GEOCHEMISTRY AND GEOLOGY MINERALIZED AND BARREN KOMATI ITES - WESTERN AUSTRAL I A BY ROBERT D. McNEIL A the9is submitted to the Department of Geol~gyin fulfilment of the requirements for the degree of Master of Science. University of Tasmania .Hobart, Tasmani a 4u.l~~.I980 This thesis contains no material which has been accepted for the award of any other degree or diploma at any university, and that to the best of my knowledge the thesis contains no copy or paraphrase of material previously published or written by another person except when due reference is made in the text of the thesis or as acknowledged in the Acknowledgements. Robert D. McNeil July 1, 1980 7 ABSTRACT d: SUM^.^^ 7 Western Austral ian Archaean komatii tes which are associated with nickel sulphide mineralization can be separated into two groups - Mineralized or Barren, based on komatiite lithogeochemistry. Mineralized komatiites may host nickel sul phi de deposi ts whereas Barren komati i tes do not. Chemical relationships were determined from a data base of approximately 3300 samples of fresh komatiite ul~arnaficfrom four nickel provinces and other greenstone be1 ts not known to contain nickel sulphides. Mean chemical values for each group of komati i tes were : Category Ni P -N i CUP--- Cu A1 -Ca EL --Zn Cr -Mn -Fe COP- -Co - (21 Mineral ized 1027 2220 36 42 1.6 2.2 19.2 69 1617 1057 6.1 49 119 Barren 429 1530 29 39 2.2 2.8 16.4 76 2260 1128 7.0 32 119 Discriminant analysis , using the above thirteen chemical determinations as variables, for each of 2775 samples from forty localities, indicated that samples could be classified as either Mineralized or Barren with an expected accuracy of greater than 80 percent. No single element or chemical deter-min- ation is definftive, but collectively, Cr, Ni, Zn, Cu, Nip,- Mg, Fe and Co can distinguish between the two groups of ul tramafics. Critical elements are Cr, Ni and Nip,- assuming that values of Zn, Cu, Mg and Fe approximate the mean value for all West Austral ian komatiites. The Ni to Cr ratio is always greater than unity (1) in Mineralized komatiites and the Ni to Nip- ratio is always less than 3.5. Sulphur is not a diagnostic element. NOTES : 1. -P indicates a partial or sulphide analysis. 2. Al, Ca, Mg and Fe results are expressed in percentages; all others in parts per mill ion. (ii) Increasing Ni/Cr ratios and decreasing Ni/NiP- ratios within a komatiite can be regarded as indicative of increasing nickel sulphide potential. Mineralized komatiites contain less Cr within the silicate lattice structure and less chromite than Barren komatiites. However, the more important relationship appears to be the lesser amount of Cr attached to the silicate mineral lattice. Correlation analysis showed that: 1. most correlations are much stronger in Barren than in Mineralized ultramaf ics; 2. the chalcophile elements, Cu, Ni, Co and Fe (constituents of nickel sulphide deposits), show moderate to strong correlations with the rock forming elements, Mg, Mn, Ca, A1 in Barren ultramaf ics, but only weak or no correlation in the Mineralized ultramafics; 3. copper has moderate positive correlation with Fe, Mn, Ca, A1 and negative correlation with Mg in Barren ultramafics but shows no correlation with these same elements in Mineralized ultramaf ics. These correlation differences suggest that in Barren komatiites Ni, Cu, Co and Fe are contained in the silicate mineral lattice whereas in Mineralized komatiites they are presently partly as a separate sulphide fraction. In addition they may also suggest that these sulphides were added or removed from Mineralized komatiites after the formation of the komatiite magma, probably by concentration and removal in an' immissible sulphide-oxide melt. -. '. Komatiites can be divided into two separate suites called volcanic and intrusive. Volcanic suites such as those at Kambalda and Windarra South may contain many individual komatiite flows. The basal section of a komatiite volcanic pile consists of a small number of thick units which may contain sulphide mineralization whereas the central and upper parts of the pile consists of multiple -thin units. Both thick and -thin units consist of an olivine cumulate derived lower part overlain by a silicate liquid derived upper part. In thick units the olivine cumulate section is dominant whereas in -thin units the (iii) silicate liquid section is dominant. Spinifex texture is characteristic of 4' ., unmetamorphosed sequences. ufl &)++ CI 0 In metamorphosed sequences such as Windarra South it is not possible to identify individual komati i tes using mineralogical or textural criteria but it can be accomplished using chemical data. Intrusive suite komati i te sequences such as Forrestania or Perseverance usually consist of a small number of high Mg, homogeneous peridoti tes and/or dunites. Equigranular, equant olivine textures are characteristic. These komatiites are often continuous over strike lengths of the order of tens of kilometers and contain relatively 1i ttle internal chemical variation. Volcanic komatiites such as those at Windarra South and Kambalda are considered to be ul tramafic lavas. Chemical differences between volcanic and intrusive sequences have been defined. Typical chemical values for the cumulate section of a volcanic komatiite and for intrusive komatiites, both with moderate to high mineral izati on coefficients are: - Classification Nip- Ni & - - @ CUP--- Cu A1 Zn Cr Mn COP- Vol can i c Komati i tes 1000 2100 30 - 90 1-2 17-24 60 1300 1000 5.5 55 120 Intrusive Komati i tes 1200 2500 5 - 60 0.5 20-26 60 1000 900 6 60 125 In general, if NiP or Ni are less than 500 and 1800 ppm respectively, or Cr greater than 2100 ppm, a komatiite can be regarded as Barren. It has been possible to define sections of greenstone belts as prospective for nickel sulfides and other parts as unprospective. For example, the Forrestania section of the Forrestania-Southern Cross greenstone belt has a different chemical signature to the Southern Cross section. The latter section is unlikely to contain economic nickel sulphide accumulations. (P v) ACKNOWLEDGEMENTS The writer acknowledges the support of the Tenneco Australia Inc. - Minops Pty. Ltd. Joint Venture and Union Oi 1 Development Corporation. Numerous companies and i ndi vi dual s a1 1 owed the wri ter to coll ect sampl es , most notably the Tenneco-Minops Joint Venture, Poseiden Ltd. and Amax Exploration (Australia) Pty. Ltd. Most of the geoloqic maps used for location purposes or for geological descriptions of specific areas are based on fie1d mapping by other geologists. The names of many of these geologists are unknown but M. Lennox, J. Noakes and M. Woodhouse, all formerly of Tenneco Australia Inc. deserve specific acknowledgement. However, a1 1 solid geology interpretive maps and sections presented here are the responsibility of and have been compiled by the writer. The writer especially wishes to thank the above three geologists for their assistance and for many stimulating discussions. Thanks are also due to E.A. Rugg formerly of Tenneco Australia Inc. and E.H. Lindsey of Union Oil Co. of Cal ifornTa for their support throughout much of the study. W. R. Guthrie, also formerly of Tenneco Australia Inc. , was responsible for most of the analytical procedures and results. The description of analytical procedures are based on personal communications and notes from Guthrie. P. Walker and N. Campbell deserve thanks for explaining the statistical procedures to the writer. The description of Mulvar in Appendix E is based on unpublished notes by J. Keays. Finally I express my sincere thanks to Pamela Strauss for the final typing of this manuscri'pt. (4 CONTENTS -PAGE ABSTRACT ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES (xiii) CHAPTER I - INTRODUCTION 1.1 Format of Thesis 1.2 Data Base 1.3 Analytical Procedures CHAPTER 2 - GEOLOGICAL ENVIRONMENT 2.1 Topography and Weathering 2.2 Greenstone Belts 2.3 Komat ii tes 2.4 Metamorphism and Alteration 2.5 Geological Subdivision of the Yilgarn Block 2.6 Nickel Provinces CHAPTER 3 - MINERALIZED AND BARREN KOMATIITES 3.1 Mean Geochemical Results for each Locality 3.2 Discriminant Analysis 3.2.1 First and second stage analyses 3.2.2 Third stage analyses 3.3 Principal Component Analysis 3.4 Correlation Analysis 3.5 Relative Importance of Each Variable CHAPTER 4 - CHEMICAL CHARACTERISTICS OF VOLCANIC AND INTRUSIVE KOMATI ITES 4.1 Volcanic Komatiite Suite - 4.1.1 Kambalda 4.1.2 Windarra South 4.1.3 Trough Wells 4.1.4 Eureka Greenstone Belt 4.1.5 Red Well - 4.1.6 Airport - Yilmia 4.2 Intrusive Komatiite Suite 4.2.1 Queen Victori a Rocks 4.2.2 Forrestania 4.2.3 Bu 1 lf inch 4.2.4 Mistake Creek CONTENTS -PAGE CHAPTER 5 - EVALUATION OF WONGANOO-BANDJAWARN GREENSTONE BELT 5.1 Geological Setting 5.2 Evaluation of individual areas within the greenstone belt 5.2.1 Dingo Range West 5.2.1.1 Geology 5.2.1.2 Komatiite Sequence 5.2.1.3 Geochemistry 5.2.2 Devines 5.2.3 Dingo Range East and Lalor North 5.2.4 Mt. Step and Collin Well 5.3 Discussion CHAPTER 6 - COMPARISON OF FORRESTANIA NICKEL PROVINCE WITH THE CONTIGUOUS SOUTHERN CROSS GREENSTONE BELT 6.1 Forrestania Nickel Province 6.2 Southern Cross Greenstone Belt 6.2.1 Marvel Lock A 6.2.2 Trough Wells 6.2.3 Southern Cross Drill Holes 6.2.4 Marvel Lock B 6.2.5 Marvel Lock C 6.2.6 Ennuin 6.2.7 Bullfinch 6.3 Discussion CHAPTER 7 - APPLICATION OF GEOCHEMICAL CRITERIA TO NEW AREAS AND SAMPLE GROUPS WITH HIGH DEGREE OF MISCLASSIFICATION 7.1 Application of geochemical criteria to new areas 7.1.1 Area B 7.1.2 Area C 7.2 Barren groups with high percentage of samples misclassified 7.2.1 AreaA 7.2.1 Yerilla 7.2.3 Heather Hill CHAPTER 8 - DISCUSSION AND CONCLUSIONS 8.1 Mean Geochemical Values 8.2 Mineralized and Barren Komatiite Characteristics 8.2.1 Volcanic Suite 8.2.2 Intrusive Suite 8.2.3 Application of Mineralized - Barren Criteria 8.3 Chemical Gradients 8.4 Regional and Stratigraphic Chemical Differences 8.5 Comments cn Individual Areas 8.6 Significance of Sulphur 8.7 Genetic Aspects (vii ) CONTENTS -PAGE REFERENCES APPENDIX A - GEOLOGY AND GEOCHEMISTRY OF KALGOORLIE-NORSEMAN NICKEL PROVINCE A.l Geological Setting A.l.l Stratigraphy A.