7219 and are reported relative to 143Nd/144Nd = 0.511850 for the La Jolla Standard. Pb-isotopic ratios were measured by static multi-collection and are reported relative to the values of Todt et al. [244] for NBS SRM 981; the long-term errors measured for this standard are ±0.008 for 206Pb/204Pb and 207Pb/204Pb, and ±0.024 for 208Pb/204Pb. The within- run errors shown refer to the last significant figure. Estimated uncertainties on concentrations are ∼ 1% for both Sm and Nd, and ∼ 2% on Rb, Sr, U, Th, and Pb. Total procedural blanks are negligible: < 10 picograms (pg) for Nd, < 35 pg for Sr, < 3 pg for Th, < 5 pg for U, and < 60 143 144 147 144 pg for Pb, ǫNd = 0 today corresponds to Nd/ Nd = 0.51264; for Sm/ Nd = 0.1967, 143 144 ǫNd(t) = 0 corresponds to Nd/ Nd=0.512486 at 120 Ma

60 Clast B4 12 12 Clast B4 Clast B7

MORB Clast C4 MORB Clast C4

10 Clast C1 10 Clast E1

Clast C4 1184 Whole Rock

Louisville Hotspot Clast E1 ) 8 8 (t) (t 1184 Whole-rock Nd Nd Nauru Basin ilon

Epsilon Nd Epsilon 6 6 Eps

Kwaimbaita Formation Kwaimbaita Formation 4 4

Singgalo Formation Singgalo Formation A) B) 2 2 0.702 0.7025 0.703 0.7035 0.704 0.7045 0.705 17.518 18.519

87Sr/86Sr (t) 206Pb/204Pb (t)

87 86 206 204 Figure 4.6. (a) Sr/ Sr(t) vs. ǫNd(t); (b) Pb/ Pb(t) vs. ǫNd(t). Site 1184 samples have ǫNd(t) within the range of Kwaimbaita-type basalts, yet show relatively large variations in both 87Sr/86Sr and 206Pb/204Pb. These variations are interpreted as being the result of secondary alteration that has changed the abundances of the ﬂuid-mobile elements Rb, Sr, Pb, and to a certain extent, U. Site 1184 whole-rock data are from White et al. [251]. Sources for ﬁelds: [31, 32, 40, 154, 157, 158, 241, 242].

4.3 Discussion

4.3.1 The inﬂuence of alteration

The basalt clasts from Site 1184 are lithic fragments derived from coeval or possibly pre-existing volcanic rocks during explosive phreatomagmatic eruptions [243]. Therefore, alteration eﬀects are a major concern in our study and here we evaluate these prior to any petrogenetic interpretation. The variability of the major-element composition of the clasts relative to the

OJP basalt groups is interpreted primarily to be the result of alteration. The gen-

61 erally high LOI values, coupled with petrographic observations, demonstrate the moderately to highly altered nature of these clasts. In Figure 4.3a, b it is seen that

CaO is negatively correlated and K2O is positively correlated with LOI, demon- strating that elements can be depleted or enriched (respectively) during alteration. Indeed, many studies in the last several decades [99] have reported that during low-temperature submarine alteration, ocean-ridge tholeiites lose SiO2, CaO, and

MgO, and gain K2O, Fe2O3, MnO, Na2O, and P2O5. It would appear that the

Site 1184 clasts have only gained K2O (and possibly Na2O) and lost CaO, SiO2, and MgO to any signiﬁcant degree, although the Group 4 samples with > 5 wt.%

LOI have the highest P2O5 contents. In terms of incompatible trace elements, and Sr- and Pb-isotopic ratios, the basalt clasts from Site 1184 exhibit much more variation than basalt compositions across the OJP (Figs. 4.5 and 4.6). This is also probably a result of sea-water alteration, which aﬀected the abundances of ﬂuid-mobile trace elements such as Rb,

Sr, Pb, and, to a certain extent, U [102, 164]. The effect of sea-water alteration on both trace elements and isotopes is demonstrated by the negative correlation between bulk-clast Sr concentrations and isotopic compositions (measured on leached splits of sample powder) (Fig. 4.7). The negative correlation suggests that the clasts with low Sr contents and high 87Sr/86Sr ratios have exchanged significant amounts of Sr with sea-water (8 ppm Sr, present day 87Sr/86Sr = 0.709, at 120 Ma 87Sr/86Sr =∼ 0.707 [68, 108]). The clasts with high 87Sr/86Sr have similar Sr-isotopic compositions to sea-water at 120 Ma, indicating that most of the alteration occurred during or soon after the eruptions. As noted earlier, the Pb- isotopic signature appears to have been affected by alteration. White et al. [251] observed that, in the bulk volcaniclastic rocks at Site 1184, Sr- and Pb-isotope

62 250 Clast B3 Clast B7 200 Clast C1 Clast C4 Clast E1 Seawater 150

) Seawater at 120 Ma Avg. Kwaimbaita

Sr (ppm 100 Avg. Singgalo

0 0.703 0.704 0.705 0.706 0.707 0.708 0.709 0.710

87Sr/86Sr

Figure 4.7. Sr (unleached, ppm) vs. 87Sr/86Sr. The negative correlation suggests low Sr and high 87Sr/86Sr clasts have exchanged with sea-water (8 ppm Sr and 87Sr/86Sr=0.709).

ratios have been affected by sea-water alteration. The same type of result was also seen in the Singgalo-type vitric tuff at Site 1183 [240]. A number of studies have shown that the concentrations of some trace elements, such as REE and HFSE, in igneous rocks are generally not affected during mild sea-water alteration [22, 151]. Yet, it has also been documented that the REE are relatively mobile during more extensive alteration compared to the HFSE, and that the LREE are preferentially mobilized over the HREE [12, 127, 151]. Over- all, the Site 1184 basalt clasts exhibit large variations in element ratios such as La/Nb, La/Ta, Th/Ta, Th/Nb, and Zr/Hf. Kurtz et al. [134] reported enrichments in alteration-resistant, insoluble trace elements (Zr, Nb, Hf, Ta, Th) in strongly weathered Hawaiian soils over parent lava values due to extensive mass lass of more soluble major elements during soil formation. The presence of wood

63 and oxidized horizons suggests that subaerial weathering of the volcaniclastic deposits did occur. However, Thordarson [243] has shown that each volcaniclastic subunit probably represents a single eruptive event and, as such, soil could only have formed at the top of the subunits well away from the clasts. The Group 4 clasts were taken from a highly oxidized horizon of the core and exhibit marked enrichments of Nb and Ta (as well as P2O5) and slight enrichments of Zr and Hf over the REE. Preferential mobilization of the LREE over the HREE [12, 127, 151] can explain the similarity between HREE abundances in the Group 2 clasts and those in average Kwaimbaita-type basalt (Fig. 4.5b) and the marked depletions and enrichments in the LREE and other incompatible elements. Enrichment of the LREE during alteration may also account for the Nb–Ta and Zr–Hf depletions relative to the REE in Group 2 Clast D1 (Fig. 4.5b) and Group 3 Clast B5 (Fig. 4.5c). In order to understand the nature of incompatible-element mobility during alteration, relatively immobile elements need to be identiﬁed for comparisons to be made. Kurtz et al. [134] concluded that Nb and Ta were the most stable in the development of soils on volcanic terrains, but that Th, Zr, and Hf were mobile during this process. The Site 1184 basalt clasts have (Nb/Ta)PM (primitive-mantle-normalized) ratios of 0.96 ± 0.06 (1σ) if Group 2 Clasts B2 and

B8 ((Nb/Ta)PM =0.51 and 1.90) are omitted. These two clasts exhibit quite dissimilar Nb abundances (2.24 and 6.56 ppm, respectively), and it is probable that in at least these two samples Nb was mobile. In examining the primitive-mantle- normalized patterns (Fig. 4.5) the data suggests that Ta was relatively immobile

(note the narrow range of normalized Ta values in Figure 4.5, especially in Group 2). In an attempt to examine mobility, elements of interest (i.e. HFSEs and

64 REEs) have been normalized to Ta and compared with the average compositions of the Kwaimbaita, Singgalo, Wairahito, and Kroenke-type basalts (Fig. 4.8a-c). The distribution seen in these ﬁgures is interpreted as primarily indicating mobility of La, Yb, Zr, and Hf during alteration, although (Yb/Ta)PM values may be changed somewhat during fractional crystallization. Also (La/Ta)PM ratios of most non-arc related, mantle derived magmas are relatively close to 1.0 [161, 234]. In magmas contaminated by continental crust, these values deviate from unity, but there is not evidence for such contamination in the OJP basalts [154, 157, 158].

1.5 PM 1 [Yb/Ta] 0.5

a) 0 0 0.511.522.5

[La/Ta]PM Group 1 1.5 Group 2 1.25 Group 3 1 PM Group 4

0.75 Avg Leg 192 primitive basalts [Th/Ta] 0.5 Avg Wairahito b) Avg Kwaimbaita 0.25 0.511.5 2 Avg Singgalo

[Nb/Ta]PM 1.5

1.25 PM 1 [Hf/Ta] 0.75

c) 0.5 0.25 0.5 0.75 1 1.25 1.5

[Zr/Ta]PM

Figure 4.8. (a) (La/Ta)PM vs. (Yb/Ta)PM; (b) (Nb/Ta)PM vs. (Th/Ta)PM; (c) (Zr/Ta)PM vs. (Hf/Ta)PM. The scatter in these plots away from typical OJP values is interpreted as showing the relative mobility of La, Yb, Nb, Th, Zr, and Hf during alteration.

65 Assuming Ta is immobile, the relative mobility of other elements can be estimated. In terms of both light and heavy REE mobility, Group 1 clasts have similar (La/Ta)PM ratios to the average composition of each OJP basalt group, and slightly higher (Yb/Ta)PM ratios (Fig. 4.8a). Group 2 clasts exhibit a wide range of (La/Ta)PM values indicative of LREE enrichment and depletion, whereas

(Yb/Ta)PM is more restricted. One Group 3 clast (Clast B5) is radically diﬀerent from the average OJP basalts, suggesting that it may have been enriched in REEs.

All Group 4 clasts have (La/Ta)PM and (Yb/Ta)PM values that overlap and extend to lower values than average OJP basalts. Overall, these relationships are consistent with the altered nature of the Site 1184 clasts and with the observation that the LREE are more mobile than the HREE during alteration of basaltic material [12]. The HFSEs Zr, Nb, Hf, and Ta exhibit varying degrees of enrichment and depletion relative to Ta. As noted above, Nb appears immobile, except in two

Group 2 clasts (Clasts B2 and B8), in which depletions and enrichments (respectively) relative to Ta are observed (Fig. 4.8b). All other clasts have (Nb/Ta)PM values of ∼ 1.0 and Group 1 clasts plot directly on top of the average OJP basalt compositions, indicating that Nb/Ta ratios are unaﬀected by alteration. However,

(Th/Ta)PM values for the remaining clasts are variable, with Group 2 clasts having the largest range (∼ 0.4–1.5), which suggests Th removal and addition relative to Ta. Whereas Th shows signs of mobility, it does not show nearly as much variation as La, Yb, Nb, and Hf. Group 3 clasts have relative Th depletions, whereas those from Group 4 show slight Th enrichments (Fig. 4.8b). This contrasting behavior between Groups 3 and 4 is also seen in Figure 4.8c, where the Group 3 clasts are relatively enriched in Zr and Hf while those from Group 4 are depleted. Relative

66 to the average OJP basalt compositions, Group 1 clasts appear to have a slight

Zr enrichment (Fig. 4.8c). Again, Group 2 clasts exhibit a wide range of values. The presence of positive Y anomalies in five of the Site 1184 basalt clasts is puzzling (Fig. 4.5a–c). This could be the product of a polyatomic interference (e.g. 52Cr37Cl, 70Zn19F) that may have been induced through incomplete removal of HCl during our initial sample preparation or through incomplete HF removal during dissolution. However, these anomalies were replicated through different dissolutions (HF-HNO3 and sodium peroxide fusion), so it is unlikely that an acid-induced polyatomic interference is the cause. Such positive Y anomalies are present in lavas that have undergone varying degrees of alteration [55, 136]. Past studies of element mobility during the weathering of igneous rocks have reported all of the REEs and Y as being collectively (although not uniformly) enriched or depleted in the weathered product [70, 160, 199]. Price et al. [199] attributed these weathering effects to the presence or absence of secondary phosphates or clay minerals. Duddy [70] concluded that the REE+Y were effectively immo- bilized by the development of vermiculite as a weathering product. Kuschel % Smith [136] suggested the relative enrichment of Y was due solely to secondary phosphate and presented SEM images and mineral analyses of these phases in support of this conclusion, but generation of the positive Y anomaly in some of the weathered samples was not supported by the REE and Y contents of the secondary phosphates. However, Cotten et al. [55] reported the presence of secondary rhabdophane-type REE-Y phosphates that did contain significant Y on order of

6–15 wt.% Y2O3 (47,000–120,000 ppm Y). If such secondary phosphates were developed during the alteration of the Site 1184 basalt clasts, it is evident that only a minute amount would dramatically alter the Y abundance of the whole-rock

67 composition. However, the REEs are also enriched in these secondary phases [55] and there are no concomitant REE enrichment in the clasts. In addition, there is no correlation between Y and P2O5 contents, and the samples with the highest

P2O5 abundances exhibit no positive Y anomaly. Only clasts with CaO > 10 wt.% exhibit a positive Y anomaly (represented as (Y/Ta)PM) (Fig. 4.3.1). This indicates that the carrier of Y is rich in CaO. The Y carrier could, therefore, be calcite (present in thin section of some clasts) but there is no correlation between (Y/Ta)PM and LOI or between (Y/Ta)PM and Sr. Mahoney et al. [155] reported the presence of zeolites in the Site 1184 volcaniclastic matrix. Examination of thin sections of four of the samples that exhibit the positive Y anomaly (Clasts B3, B8, E1, and E2) shows that zeolite minerals are present as vesicle ﬁll and veins both within and surrounding the clasts in the volcaniclastic matrix. If a Ca-rich zeolite (e.g. chabazite) was present in some of the basalt clasts, it could preferentially enrich the bulk composition in

Y if signiﬁcant interaction with sea-water had occurred. For example, Wheat et al. [250] noted that relative to ocean-ridge basalt, sea-water exhibits a positive Y anomaly. Clast C1 contains zeolite minerals but does not show the Y anomaly.

This sample has low CaO (4.83 wt.%) and high Na2O (5.73 wt.%), suggesting that the zeolite present in Clast C1 is natrolite, a Na-rich zeolite. Formation of a Ca-rich zeolite during alteration by seawater in the other clasts with zeolites could produce a positive Y anomaly, as it appears that Na-rich zeolites are not Y carriers. Further work needs to be conducted in order to test this tentative conclusion.

In summary, we conclude that on the basis of primitive-mantle-normalized diagrams (Fig. 4.5a-d), Ta is the most immobile trace element in the suite of

68 2.5 Group 1

Group 2

Group 3

2 Group 4 PM 1.5 Y/Ho

0.5 0 5 10 15

CaO

Figure 4.9. CaO (wt.%) vs. (Y/Ho)PM. (Y/Ho)PM is used to represent the positive Y anomaly present in ﬁve Site 1184 clasts. This anomaly is present in samples with CaO abundances over 10 wt.%. This implies that the Y carrier is a Ca-rich phase, such as chabazite, a Ca-rich zeolite. See text for full details.

14 basalt clasts studied from Site 1184. Niobium is generally immobile but two Group 2 clasts exhibit either Nb enrichment or depletion, indicating that Nb was mobilized during alteration. Zirconium and Hf also exhibit mobility and the LREE are more mobile than the HREE. Yttrium has been added to some clasts and is manifest as positive Y anomalies on Figure 4.5. We tentatively suggest that this is the result of secondary zeolite formation as a result of interaction of clasts with sea-water.

4.3.2 Petrogenetic interpretations

Although alteration has aﬀected the major- and trace-element compositions of the Site 1184 basalt clasts, their incompatible-element proﬁles can help indicate

69 similarities to other OJP basalt types (Fig. 4.5a-d). In order to highlight these similarities, Figure 4.10 presents normalized data of the Site 1184 basalt clasts relative to the OJP basalt type that they most closely resemble. Group 1 clasts show clear similarities to Kwaimbaita-type basalt [172, 242]. Depletions in Rb and Ba and variable U concentrations can be attributed to alteration [222]. The relatively low Sr abundances in these clasts relative to the average Kwaimbaita- type basalt may also be a result of greater amount of plagioclase fractionation in the Site 1184 magmas, but there is no corresponding Eu depletion (Figs. 4.5a and 4.10a). It is much more likely to be due to preferential Sr removal during alteration. The Group 2 clasts, the most variable group, are tentatively interpreted to be akin to the Kwaimbaita-type basalt on the basis of ǫNd(t) and HREE abundances. Group 2 clasts are slightly more enriched in the HREE than are the average Kwaimbaita-type basalt, and elements more incompatible than Ti in Figure 4.10b are highly variable. However, Nb and Ta exhibits little variation and values of these two elements are similar to the Ta and Nb values of the Kwaimbaita-type basalt (Group 2 Ta = 0.20–0.25 ppm vs. 0.24 ppm for the average Kwaimbaita-type basalt). The two clasts that make up Group 3 are broadly similar to that Kroenke-type basalts (Fig. 4.10c), although Clast B7 has Ti, Zr, and Nb concentrations similar to those of average Kwaimbaita-type basalt. Clast B5 exhibits the greatest variability with depletions in Nb, Ta, Zr, and Hf, and a large enrichment in Sr. Clasts B7 appears to be intermediate between Kroenke- type and Kwaimbaita-type basalt compositions. Finally, Group 4 clasts are more evolved than the Wairahito basalts (the most evolved rock type so far recorded from the OJP) of Birkhold [24], and are signiﬁcantly enriched in Th, Nb, Ta, and the HREE relative to average Wairahito basalt. This group also displays negative

70 Eu and Ti anomalies (Figs. 4.5d and 4.10d), and clasts forming this group have the highest LOI values (∼ 7.2–11.8 wt.%). These samples have similarly shaped profiles to the whole-rock analyses of those of four samples of bulk tuff from Subunit IIC [78, high-Nb group], although they are significantly more enriched than the high-Nb group. However, analysis of the volcaniclastic matrix surrounding the clasts from Subunit IIC show strong similarities to the high-Nb group (Fig. 4.11). The compositional differences between the Group 4 clasts and the high-Nb group and Group 4 matrix indicate that they clasts and matrix were not formed during the same eruptive episode, although they are probably related because they both show high Nb, Ta, and Th relative to the REE, which is unique among OJP samples. The Group 4 basalt clasts and matrix and the high-Nb group of Fitton & Godard [78] represent a magma type not seen elsewhere in the region.

71 3 Clast B1 Clast E1

Clast B6 Clast E2 2

a) Group 1 0 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 3 Clast B2 Clast B8 Clast B3 Clast D1 2 Clast B4

b) Group 2 0 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 3 Clast B5

2 Clast B7

c) Group 3 0 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu 3 Clast C1

Clast C2 2 Clast C3

Sample/Average WairahitoSample/Average primitive basalts Sample/Leg 192 Kwaimbaita Sample/Average Kwaimbaita Sample/Average d) Group 4 0 Rb Ba Th U Nb Ta La Ce Pr Sr Nd Sm Zr Hf Eu Ti Gd Tb Dy Y Ho Er Tm Yb Lu

Figure 4.10. (a) Group 1 basalt clasts normalized to average Kwaimbaita-type basalt; (b) Group 2 basalt clasts normalized to average Kwaimbaita-type basalt; (c) Group 3 basalt clasts normalized to average Kroenke-type basalt; and (d) Group 4 basalt clasts normalized to average Wairahito basalt. These proﬁles highlight the Site 1184 group aﬃnities shown in Figure 4.5. Data from which average basalt compositions were calculated are from: [24, 78, 155, 172, 241, 242].

72 100 Avg Kwaimbaita Avg Wairahito Avg high-Nb Group Clast C2 Clast C3 Clast C1 Avg Group 4 Matrix

10 Sample/Primitive Mantle

1 Rb Ba Th UNbTaLaCePrSrNdSmZrHfEuTiGdTbDyYHo Er Tm Yb Lu

Figure 4.11. Average Group 4 clast composition, average Group 4 volcaniclastic matrix composition, average high-Nb group of Fitton & Godard [78], and average Kwaimbaita-type basalt, all normalized to primitive mantle. Note the similarities in the high-Nb group to the Group 4 volcaniclastic matrix and the significant differences of that to the Group 4 clasts. However, the incompatible-element profiles are similar (enriched Nb and Ta over the LREE), although distinct differences do occur (such as the significant depletion in Eu and Ti in the Group 4 clasts). These data are interpreted to mean that the Group 4 matrix and clasts are not from the same eruptive episode, yet they are probably related, as they both represent a distinctive high-Nb magma type not seen before on the OJP.

4.3.3 Source region

Compositions of unaltered glass present in the lower sections of the Site 1184 core are similar to those of the widespread Kwaimbaita-type lavas [251]. The basalt clasts studied here define a narrow range of initial ǫNd values from +6.0 to +6.5, within the field defined by the Kwaimbaita-type basalt [154, 158, 240–242]. Combining this observation with similarities in normalized incompatible element profiles, it is probable that Groups 1–4 were derived from the Kwaimbaita-type source, and represent basalts with affinities to both the Kwaimbaita-type and the MgO-rich Kroenke-type basalts [155, 240–242]. Group 4 clasts are relatively evolved, more so than any other lava type yet recovered from the OJP, and thus

73 are unique OJP samples. However, the similarity of Nd isotopic ratios with those of the Kwaimbaita- and Kroenke-type basalts indicate derivation from an isotopically similar or identical source.

4.4 Conclusions

1. Ocean Drilling Program Leg 192 recovered the ﬁrst evidence that at least part of the eastern salient of the Ontong Java Plateau was erupted at or above sea level. Unlike the OJP basalt sequence cored at ODP Leg 130 Site 807, and the now subaerial outcrops on Malaita and Santa Isabel (Solomon

Islands), the Singgalo type of basalt ﬂows is absent.

2. Major-element, trace-element and isotope variations within the suite of basalt clasts extracted from the volcaniclastic sequence at Site 1184 show the effects of secondary alteration on basalt derived from the Kwaimbaita-type

mantle source. On the basis of incompatible-element profiles (normalized to estimated primitive mantle values), four groups of clasts are defined. Most of the clasts (Groups 1 and 2) have affinities with Kwaimbaita-type basalt, but two show some similarities to the relatively Mg-rich Kroenke-type lavas (Group 3). The three Group 4 clasts are unique among samples recovered

from the OJP in that their incompatible-element abundances are relatively enriched. Secondary alteration has mobilized elements that are immobile during moderate alteration, namely Zr, Hf, Th, the REE and, in two clasts, Nb. Of the HFSEs, only Ta appears unaﬀected by this secondary alteration.

Several samples also exhibit curious positive Y anomalies, which we tentatively interpret to be the result of secondary Ca-rich zeolite (chabazite?) development through interaction with sea-water.

74 3. Ages estimated for the Site 1184 sequence from 40Ar/39Ar measurements

[33] strongly suggest that the eastern salient formed concurrently with the high plateau and that there is not a younging trend from the high plateau to the eastern salient [132]. Evidence for emergence of a portion of the plateau during it’s formation has potentially important implications for the eﬀects

of OJP emplacement on the local and, perhaps, global environment during the Aptian, the magnitude of which much be addressed by future drilling expeditions to the OJP and other parts of the western Paciﬁc. Furthermore, the discovery that the Kwaimbaita-type source was also producing basaltic volcanism both on the main plateau and on the eastern salient has important

implications for the modeling of magma dynamics and evolution during the growth of the OJP.

75 CHAPTER 5

ONTONG JAVA PLATEAU XENOLITHS FROM THE ISLAND OF MALAITA

Background information, geological setting, and introductory material covering the Ontong Java Plateau (OJP) are presented in detail in Chapter 2. Detailed descriptions of analytical techniques can be found in Chapter 9.

5.1 Geology of Malaita and the Solomon Islands

The geology of Malaita is complex and is related to the obduction of the OJP during its collision with the Solomon Arc. This collision has folded and uplifted the southern margin of the OJP, creating several of the Solomon Islands including Malaita. The Solomon Islands form a linear double chain of islands oriented roughly

NW-SE in the south-west Paciﬁc Ocean [194] (Fig 5.1). The only subaerial outcrops of OJP basalt are found on the islands of Malaita, Santa Isabel, and San Cristobal (Makira), which were formed as a result of the collision of the OJP and the old Solomon Arc during the late Tertiary [51, 54]. The Solomon Islands are the subaerial tip of an upstanding topographic block called the Solomon Island Block [194]. The modern Solomon Islands are a complex arc system that represents not only a volcanic arc sequence, but also uplifted portions of the OJP and

76 Cretaceous and Tertiary sea ﬂoor [106, 191]. The Solomon Islands were originally divided into three distinct regions [48–50] (Fig. 5.1):

1. The Paciﬁc Province (PP in Figure 5.1) appears to be an uplifted and largely unmetamorphosed portion of the OJP and forms the basement of Malaita, Ramos, Ulawa, and north of the Kaipito-Korighole fault system on Santa

Isabel (Makira) [106]

2. The Central Province (CP in Figure 5.1) is southwest of the Paciﬁc Province and is composed of variably metamorphosed Cretaceous and Tertiary sea ﬂoor and remnants of the Old Solomon Island Arc.

3. The Volcanic Province (VP in Figure 5.1) extends southwest to the Central

Province and is an island arc sequence of volcanic and intrusive rocks and active volcanos.

Malaita is part of the tectonic structure termed the “Malaita Anticlinorium”, which has resulted in the uplift and exposure of OJP basement and has formed several of the Solomon Islands. In the original province model of Coleman et al. [49, 50, 52], Malaita is part of the Pacific Province, or the part of the Solomon Is- lands composed of largely unmetamorphosed OJP basalt. Several broad anticlines are exposed on Malaita along the Kwaimbaita and Singgalo rivers that reveal upwards of 3.5 km of basement section. Later research indicated that province model of Coleman et al. [191, 194] was overly simplistic and that the Solomon Islands are composed of five distinct terrains [191, 194] on the basis of basement age and composition. These terrains do not always overlap with the classic definition of accretionary terranes. In the Petterson model, Malaita is part of the Ontong Java Plateau Terrain (OJPT). This terrain is characterized by the basement being

77 Figure 5.1. Map of Solomon Island Chain composed of essentially unmetamorphosed OJP basalts and no arc development. The islands comprising the OJPT (Malaita, Santa Isabel, and Ulawa) have ex- ceptionally thick basement sequences with little to no inter-sheet sediment beds.

Comparison of basement sequences on Malaita with drill cores from the OJP reveals that they are essentially identical geochemically. In fact, samples recovered from the Singgalo River are identical to those of Unit A drilled from ODP Site 803 and samples from the Kwaimbaita River are identical to those of Units C- G of Site 803. Malaita and Site 803 are separated by approximately 1500 km

(Fig. 1.1). These two lava types have been termed the Singgalo (Unit A, Site 803) and Kwaimbaita (Units C-G, Site 803) formations after Tejada et al. [242]. See Chapter 2, Section 2.4 for description of the diﬀerences between these two volcanic units. The presence of these geochemically identical lava units have led to the interpretation that the OJPT was formed as an integral part of the OJP [193].

78 5.2 The Malaitan Aln¨oite

A unique intrusion of alnöite that contains deep-seated xenoliths of garnet- bearing lherzolite and discrete megacrysts [184] is present on Malaita. Intrusions of this type are rare in oceanic settings. The intrusion occurred when the OJP was situated some 2000 km east of the Solomon Arc as the OJP passed over a hotspot (other theories include low degree decompression melting during flexure of the OJP as the OJP collided with the Australian plate) resulting in minor alkalic volcanism and emplacement of the alnöite intrusion at 34 Ma [64]. Alnöites are related to continental kimberlites and are a type of intrusive alkaline volcanism that is defined by the presence of melilite, a solid solution of the akermanite-gehlenite series. The xenoliths are hosted in an alnöite breccia. The alnöite contains the mineral assemblage phlogopite-melilite-olivine-pyroxene and is commonly quite weathered and friable. The proportion of xenoliths contained in the breccia varies but is usually greater than 10%. There are two main types of xenoliths: ultrabasic nodules (lherzolites and pyroxenites) and discrete nodules (typically monominer- alic megacrysts of ilmenite, black pyroxene, and red-brown garnet) [183]. The ultrabasic nodules are typically weathered; olivine is almost completely altered in most samples, although the pyroxenes and aluminous phases remain in tact.

They are similar to upper mantle nodules found at continental localities. Garnet is present in some samples. A third and minor type of xenolith group is composed of metamorphosed basalts and gabbros, presumably extracted from the OJP itself. Samples of the alnöite volcanics were initially collected from stream beds in north-central Malaita in the early 1950’s where the unique kimberlitic mineral assemblage led to brief exploration for diamonds [169]. The alnöite outcrops (sills and pipes) are known from at least five locations scattered amongst the NW-

79 SE trending anticlines and synclines of northern Malaita (Fig. 5.2). Boulders of alnöite breccia found in beach gravel on the extreme northern tip of Malaita is suggestive of an alnöite volcanic province that extends over the whole of northern Malaita [169]. It wasn’t until the mid 1970’s that a substantial collection of Malaitan alnöite samples were collected by Dr. P.H. Nixon. These samples were described in a series of papers [182–185] that published major element and trace element analyses for the alnöite and discrete nodule suite. These works suggested that the alnöite is a primary mantle melt of 4-4.5% partial melting of a LREE enriched garnet peridotite source. Neal and Davidson (1989) [171] alternatively suggested that the source of the alnöites was a LREE depleted source. The megacryst suite is the result of fractional crystallization of the the “proto-alnöite” magma was produced by a zone refining process in a rising diapir [171].

5.3 Previous Work on Aln¨oite Xenoliths

The proportion of xenoliths contained in the alnöite varies but is generally greater than 10% [183]. Nearly all of the non-megacryst xenoliths previously studied have been lherzolitic in composition. Alteration levels of the alnöite breccia and xenoliths are very high due to small sample size, large grain size, and heavy weathering due to the tropical climate of Malaita. Olivine is commonly altered to serpentine and has been partially replaced by calcite and clay minerals [169]. Clinopyroxene is fresher than orthopyroxene, the latter typically being more serpentized. The ultramafic xenoliths have retained their original textures despite being heavily altered [169]. The peridotite xenoliths are presumed to represent upper mantle compositions that were present below the OJP at approximately 34 Ma.

80 Figure 5.2. Map of north-central Malaita showing locations of aln¨oite pipes and sills. From Neal, 1985 [169].

The megacryst xenolith suite is dominated by large and smooth red-brown garnets that can weigh up to 8.2 kg [183]. Large single crystals of clinopyroxene, orthopyroxene, and ilmenite are also present [182]. Granular aggregates of these minerals are much less common than the single crystal occurrences. Paleotemperature calculations from the ultramafic xenoliths record higher heat flow relative to continental shield areas that have similar intrusions, for example, South Africa. The evidence suggests intrusion from the base of the lithosphere (∼ 110 km). These data suggest that the alnöite was emplaced in a tectonic situation similar to kimberlite emplacement, i.e., areas of low geothermal gradient that have been stable for long periods of time [183]. This evidence points to the fact that the geothermal gradient under the Ontong Java Plateau is steeper than that of continental shield regions, but not as steep as oceanic geotherms. This

81 reﬂects the fact that the OJP crustal thickness could be as thick as 40 km and the lithospheric base at approximately 110 km [183].

5.4 The Xenolith Suite Studied for this Work

In contrast to the typical Malaitan aln¨oite xenolith suite, which consists predominantly of upper mantle lherzolites and mono-mineralic megacrysts, the suite studied in this Chapter is composed primarily of basalts and gabbros presumed to be from the OJP. Three upper mantle xenoliths are also described, but the primary focus of this chapter is to determine the relationship of the basalt and gabbro samples to the OJP.

5.5 Analytical Methods

Sample preparation, loss on ignition (LOI) analysis, major and trace element analyses, and some of the electron microprobe data presented in this chapter were performed at the University of Notre Dame. Parts of the electron microprobe data were conducted by Dr. Clive Neal at the University of Leeds. Isotopic analyses were conducted by John Mahoney at the University of Hawaii. Please see

Chapter 9 for complete descriptions of the methods used in this study.

5.6 Results

5.6.1 Petrography

In all of the following petrographic descriptions, the estimated modal percentages are generally accurate to ±5% for plagioclase and clinopyroxene and less than 5% for orthopyroxene and olivine. Modal abundances of accessory minerals

82 are generally not estimated unless they make up greater than 5% of the total mode.

5.6.1.1 BM/SI-7

BM/SI-7 is a coarse-grained gabbro (Fig. 5.3). In hand sample, plagioclase and pyroxene can be identiﬁed (generally 1-2 mm). Veins (0.5 mm) of white minerals are present, possibly carbonate, although calcite was not identiﬁed in thin section. Plagioclase makes up 50% of the modal abundance, clinopryoxene 35%, orthopyroxene 10%, and olivine < 5%. Crystal grains are generally subhe- dral to euhedral. Orthopyroxene and plagioclase are both largely unaltered, while clinopyroxene shows alteration to clay minerals along grain boundaries and fractures. Olivine is largely altered to serpentine and clay minerals, especially along fractures.

5.6.1.2 PHN 5802

There is no hand sample for PHN 5802, but thin section inspection shows that this sample is a garnet biotite clinopyroxenite, possibly of eclogitic heritage (Fig. 5.4). Garnet is present as large euhedral (up to 1 cm) crystals mantled by kelphytized rims (∼ 0.1 mm). Occasional large biotite crystals are present (up to several millimeters across) as well. The thin section is split approximately 50:50 into two distinct zones: a biotite-rich and a biotite-poor zone. Clinopyroxene is by far the most abundant mineral present with approximately 75% of the modal abundance. A single clinopyroxene crystal was identiﬁed as omphacite by electron microprobe analysis. Omphacite and pyrope rich garnet are the identifying minerals for eclogite, although the paucity of omphacite suggests that this sample is

83 CPX PLAG

Figure 5.3. Photomicrograph of BM/SI-7 under cross polars. The calibration mark is 1 mm. Relatively coarse-grained plagioclase (PLAG) and clinopyroxene (CPX) are visible in the photomicrograph.

not a true eclogite. Biotite is 15% of the modal abundance, garnet 5%, opaques

(probably ilmenite as this sample has extremely high TiO2 abundances, see Ta- ble 5.1) 3-5%, and there is minor amounts of hornblende. The biotite crystals are strongly pleochroic. There is no olivine in this sample.

5.6.1.3 PHN 3939

PHN 3939 is a coarse-grained ultramaﬁc upper mantle xenolith (Fig. 5.5). Gar- net is present as small (< 5 mm) round crystals mantled by thin kelphytized rims. The perimeters of the garnet crystals often enclose orthopyroxene crystals (see

Fig 5.5). Clinopyroxene and orthopyroxene comprise the majority of the rock, with clinopyroxene being approximately 50% and orthopyroxene approximately 30%. Biotite, garnet, and hornblende comprise the rest of the mode. Alteration

84 CPX BIO

GRT

Figure 5.4. Photomicrograph of PHN 5802 under plane polarized light. The calibration mark is 1 mm. A large euhedral garnet (GRT) crystal with a kelphytized rim is visible in the lower left hand corner. Large biotite (BIO) crystals and groundmass clinopyroxene (CPX) are visible as well.

is limited to veins of calcite and the minor alteration of clinopyroxene along fractures. Olivine is absent, similar to sample PHN 5802. Classiﬁed as a garnet biotite websterite.

5.6.1.4 PHN 5765

Sample PHN 5765 is a gabbro with granoblastic texture (Fig. 5.6). PHN 5765 is sample limited, i.e., no hand sample and only one thin section. The mineralogy of this sample appears to be quite simple; PHN 5765 is composed of approximately equal proportions of plagioclase and clinopyroxene. The plagioclase and clinopyroxene are in a ﬁne-grained (< 1 mm) granoblastic texture. Triple junction grain boundaries are common. Based on the major element chemistry (Table 5.1), the

85 BIO

GRT

OPX

CPX

Figure 5.5. Photomicrograph of PHN 3939 under crossed polars. Calibration mark is 1 mm. In polarized light, garnet (GRT) is isotropic. Small crystals of orthopryoxene (OPX) are enclosed within a garnet crystal at the top right corner. Biotite (BIO) and clinopyroxene (CPX) crystals are visible as well.

protolith was basaltic, making this sample possibly a metamorphosed OJP basalt from deep regions of the plateau.

5.6.1.5 PHN 5792

There is no hand sample available for PHN 5792. A very small (< 1 cm) chip was completely consumed in the process of creating the thin section. Parts of the sample started to ﬂake oﬀ the slide while thinning and polishing the sample. In order to preserve what was left of the slide, all polishing was stopped. Therefore the thin section is slightly thicker than 30 µm. No photomicrograph is included because of these issues. Please see the description of ST-7 for further explanation. PHN 5792 is composed of clinopyroxene and plagioclase. This sample is a gabbro.

86 CPX CPX

PLAG PLAG

PLAG

Figure 5.6. Photomicrograph of PHN 5765 under crossed polars. Calibration mark is 1 mm. Plagioclase (PLAG) crystals tend to have triple-junction grain boundaries. Clinopyroxene (CPX) crystals are visible in the upper part of the photomicrograph.

5.6.1.6 PHN 5793

Several inch-sized chips of PHN 5793 were examined. In hand sample, banding of light and dark minerals is apparent. The rock is phaneritic, although grain size is extremely small (< 0.5 mm) and crystals are barely visible to the naked eye.

Thin section examination showed this sample to be a gabbro with granoblastic texture, similar to PHN 5765 (Fig. 5.7). Plagioclase and clinopyroxene comprise the majority of the mineral phases and in roughly 35:65 proportions. Small (less than 5%) amounts of K-feldspar and orthopyroxene are present. The light and dark banding in the hand sample is evident in the thin section. The granoblastic texture and the common triple-junction grain boundaries suggest textural equilibrium. The major element chemistry (Table 5.1) of this sample is basaltic so this

87 CPX

CPX

PLAG

CPX CPX

Figure 5.7. Photomicrograph of PHN 5793 under crossed polars. Calibration mark is 1 mm. Plagioclase (PLAG) and clinopyroxene (CPX) are the only crystal phases visible in this photomicrograph.

sample could represent basalts erupted during the early stages of OJP volcanism, which was subsequently buried and metamorphosed.

5.6.1.7 PHN 5843

PHN 5843 is similar to PHN 5765 and PHN 5793 in mineralogy and texture (see Figure 5.6). There is no hand sample to examine, but the thin section shows that plagioclase and clinopyroxene are in approximately equal proportions making up about 80% of the modal mineralogy with orthopyroxene and olivine comprising the rest of the sample. Triple-junction grain boundaries are common, especially between clinopyroxene crystals. Olivine is very altered, often being completely serpentized. Fresh olivine that remains is often mantled by serpentized rims. There is some banding of light and dark minerals, but not as prominent as in

88 PHN 5793. Once again, this sample has a basaltic major element composition

(Table 5.1) and could be a OJP crustal xenolith.

5.6.1.8 PHN 5853

PHN 5853 is a biotite-bearing harzburgite. In hand sample, PHN 5853 is medium grained (0.1-0.5 mm) and the grains appear to be roughly equant. Olivine is the predominant crystal phase, comprising approximately 65% of the modal abundance (Fig. 5.8). Orthopyroxene, biotite, and hornblende make up the remaining primary mineralogy. Olivine is largely decomposed to alteration phases and only small fragments of fresh olivine are left in the centers of the largest grains, although the original grain boundaries are preserved.

BIO

OL OL OPX OPX BIO

OL BIO OPX

Figure 5.8. Photomicrograph of PHN 5853 under crossed polars. Calibration mark is 1 mm. Crystals of olivine (OL), biotite (BIO) and orthopyroxene (OPX) are present in the photomicrograph.

89 5.6.1.9 PHN 5854

PHN 5854 is a gabbro. Plagioclase is the predominantly phase, with clinopyroxene, orthopyroxene, and olivine comprising the rest of the sample in approximately 60:20:15:5 proportions (Fig. 5.9). Olivine is largely replaced by serpentine and brown clay alteration products, generally only small (< 0.2 mm) cores of fresh olivine surrounded by alteration products. Clinopyroxene crystals are often rimmed by clay mineral alteration phases. The modal abundance of olivine could be as high as 15% because it’s diﬃcult to tell if the ubiquitous brown patches of alteration minerals were originally olivine crystals. Triple-junctions and banding of light and dark minerals are poorly developed. Electron microprobe data indicated the presence of garnet, which is present as anhedral crystals with numerous inclusions. The garnet appears to be out of equilibrium.

5.6.1.10 ST-7

ST-7 is limited by not having a usable hand sample to examine except for a small chip that was used to make the thin section. The chip is so small that no real usable information can be gleaned from it other than the sample is fine-grained (grain size < 1 mm). In addition, during creation of the thin section fragments of the section started to flake off (samples PHN 5792 and ST-7 were prepared at the same time using the same technique and I believe that I failed to adequately scour the side of the glass slide that the section is attached to and, during the polishing phase, the section failed to adhere to the glass slide as well as hoped), and since the sample was so limited, polishing on the slide was ceased. Because of this, no photomicrograph is included. Therefore, the thin section thickness is thicker than desired which makes unambiguous identification of minerals difficult.

90 PLAG

PLAG

OPX CPX

PLAG

Figure 5.9. Photomicrograph of PHN 5854 under crossed polars. Calibrations mark is 1 mm. Crystals of plagioclase (PLAG), clinopryoxene (CPX), orthopyroxene (OPX) are visible). One small olivine (OL) crystal surrounded by serpentine alteration products is present in the lower right hand corner.

This sample is similar to PHN 5765, being composed of essentially clinopyroxene and plagioclase. Orthopyroxene and olivine may be present as well. The crystals are mostly anhedral and equant giving this sample a somewhat aplitic texture. Triple-junction grain boundaries are present giving the samples a granoblastic texture. ST-7 appears to be a basalt/gabbro that has undergone burial metamorphism. Major element chemistry conﬁrms that ST-7 has a basaltic composition (Table 5.1).

5.6.2 Major Element Data

Major element concentrations and LOI of the 10 xenoliths analyzed by ICP- OES are presented in Table 5.1. Ignited powders were fused using the lithium

91 metaborate fusion technique (see Chapter 9 for description) and were brought into solution using 5% v/v HNO3. Therefore, the major element total shown in Table 5.1 should total 100 wt.% before addition of LOI. The range in major elements totals vary from 97.15 wt.% to 102.30 wt.%. The majority of the samples (6 of 10) are with ±2 wt.% of 100%, which is generally considered acceptable. The other four are within ±3 wt.% of 100%. LOI values are generally greater than 1.5 wt.% with a maximum of 7.31 wt.%. Half of the samples have LOI values greater than 2 wt.% which reﬂects the overall altered nature of these samples.

MgO correlates positively with LOI while SiO2 correlates negatively (Fig. 5.10). The MgO-LOI correlation is not surprising considering that samples with high

MgO contents also contain a large proportion of olivine and clinopyroxene that has largely been altered during weathering. SiO2 has a rough negative correlation with LOI with sample PHN 5802 falling dramatically off the trend defined by the other samples. SiO2 (as well as CaO and MgO) has been reported to be lost from submarine tholeiites during low-temperature alteration [99]. The alnöite samples, in addition to subaerial weathering, would have undergone submarine alteration earlier in their history. While it’s not clear why samples PHN 5802 would have experienced that alteration to a much greater degree than the other samples, it’s a possibility.

92 TABLE 5.1 MAJOR ELEMENT CONCENTRATIONS AND LOI VALUES (ALL VALUES IN WT.%) OF MALAITAN XENOLITHS

SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total LOI

BM/SI-7 47.49 0.63 14.20 8.30 0.16 8.89 12.97 4.49 0.26 0.13 97.54 1.83

PHN 5802 40.82 6.69 11.02 14.77 0.15 14.44 8.14 0.93 1.96 0.11 99.02 2.78

PHN 3939 48.36 2.80 8.76 10.58 0.13 17.85 8.14 0.99 1.46 0.07 99.13 1.45 93

PHN 5765 46.97 0.26 17.32 5.29 0.08 12.86 13.49 2.98 0.47 0.00 99.69 4.37

PHN 5792 49.89 0.44 17.49 5.13 0.11 10.07 14.44 4.25 0.47 0.00 102.30 1.24

PHN 5793 51.18 0.44 16.37 5.74 0.08 9.77 14.45 2.24 0.05 0.30 100.61 1.65

PHN 5843 51.29 0.41 15.58 6.10 0.09 9.68 14.01 2.11 0.28 0.32 99.88 2.46

PHN 5853 45.98 0.51 14.11 5.94 0.08 18.70 10.70 1.00 0.05 0.08 97.15 7.31

PHN 5854 47.52 0.20 19.28 4.71 0.06 11.25 12.04 1.99 0.65 0.03 97.72 3.83

ST-7 49.74 0.75 12.38 8.74 0.13 9.68 14.32 2.18 0.41 0.08 98.41 1.54 TABLE 5.1 (CONTINUED)

SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total LOI

BHVO-1 average (n=8) 49.36 2.70 13.54 12.10 0.14 7.12 11.28 2.24 0.52 0.28 99.28

Recommended values1 49.94 2.71 13.8 12.23 0.17 7.23 11.4 2.26 0.52 0.273 100.53

BHVO-2 average (n=13) 49.65 2.70 13.54 12.25 0.15 7.25 11.35 2.22 0.51 0.27 99.90 94

Recommended values1 49.9 2.73 13.5 12.3 0.129 7.23 11. 4 2.22 0.52 0.27 100.2

1Certified values from USGS te eoih,poal eetn h rsneo ilmeni of presence the reflecting probably elevate are xenoliths, 3939 other PHN and 5902 PHN in concentrations Titanium n ieo aDnl n asr 13,wihams certai Na almost which [153], Katsura al and MacDonald the of above line field ing un basalt picritic contains the 5802 in plots PHN and abundances 5.11). (Fig. [153] diagram crimination 2 h ape ltlreyi h aatfil naSiO a on field basalt the in largely plot samples The rK or O eoihsie nySiO Only suite. xenolith iue51.SiO 5.10. Figure htMOcreae oiieywt O n SiO and LOI with positively correlates MgO that 2 rbt aebe lvtddrn eodr eteigpro weathering secondary during elevated been have both or O

SiO2 (wt.%) 2 n g w.)v O w.)frteMalaitan the for (wt.%) LOI vs (wt.%) MgO and 2 n g hwcreain ihLI Notice LOI. with correlations show MgO and 95 eo titanomagnetite. or te 2 2 l en hteither that means nly ai/hlitcdivid- kalic/tholeiitic negatively. sttlaklsdis- alkalis total vs sal o SiO low usually eaiet the to relative d cesses. 2 Several other samples plot above the tholeiitic/alkalic dividing line, which may also reflect increase in Na2O or K2O or both during alteration or perhaps a degree of chemical exchange with the alkalic host alnöite magma. Harker variation diagrams with MgO on the abscissa show two distinct groups with the xenolith suite. The first group (samples PHN 5802, PHN 5853, and PHN 3939) is defined by MgO abundances greater than 14 wt.%. These samples are the garnet bearing pyroxenites (PHN 5802 and PHN 3939) and the peridotite (PHN 5853) that are likely to be xenoliths of ultramafic upper mantle from below the OJP. The

T pyroxenites have low SiO2, Al2O3, and CaO abundances and high TiO2, Fe2O3 , and MnO for a given MgO value relative to the rest of the suite. Samples with

MgO less than 14 wt.% are generally gabbroic granulites and gabbros. Within this subset of samples there is some variation as well. ST-7 has much lower Al2O3 contents than the other low MgO samples, most likely due to the lower abundances

T of plagioclase in the sample and higher Fe2O3 , which is indicative of the olivine in the sample. CaO abundances are all above 12% reﬂecting the large proportion of anorthitic plagioclase and clinopyroxene in the low MgO samples.

5.6.3 Trace Element Data

Trace element abundances of the 12 xenolith samples analyzed by ICP-MS are presented in Table 5.2. There is a wide range in the incompatible trace element concentrations of the xenoliths analyzed in this study. This is not surprising considering that ultramaﬁc pyroxenites and metamorphosed basalts are represented in the suite. Concentrations of the more mobile trace elements, such as Rb, Br, and Sr, show a wide range in concentrations but are generally elevated relative to MORB. This suggests seawater alteration [102, 164] has aﬀected the xenoliths. Nb

96 r eni h aoiyo h ape;i seieti H 58 PHN in evident is it samples; the of Posi majority 5792. the PHN of in that seen to are identical nearly is it RE that the shows in Mala 211 Ce-anomaly the negative the from of 211) Examination (CRN example. peridotite an mantle sed a pelagic cite of They incorporation mantle. and & recycling Neal of in result discussed the were ing feature this of nega pronounced origins a possible shows The 5792 xeno PHN the of 5.12). proﬁles (Fig. REE anomalies The ious unlikely. is this pha so This section, thin zircon. of dissolution Figure incomplete on of Eu and indicative Sm be to relative 5.13 depletion (Fig. slight elements a show these dances in enriched relatively is 5802 PHN normalized mantle primitive 58 on PHN anomalies Ta and and 5842, Nb PHN negative 5792, (PHN samples several with suite 0.02–16. the and (0.19–15.2 range large a show abundances Ta and iue51.SiO 5.11. Figure iciiaindarm fe aDnl n asr [153] Katsura and MacDonald After diagram. discrimination

Na2O + K2O (wt.%) 2 w.)v oa lai (Na Alkalis Total vs (wt.%) SiO 97 2 (wt.%) 2 O+K ewsntosre in observed not was se mn noteupper the into iment 2 .Z n fabun- Hf and Zr ). ,rsetvl)over respectively) 8, ihsieso var- show suite lith )(wt.%) O) alr[7]a be- as [173] Taylor sample while plots, ieE anomalies Eu tive .3 hc could which 5.13, ieC anomaly. Ce tive rfieo CRN of profile E tnalnöite as itan 3 exhibiting 43) 2 BM/SI-7, 02, . Figure 5.12. Chondrite normalized REE profiles for the Malaitan Xenoliths and average Kroenke, Kwaimbaita, and Singgalo basalts

PHN 5843, PHN 5853, PHN 5854, and ST-7. This feature will be discussion in Section 7.4. The REE proﬁles are generally negative (LREE depleted), which is similar to OJP type basalt proﬁles. Only PHN 5843 shows a LREE depletion relative to the HREE. PHN 5854 is REE depleted relative the rest of the suite.

98 TABLE 5.2 TRACE ELEMENT ABUNDANCES (ALL VALUES IN PPM) OF MALAITAN XENOLITHS

BM/SI-7 PHN 5802 PHN 3939 PHN 5765 PHN 5792 PHN 5793 PHN 5843 PHN 5853 PHN 5854 ST-7 BHVO-1 BHVO-1 BHVO-2 BHVO-2

1 1 AVG (n=6) CERT AVG (n=6) CERT

Li 4.32 14.1 8.26 23.5 11.8 14.5 11.8 109 16.2 5.15 6.08 4.6 5.83 5

2 Be bdl bdl 0.23 0.30 0.23 0.18 0.12 0.59 0.07 0.26 0.77 1.1 0.84 1.1 99

Sc 62.7 22.2 34.6 47.8 81.6 71.2 56.2 44.5 27.8 75.8 40.2 31.8 40.0 32

V 322 521 510 402 519 476 437 346 149 586 423 317 464 317

Cr 903 634 2082 3614 1516 1970 1018 1187 511 321 690 289 746 280

Co 43.5 117 100 85.3 41.3 44.1 41.6 59.1 49.1 58.1 56.8 45 56.7 45

Ni 78.1 192 526 552 412 290 155 334 213 83.6 134 121 139 119

Cu 36.9 150 199 99.9 41.7 55.8 46.4 77.6 39.9 23.7 196 135 184 127

Zn 70.4 309 157 46.1 50.7 51.6 42.3 44.9 31.4 63.2 183 105 247 103

Ga 101 76.7 10.2 11.2 15.7 15.2 165 69.9 55.8 104 25.5 21 25.1 21.7

Rb 11.3 37.8 54.0 1.02 0.62 0.66 28.1 5.82 24.7 31.4 10.0 11 10.7 9.8 TABLE 5.2 (CONTINUED)

BM/SI-7 PHN 5802 PHN 3939 PHN 5765 PHN 5792 PHN 5793 PHN 5843 PHN 5853 PHN 5854 ST-7 BHVO-1 BHVO-1 BHVO-2 BHVO-2

1 1 AVG (n=6) CERT AVG (n=6) CERT

Sr 1719 128 101 119 144 149 704 1367 2163 2264 408 403 438 389

Y 16.0 14.2 15.8 8.55 17.3 12.2 12.0 10.2 32.4 24.9 26.4 27.6 26.2 26

Zr 18.2 48.5 27.4 10.4 11.4 12.0 17.3 23.8 71.8 22.3 182 179 187 172 100

Nb 4.31 15.2 5.39 4.28 0.57 1.60 0.19 3.48 0.19 4.63 20.1 19 20.0 18

Mo 0.58 0.43 0.34 0.34 0.11 0.18 1.93 2.40 1.25 2.16 1.73 1.02 5.87 1.02

Sn bdl bdl 6.51 0.19 0.17 0.11 0.52 0.78 0.81 0.56 3.15 2.1 1.76 1.9

Sb 0.05 0.30 0.05 0.07 0.08 0.07 0.09 0.11 bdl 0.06 0.18 0.16 0.34 0.16

Cs bdl 0.15 0.40 0.14 0.10 0.22 0.17 0.63 0.52 0.37 0.10 0.13 0.12 0.13

Ba 1729 1300 507 407 56.2 129 3116 1284 860 2700 141 139 142 130

La 2.14 3.33 2.77 2.85 6.81 3.85 1.16 3.27 3.13 3.40 16.6 15.8 16.6 15

Ce 4.98 8.19 6.26 6.43 4.16 7.79 2.48 6.43 5.48 6.57 40.5 39 39.6 38

Pr 0.76 1.22 0.88 0.81 1.48 0.96 0.45 0.89 0.69 0.99 5.64 5.7 5.52 5.7 TABLE 5.2 (CONTINUED)

BM/SI-7 PHN 5802 PHN 3939 PHN 5765 PHN 5792 PHN 5793 PHN 5843 PHN 5853 PHN 5854 ST-7 BHVO-1 BHVO-1 BHVO-2 BHVO-2

1 1 AVG (n=6) CERT AVG (n=6) CERT

Nd 4.02 6.48 4.35 3.50 7.45 4.42 2.50 4.18 2.22 4.71 25.2 25.2 24.6 25

Sm 1.55 2.33 1.45 1.25 2.07 1.48 1.23 1.46 0.49 1.85 6.37 6.2 6.21 6.2

Eu 1.27 1.26 0.49 0.47 0.90 0.65 1.53 1.00 0.59 1.27 2.20 2.06 2.13 2.1 101

Gd 1.82 2.42 2.17 1.91 3.12 2.25 3.53 2.42 1.12 3.35 6.55 6.4 6.65 6.3

Tb 0.45 0.53 0.30 0.27 0.45 0.35 0.33 0.31 0.14 0.50 1.03 0.96 0.98 0.9

Dy 3.02 2.98 1.76 1.81 2.93 2.33 2.37 1.92 0.89 3.12 5.57 5.2 5.46 5.2

Ho 0.63 0.58 0.39 0.36 0.61 0.49 0.53 0.42 0.22 0.66 1.05 0.99 1.02 1.04

Er 1.81 1.50 1.12 0.95 1.69 1.40 1.46 1.18 0.59 1.81 2.65 2.4 2.61 2.4

Tm 0.25 0.19 0.16 0.15 0.20 0.18 0.19 0.17 0.09 0.28 0.35 0.33 0.34 0.33

Yb 1.58 1.13 0.96 0.81 1.35 1.16 1.37 0.98 0.47 1.81 2.12 2.02 2.01 2

Lu 0.23 0.15 0.11 0.10 0.16 0.16 0.19 0.13 0.10 0.25 0.29 0.291 0.28 0.28

Hf 0.69 1.70 1.01 0.35 0.50 0.54 0.82 0.77 1.26 0.80 4.59 4.38 4.50 4.1 TABLE 5.2 (CONTINUED)

BM/SI-7 PHN 5802 PHN 3939 PHN 5765 PHN 5792 PHN 5793 PHN 5843 PHN 5853 PHN 5854 ST-7 BHVO-1 BHVO-1 BHVO-2 BHVO-2

1 1 AVG (n=6) CERT AVG (n=6) CERT

Ta 0.18 1.68 0.53 0.18 0.03 0.18 0.08 0.21 0.02 0.23 1.23 1.23 1.19 1.4

W 0.19 345 0.17 0.10 0.09 0.10 0.19 0.15 0.14 0.26 0.21 0.2 0.23 0.2

Pb 1.32 0.70 0.08 0.23 0.11 0.22 0.18 0.22 0.11 0.36 1.16 2.6 1.06 2.6 102

Bi 0.01 0.02 0.00 0.01 0.02 0.01 0.02 0.07 0.00 0.02 0.01 0.008 0.02 0.008

Th 0.27 0.34 0.31 0.52 0.16 0.97 0.01 0.35 0.00 0.40 1.25 1.08 1.21 1.2

U 0.09 0.15 0.13 0.35 0.10 0.30 0.04 0.13 0.05 0.16 0.44 0.42 0.43 0.42

1Certiﬁed values from USGS

2Below detection limit Figure 5.13. Primitive mantle normalized incompatible element proﬁles for the Malaitan Xenoliths and average Kroenke, Kwaimbaita, and Singgalo basalts

5.6.4 Sr- and Nd-Isotopic Data

Sr- and Nd-isotopic ratios were analyzed by Dr. John Mahoney at the Univer- sity of Hawaii. Whole rock Sr- and Nd-isotopic data were collected from samples

PHN 5802, PHN 5853, and ST-7. Sr- and Nd-isotopic data from garnet and clinopyroxene crystals were performed on mineral separates from PHN 5802 and PHN 3939. No whole rock isotope data were collected from PHN 3939. The Sr- and Nd-isotopic data from the xenolith suite is presented in Table 5.3.

103 TABLE 5.3 SR- AND ND-ISOTOPIC DATA (ELEMENTAL ABUNDANCES IN PPM) OF MALAITAN XENOLITHS

Sample PHN 5802 PHN 5802 PHN 5802 PHN 5802 PHN 5802 PHN 3939 PHN 3939 PHN 5853 ST-7

1 2 3 Type WR CPX CPX GRT GRT CPX GRT WR WR

Rb 20.85 10.21 2.804 0.1711 0.0262 87.00 11.70 123.5 139

Sr 124.5 106.4 69.65 74.2 2.079 0.20 0.58 0.1765 0.2054

87 86 Sr/ Sr 0.704830 0.704479 0.704293 0.705016 0.704280 0.703872 0.704690 0.704887 0.705942

87 86 4 Sr/ Sr120 0.704005 0.704005 0.704094 0.705005 0.704218 0.703803 0.704469 0.704674 0.705150

87 86 5 Sr/ Sr35 0.704589 0.704341 0.704235 0.705013 0.704262 0.703852 0.704626 0.704825 0.705711

Sm 1.791 3.088 2.447 5.975 3.561 1.946 1.545 1.300 1.779

Nd 4.856 8.901 6.711 5.346 2.978 5.82 1.621 4.454 5.236

143 144 Nd/ Nd 0.512780 0.512772 0.512771 0.512871 0.512894 0.512829 0.512908 0.512892 0.512959

143 144 4 Nd/ Nd120 0.512605 0.512607 0.512598 0.512341 0.512326 0.512670 0.512456 0.512753 0.512798

ǫNd 120 Ma 2.3 2.4 2.2 -2.8 -3.1 3.6 -0.6 5.2 6.1

143 144 5 Nd/ Nd35 0.512729 0.512724 0.512721 0.512716 0.512728 0.512783 0.512776 0.512852 0.512912

ǫNd 35 Ma 2.6 2.5 2.4 2.4 2.6 3.7 3.5 5.0 6.2

1Whole-rock

2Clinopyroxene mineral separates leached in hot HCl

3Garnet mineral separates leached in hot HCl

4Corrected to 120 Ma, the age of emplacement of the OJP

5Corrected to 35 Ma, the age of emplacement of the aln¨oite

104 The three whole rock samples do not directly overlap either of the two main isotopic fields for OJP lavas (Fig. 5.14). Sr- and Nd-isotopic data for clinopyroxene mineral separates closely overlap the whole-rock data for PHN 5802, an unsurprising fact given that PHN 5802 is almost solely composed of clinopyroxene, although this demonstrates the effectiveness of the whole-rock leaching procedure in removing secondary contamination. Clinopyroxene mineral separates from another pyroxenite, PHN 3939, plot near those from PHN 5802. The garnet mineral separates from PHN 5802 and PHN 3939 have similar negative ǫNd values relative to the rest of the data. The whole rock Sr- and Nd-isotopic data for PHN 5853, the only peridotite in the xenolith suite analyzed in this paper, plot relatively close the Singgalo field and within the Malaitan alnöite field. This sample has perhaps undergone isotopic equilibrium with the host alnöite. ST-7, interpreted as a metamorphosed gabbro, has an ǫNd value within the Kwaimbaita/Kroenke OJP field but much higher 87Sr/86Sr. The alnöite data are from Bielski-Zyskind et al. and Neal & Davidson [21, 171] and extend to slightly more enriched Sr- and Nd-isotopic compositions than the Kwaimbaita/Kroenke OJP lavas. The alnöite field overlaps the Singgalo field, but also extends to more enriched isotopic compositions.

5.6.5 Electron Microprobe Data

Electron microprobe data collected on clinopyroxene, orthopyroxene, garnet, and olivine crystals are presented in Table 5.4. Additional unpublished EMP data (collected at the University of Leeds by Professor Peter H. Nixon and were provided to me as part of this project) are plotted along with the data presented here in all of the following ﬁgures. Figure 5.15 shows pyroxene composition in terms

105 Kwaimbaita Singgalo ST-7 WR Kroenke PHN 5853 WR

PHN 3939 CPX

5802 CPX 5802 WR

PHN 3939 GRT

5802 GRT

87Sr/86Sr (t)

Figure 5.14. 87/86 vs eNd for the Malaitan xenoliths and the host alnöite. Fields for OJP basalt types (Kwaimbaita/Kroenke and Singgalo) and MORB are shown for comparison purposes. Data fields for alnöite from Bielski-Zyskind et al. and Neal & Davidson [21, 171]. Fields for MORB and OJP lava types from: [31, 32, 40, 154, 157, 158, 241, 242]

of the pyroxene ternary classification diagram. Clinopyroxene from PHN 3939 encompass a narrow compositional range in the augite field from En51Fs14Wo35 to En54Fs12Wo34, which overlaps the clinopyroxene megacryst xenolith field (not shown) (Fig. 5.15). BM/SI-7 contains both clinopyroxene and orthopyroxene. Clinopyroxene from BM/SI-7 overlaps the high-Wo section of the augite field and the diopside field with very little compositional variation. BM/SI-7 orthopyroxene crystals plot in the enstatite field. PHN 5802 is the only sample that shows a distinct range in clinopyroxene composition, ranging from approximately Wo30 to Wo45. Three pyroxene crystals from PHN 5802 plot as pigeonite and there is

106 one omphacitic (not shown on Figure 5.15) pyroxene that was analyzed. Based on the large number of clinopyroxene analyses (n = 29) on PHN 5802, omphacitic pyroxene appears to be rare. The pigeonite analyses may be the exsolution lamel- lae in augite. There is no record of the location of the EMP analyses for most of the data points from PHN 5802, so this is impossible to verify. Clinopyroxene and orthopyroxene crystals from PHN 5843 plot in narrow compositional ranges in the diopside and enstatite ﬁelds, respectively. Clinopyroxene analyses extend to slightly less Fe rich compositions in PHN 5843 than BM/SI-7, PHN 3939 and PHN 5802. PHN 5853 clinopyroxene analyses closely overlap those of BM/SI-7, also showing very little compositional variation. The orthopyroxene crystals from

PHN 5854 are the most enstatite rich of the xenolith suite orthopyroxene crystals (Fig. 5.15). Similarly, the clinopyroxene crystals from PHN 5854 are also the most wollastonite-rich and ferrosillite-poor of the all the clinopyroxene crystals. They also exhibit a narrow compositional range.

A single olivine crystal from PHN 3939 was analyzed. This crystal has a high forsterite content (Fo91), which is consistent with the petrographic interpretation that this sample is an upper mantel xenolith. EMP data from garnet crystals in PHN 5802 show a narrow compositional range. These pyrope-rich garnets directly overlap the range for garnets from eclogites [52] (Fig. 5.17).

107 TABLE 5.4 ELECTRON MICROPROBE DATA ON CLINOPYROXENE, ORTHOPYROXENE, OLIVINE, AND GARNET CRYSTALS FROM MALAITAN XENOLITHS

1 2 Sample Phase No. of Points SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 Cr2O3 Total

PHN 3939 CPX1 10 52.88 0.90 5.72 6.50 0.11 16.73 15.02 1.89 0.01 0.17 0.11 100.04

PHN 3939 CPX2 10 52.17 1.04 5.79 7.27 0.14 16.50 14.59 1.91 0.01 0.16 0.07 99.63 108

PHN 3939 CPX3 16 52.41 1.03 5.73 6.88 0.12 16.43 14.87 1.90 0.01 0.15 0.09 99.61

PHN 3939 CPX4 9 51.11 1.25 6.47 7.38 0.13 15.94 14.78 1.86 0.01 0.15 0.10 99.18

PHN 3939 CPX5 8 51.25 1.25 6.48 7.39 0.13 15.95 14.75 1.86 0.01 0.15 0.10 99.32

PHN 3939 OL1 3 41.53 0.00 0.03 8.35 0.11 49.98 0.09 0.00 0.00 0.01 0.04 100.14

PHN 3939 CPX6 10 51.87 1.21 5.97 7.34 0.12 15.83 15.05 1.97 0.01 0.14 0.07 99.57

PHN 5843 OPX1 6 55.72 0.01 1.27 15.51 0.34 28.20 0.23 0.01 0.01 0.01 0.00 101.32

PHN 5843 CPX1 3 53.12 0.53 2.81 5.35 0.17 15.62 22.71 0.53 0.02 0.25 0.12 101.32

PHN 5843 OPX2 3 55.20 0.30 1.86 15.59 0.37 27.09 0.75 0.01 0.00 0.01 0.09 101.26 TABLE 5.4 (CONTINUED)

1 2 Sample Phase No. of Points SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 Cr2O3 Total

PHN 5843 OPX3 6 55.30 0.26 1.88 16.03 0.37 27.09 0.64 0.02 0.01 0.01 0.08 101.67

PHN 5843 CPX2 10 51.82 0.91 3.48 6.36 0.19 14.98 22.35 0.62 0.00 0.25 0.15 101.13

PHN 5843 OPX4 15 54.19 0.29 1.86 16.54 0.38 26.76 0.75 0.02 0.00 0.01 0.07 100.88

PHN 5843 CPX3 4 52.53 0.92 3.46 6.32 0.19 14.92 21.90 0.64 0.01 0.22 0.14 101.26 109

PHN 5802 GRT1 5 38.30 2.16 20.85 18.20 0.35 13.93 6.59 0.14 0.00 0.13 0.00 100.67

PHN 5802 GRT2 5 38.17 2.02 22.13 17.67 0.38 13.71 6.37 0.15 0.00 0.12 0.01 100.75

PHN 5802 GRT3 2 39.16 1.84 22.31 17.42 0.30 13.85 6.31 0.13 0.00 0.13 0.00 101.44

PHN 5802 CPX1 15 47.80 2.07 11.21 10.03 0.14 12.15 13.38 2.68 0.01 0.15 0.01 99.63

PHN 5802 CPX2 7 49.91 1.21 8.51 10.80 0.19 14.18 12.45 2.39 0.01 0.15 0.01 99.79

BM/SI-7 OPX1 4 54.01 0.16 0.94 18.75 0.47 24.89 0.82 0.05 0.01 0.01 0.00 100.12

BM/SI-7 CPX1 6 52.70 0.68 1.93 9.16 0.26 14.81 20.27 0.41 0.01 0.22 0.02 100.46 TABLE 5.4 (CONTINUED)

1 2 Sample Phase No. of Points SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 Cr2O3 Total

BM/SI-7 CPX2 10 52.09 0.75 2.36 8.54 0.25 14.28 21.34 0.40 0.01 0.24 0.02 100.27

BM/SI-7 CPX3 12 52.08 0.70 2.32 9.30 0.27 15.05 19.60 0.39 0.09 0.22 0.03 100.06

1CPX–clinopyroxene, OPX–orthopyroxene, OL–olivine, GRT–garnet 110 2Number of individual spot analyses per crystal Figure 5.15. Pyroxene ternary classiﬁcation diagram. Composition is shown in terms of the three pyroxene endmembers: enstatite, wollastonite, and ferrosillite. For most samples, the clinopyroxene and orthopyroxene show little compositional variation, however, PHN 5802 shows a large range in clinopyroxene composition.

5.7 Discussion

5.7.1 Origin of REE Anomalies

The strong negative Ce anomaly present in PHN 5792 is similar to the negative Ce anomaly seen in a peridotite xenolith from the Malaitan aln¨oite studied by Neal & Taylor [173]. The Ce4+ species is unlikely to form naturally in magmatic

111 CPX

GRT

OPX

Figure 5.16. Ca-Mg-Fe* compositions of clinopyroxene, orthopyroxene, and garnet in Malaitan xenoliths. The relative ﬁelds for each phase are marked (CPX–clinopyroxene, OPX–orthopyroxene, GRT–garnet). Note the

systems due to the reaction of Ce4+ with ferrous iron to form Ce3+ and ferric iron [220]. While negative Ce anomalies should not occur due to any igneous process, there are secondary processes that can generate such an anomaly. Ce4+ is insoluble, so oxidation of Ce3+ in seawater leads to negative anomalies of Ce in seawater [65, 66, 74]. Therefore, pelagic sediments and basalts that have undergone seawater alteration will also display a negative Ce anomaly. Neal & Taylor [173] showed that incorporation of pelagic containing a negative Ce anomaly sediment into the mantle (via subduction) could be responsible for the strong negative Ce anomaly in the peridotite xenolith from Malaita. As PHN 5792 is not a mantle xenoliths (it

112 Figure 5.17. Garnet proportions in eclogites from various locations. The pyrope-rich garnets from PHN 5802 directly overlap the range of garnets in eclogites from gneisses and migmatitic terrains.

is a gabbroic basalt) it appears to have inherited the anomaly through seawater alteration. As discussed in the petrography subsection 5.6.1.5, PHN 5792 appears to be highly altered as most of the clinopyroxene crystals are embayed with alteration products on the grain boundaries and along fractures. Another possibility is that this sample is the product of melting of a source with an inherent negative Ce anomaly. If the work of Neal & Taylor [173] is correct and the peridotite xenolith with a negative Ce anomaly is the result of recycling of crustal materials, then there may be a reservoir with a negative Ce anomaly beneath the OJP capable of producing lavas with this anomaly. However, as no OJP basalt sampled to date contains a negative Ce anomaly, leads to conclusion that the negative Ce

113 anomaly in PHN 5792 is the product of seawater alteration prior to this sample being buried under the lava pile and subsequently metamorphosed. The REE profiles (Fig. 5.18) of samples BM/SI-7, PHN 5802, PHN 5843, PHN 5853, and ST-7 show general concave upward REE profiles. In samples BM/SI-7, PHN 5843, and ST-7, the apparent positive Eu anomaly is most likely from accumulation of plagioclase as petrographic inspection of these samples reveals that these samples are rich in plagioclase. However, this explanation cannot explain the REE profiles of PHN 5802 and PHN 5853. PHN 5802 is a clinopyroxenite and PHN 5853 is a biotite-bearing harzburgite. Plagioclase is absent in these samples, so the apparent positive Eu anomaly may be more a function of depleted

LREE concentrations relative to the MREE and the HREE. REE from clinopyroxene mineral separates from several Malaitan xenoliths have similar REE patterns with depleted LREE ([La/Lu]N < 1) relative to the HREE or even “kinked” REE patterns with depleted LREE except for La ([La/Lu]N ∼ 1) [169]. So the clinopyroxene rich PHN 5802 and PHN 5853 samples may have inherited this REE proﬁle from the large proportion of pyroxene in the samples. Another possibility is that the Eu anomaly is the result of metamorphism of plagioclase to pyroxene and the pyroxene inherited the positive Eu anomaly from the plagioclase [173, 174].

Sample PHN 5854 is signiﬁcantly more depleted in REE than the rest of the xenolith suite (Fig. 5.12). This sample contains grossular garnet crystals (from unpublished EMP data analyzed by Prof. Peter H. Nixon at the University of Leeds), which could be from the metamorphic reaction of anorthite-rich plagioclase and olivine to form grossular. This would explain the inherited positive Eu anomaly and, as there are examples of anorthite-rich plagioclase cumulate xenoliths, this is a simple explanation of the presence of grossular garnet.

114 Figure 5.18. Chondrite normalized REE proﬁles of samples BM/SI-7, PHN 5802, PHN 5843, PHN 5853, and ST-7.

5.7.2 Thermobarometry

Electron microprobe data on clinopyroxene and orthopyroxene crystals can be used to determine crystallization temperature and pressure. Using the single clinopyroxene thermometer [176, 177], the crystallization temperature was estimated. When available, the two-pyroxene enstatite-in-clinopyroxene thermometer was used to check the results of the single clinopyroxene method. The two- pyroxene method generally produced temperatures approximately 100 ◦C lower than the single clinopyroxene method. Pressure determinations were made by utilizing the CpxBar excel spreadsheet provided by P. Nimis using the clinopyroxene geobarometers presented in Nimis [175]. Temperature and pressure data for samples with clinopyroxene EMP data is presented in Table 5.5. Unfortunately, the single clinopyroxene thermometer and barometers of Nimis & Taylor [176] and

Nimis [175] are calibrated for mantle compositions. Therefore, the samples with no or partial equilibration pressures are those that are outside of the calibration range of the geobarometer.

115 TABLE 5.5 TEMPERATURE ( ◦C) AND PRESSURE (KBAR) OF CLINOPYROXENE FROM MALAITA XENOLITHS

Sample Phase T–low1 T–high2 P–low1 P–high2

PHN 3939 CPX1 1202 1474 14.9 4.9

PHN 3939 CPX2 1192 1439 15.0 6.1

PHN 3939 CPX3 1196 1454 14.8 5.4

PHN 3939 CPX4 1188 1425 15.3 6.7

PHN 3939 CPX5 1188 1425 15.4 6.9

PHN 3939 CPX6 1188 1425 14.8 6.2

PHN 5843 CPX1 1212 1508

PHN 5843 CPX2 1195 1450

PHN 5843 CPX3 1196 1451

PHN 5802 CPX1 1132 1251 24.7 21.0

PHN 5802 CPX2 1141 1276 21.9 17.5

BM/SI-7 CPX1 1161 1340

BM/SI-7 CPX2 1165 1350

BM/SI-7 CPX3 1162 1340

See footnote:3

PHN 5854 CPX1 1237 1598

PHN 5854 CPX2 1241 1615

PHN 5854 CPX3 1236 1594

PHN 5854 CPX4 1239 1607 3.5

116 TABLE 5.5 (CONTINUED)

Sample Phase T–low1 T–high2 P–low1 P–high2

PHN 5854 CPX5 1240 1612 0.6

PHN 5854 CPX6 1240 1610

PHN 5854 CPX7 1237 1599

PHN 5854 CPX8 1242 1619

PHN 5854 CPX9,RIM 1242 1618

PHN 5854 CPX10 1240 1612

PHN 5854 CPX11,CORE 1236 1597

PHN 5854 CPX12,CENTER 1237 1600

PHN 5854 CPX12,MIDDLE RING 1241 1616

PHN 5854 CPX12,OUTER RING 1232 1581 6.0

PHN 5802 CPX1 1129 1243 21.8 18.1

PHN 5802 CPX2 1136 1262 10.7 6.2

PHN 5802 CPX3 1144 1285 19.3 14.4

PHN 5802 CPX4 1142 1280 20.3 15.6

PHN 5802 CPX5 1147 1295 21.3 16.3

PHN 5802 CPX6 1143 1283 20.4 15.7

PHN 5802 CPX7 1147 1295 20.8 15.7

PHN 5802 CPX8 1128 1240 22.6 19.0

PHN 5802 CPX9 1129 1242 22.1 18.4

PHN 5802 CPX10 1133 1253 22.1 18.2

PHN 5802 CPX11 1134 1256 23.1 19.2

117 TABLE 5.5 (CONTINUED)

Sample Phase T–low1 T–high2 P–low1 P–high2

PHN 5802 CPX12 1138 1269 19.8 15.4

PHN 5802 CPX13 1128 1239 22.6 19.0

PHN 5802 CPX14 1129 1242 23.4 19.7

PHN 5802 CPX15 1129 1242 23.0 19.3

PHN 5802 CPX16 1146 1292 9.6 4.3

PHN 5802 CPX17 1121 1220 23.6 20.5

PHN 5802 CPX18 1141 1276 19.7 15.1

PHN 5802 CPX19 1139 1271 23.2 18.9

PHN 5802 CPX20 1130 1243 23.5 19.8

PHN 5802 CPX21 1129 1243 22.8 19.1

PHN 5802 CPX22 1132 1251 21.8 17.9

PHN 5802 CPX23 1141 1278 20.3 15.7

PHN 5802 CPX24 1134 1256 20.6 16.6

PHN 5802 CPX25 1147 1294 17.5 12.4

PHN 5802 CPX26 1142 1279 22.2 17.8

PHN 5802 CPX27 1140 1275 22.7 18.3

PHN 5853 CPX1,CORE 1159 1331 12.7 6.5

PHN 5853 CPX1,RIM 1170 1367 8.2 0.7

PHN 5853 CPX2,CORE 1149 1301 14.1 8.6

PHN 5853 CPX2,RIM 1160 1336 12.1 5.7

PHN 5853 CPX3,CORE 1155 1318 11.3 5.3

118 TABLE 5.5 (CONTINUED)

Sample Phase T–low1 T–high2 P–low1 P–high2

PHN 5853 CPX3,RIM 1153 1315 10.8 4.8

PHN 5853 CPX4,RIM 1166 1353 9.7 2.6

PHN 5853 CPX5,CORE 1166 1355 10.7 3.7

PHN 5853 CPX5,RIM 1172 1372 9.9 2.4

1Lower limit temperature estimation. Pressure calculated using the low temperature

2Upper limit temperature estimation. Pressure calculated using the high temperature

3These data were collected at the University of Leeds by Prof. Peter H. Nixon at the Uni- versity of Leeds and are unpublished

These data show that PHN 5802, PHN 3939, and PHN 5853 equilibrated at fairly high pressures: 20-24 kbar (2–2.4 GPa), 8-14 kbar (0.8–1.4 GPa), and 8-13

(0.8–1.3 GPa) kbar respectively. The other samples are gabbros that appear to have experienced burial metamorphism and equilibrated at lower pressures. Pressure increases with depth in the earth at a rate of approximately 1 kbar per 3.3 km of depth. Therefore, crystallization pressures of 25 kbar equates to approximately 82.5 km depth. The OJP is approximately 33 km thick, but could be as thick as 38 km under the high plateau [211]. Estimated crystallization depths of 66-79.2 km for PHN 5802 reinforces the previous interpretation based on the upper mantle like mineralogy that this sample is not part of the OJP and is an upper mantle xenolith. On the other hand, PHN 3939 crystallized between

46.2 km and 26.4 km depth. The 46 km depth is also below the lowermost limits of the OJP. PHN 3939 is a garnet-bearing websterite which suggests relatively

119 deep crystallization pressures and upper mantle conditions. The work of Ishikawa et al. [116] on the the nature of the mantle beneath of the OJP suggests that garnet websterites are present in vein-like structures beneath the OJP from approximately 70 km depth to less than 50 km, which is roughly consistent with our pressure estimation. Like PHN 5802, PHN 3939 is not considered to represent the metamorphosed lower portion of the OJP. PHN 5853 crystallized between 42.9 and 26.4 km, which is at slightly lower pressure than PHN 3939. PHN 5853 is a harzburgite. The sub-OJP mantle model proposed by Ishikawa et al. [116] suggests two locations for harzburgite; one at ∼ 90 km and at 30–50 km. Our pressure estimation and calculated equilibration depth is consistent with origination from this upper harzburgite zone. The lower harzburgite zone is predicted to be the result of high degrees of melting and therefore, should be strongly depleted in incompatible and REE elements. However, PHN 5853 is not depleted, which further strengthens the interpretation that it is from the upper harzburgite zone.

Calibration of the geobarometer is not valid at crustal pressure, so the absolute values of the equilibrium crystallization pressure for samples PHN 5843, PHN 5854, and BM/SI-7 are of little consequence other than providing further evidence that these predominantly gabbroic samples crystallized in upper level magma chambers of the OJP. The barometer used is not precise enough to give relative locations within the OJP for these samples.

5.7.3 Relationship of the Basalt/Gabbro Samples to the OJP

The crystallization temperature and estimations presented in sub section 5.7.2 suggest that samples PHN 5843, PHN 5854, and BM/SI-7 crystallized in upper level magmas chambers with the OJP lava pile under relatively low pressure

120 conditions. Two groupings were made on the basis of hand and thin section examination of the basalt/gabbro samples. The basalt/gabbro group is composed of PHN 5854, PHN 5792, and BM/SI-7 and the gabbro group with granoblastic textures is composed of PHN 5765, PHN 5793, PHN 5843, and ST-7. The granulites are interpreted to be basalts or gabbros that underwent varying degrees of metamorphism resulting in a granoblastic texture. While the exact metamorphic mechanism is equivocal, a reasonable interpretation would be a combination of burial metamorphism due to the immense size and thickness of the OJP and high temperatures from the large amount of heat required to achieve the upwards of 30% partial melting associated with the formation of the OJP [35, 78]. The greater degree of formation of the triple-junction grain boundaries in the granulite samples would lead to the conclusion that the granulite samples are from deeper in the OJP than the apparently non-metamorphosed basalts. In the scope of this paper, the exact metamorphic conditions are unimportant, but the relative location in the stratigraphy of the OJP of the gabbros and granulites is of consequence to enhance our understanding of the structure and geochemistry of deep regions of the OJP. On the basis of the pressure estimations presented in Table 5.5 and the petrographic examinations in sub section 5.6.1, it is certain that a number of these xenoliths represent portions of the OJP that are much deeper than has yet been sampled by ocean drilling or ﬁeld work. In order to determine the relationship of these samples to the OJP, the incompatible trace elements and REE were normalized to Kroenke and Kwaimbaita type basalts (Fig. 5.19). By doing so, it’s possible to identify if these samples have similar trace element compositions to the normalized values.

121 lc ieidct opstoa ieecsfo h norm the from diﬀerences compositional indicate line black h noptbetaeeeet n E oecoeymtht match closely more REE and elements trace incompatible The wibiabsl.Tetptopos( n )aenormalized are B) and (A plots two top The basalt. Kwaimbaita omlzdt vrg wibiabsl.Tedvain fr deviations The basalt. Kwaimbaita average to normalized iue51.Cutlsmlsnraie oaeaeKroenke average to normalized samples Crustal 5.19. Figure vrg rek aat h otmtopos( n )are D) and (C plots two bottom The basalt. Kroenke average

Sample/Avg Kwaimbaita Basalt rek aat hnKamat basalts. Kwaimbaita than basalts Kroenke 122 D B C A lzdvalues. alized mthe om oeof hose and to Figure 5.19 indicates these samples have a greater affinity to Kroenke-type basalts than Kwaimbaita-type. Samples BM/SI-7 and ST-7 normalized REE profiles (Fig. 5.19) most resemble Kroenke-type basalts. Additionally, both samples plot inside or near the field for Kroenke type basalts on a TiO2 vs. MgO plot

(Fig. 5.20). Isotopically, ST-7 has a similar ǫNd (corrected to 120 Ma) as Kwaim- baita and Kroenke basalts (Fig. 5.14). An increase in the 87Sr/86Sr ratio of seawater altered OJP basalts is commonly seen [223] due to the high 87Sr/86Sr ratio of seawater, so the relatively radiogenic Sr isotopic composition of ST-7 is likely the result of seawater alteration products that were not completely removed by leaching. Incompatible trace element and REE proﬁles of BM/SI-7 and ST-7 samples (Fig. 5.19) show that they show little deviation from Kroenke-type basalt, especially among the HREE, except for the positive Eu anomaly. Both of these samples have at least 50% plagioclase modal abundances, which can account for the Eu anomaly. These samples appear to be plagioclase-rich cumulates from fractionating Kroenke-type lavas. This explanation accounts for the coarse mineralogy and the positive Eu anomaly. Samples PHN 5765, PHN 5792, PHN 5793, and PHN 5843 are unlike, and generally depleted relative to Kroenke type lavas. Sm-Nd isotopic data for PHN

5793 is within the range for Kwaimbaita-Kroenke basalts. However, there are some key diﬀerences between these samples and the known OJP basalt types.

The xenolith samples are generally more fractionated (La/YbN between 2.38 and

3.61) than those from Kroenke-type lavas, which typically have La/YbN values of around 0.78. In fact, most of the xenoliths have REE proﬁles more fractionated than those from the Singgalo formation, which contains the most evolved lavas found on the OJP [241, 242]. Figure 5.20 shows that PHN 5765, PHN 5792, PHN

123 ST-7

BM/SI-7 (wt.%) 2 TiO

Figure 5.20. TiO2 (wt.%) vs MgO (wt.%) for crustal xenoliths. The ﬁeld for Kroenke-type basalts is show for comparison. Kroenke-type basalt data from Fitton & Godard [78]

5793, and PHN 5843 all have lower TiO2 abundances for a given MgO value. The overall MgO abundances are slightly greater for samples PHN 5792 (MgO =

10.07 wt.%) and PHN 5854 (MgO = 11.25 wt.%) and the TiO2 abundances lower than in Kroenke-type basalts. Cr and Ni abundances are quite high (Cr between approximately 1000 and 3600 ppm and Ni between 155 and 550 ppm, compared to approximately 500 and 175 ppm for Kroenke-type basalts, respectively), which suggests that these samples may be more primitive than Kroenke-type basalts.

There is some evidence that the LREE can be mobilized during seawater alteration [12, 127, 151] (recall that these samples have experienced submarine alteration, subaerial alteration in an rain forest environment, and potential trace element equilibration with the host aln¨oite) resulting in either enrichments or depletions in overall abundances [12, 127, 151, 223]. Enrichment of the LREE over the HREE were seen in small basaltic clasts contained with a porous volcaniclastic matrix at

124 Site 1184 on the OJP and were attributed to aqueous alteration of the samples

[223]. The aln¨oite breccia probably behaves similarly to the volcaniclastic material at Site 1184 and is very porous and often highly altered itself (see Section ??). Therefore it is possible that the more fractionated REE proﬁles of these samples are due to preferential enrichment of the LREE over the HREE during alteration.

The major element, compatible trace element, heavy REE, and petrography data suggests that these samples are more primitive than Kroenke-type basalt, while the Sm-Nd isotopic data from PHN 5793 suggests that they are from the same source. The interpretation favored here is that these samples represent parental magmas to Kroenke-type lavas.

5.8 Conclusions

Xenoliths representing upper mantle pyroxenites and peridotites and OJP crustal basalts/gabbros were analyzed for major and trace elements. Electron microprobe data on mineral phases was used to determine metamorphic conditions and the crystallization pressure and temperatures. Sm-Nd and Rb-Sr isotopic data identify certain samples as having an OJP affinity. The presence of pyrope-rich garnet in samples PHN 5802 and PHN 3939 combined with the high modal abundance of pyroxene suggested that these samples are upper mantle pyroxenites. High crystallization pressures determined from EMP data on clinopyroxene and orthopyroxene compositions confirmed that they originated from at least 50 km depth. They are similar to other pyroxenites found in the Malaitan alnöite and represent fragments of the mantle root below the OJP. Grossular-rich garnet in sample PHN 5854 suggests metamorphic reaction of anorthite-rich plagioclase and olivine to grossular garnet. This fluid may have

125 also mobilized the HREE of this sample causing apparent REE depletions. PHN

5853 is a harzburgite, also interpreted from be from the upper mantle below the OJP at 34 Ma. Samples BM/SI-7 and ST-7 appear to be samples of Kroenke-type basalt and therefore show that the aln¨oite has sampled wall-rock from a large range of depths: from upper mantle at least 10 km below the crustal bottom of the OJP to basalts that were erupted on the surface. PHN 5765, PHN 5792, PHN 5793, and PHN 5843 appear to be slightly more primitive basalt/gabbros than Kroenke-type lavas, which was the previously most primitive OJP lava type. Additional Sm-Nd isotopic data (and to a lesser degree Rb-Sr) for samples without isotopic data (especially for the basalt/gabbro samples) would be extremely useful to further identify the provenance of these samples. Major and trace element data suggest origination from a OJP source, but isotopic data would provide further evidence.

126 CHAPTER 6

HAWAIIAN PLUME SOURCE HYBRIDIZATION: EVIDENCE FROM NINTOKU SEAMOUNT, EMPEROR SEAMOUNT CHAIN

Background information regarding the Hawaiian Ridge-Emperor Seamount Chain are presented in Chapter 3. Detailed descriptions of analytical techniques can be found in Chapter 9.

6.1 Nintoku Seamount

Nintoku Seamount is located at 41 ◦N, 175 ◦E in the central part of the ESC

(Fig. 3.2), and formed at approximately 56 Ma [61, 72]. The upper 32 m of the volcanic basement of this seamount was previously drilled during Deep Sea Drilling Project (DSDP) Leg 55, Site 432, where three ﬂow units of alkali basalt were recovered [130]. Site 1205 of ODP Leg 197 is located at the north-west edge of the summit region of Nintoku Seamount in 1310 m water depth, about 100 m south-west of Site 432. Hole 1205A encountered basement at 42.7 meters below sea level (mbsf) and penetrated 283.3 m into a sequence of basaltic lavas and interbedded sediments and soil horizons (Fig. 6.1). The minimum age of the basement rocks at this site is constrained to be 53.6–54.7 Ma on the basis of the nanofossil assemblage (zone NP10, Early Eocene) recovered in Core 197-1205A- 5R immediately overlying volcanic basement. Dalrymple et al. [61] obtained an

127 ’ n aoheflwtps hc eecmol eiua an vesicular commonly were which types, flow pahoehoe and a’a 40 ihn infiatdffrnei g ewe ape rmthe from section. samples the between age in difference significant no with n n edprsprt rmSt 25gv enaeo 55 of age mean a gave 1205 Site from separate feldspar one and aatfo SPSt 432. Site DSDP from basalt Ar tlat2 aaflwuiswr eoee tSt 25 hs i These 1205. Site at recovered were units flow lava 25 least At / itn hcns,mnrlg,adptooyo ocncu volcanic of petrology and mineralogy, thickness, listing 39 iue61 tairpi rfiefrHl 25 itk Sea Nintoku 1205A Hole for profile Stratigraphic 6.1. Figure raeo 56 of age Ar

De htp ( mb )fs 200 150 100 50 0 14R 13R 12R 11R 10R 29R 28R 27R 26R 25R 24R 23R 22R 21R 20R 19R 18R 17R 16R 15R 7R 6R 5R 4R 3R 2R 1R 9R 8R

. C ero 2

± Recov yre Hol e 1

0 Se id me tn htiL olo gy 205A . afrasml foiieadpaicaepyi alkali phyric plagioclase and olivine of sample a for Ma 6 12b 12a 15b 14b 14a 11b 11a 13b 13a 15a 10 5a 8a 8b 5b 3b 3a 9 4 2 7 6 1

U tin Seamount Nintoku Mod Plag-Ol Mod Plag-Ol Mod Plag-Ol Mod-Hi Plag Mod-Hi Sp Plag-Ol Sp Aphyric Sandstone to Aphyric rmTarduno from Sp-Mod Ol- Plag Aphyric Sp-Mod Plag-Ol Breccia Breccia Aphyric Aphyric Sp-Mod/ Plag-Ol Soil Soil Soil Soil Soil Soil Soil Soil 40 Rock yt pe Ar 0.9 1.0 0.6 0.4 0.9 1.0 1.1 2.0 1.4 1.2 2.5 1.5 klA ytinila ni dex p 3-15 o 1- o 1- p 1- p 1- o 1- p 3- o 1- p 5- o 1- p 4- o 1- p 1- / p 3 o 1 p 1 39 2 3 8 5 2 5 3 7 3 7 3 2 Pheno yrc sts ( %) 1-15 3-10 0-30 1-15 0-30 1-40 2-25 0-10 1-10 1-7 1-8 0-8

rpaeuae rmsxwoerc samples whole-rock six from ages plateau Ar Vesi elc s (%) m-h s-m s-m s-m s-m s-h s-h s-h s-h s-h s-m m oitaretlA n 128 6 1 1 1 1 1 2 1 1 1 2 1 1 # of samples 300 250 tal. et TD 326.0mbsf 35R 34R 33R 32R 30R 44R 43R 42R 41R 40R 37R 36R p o Phen Weat Plag Ol Hi Mod Sp Roc 45R 39R 38R 31R C ero lagioclase livine k N hered ocr Basalt Breccia Recov yre ame yst [236] weath plagioclase oliv highly moder sparsely s (%) htiL olo gy ine ered flo ered ately 29a 28a 24a 19b 19a 18b 18a 30b 26b 26a 19c 28b 24b 30a 20 27 23 22 21 17 16 25

–c U tin Weathered Sp-M Mod Mod Mod Mod-Hi Ol Mod-Hi Aphyric to Aphyric Aphyric to Aphyric Aphyric to Aphyric Mod Plag Aphyri Aphyric Sp Pl Sp-M Sp-M Breccia Breccia Breccia Breccia Breccia Plag-Ol Plag-Ol Plag-Ol Mod Ol Mod Ol Mod-Hi w top Soil Soil Soil Ol- Ol- Ol- Soil Sediment od c t ag od od Plag Plag Plag Rock yt pe o Ol h m s Alt negative saturatio Devia Alkal c numbers ar numbers thol -0.8 -0.5 1.2 1.5 1.3 0.7 0.7 0.8 1.1 1.3 1.4 0.6 0.6 erat sli com mod hig eiit

ght klA ytinila ni dex inity tion fromsilica- h ic) o 5- o 3- o 4- p 1- o 3- o 1- p 1- o 3- p 0- p 1- p 1- o 1- p 1- o 1- plete ion erate p 1 p 1 o 2 o 1 numbers are numbers n li In 10 20 11 7 5 7 2 6 2 2 3 2 4 6 Pheno yrc sts ( %) e al ne (positive dex 1-10 0-25 1-15 -0s-h 2-10 1-25 2-25 1-25 1-50 0-3 1-5 0-7 0-5 7 kalic, kalic,

is Adapted nits. Ve elcis s ( %) o n otmof bottom and top . m-h m-h s-m s-m s-m m-c s-h s-h s-h s-h s-h s-c 51 aeweathered have d tlA are oit n 1 1 1 1 1 1 2 1 1 1 1 2 # of samples ± mount, cue both ncluded 0 . 2M [72], Ma 22 flow top and inter-flow soil horizons indicating that the sequence was erupted subaerially and that eruption rates were relatively low [236]. The lavas drilled at Site 1205 vary from aphyric to plagioclase-phyric to plagioclase- olivine-phyric basalts with texture varying from subophitic to subtrachytic [236]. The groundmass is typically composed of plagioclase, olivine, clinopyroxene, titanomagnetite, and glass. Preliminary shipboard major element analyses of the lava flow units showed that most are composed of alkali basalt. A conglomerate containing clasts of alkali basalt, hawaiite, and mugearite overlies the basement at Sites 1205 and 432. Alkali basalts and hawaiites with similar chemistry and mineralogy characterize the post-shield alkalic stage of many young Hawaiian volcanos, including Mauna Kea [88], Kohala [138], and Haleakala [38].

6.2 Samples

A total of 33 samples were analyzed (Table 6.1). Of these, six samples are clasts of volcanic rock that were extracted from the conglomerate unit directly overlying the volcanic basement. Drill core samples were crushed to 0.5–1.0 cm sized chips and the freshest material was hand-picked. Vesicle ﬁllings, veins, and highly altered material were avoided. The chips were rinsed in distilled water, dried, and powdered in an agate swing mill at the Max Plank Institut fur Chemie, Mainz.

129 TABLE 6.1 SUMMARY OF UNIT NUMBER, RECOVERY DEPTH, AND ROCK TYPE OF SAMPLES RECOVERED FROM SITE 1205A

Sample Unit # Depth (mbsf) Rock Type

5R-2, 21-35 SE II 34.09 mugearite

5R-2, 31-34 SE II 34.19 mugearite

5R-2, 74-77 SE II 34.62 mugearite

5R-2, 90-94 SE II 34.78 alkalic basalt

5R-2, 96-100 SE II 34.84 mugearite

5R-2, 114-116 SE II 35.02 hawaiite

6R-2, 8-12 1 43.88 moderately to highly plagioclase-phyric basalt

6R-2, 114-119 1 45.02 moderately to highly plagioclase-phyric basalt

10R-2, 0-5 3B 63.12 sparsely to moderately olivine-plagioclase-phyric basalt

15R-3, 10-12 5B 93.65 moderately plagioclase-olivine-phyric basalt

19R-1, 80-86 6 110.70 moderately plagioclase-olivine-phyric basalt

20R-5, 67-72 8A 125.80 moderately plagioclase-olivine-phyric basalt

21R-3, 71-76 8B 132.83 plagioclase-olivine basalt autoclastic breccia

23R-1, 85-90 10 139.75 sparsely plagioclase-olivine-phyric to aphyric basalt

24R-1, 51-56 10 149.01 sparsely plagioclase-olivine-phyric to aphyric basalt

26R-2, 30-34 11B 169.50 aphyric basalt

27R-6, 124-129 12B 185.66 aphyric basalt

28R-3, 10-15 13B 189.75 aphyric basalt

29R-3, 92-95 14B 199.01 sparsely to moderately plagioclase-olivine-phyric basalt

29R-4, 117-121 15B 201.30 sparsely to moderately plagioclase-olivine-phyric basalt

130 TABLE 6.1 (CONTINUED)

Sample Unit # Depth (mbsf) Rock Type

32R-3, 17-22 17 220.47 sparsely to moderately plagioclase-olivine-phyric basalt

33R-2, 70-75 17 226.61 sparsely to moderately plagioclase-olivine-phyric basalt

34R-3, 79-84 18B 238.29 aphyric to moderately olivine-phyric basalt

35R-3, 108-114 19B 247.66 moderately to highly olivine-phyric basalt

36R-5, 135-139 20 260.44 sparsely to moderately olivine-phyric basalt

37R-5, 38-42 21 270.02 moderately to highly plagioclase-olivine-phyric basalt

38R-1, 4-7 22 273.64 moderately olivine-plagioclase-phyric basalt aphyric basalt breccia

41R-3, 18-23 24B 285.02 aphyric basalt breccia

41R-5, 46-50 26A 288.24 aphryic to sparsely plagioclase-phyric basalt

42R-4, 12-16 27 296.96 aphyric to sparsely plagioclase-phyric basalt

44R-2, 14-19 28 313.34 olivine-plagioclase basalt breccia

44R-2, 116-121 29A 314.36 olivine basalt breccia

45R-1, 115-118 30B 322.65 aphyric to moderately olivine-plagioclase-phyric basalt

131 Detailed descriptions of the techniques used to acquire LOI, major element, trace element, Sr-, Nd-, and Pb- isotopic ratios can be found in Chapter 9.

6.3 Results

We divide the Site 1205 Nintoku Seamount lavas into an upper and lower alkalic series, which are separated by Unit 19b (tholeiitic basalt).

6.3.1 Major Elements

Whole-rock major element concentrations and LOI values of 33 samples of volcanic rock from Nintoku Seamount are presented in Table 6.2. Major element totals (total + LOI) vary from 97.5–105.38, although much of the variation is caused by 5 samples (Table 6.2). The major element totals of the other 28 samples are 100 ± 2 wt.%. Sample 35R-3, 108-114 from Unit 19B, which contains accumulated olivine, is the only sample that plots in the tholeiitic basalt ﬁeld (Fig.

6.2). Although alteration may increase Na2O and K2O contents (see Discussion), the high abundances of the more immobile elements such as TiO2 and P2O5 and incompatible trace element concentrations (e.g., Nb and Zr) of the Nintoku lavas compared to Hawaiian tholeiites conﬁrm that the former are indeed alkalic. The most evolved rock types are the conglomerate clasts from Core 5R, which were not present as ﬂow units deeper in the recovered basement section at Site 1205.

The signiﬁcant variation in composition of the lava clasts from the conglomerate indicate that they were derived from several discrete ﬂow units.

132 TABLE 6.2 MAJOR ELEMENT ABUNDANCES FOR LEG 197, SITE 1205 BASALTS

1 2 ID Depth (mbsf) SiO2 Al2O3 TiO2 MnO MgO Fe2O3 Na2O K2O P2O5 CaO Total LOI

5R-2, 21-25 -34.09 49.0 21.0 3.47 0.10 1.85 9.61 5.58 2.20 1.21 5.99 94.18 7.30

5R-2, 31-34 -34.19 49.6 18.1 3.01 0.13 1.80 12.5 5.09 3.61 0.91 5.33 101 3.14

5R-2, 74-77 -34.62 48.5 17.1 2.84 0.19 3.79 12.9 5.07 2.33 0.83 6.44 101 4.31 133

5R-2, 90-94 -34.78 47.5 17.1 2.66 0.15 5.90 14.0 4.20 0.86 0.44 7.13 98.5 3.58

5R-2, 96-100 -34.84 49.5 18.0 2.93 0.14 3.44 11.5 5.23 2.43 0.84 6.04 98.9 3.30

5R-2, 114-116 -35.02 48.6 17.5 2.96 0.13 3.04 14.1 4.84 2.12 0.84 5.87 96.0 4.22

6R-2, 8-12 -43.88 47.7 16.7 2.71 0.14 6.17 11.6 3.53 1.13 0.80 9.53 97.1 3.41

6R-2, 114-119 -45.02 48.2 17.0 2.68 0.17 6.31 10.7 3.83 1.35 0.71 9.03 97.3 3.62

10R-2, 0-5 -63.12 47.2 15.6 2.95 0.23 5.70 14.3 4.15 1.15 0.64 8.09 98.7 1.41

15R-3, 10-12 -93.65 46.9 15.7 2.47 0.20 6.84 13.8 3.37 0.73 0.43 9.59 97.8 2.50

19R-1, 80-86 -110.70 47.3 16.5 2.49 0.17 6.60 12.5 3.49 1.00 0.60 9.37 95.3 3.04

20R-5, 67-72 -125.80 47.4 15.2 2.63 0.20 5.97 13.9 3.83 0.82 0.52 9.42 96.5 1.29 TABLE 6.2 (CONTINUED)

1 2 ID Depth (mbsf) SiO2 Al2O3 TiO2 MnO MgO Fe2O3 Na2O K2O P2O5 CaO Total LOI

21R-3, 71-76 -132.83 47.9 15.6 2.54 0.17 6.65 13.2 3.35 0.61 0.49 9.50 96.7 2.52

23R-1, 85-90 -139.75 47.8 16.1 2.61 0.19 7.02 12.7 3.44 1.08 0.49 8.60 96.6 3.27

24R-1, 51-56 -149.01 47.4 16.3 3.02 0.20 7.21 14.3 3.39 0.86 0.46 6.90 93.8 4.26

26R-2, 30-34 -169.50 47.5 15.6 2.65 0.20 6.96 13.3 3.32 0.54 0.48 9.53 96.0 3.13 134

27R-6, 124-129 -185.66 47.8 15.6 2.20 0.17 7.19 12.7 3.29 0.61 0.40 9.99 96.0 3.13

28R-3, 10-15 -189.75 47.5 14.8 2.19 0.20 7.17 14.0 3.21 1.01 0.34 9.57 97.0 2.22

29R-3, 18-23 -199.0 47.7 15.7 2.72 0.32 6.24 13.7 3.54 0.47 0.46 9.16 96.1 3.01

29R-4, 117-121 -201.30 48.2 16.6 2.16 0.19 6.66 12.3 3.57 0.73 0.40 9.15 96.6 4.79

32R-3, 17-22 -220.47 47.8 16.2 2.04 0.17 6.82 12.4 3.28 1.09 0.37 9.75 97.2 3.05

33R-2, 70-75 -226.61 47.6 15.7 2.38 0.19 6.74 13.1 3.23 1.07 0.43 9.60 97.6 3.26

34R-3, 79-84 -238.29 46.6 15.6 2.33 0.20 8.36 14.0 3.13 0.40 0.53 8.88 91.4 6.02

35R-3, 108-114 -247.66 46.5 12.1 1.16 0.19 16.3 13.7 1.93 0.15 0.24 7.69 91.4 7.18

36R-5, 135-139 -260.44 47.1 14.0 2.58 0.19 9.03 13.9 3.25 0.46 0.53 8.91 96.2 2.07 TABLE 6.2 (CONTINUED)

1 2 ID Depth (mbsf) SiO2 Al2O3 TiO2 MnO MgO Fe2O3 Na2O K2O P2O5 CaO Total LOI

37R-5, 38-42 -270.02 46.8 14.9 2.84 0.19 7.37 13.9 3.36 1.11 0.71 8.86 100 1.77

38R-1, 4-7 -273.64 46.8 15.0 2.85 0.19 7.28 13.8 3.39 1.15 0.67 8.87 102 1.47

41R-3, 18-23 -285.02 47.6 15.1 2.81 0.20 7.02 14.3 3.42 0.49 0.54 8.54 94.0 3.08

41R-5, 46-50 -288.24 47.5 15.5 2.70 0.22 6.57 13.9 3.37 1.05 0.50 8.72 98.0 3.10 135

42R-4, 12-16 -296.96 47.5 15.5 2.68 0.23 5.95 15.2 3.53 0.26 0.44 8.68 98.8 2.90

44R-2, 14-19 -313.34 47.3 16.2 3.22 0.26 5.70 14.2 3.40 0.69 0.60 8.40 93.5 5.78

44R-2, 116-121 -314.36 46.5 16.6 3.42 0.28 6.24 15.6 3.00 0.61 0.58 7.21 94.7 3.31

45R-1, 115-118 -322.65 47.1 16.3 3.19 0.21 5.35 14.3 3.56 0.93 0.59 8.47 97.5 2.21

BHVO-2 AVG 49.0 13.3 2.69 0.17 7.42 11.4 2.10 0.42 0.38 11.3

Recommended Values3 49.9 13.5 2.73 0.20 7.23 12.3 2.22 0.52 0.27 11.4

Std. Dev. (n = 15) 2.00 0.65 0.10 0.01 0.25 0.58 0.13 0.31 0.02 0.43 TABLE 6.2 (CONTINUED)

1 2 ID Depth (mbsf) SiO2 Al2O3 TiO2 MnO MgO Fe2O3 Na2O K2O P2O5 CaO Total LOI

BIR-1 AVG 48.9 15.9 0.99 0.18 9.96 10.9 2.10 bdl 0.08 13.7

Recommended Values3 48.0 15.5 0.96 0.02 9.70 11.3 1.82 0.03 0.02 13.3

Std. Dev. (n = 10) 3.81 1.24 0.08 0.18 0.85 0.92 0.17 0.44 0.01 1.07 136

BPL-1 AVG 47.3 15.3 2.35 0.20 8.25 12.7 2.40 1.14 0.69 9.96

Recommended Values4 47.3 15.3 2.40 0.20 8.30 12.5 2.40 1.10 0.50 10.0

Std. Dev. (n = 10) 0.90 0.35 0.05 0.01 0.23 0.48 0.08 0.12 0.04 0.16

1Sample data presented in this table have been normalized to 100 wt.%

2Un-normalized major element total

3BHVO-2 and BIR-1 recommended values from USGS preliminary certiﬁcate of analysis and USGS certiﬁcate of analysis, respectively

4BPL-1 recommended values from Scott Hughes, Idaho State University, personal communication 10

Mugearite 8

Alkalic

Hawaiite 6 O 2 K

+ Tholeiitic

O 2 a

N 4

Basalt

Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts Tholeiitic Sample 0 40 45 50 55 60

SiO 2

Figure 6.2. Classiﬁcation of Site 1205 basalts, SiO2 (wt.%) vs. Total Alkalis (Na2O+K2O) (wt.%), where the dashed line separates Hawaiian tholeiite and alkali lavas [165]. Fields for rock types are from Cox et al. [56]. All samples except for 35R-3, 108-114 are alkalic. Samples from Core 5R-2 in Unit SE II (the conglomerate overlying the volcanic basement) are predominantly more evolved than underlying alkalic basalts

MgO contents vary from 1.74 to 14.9 wt.%. As MgO abundances decrease,

CaO, MnO, and FeOT (total Fe as FeO) concentrations decrease whereas P2O5,

K2O, Na2O, SiO2, Al2O3, and TiO2 increase (Fig. 6.3). The hawaiite/mugearite clasts from the conglomerate have higher TiO2 and P2O5 contents than their younger counterparts from Mauna Kea, Haleakala, and Kohala volcanos.

6.3.2 Trace Elements

Trace element contents of the Nintoku Seamount lavas from Site 1205 are presented in Table 6.3. Nickel, Cr, Sr, and V generally increase down the section.

137 50 4 (a) (b) 49.5 3.5 49 ) )

% 3 % .t 48.5 .t w w 48 ( ( 2.5 ( ( 2 2 O

O 47.5 iT i 2 S 47 1.5 46.5

46 1 0 5 10 15 20 0 5 10 15 20

22 16 (c) (d) 15 20

) 14 ) %

% .t 18 .t 13 w w ( ( ( ( 3 T

O 12

2 16 lA O e

F 11 14 10

12 9 0 5 10 15 20 0 5 10 15 20 0.35 11 (e) (f)

0.3 10 ) ) 0.25 9 % % .t .t w w ( ( 0.2 ( 8 O O n 0.15 a 7 C M

0.1 6

0.05 5 0 5 10 15 20 0 5 10 15 20 6 1.4 (g) (h)

1.2 5 ) )

% 1 % .t 4 .t w w ( ( (

0.8 5 O 2

3 O 2 a 0.6 P N

2 0.4

1 0.2 0 5 10 15 20 0 5 10 15 20 MgO (wt. %) MgO (wt. %)

Figure 6.3. MgO variation diagrams. The clasts from the conglomerate form a roughly linear extension from the Core 6R-2 alkalic basalts (MgO = 6.17 and 6.31 wt.%), suggesting that they formed by fractional crystallization of melts of alkalic basalt. Symbols are the same as in Figure 6.2.

One sample from Unit 19b, which contains accumulated olivine, has the highest Ni content (1135 ppm) while the hawaiites and mugearites contain the lowest concentrations of compatible trace elements. The Ni contents of the alkali basalts (< 234 ppm) are too low for these lavas to represent primary melts of mantle peridotite, indicating that they have undergone fractional crystallization within the crust.

The concentration of the highly incompatible elements that are immobile during alteration (e.g., Zr and Nb) vary by a factor of ∼ 13, with the highest concen-

138 trations in the most evolved lavas (up to 91 times primitive mantle abundances).

Chondrite-normalized REE patterns of the Nintoku lavas are similar to those of similar rock types from the Hawaiian Islands (Fig. 6.4). Hawaiites and mugearites have higher concentrations of incompatible trace elements and steeper REE patterns (La/Yb varies from 12.6–17.0) compared to the alkali basalts (La/Yb <

10.0). Ba/Th ratios (113–162) of the Nintoku lavas are high compared to most other intraplate oceanic lavas, but are similar to other Hawaiian lavas [259].

139 TABLE 6.3 TRACE ELEMENT ABUNDANCES (PPM) OF LEG 197, SITE 1205 BASALTS, PART 1

ID Li Be Sc V Cr Co Ni Cu Zn Ga Rb Sr Y Zr Nb Mo Sn Sb Cs Ba La Ce

5R-2, 21-25 22.2 1.42 15.2 114 30.7 41.4 14.9 30.0 170 0.26 11.3 1008 36.5 276 57.4 1.20 1.69 0.35 0.20 524 43.4

5R-2, 31-34 11.4 1.83 13.9 89.3 45.8 35.5 8.32 21.7 140 12.2 30.6 907 29.0 230 57.7 2.03 1.68 0.09 0.27 617 39.6

5R-2, 74-77 14.8 1.70 13.8 98.5 44.1 28.9 9.74 23.7 175 17.0 26.7 899 26.3 233 54.5 2.20 1.37 0.18 0.22 651 36.4 140

5R-2, 90-94 11.7 1.16 28.2 373 194 53.6 76.4 71.3 201 29.5 13.4 432 22.1 137 25.7 1.47 1.29 0.16 0.08 233 18.2

5R-2, 96-100 17.9 1.93 14.8 100 47.0 34.9 11.1 32.1 157 13.2 48.6 920 32.3 234 64.7 3.23 1.99 0.31 0.83 659 41.5

5R-2, 114-116 16.3 1.47 14.0 85.9 50.1 37.9 9.61 19.9 149 75.0 33.3 908 27.2 217 59.8 2.98 1.51 0.13 0.43 499 30.4

1 6R-2, 8-12 13.3 1.59 23.6 270 140 52.9 76.5 89.6 155 49.3 19.2 728 23.5 209 40.3 1.55 6.09 bdl 0.06 383 25.4

6R-2, 114-119 14.8 1.60 23.8 267 123 60.2 76.0 62.4 164 12.8 17.1 761 21.9 166 34.9 1.70 1.11 0.12 0.11 390 28.5

10R-2, 0-5 14.0 1.67 23.2 233 147 49.2 57.7 67.4 159 44.1 22.7 547 28.5 216 42.4 2.43 8.08 0.09 0.17 392 25.5

15R-3, 10-12 9.40 1.01 29.6 367 150 56.2 76.6 97.3 153 31.2 12.4 489 23.6 132 22.6 1.24 5.25 bdl 0.08 212 15.0

19R-1, 80-86 6.61 1.24 28.3 278 105 52.2 80.2 72.6 149 35.1 17.0 773 26.1 158 30.1 1.23 6.51 0.02 0.10 281 20.2 TABLE 6.3 (CONTINUED)

ID Li Be Sc V Cr Co Ni Cu Zn Ga Rb Sr Y Zr Nb Mo Sn Sb Cs Ba La Ce

20R-5, 67-72 9.89 1.22 27.7 311 99.5 50.4 61.2 92.0 169 27.0 18.0 560 271. 155 28.1 1.58 6.15 0.03 0.09 246 17.7

21R-3, 71-76 12.0 1.05 29.6 351 180 56.2 78.7 83.1 151 33.1 12.9 399 24.5 146 25.4 1.17 6.05 0.03 0.11 216 14.6

23R-1, 85-90 13.3 1.16 31.2 372 174 59.1 75.4 99.7 168 37.5 14.5 399 26.4 158 27.1 1.28 6.49 0.04 0.14 225 15.8

24R-1, 51-56 19.6 1.09 37.0 440 121 63.6 68.1 107 181 37.5 15.6 421 30.6 160 27.7 1.05 6.75 0.05 0.10 239 18.3 141

26R-2, 30-34 6.13 1.12 28.8 363 149 52.9 93.3 92.5 159 24.8 12.7 418 24.2 143 26.1 1.15 5.59 0.02 0.04 227 16.5

27R-6, 124-129 4.93 0.89 31.7 353 246 61.0 112 99.3 145 28.5 9.61 360 22.6 121 18.5 1.09 5.50 0.02 0.03 157 11.1

28R-3, 10-15 6.44 0.81 33.2 363 134 60.4 74.7 103 160 28.3 6.95 341 25.3 115 17.5 1.08 4.91 0.03 0.02 142 11.6

29R-3, 18-23 4.29 1.01 29.7 379 146 50.4 66.4 76.9 162 29.1 9.18 351 27.8 152 26.2 1.42 5.80 0.03 0.03 197 16.5

29R-4, 117-121 5.39 0.74 29.1 313 241 51.8 108 89.0 146 27.4 9.15 315 22.1 112 18.3 0.95 4.60 bdl 0.02 157 12.1

32R-3, 17-22 5.06 0.89 28.1 309 186 55.2 111 88.2 132 23.9 9.95 366 21.4 112 18.1 0.91 4.50 bdl 0.04 155 12.1

33R-2, 70-75 2.66 1.05 30.6 371 2.06 58.8 89.5 94.6 147 33.1 11.0 395 25.9 144 25.1 1.17 4.88 0.03 0.02 218 15.4 TABLE 6.3 (CONTINUED)

ID Li Be Sc V Cr Co Ni Cu Zn Ga Rb Sr Y Zr Nb Mo Sn Sb Cs Ba La Ce

34R-3, 79-84 4.21 1.02 25.2 257 299 65.8 185 100 144 28.9 1.44 370 26.3 160 25.4 1.10 6.13 0.05 0.05 206 16.2

35R-3, 108-114 5.09 0.67 42.3 361 1489 176 1135 135 187 24.9 1.44 335 24.2 91.1 11.4 0.73 13.2 0.05 0.02 98.7 8.10

36R-5, 135-139 14.2 1.31 25.6 298 270 65.4 234 103 179 20.9 21.9 582 22.8 136 26.9 1.38 1.21 0.02 0.04 264 20.5

37R-5, 38-42 7.25 1.53 25.6 300 173 56.9 152 103 169 28.9 18.7 632 25.7 193 34.2 1.76 6.90 0.06 0.24 313 23.7 142

38R-1, 4-7 6.24 1.46 26.6 336 217 63.4 149 105 171 45.0 16.2 647 25.6 199 33.7 1.89 7.14 0.04 0.13 333 22.0

41R-3, 18-23 6.09 1.23 33.4 298 134 49.1 81.6 84.6 162 27.3 14.1 359 33.0 154 25.1 1.18 6.69 0.04 0.06 219 19.2

41R-5, 46-50 3.81 0.98 32.8 379 178 61.7 87.6 79.3 187 36.2 12.5 362 29.7 171 24.1 1.35 6.62 bdl 0.02 207 15.2

42R-4, 12-16 1.69 1.07 30.2 344 145 60.0 84.5 71.4 163 32.4 12.9 348 28.9 166 24.4 1.44 5.87 bdl 0.03 213 14.6

44R-2, 14-19 4.20 1.05 32.3 393 168 66.0 81.7 92.7 205 41.1 12.8 398 34.2 219 29.9 1.20 8.16 0.03 0.03 255 18.9

44R-2, 116-121 6.41 1.20 31.7 311 154 55.3 70.5 81.2 187 31.2 12.7 310 34.0 190 27.2 1.33 7.41 bdl 0.04 226 19.9

45R-1, 115-118 4.98 1.35 27.0 361 173 56.8 78.1 83.8 190 33.1 16.6 393 30.3 215 28.2 1.85 7.24 0.03 0.12 244 18.4 TABLE 6.3 (CONTINUED)

ID Li Be Sc V Cr Co Ni Cu Zn Ga Rb Sr Y Zr Nb Mo Sn Sb Cs Ba La Ce

BHVO-2 3.86 1.03 38.0 345 351 56.2 137 151 138 25.3 10.1 407 25.8 174 19.2 3.94 5.78 0.15 0.11 142 15.5

2 BHVO-2 Recomm. Values 5.00 1.10 32.0 317 280 45.0 119 127 103 21.7 9.80 389 26.0 172 18.0 1.02 1.90 0.16 0.13 130 15.0

Std. Dev. (n = 5) 1.07 0.08 2.46 47.8 46.3 3.93 10.7 18.7 19.3 2.26 0.63 19.6 0.95 16.7 0.91 1.53 3.86 0.03 0.04 6.20 0.67 143 1Below detection limit

2BHVO-2 recommended values from USGS preliminary certiﬁcate of analysis TABLE 6.3 TRACE ELEMENT ABUNDANCES (PPM) OF LEG 197, SITE 1205 BASALTS, PART 2

ID Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta W Pb Th U

5R-2, 21-25 11.7 44.4 9.33 3.38 9.89 1.32 7.62 1.43 3.70 0.51 2.97 0.43 6.93 3.87 0.14 2.27 4.63 1.89

5R-2, 31-34 10.2 38.2 7.63 2.61 8.27 1.08 6.08 1.10 2.93 0.40 2.47 0.33 5.94 3.62 0.22 1.63 3.91 1.06

5R-2, 74-77 9.30 35.2 7.13 2.61 7.89 0.98 5.36 1.04 2.75 0.37 2.20 0.31 5.90 3.68 0.12 1.18 4.02 1.04 144

5R-2, 90-94 4.99 19.1 4.45 1.85 5.69 0.77 4.33 0.79 2.14 0.29 1.88 0.23 3.91 1.52 0.11 0.38 1.73 0.46

5R-2, 96-100 11.0 41.1 8.33 3.01 8.99 1.17 6.48 1.27 3.39 0.45 2.71 0.35 6.38 3.89 0.45 1.99 4.40 1.37

5R-2, 114-116 8.01 29.9 6.20 2.19 7.15 0.89 5.04 0.95 2.34 0.31 2.00 0.26 5.14 3.21 0.31 0.67 3.40 0.96

6R-2, 8-12 7.10 29.2 6.28 2.21 6.53 0.82 4.47 0.87 1.87 0.22 1.49 0.17 4.23 2.30 0.54 1.65 2.64 0.77

6R-2, 114-119 7.77 29.7 6.33 2.27 7.21 0.93 5.06 0.92 2.32 0.40 1.81 0.24 4.40 2.29 0.18 1.52 2.61 0.96

10R-2, 0-5 7.04 28.5 6.58 2.21 6.88 0.91 5.06 1.01 2.19 0.29 2.03 0.29 4.66 2.58 0.72 2.03 3.01 0.90

15R-3, 10-12 4.64 19.5 4.82 1.79 5.25 0.78 4.87 0.97 2.25 0.31 2.09 0.27 3.14 1.35 0.16 1.02 1.46 0.41

19R-1, 80-86 5.95 24.2 5.77 1.99 5.93 0.89 5.15 1.01 2.36 0.31 2.04 0.29 3.80 1.69 0.21 1.28 2.07 0.23 TABLE 6.3 (CONTINUED)

ID Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta W Pb Th U

20R-5, 67-72 5.36 22.8 5.70 1.96 6.08 0.89 5.66 1.13 2.63 0.35 2.05 0.31 3.80 1.58 0.17 1.15 1.70 0.54

21R-3, 71-76 4.39 18.7 4.73 1.60 4.92 0.72 4.23 0.86 1.99 0.27 1.72 0.24 3.30 1.43 0.19 1.15 1.61 0.38

23R-1, 85-90 4.81 20.1 5.02 1.85 5.38 0.78 4.66 0.94 2.32 0.26 1.79 0.26 3.51 1.53 0.26 1.01 1.69 0.27

24R-1, 51-56 5.44 22.8 5.66 2.01 6.03 0.89 5.30 1.11 2.66 0.36 2.07 0.29 3.60 1.67 0.22 1.49 2.10 0.31 145

26R-2, 30-34 5.08 21.4 5.24 1.85 5.40 0.83 5.05 0.98 2.35 0.31 1.92 0.27 3.42 1.53 0.15 1.12 1.61 0.49

27R-6, 124-129 3.60 15.5 4.05 1.43 4.53 0.64 3.72 0.80 1.72 0.23 1.53 0.22 2.81 1.05 0.20 0.83 1.14 0.32

28R-3, 10-15 3.77 16.3 4.29 1.58 4.84 0.76 4.58 0.95 2.26 0.32 1.97 0.29 2.75 1.02 0.12 0.95 1.12 0.36

29R-3, 18-23 5.03 21.2 5.28 1.87 5.83 0.89 5.46 1.10 2.70 0.37 2.35 0.32 3.52 1.45 0.11 1.11 1.58 0.25

29R-4, 117-121 3.92 16.7 4.34 1.56 4.68 0.75 4.50 0.88 2.08 0.28 1.74 0.24 2.72 1.01 0.15 0.76 1.16 0.22

32R-3, 17-22 3.87 15.8 4.10 1.49 4.61 0.72 4.18 0.86 2.03 0.27 1.76 0.24 2.81 1.00 0.11 0.76 1.11 0.33

33R-2, 70-75 4.70 19.4 4.68 1.76 5.44 0.77 4.76 0.96 2.24 0.28 1.86 0.25 3.48 1.49 0.20 0.95 1.63 0.48 TABLE 6.3 (CONTINUED)

ID Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta W Pb Th U

34R-3, 79-84 4.90 20.3 4.88 1.66 5.11 0.74 4.36 0.89 2.09 0.29 1.84 0.28 3.43 1.43 0.24 1.20 1.70 0.54

35R-3, 108-114 2.74 12.3 3.61 1.35 4.31 0.67 4.14 0.91 2.08 0.28 1.90 0.26 2.49 0.78 0.12 0.47 0.79 0.18

36R-5, 135-139 5.81 22.2 5.16 1.91 5.65 0.86 4.55 0.84 2.18 0.30 1.60 0.24 3.63 1.55 0.10 0.08 1.80 0.56

37R-5, 38-42 6.89 27.2 6.19 2.14 6.33 0.94 5.22 1.01 2.41 0.32 1.99 0.26 4.43 1.90 0.37 1.74 2.37 0.71 146

38R-1, 4-7 6.38 26.2 5.76 2.06 6.13 0.82 4.79 0.99 2.14 0.26 1.80 0.21 4.08 2.01 0.49 1.83 2.48 0.70

41R-3, 18-23 6.16 25.9 6.47 2.26 7.05 1.08 6.30 1.25 2.94 0.42 2.65 0.37 3.99 1.54 0.29 1.74 1.88 0.34

41R-5, 46-50 4.96 21.3 5.36 1.91 6.04 0.87 5.31 1.10 2.65 0.33 2.12 0.28 3.65 1.35 0.29 1.18 1.49 0.27

42R-4, 12-16 4.69 20.6 5.32 1.83 5.85 0.86 5.02 1.03 2.42 0.32 2.18 0.30 3.95 1.45 0.35 0.93 1.57 0.46

44R-2, 116-121 6.63 28.0 6.95 2.47 7.62 1.13 6.80 1.39 3.23 0.43 2.89 0.38 4.85 1.60 0.35 1.43 1.89 0.35

44R-2, 14-19 6.08 27.0 6.84 2.34 7.40 1.05 6.36 1.32 3.00 0.37 2.40 0.34 4.90 1.78 0.14 1.55 1.97 0.30

45R-1, 115-118 6.18 26.2 6.76 2.34 6.83 1.08 6.47 1.26 3.02 0.43 2.44 0.35 4.85 1.52 0.40 1.75 1.65 0.52 TABLE 6.3 (CONTINUED)

ID Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta W Pb Th U

BHVO-2 5.72 24.8 5.99 2.07 6.48 0.90 5.25 1.03 2.49 0.34 2.01 0.28 4.58 1.23 0.29 1.22 1.32 0.45

1 BHVO-2 Recomm. Values 5.70 25.0 6.20 2.10 6.30 0.90 5.20 1.04 2.40 0.33 2.00 0.28 4.10 1.40 0.20 2.60 1.20 0.42

Std. Dev. (n = 5) 0.03 0.48 0.30 0.05 0.28 0.01 0.13 0.02 0.13 0.02 0.18 0.01 0.26 0.11 0.09 0.53 0.11 0.05 147 1BHVO-2 recommended values from USGS preliminary certiﬁcate of analysis 1000 Mauna Kea Haleakala Hualalai Kohala Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts Tholeiitic Sample 100 Samples/Chondrite

10 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 6.4. Chondrite-normalized REE diagram showing the average composition of the conglomerate clasts, the upper series and lower series alkalic basalts, tholeiitic sample 35R-3, 108-114, Mauna Kea, Haleakala, Hualalai, and Kohala post-shield basalts. The compositions of the Nintoku samples are similar to those of similar rock types from modern Hawaiian volcanos

6.3.3 Sr and Nd Isotopes

Age-corrected 87Sr/86Sr ratios for the leached sample powders range from 0.703130–0.703245 and initial 143Nd/144Nd ratios range from 0.512927–0.512990

(Table 6.4). Epsilon-Nd and ǫSr values of the Site 1205 lavas are within the range of post-shield lavas from Mauna Kea, Haleakala, and Kohala volcanos, although

ǫSr is at the lower end of this range (Fig. 6.5). One sample of alkali basalt from

Site 432 analyzed by Regelous et al. [201] has lower ǫSr than any post-shield lavas from the Hawaiian Islands. There is no systematic variation of Sr or Nd isotope composition with degree of diﬀerentiation (Fig. 6.6). The older lavas from deeper in the borehole tend to have the highest ǫSr values (Fig. 6.7), as discussed in the next section. Epsilon-Sr correlates negatively with ratios of more to less incompatible trace elements, such as La/Yb (Fig. 6.8).

148 TABLE 6.4 STRONTIUM AND NEODYMIUM ISOTOPE DATA FOR LEG 197, SITE 1205 BASALTS

87 86 87 86 143 144 143 144 ID Rb Sr Sr/ Sri Sr/ Srt ǫSr Sm Nd Nd/ Ndi Nd/ Ndt ǫNd

5R-2,21-25 11.3 1008 0.703145 0.703140 -18.4 9.33 44.4 0.512973 0.512927 7.04

5R-2,90-94 13.4 432 0.703280 0.703220 -17.3 4.45 19.1 0.513040 0.512989 8.25

5R-2,96-100 48.6 920 0.703192 0.703130 -18.5 8.33 41.1 0.512986 0.512941 7.32 149 6R-2,8-12 19.2 735 0.703195 0.703156 -18.2 6.28 29.2 0.512995 0.512948 7.45

10R-2,0-5 22.7 552 0.703219 0.703141 -18.4 6.58 28.5 0.513019 0.512968 7.85

15R-3,10-12 21.1 533 0.703244 0.703174 -17.9 4.51 19.0 0.513003 0.512951 7.50

19R-1,80-86 17.0 781 0.703247 0.703174 -17.9 6.14 24.4 0.513001 0.512946 7.41

20R-5, 67-72 6.06 23.0 0.513016 0.512958 7.67

21R-3, 71-76 4.73 18.7 0.512998 0.512942 7.34 23R-1,85-90 14.5 404 0.703232 0.703162 -18.1 5.02 20.1 0.513025 0.512970 7.88

24R-1,51-56 15.6 426 0.703236 0.703161 -18.1

26R-2,30-34 12.7 422 0.703249 0.703180 -17.8 5.57 21.6 0.513014 0.512957 7.63 TABLE 6.4 (CONTINUED)

87 86 87 86 143 144 143 144 ID Rb Sr Sr/ Sri Sr/ Srt ǫSr Sm Nd Nd/ Ndi Nd/ Ndt ǫNd

27R-6,124-129 9.61 363 0.703259 0.703199 -17.5 4.05 15.5 0.513002 0.512944 7.38 28R-3,10-15 6.37 308 0.703246 0.703202 -17.5 4.45 16.2 0.513015 0.512954 7.58

29R-3,18-23 9.18 355 0.703244 0.703189 -17.7

29R-3, 92-95 5.61 21.3 0.513032 0.512974 7.96 150 29R-4, 117-121 4.61 16.8 0.513014 0.512954 7.56

32R-3,17-22 15.1 349 0.703288 0.703231 -17.1 3.84 15.6 0.513011 0.512957 7.62

33R-2,70-75 11.0 399 0.703255 0.703190 -17.7 4.68 19.4 0.513016 0.512963 7.74 34R-3,79-84 11.5 374 0.703320 0.703236 -17.0 4.88 20.3 0.513011 0.512958 7.67

35R-3,108-114 1.44 190 0.703208 0.703200 -17.5 2.04 7.08 0.513053 0.512990 8.26

37R-5,38-42 18.8 638 0.703204 0.703151 -18.2 6.58 27.4 0.513020 0.512967 7.83

38R-1,4-7 16.2 653 0.703204 0.703153 -18.2 5.76 26.2 0.513007 0.512958 7.66

41R-3,18-23 14.1 363 0.703285 0.703200 -17.5

42R-4,12-16 12.9 351 0.703331 0.703245 -16.9 5.32 20.6 0.513012 0.512955 7.59 TABLE 6.4 (CONTINUED)

87 86 87 86 143 144 143 144 ID Rb Sr Sr/ Sri Sr/ Srt ǫSr Sm Nd Nd/ Ndi Nd/ Ndt ǫNd

44R-2,116-121 21.9 309 0.703333 0.703233 -17.1 6.21 25.4 0.513005 0.512951 7.51 44R-2, 115-118 7.19 26.5 0.513003 0.512943 7.36 151 10 Hawaii (post-shield) Hawaii (shield) Nintoku (post-shield) 8

6 Nd ε 4

0 -20 -10 0 10

εSr

Figure 6.5. ǫSr vs. ǫNd showing Hawaiian and Emperor post-shield and shield basalts and the Nintoku post-shield basalts. The Nintoku post-shield basalts largely overlap post-shield basalts from Hawaiian volcanos, although ǫSr of the Nintoku samples extend towards the lower end of this range.

6.4 Discussion

6.4.1 Chemical and Isotopic Variations with Depth (Stratigraphic Age) in the Site 1205 Drill Core

There are signiﬁcant variations in lava composition that correlate with stratigraphic position in the Site 1205 drill core. With decreasing depth (decreasing age), incompatible element concentrations and ratios of more- to less-incompatible elements (e.g., La/Yb and Nb/Zr) generally increase, and initial 87Sr/86Sr decreases (Fig. 6.7). Epsilon-Nd shows no consistent correlation with depth. Simi- lar compositional variations with age have been reported for the Kula post-shield

152 0.7033

0.7032

0.7032 i

r 0.7032 S 6 8 r S 7

8 0.7032

0.7032

Conlomerate Clast Samples 0.7031 Upper Series Alkalic Basalts Lower Series Alkalic Basalts Tholeiite Sample 0.7031 0 5 10 15 20

MgO (wt. %)

8.4

8.2

7.8 d N ε 7.6

7.4

7.2 Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts Tholeiite Sample 7 0 5 10 15 20

MgO (wt. %)

87 86 Figure 6.6. MgO (wt.%) vs. Sr/ Sri and ǫNd. Neither isotope composition varies systematically with degree of diﬀerentiation. lavas of Haleakala volcano [38], although the variation in Sr and Nd isotope ratios of the Nintoku lavas is less than half that observed in the Kula volcanics. Samples with the lowest 87Sr/86Sr values occur in the upper part of the Site 1205 drill core (see next section), tend to have higher incompatible trace element contents, and have higher La/Yb and Nb/Zr ratios (Fig. 6.8).

As on Mauna Kea, and in contrast to Haleakala, alkali basalt and their diﬀer- entiates (hawaiites and mugearites) are not intercalated in the post-shield stage of Nintoku Seamount, but instead the more evolved rocks appear to have formed a discontinuous cap on the alkali basalts. Although the more evolved rock types are

153 0 0

-50 -50

-100 -100 )fs )fs

b -150 b -150 m m ( htp ( htp (

-200 -200 e e D D

-250 -250

-300 -300

-350 -350 0 5 10 15 20 1 1.5 2 2.5 3 3.5

MgO (wt. %) TiO (wt. %) 2

0 0

-50 -50

-100 -100 )fs )fs

b -150 b -150 m m ( htp ( htp (

-200 -200 e e D D

-250 -250

-300 -300

-350 -350 10 20 30 40 50 60 70 4 6 8 10 12 14 16 18

Nb (ppm) La/Yb

0 0

-50 -50

-100 -100 )fs )fs

b -150 b -150 m m ( htp ( htp (

-200 -200 e e D D

-250 -250

-300 -300

-350 -350 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28 0.70312 0.70314 0.70316 0.70318 0.7032 0.70322 0.70324 0.70326

87 86 Nb/Zr Sr/ Sr I

Figure 6.7. Depth proﬁles from Hole 1205A. With decreasing depth (decreasing age), incompatible element concentrations tend to increase, ratios of more- to less-incompatible elements (e.g., La/Yb and Nb/Zr) increase, and initial 87Sr/86Sr decreases. present only as clasts within the conglomerate unit immediately above basement at Site 1205, the predominance of hawaiite and mugearite compositions suggest that the stratigraphically highest lavas units in this area were dominated by these rock types. The origin and petrogenetic signiﬁcance of the temporal variations in lava compositions are discussed further in the following sections.

6.4.2 Inﬂuence of Alteration of Chemistry

Many of the Site 1205 lava units have undergone subaerial weathering immediately after eruption and are highly altered. Flow tops are variably weathered

154 18 Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts 16 Tholeiitic Sample

12 b Y/a L 10

4 0.70312 0.70314 0.70316 0.70318 0.7032 0.70322 0.70324 0.70326

87Sr/86Sr I

0.24 Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts 0.22 Tholeiitic Sample

0.2 rZ/b 0.18 N

0.16

0.14

0.12 0.70312 0.70314 0.70316 0.70318 0.7032 0.70322 0.70324 0.70326

87Sr/86Sr I

Figure 6.8. Initial 87Sr/86Sr vs. La/Yb, Nb/Zr. Samples with the lowest 87 86 Sr/ Sri tend to have higher La/Yb and Nb/Zr suggesting a heterogeneous source area. and oxidized, and soil horizons separate many of the flows [236] (Fig. 6.1). The lava flow units subsequently underwent seawater alteration after subsidence of the summit of Nintoku Seamount below sea level. Secondary alteration minerals are abundant. For example, olivine is often completely replaced by iddingsite and calcite; glass in the groundmass and in flow margins is often replaced by clay minerals.

Vesicles are partially to completely ﬁlled with clay minerals, Fe-oxyhydroxides or zeolites, and calcite veins occur in basalts throughout the section. Loss on ignition (LOI) values range from 1.3 to 7.3%, reﬂecting the presence of secondary carbonate and water-bearing clay minerals in most of the samples, despite great care taken to sample only the freshest material. Given the abundance of secondary minerals, al-

155 5 Haleakala Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts 4 Tholeiitic Sample

3 5 O 2 P/ O 2 K 2

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4

P O (wt. %) 2 5

Figure 6.9. P2O5 (wt.%) vs. K2O/P2O5. The Nintoku samples tend to have lower and more variable K2O/P2O5 values than Haleakala samples, reﬂecting preferential loss of K2O during subaerial and submarine weathering. teration is likely to have aﬀected the concentrations of K, Na, and to a lesser extent

Ca, Mg, and P in the most altered samples. K2O/P2O5 ratios of the Site 1205 lavas are more variable, and generally lower, than fresh post-shield lavas from Haleakala (Fig. 6.9), probably reﬂecting preferential loss of K during alteration [84]. Few of the sample analyzed here have Ce/Pb, Rb/Ba, or Nb/U within range of fresh oceanic basalts (Fig. 6.10), indicating that Pb, U, Rb, and possibly Ba were mobile during alteration. The relatively high Ce/Pb and Nb/U ratios suggest that most samples have lost U and Pb (the highest measured Nb/U ratios indicate 65% of original U content removed). Therefore, noting the inﬂuence of secondary processes on the primary chemistry of the Site 1205 lavas, we use those elements that are immobile during alteration (REE and HFSE) to investigate the petrogenesis of the Nintoku basalts [22, 151]

156 140 Conglomerate Clast Samples Upper Series Alkalic Basalts Lower Series Alkalic Basalts Tholeiitic Sample 120

100 U /b 80 N

Range of unaltered 60 oceanic basalts

20 0 20 40 60 80 100

Ce/Pb

Figure 6.10. Ce/Pb vs. Nb/U. Box shows the typical range of fresh oceanic basalts. Nintoku samples extend to higher Ce/Pb and Nb/U suggesting that Pb and U were mobile during alteration.

6.4.3 Role of Fractional Crystallization and Crystal Accumulation

Hawaiites and mugearites are considered to be the products of fractional crystallization of clinopyroxene-rich cumulates from alkalic lavas at depth of ∼ 15 km [39, 88, 249]. The major element characteristics of the hawaiite and mugearite samples from Nintoku Seamount can be modeled by using methods adapted from

Albarede [3]. The hawaiite/mugearite samples deﬁne a fairly linear extension in MgO variation diagrams (Fig. 6.3) from the alkalic basalts from Core 6R-2 (the youngest igneous basement recovered), suggesting that the evolved samples formed by fractional crystallization of a parental melt with major element characteristics similar to Core 6R-2 alkalic basalts. After applying a least squares ﬁt to the major element versus MgO content of the hawaiite/mugearite and Core 6R-2 alkalic basalts, an estimate of the general parental melt was made. Using the concentrations of highly incompatible elements Nb, Th, and La from sample 5R-2, 90-94,

157 alkali basalt clast contained in the conglomerate with the hawaiite/mugearite samples that has major element abundances similar to the parental melt estimate, an estimate of the amount of fractional crystallization was made assuming that the bulk partition coeﬃcient, D, for these elements is approximately 0. If this is the case, then the standard Rayleigh fractional crystallization equation reduces to

CL = CO/F , where CO is the trace element concentration of 5R-2, 90-94 and CL is the average trace element concentration in the hawaiite/mugearite samples. Once an estimate of F is made, the major element composition of the cumulate that fractionated from the hawaiite/mugearite parental melt can also be estimated using simple mass balance approaches. Using a constrained linear least-squares method presented by Albarede [3], the cumulate composition was recast into proportions of mineral phases present as phenocrysts in the hawaiite/mugearite lavas. The Site 1205 lavas commonly contain olivine, plagioclase, clinopyroxene, and titanomagnetite as phenocryst phases. The theoretical cumulate composition can be represented as the vector y contained within the matrix A, which has 5 columns representing the cumulate and mineral (olivine, plagioclase, clinopyroxene, and Fe-Ti oxide) compositions and 11 rows for the 11 major elements modeled. The compositions of the mineral phases were taken from electron microprobe data collected on the phenocryst phases from the Nintoku hawaiite/mugearite samples during DSDP Leg 55 [42]. Manipulation of matrix A leads to the unconstrained solution, the vector x. The constrained solution (the proportions of mineral phases needed to reproduce the cumulate composition) can be calculated after recalcu- lating the unconstrained solution to force the mineral phase proportions to sum to unity. The fractional crystallization (37–58%) of a crystal assemblage composed of plagioclase, clinopyroxene, olivine, and Fe-Ti oxides in the proportions 46:30:14:10

158 50 4

49.5

49 3.5

48.5 2 2 O O iS 48 iT 3

47.5

47 2.5

46.5

46 2 01234567 0 1 2 3 4 5 6 7

MgO MgO

22 10

20 9

18 8 3 O O a 2 16 7 lA C

14 6

12 5

10 4 01234567 0 1 2 3 4 5 6 7

MgO MgO

6 4

3.5 5.5

5 2.5 O O 2

4.5 2 2 a K N

1.5 4

3.5 0.5

3 0 01234567 01234567

MgO MgO

Figure 6.11. MgO variation diagrams showing only the conglomerate clasts and the alkalic basalts from the uppermost igneous basement. Symbols are the same as in Figure 6.2. The + symbols are the hypothetical liquid line of descent if a cumulate with the composition given in Table 5 was removed from a liquid with a composition similar to the basalts from the uppermost igneous basement. can reproduce the trends seen in the hawaiite/mugearite samples, except for MnO (Fig. 6.11). Model parameters are presented in Table 6.5.

This model predicts significant plagioclase, clinopryoxene, and Fe-Ti oxide fractionation, the effects of which are only partially borne out in the trace element characteristics of the hawaiites/mugearites. Scandium is compatible in clinopyroxene [94, 101, 118] and the hawaiite/mugearite lavas have corresponding low Sc concentrations (< 15.5 ppm) that is consistent with significant clinopyrox-

159 ene fractionation. However, Sr and Eu concentrations in the hawaiite/mugearite lavas are not depleted, which is inconsistent with extensive fractionation of plagioclase. Certain characteristics of the Nintoku samples suggest mitigating factors.

The high Sr abundances are coupled with high P2O5 abundances, which suggest the presence of minor mineral phases such as apatite may be controlling the Sr budget in these samples. Petrography indicates that the dominate Fe-Ti oxide is titanomagnetite [236] which require relatively high oxygen fugacity conditions to crystallize (e.g, QFM). This would keep Eu in the 3+ oxidation state, thereby limiting the uptake of Eu into the fractionating plagioclase.

TABLE 6.5 MAJOR ELEMENT MODEL PARAMETERS

Initial1 Final Cumulate PLAG2 OL3 PYR4 OX5 Cum∗

SiO2 47.3 49.1 45.8 54.1 39.5 50.0 0.0 45.6

TiO2 2.59 3.18 2.10 0.0 0.0 1.67 15.3 2.00

Al2O3 16.4 19.0 14.3 29.0 0.15 2.77 1.43 14.4

Fe2O3 3.67 3.54 3.79 0.0 0.0 0.0 33.3 3.25 FeO 9.36 9.02 9.67 0.64 15.0 9.36 47.8 9.85

MnO 0.16 0.12 0.20 0.0 0.14 0.16 0.44 0.11 MgO 6.50 1.50 10.6 0.01 44.7 13.9 1.72 10.5

CaO 8.71 5.01 11.8 11.4 0.26 21.5 0.0 11.8

Na2O 3.74 5.54 2.26 4.78 0.0 0.39 0.0 2.32

K2O 0.97 3.01 -0.70 0.31 0.0 0.0 0.0 0.14

P2O5 0.60 1.03 0.26 0.0 0.0 0.0 0.0 0.0

160 TABLE 6.5 (CONTINUED)

Initial1 Final Cumulate PLAG2 OL3 PYR4 OX5 Cum∗

Modal 0.46 0.14 0.30 0.10 Proportions

1The “Initial” and “Final” compositions refer to the end-point compositions of the least- squares ﬁt of the Nintoku data presented in Figure 6.11. The “Cumulate” is the composition of the solid that needs to be removed to account for the major element variations in the Nintoku data assuming the fraction of liquid remaining at the Final composition is 0.45 (see text for justiﬁcation). This composition was recast into proportions of the four mineral phases listed above, which was then used to recalculate the cumulate composition, given as Cum∗ (see text for full explanation)

2Plagioclase

3Olivine

4Pyroxene

5Oxide

This model can be quantitatively tested by modeling the trace element variations between the hawaiite/mugearite and the alkali basalt clasts using the mineral proportions and percent fractional crystallization derived from the major element model. Figure 6.12 shows that the majority of incompatible trace elements can be adequately predicted using calculated mineral proportions, except for Sr, possible reasons for which were discussed above. Modeled concentrations for all elements except Sr, Y, and Yb are with 25% of measured concentrations for samples 5R-2, 21-25 and 5R-2, 96-100. For sample 5R-2, 31-34 only Sr, Y, Er, and Yb are outside this range. Samples 5R-2, 74-77 and 5R-2, 114-116 show worse agreement; the predicted HREE concentrations are systematically higher in these samples than

161 10 Conglomerate Clast Samples e c n a d n u b A d A

etalu 1 cla C /la utc A

0.1 Th Nb Ta La Ce Pr Sr Nd Zr Hf Sm EuGd Tb Dy Y Ho Er TmYb Lu

Figure 6.12. Major element model test ﬁgure. In this ﬁgure, the actual trace element compositions of the conglomerate clasts were divided by concentrations predicted for each sample by the major element model (see text). In general, agreement between actual and predicted concentrations was good, except for Sr. the actual concentrations, however, increasing clinopyroxene fractionation could lower the model HREE predictions.

6.4.4 Role of Melting Processes and Source Heterogeneity in Creating Variation

Within Nintoku Lavas

Major and trace element and isotopic compositions indicate that the Nintoku

Lavas cannot all be related simply by closed-system fractional crystallization from a single parental magma composition. At MgO contents of ∼ 7 wt.%, there is a considerable range in SiO2, TiO2, FeOT, MnO, CaO, P2O5, and K2O in alkali basalt samples (Fig. 6.3). To some extent, these major element variations could reﬂect variations in the degree and pressure of melting, such that the Nintoku lavas are derived from crystallization of a range of melt compositions. However, the sig- niﬁcant variation in Sr isotope composition, and the correlation of 87Sr/86Sr with

162 incompatible trace element ratios (Fig. 6.8) indicates that source heterogeneity was an important factor in creating variation with the Nintoku lavas. Mixing of two sources, or melts derived from them, followed by variable amount of fractional crystallization could account for the trace element and isotopic variations, although the scatter in plots such as Nb/Zr versus 87Sr/86Sr requires both end-members to be heterogeneous (Fig. 6.8). Most importantly, simple two- component mixing does not readily explain the systematic variations in trace element and isotopic composition with time that are observed in the lavas from Site 1205.

6.4.5 Model for the Formation of the Nintoku Post-Shield Basalts

Two characteristics of the Nintoku alkali basalts must be accounted for in any model of their genesis. First, there are systematic variations in incompatible trace element and Sr isotope ratios with age: 87Sr/86Sr ratios decrease and La/Yb, Nb/Zr ratios increase with decreasing age (Fig. 6.7). Second, lavas with the most enriched trace element compositions (highest La/Yb, Nb/Zr) were derived from a source with long-term depletion in highly incompatible elements (low 87Sr/86Sr). Similar trace element-isotope systematics characterize the transition from shield tholeiitic lavas through post-shield alkali lavas to rejuvenated stage alkalic lavas of many Hawaiian volcanoes [37], and have also been recognized within the post- shield alkalic lavas of Haleakala [38]. Several models have been proposed to account for the combination of trace element enriched lavas (high Rb/Sr, low Sm/Nd) with depleted isotope signatures (low 87Sr/86Sr, high 143Nd/144Nd) from Hawaii [36, 37, 87, 205, 259]. Very small (< 0.2%) degrees of melting of depleted mantle sources, or reactive melt transport

163 processes are likely to result in decoupling between trace element and isotopic ratios and non-linear variations between concentrations of incompatible elements (such as Nb and Zr), unlike those observed in Hawaiian post-erosional stage lavas [205]. Instead, many models attempting to explain the origin of Hawaiian alkali lavas invoke small degree melts of a source that was relatively recently (< 400 m.y. ago) enriched in incompatible trace elements. Chen and Frey [36, 37] proposed that partial melts of the Hawaiian plume are mixed with 1–4% of very small degree melts (0.1–1.0%) of a depleted source (oceanic lithosphere or the upper mantle source of mid-ocean-ridge basalts (MORB)). This multicomponent mixing model could account for much of the trace element and isotopic variations in lavas from Haleakala. Mixing of approximately constant proportions of the enriched source with variable degree melts of the depleted source will result in negative correlations in La/Ce or La/Yb vs. 87Sr/86Sr diagrams. A similar model, in which small degree melts of the MORB source mantle are added to the Hawaiian plume source, and this mixed source then melted to various degrees, has been used to explain the compositions of Koloa post-erosional lavas from Kauai [205] and the post-erosional Honolulu Volcanics of Oahu [215].

An alternative model involves mixing a depleted source with 0.5–2.0% of small degree melts (∼ 2%) from a more enriched source [41]. This is similar to metasomatic models in which small degree partial melts of the enriched source are added to the depleted oceanic lithosphere, which subsequently undergoes partial melting to form the parental magmas of Hawaiian lavas [257]. In this model, the melt with the greatest contribution from the low 87Sr/86Sr component must have been enriched by the smallest degree melts from the plume in order to obtain a negative

164 correlation between 87Sr/86Sr and La/Ce or Nb/Zr. Yang et al. [259] favored such a model to explain the compositions of the Honolulu Volcanics because recently determined partition coeﬃcients imply that extremely small degree melts (< 0.1%) of the depleted source would be required to explain the high La/Ce ratios of the Hawaiian post-erosional lavas when using the Chen and Frey [36, 37] model. An alternative possibility, explored more fully below, is that the low 87Sr/86Sr end- member (oceanic lithosphere) is more enriched in trace elements than assumed by Chen and Frey [37], perhaps by incorporation of small degree melts of MORB source soon after its formation at the MOR axis. Below, we evaluate which of these models can best explain the trace element and isotope compositions of the

Nintoku lavas.

Model 1—The Chen and Frey-type Model: Considering the compositional similarities between Nintoku and Haleakala post-shield lavas, we ﬁrst applied a multicomponent mixing model similar to that proposed by Chen and Frey [37]. Dis- tribution coeﬃcients and model parameters are given in Table 6.6. Figure 6.13a illustrates that the Nintoku lavas can be generated by mixing of 88.4–89.2% undepleted mantle (UD) and 11.6–10.8% low degree (0.79–1.93%) partial melts of

MORB source. The amount of incipient melt mixed with UD is greater at Nin- toku than for the Kula Series post-shield basalts from Haleakala (97–98% UD, 2–3% MORB source), and the amount of partial melting of the MORB source required to ﬁt the Nintoku data (0.79–1.93%) is higher than predicted for Haleakala

(0.2–0.5%) [37].

165 TABLE 6.6 RESERVOIR COMPOSITIONS (PPM) AND PARTITION COEFFICIENTS

Reservoir1 Rb Th Nb Sr Nd Sm Hf La Ce La/Ce 87Sr/86Sr

UD 0.73 0.09 0.86 23.7 1.2 0.39 0.31 0.71 1.9 0.374 0.7047

MS 0.11 0.02 0.31 13.2 0.86 0.32 0.25 0.31 0.95 0.326 0.7023

N-MORB 0.56 0.12 2.33 90.0 7.3 2.63 2.05 2.5 7.5 0.333 0.7023 166 RDS 0.61 0.14 2.51 92.2 7.42 2.64 2.06 2.66 7.73 0.345 0.7023

DML 0.041 0.005 0.11 7.0 0.49 0.31 0.17 0.11 0.34 0.326 0.7025

HP 0.85 0.11 1.4 35.9 2.2 0.69 0.53 1.4 3.4 0.407 0.7036

OL2 0.003 0.00001 0.00005 0.00005 0.0003 0.0009 0.04 0.0004 0.009

CPX 0.0004 0.0021 0.004 0.282 0.35 0.736 0.44 0.032 0.057 OPX 0.0002 0.00003 0.003 0.002 0.003 0.03 0.01 0.0005 0.001

GRT 0.0002 0.0021 0.01 0.0007 0.027 0.22 0.23 0.0164 0.065

PHL 1.7 0.14 0.044 0.0063 0.0059 0.003 0.021 TABLE 6.6 (CONTINUED)

Reservoir Rb Th Nb Sr Nd Sm Hf La Ce La/Ce 87Sr/86Sr

AMP 0.023 0.001 0.08 0.27 0.23 0.32 0.3 0.075 0.11

1UD, undepleted mantle [37]; MS, MORB source [37]; N-MORB, normal MORB [234], except that 87Sr/86Sr was left the same as in MS; FDS, fertilized depleted source, created by mixing 0.5% incipient melts of MS with N-MORB in 0.5:99.5 proportions (see text for full description); DML, depleted mantle lithosphere [259]; HP, Hawaiian plume [259] 167 2OL, olivine; CPX, clinopyroxene; OPX, orthopyroxene; GRT, garnet; PHL, phlogopite; AMP, amphibole. Mineral modes for calculation of the bulk partition coeﬃcients for the above reservoirs were adapted from those found in Chen and Frey [37], Yang et al. [259], and Halliday et al. [97]. The ol:opx:cpx:gt:phl:amp proportions in the source and melt for Models 1 and 2 are: Model 1—0.6:0.25:0.1:0.05:0:0 and 0.1:0.1:0.4:0.4:0:0; Model 2—0.55:0.25:0.15:0.05:0:0 and 0.1:0.1:0.4:0.4:0:0. Mineral proportions for MS, FDS, and melt for Model 3, respectively, are 0.706:0.2:0.05:0.02:0.002:0.02, 0.6:0.25:0.1:0.05:0:0, and 0.1:0.1:0.4:0.4:0:0. Partition coeﬃcients from: [16, 82, 96, 97, 115, 118, 121, 122, 126, 131, 146, 163, 168, 221, 227, 229, 232, 245, 260] 0.65 0.7 (a) (b) 0.6% 10% 20% 0.65 0.6 5.0% 1.0%

2.0% 10% 0.8% 0.6 0.55 5.0% 2.0% 0.55 0.5% 1.2% 0.5 3.0% e e C C / /a 0.5

a 4.0% L L 0.45 2.0% 1.0% 0.45 10% 0.1% 0.4 0.4 Hawaiian UD Plume 0.35 0.35 MORB Source DM Lithosphere 0.3 0.3 0.702 0.7025 0.703 0.7035 0.704 0.7045 0.705 0.7024 0.7026 0.7028 0.703 0.7032 0.7034 0.7036 0.7038

87 86 Sr/ Sr 87Sr/86Sr

0.65 (c)

20% 0.6 0.8% 10% 5.0% 3.0% 1.0% 0.55 1.0%

0.5 1.5% e C /a L 0.45 2.3%

0.4

UD 0.35 Re-fertilized Oceanic Lithosphere

MORB Source 0.3 0.702 0.7025 0.703 0.7035 0.704 0.7045 0.705

87Sr/86Sr

87 86 Figure 6.13. Sr/ Sri vs. La/Ce. (a) Model 1: UD is undepleted mantle (composition is given in Table 6.6). Curved horizontal lines are mixing lines between incipient melts of MORB source and UD, while vertical lines are lines of equal mixing proportions. (b) Model 2: Mixing of incipient melts of Hawaiian plume (composition given in Table 6.6) with DM. Notice that the trend of the data crosscut the lines of equal mixing proportions. (c) Model 3: Mixing of incipient melts of MORB source fertilized at time of formation by small degree melts at the mid-ocean ridge with UD. The degree of partial melting is more realistic than the extremely small degrees of melting required by Model 1.

168 Following Chen and Frey [37], four samples that have olivine as the only phenocryst phase were selected to quantitatively test the model. Since the model predicts only the hypothetical parental magma composition, it is necessary to correct the trace element concentrations of the selected samples for olivine fractionation in order to estimate the parental melt of each sample. Olivine addition dilutes incompatible trace element concentrations but does not radically alter trace element ratios. Equilibrium olivine was added in 1% increments until the major element composition of each sample was in equilibrium with Fo90 [37]. The Fe/Mg exchange coeﬃcient between olivine and melt was 0.30 [81]. This addition of olivine was then used to back-calculate the trace element abundances of the parental magma assuming that olivine was the only fractionating phase. Using the corrected La abundance, we calculated how much partial melting of the hybrid UD-MORB source was required to achieve the corrected La abundance of each sample. After partially melting the hybrid source by the amount calculated above, the model results were then compared to the olivine-fractionation-corrected trace element abundances. The results are presented in Table 6.7.

169 TABLE 6.7 TRACE ELEMENT MODEL RESULTS

10R-2 33R-2 34R-3 41R-3

0-5 70-75 79-84 18-23

Model 11 Model 22 Model 33 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

f%4 1.10 2.75 1.37 1.61 4.99 2.04 1.38 4.37 1.74 1.56 4.85 1.97

%x5 11.4 1.78 1.91 11.6 2.76 2.00 10.32 3.09 1.75 11.3 2.83 1.94 170 F%6 18.9 4.97 7.53 22.1 11.1 9.77 20.2 11.2 8.83 18.4 9.41 8.07

Rb 0.61 7 0.78 0.85 0.78 0.96 1.11 0.79 0.96 1.12 0.77 0.94 1.09

Th 0.80 0.79 0.99 0.85 0.85 1.08 0.88 0.86 1.11 0.93 0.93 1.18

Nb 0.73 0.74 0.75 0.71 0.72 0.74 0.75 0.75 0.78 0.89 0.91 0.93

Sr 0.91 0.56 0.96 0.90 0.54 0.93 0.96 0.60 1.01 1.20 0.71 1.22

Nd 0.92 0.58 0.89 0.98 0.61 0.93 0.92 0.59 0.90 0.88 0.53 0.83

Sm 0.85 0.60 0.80 0.93 0.65 0.88 0.86 0.61 0.83 0.77 0.52 0.71

Hf 0.93 0.60 0.91 0.96 0.61 0.94 0.95 0.63 0.94 0.97 0.61 0.92 TABLE 6.7 (CONTINUED)

10R-2 33R-2 34R-3 41R-3

0-5 70-75 79-84 18-23

Model 11 Model 22 Model 33 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

La 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Ce 0.98 0.85 0.99 0.98 0.83 0.99 0.98 0.85 1.00 0.98 0.82 0.97 171

1Adaptation of model presented in Chen and Frey [37]

2Adaptation of model presented in Yang et al. [259]

3Adaptation of Chen and Frey [37] model with enrichment of depleted source as suggested in Halliday et al. [97]

4Degree of incipient melting of MS in Model 1, HP in Model 2, and FDS in Model 3

5Amount of incipient melt in the hybrid source

6Degree of partial melting of the hybrid source

7Results are presented as the ratio of predicted abundances to olivine fractionation corrected abundances. Abundances that do not agree within 25% are underlined and in italics The agreement of the actual data with model predictions is generally good. In the following discussion of the model results, samples which show discrepancies of greater or less than 25% of the actual data are considered to be in poor agreement. Predicted Rb concentrations are systematically low (21-39%) while Nb predictions are low for two of the four samples quantitatively studies (by 27 and 29% for samples 10R-2, 0-5 and 33R-2, 70-75, respectively). Thorium predictions are within 7–20% agreement and Sr predictions are within 9%, except for sample 41R-2, 18- 23 (20%). Neodymium, Sm, and Hf all agree within 15%, many within 5% of the actual data. Overall, the model adequately predicts trace element concentrations.

Model 2—The Yang et al.-type Model: We next applied the model of Yang et al. [259] to the Nintoku data. In this model, the procedure is basically the same as that used by Chen and Frey [37], except incipient melts of Hawaiian plume source are mixed with depleted mantle lithosphere. The negative correlation between 87Sr/86Sr and La/Ce for the Nintoku data requires those samples created by the largest degree of melting to contain the smallest contribution from the low 87Sr/86Sr source (Fig. 6.13b). Mixing proportions vary from 96.1–98.5% MORB source and 3.9–1.5% incipient melts from the Hawaiian plume (at 1.84–7.15% partial melting). Quantitative testing was performed in the same manner as in Model 1. The full results of the model are presented in Table 6.7. The predicted Sr, Nd, Sm, and Hf concentrations were substantially lower than those measured in the lavas (> 29% lower). Predicted Rb concentrations agreed quite well with the actual data, although Rb concentrations are likely to have been aﬀected by alteration. Thorium and Nb also show close appoximation to the actual data. However, the poor reproducibility of the REE and Hf forces us to conclude that this model is not applicable to the petrogenesis of the Nintoku lavas.

172 Model 3—Modified Chen and Frey-type Model: The third model is a modified version of that proposed by Chen and Frey [37]. In this model, a more enriched composition of the depleted endmember is used. Undepleted mantle is used as the enriched endmember. There is some evidence, discussed more fully in the next section, that oceanic lithosphere may be enriched by small degree melts of MORB source soon after its formation at the ridge axis. In this model, oceanic lithosphere (N-MORB) is metasomatized by a 0.5% degree partial melt of MORB source (in 99.5:0.5 proportions, respectively) to create the “fertilized” depleted end-member. Melting of this source varies from 0.98–2.49% (as opposed to 0.79–1.93% in the Model 1) and mixing proportions vary from 1.79 to 2.03% fertilized depleted end- member and 98.21 to 97.97% Hawaiian plume source (Fig. 6.13c). The model was quantitatively tested in the same manner as the other two models and the results are presented in Table 6.7. Predicted trace element concentrations are significantly better than in the previous two models. Predicted Rb concentrations were all within 25% of the actual concentrations (all but one sample was within 8%), which is substantially better than Model 1 and similar to the predicted concentrations from Model 2. Niobium was consistently low (> 25% for two of the four samples), which is similar to Model 1. Neodymium, Sm, Hf, and Th all agree within 16% of the actual concentration, many within 5%. The degree of incipient melting of the depleted source is higher than Model 1 (1.78% partial melting on average vs. 1.41% for Model 1) and the proportion of incipient melt mixing with Hawaiian plume source is much less than Model 1 (∼ 2% incipient melt vs. ∼ 11% incipient melt in Model 1). As a result, smaller volumes of the oceanic lithosphere are required to melt. The amount of partial melting of the hybrid source is also less than in Model 1 (8.55% vs. 19.88% for Models 3

173 10 10 10R-2, 0-5 Model 1 33R-2, 70-75 Model 1 Model 2 Model 2 Model 3 Model 3 n n oitartn oitartn e e c c n n o o C C la la 1 1 utc utc A/d A/d etcid etcid er er P P

0.1 0.1 Rb Th Nb Sr Nd Sm Hf La Ce Rb Th Nb Sr Nd Sm Hf La Ce

10 10 34R-3, 79-84 Model 1 41R-3, 18-23 Model 1 Model 2 Model 2 Model 3 Model 3 n n oitartn oitartn e e c c n n o o C C la la 1 1 utc utc A/d A/d etcid etcid er er P P

0.1 0.1 Rb Th Nb Sr Nd Sm Hf La Ce Rb Th Nb Sr Nd Sm Hf La Ce

Figure 6.14. Graphical representation of the results of the three models presented in Figure 6.13. Overall, Model 3 more accurately predicts the concentrations of the elements modeled. and 1, respectively). The results of the three models are summarized in Figure 6.14. The chemical and isotopic variation with time recorded in the Site 1205 lavas indicate that an increasing proportion of the incompatible trace element enriched, low 87Sr/86Sr end-member contribued to melting over the period of time represented by the lavas, coupled with a lower degree of melting of this component. These temporal trends are perhaps the result of the drift of Nintoku Seamount away from the hottest part of the Hawaiian plume.

6.4.6 A Physical Model for Nintoku Post-Shield Magmatism

Chen and Frey [37] envisaged that Hawaiian post-shield lavas were formed by melts from the Hawaiian plume mixing with incipient melts of oceanic lithosphere,

174 which formed the wallrock of the plume conduit. The relative contribution from the radiogenic 87Sr/86Sr plume source decreased with time as the oceanic plate drifted away from the axis of the plume [37]. A role for ∼ 90 Ma old Paciﬁc lithosphere in the generation of Hawaiian alkalic lavas is supported by evidence fro residual phlogopite or amphibole in the source of the Honolulu Volcanics and North Arch lavas [44, 45, 83, 259]. These phases are unlikely to be stable in the convecting upper mantle, but may exist in the lower P-T conditions of the oceanic lithosphere [45]. The relatively radiogenic Os isotopic composition of many Hawaiian post-erosional lavas is inferred to have been inherited from pyroxenite veins within the ∼ 90 Ma old Paciﬁc oceanic lithospheric mantle beneath Hawaii [144]. Basu and Faggart [15] noted that compared to

MORB, Hawaiian post-shield lavas are oﬀset to higher ǫSr for a given ǫNd, and suggested that they contain a seawater-altered oceanic lithosphere component. Several previous studies have argued that the lowermost parts of the oceanic lithosphere may be enriched in incompatible trace elements soon after formation at the ridge axis by small degree melts of the depleted upper mantle [93, 97, 144]. This may then act as an important reservoir for incompatible trace elements if remobilized when later passing over a plume heat source [45, 97]. Ancient (recycled), fertilized oceanic lithosphere may represent a source for tholeiitic intraplate oceanic lavas [97, 181], including those from the Hawaiian hotspot [143]. We recognize that the modeling presented above is dependent upon the partition coeﬃcients, assumed end-member compositions, and modal mineral proportions of sources and crystallizing melts. These basic assumptions make the model predictions non-unique, however, the modeling does serve to show that fertilization of the oceanic lithosphere is a viable process for generating the post-shield

175 basalts of Nintoku seamount. In particular, fertilized oceanic lithosphere removes the need to invoke very small degrees of melting to account for the incompatible trace element enrichment of Hawaiian post-shield and post-erosional lavas. The fertilization model satisfies constraints imposed by Os isotopic compositions, which imply a significant age for the pyroxenite veins that contribute to post-shield magmatism [144] and complies with evidence for residual phlogopite/amphibole in the source of the Hawaiian post-shield lavas. Recently, Frey et al. [85] have suggested that the depleted (low 87Sr/86Sr) component in both the oldest Emperor Seamount lavas, and younger Hawaiian post erosional lavas is the same, and that it is intrinsic to the Hawaiian plume and distinct from the source of MORB. This model has some difficultly in explaining the incompatible trace element enriched compositions of Hawaiian post shield alkalic lavas, although it is possible that heat consumed by melting of the high 87Sr/86Sr “veins” inhibits melting of the more refractory matrix until the veins are exhausted, at which point they are more deplted in incompatible trace elements than the matrix.

6.4.7 Constraints from Sr Isotope Variations with Time on Petrogenesis of Hawaiian-

Emperor Post-Shield Lavas

Previous studies have show that ǫSr values of tholeiitic lavas decrease northwards along the ESC (Fig. 6.15). The origin of the depleted isotopic compositions of the older Emperor tholeiites is debated, but likely reﬂects the fact that the oldest Emperor Seamounts were formed close to a former spreading center (Fig. 6.15). This is supported by the major and trace element compositions of Detroit Seamount lavas indicate that they were derived by melting at a lower average

176 pressure than young Hawaiian lavas. Keller et al. [124] suggest that the depleted compositions of older Emperor lavas are the result of plume-ridge interaction, such that the Hawaiian plume entrained depleted upper mantle and young, hot lithosphere. Alternatively, the depleted compositions of these of these lavas may result from melting of a depleted, refractory mantle component that contribute to melting only when the Hawaiian plume is situated beneath young, thin lithosphere where decompression melting extended to shallower depths [85, 112, 201].

10 Post-shield Shield

-10 Sr ε

-20 u Nintok

-30 6000 5000 4000 3000 2000 1000 0

Distance from Kilauea (km)

Sr -10 ε

-20 oku Nint

-30 0 20 40 60 80 100

Age of crust at time of seamount formation (Ma)

Figure 6.15. Distance from Kilauea (km) vs. ǫSr and age of crust at time of seamount formation vs. ǫSr. Shield and post-shield data for Hawaiian volanoes and ESC seamounts are from: [15, 38, 111, 112, 125, 201].

177 Together with other Leg 197 data [112], our new results show that ǫSr values of post-shield lavas also decrease northwards along the ESC (Fig. 6.15). Alkalic lavas from ODP Site 883 and 1204 on Detroit Seamount, which are considered to represent post-shield lavas [112] have lower ǫSr than any post-shield or post- erosional lava reported from the Hawaiian Islands. Two samples of alkali basalt from Suiko Seamount have ǫSr values that are intermediate between those of post- shield alkalic lavas from the Hawaiian Islands and from Detroit Seamount [137]. Epsilon-Sr values of most Nintoku lavas are wtihin the range of Hawaiian post- shield lavas, but lie at the depleted end of this range (Figs. 6.5 and 6.15).

The lower ǫSr values of older Emperor post-shield lavas could reﬂect the fact that ǫSr values of the plume-derived end-member was lower when the Hawaiian plume was situated close to a ridge axis, so that plume-lithosphere melt mixtures from which the alkali basalt were derived also had less radiogenic Sr. However, in this case, tholeiitic lavas would be expected to have a greater range in ǫSr, and to show a more pronouned decrease in ǫSr with distance northward along the ESC. The available data suggest that this is not the case (Fig. 6.15). We can rule out the possibility that post-shield lavas from the older Emperor Seamounts contain a larger contribution from the low 87Sr/86Sr, incompatible trace element enriched, lithospheric end-member, because this does not explain the low incompatible trace element concentrations and low La/Ce, Nb/Zr in alkalic lavas from Detroit, compared to young Hawaiian post-shield lavas. Similarity, ingrowth of 87Sr in old, metasomatized lithosphere with high Rb/Sr does not explain the low incompatible element concentrations of the Detroit lavas.

One possible explanation for the observed temporal variations in ǫSr of post- shield lavas is that younger oceanic lithosphere has lower incompatible trace el-

178 ement concentrations and lower 87Sr/86Sr. If the oceanic lithosphere is fertilized over periods of several tens of m.y. [180, 181], by small degree melts with higher 87Sr/86Sr than normal depleted upper mantle, then the 87Sr/86Sr and incompatible trace element content of the bulk oceanic lithosphere will increase with time, and post-shield lavas from seamounts formed on older crust will inherit higher ǫSr and La/Ce. Alternatively, a higher degree of melting of the low 87Sr/86Sr, lithospheric end- member might also explain the lower ǫSr, La/Ce, and incompatible trace element concentrations of the older Emperor post-shield lavas, if more deplted components, with lower 87Sr/86Sr, have higher melting temperatures. Increased decompression melting might be expected in the lower portions of young, thin oceanic lithosphere which would be uplifted to a greater extend above the Hawaiian plume. These two hypotheses could perhaps be tested by comparing compositions of post-shield stage lavas along the ESC, although we note that post-shield lavas on the Hawaiian

Islands are highly variable.

6.5 Conclusions

1. Alkali basalts and their diﬀerentiates from Nintoku Seamount have similar

major and trace element compositions to lavas from the post-shield stage of young volcanoes from the Hawaiian Islands. Least-squares modeling indicates that fractional crystallization of olivine, plagioclase, clinopyroxene, and Fe-Ti oxides can account for much of the chemical variation within the

lava sequence.

179 2. A modiﬁed version of the Chen and Frey [37] model for the origin of Hawaiian

post-shield lavas, in which melts of the Paciﬁc oceanic lithosphere, previously metasomatized by small degree upper mantle melts, are mixed with melts of Hawaiian plume mantle is proposed to explain the incompatible trace element and isotopic compositions of the Nintoku lavas, without the need to

invoke very small degrees of melting.

3. Sr and Nd isotope compositions of post-shield lavas from Nintoku Seamount are within the range of post-shield lavas from the Hawaiian Islands, but lie

at the depleted end of this range. ǫSr values of post-shield, as well as shield lavas decrease northwards along the ESC. This could reﬂect either increased

melting of the low 87Sr/86Sr end-member beneath volcanoes situated on younger oceanic lithosphere, or less incompatible trace element enrichment (metasomatism) in the lower portions of younger oceanic lithosphere.

180 CHAPTER 7

THE PLATINUM GROUP ELEMENT AND RE–OS ISOTOPIC COMPOSITION OF THE EMPEROR SEAMOUNT CHAIN

7.1 Introduction

The intraplate Hawaiian Ridge-Emperor Seamount Chain (HR-ESC) is an approximately 6,000 km long, age-progressive chain of volcanic islands and seamounts that has been created as the Paciﬁc plate moved over the Hawaiian hotspot (Fig- ure 3.2). Active volcanism occurs on the Hawaiian Islands and the volcanoes become progressively older to the west and north. The oldest preserved volcanoes, Meiji and Detroit Seamounts, have an age of approximately 76–85 Ma [72, 123, 201]. Lavas from the ESC thus preserve a record of Hawaiian hotspot geochemistry since the Late Cretaceous. Recent studies have shown that the geochemistry of lavas from the ESC vary systematically from one end to the other. Tholeiitic and alkalic lavas from Meiji and Detroit Seamounts are depleted in highly incompatible trace elements compared to their younger counterparts from the Hawaiian Islands [124, 201]. Initial 87Sr/86Sr ratios of tholeiitic and alkalic lavas from Detroit and Meiji Seamounts extend to lower values than lavas from younger Hawaiian Ridge seamounts [112, 124, 201]. In contrast, the available geochemistry data for lavas from younger Emperor Seamounts (between Suiko and Daikakuji, 65–42 Ma) show that these

181 seamounts have major and trace element compositions similar to those of young

Hawaiian lavas [18, 41, 43, 59, 60, 124, 130, 201]. However, with the exception of Suiko Seamount, samples from the younger Emperor Seamounts are limited to a few dredge and shallow drill sites. Dredged samples in particular are generally highly altered, so the primary geochemistry of lavas from most of these sites is poorly known. The primary goal of this chapter is to determine if the PGE content of the ESC varies systematically in a similar manner to the ESC trace element and isotopic composition. To do so we have analyzed picritic and high-MgO basalts from three Emperor Seamount (Detroit, Nintoku, and Koko) for PGE abundances and

187Os/188Os isotopic ratios.

7.1.1 Detroit Seamount, ODP Leg 197, Site 1203

Detroit Seamount is centered at 51 ◦N, 167 ◦E in the northern part of the ESC (Figure 3.2) between 2,400 and 3,000 mbsl (Figure 7.1). Detroit was previously drilled during Ocean Drilling Program Leg 145 (Sites 883 and 884). At

Site 883, only alkalic basalts were recovered that are probably post-shield lavas. Tholeiitic lavas were recovered from Site 884, however, only 84 m of basement was recovered. 40Ar/39Ar age dating of Site 884 tholeiites yielded an age of 81 Ma [123]. Detroit seamount is similar in size to the subaerial part of the big islands of Hawaii and probably represents several coalesced volcanic shields [72, 112]. Hole 1203A is located in a valley on the eastern ﬂank of Detroit Seamount and penetrated 452.6 m into the volcanic basement. At least 18 lava ﬂow units were recovered and 40Ar/39Ar dating yielded an age of 75.8 ± 0.6(2σ) Ma [72]. Lavas

182 51°36' N

51°00' Site

160° 180° 200° 220°

60°

Site 50°

Seamount

40°

° 50 00' ° 168°42'E 167°00' 30

20°

10°

Figure 7.1. Bathymetry of the Detroit Seamount region. Adapted from Tarduno et al. [236]

from Hole 1203A are aphyric to olivine- or olivine/plagioclase-phryic basalt with trace clinopyroxene phenocrysts [236].

7.1.2 Nintoku Seamount, ODP Leg 197, Site 1205

Nintoku Seamount is located at 41 ◦N, 175 ◦E in the central part of the ESC (Figure 7.2). The upper 32 m of the volcanic basement was previously drilled during Deep Sea Drilling Project (DSDP) Leg 55, Site 432, where three ﬂow units of alkali basalt were recovered [130] and dated by 40Ar/39Ar to be ∼ 56 Ma [61]. More recent 40Ar/39Ar age dating from lava ﬂows at Site 1205 have yielded a mean

183 42° N

Site 432 Site 1205

41°

40°

39°

168°E 169° 170° 171° 172°

Figure 7.2. Bathymetry of the Nintoku Seamount region. Adapted from Tarduno et al. [236]

age of 55.6 ± 0.2 Ma [72]. Hole 1205 is located on the north-west edge of the summit region of Nintoku

Seamount at 1310 mbsl. 282.3 m of volcanic basement was recovered, comprised of twenty-five lava flow units, including both a’a and pahoehoe flow types. Lavas vary from aphyric to plagioclase-phyric to plagioclase-olivine-phyric basalt. Ground- mass is typically composed of plagioclase, olivine, clinopyroxene, titanomagnetite, and glass. The basalts are predominantly alkalic [224], with only one tholeiitic picrite flow unit.

184 37° N 4000 1000 5000

2000 5000 5000

200

36° 4000 4000

5000 3000

1000 2 000

5000

20 0 0 5000 LEE8-76-NP 4000

10 Site 308 3 0 35° 00 00 Site 1206

5000 4000 5000 Site 309 3000 3000 4000 4000 2000

00 20 3000 5000

5000 34°

3000 2000

170°E 171° 172° 173° 174°

Figure 7.3. Bathymetry of the Koko Seamount region. Adapted from Tarduno et al. [236]

7.1.3 Koko Seamount, ODP Leg 197, Site 1206

Koko Seamount is located at at 35 ◦N, 173 ◦E ∼ 200 km north of the bend in the HR-ESC (Fig. 3.2). Koko Seamount is most likely a plexus of coalesced volcanoes, however it is covered by a thick carbonate cap that precludes identiﬁcation of distinct volcanic centers [236] (Fig. 7.3). Volcanic basement was encountered at 57 mbsf at Hole 1206A and penetrated 278 m, of which 144 m was recovered. This marked the ﬁrst time that Koko Seamount had been drilled, although dredge samples were recovered during the

Thomas Washington expedition Aries 7 [63]. The basement recovered from Hole 1206A was divided into 22 lithologic units, predominantly olivine-phyric to aphyric tholeiitic basalt, although in two units alkalic basalt was recovered.

185 7.1.4 PGEs and the Hawaiian Plume

Early work on the PGE systematics of the Hawaiian plume focused on the 187Os/188Os composition of the plume. Many workers [103, 105, 142, 216] have proposed that the elevated 187Os/188Os ratios of Hawaiian lavas was due to recycling of oceanic crust, a hypothesis that had already been suggested on the basis of other isotopic systems. During mantle melting, Re is strongly partitioned into the melt while Os is retained in the mantle residue. Because of the large diﬀerence in Dliquid/solid between Re and Os, melting of MORB source to form oceanic crust leads high Re/Os ratios in the crust and, over time, suprachondritic 187Os/188Os ratios. Other isotopic systems (i.e. Pb and O isotopes) had already indicated that there may be a recycled oceanic crust + pelagic sediments component in the Hawaiian plume. It was generally accepted that the elevated 187Os/188Os ratios supported this idea. However, Walker et al. [248] suggested that the elevated 187Os/188Os ratios of the Hawaiian plume (and other mantle plumes) could better be modeled by incorporation of a small percentage of outer core material (1–3%) and suggested that examination of the 186Os/188Os isotopic ratio could provide further evidence of this interaction. Soon after, Bennett et al. [19] demonstrated negatively correlated Os–Pb isotopic ratios from Hawaiian picrites. The linear arrays suggested binary-mixing, but mixing of Hawaiian plume source and recycled oceanic crust did not adequately predict the arrays. Bennett et al. [19] suggested that a better ﬁt would be achieved by mixing between two distinct mantle components within in the Hawaiian plume. As analytical techniques improved for measur- ing the 186Os/188Os in primitive magmas, it became possible to precisely measure the 186Os/188Os in Hawaiian primitive basalts. Brandon et al. [25, 26] reported

186 coupled enrichments of 186Os/188Os and 187Os/188Os in Hawaiian picrites and proposed that the Hawaiian plume originated at the core-mantle boundary and incorporated a small amount of outer core material. In addition, they demonstrated that oceanic crust, while having a relatively high Pt/Os ratio, simply did not have enough Pt to be able to substantially change the 186Os/188Os composition of the

Hawaiian plume. Calculations showed that the plume would have to be composed of 70–90% oceanic crust, which would make the resulting melts tonalitic rather than picritic. The calculations made by Brandon et al. [25] did not take into account any pelagic sediment that could contribute to the overall 186Os/188Os composition of the recycled oceanic crust. The now extinct 182Hf–182W isotopic system was proposed as a new test of core-mantle interaction because W is siderophile and would have preferentially partitioned into the core. Hf, on the other hand, is lithophile and would have remained in the silicate Earth. Formation of the core would have fractionated Hf from W, and due to the very small amounts of Hf in the core, the outer core would develop highly unradiogenic 182W/184W. Any incorporation of outer core material should have imparted the negative ǫW signature of the core. Tungsten isotope data from Hawaiian samples indicated that the Hawaiian samples did not have negative ǫW as would be predicted from the core contribution model and instead were broadly similar to silicate Earth [219]. Schersten et al. [219] proposed that the elevated 186Os/188Os signature of the Hawaiian plume that is inadequately accounted for by recycling of oceanic crust, could be achieved by the addition of Mn-nodules being subducted along with other pelagic sediments.

Approximately 2–4% contamination of these Mn-nodules would be enough to elevate the 186Os/188Os ratio to observed levels; in addition, the model predicts a

187 5–9 times Mn increase in the mantle source and a decrease in the overall plume

Fe/Mn ratio [219]. While the Mn-nodule hypothesis predicts an overall lowering of the plume source Fe/Mn ratio, the core contribution model predicts an excess of Fe in the Hawaiian plume. However, major element compositions (such as Fe and Mn) are subject to significant modification due to differences in melting and eruption conditions. Despite this problem, Humayan et al. [113] presented detailed and precise measurements of the Fe/Mn ratio of Hawaiian lavas. Their data revealed a 1 to 2% increase in the mole fraction of Fe beneath Hawaii that corresponds to a density anomaly of approximately 0.5%, which is about the same magnitude observed by seismic tomography around Hawaii. The gravity data seem to rule out any contribution from Mn-nodules and Humayan et al. [113] propose that the nature of the interaction between the lower mantle and outer core is in the form of Fe-FeS metasomatic liquids. These fluids would carry the suprachondritic

Os isotopic signature, but since W has a lower preference for solid metal than Os, W and Os could be decoupled from each other. Fe-FeS metasomatism of the lower mantle by ﬂuids from the outer core could account for the silicate earth ǫW composition of the Hawaiian plume.

No matter which model is used to account for the suprachondritic Os isotopic compositions of Hawaiian magmas, each model predicts consequences for the PGE abundances of the plume. PGE abundances are strongly influenced by melting conditions, especially the presence or absence of sulfide as the PGEs are partitioned strongly into any sulfide phase. If sulfides are retained in the plume source during melting, the overall PGE abundances of the resulting melt will be substantially less than in the original source material. If sulfide becomes immisci-

188 ble during fractional crystallization, the fractionating sulﬁde phases will scavenge the PGEs from the melt and dramatically lower the overall PGE content of the remaining melt. Bennett et al. [20] measured the PGE and Re composition of picritic lavas from the Hawaiian Islands and found that the PGE compositions are consistent with melting of a source under varying degrees of residual sulﬁde.

Their data indicated that addition of Re and PGEs through the contamination of the plume source with outer core material was not needed to explain their data, but did not rule out that possibility.

7.2 Analytical Techniques

Detailed descriptions of the techniques used to prepare the sample powders, acquire LOI, major element, trace element, and Os- isotopic ratios can be found in Chapter 9.

7.3 Results

7.3.1 Petrography

7.3.1.1 Detroit Seamount

Samples 1203 32R-3 64-72, 1203 37R-3 80-89, and 1203 42R-5 28-37 are olivine- phyric basalts while sample 1203 59R-2 74-80 is aphyric. Olivine phenocrysts range from 0.5 mm to 4 mm in diameter. Sample 1203 42R-5 28-37 contains small (< 1 mm) plagioclase phenocrysts in addition to olivine phenocrysts. The groundmass in these samples is typically ﬁne-grained and composed of interlocking laths of plagioclase and clinopyroxene. In samples 1203 37R-3 80-89 and 1203 59R- 2 74-80, clinopyroxene in the groundmass tends to form in a poikilitic texture with

189 plagioclase. This textural evidence suggests that clinopyroxene crystallization was fairly late. Alteration tends to be conﬁned to degradation of olivine to brown clays and serpentine, although samples 1203 32R-3 64-72 and 1203 59R-2 74-80 have some alteration products in the groundmass (probably altered clinopyroxene). We attempted to minimize the inclusion of the alteration products in the powders by careful sample preparation. Please see Chapter 9.

7.3.1.2 Nintoku Seamount

Sample 1205 35R-4 104-110 is an olivine-phyric basalt. The olivine phenocrysts are heavily altered. The groundmass is ﬁne-grained and composed of plagioclase laths and clinopyroxene. The clinopyroxene in the groundmass is variably altered. Isotropic green clay minerals (probably smectite groups clays such as saponite and nontronite) are abundant in thin section and give the section a greenish tint. In thin section, very little to no fresh olivine is present. Olivine is typically completely replaced by clays and serpentine. As with the Detroit Seamount samples, we attempted to minimize the inclusion of the alteration products in the powders by careful sample preparation.

7.3.1.3 Koko Seamount

Samples 1206 04R-3 73-80 and 1206 18R-1 51-54 are olivine-phyric basalts with olivine typically being less than 10% of the mode. Samples 1206 38R-1 17-25 and 1206 39R-3 81-90 are highly olivine-phyric basalts where olivine is typically greater than 40% of the mode. Sample 1206 15R-1 7-10 is nearly aphyric and has only the occasional olivine phenocryst. The groundmass in all samples is composed of interlocking plagioclase and clinopyroxene laths and signiﬁcant proportion of

190 opaque oxides. Sample 1206 04R-3 73-80 is the only sample where the groundmass crystals are visible to the unaided eye; all the rest of the samples are aphaneritic. Alteration is often visible in hand sample as very little fresh olivine is observed in these samples. In thin section, olivine is often completely replaced by brown clay minerals and serpentine. Larger clinopyroxene crystals are also often highly altered. Sample 1206 38R-1 17-25 contained some fresh olivine that is typically mantled by serpentine. Rims of a reddish mineral (perhaps Fe oxyhydroxide) occasionally encircle olivine phenocysts as well. Sample 1206 39R-3 81-90 has veins of calcite permeating the sample. In general, the level of alteration is greater in the Koko samples than the other ESC samples. However, as with the Detroit and Nintoku Seamount samples, we attempted to minimize the inclusion of the alteration products in the powders by careful sample preparation.

7.3.2 Major and Trace Elements

Major element data and loss on ignition (LOI) values for Emperor Seamount lavas are presented in Table 7.1. Trace element data are presented in Table 7.2.

191 TABLE 7.1 MAJOR ELEMENT CONCENTRATIONS AND LOI VALUES (ALL VALUES IN WT.%) OF ESC SAMPLES

Site 1203 1203 1203 1203 1205 1206 1206 1206 1206 1206 BIR-1

Core 32R-3 37R-3 42R-5 59R-2 35R-4 04R-3 15R-1 18R-1 38R-1 39R-3

Interval 64–72 80–89 28–37 74–80 104–110 73–80 7–10 51–54 17–25 81-90

SiO2 45.6 43.2 45.2 46.0 44.1 46.9 47.6 45.0 46.8 46.7 47.4

Al2O3 12.9 7.5 15.1 14.7 14.2 12.1 13.3 9.3 10.4 9.65 15.1

TiO2 0.97 0.71 1.43 1.65 1.34 1.94 2.33 1.68 1.76 1.76 0.95

MnO 0.17 0.20 0.15 0.15 0.14 0.20 0.17 0.19 0.17 0.20 0.20

MgO 13.7 24.1 9.04 7.44 9.20 10.4 7.85 17.3 15.7 15.9 9.27

Fe2O3 10.2 12.4 10.1 11.6 12.6 13.0 12.6 12.9 13.0 12.9 11.2

CaO 10.8 5.31 10.5 11.1 8.24 8.03 10.5 7.01 7.50 7.32 13.1

Na2O 2.06 1.29 2.17 2.54 2.30 3.17 2.39 1.32 1.75 2.08 2.61

K2O 0.14 0.07 0.05 0.08 0.17 0.18 0.26 0.15 0.20 0.21 0.02

P2O5 0.11 0.08 0.12 0.12 0.07 0.24 0.26 0.19 0.20 0.21 0.04

LOI 2.43 1.97 4.91 3.75 6.78 2.35 3.22 3.67 1.89 1.17

Total+LOI 99.1 96.9 98.7 99.1 99.2 98.4 100.5 98.7 99.3 98.0 99.9

192 TABLE 7.2 TRACE ELEMENT ABUNDANCES (ALL VALUES ARE IN PPM) OF ESC SAMPLES

Site 1203 1203 1203 1203 1205 1206 1206 1206 1206 1206 BHVO-2

Core 32R-3 37R-3 42R-5 59R-2 35R-4 04R-3 15R-1 18R-1 38R-1 39R-3

Interval 64–72 80–89 28–37 74–80 104–110 73–80 7–10 51–54 17–25 81-90

Li 10.1 10.3 16.3 10.6 3.45 10.6 8.4 14.4 7.78 6.09 5.01

Be 0.27 0.22 0.48 0.52 0.40 0.87 0.71 0.53 0.58 0.59 0.08

Sc 35.8 26.9 40.0 41.6 26.9 31.4 31.4 28.1 27.5 26.7 48.3

V 171 159 286 299 200 271 245 202 211 211 340

Cr 686 1734 475 337 515 415 339 851 756 904 416

Ni 335 1086 244 139 295 339 163 844 732 740 204

Cu 77.0 77.7 106 143 113 102 154 110 82.4 90.8 151

Zn 75.5 98.2 106 109 129 148 143 131 139 146 93.7

Ga 16.2 11.8 21.7 23.0 24.3 23.8 24.2 16.9 19.0 19.8 20.0

Rb 2.53 1.30 0.44 0.42 1.07 3.00 3.93 1.73 1.55 2.46 0.22

Sr 205 113 239 265 297 306 405 220 248 257 126

Y 16.7 13.5 24.4 28.7 20.5 31.8 31.9 19.3 24.8 25.8 17.3

Zr 53.8 39.1 83.2 96.5 64.8 135 135 95.6 107 113 15.2

Nb 3.79 2.12 5.88 6.42 8.30 16.5 13.8 8.91 10.6 11.1 0.60

Mo 0.51 0.37 0.87 0.63 0.43 0.81 0.59 0.29 0.31 0.33 0.18

Sn 0.36 0.32 0.81 0.89 0.69 1.43 1.45 0.93 1.21 1.21 0.85

Sb bdl1 bdl 0.01 0.01 bdl 0.03 0.01 0.01 0.01 0.01 0.70

Cs bdl bdl bdl bdl bdl 0.10 0.02 0.05 0.01 0.03 bdl

193 TABLE 7.2 (CONTINUED)

Site 1203 1203 1203 1203 1205 1206 1206 1206 1206 1206 BHVO-2

Core 32R-3 37R-3 42R-5 59R-2 35R-4 04R-3 15R-1 18R-1 38R-1 39R-3

Interval 64–72 80–89 28–37 74–80 104–110 73–80 7–10 51–54 17–25 81-90

Ba 26.4 13.4 16.9 35.6 83.1 96.1 81.5 40.7 72.2 76.0 7.74

La 4.03 2.31 5.99 6.66 8.06 15.4 13.2 8.26 9.97 10.5 0.72

Ce 9.93 6.52 15.7 17.6 16.7 30.2 29.9 20.8 23.1 22.9 2.29

Pr 1.52 1.06 2.47 2.79 2.43 4.53 4.76 3.26 3.58 3.56 0.46

Nd 6.74 4.56 10.4 12.1 9.55 17.7 19.7 13.5 14.6 14.4 2.49

Sm 2.01 1.42 3.11 3.54 2.66 4.76 5.23 3.62 4.03 3.92 1.13

Eu 0.83 0.57 1.25 1.41 1.12 1.68 1.92 1.30 1.46 1.43 0.55

Gd 2.57 1.94 3.95 4.63 3.50 5.60 5.99 4.04 4.84 4.70 1.99

Tb 0.43 0.33 0.67 0.78 0.58 0.91 0.91 0.62 0.75 0.73 0.39

Dy 2.60 2.12 5.24 4.92 3.51 5.29 5.48 3.58 4.38 4.32 2.76

Ho 0.56 0.44 0.88 1.03 0.72 1.07 1.04 0.68 0.85 0.85 0.63

Er 1.53 1.26 2.47 2.90 1.85 2.89 2.80 1.83 2.27 2.27 1.92

Tm 0.22 0.18 0.37 0.42 0.26 0.41 0.38 0.26 0.33 0.32 0.29

Yb 1.24 1.13 2.09 2.42 1.46 2.26 2.10 1.37 1.83 1.78 1.78

Lu 0.16 0.14 0.28 0.34 0.18 0.30 0.28 0.17 0.24 0.23 0.24

Hf 1.32 1.01 2.28 2.65 1.80 3.65 3.89 2.60 3.01 2.91 0.58

W bdl bdl 0.02 0.01 0.03 0.15 0.06 0.06 0.06 0.07 0.01

Pb 0.13 0.17 0.36 0.44 0.39 0.85 0.74 0.44 0.54 0.54 4.03

Th 0.22 0.11 0.34 0.37 0.55 0.99 0.71 0.49 0.60 0.57 0.02

194 TABLE 7.2 (CONTINUED)

Site 1203 1203 1203 1203 1205 1206 1206 1206 1206 1206 BHVO-2

Core 32R-3 37R-3 42R-5 59R-2 35R-4 04R-3 15R-1 18R-1 38R-1 39R-3

Interval 64–72 80–89 28–37 74–80 104–110 73–80 7–10 51–54 17–25 81-90

U 0.07 0.03 0.12 0.14 0.13 0.31 0.22 0.14 0.16 0.19 0.01

1Below detection limit

The major element totals plus LOI vary from 96.9 to 100.5 wt.%. All samples except 1206 04R-3 73–80 and 1205 35R-4 104–110 are tholeiitic (Fig 7.4). The samples vary from nearly fresh (< 1 wt.% LOI) to moderately altered (∼ 5+ wt.%

LOI). MgO contents vary from 7.64 to 24.1 wt.%. These values represent the most magnesian samples available from each site. Subaerial and aqueous alteration (based on visual inspection and LOI values) has likely affected the concentrations of fluid mobile major and trace elements. The Sr abundances of the Site 1203, 1205 and one of the 1206 lavas show a distinct spike (Fig. 7.5). This is likely due to low-temperature seawater interaction and the fluid-mobile nature of Sr [102, 164].

7.3.2.1 Detroit Seamount (∼ 81 Ma)

The four samples analyzed from Detroit Seamount (ODP Leg 197, Site 1203) are tholeiitic basalts with MgO contents between 7.64 and 24.1 wt.%. On a SiO2 vs. total alkalis (Na2O+K2O) diagram (Fig. 7.4), one sample (1203 37R-3 80–89) plot within the ﬁeld for picritic basalts. The other three plot within the basalt

195 ed(i.74.TeDtottoeie aelwrconcentr lower have tholeiites Detroit The 7.4). (Fig. ﬁeld it ewe aainIlnstoeie frsmlswit samples (for tholeiites r Islands La/Yb Hawaiian between 7.7a). diate (Fig. range (MORB) basalt ridge mid-ocean the oe aY ais ocnrtoso hadN i ttedep with the samples at (for lie tholeiites Nb Islands and Hawaiian Th wh of Concentrations Islands, Hawaiian ratios. the La/Yb [112, lower from papers those than previous higher in are reported concentrationselements those The to similar 7.5a). are (Fig. study wt.% 10 than la greater tholeiitic dances Island Hawaiian and lavas ESC younger the than Al OB(i.7.7b). (Fig. MORB P 2 2 O O ocnrtoso ihyicmail rc lmnsi De in elements trace incompatible highly of Concentrations 5 iue74 SiO 7.4. Figure 3 n TiO and , h ahdln eaae aaintoeii n lai l alkalic and tholeiitic Hawaiian separates line dashed the n a Fg 7.6). (Fig. CaO and 2

hnKk emutwiecnann rae concentration greater containing while Seamount Koko than Na2O + K2O (wt. %) oktp ed r rmLeBas from are fields type Rock 2 w.)v.ttlakls(Na alkalis total vs. (wt.%) SiO 196 2 (wt.%) > 0w. g)rneadwithin and range MgO) wt.% 10 2 K + O tal. et [145]. 2 h )(t%,where (wt.%), O) toso Fe of ations > a ihMOabun- MgO with vas ri aa r lower are lavas troit 0w. g)and MgO) wt.% 10 0] h HREE The 201]. to r interme- are atios c srflce in reflected is ich vs[153]. avas eotdi this in reported ee n fthe of end leted 2 O 3 T K , of s 2 O, a)

Figure 7.5. Incompatible trace element concentrations of Detroit (a), Nintoku (b), and Koko (c) seamounts normalized to primitive mantle [161]. Shaded ﬁeld represents Hawaiian Chain tholeiitic basalts with MgO contents greater than 10 wt.% (GEOROC database: http://georoc.mpch-mainz.gwdg.de/georoc/Entry.html).

7.3.2.2 Nintoku Seamount (∼ 56 Ma)

Only one sample was analyzed from Nintoku Seamount. This sample (1205 35R-4 104-100) is from a moderately to highly olivine-phyric a’a flow containing accumulated olivine [236]. It is the only tholeiitic lava flow unit (Unit 19b) recovered from Nintoku Seamount, although it actually falls within the alkalic basalt field (Fig. 7.4) because of subaerial and submarine alteration that has increased the Na2O and K2O concentrations of the Nintoku lavas above primary levels [224].

Abundances of Al2O3 and K2O fall intermediate between Detroit and Koko lavas.

T CaO and Fe2O3 are similar to Koko lavas, while P2O5 and TiO2 abundances are similar to Detroit lavas.

197 fHwia sadtoeie smlswith (samples tholeiites wit Island plot Hawaiian they of and Seamount, Koko and Islands Hawaiian the for iia oKk aa n aainIln hlits(i.7 (Fig. tholeiites La/ Island range. Hawaiian lava and Koko lavas the Koko of to end similar depleted thol the at Island and Hawaiian MgO) of wt.% range Hist 10 the Kilauea within the well of but end 7.7a) depleted (Fig. the at fall abundances Nb and ocnrtoso noptbetaeeeet r general are elements trace incompatible of Concentrations c,K (c), itk,adKk emut.Nt htDtotSaon has Seamount Detroit that Note Seamounts. Koko and Nintoku, iue76 akrdarm hwn ao lmn rnsof trends element major showing diagrams Harker 7.6. Figure bnacso Al of abundances 2 d,P (d), O

T Al O (wt.%) P2O5 (wt.%) Fe2O3 (wt.%) 2 3 )f) e) c) )b) a) 2 O 2 5 O e,adTiO and (e), 3 a n a b n oe bnacso Fe of abundances lower and (b) CaO and (a) 2 198 f hnteyugrECseamounts. ESC younger the than (f)

K2O (wt.%) TiO2 (wt.%) > d) 0w. g)(i.75) Thorium 7.5b). (Fig. MgO) wt.% 10 .7b). ylwrteaverage the lower ly its(rae than (greater eiites i h oe limit lower the hin rclLv ﬁeld Lava orical Detroit, brto are ratios Yb higher 2 O 3 Kilauea Historical Lavas

MORB

Kilauea Historical Lavas

MORB b)

Figure 7.7. Th (ppm) vs Nb (ppm) (a) and La/Yb vs Nb (ppm) (b). The Detroit lavas have lie at the depleted end of the range of Hawaiian Island tholeiites (greater than 10 wt.% MgO), but are similar to MORB (MORB data from Regelous et al. [202])(a). On the La/Yb vs Nb plot, the Detroit Lavas are intermediate between the depleted end of the Hawaiian Island tholeiites (greater than 10 wt.% MgO) and the MORB ﬁeld (b). Kilauea historical lava data from Pietruszka & Garcia [197].

7.3.2.3 Koko Seamount (∼ 48 Ma)

A total of five samples were analyzed from Koko Seamount. Four samples plot within the tholeiitic basalt and picritic basalt fields (Fig. 7.4), while the fifth is classified as an alkalic basalt. Similarly to Site 1205, it’s likely that post- eruption alteration has increased either (or both) the Na2O and K2O concentrations in this sample. In addition, the TiO2 and Zr concentrations are similar to the other tholeiitic samples, which suggests that this sample is in fact tholeiitic. The four samples with MgO concentrations greater than 10 wt.% follow a

199 linear trend on Figure 7.6b and Figure 7.6f suggesting that they evolved from a single parental magma under olivine control. The 5th sample plots close to the trends deﬁned by these four samples and probably experienced plagioclase fractionation. Concentrations of incompatible trace elements are similar to Hawaiian Island tholeiites (greater than 10 wt.% MgO) (Fig. 7.5) and Kilauea historical lavas (not shown on Fig. 7.5). Thorium values are slightly lower at a given Nb value than in historical lavas from Kilauea (Fig. 7.7a). However, they are within the range of Hawaiian Island tholeiites (greater than 10 wt.% MgO). La/Yb values are similar to Kilauea historical lavas (Fig. 7.7b).

7.3.3 Platinum Group Elements and Re–Os Isotopes

Platinum-group element (PGE) data are presented in Table 7.3. Primitive mantle-normalized PGE profiles of the ESC basalts analyzed in this study generally overlap the fields for komatiites and the upper range of picritic basalts (Fig. 7.8). All of the samples analyzed show a flattening or even depletion of Pd relative to Pt and an enrichment of Ir relative to Ru. Nintoku and Koko lavas have similar concentrations of PGEs, approximately at the upper limit of the picrite field. Detroit lavas are slightly more depleted than the Nintoku and Koko lavas, approximately in the middle of the picrite field. All of the ESC lavas are more fractionated than komatiitic lavas (Fig. 7.8).

200 TABLE 7.3 PLATINUM GROUP ELEMENT ABUNDANCES (ALL VALUES ARE IN PPB) OF ESC SAMPLES

Site 1203 1203 1203 1203 1205 1206 1206 1206 1206 1206 UMT-1

Core 32R-3 37R-3 42R-5 59R-2 35R-4 04R-3 15R-1 18R-1 38R-1 39R-3

Interval 64–72 80–89 28–37 74–80 104–110 73–80 7–10 51–54 17–25 81-90

Ir 1.60 1.61 0.84 1.45 3.12 2.10 2.26 5.23 6.52 2.06 8.93

2σ 0.24 0.49 0.42 0.06 2.20 0.88 1.27 1.74 1.69 0.20

Ru 1.48 1.91 0.87 0.95 2.81 2.01 1.91 2.81 6.41 3.02 10.9

2σ 0.18 0.62 3.82 0.37 0.09 0.43 0.60 0.35 1.88 0.89

Rh 0.57 0.59 0.38 0.45 0.97 0.92 0.74 1.02 1.96 1.20 9.86

2σ 0.09 0.06 0.84 0.15 0.53 0.30 0.07 0.37 0.73 0.32

Pt 12.77 13.99 8.73 10.39 30.00 31.88 21.15 57.87 52.36 18.00 117

2σ 2.06 0.84 2.65 0.81 15.59 7.09 14.33 2.07 36.78 4.16

Pd 9.99 5.47 2.36 5.56 6.29 11.90 7.75 9.23 32.10 10.26 90.1

2σ 4.73 0.99 5.17 0.56 1.54 5.11 2.43 3.29 5.01 1.59

201 Detroit Seamount a)

Koko Seamount c)

Nintoku Seamount b)

Figure 7.8. Primitive mantle-normalized PGE proﬁles of Detroit (a), Nintoku (b), and Koko (c) basalts analyzed in this study. The range in the PGE composition of Hawaiian picrites is shown as the shaded ﬁeld. Data for Hawaiian picrites from: [20, 239]

Re–Os isotopic data are presented in Table 7.4. Five samples were analyzed for 187Os/188Os isotopes, two from Site 1203 and 3 from Site 1206. Age-corrected 187Os/188Os values of these samples range from 0.13536 to 0.15176. Figure 7.9 shows that initial 187Os/188Os increases with increasing seamount age. The data extend the range of Hawaiian picrite data collected by Brandon et al. [25] but extend towards MORB-like values (Fig. 7.10).

202 TABLE 7.4 RE/OS ISOTOPIC DATA (CONCENTRATION VALUES IN PPB) OF ESC SAMPLES

Site 1203 1203 1206 1206 1206

Core 32R-3 37R-3 18R-1 38R-1 39R-3

Interval 64–72 80–89 51–54 17–25 81–90

Age (Ma) 81 81 48 48 48

Re 2.13 1.62 2.87 1.83 1.18

Os 0.317 0.644 0.162 0.206 0.365

187Os/188Os (i) 0.15177±12 0.14336±10 0.14014±50 0.13681±15 0.13536±13 Os (i) 188 Os/ 187

Figure 7.9. 187Os/188Os vs Age (Ma). Hawaiian data from Brandon et al. [25].

203 h erae G otnsi h ie10 n 25lvsrel lavas 1205 and 1203 Site the in contents PGE decreased the componen mantle depleted element trace incompatible tory, averag contrib increased higher an of in result resulting lithosphere the thin or beneath [124] ridge the melt to partial close to mantle response in be also to hypothesized been in have elements incompatible of levels Decreased t [159]. close center lithosphere thin young, on formed Seamount was Seamount Koko troit to mixin Seamount decreased Detroit to from PGE lavas due the plume possibly Hawaiian of ESC, concentrations the The f of the formation Ridge. during during Hawaiian evolved the plume and Hawaiian ESC the the of budget PGE the that . Discussion 7.4 h G n eO stpccnetain fteECseamoun ESC the of concentrations isotopic Re–Os and PGE The iue7.10. Figure al. 2] OBdt rmmn okr 6,24 239]. 204, [69, workers many from data MORB [25]. 187Os/188Os (i) 187 Os/ 188 sv s(p) aaindt rmBrandon from data Hawaiian (ppt). Os vs Os 204 to frltvl refrac- relatively of ution eri n ej lavas Meiji and Detroit n fdpee upper depleted of ing s[0] naddition, In [201]. ts omrspreading former a o ere fmelting of degrees e tv otoefrom those to ative fMR and MORB of g 19 ic De- since [129] rainof ormation increased s sindicate ts et Site 1206 could be due to sulﬁde saturation and subsequent PGE scavenging or to changes in the source material of the Hawaiian plume source during development of the ESC. These possibilities will be explored in greater detail below.

7.4.1 Signiﬁcance of Pd and Os Depletions

Many of the PGE profiles from the Site 1203, 1205, and 1206 lavas show a relatively flattening or depletion of Pd relative to Pt when normalized to primitive mantle (Fig. 7.8). As discussed at length in Ely & Neal [75] such a flattening/depletion could be the effect of several different processes. Such a depletion is unlikely to be produced by the cation exchange procedure as profiles showing no flattening were produced (1203 32R-3 64-72, 1206 04R-3 73-80, and the reference material UMT-1). Ely & Neal [75] explains that Pd depletions could be caused by chromite fractionation/accumulation, sulfide immiscibility, and weathering. Chromite fractionation/accumulation is unlikely since there is no correlation between Pd and Cr abundances (Fig. 7.11). Extraction of a sulfide liquid could cause a relatively depletion in Pd since Pd has a partition coefficient roughly

1.5–2 times that of Pt for sulfide liquid/silicate melt. However, extraction of a sulfide liquid is also unlikely given the relatively high abundances of PGEs in the samples and the large values of the PGE-sulfur partition coefficients. Therefore, we conclude that the most likely candidate for the cause of the Pd depletions is weathering. As Ely & Neal [75] state, Pd has been found to be more mobile than Pt in surficial ore deposits and is often transported away in solution. In addition, hydrothermal alteration preferentially mobilizes Pd [14, 256]. The ODP samples analyzed in this study have undergone both subaerial and submarine alteration [207, 236] and have LOI values ranging from 1.17 to 6.78 wt.%. Sample 1206 04R-3

205 Figure 7.11. Pd (ppb) vs Cr (ppb) for Site 1203, 1205, and 1206 lavas. Note that there is no correlation of Pd and Cr abundances which suggest that chromite fractionation/accumulation does not control Pd abundances.

73-80, which does not show the relative Pd depletion, has a low LOI, suggesting a correlation between increased weathering and the degree of Pd depletion. Pt/Pd does tend to increase with increasing LOI (Fig. 7.12), despite a weak correlation. However, this sample with the highest LOI does not have the highest Pt/Pd ratio. This may be because of the nature of the alteration that has aﬀected the samples. For example, if alteration occurred under oxidizing conditions, it would promote the formation of Pd cations, making it more bioavailable [58]. If the alteration conditions were reducing, metallic Pd species would be formed, inhibiting mobility. Similarly, the ﬁve samples analyzed for Os abundances and isotopic ratios show fractionations of Os from Ir (Fig. 7.13). These elements are expected to behave in a similar manner geochemically and fractionations substantially away from the chondritic C1 Os/Ir ratio of 1.05 are not expected. The range in the Os/Ir ratio of

206 Figure 7.12. LOI (wt.%) vs Pt/Pd from ESC high-MgO and picritic samples. There is a rough positive trend to the data. Lower Pt/Pd ratios at LOI values greater than 4 wt.% may be a function of the high amounts of alteration also mobilizing Pt.

terrestrial rocks and minerals is 0.8 to 1.3 × C1, with the majority of the samples having a slightly lower Os/Ir ratio [2]. Handler et al. [98] found that massif peridotites and xenoliths from Australia also tended to have lower Os/Ir ratios than C1 chondrite. There are several processes which could potentially fractionate Os and Ru from Ir. Crystal/Liquid distribution coeﬃcients for Os and Ir in chromite are

2700 and 1600, respectively [89], which suggests that chromite crystallization could fractionate Os from Ir. Figure 7.14 shows that the samples with the lowest Cr concentrations also tend to have more fractionated Os/Ir ratios, which could be consistent with chromite fractionation. Figure 7.14 shows that the Site 1206 lavas tend to have the lowest Cr/La ratios, which match with the lowest Os/Ir ratios, again suggesting that chromite removal could be the cause of the fractionated Os/Ir ratios. However, if early chromite fractionation occurred, it would

207 Figure 7.13. Primitive mantle normalized plot of the ﬁve samples with Os concentrations. Notice the pronounced fractionation of Os from Ir.

be expected that both Os and Ru would be fractionated from Ir. This is not observed in the samples, so it’s unlikely that early chromite fractionation could cause the observed primitive mantle normalized profiles (Fig. 7.13) and Os/Ir ratios. Osmium can fractionate from Ir by several different post-emplacement secondary processes. Hydrothermal fluids could facilitate the volatilization of Os

2− through the formation of OsO4 or OsO2Cl4 [2]. This mode of fractionation has been cited in the Troodos Ophiolite Complex of Cyprus [17]. Handler et al. [98] also invoke a post-emplacement secondary process to subchondritic Os/Ir ratios from massif peridotites and xenoliths in Australia. In this case, they attributed the loss of S by volatilization under oxidizing conditions to also lead to

Os volatilization. Oxidizing conditions may have had a role in low Os abundances of the Site 1206 lavas. Examination of the alteration phases of the Site 1206 lavas shows that olivine phenocrysts are occasionally rimmed with a reddish mineral

208 that is reported by Tarduno et al. [236] to be Fe oxyhydroxide, which forms under oxidizing conditions. The Site 1203 samples appear to have experienced low temperature alteration under at least mildly reducing oxygen fugacity as no Fe oxyhydroxide is present in these samples. Therefore, it is believed that the low Os abundances and resulting Os/Ir ratios of the Site 1206 samples are due to the formation of tetravalent Os under oxidizing conditions and it’s subsequent removal due by hydrothermal ﬂuids or volatilization.

209 a)

Figure 7.14. a: Os/Ir vs Cr (ppm) for the ESC lavas. Note that the Cr abundances are lowest for the Site 1206 samples with the lowest Os/Ir ratio. b: Os/Ir vs Cr/La for the ESC lavas. Note that Cr/La ratios are lowest in the Site 1206 (the ratio of Cr to La should be independent of diﬀerences in overall trace element concentrations) and correlate with the lowest Os/Ir ratios. Since the data presented in these plots are from diﬀerent seamounts, chromite fractionation trends cannot be predictively modeled.

7.4.2 PGE and Re–Os Systematics

Mixing of MORB and Hawaiian plume lavas has been invoked to explain the depleted incompatible element and isotopic signatures of Detroit and Meiji lavas [124, 129]. 87Sr/86Sr and 143Nd/144Nd ratios progressively become more depleted

210 and MORB-like with age along the ESC, culminating in the nearly identical to

MORB Sr and Nd isotopic values of Detroit and Meiji Seamounts [124, 201]. In addition, plate reconstructions place Detroit and Meiji seamounts near an ancient spreading center during their formation at approximately 80 Ma [159]. MORB is formed from the partial melting of the isotopically and incompatible element depleted MORB source material (depleted mantle material or DMM). DMM has experienced many melting episodes and therefore, the ratios of incompatible to compatible elements is quite low and 87Sr/86Sr and 187Os/188Os will be unradiogenic in DMM while 143Nd/144Nd will be elevated. These isotopic properties are passed along to the melts that form MORB. So, initial MORB will have a relatively unradiogenic 187Os/188Os isotopic ratio. However, unlike the Sr and Nd isotopic systems where Rb and Sr and Sm and Nd are similar in incompatibility to each other (relative to the much larger diﬀerence in bulk partition coeﬃcients between Re and Os), Re is more incompatible than Os (Os is actually mildly compatible during mantle melting), so MORB lavas will have very high Re/Os ratios (Re abundances of N-MORB average nearly 1000 ppt, while Os averages approximately 50 ppt). These high ratios will develop into extremely high 187Os/188Os ratios over time. The PGE data presented in this paper indicate that Detroit has lower abundances of PGEs and 187Os/188Os isotopic values that are elevated relative to the rest of the ESC and Hawaii. Is MORB (either young or ancient)–plume mixing a viable process for explaining these data?

187 188 The ( Os/ Os)I ratios of the Site 1203 lavas analyzed in this study are more radiogenic than those found at Site 1206 or the Hawaiian Islands (Figs. 7.9 and

7.10), indicating that mixing between ancient MORB and plume material may be a viable cause for the highly radiogenic 187Os/188Os ratios at Site 1203. However,

211 because of the compatible nature of Os during mantle melting, the overall Os budget in MORB is extremely low (average < 10 pg/g) and therefore, for addition MORB crust, large proportions of MORB are needed to signiﬁcantly alter the 187Os/188Os ratio of the plume material. Using the binary mixing equation for isotopic ratios:

187Os 187Os Os 187Os Os A − B 188 = 188 fA + 188 (1 fA) (7.1) OsM OsA OsM OsB OsM

where fA is the percentage of component A in the mixture (from 1 to 0) and component A is plume material and component B is MORB. Young MORB has initial 187Os/188Os ratios that are actually lower than the Hawaiian plume material (∼ 0.12 [100] vs ∼ 0.135), so mixing of young MORB and Hawaiian plume material would be incapable of elevating the 187Os/188Os ratio of the mix. If the mixing scenario proposed in Keller et al. [124] had occurred and Hawaiian plume lavas had entrained DMM, the 187Os/188Os ratios of the mixture should have been lowered towards young MORB values. Therefore, we concluded that the elevated 187Os/188Os ratios of the Site 1203 lavas were not the result of entrainment of

DMM into the Hawaiian plume. Because of the high Re and low Os abundances in oceanic crust, ancient oceanic crust will develop extremely suprachondritic 187Os/188Os isotopic ratios over time. Plume-crust mixing models presented in Lassiter & Hauri [142] and Brandon et al. [25] used 187Os/188Os values of 2.74 and 2.88, respectively, for ancient (1.8 and 2 Ga, respectively) recycled oceanic crust. These extremely suprachondritic 187Os/188Os ratios would be capable of raising the overall 187Os/188Os ratio despite low Os abundances. As mentioned in sub section 7.1.4, Brandon et al. [25] showed that Hawaiian lavas have coupled enrichments in 187Os/188Os and

212 186Os/188Os isotopic ratios that could be caused by incorporation of outer core material. Coupled enrichments in 186Os/188Os and 187Os/188Os are unlikely to be caused by the incorporation of recycled oceanic crust due to the extremely low Pt concentrations (e.g., < 6 ppb) in oceanic crust and MORB. Even though oceanic crust can develop elevated Pt/Os ratios (due predominantly to extremely low Os concentrations (e.g., < 0.5 ppb)), the overall low Pt concentrations require unten- ably large proportions of ancient recycled crust to be added to the plume source to appreciably change the 186Os/188Os isotopic ratios [25]. The core incorporation model of Brandon et al. [25] assumes that if initial bulk core 187Os/188Os ratios and Re and Os abundances were similar to chondritic metal and that a rapid, complete crystallization of the inner core was completed by 4.3 Ga, then

187 188 the present day Os/ Os ratio would be 0.13709 (or a γOs = +7.95). The very high Os abundance of the core (> 1000 ppb) means that even a small percentage of outer core material being added to the Hawaiian plume source will dominate the Os isotopic composition (see Brandon et al. [25] and equation 7.1). In this case, only 0.8% outer core material added to the Hawaiian plume will elevate the 187Os/188Os and 186Os/188Os ratios of the source to those found in some Hawaiian lavas.

Table 7.5 shows the 187Os/188Os isotopic compositions resulting from mixing diﬀerent combinations of mantle components. The Upper Mantle component (UM) represents a mantle source (187Os/188Os = 0.1264) that is intermediate between Hawaiian lavas (0.133) and DMM (0.12). Hawaiian-type mantle (HM) composition is an average of the 187Os/188Os and Os abundances of Hawaiian lavas from Mauna Kea, Mauna Loa, Hualalai, and Kilauea [25]. The recycled oceanic crust (ROC) component is an 187Os/188Os enriched basaltic crust. The

213 reducing sediment (SED) component represents oceanic pelagic sediment. The composition of the UM and ROC components are from Lassiter & Hauri [142] while the composition of the HM, SED and OC components are from Brandon et al. [25]. These mantle components were combined in various ways to model the eﬀect on the Re-Os system of addition of recycled oceanic crust, recycled oceanic crust plus sediment, and outer core to an upper mantle and Hawaiian-type mantle compositions. Three diﬀerent 187Os/188Os starting compositions were chosen to model: one to represent the maximum 187Os/188Os ratios observed at Site 1203 (0.150), another to represent Koolau source (0.143), and another to represent the average Hawaiian tholeiite value (0.135). The value of fA (the percentage of the enriched component needed to achieve the required 187Os/188Os ratio given the initial component compositions) was found by solving the isotopic binary mixing

187 188 187 188 equation for fA. In this case, the value of Os/ OsM (where Os/ OsM is the 187Os/188Os ratio of the mixed source) is one of the three ratios given above.

OsM is calculated using a simple mass balance approach. The ROCSED hybrid component is a 97:3 mix of recycled oceanic crust and reducing sediment and the UMOC hybrid component is a 99.2:0.8 mix of UM and OC (the mixing parameters were chosen to represent the favored mixing parameters proposed by

Lassiter & Hauri [142] and Brandon et al. [25] for ROCSED and UMOC, respectively). A graphical representation of the results is shown in Figure 7.15. The final source compositions with the highest Os abundances, i.e., UMOC-ROC, UMOC-ROCSED, HM-ROC, and HM-ROCSED, all show significant variations in the amount of depleted component required to produce the modeled 187Os/188Os compositions (Table 7.5). This is because the higher Os abundances in UM and HM make these components more insensitive to modification through assimilation

214 of a different source. The UM component has 3.1 ppb Os which leaves this component relatively more susceptible to change than the HM component. This is also reflected in the limited variation in amount of depleted component required (from a maximum of 36% to a minimum of 8% depleted component). If the outer core contribution model of Brandon et al. [25] is assumed, then the HM component can be considered to represent a combination of a peridotite source plus outer core. In this light, the similarity between the UCOC and HM model source regions are not surprising. The dramatic difference between the HM-ROCSED and HM-ROC model source region mixing percentages, especially at the higher 187Os/188Os isotopic ratios, suggest that mixing of recycled oceanic basalt alone is not feasible to model the Site 1203 187Os/188Os ratios. If a small amount of pelagic sediments (3%) are added to ancient oceanic basalt, then the mixing parameters are much more realistic (e.g., to generate the 187Os/188Os ratios of Detroit Seamount lavas by mixing UM and ROC, it requires 37% ROC. Adding 3% SED to the ROC lowers the proportion of the ROCSED mix to 19%). Extremely high (> 30) amounts of recycled oceanic crust are not realistic because melting such a large proportion of what would essentially be eclogite would cause the produced magmas to be andesitic to tonalitic rather than tholeiitic. Therefore, the maximum contribution from recycled oceanic crust can only be approximately 30%. The data presented in this paper make it difficult to unambiguously determine the exact nature of the source components, but it is certain that addition of only outer core material to the Hawaiian plume cannot produce the elevated 187Os/188Os ratios seen at Site 1203. Addition of only recycled oceanic crust is also insufficient to produce the elevated 187Os/188Os ratios at Site 1203 without addition of prohibitively large amounts of the depleted component. Therefore, the most realistic model that our

215 data supports is a Hawaiian plume type mantle component (represented by Mauna

Kea, Mauna Loa, Hualalai, and Kileaua) mixed with recycled oceanic crust plus a small amount of pelagic sediment. This conclusion is supported by multiple studies that attribute parts of the geochemical signature of the Hawaiian plume to a depleted and recycled component (e.g., [104, 142]).

TABLE 7.5 MIXING PARAMETERS FOR 187OS/188OS MODEL

Component1 Os (ppb) 187Os/188Os

Upper Mantle (UM) 3.1 0.1264

Hawaiian-type Mantle (HM) 8.5 0.133

Recycled Oceanic Crust (ROC) 0.05 2.74

Reducing Sediment (SED) 0.244 27.0

Outer Core (OC) 1112 0.1371

Mixed Components fA OsM (ppb)3 187Os/188OsM 4

UM–OC (UMOC) 0.992 11.97 0.13435

ROC–SED (ROCSED) 0.97 0.0558 5.922

Final Source fA5 187Os/188OsM

HM–ROC 0.88 0.135

0.60 0.143

0.47 0.150

HM–ROCSED 0.95 0.135

0.79 0.143

216 TABLE 7.5 (CONTINUED)

Final Source fA2 187Os/188OsM

0.69 0.150

UM–ROC 0.83 0.135

0.72 0.143

0.64 0.150

UM–ROCSED 0.92 0.135

0.86 0.143

0.81 0.150

UMOC–ROC 0.94 0.135

0.56 0.143

0.41 0.150

UMOC–ROCSED 0.97 0.135

0.76 0.143

0.63 0.150

1UM and ROC are from Lassiter & Hauri [142]. HM, SED, and OC is from Brandon et al. [25]

2Calculated by solving the isotopic binary mixing equation for fA using the 187Os/188Os values of 0.135 (∼average Site 1206), 0.143 (lower limit of Site 1203), and 0.150 (upper limit of Site 1203)

3OsM is calculated using a simple mass balance approach

4187 188 Os/ OsM is calculated using the isotopic binary mixing equation shown above

5Calculated by solving the isotopic binary mixing equation for fA using the 187Os/188Os values of 0.135 (∼average Site 1206), 0.143 (lower limit of Site 1203), and 0.150 (upper limit of Site 1203)

217 Figure 7.15. Graphical representation of the results of the 187Os/188Os mixing model. The top bar for each component is for a 187Os/188Os ratio of 0.135, the middle bar is for a 187Os/188Os ratio of 0.143 and the bottom bar for a 187Os/188Os ratio of 0.150.

How does this potential mixing of recycled oceanic crust and Hawaiian plume material take place? Regelous et al. [201] proposed that the MORB-like Sr and Nd isotopic compositions of the Detroit lavas were not the result of mixing between

MORB magmas or DMM due to the plume being close to an ancient spreading center, but instead was the result of increased melting of a heterogeneous mantle composed of a low melting point, incompatible element enriched heterogeneities (the enriched mantle plume component) contained within a depleted, refractory matrix underneath increasingly thin lithosphere. Therefore, the depleted matrix is

218 inherent to the plume and is not the result of a late-stage mixing. Frey et al. [85] examined Sr, Nd, Pb, and Hf isotopic data from Detroit Seamount and concluded that the depleted signature was dissimilar in some key areas to oceanic crust in the vicinity of the Hawaiian hotspot at the time and that the depleted component was intrinsic to the plume. An inherent depleted component was also proposed for Galapagos Plume [109] and for Iceland [226]. In these cases, the depleted component was speculated to be recycled oceanic crust. It was mentioned above that numerous workers have attributed some key geochemical signatures of the Hawaiian plume to recycled oceanic crust. So this strongly depleted component in the Hawaiian plume (and the other plumes) could be ancient recycled oceanic crust that has not fully mixed with Hawaiian plume material and exists as a type of refractory matrix that envelopes the more enriched Hawaiian plume heterogeneities. This concept of mantle structure (i.e., the plum pudding mantle) has been proposed by several workers [4, 109, 178, 179, 196, 228, 233]. When a plum pudding mantle begins to melt, the enriched “plums” will preferentially enter the melt ﬁrst and only in relatively high degrees of melting will the depleted refractory component begin to melt. Under thin lithosphere and corresponding reduced pressure (i.e., enhanced decompression melting), the average melting percentage is elevated and more of the depleted refractory matrix is incorporated into the overall melt. This produces melts with lower incompatible trace element abundances, lower ratios of more- to less-incompatible elements, lower 87Sr/86Sr, and higher 143Nd/144Nd [195, 201]. In addition, the overall 187Os/188Os ratios of the magmas generated during high degrees of melting under thin lithosphere will be elevated.

Frey et al. [85] proposed that the depleted component is intrinsic to the plume and has persisted for over 80 Myr, but is only tapped under special conditions such

219 as rejuvenation stage melting and under thin lithosphere when the plume is close to a spreading axis [159, 201]. Since the thickness of oceanic lithosphere increases the further the crust moves from the spreading center until the lithosphere is approximately 70 Myr old, after which the thickness remains constant [189, 201], Koko Seamount and the Hawaiian Ridge were not formed from the relatively high degrees of partial melting that formed Detroit and presumably Meiji Seamounts. However, one problem with this scenario is that recycled oceanic crust should be eclogitic in composition. Eclogite would preferentially melt over peridotite. So it appears that if the depleted component is recycled oceanic crust, it would have to have undergone an earlier melting event (possibly soon after subduction) and only the residue would contribute to the plume source. Another possibility is that the early plume products sampled a more heterogeneous source and after a period of time the composition of the source became more static.

7.4.3 Eﬀect of Sulﬁde Retention in the Source of the Hawaiian Plume

Bennett et al. [20] explained the PGE and Re characteristics of the Hawaiian plume by suggesting that variable amounts of sulfide retention in the plume source could generate the range of PGE compositions seen in Hawaiian lavas. High PGE and Re lavas, such as those found at Kilauea and Loihi, were the result of relatively small amounts of residual sulfide in the plume while the relatively lower PGE abundances in Koolau picrites were the result of higher amounts of residual sulfide. We decided to test the model outlined in their paper with our ESC picrites (Table 7.6). Bennett et al. [20] modeled the effects of adding a small amount of outer core material to the Hawaiian plume, but determined that it was not necessary due to the fact that there were no samples with Re/Yb high enough

220 to warrant it (see Fig. 6 from Bennett et al. [20]). Our new data from the ESC

ﬁll this void and suggest that mixing of mantle material with a small amount (0.5%) of outer core material combined with very little to no residual sulﬁde is a viable source for the source of the Hawaiian plume during formation of the ESC (Fig. 7.16).

TABLE 7.6 PARAMETERS FOR RESIDUAL SULFIDE MODEL

Component Pt Re Yb Sm Nd Re/Yb Pt/Yb Sm/Nd

UM1 6.54 0.28 0.49 0.44 1.36 0.57 13.4 0.32

Core 6888 164 0 0 0 0 0 0

UM + 5% 40.9 1.10 0.49 0.44 1.35 2.25 84.0 0.32

Outer Core

Dist. Coeﬀ. OL2 OPX3 CPX4 GRT5 Sulﬁde

Pt 0.08 2.2 2.2 3 6900

Re 0.01 0.18 0.03 2.7 900

Yb 0.03 0.03 0.45 5.0 0

Sm 0.0006 0.01 0.28 0.25 0

Nd 0.0002 0.007 0.187 0.087 0

Model OL OPX CPX GRT Sulﬁde

0% Sulﬁde 0.5 0.03 0.15 0.05 0

0.05% Sulﬁde 0.4995 0.3 0.15 0.05 0.0005

0.1% Sulﬁde 0.499 0.3 0.15 0.05 0.001

0.5% Sulﬁde 0.495 0.3 0.15 0.05 0.005

221 TABLE 7.6 (CONTINUED)

Dist. Coeﬀ. OL2 OPX3 CPX4 GRT5 Sulﬁde

Bulk Dist. 0% S 0.05% S 0.1% S 0.5% S

Coeﬀ.

Pt 1.18 4.630 8.080 35.68

Re 0.1985 0.6485 1.0985 4.6985

Yb 0.3415 0.3415 0.3415 0.3414

Sm 0.0578 0.0578 0.0578 0.0578

Nd 0.0346 0.0346 0.0346 0.0346

1Upper mantle garnet lherzolite source

2Olivine

3Orthopyroxene

4Clinopyroxene

5Garnet

Figure 7.16 indicates that variable degrees of residual sulﬁde in the plume source may be responsible for the variations in the PGE and Re contents of the

ESC and Hawaiian lavas. The Site 1206 lavas require approximately 0.05 and 0.1% sulﬁde to be retained while the Site 1203 and 1205 lavas require between 0 and 0.08% residual sulﬁde. This suggests that the degree of melting was, on average, greater in the 1203 and 1205 lavas than in the 1206 lavas, which is consistent with the heterogeneous melting hypothesis of Regelous et al. [201]. Site 1203 formed on thin lithosphere near a spreading center that would have had a larger melting

222 column and relatively high degrees of melting. As such, this would have melted a greater proportion of refractory material and would have melted a greater amount of any sulfide in the plume. Site 1206 would have formed on thicker and cooler lithosphere which would have had a smaller melting column, decreased melting of the refractory depleted component, and some residual sulfide. As the Hawaiian plume continued to create volcanoes on thicker and cooler lithosphere, the amount of melting of this refractory depleted component would continue to decrease and the amount of residual sulfide would also increase as seen in Hawaiian lavas [20].

7.5 Conclusions

The elevated PGE concentrations and suprachondritic 187Os/188Os ratios of picrites and high-Mg basalts from the ESC are consistent with the source of the Hawaiian plume originating at the core-mantle boundary during formation of the ESC (Fig. 7.16). In addition to our geochemical evidence, there is some seismic evidence that the plume extends to the core-mantle boundary [217]. The 187Os/188Os data indicate that the composition of the source material changed during formation of the ESC. The elevated 187Os/188Os ratios of Site 1203 lavas cannot be due to shallow level assimilation of oceanic crust or due to mixing with

MORB lavas formed at the nearby spreading ridge at time of formation. In addition, the elevated 187Os/188Os ratios of the Site 1203 lavas cannot be wholly attributed to increased addition of outer core material. Instead, it is tentatively proposed that the suprachondritic Os isotopic ratios of the Site 1203 lavas are the result of increased melting of a refractory recycled oceanic crust (plus pelagic sediment) due to higher degrees of melting under thin lithosphere. Varying the thickness of the lithosphere caused the proportion of the refractory component to

223 vary and changed the amount of residual sulﬁde left in the plume from nearly zero under thin lithosphere at Sites 1203 and 1205 to almost 0.1% in the Site 1206 lavas to up to 0.5% at the Hawaiian Islands. These data suggest that the time-resolved changes in incompatible elements and isotopic ratios during formation of the ESC are due primarily to changes in lithosphere thickness and that after approximately

70 Ma lithosphere thickness stabilized and lead to the relatively unchanged geochemical signature of the Hawaiian Ridge, supporting the hypothesis of Regelous et al. [201].

224 0% S ore r C

0.05% S .5% Oute + 0 0.1% S

0.5% S a) .5% Outer Core

0% S + 0 UM

0.05% S

0.1% S

0.5% S

Figure 7.16. Re/Yb (a) and Pt/Yb (b) vs Sm/Nd of ESC samples from this study and Hawaiian Island samples [20]. UM is an upper mantle garnet lherzolite source (composition outlined in Table 7.6). The curves are the eﬀect of melting UM that has 0.5% of outer core material added to it at various amounts of residual sulﬁde.

225 CHAPTER 8

CONCLUSIONS

8.1 The Ontong Java Plateau

The vast size (2 × 106 km2) and the homogeneous nature of the Ontong Java

Plateau is suggestive of a magmatic event of epic proportions. Due to the submarine nature of the plateau, our understanding of the lava types that compose the OJP is limited by the diﬃculty in retrieving samples and by the limited penetration depth of deep sea drilling compared to the overall thickness of the plateau (< 5 km vs. ∼ 40 km, respectively). With each expedition to the OJP, we gather more information that can hopefully lead to unravelling the apparent paradoxes relating to the formation and evolution of the OJP. How can a magmatic event the size of which emplaced nearly 6 × 107 km3 [79] of magma nearly all at one time not result in uplift and shallow water to subaerial eruption? Trace element modeling suggests that the OJP was formed from ∼ 30% melting of mildly depleted peridotite source [35, 79]. High degrees of melting such as this requires decompression melting under thin lithosphere and an actively fed melt zone [79]. Even under extraordinary melting conditions such as this, it is astounding that the geochemistry of the known rock types on the OJP are so uniform. Fractional crystallization and slightly diﬀering degrees of partial melting at the edges of the plateau would be expected to generate more

226 differentiated lavas and introduce a range of compositions. Prior to Leg 192, much of the OJP had yet to be sampled, including the entire eastern salient of the plateau. The sequence of volcaniclastic rocks recovered from the eastern salient indicated subaerial eruption conditions for the first time. My work on the basaltic clasts contained within the volcaniclastic succession from the eastern salient of the OJP is the first detailed geochemical study of the volcanic products of the eastern salient. My data indicates that the Kwaimbaita-type basalts were erupted on part of the eastern salient. This is important because the presence of Kwaimbaita-type lavas combined with identical eruption ages as the rest of the plateau shows that the eastern salient is not part of a potential plume tail. In addition, the discovery of a novel lava type suggests that at the margins of the plateau, melting and crystallization conditions were such that more fractionated lavas were produced. The high-Nb group at Site 1184 appears to have been derived from the same mantle source as the Kwaimbaita- and Kroenke-type basalts, yet are relatively evolved compared to those two lava types. The alnöite pipes on the island of Malaita are uniquely suited to examining the deep regions of the OJP that drilling may never be able to sample.

Xenoliths contained within the alnöite have been identified previously as containing many different ultramafic mantle nodules, megacrysts of single minerals that are also suggested to be of mantle origin, and mafic crustal samples. A suite of these xenoliths was examined and it was determined that three samples were of upper mantle origin and the others were OJP related crustal gabbros and granulites. The xenolith samples represent the first samples recovered from the OJP from depths greater than 3.5 km. These samples show evidence of

227 burial metamorphism, some have triple-junction grain boundaries which is suggestive of equilibration in high pressure environment. Two samples appear to be closely related (possibly cumulates of a fractionating Kroenke-type lava) to Kroenke-basalts. The remaining samples are unlike the three main OJP lava types in terms of their incompatible trace element compositions, although post- emplacement alteration may have enriched the LREE proﬁles of these samples. Major element and compatible trace element data suggests that they are more primitive than Kroenke-type basalts. This interpretation is also supported by the generally lower HREE abundances. These clues suggest that these samples are more primitive than Kroenke-type basalt and are possibly parental magmas to the Kroenke-Kwaimbaita lava sequence. Isotopic data from one sample is within the range of Kroenke-Kwaimbaita lavas which is consistent with this interpretation.

8.2 Hawaiian Ridge-Emperor Seamount Chain

The ∼6,000 km long HR-ESC is one of the best studied ocean island chains in the world, but until ODP Leg 197, the geochemistry of many of the seamounts comprising the ESC were unknown to poorly constrained. Early drilling expeditions to the ESC recovered only small amounts of basement (e.g., drilling at Nintoku Seamount recovered on 32 m of basement) or recovered only dredge samples, which tend to be highly altered and of unknown provenance. Leg 197 marked the ﬁrst time that substantial amounts of drill core were recovered from

Detroit, Nintoku, and Koko seamounts. Crude estimates made from the limited drilling and dredge samples had previously been interpreted to show that the ESC was broadly similar to the Hawaiian Islands in terms of major and trace

228 elements, except for Detroit Seamount which was known to be more depleted.

The greater than 300 m of drill core recovered from Nintoku Seamount during Leg 197 allowed, for the first time, extensive interpretations on the geochemistry and source region of this seamount. The thick post-shield alkalic cap that was predominantly sampled showed evidence of being geochemically similar to post-shield basalts from the Hawaiian Islands in terms of major and trace elements, but slightly isotopically depleted relative to Hawaiian post-shield basalts. The trace element and isotopic composition of the Nintoku post-shield lavas was explained by using a modified version of the trace element model for the formation of the Haleakala post-shield basalts [37], where small degree melts of metasomatised Pacific ocean lithosphere are mixed with Hawaiian plume melts to form a hybrid source with a distinct trace element and isotopic composition. The slightly more depleted isotopic ratios compared to Hawaiian basalts is interpreted as increased melting of the depleted component or as incorporation of less metasomatised oceanic lithosphere. The data presented in Chapter 6 indicate that the Hawaiian plume was producing geochemically similar magmas at 56 Ma and today. The drift in the position of the Hawaiian plume during the formation of the ESC [237] appears to have had no impact upon the trace element and isotopic composition of the magmas produced by the plume at 56 Ma. The role of recycled oceanic crust and outer core material in the PGE and Re/Os isotopic composition of the Hawaiian plume has been an ongoing controversy. Suprachondritic Re/Os ratios in Hawaiian picrites has been used as evidence for both components, while various other isotopic ratios have been used to argue both for and against the incorporation of either component. Pt/Os isotopes

229 positively coupled with Re/Os isotopes has been regarded as a strong argument for the incorporation of outer core material [25, 26], although this argument has not gone without challenge [219]. Oxygen and Pb isotopes are commonly used to supplement the recycled oceanic crust hypothesis. The one thing in common with these theories is that they are all based on picritic and high-MgO basalts from the

Hawaiian Islands. No study has been done that addresses the PGE and Re/Os isotopic composition of the ESC, which is known from trace element and Rb/Sr and Sm/Nd isotopes to be more heterogeneous than the Hawaiian Islands. The PGE and Re/Os isotope data presented in this dissertation from Detroit, Nintoku, and Koko seamounts represents the ﬁrst data of this nature from the ESC and illuminates parts of the controversy. The PGE and Re/Os ratios from Nintoku and Detroit are largely similar to those of the Hawaiian Islands, although the absolute PGE abundances are slightly higher. The interesting data comes from Detroit Seamount which is well known for its MORB-like Sr- and Nd-isotopic ratios and depleted trace element abundances. The PGE abundances from Detroit are depleted relative to those from the rest of the ESC, but largely similar to those from Hawaii. However, the Re/Os isotope data shows evidence of a strongly suprachondritic Re/Os source in the Detroit samples. The 187Os/188Os ratios are elevated beyond that which the outer core incorporation models can generate, which leaves the recycled oceanic lithosphere model. Due to the extremely high Re/Os ratios in oceanic crust over time such material will develop extremely suprachondritic 187Os/188Os ratios. The issue is that the overall low amounts of Os make changing the 187Os/188Os ratio of the source that the recycled oceanic crust is added to very diﬃcult. Modeling suggests that the recycled oceanic crust component is only apparent in unusual melting regimes. Detroit Seamount was

230 near an ancient spreading center during its formation and therefore, melting took place under thin lithosphere. This thin lithosphere allowed for greater degrees of decompression melting of a refractory component which lowered the overall trace element abundances and isotopic ratios. If this refractory component was highly enriched in Re/Os and had developed a correspondingly high 187Os/188Os ratio over time, this scenario would result in the overall depleted trace element and Sr- and Nd-isotopic ratios while creating extremely suprachondritic 187Os/188Os ratios in Detroit Seamount lavas. The simplest explanation for the refractory material would be ancient recycled oceanic crust as it would meet all of these geochemical requirements. However, recycled oceanic crust is assumed to be primarily eclogitic, which probably would not be refractory under mantle melting conditions. So it appears that the refractory material would have had to have undergone some sort of phase change to make it more refractory in relation to the Hawaiian plume material.

The work presented in this dissertation on the HR-ESC indicates that core- mantle interaction has occurred over the life of the plume, the Hawaiian plume is composed of enriched heterogeneities contained within a more depleted matrix, and that the apparent ﬁxing in the mantle of the Hawaiian plume at the bend in the HR-ESC did not have any geochemical eﬀects.

8.3 Summary

1. Discovery of a novel OJP lava type.

2. My samples from Site 1184 and the Malaitan aln¨oite xenoliths exhibit more variation in incompatible trace element chemistry than has been documented by ocean drilling and ﬁeldwork.

231 3. First detailed geochemical study of Nintoku Seamount. Found that the

Hawaiian plume was producing magma that were geochemically similar to Hawaiian post-shield magmas at 56 Ma.

4. PGE abundances from the ESC require an outer core component, indicating that the Hawaiian plume has been sampling the outer core throughout its

known lifetime.

5. The Hawaiian plume is composed of enriched heterogeneities contained within a more depleted matrix. This depleted matrix only contributes to the melt during unusual melting conditions.

6. Detroit Seamount’s unique geochemistry is not the result of mixing between

plume and MORB magmas.

7. The bend in the HR-ESC appears to have had no eﬀect on the geochemistry of the plume magmas.

8.4 Relevance and Future Work

The OJP and the HR-ESC are two of the largest igneous structures on the Earth. Understanding their formation is critical to understanding how heat es- capes from the mantle, large scale mantle melting regimes, long-term magma sources, and the interaction between the mantle and outer core. In addition, understanding how various mantle components are incorporated into magma source regions has important implications for mantle structure and dynamics. The PGE model outlined in Chapter 7 would beneﬁt from additional data.

Pt–Os and O isotopes would be the most useful to determine if the ESC samples exhibit the coupled enrichments in 186Os/188Os and 187Os/188Os ratios. The

232 O-isotope data is useful to determine if the samples are degassed and can also be used to identify the presence of recycled oceanic crust. Rb–Sr and Sm–Nd isotopic data on the samples analyzed in this dissertation would also be useful to further compare these samples with others from Detroit, Nintoku, and Koko Seamounts.

233 CHAPTER 9

ANALYTICAL TECHNIQUES

9.1 Overview of Techniques

9.2 Sample Preparation

A large and potentially severe source of contamination can occur during sample preparation. Metal contamination during sample cutting and cross-contamination during grinding, crushing, and powdering all are easy contaminants. Great care needs to be utilized during sample preparation to ensure measured concentrations are representative of the actual rock sample. Samples recovered from Ocean Drilling Program (ODP) sites are presented to the analyst as drill core. Sections of samples are removed from the core by cutting the desired piece from the bulk sample by a diamond-impregnated cutting saw. In general, contamination from the diamond saw needs to be minimized, although this is not critical for major element and loss on ignition (LOI) analysis, as metal pieces could be present on the cut surfaces. This contamination is removed by grinding all the surfaces of the rock sample on diamond-impregnated lapping wheels and thoroughly washing with water. This removes any potential metal contamination from the surface of the sample. After grinding, the samples are broken into 1–2 cm sized chips within a duct tape reinforced plastic bag by hitting with a duct tape cushioned hammer. The duct tape is employed during crushing in

234 order to minimize the contact of the metal hammer with the sample. These small chips are then loaded into an alumina SPEX jaw crusher which further reduces the size of the chips to < 0.5 cm. The jaw crusher was thoroughly cleaned between each sample and a small aliquot of sample chips were introduced, processed, and discarded to minimize the chances of cross-contamination. The chips were then rinsed in 18 MΩ water repeatedly and dried on a hotplate. Once dry, the chips were inspected visually and chips showing greater than average alteration were discarded. The remaining chips were powdered in a SPEX Mixer/Mill in alumina shaker bombs or by hand in an agate mortar and pestle (depending upon the amount of sample available).

This procedure was modiﬁed slightly if the sample was to be analyzed for platinum group elements (PGEs). The PGEs are highly siderophile and also are in signiﬁcant quantities in gold and platinum jewelry. When handling samples intended for PGE analysis, all jewelry was removed from the hands and wrists or gloves were worn. In addition, great care was taken to ensure that any surfaces of the rock sample that may have been in contact with metal are removed and that no further contact is made.

9.3 Loss on Ignition

The loss on ignition (LOI) analysis is a measure of the amount of hydrated and non-primary (i.e. not original to the rock) phases. Compounds that typically contribute to the total LOI are: volatile compounds, including H2O, CO2, Cl, S, and other added compounds, such as the oxidation of FeO to Fe2O3 and later ◦ CO2 (from CaCO3). Carbonate is ignited above temperatures of 850 C, so this procedure also eﬀectively removes any primary or secondary carbonate (primary

235 carbonate is rare in basaltic samples). This analysis is performed for several reasons:

• To obtain an estimation of the amount of alteration of a sample.

• To prepare a sample powder for major and trace element analysis.

• To remove phases that are not original to the rock, which could inﬂuence

the measured composition.

An aliquot of sample powder is carefully weighed into a platinum (or, alternatively, ceramic) crucible and the weight is recorded. The crucible is placed inside a furnace capable of reaching and maintaining temperatures in excess of 1050 ◦C.

The crucible is then heated for four hours, removed, allowed to cool, then immediately weighed again. The diﬀerence in measured weight is converted into LOI by the equation:

LOI =(Wf − Wi)/Wi × 100

where Wi is the initial measured weight and Wf is the measured weight after ignition.

The major element oxide composition (major element oxides = SiO2, Al2O3,

TiO2, MnO, MgO, CaO, Na2O, K2O, P2O5) should total 100% for normal, unaltered basalts. When the analyzed major element composition of un-ignited sample powder is not 100%, the diﬀerence can usually be accounted for by calculating the LOI on the un-ignited powder. This is the dominant way in which LOI will be used.

236 9.4 Major Element Analysis

The major elements (see section 9.3) are deﬁned as those that are present in concentrations greater than one part per hundred (or > 1 wt.%), although

MnO and K2O are commonly present in basalts in concentrations less than 1 wt.%. Major element concentrations are used to differentiate rock types, degree of differentiation, and trends in magma evolution [253]. Sample solutions for major element analysis are prepared as follows: Approxi- mately 0.1 g of sample powder is fused with 0.5 g of lithium metaborate in carbon crucibles at 1050 ◦C for 30 minutes in Notre Dame’s Fisher Scientific Isotemp pro- grammable muffle furnace. Immediately upon removal from the oven, the molten pellets are quenched in 5% HNO3, which is then transferred to polypropylene bottles and brought to 100 g with 5% HNO3. This solution is then placed in an ultrasonic bath and any remaining solid material is allowed to completely dissolve.

Both ignited and un-ignited powders were used for major element analysis. Major element concentrations were determined by inductively coupled plasma- optical emission spectroscopy (ICP-OES) at the University of Notre Dame. The protocol for ICP-OES analyses is outlined in Table 9.1. This protocol is mod- iﬁed from the one used for shipboard ICP-OES analyses during ODP Leg 192

[155] and a novel protocol developed by the candidate for use on the Center for Environmental Science’s (CEST) ICP-OES.

237 TABLE 9.1 PROTOCOL FOR ICP-OES ANALYSIS

Sequence Analyte Counter Sequence Analyte Counter

1 5% HNO3 – 10 DCS-3 0 2 DCS-11 0 11 UNK-3 1

3 PB2 1 12 UNK-2 2

4 SRM-13 2 13 UNK-1 3

5 SRM-2 3 14 DCS-4 0

6 DCS-2 0 15 SRM-2 1

7 UNK-14 1 16 SRM-1 2 8 UNK-2 2 17 PB 3

9 UNK-3 3 18 DCS-5 0

1Drift corrector solution

2Procedural blank

3Standard reference material

4Unknown

The protocol is initiated by a 5% HNO3 solution to ﬂush the system and to determine instrument background levels. Machine drift is accounted for analyzing a drift corrector solution (DCS) every four test tubes. The concentration of this solution needs not be known exactly, but it must contain all analytes of interest and be in concentrations comparable to the solutions of interest. Basaltic standard reference material (SRM) BHVO-1 was usually used as the DCS in the ICP-OES analyses performed as part of this dissertation. The correction equation

238 for machine drift is:

DF =(DCSn + [Xcounter × (DCSn+1 − DCSn)])/DCSn

where Xcounter is the proportion of the number of samples between DCSs. Four test tubes make one “block” of samples (Table 9.1), starting with a DCS and followed by three test tubes containing PBs, SRMs, or UNKs. So, Xcounter for DCS-1 would be 0/4, PB would be 1/4, and so on until DSC-2 at which point the

Xcounter resets to 0/4. DF is the “drift factor”. This factor is calculated for every test tube and is used to calculated the correction factor (CF ) by the equation:

DFn CFDCS = DFfirst

The CFDCS is then used to calculated the drift corrected counts by the equation:

CPSuncorr CPScorr = CFDCS

where CPScorr is the drift corrected counts per second (cps) and CPSuncorr is the uncorrected counts. All major element analyses performed as part of the research presented in this dissertation use at least two of the SRM’s BHVO-1, BHVO-2, and BIR-1. In an individual major element run, typically BHVO-2 and BIR-1 were analyzed and, in some cases, BPL-1 was also analyzed. With the exception of the BPL- 1, all SRM’s have major element concentrations certiﬁed by the United States Geological Society (USGS). In lieu of analyzing standard solutions and creating a calibration curve, calibration is performed by taking the blank- and drift-corrected

239 counts of a selected SRM and dividing by the accepted elemental concentration.

The procedural blank is used for the blank subtraction instead of 5% HNO3 as the lithium metaborate may contribute to the counts of some elements. This calculation resulted in a simple calibration that can then be used to calculate unknown concentration by the equation:

CPSunk Xunk = (CPSsrm/Xsrm) where CPS is counts per section and X is concentration. This procedure is similar to that used for shipboard ICP-OES reduction during ODP Leg 192 [155]. Since the samples that will be analyzed during the course of this study are basaltic, the use of basaltic SRM’s to provide calibration is justiﬁed as matrix eﬀects in the plasma of the ICP will be similar in both sample and SRM. Figure 9.1 diagrams the data reduction process.

9.5 Trace Element Analysis

Trace elements are present in natural samples in concentrations < 1 wt.%. One of the fundamental aims of petrogenetic studies of igneous rocks is to distinguish source characteristics from those arising from subsequent processes [253]. Trace elements, along with isotopes, have been the main tool for petrologists interested in source regions of basaltic systems for the last 50 years.

Trace element abundances were quantiﬁed by ICP-mass spectrometry (ICP- MS) at the University of Notre Dame. Solutions used for trace element analysis were prepared by dissolving the sample powder in a HF-HNO3 acid attack. The series of steps for dissolution of the sample powder are as follows:

240 Raw cps

blank subtraction

blank-subtraced cps

drift correction

blank- and drift- corrected cps

calibration

reduced data

Figure 9.1. Flowchart diagraming reduction of raw ICP-OES data

1. Weigh 0.05 grams of sample powder into a screw-top Teﬂon beaker. Sixty

drops of HNO3 and 50 drops of HF is added to each beaker. The beaker is agitated so that the powder is in suspension. The beaker is capped and

placed on a hotplate for 48 hours at 100 ◦C.

2. The beaker is removed the hotplate and allowed to cool. The beaker is uncapped and several drops of 18 MΩ is added to the cap. The water droplets are swirled around the cap so that any liquid adhered to the cap

can be added back to the beaker.

3. The beaker is placed on the hotplate at 150–180 ◦C and the solution is allowed to dry.

241 4. Once dry, 40 drops of HNO3 is added, the beaker is recapped, and placed back on the hotplate for 24 hours at 100 ◦C.

5. Repeat steps 2 and 3.

6. Once dry, add 40 drops of HNO3 and immediately dry.

7. Once dry, add 150 drops of HNO3 and make sure that all solid material is dissolved. Transfer this solution to a 125 ml polypropylene bottle. 18 MΩ water is added to the bottle until the total solution weight is 100 g. This

results in a ∼5% HNO3 solution.

All acids used in the dissolution procedure were purchased at reagent grade, then double-distilled to ensure purity. The protocol for trace element analysis on the ICP-MS is modiﬁed from that presented in Jenner et al. [119] and Neal [170]. Arsenic, Rh, In, Re, and Tl were used as internal standards at concentrations of 20 ppb in all solutions (Table 9.2). The spike solution was only added to unknown samples in a 50:50 ratio. The

1 concentration of spike in the ﬁnal solution should be approximately 2 of that presented in Table 9.2.

TABLE 9.2 STANDARD AND SPIKE CONCENTRATIONS (PPB) USED IN TRACE

ELEMENT ICP-MS ANALYSIS

Element Std A Std B Spike Element Std A Std B Spike

Li 40 – – Ce 20 20 –

Be 40 – 40 Pr 20 – 40

242 TABLE 9.2 (CONTINUED)

Element Std A Std B Spike Element Std A Std B Spike

Sc 40 – 80 Nd – 40 80

V 40 – 80 Sm – 40 80

Cr – 100 – Eu – 40 20

Co 40 – 80 Gd – 40 40

Ni 40 – 80 Tb 40 – 20

Cu 40 – 80 Dy 20 – 40 Zn 40 – 80 Ho 20 – 20

Ga – 40 80 Er 40 – 80

Rb 40 40 – Tm 20 – 40

Sr 100 – 200 Yb 20 – 20

Y 20 – 20 Lu – 40 20

Zr 40 – 80 Hf – 20 40 Nb 20 – 20 Ta – 20 20

Mo 20 – 20 W – 20 20

Sb 20 – 20 Pb 40 – 40

Cs 20 20 20 Th 20 20 20

Ba 200 – 400 U 20 – 20

La 20 – 40

Two test tubes were prepared for each SRM and unknown: a spiked and unspiked tube. Each test tube contained 5 ml of SRM or unknown sample solution

243 and 5 ml of spike or internal standard for spiked tubes and unspiked tubes, respectively. Procedural blanks and acid blanks were prepared in the same way. The protocol for ICP-MS analysis is presented in Table 9.3.

TABLE 9.3 PROTOCOL FOR ICP-MS ANALYSIS

Sequence Analyte Sequence Analyte

1 5% HNO3 15 Standard B 2 Standard A 16 Flush

3 Standard B 17 Acid Blank

4 Flush 18 UNK-1 Unsp 5 Acid Blank 19 UNK-1 Sp

6 PB Unsp1 20 Flush

7 PB Sp2 21 UNK-2 Unsp

8 Flush 22 UNK-2 Sp

9 SRM-1 Unsp 23 Flush

10 SRM-1 Sp 24 UNK-3 Unsp 11 Flush 25 UNK-3 Sp

12 SRM-2 Unsp 26 Standard A

13 SRM-2 Sp 27 Standard B

14 Standard A 28 5%HNO3

1All acronyms are the same as in Table 9.1, Unspiked

2Spiked

244 An aggressive wash protocol was utilized to ensure memory effects were effectively eliminated. Between each test tube a three-stage wash sequence was employed, the first wash being 10% v/v HNO3 + 10 drops 29 N HF acid, the second being 10% v/v HNO3, and the third being 5% v/v HNO3. Each wash was run for 90 seconds. This procedure was shown to eliminate any memory effects [162] and allows for monitoring of sensitivity fluctuations. The final data was calculated using a combination of external calibration and standard addition. Figure 9.2 diagrams the ICP-MS data reduction method.

Raw cps

blank subtraction

blank-subtraced cps

matrix corrected sensitivity

sensitivity by external calibration

calculation of matrix factors

blank-,drift-, and calibrated-cps

Figure 9.2. Flowchart diagraming reduction of raw ICP-MS data

245 Trace element abundances are calculated using the following equation:

X = CPSblanksubtracted × MF/DILSS/Sext.calib × EF where

• MF is the matrix factor for the element of interest;

• DILSS is the concentration of sample powder in the sample solution times 1000;

• Sext.calib is the sensitivity of the element by external calibration;

• EF is the “element factor”, which varies depending on the mass of the element in interest

Each of these parameters are calculated for each sample and each element. Blank subtraction is applied by using the blank subtracted cps, machine drift is corrected by the Sext.calib factor, sample powder concentration in the sample solution is addressed by the DILSS, ratio of sample solution to spike or internal standard is applied in the MF, and elemental variations due to mass are addressed by the EF.

9.6 Platinum Group Element Analysis

The platinum group elements (PGEs; Pt, Ir, Os, Ru, Rh, and Pd) are present in maﬁc samples generally as metallic alloys, tellurides, selenides, arsenides, sul- pharsenides, and sulﬁdes [57]. These phases can concentrate PGEs in the ppm range, while the bulk-rock will generally contain PGE levels in the ppb to ppt levels [57]. The role of silicate phases as PGE hosts is unclear, as there have been studies indicating a strong correlation between a PGE and a major element oxide,

246 however, the cause of the correlations is not clear [29, 57]. Some workers have proposed that olivine can incorporate Ir [29] while others have suggested that Or, Ir, and Rh could substitute in the chromite structure [1]. Other possibilities include that the PGEs are present in silicates as PGE-rich inclusions [13]. These factors are of considerable importance to PGE analysis because the PGEs have a specific gravity 2 to 8 times that of the other silicate minerals that make up the rock [28, 110]. Because of this, PGEs tend to settle in sample powders making it difficult to prepare a homogeneous powder for analysis. This phenomena is termed the “nugget effect”. Because of the heterogeneous distribution of PGEs, large sample sizes have traditionally employed to mitigate the nugget effect.

Sample preparation for PGE analysis needs to address these issues. The original method for PGE analysis at the University of Notre Dame was presented in Ely et al. [76]. All protocols outlined in Section 9.2 were strictly followed during crushing, grinding, and powdering of samples. The nugget eﬀect problem is addressed in the Ely et al. [76] by thoroughly mixing the sample powder directly prior to dissolution. In contrast to NiS ﬁre assay methods which typically requires at least 5 g of sample powder to be processed during dissolution [200, 214], the Ely et al. [76] requires only 50-200 mg of sample and yields extremely low blank levels. Sample dissolution is straightforward and is outlined below.

1. 250 mg of sample powder is weighed into a teﬂon liner designed to ﬁt inside of high pressure “Parr” dissolution vessels.

2. HNO3 and HF (3:4 ratio) are added to the sample powder and mixed to ensure the acid accesses all of the sample powder.

247 3. The solution is placed on a hotplate set at 150 ◦C and immediately evapo-

rated.

4. The now dried sample is treated with aqua regia (HCl:HNO3—3:1) and the teﬂon digestion vessels are capped and placed inside the Parr vessels.

5. The samples are heated at 150 ◦C for 24–48 hours depending upon the min-

eralogy of the samples.

6. The resulting solution is evaporated, treated with 100 drops of 12 ml HCl, evaporated again, then fully dissolved in 2 ml of 5% v/v HCl.

7. This solution is loaded onto cation exchange columns and the eﬄuent is collected and prepared for analysis.

The PGEs form chlorocomplexes in HCl media which are not retained on cation exchange resin [231] as opposed to anion exchange resins which bind PGEs [203]. PGEs were quantiﬁed at the University of Notre Dame using a VG Elemental PQII STE ICP-MS in combination with a CETAC U-5000AT ultrasonic nebulizer

(USN). The use of the was utilized to achieve the desired machine detection limits (< 1 pg/g) which allows for quantitative determination of PGEs in rock samples [76]. However, sample passing through the USN is exposed to a large amount of glass surface, which in turn, retains elements being quantified. Therefore, a rigorous four-stage wash protocol was developed by Ely et al. [76] which effectively eliminated memory effects.

9.7 Isotope Analyses

Sr (87Sr/86Sr), Nd (143Nd/144Nd), and Pb (206Pb/204Pb, 207Pb/204Pb, 208Pb/204Pb) isotope analyses were performed by collaborators on several projects. Os (187Os/188Os)

248 analyses were conducted at Johnson Space Center under the supervision of Dr.

Alan D. Brandon.

9.7.1 Sr-, Nd-, & Pb-Isotope Analysis as part of Chapter 4

Strontium-, Nd-, and Pb-isotopic measurements conducted as part of Chap- ter 4 were performed by Dr. Paterno R. Castillo at Scripps Institute of Oceanog- raphy using a Micromass Sector 54 multi-collector thermal ionization mass spectrometer. In order to minimize the eﬀects of sea-water alteration on the Sr and Nd isotopic composition, the sample powders were subjected to a harsh multi- step leaching procedure [31, 154]. Unleached powders were used for Pb-isotope analysis because almost all the leached powders yielded very low concentrations of Pb, indicating that most of the Pb was in leachable phases. Both the leached and unleached powders were dissolved in clean Teﬂon vessels using ∼ 1 ml of a

2:1 mixture of concentrated Hf and HNO3, and heated on a hotplate at low power for 16-24 h. The resulting solutions were evaporated to dryness, redissolved in a small amount of concentrated HNO3, evaporated to dryness and the procedure re- peated. Strontium and rare earth elements (REE) were ﬁrst separated in primary cation-exchange columns; Nd was from the rest of the REE by passing the REE aliquots through small EDTA ion-exchange columns. Lead was separated using a standard anion-exchange method [117, 152] in a HBr-HNO3 medium. Concentra- tions of Rb, Sr, Sm, Nd, U, Th, and Pb were measured on separate dissolutions using a high-resolution ICP-MS at the Scripps Institute of Oceanography [117].

249 9.7.2 Sr- & Nd-Isotope Analysis as part of Chapter 6

Strontium and Nd isotope analyses were carried out by Dr. Marcel Regelous at the University of Bristol. Before dissolution, sample powders designated for

Sr isotope analysis were subjected to a sequential leaching procedure in order to remove alteration products. About 4 g of rock powder was weighed into a Teﬂon beaker to which 5–10 ml of distilled 6M HCl was added. The samples were placed in an ultrasonic bath for 2 hours and the acid was changed every 30 minutes. The samples were then leached in hot (80 ◦C) 6M HCl for a further 5 hours with the acid being changed every 60 minutes. The residue was soaked in deionized water at 80 ◦C for 30 minutes, rinsed twice, and dried. This leaching procedure has been shown to yield reliable initial Sr isotope data for old, altered Hawaiian lavas [112, 201].

For Sr isotope analysis, about 0.1 g of leached rock powder was spiked with

87 84 Rb and Sr and digested in HF-HNO3. After treating with 15M HNO3 and 6M HCl, the sample was dissolved completely in dilute HCl. Rubidium and Sr were separated from the rock matrix on cation exchange columns in 2.5M HCl. The

Rb fraction was evaporated to dryness, redissolved in 3M HNO3, and cleaned of Sr by passing it through a second column containing 50 µl of Sr Spec resin. Neodymium isotope analyses were conducted on unleached sample powders.

Approximately 0.1 g of rock powder was digested using HF-HNO3-HCl, dissolving in 3M HNO3, and passed through a column containing 0.1 ml of TRU spec resin.

After washing with 3M HNO3, the REE were recovered in 2.5M HCl. All reagents used were quartz distilled and total procedural blanks for Sr and Nd were below 400 pg and 50 pg, respectively.

250 Strontium and Nd isotope measurements were conducted using a Finnigan

Triton thermal ionization mass spectrometer in static mode. These measurements were corrected for instrumental mass fractionation using exponential fractionation corrections and assuming 87Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219. The NBS-987 Sr and J&M Nd standards gave 0.710247 ± 8(n = 12) and 0.511114 ±

5(n = 8) over the period of the analysis. Rubidium isotope compositions were determined by multicollector ICP-MS using methods adapted from Waight et al. [247]. The Rb fraction was dissolved in 0.6M HCl + 0.02M HF in order to obtain a Rb concentration of ∼ 100 ppb. About 2 ml of this solution was transferred to a small vial and approximately 250 ng of natural Zr was added. Sample solutions were introduced into a Finnigan Neptune double focusing MC-ICP-MS via a Cetac Aridus microconcentric nebulizer and desolvating system equipped with a PFA spray chamber, at an uptake rate of 50 µl/minute using a sample Ar gas ﬂow rate of ∼ 0.9 L/minute. The spray chamber was maintained at a temperature of 105 ◦C. Typical Ar (sweep gas) and N2 ﬂows were 3 L/minute and 20 ml/minute, respectively. At the start of each analytical session, the instrument was tuned using a solution of natural Rb. Typical sensitivity for Rb was ∼ 200–400 Volts/ppm. Masses 85, 87, 88,

90, and 91 were measured simultaneously on Faraday cups. The contribution of Sr to mass 87 was monitored via mass 88 and in all cases the correction applied (using the 88Sr/86Sr ratio obtained for individual samples by TIMS) was negligible. An individual analysis typically consisted of 50 integrations of 8 seconds. 0.6M

HNO3 followed by a 0.6M HCl + 0.02M HF was used to ﬂush the system between analyses. The total washout time was about 15 minutes. The measured 85Rb/87Sr ratio was corrected for instrumental mass fractionation (0.3–0.4%/amu) using an

251 exponential correction and assuming 90Zr/91Zr=4.5882. Based on repeat analysis of rock standards, the reproducibility of Rb concentration measurements is better than 0.6%. Age corrections have been applied to the Sr and Nd isotope data assuming an age of 56 Ma for Nintoku Seamount [60, 71, 72].

9.7.3 Os Isotope Analyses

Os (187Os/188Os) isotope measurements were conducted at Johnson Space Center using a Finnigan Triton negative thermal ionization mass spectrometer (NTIMS). The protocol for 187Os/188Os isotope analysis was originally presented in Bran- don et al. [25] and was used for 187Os/188Os analyses with only minor diﬀerences.

Approximately 0.5 g of sample powder was spiked with a multi-element PGE spike

◦ and dissolved in reverse aqua regia (HCl:HNO3 = 1:2) at 240 C in quartz glass Carius tubes to obtain sample-spike equilibrium [25]. The Os was purified using carbon tetrachloride solvent extraction [47] followed by microdistillation purifica- tion [23]. The purified Os was loaded onto Pt filaments, allowed to evaporate on the filament and covered with Ba hydroxide and analyzed using negative thermal ion mass spectrometry (NTIMS) in negative mode at Johnson Space Center, Houston.

252 BIBLIOGRAPHY

1. G. Agiorgitis and R. Wolf, Aspects of osmium, ruthenium and iridium contents in some Greek chromites. Chemical Geology, 23: 267–272 (1978).

2. G. Agiorgitis and R. Wolf, Variations of the Os/Ir Ratio in Terrestrial Rocks and Minerals. Chemical Geology, 42: 277–286 (1984).

3. F. Albarede, Introduction to Geochemical Modeling. University of Cam- bridge, Cambridge (1995).

4. C. Allegre and D. Turcotte, Implications of a two-component marble-cake mantle. Nature, 323: 123–127 (1986).

5. D. Anderson, Superplumes or supercontinents. Geology, 22: 39–42 (1994).

6. D. Anderson, The sublithospheric mantle as the source of continental ﬂood basalts; the case against the continental lithosphere and plume head reservoirs. Earth and Planetary Science Letters, 123: 269–280 (1994).

7. D. Anderson, The EDGES of the Mantle. Geodynamics, 28: 255–271 (1998).

8. D. Anderson, The Thermal State of the Upper Mantle; No Role for Mantle Plumes. Geophysical Research Letters, 27(22): 3623–3626 (2000).

9. D. Anderson, Top-Down Tectonics. Science, 293: 2016–2018 (2001).

10. D. Anderson, Y.-S. Zhang and T. Tanimoto, Plume heads, continental lithosphere, ﬂood basalts and tomography. In B. Storey, T. Alabaster and R. Pankhurst, editors, Magmatism and the Causes of Continental Break-up, volume 68 of Special Publications, pages 99–124, Geological Society (1992).

11. M. Antretter, P. Riisager, S. Hall, X. Zhao and B. Steinberger, Modelled palaeolatitudes for the Louisville hot spot and the Ontong Java Plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 21–30, Geological Society, London (2004).

253 12. W. Bach and W. Irber, Rare earth element mobility in the oceanic lower sheeted dyke complex: evidence from geochemical data and leaching experiments. Chemical Geology, 151: 309–326 (1998).

13. S.-J. Barnes and A. Naldrett, Fractionation of the platinum-group elements and gold in some komatiites of the Abitibi Greenstone Belt, northern Ontario. Economic Geology, 82: 165–183 (1987).

14. S.-J. Barnes, A. Naldrett and M. Gorton, The origin of the fractionation of platinum-group elements in terrestrial magmas. Chemical Geology, 53: 303–323 (1985).

15. A. Basu and B. Faggart, Temporal isotopic variations in the Hawaiian mantle plume: The Lanai anomaly, the Molokai fracture zone and a seawater-altered lithospheric component in Hawaiian volcanism. In A. Basu and S. Hast, editors, Earth Processes: Reading the Isotopic Code, volume 95 of Geophysical Monograph, pages 149–159, American Geophysical Union (1996).

16. P. Beattie, Uranium-thorium disequilibrium and partitioning on melting of garnet peridotite. Nature, 363: 63–65 (1993).

17. R. Becker and G. Agiorgitis, Iridium, osmium, and palladium distribution in rock of the Troodos Complex, Cyprus. Chemical Erde, 37: 302–306 (1978).

18. A. Bence, S. Taylor and M. Fisk, Major- and trace-element geochemistry of basalts from Ojin, Nintoku, and Suiko Seamounts of the Emperor Seamount Chain: DSDP-IPOD Leg 55. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, pages 599–606, U.S. Government Printing Oﬃce, Washington (1980).

19. V. Bennett, T. Esat and M. Norman, Two mantle plume components in Hawaiian picrites inferred from correlated Os-Pb isotopes. Nature, 381: 221– 224 (1996).

20. V. Bennett, M. Norman and M. Garcia, Rhenium and platinum group element abundances correlated with mantle source components in Hawaiian picrites: sulphides in the plume. Earth and Planetary Science Letters, 183: 513–526 (2000).

21. M. Bielski-Zyskind, G. Wasserburg and P. Nixon, Sm-Nd and Rb-Sr System- atics in Volcanics and Ultramaﬁc Xenoliths From Malaita, Solomon Island, and the Nature of the Ontong Java Plateau. Journal of Geophysical Research, 89(B4): 2415–2424 (1984).

254 22. P. Bienvenu, H. Bougault, J. Joron, M. Treuil and L. Dmitriev, MORB alteration; rare-earth element/non-rare-earth hygromagmaphile element fractionation. Chemical Geology, 82: 1–14 (1990).

23. J. Birck, M. Roy-Barman and F. Capmas, Re-Os isotopic measurements at the femtomole level in natural samples. Geostandards Newsletter, 20(1): 19– 27 (1997).

24. A. Birkhold, A geochemical investigation of Ontong Java Plateau basement exposed on the Island of Makira (San Cristobal), Solomon Islands, South Paciﬁc. Phd thesis, University of Notre Dame (2000).

25. A. Brandon, M. Norman, R. Walker and J. Morgan, 186Os/187Os systematics of Hawaiian picrites. Earth and Planetary Science Letters, 174: 25–42 (1999).

26. A. Brandon, R. Walker, J. Morgan, M. Norman and H. Prichard, Coupled 186Os and 187Os evidence for core-mantle interaction. Science, 280: 1570– 1573 (1998).

27. A. Brandon, R. Walker, I. Puchtel, H. Becker, M. Humayan and S. Revillon, 186Os/187Os systematics of Gorgona Island komatiites: implications for early growth of the inner core. Earth and Planetary Science Letters, 206: 411–426 (2003).

28. C. Brooks, Analysis? Which analysis, of what and why. In 20th International Geochemical Exploration Symposium (2001).

29. G. Brugmann, N. Arndt, A. Hofmann and H. Tobschall, Noble metal abundances in komatiite suites from Alexo, Ontario and Gorgona Island, Columbia. Geochimica et Cosmochimica Acta, 51: 2159–2169 (1987).

30. I. Campbell and R. Griﬃths, Implications of mantle plume structure for the evolution of ﬂood basalts. Earth and Planetary Science Letters, 99: 79–93 (1990).

31. P. Castillo, R. Carlson and R. Batiza, Origin of the Nauru Basin igneous complex: Sr, Nd, and Pb isotope and REE constraints. Earth and Planetary Science Letters, 103: 200–213 (1991).

32. P. Castillo, M. Pringle and R. Carlson, East Mariana Basin tholeiites: Juras- sic ocean crust or Cretaceous rift basalts related to the Ontong Java Plateau. Earth and Planetary Science Letters, 123: 139–154 (1994).

33. L. Chambers, J. Fitton and M. Pringle, Phreatomagmatic eruptions on the Ontong Java Plateau: an Aptian 40Ar/39Ar age for volcaniclastic rocks at

255 ODP Site 1184. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 325–332, Geological Society, London (2004).

34. L. Chambers, M. Pringle and J. Fitton, Age and duration of magmatism on the Ontong Java Plateau: 40Ar/39Ar results from ODP Leg 192. EOS Transactions, 83: F47 (2002), Abstract V71B-1271.

35. W. Chazey and C. Neal, LIP magma petrogensis from source to surface: Platinum-group element evidence from Ontong Java Plateau basalts recovered during ODP Legs 130 and 192. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 219–238, Geological Society, London (2004).

36. C. Chen and F. Frey, Origin of Hawaiian tholeiite and alkalic basalt. Nature, 302: 785–789 (1983).

37. C. Chen and F. Frey, Trace element and isotopic geochemistry of lavas from Haleakala volcano, east Maui, Hawaii: Implications for the origin of Hawaiian basalts. Journal of Geophysical Research, 90(B10): 8743–8768 (1985).

38. C. Chen, F. Frey and M. Garcia, Evolution of alkalic lavas at Haleakala Volcano, East Maui, Hawaii; major, trace element and isotopic constraints. Contributions to Mineralogy and Petrology, 105: 197–218 (1990).

39. C. Chen, F. Frey, M. Garcia, G. Dalrymple and R. Hart, The tholeiite to alkalic basalt transition at Haleakala Volcano, Maui, Hawaii. Contributions to Mineralogy and Petrology, 106: 183–200 (1991).

40. Q. Cheng, K. Park, J. Macdougall, A. Zindler, G. Lugmair, H. Staudi- gel, J. Hawkins and P. Lonsdale, Isotopic evidence for a hotspot origin of the Louisville seamount chain. In B. Keating, P. Fryer, R. Batiza and G. Boehlert, editors, Seamounts, Islands, and Atolls, volume 43 of Geophys- ical Monograph, pages 283–296, American Geophysical Union (1987).

41. D. Clague and G. Dalrymple, Age of Koko Seamount, Emperor Seamount Chain. Earth and Planetary Science Letters, 17: 411–415 (1973).

42. D. Clague, M. Fisk and A. Bence, Mineral chemistry of basalts from Ojin, Nintoku, and Suiko Seamounts, Leg 55 DSDP. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, pages 607–638, U.S. Government Printing Oﬃce, Washington (1980).

256 43. D. Clague and F. Frey, Trace-element geochemistry of tholeittic basalts from Site 433C, Suiko Seamount. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, pages 559–570, U.S. Government Printing Office, Washington (1980). 44. D. Clague and F. Frey, Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu; implications for the oceanic mantle below Hawaii. Journal of Petrology, 23(3): 447–504 (1982). 45. C. Class and S. Goldstein, Plume-lithosphere interactions in the ocean basins: constraints from the source mineralogy. Earth and Planetary Science Letters, 150: 245–260 (1997). 46. M. Coffin and O. Eldholm, Large igneous provinces: crustal structure, dimen- sions, and external consequences. Reviews of Geophysics, 32: 1–36 (1994). 47. A. Cohen and F. Waters, Separation of osmium from geological materials by solvent extraction for analyses by TIMS. Analytica Chimica Acta, 332: 269–275 (1996). 48. P. Coleman, Stratigraphical and structural notes on the British Solomon Is- lands with refernce to the first geological map, 1962. Rep. 29, British Solomon Islands Geological Record (1965). 49. P. Coleman, The Solomon Islands as an island arc. Nature, 211: 1249–1251 (1966). 50. P. Coleman, Geology of the Solomon and New Hebrides Islands, as part of the Melanesian Re-entrant, SW Pacific. Pacific Science, 24: 289–314 (1970). 51. P. Coleman and L. Kroenke, Subduction without volcanism in the Solomon Islands arc. Geo-Marine Letters, 1: 129–134 (1981). 52. R. Coleman, E. Lee, L. Beatty and W. Brannock, Eclogites and eclogites: Their differences and similarities. Geological Society of America Bulletin, 76: 483–508 (1965). 53. K. Condie, Plate Tectonics and Crustal Evolution. Butterworth-Heinemann, fourth edition (1997). 54. P. Cooper and B. Taylor, Polarity reversal in the Soloman Islands arc. Na- ture, 314: 428–430 (1985). 55. J. Cotten, A. L. Dez, M. Bau, M. Caroff, R. Maury, P. Dulski, S. Four- cade, M. Bohn and R. Brousse, Origin of anomalous rare-earth element and yttrium enrichments in subaerially exposed basalts: Evidence from French Polynesia. Chemical Geology, 119: 115–138 (1995).

257 56. K. Cox, J. Bell and R. Pankhurst, The Interpretation of Igneous Rocks. Allen and Unwin, London (1979).

57. J. Crocket, Platinum-Group Element Geochemistry of Mafic and Ultramafic Rocks. In L. Cabri, editor, The Geology, Geochemistry, Mineralogy and Mineral Beneficiation of Platinum-Group Elements, volume 54 of CIM Spe- cial Volume, pages 177–210, Canadian Institute of Mining, Metallurgy and Petroleum (2002).

58. S. Dalheimer, C. Neal and J. Fein, Potential siderophore mobilization of platinum-group elements in surface environments. Environmental Science and Technology, in press (2006).

59. G. Dalrymple and D. Clague, Age of the Hawaiian-Emperor bend. Earth and Planetary Science Letters, 31: 313–329 (1976).

60. G. Dalrymple and M. Garcia, Age and chemistry of volcanic rocks dredged from Jingu Seamount, Emperor Seamount Chain. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, U.S. Government Printing Oﬃce, Washington (1980).

61. G. Dalrymple, M. Lanphere and D. Clague, Conventional and 40Ar/39Ar K- Ar ages of volcanic rocks from Ojin (Site 430), Nintoku (Site 432), and Suiko (Site 433) seamounts and the chronology of volcanic propagation along the Hawaiian-Emperor chain. In E. Jackson and I. Koisumi, editors, Initial Re- ports of the Deep Sea Drilling Project, volume 55, U.S. Government Printing Oﬃce, Washington (1980).

62. G. Davies, Mantle plumes, mantle stirring and hotspot chemistry. Earth and Planetary Science Letters, 99: 94–109 (1990).

63. T. Davies, P. Wilde and D. Clague, Koko seamount: a maor guyot at the southern end of the Emperor seamounts. Marine Geology, 13: 311–321 (1972).

64. G. Davis, The ages and uranium contents of zircons from kimberlites and as- sociates rocks. In Extended Abstracts from the Second Kimberlite Conference, Arizona (1977).

65. H. DeBaar, M. Bacon and P. Brewer, Rare earth elements in the eastern equatorial Paciﬁc Ocean. EOS Transactions, page 1030 (1983).

66. H. DeBaar, M. Bacon, P. Brewer and K. Bruland, Rare earth elements in the Paciﬁc and Atlantic Oceans. Geochimica et Cosmochimica Acta, 49: 1943– 1959 (1985).

258 67. D. DePaolo, J. Bryce, A. Dodson, D. Shuster and B. Kennedy, Iso- topic evolution of Mauna Loa and the chemical structure of the Hawai- ian plume. Geochemistry, Geophysics, Geosystems, 2 (2001), Paper number: 2000GC000139.

68. D. DePaolo and B. Ingram, High-resolution stratigraphy with Sr isotopes. Science, 227: 938–941 (1985).

69. C. Devey, C.-D. Garbe-Schönberg, P. Stoffers, C. Chauvel and D. Mertz, Geochemical effects of dynamic melting beneath ridges: Reconciling major and trace element variations in Kolbeinsey (and global) mid-ocean ridge basalt. Journal of Geophysical Research, 99(B5): 9077–9095 (1994).

70. I. Duddy, Redistribution and fractionation of rare-earth and other elements in a weathered proﬁle. Chemical Geology, 30: 363–381 (1980).

71. R. Duncan, A time frame for construction of the Kerguelen Plateau and Broken Ridge. Journal of Petrology, 43(7): 1109–1119 (2002).

72. R. Duncan and R. Keller, Radiometric ages for basement rocks from the Emperor Seamounts, ODP Leg 197. Geochemistry, Geophysics, Geosystems, 5 (2004), Paper number: 10.1029/2004GC000704.

73. J. Eiselle, W. Abouchami, S. Galer and A. Hofmann, The 320 kyr Pb isotope evolution of Mauna Kea lavas recorded in the HSDP-2 drill core. Geochemistry, Geophysics, Geosystems, 4 (2003), Paper number: 10.1029/2002GC000339.

74. H. Elderﬁeld and M. Greaves, The rare earth elements in seawater. Nature, 296: 214–219 (1982).

75. J. Ely and C. Neal, Using platinum-group elements to investigate the origin of the Ontong Java Plateau, SW Paciﬁc. Chemical Geology, 196: 235–257 (2003).

76. J. Ely, C. Neal, J. O’Neill and J. Jain, Quantifying the platinum group elements (PGEs) and gold in geological samples using cation exchange pre- treatment and ultrasonic nebulization inductively coupled plasma-mass spectrometry (USN-ICP-MS). Chemical Geology, 157: 219–234 (1999).

77. M. Feigenson, L. Bolge, M. Carr and C. Herzberg, REE inverse modeling of HSDP-2 basalts: Evidence for multiple sources in the Hawaiian plume. Geochemistry, Geophysics, Geosystems, 4 (2003), Paper number: 10.1029/2001GC000207.

259 78. J. Fitton and M. Godard, Origin and evolution of magmas on the Ontong Java Plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 151–178, Geological Society, London (2004).

79. J. Fitton, J. Mahoney, P. Wallace and A. Saunders, Origin and evolution of the Ontong Java Plateau: introduction. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 1–8, Geological Society, London (2004).

80. F. Flanagan, Three USGS maﬁc rock reference samples, W-2, DNC-1, and BIR-1. US Geological Survey Bulletin, 1623: 54 (1984).

81. C. Ford, D. Russell, J. Craven and M. Fisk, Olivine liquid equilibria: Tem- perature, pressure and composition dependence of the crystal/liquid cation partition coeﬃcients for Mg, Fe2+, Ca and Mn. Journal of Petrology, 24: 256–265 (1983).

82. F. Frey, Rare earth abundances in a high-temperature peridotite intrusion. Geochimica et Cosmochimica Acta, 33(11): 1429–1447 (1969).

83. F. Frey, D. Clague, J. Mahoney and J. Sinton, Volcanism at the edge of the Hawaiian Plume: petrogenesis of submarine alkalic lavas from the North Arch Volcanic Field. Journal of Petrology, 41: 667–691 (2000).

84. F. Frey, M. Garcia and M. Roden, Geochemical characteristics of Koolau Volcano; implications of intershield geochemical diﬀerences among Hawaiian volcanoes. Geochimica et Cosmochimica Acta, 58(5): 1441–1462 (1994).

85. F. Frey, S. Huang, J. Blichert-Toft, M. Regelous and M. Boyet, Origin of depleted components in basalt related to the Hawaiian hot spot: Evidence from isotopic and incompatible element ratios. Geochemistry, Geophysics, Geosystems, 6(1) (2005).

86. F. Frey and J. Rhodes, Intershield geochemical diﬀerences among Hawai- ian volcanoes: Implications for source compositions, melting processes and magma ascent paths. Philosophical Transactions of the Royal Society of Lon- don, 352: 121–136 (1993).

87. F. Frey and M. Roden, The mantle source for the Hawaiian Islands; constraints from the lavas and ultramaﬁc inclusions. In ??, editor, Mantle Meta- somatism, volume ??, pages ??–??, Academic Press, London (1987).

260 88. F. Frey, W. Wise, M. Garcia, H. West, S. Kwon and A. Kennedy, Evolution of Mauna Kea Volcano, Hawaii; petrologic and geochemical constraints on postshield volcanism. Journal of Geophysical Research, 95(B2): 1271–1300 (1990).

89. R. Gijbels, H. Millard, G. Desborough and A. Bartel, Osmium, ruthenium, iridium and uranium in silicates and chromites from the Eastern Bushveld Complex, S. Africa. Geochimica et Cosmochimica Acta, 38: 319–337 (1974).

90. T. Gladczenko, M. Coﬃn and O. Eldholm, Crustal structure of the Ontong Java Plateau: modeling of new gravity and existing seismic data. Journal of Geophysical Research, 102: 22711–22729 (1997).

91. E. Gladney and I. Roelandts, 1987 compilation of elemental concentration data for USGS BIR-1, DNC-1, and W-2. Geostandards Newsletter, 12: 63– 118 (1988).

92. K. Govindaraju, 1994 compilation of working values and sample description for 383 geostandards. Geostandards Newsletter, 18: 158 (1994).

93. D. Graham, A. Zindler, M. Kurz, W. Jenkins, R. Batiza and H. Staudigel, He, Pb, Sr and Nd isotope constraints on magma genesis and mantle heterogeneity beneath young Paciﬁc seamounts. Contributions to Mineralogy and Petrology, 99(4): 446–463 (1988).

94. T. Green, Experimental studies of trace-element partitioning applicable to igneous petrogenesis-Sedona 16 years later. Chemical Geology, 117: 1–36 (1994).

95. R. Griﬃths and I. Campbell, Stirring and structure in mantle starting plumes. Earth and Planetary Science Letters, 99: 66–78 (1990).

96. P. Hack, R. Nielson and A. Johnston, Experimentally determined rare-Earth element and Y partitioning behavior between clinopyroxene and basaltic liquids at pressures up to 20 kbar. Chemical Geology, 117: 89–105 (1994).

97. A. Halliday, D. Lee, S. Tommasini, G. Davies, C. Paslick, J. Fitton and D. James, Incompatible trace elements in OIB and MORB and source enrichment in the sub-oceanic mantle. Earth and Planetary Science Letters, 133: 379–395 (1995).

98. M. Handler, V. Bennett and G. Dreibus, Evidence from correlated Ir/Os and Cu/S for late-stage Os mobility in peridotite xenoliths: Implications for Re-Os systematics. Geology, 27(1): 75–78 (1999).

261 99. R. Hart, Chemical exchange between sea water and deep ocean basins. Earth and Planetary Science Letters, 95: 269–279 (1970).

100. S. Hart, J. Blusztajn, H. Dick, P. Meyer and K. Muehlenbachs, The ﬁnger- print of seawater circulation in a 500-meter section of ocean crust gabbros. Geochimica et Cosmochimica Acta, 63(4): 4059–4080 (1999).

101. S. Hart and T. Dunn, Experimental cpx/melt partitioning on melting of garnet peridotite. Nature, 363: 63–65 (1993).

102. S. Hart, A. Erlank and E. Kable, Sea ﬂoor basalt alteration: Some chemical and Sr isotopic eﬀects. Contributions to Mineralogy and Petrology, 44: 219– 230 (1974).

103. E. Hauri and S. Hart, Re-Os isotope systematics of HIMU and EMII oceanic island basalts from the south Paciﬁc Ocean. Earth and Planetary Science Letters, 114: 353–371 (1993).

104. E. Hauri and S. Hart, Rhenium abundances and systematics in oceanic basalts. Chemical Geology, 139: 185–205 (1997).

105. E. Hauri, J. Lassiter and D. DePaolo, Osmium isotope systematics of drilled lavas from Mauna Loa, Hawaii. Journal of Geophysical Research, 101(B): 11793–11806 (1996).

106. M. Hawkins and A. Barron, The geology and mineral resources of Santa Is- abel, Solomon Islands. Technical Report 114 pp + maps, Ministry of Natural Resources, Honiara (1991).

107. C. Herzberg, Partial melting below the Ontong Java Plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 179–183, Geological Society, London (2004).

108. J. Hess, M. Bender and J. Schilling, Evolution of the ratio of strontium-87 to strontium-86 in seawater from Cretaceous to present. Science, 231: 979–984 (1986).

109. K. Hoernle, R. Werner, J. P. Morgan, D. Garbe-Sch¨onberg, J. Bryce and J. Mrazek, Existence of complex spatial zonation in the Galapagos plume for at least 14 m.y. Geology, 28: 435–438 (2000).

110. E. Hoﬀman and B. Dunn, Sample Preparation and Bulk Analytical Meth- ods for PGE. In L. Cabri, editor, The Geology, Geochemistry, Mineralogy and Mineral Beneﬁciation of Platinum-Group Elements, volume 54 of CIM

262 Special Volume, pages 1–12, Canadian Institute of Mining, Metallurgy and Petroleum (2002).

111. A. Hofmann, M. Feigenson and I. Raczek, Kohala revisited. Contributions to Mineralogy and Petrology, 95(1): 114–122 (1987).

112. S. Huang, M. Regelous, T. Thordarson and F. Frey, Petrogenesis of Lavas from Detroit Seamount: Geochemical diﬀerences between Emperor Chain and Hawaiian Volcanoes. Geochemistry, Geophysics, Geosystems, 6(Q01L06) (2005).

113. M. Humayan, L. Qin and M. Norman, Geochemical Evidence for Excess Iron in the Mantle Beneath Hawaii. Science, 306: 91–94 (2004).

114. S. Ingle and M. Coﬃn, Impact origin for the greater Ontong Java Plateau? Earth and Planetary Science Letters, 218: 123–134 (2004).

115. A. Irving and F. Frey, Trace-element abundances in megacrysts and their host basalts - constraints on partition-coeﬃcients and megacryst genesis. Geochimica et Cosmochimica Acta, 48(6): 1201–1221 (1984).

116. A. Ishikawa, S. Maruyama and T. Komiya, Layered Lithospheric Mantle Beneath the Ontong Java Plateau: Implications from Xenoliths in Aln¨oite, Malaita, Solomon Islands. Journal of Petrology, 45(10): 2011–2044 (2004).

117. P. Janney and P. Castillo, Basalts from the central Paciﬁc Basin; evidence for the origin of Cretaceous igneous complexes in the Jurassic western Paciﬁc. Journal of Geophysical Research, 101: 2875–2893 (1996).

118. G. Jenner, S. Foley, S. Jackson, T. Green, B. Fryer and H. Longerich, Determination of partition coeﬃcients for trace elements in high pressure- temperature experimental run products by laser ablation microprobe- inductively coupled plasma-mass spectrometry (LAM-ICP-MS). Geochimica et Cosmochimica Acta, 57(23/24): 5099–5103 (1993).

119. G. Jenner, H. Longerich, S. Jackson and B. Fryer, ICP-MS: A powerful tool for high-precision trace-element analysis in Earth sciences: Evidence from analysis of selected USGS reference samples. Chemical Geology, 83: 133–148 (1990).

120. Y. Ji and H. Hataf, Detection of mantle plumes in the lower mantle by diﬀraction tomography: Hawaii. Earth and Planetary Science Letters, 159: 99–115 (1998).

263 121. K. Johnson and R. Kinzler, Partitioning of REE, Ti, Zr, Hf, and Nb between clinopyroxene and basaltic liquid: an ion microprobe study. EOS Transac- tions, 70: 1388 (1989).

122. P. Keleman, N. Shimizu and T. Dunn, Relative depletion of niobium in some arc magmas and the continental crust: partitioning of K, Nb, La, and Ce during melt/rock reaction in the upper mantle. Earth and Planetary Science Letters, 120: 111–134 (1993).

123. R. Keller, R. Duncan and M. Fisk, Geochemistry and 40Ar/39Ar geochronology of basalts from ODP Leg 145 (North Paciﬁc Transect). In D. Rea, I. Basov, D. Scholl and J. Allan, editors, Proceedings of the Ocean Drilling Program, Scientiﬁc Results, volume 145, pages 333–344, Ocean Drilling Pro- gram, College Station, TX (1995).

124. R. Keller, M. Fisk and W. White, Isotopic evidence for Late Cretaceous plume-ridge interaction at the Hawaiian hotspot. Nature, 405: 673–676 (2000).

125. A. Kennedy, S. Kwon, F. Frey and H. West, The isotopic composition of post- shield lavas from Mauna Kea volcano, Hawaii. Earth and Planetary Science Letters, 103: 339–353 (1991).

126. A. Kennedy, G. Lofgren and G. Wasserburg, An experimental study of trace element partitioning between olivine, orthopyroxene, and melt in chondrules: equilbrium values and kinetic eﬀects. Earth and Planetary Science Letters, 115: 177–195 (1993).

127. Y. Kikawada, T. Ossaka, T. Oi and T. Honda, Experimental studies on the mobility of lanthanides accompanying alteration of andesite by acidic hot spring water. Chemical Geology, 176: 137–149 (2001).

128. S. King and D. Anderson, An alternative mechanism of ﬂood basalt formation. Earth and Planetary Science Letters, 136: 269–279 (1995).

129. W. Kinman and C. Neal, The inﬂuence of MORB on a plume-generated seamount: the story told by plagioclase phenocrysts from Detroit Seamount Lavas, Emperor Seamount Chain. In AGU Fall Meeting, volume 83, Ameri- can Geophysical Union, San Francisco (2002), Abstract T61C.

130. R. Kirkpatrick, D. Clague and W. Freisen, Petrology and geochemistry of volcanic rocks, DSDP Leg 55, Emperor Seamount Chain. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, pages 509–558, U.S. Government Printing Oﬃce, Washington (1980).

264 131. J. Kramers, J. Roddick and J. Dawson, Trace element and isotopic studies on veined, metasomatic, and “MARID” xenoliths from Bultfontein, South Africa. Earth and Planetary Science Letters, 65: 90–106 (1983).

132. L. Kroenke and J. Mahoney, Rifting of the Ontong Java Plateau’s eastern salient and seaﬂoor spreading in the Ellis Basin; relation to the 90 Ma eruptive episode on the plateau. EOS Transactions, 46(77): 713 (1996).

133. L. Kroenke, P. Wessel and A. Sterling, Motion of the Ontong Java Plateau in the hot-spot frame of reference: 122 Ma–present. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 9–20, Geological Society, London (2004).

134. A. Kurtz, L. Derry, O. Chadwick and M. Alfano, Refractory element mobility in volcanic soils. Geology, 28(8): 683–686 (2000).

135. M. Kurz, T. Kenna, J. Rhodes and M. Garcia, Isotopic evolution of Mauna Loa volcano: A view from the submarine southwest rift zone. In J. Rhodes and J. Lockwood, editors, Mauna Loa Revealed: Structure, Composition, History, and Hazards, volume 92 of Geophysical Monograph, American Geo- physical Union (1995).

136. E. Kushcel and I. Smith, Rare earth mobility in young arc-type volcanic rocks from northern New Zealand. Geochimica et Cosmochimica Acta, 56: 3951–3955 (1992).

137. M. Lanphere, G. Dalrymple and D. Clague, Rb-Sr sytematics of basalts from the Hawaiian-Emperor Volcanic Chain. In E. Jackson and I. Koisumi, editors, Initial Reports of the Deep Sea Drilling Project, volume 55, pages 695–706, U.S. Government Printing Oﬃce, Washington (1980).

138. M. Lanphere and F. Frey, Geochemical evolution of Kohala Volcano, Hawaii. Contributions to Mineralogy and Petrology, 95(1): 100–113 (1987).

139. R. Larson, Geological consequences of superplumes. Geology, 19: 963–966 (1991).

140. R. Larson, Latest pulse of Earth: Evidence for a Mid-Cretaceous superplume. Geology, 19: 547–550 (1991).

141. J. Lassiter, D. DePaolo and M. Tatsumoto, Isotopic evolution of Mauna Kea volcano: results from the initial phase of the Hawaii Scientiﬁc Drilling Project. Journal of Geophysical Research, 101(B): 11769–11780 (1996).

265 142. J. Lassiter and E. Hauri, Osmium-isotope variations in Hawaiian lavas: evidence for recycled oceanic lithosphere in the Hawaiian plume. Earth and Planetary Science Letters, 164: 483–496 (1998).

143. J. Lassiter, E. Hauri, I. Nikogosian and H. Barsczus, Chlorine-potassium variations in melt inclusions from Raivavae and Rapa, Austral Islands: constraints on chlorine recycling in the mantle and evidence for brine-induced melting of oceanic crust. Earth and Planetary Science Letters, 202: 525–540 (2002).

144. J. Lassiter, E. Hauri, P. Reiners and M. Garcia, Generation of Hawaiian post- erosional lavas by melting of a mixed lherzolite/pyroxenite source. Earth and Planetary Science Letters, 178: 269–284 (2000).

145. M. Le Bas, R. Le Maitre, A. Streckeisen and B. Zanettin, Chemical classi- ﬁcation of igneous rocks based on the total alkali-silica diagram. Journal of Petrology, 27: 745–750 (1986).

146. D. Lee, A. Halliday, G. Davies, E. Essene, J. Fitton and R. Temdjin, Melt enrichment of shallow depleted mantle: a detailed petrological, trace element, and isotopic study of mantle derived xenoliths and megacrysts from the Cameroon line. Journal of Petrology, ??: ??–?? (1995).

147. W. Leeman, J. Budahn, D. Gerlach, D. Smith and B. Powell, Origin of Hawaiian tholeiites: trace element constraints. American Journal of Science, 280-A: 794–819 (1980).

148. H. Longerich, G. Jenner, B. Fryer and S. Jackson, Inductively coupled plasma-mass spectrometric analysis of geological samples: A critical evaluation based on case studies. Chemical Geology, 83: 105–118 (1990).

149. P. Lonsdale, J. Dier and J. Natland, Posterosional volcanism in the Creta- ceous part of the Hawaiian Hotspot Trail. Journal of Geophysical Research, 98(B3): 4081–4098 (1993).

150. M. Loubet, R. Sassi and G. D. Donato, Mantle heterogeneities: a combined isotope and trace element approach and evidence for recycled continental crust materials in some OIB sources. Earth and Planetary Science Letters, 89: 299–315 (1988).

151. J. Ludden and G. Thompson, An evaluation of the behavior of the rare earth elements during the weathering of sea-ﬂoor basalt. Earth and Planetary Science Letters, 43: 85–92 (1979).

266 152. G. Lugmair and S. Galer, Age and isotopic relationships among the angrites Lewis Cliﬀ 86010 and Angra dos Reis. Geochimica et Cosmochimica Acta, 56: 85–92 (1992).

153. G. Macdonald and T. Katsura, Chemical compositions of Hawaiian lavas. Journal of Petrology, 5(1): 82–133 (1964).

154. J. Mahoney, An isotopic survey of Paciﬁc oceanic plateaus: implications for their nature and origin. In B. Keating, P. Fryer, R. Batiza and G. Boehlert, editors, Seamounts, Islands, and Atolls, volume 43 of Geophysical Mono- graph, pages 207–220, American Geophysical Union (1987).

155. J. Mahoney, J. Fitton and P. Wallace, Proceedings of the Ocean Drilling Pro- gram, Initial Reports, 192. [CD-ROM], Available from: Ocean Drilling Pro- gram, Texas A&M University, College Station TX 77845-9547, USA (2001).

156. J. Mahoney, J. Sinton, M. Kurz, J. Macdougall, K. Spencer and G. Lugmair, Isotope and trace element characteristics of a super-fast spreading ridge; East Paciﬁc Rise, 12-23 degrees S. Earth and Planetary Science Letters, 121(1-2): 173–193 (1994).

157. J. Mahoney and K. Spencer, Isotopic evidence for the origin of the Manihiki and Ontong Java oceanic plateaus. Earth and Planetary Science Letters, 104: 196–210 (1991).

158. J. Mahoney, M. Storey, R. Duncan, K. Spencer and M. Pringle, Geochemistry and age of the Ontong Java Plateau. In M. Pringle, W. Sager, W. Sliter and S. Stein, editors, The Mesozoic Paciﬁc: Geology, Tectonics, and Volcanism, volume 77 of Geophysical Monograph, pages 233–262, American Geophysical Union (1993).

159. J. Mammerickx and G. Sharman, Tectonic evolution of the North Paciﬁc during the Cretaceous Quiet Period. Journal of Geophysical Research, 93: 3009–3024 (1988).

160. J. Marsh, REE fractionation and Ce anomalies in weathered Karoo dolerite. Chemical Geology, 90: 189–194 (1991).

161. W. McDonough and S. Sun, The composition of the Earth. Chemical Geol- ogy, 120: 223–253 (1995).

162. C. McGinnis, J. Jain and C. Neal, Characterization of memory eﬀects and development of an eﬀective protocol for the measurement of petrogeneti- cally critical trace elements in geological samples by ICP-MS. Geostandards Newsletter, 21: 289–305 (1997).

267 163. D. McKenzie and R. O’Nions, Partial Melt Distributions from Inversion of Rare Earth Elements Concentrations. Journal of Petrology, 12(5): 1031–1091 (1991).

164. M. Menzies and W. Seyfried, Basalt-seawater interactions; trace element and strontium isotopic variations in experimentally altered glassy basalt. Earth and Planetary Science Letters, 44: 463–472 (1979).

165. A. Miyashiro, Nature of alkalic volcanic rock series. Contributions to Miner- alogy and Petrology, 66(1): 91–104 (1978).

166. P. Molnar and T. Atwater, Relative motion of hotspots in the mantle. Nature, 246: 288–291 (1973).

167. J. Morgan, Convective plumes in the lower mantle. Nature, 230: 42–43 (1971).

168. H. Nasagawa, Rare Earth concentrations in zircon and apatite and their dacite and granites. Earth and Planetary Science Letters, 9: 359–364 (1970).

169. C. Neal, Mantle Studies in the Western Paciﬁc and Kimberlite-type Intru- sives. Ph.D. thesis, University of Leeds (1985).

170. C. Neal, Interior of the Moon: The presence of garnet in the primitive deep lu- nar mantle. Journal of Geophysical Research, 106(E11): 27865–27885 (2001).

171. C. Neal and J. Davidson, An unmetasomatized source for the Malaitan alnoite (Solomon Islands): Petrogenesis involving zone reﬁning, megacryst fractionation, and assimilation of oceanic lithosphere. Geochimica et Cos- mochimica Acta, 53: 1975–1990 (1989).

172. C. Neal, J. Mahoney, L. Kroenke, R. Duncan and M. Petterson, The On- tong Java Plateau. In J. Mahoney and M. Coﬃn, editors, Large Igneous Provinces: Continental, Oceanic, and Planetary Flood Volcanism, volume 100 of Geophysical Monograph, pages 183–216, American Geophysical Union (1997).

173. C. Neal and L. Taylor, A negative Ce anomaly in a peridotite xenolith: evidence for crustal recycling into the mantle or mantle metasomatism? Geochimica et Cosmochimica Acta, 53: 1035–1040 (1989).

174. C. Neal, L. Taylor, J. Davidson, P. Holden, A. Halliday, P. Nixon, J. Paces, R. Clayton and T. Mayeda, Eclogites with oceanic crustal and mantle signatures from the Bellsbank kimberlite, South Africa, part 2: Sr, Nd, and O isotope geochemistry. Earth and Planetary Science Letters, 99: 362–379 (1990).

268 175. P. Nimis, Clinopyroxene geobarometry of magmatic rocks. Part 2. Structural geobarometers for basic to acid, tholeiitic and mildly alkaline magmatic systems. Contributions to Mineralogy and Petrology, 135: 62–74 (1999). 176. P. Nimis and W. Taylor, Single clinopyroxene thermobarometry for garnet peridotites. Part I. Calibration and testing of a Cr-in-Cpx barometer and an enstatite-in-Cpx thermometer. Contributions to Mineralogy and Petrology, 139: 541–554 (2000). 177. P. Nimis and P. Ulmer, Clinopyroxene geobarometry of magmatic rocks. Part I. An expanded structural geobarometer for anhydrous and hydrous, basic and ultrabasic systems. Contributions to Mineralogy and Petrology, 133: 122–135 (1998). 178. Y. Niu, D. Bideau, R. Hekinian and R. Batiza, Mantle compositional control on the extent of mantle melting, crust production, gravity anomaly, ridge morphology, and ridge segmentation: a case study at Mid-Atlantic Ridge 33-35 ◦N. Earth and Planetary Science Letters, 186: 383–399 (2001). 179. Y. Niu, K. Collerson, R. Batiza, J. Wendt and M. Regelous, Origin of enriched-type mid-ocean ridge basalt at ridges far from mantle plumes: the East Pacific Rise at 11◦20’N. Journal of Geophysical Research, 104: 7067– 7087 (1999). 180. Y. Niu and M. O’Hara, Origin of ocean island basalts: a new perspective from petrology, geochemistry and mineral physics consideration. Journal of Geo- physical Research, 108(B) (2003), Paper number: 10.1029/2002JB002048. 181. Y. Niu, M. Regelous, I. Wendt, R. Batiza and M. O’Hara, Geochemistry of near-EPR seamount: importance of source vs. process and the origin of enriched mantle component. Earth and Planetary Science Letters, 199: 327– 345 (2002). 182. P. Nixon and F. Boyd, Garnet bearing lherzolites and discrete nodule suites from the Malaita Alnoite, Solomon Islands, S.W. Pacific, and their bearing on oceanic mantle composition and geotherm. Bulletin of the Australian Society of Exploratory Geophysics, 9(3): 103–107 (1978). 183. P. Nixon and P. Coleman, Garnet-Bearing Lherzolites and Discrete Nodule Suites from the Malaita Alnoite, Solomon Islands, and their Bearing on the Nature and Origin of the Ontong Java Plateau. Bulletin of the Australian Society of Exploratory Geophysics, 9(3): 103–107 (1978). 184. P. Nixon, R. Mitchell and N. Rogers, Petrogenesis of alnöitic rocks from Malaita, Solomon Islands, Melanesia. Mineralogical Magazine, 43: 587–596 (1980).

269 185. P. Nixon and C. Neal, Ontong Java Plateau: deep-seated xenoliths from thick oceanic lithosphere. In P. Nixon, editor, Mantle Xenoliths, chapter 24, pages 335–345, John Wiley & Sons Ltd (1987).

186. I. Norton, Plate motions in the North Paciﬁc: The 43 Ma Nonevent. Tecton- ics, 14: 1080–1094 (1995).

187. I. Parkinson, R. Arculus, E. McPherson and R. Duncan, Geochemistry, tectonics and the peridotites of the northeastern Solomon Islands. EOS Trans- actions, 76(Fall Meeting Supplement): 642 (1996).

188. I. Parkinson, B. Schaefer and R. Arculus, A lower mantle origin for the world’s biggest LIP? A high precision Os isotope isochron from Ontong Java Plateau basalts drilled on ODP Leg 192. Geochimica et Cosmochimica Acta, 66(Supplement, A580) (2002).

189. B. Parsons and J. Sclater, An analysis of the variation of ocean ﬂoor bathymetry and heat ﬂow with age. Journal of Geophysical Research, 82: 803–827 (1977).

190. D. Peterson and R. Moore, Geologic history and evolution of geologic con- cepts, island of Hawaii. Professional Paper 1350, U.S. Geological Survey (1987).

191. M. Petterson, The geology of north and central Malaita, Solomon Is- lands. (Including implications of geological research on Makira, Savo, Is- abel, Guadalcanal, and Choiseul between 1992 and 1995). Geol. Mem., 1/95 (1995).

192. M. Petterson, The geology of north and central Malaita, Solomon Islands: the thickest and most accessible part of the world’s largest (Ontong Java) ocean plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 63–81, Geological Society, London (2004).

193. M. Petterson, T. Babbs, C. Neal, J. Mahoney, A. Saunders, R. Duncan, D. Tolia, R. Magu, C. Qopoto, H. Mahoa and D. Natogga, Geological- tectonic framework of Solomon Islands, SW Paciﬁc: crutal accretion and growth within an intra-oceanic setting. Tectonophysics, 301: 35–60 (1999).

194. M. Petterson, C. Neal, J. Mahoney, L. Kroenke, A. Saunders, T. Babbs, R. Duncan, D. Tolia and B. McGrail, Structure and deformation of north and central Malaita, Solomon Islands: tectonic implications for the Ontong Java Plateau—Solomon arc collision, and for the fate of oceanic plateaus. Tectonophysics, 283: 1–33 (1997).

270 195. J. Phipps Morgan, Isotope Topology of Individual Hotspot Basalt Arrays: Mixing Curves or Melt Extraction Trajectories. Geochemistry, Geophysics, Geosystems, 1 (1999), Paper number:1999GC000004. 196. J. Phipps Morgan and J. Morgan, Two-stage melting and the geochemical evolution of the mantle: a recipe for mantle plum-pudding. Earth and Plan- etary Science Letters, 170: 215–239 (1999). 197. A. Pietruszka and M. Garcia, A rapid fluctuation in the mantle source and melting history of Kilauea Volcano inferred from the geochemistry of its historical summit lavas (1790-1982). Journal of Petrology, 40(8): 1321–1342 (1999). 198. H. Powers, Composition and origin of basaltic magma of the Hawaiian Is- lands. Geochimica et Cosmochimica Acta, 7(1-2): 77–107 (1955). 199. R. Price, C. Gray, R. Wilson, F.A and S. Taylor, The effects of weathering on rare-earth element, Y and Ba abundances in Tertiary basalts from southeastern Austrailia. Chemical Geology, 93: 245–265 (1991). 200. G. Ravizza and D. Pyle, PGE and Os isotopic analyses of single sample aliquots with NiS fire assay preconcentration. Chemical Geology, 141: 251– 268 (1997). 201. M. Regelous, A. Hofmann, W. Abouchami and S. Galer, Geochemistry of lavas from the Emperor Seamounts, and the geochemical evolution of Hawai- ian magmatism from 85 to 42 Ma. Journal of Petrology, 44(1): 113–140 (2003). 202. M. Regelous, Y. Niu, I. Wendt, R. Batiza, A. Greig and K. Collerson, Varia- tions in the geochemistry of magmatism on the East Pacific Rise at 10, 30′ ◦N since 800 ka. Earth and Planetary Science Letters, 168: 45–63 (1999). 203. M. Rehkamper and A. Halliday, Development and application of new ion- exchange techniques for the separation of the platinum group and other siderophile elements from geological samples. Talanta, 44: 663–672 (1997). 204. M. Rehkamper, A. Halliday, J. Fitton, D. Lee, M. Wieneke and N. Arndt, Ir, Ru, Pt, and Pd in basalts and komatiites: New constraints for the geochemical behavior of the platinum-group elements in the mantle. Geochimica et Cosmochimica Acta, 63(22): 3915–3934 (1999). 205. P. Reiners and B. Nelson, Temporal-compositional-isotopic trends in rejuvenated-stage magmas of Kauai, Hawaii, and implications for mantle melting processes. Geochimica et Cosmochimica Acta, 62(13): 2347–2368 (1998).

271 206. Z.-Y. Ren, S. Ingle, E. Takahashi, N. Hirano and T. Hirata, The chemical structure of the Hawaiian mantle plume. Nature, 436(7052): 837–840 (2005).

207. S. Revillon, D. Teagle, P. Boulvais, J. Shafer and C. Neal, Geochemical ﬂuxes associated with the alteration of a subaerially-emplaced seamount: Nintoku Seamount, ODP Leg 197, Site 1205. In preparation (2006).

208. N. Ribe and U. Christensen, Three-dimensional modeling of plume- lithosphere interaction. Journal of Geophysical Research, 99: 669–682 (1994).

209. N. Ribe and U. Christensen, The dynamical origin of Hawaiian volcanism. Earth and Planetary Science Letters, 171: 517–531 (1999).

210. M. Richards, R. Duncan and V. Courtillot, Flood basalts and hot-spot tracks: plume heads and tails. Science, 246: 103–107 (1989).

211. W. Richardson, E. Okal and S. Van der Lee, Rayleigh-wave tomography of the Ontong Java Plateau. Physics of the Earth and Planetary Interiors, 118: 29–61 (2000).

212. P. Riisager, S. Hall, M. Antretter and X. Zhao, Early Cretaceous Paciﬁc palaeomagnetic pole from Ontong Java Plateau basement rocks. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 31–44, Geological Society, London (2004).

213. J. Roberge, R. White and P. Wallace, Volatiles in submarine basaltic glasses from the Ontong Java Plateau (ODP Leg 192): implications for magmatic processes and source region compositions. In J. Fitton, J. Mahoney, P. Wal- lace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 329–257, Geological So- ciety (2004).

214. R. Robert, E. V. Wyk and T. Steele, Concentration of the noble metals by a ﬁre assay technique using nickel sulphide as the collector. Report 1371, National Institute of Metallurgy, Johannesburg, RSA (1971).

215. M. Roden, T. Trull, S. Hart and F. Frey, New He, Nd, Pb, and Sr isotopic constraints on the constitution of the Hawaiian plume: results from Koolau Volcano, Oahu, Hawaii, USA. Geochimica et Cosmochimica Acta, 58(5): 1431–1440 (1994).

216. M. Roy-Barman and C. Allegre, 187Os/186Os in oceanic island basalts: trac- ing oceanic crust recycling in the mantle. Earth and Planetary Science Let- ters, 129: 145–161 (1995).

272 217. S. Russel, T. Lay and E. Garnero, Seismic evidence for small-scale dynamics in the lowermost mantle at the root of the Hawaiian hotspot. Nature, 396: 255–258 (1998).

218. T. Sano and S. Yamashita, Experimental petrology of basement lavas from Ocean Drilling Program Leg 192: implications for diﬀerentiation processes in Ontong Java Plateau magmas. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 185–218, Geological Society, London (2004).

219. A. Schersten, T. Elliot, C. Hawkesworth and M. Norman, Tungsten isotope evidence that mantle plumes contain no contribution from the Earth’s core. Nature, 427: 234–237 (2004).

220. H. Schreiber, H. Lauer and T. Thitinant, The redox state of cerium in basaltic magmas: An experimental study of iron-cerium interactions in silicate melts. Geochimica et Cosmochimica Acta, 44: 1599–1612 (1980).

221. C. Schwandt and G. McKay, Rare earth element partition coeﬃcients from enstatite/melt synthesis experiments. Geochimica et Cosmochimica Acta, 62(16): 2845–2848 (1998).

222. W. Seyfried, X. Chen and L. Chan, Trace element mobility and lithium isotope exahcnge during hydrothermal alteration of seaﬂoor weathered basalt: an experimental study at 350 ◦C, 500 bars. Geochimica et Cosmochimica Acta, 62(6): 949–960 (1998).

223. J. Shafer, C. Neal and P. Castillo, Compositional variability in lavas from the Ontong Java Plateau: results from basalt clasts within the volcaniclastic succession at Ocean Drilling Program Site 1184. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 333–351, Geological Society, London (2004).

224. J. Shafer, C. Neal and M. Regelous, Petrogenesis of Hawaiian postshield lavas: Evidence from Nintoku Seamount, Emperor Seamount Chain. Geo- chemistry, Geophysics, Geosystems, 6(5) (2005).

225. P. Sikora and J. Bergen, Lower Cretaceous biostratigraphy of Ontong Java Sites from DSDP Leg 30 and ODP Leg 192. In J. Fitton, J. Mahoney, P. Wal- lace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 83–112, Geological Soci- ety, London (2004).

273 226. A. Skovgaard, M. Storey, J. Baker, J. Blusztajn and S. Hart, Osmium-oxygen isotopic evidence for a recycled and strongly depleted component in the Iceland mantle plume. Earth and Planetary Science Letters, 194: 259–275 (2001). 227. T. Skulski, W. Minarik and E. Watson, High-pressure experimental trace- element partitioning between clinopryoxene and basaltic melts. Chemical Ge- ology, 117(1-4): 127–147 (1994). 228. N. Sleep, Tapping of magmas from ubiquitous mantle heterogeneities: an alternative to mantle plumes? Journal of Geophysical Research, 89: 10029– 10041 (1984). 229. A. Sobolev, A. Migdisov and M. Portnyagin, Incompatible element partitioning between clinopyroxene and basalt liquid revealed by the study of melt inclusions in minerals from Troodos lavas, Cryprus. Petrology, 4(3): 307–317 (1996). 230. S. Solomon, N. Sleep and D. Jurdy, Mechanical models for absolute plate motions in the Early Tertiary. Journal of Geophysical Research, 82: 203–213 (1977). 231. F. Strelow, An ion exchange selectivity scale of cations based on equilibrium distribution coeﬃcients. Analytical Chemistry, 32: 1185–1188 (1960). 232. C.-O. Sun, R. Williams and S. Sun, Distribution coeﬃcients of Eu and Sr for plagioclase-liquid and clinopryoxene-liquid equilibria in oceanic ridge basalt; an experimental study. Geochimica et Cosmochimica Acta, 38(9): 1415–1433 (1974). 233. S. Sun and G. Hanson, Origin of Ross Island basanitoids and limitations upon the heterogeneity of mantle sources or alkali basalts and nephelinites. Contributions to Mineralogy and Petrology, 52: 77–106 (1975). 234. S. Sun and W. McDonough, Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In A. Saunders and M. Norry, editors, Magmatism in the Ocean Basins, volume 42 of Special Publications, pages 313–345, Geological Society, London (1989). 235. J. Tarduno and R. Cottrell, Paleomagnetic evidence for motion of the Hawai- ian hotspot during formation of the Emperor Seamounts. Earth and Plane- tary Science Letters, 153: 171–180 (1997). 236. J. Tarduno, D. Duncan and R. Scholl, Proceedings of the Ocean Drilling Pro- gram, Initial Reports, 197. [CD-ROM], Available from: Ocean Drilling Pro- gram, Texas A&M University, College Station TX 77845-9547, USA (2002).

274 237. J. Tarduno, R. Duncan, D. Scholl, R. Cottrell, B. Steinberger, T. Thordarson, B. Kerr, C. Neal, F. Frey, M. Torr and C. Carvallo, The Emperor Seamounts: southward motion of the Hawaiian Hotspot plume in Earth’s mantle. Science, 301: 1064–1069 (2003).

238. J. Tarduno, W. Sliter, L. Kroenke, M. Leckie, J. Mahoney, R. Musgrave, M. Storey and E. Winterer, Rapid formation of the Ontong Java Plateau by Aptian mantle plume volcanism. Science, 254: 399–403 (1991).

239. Y. Tatsumi, K. Oguri and G. Shimoda, The behavior of platinum-group elements during magmatic diﬀerentiation in Hawaiian tholeiites. Geochemical Journal, 33: 237–247 (1999).

240. M. Tejada, J. Mahoney, P. Castillo, S. Ingle, H. Sheth and D. Weis, Pin- pricking the elephant: evidence on the origin of the Ontong Java Plateau from Pb-Sr-Hf-Nd isotopic characteristics of ODP Leg 192 basalts. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 133–150, Geological Society, London (2004).

241. M. Tejada, J. Mahoney, R. Duncan and M. Hawkins, Age and geochemistry of basement and alkalic rocks of Malaita and Santa Isabel. Journal of Petrology, 177: 361–393 (1996).

242. M. Tejada, J. Mahoney, C. Neal, R. Duncan and M. Petterson, Basement Geochemistry and Geochronology of Central Malaita, Solomon Islands, with Implications for the Origin and Evolution of the Ontong Java Plateau. Jour- nal of Petrology, 43(3): 449–484 (2002).

243. T. Thordarson, Accretionary-lapilli bearing volcaniclastic rocks at ODP Site 1184A: A record of large Surtseyan-like prhreatomagmatic eruptions on the Ontong Java Plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 275–306, Geological Society, London (2004).

244. W. Todt, R. Cliﬀ, A. Hanser and A. Hofmann, Evaluation of a 202Pb + 205Pb double spike for high-precision lead isotopic analysis. In A. Basu and S. Hart, editors, Earth Processes: Reading the Isotopic Code, volume 95 of Geophys- ical Monograph, pages 429–437, American Geophysical Union (1996).

245. B. Villemant, H. Jaﬀrezic, J. Joron and M. Treuil, Distribution coeﬃcients of major and trace-elements—fractional crystallization in the alkali basalt series of Chaine-Des-Puys (Massif Central, France). Geochimica et Cosmochimica Acta, 45(11): 1997–2016 (1981).

275 246. F. Vine and D. Matthews, Magnetic anomalies over oceanic ridges. Nature, 199: 947–949 (1963).

247. T. Waight, J. Baker and B. Willigers, Rb isotope dilution analyses by MC- ICPMS using Zr to correct for mass fractionation; towards improved Rb-Sr geochronology. Chemical Geology, 186(1-2): 99–116 (2002).

248. R. Walker, J. Morgan and M. Horan, 187Os enrichment in some plumes: Evidence for core-mantle interaction? Science, 269: 819–822 (1995).

249. H. West, M. Garcia, F. Frey and A. Kennedy, Nature and cause of compositional variation among the alkalic cap lavas of Mauna Kea Volcano, Hawaii. Contributions to Mineralogy and Petrology, 100(3): 383–397 (1988).

250. C. Wheat, M. Mottl and M. Rudnicki, Trace elmement and REE composition of a low-temperature ridge-ﬂank hydrothermal spring. Geochimica et Cosmochimica Acta, 66(21): 3693–3705 (2002).

251. R. White, P. Castillo, C. Neal, J. Fitton and M. Godard, Phreatomagmatic eruptions on the Ontong Java Plateau: chemical and isotopic relationship to Ontong Java Plateau basalts. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume 229 of Special Publications, pages 307–323, Geological Society, London (2004).

252. J. Wilson, A possible origin of the Hawaiian Islands. Canadian Journal of Physics, 41: 863–870 (1963).

253. M. Wilson, Igneous Petrogenesis. Unwin Hyman, London (1989).

254. S. Wilson, Data Compilation for USGS Reference Material BHVO-2. Open- File Report in prep, US Geological Survey (1997).

255. J. Winchester and P. Floyd, Geochemical magma type discrimination: Ap- plication to altered and metamorphosed basic igneous rocks. Earth and Plan- etary Science Letters, 285: 459–469 (1976).

256. S. Wood, B. Mountain and P. Pan, The aqueous geochemistry of Platinum, Palladium, and Gold: Recent experimental constraints and a re-evaluation of theoretical predictions. Canadian Mineralogist, 30: 955–982 (1992).

257. T. Wright, Origin of Hawaiian tholeiite; a metasomatic model. Journal of Geophysical Research, 89(B5): 3233–3252 (1984).

258. T. Wright and R. Fiske, Origin of the diﬀerentiated and hybrid lavas of Kilauea volcano. Journal of Petrology, 12: 1–65 (1971).

276 259. H. Yang, F. Frey and D. Clague, Constraints on the source components of lavas forming the Hawaiian North Arch and Honolulu Volcanics. Journal of Petrology, 44(4): 603–627 (2003).

260. A. Zindler and E. Jagoutz, Mantle cryptology. Geochimica et Cosmochimica Acta, 52: 319–333 (1988).

This document was prepared & typeset with LATEX2ε, and formatted with nddiss2ε classﬁle (v1.0[2004/06/15]) provided by Sameer Vijay.

277