4 Lithology, structure and geochemistry of the ca 3.5 Ga North Star Basalt in the Marble Bar Greenstone Belt, Archaean Pilbara Craton, Western Australia.
This study was undertaken to investigate a possible oceanic origin for the North Star Basalt in the Marble Bar Belt. This is the lowermost formation of the Warrawoona Group and one of the oldest greenstones sequences in the Archaean Pilbara Craton in Western Australia. It consists mainly of pillowed and massive basalts, minor gabbro, and comprises a large number of mafic and ultramafic dykes. Geochemical studies have shown that the upper part of the North Star Basalt comprises enriched tholeiitic basalts, probably due to contamination of the magmas by assimilation of crustal material. They do not resemble modern MORB. The lower, ultramafic part of the stratigraphy may not be part of the North Star Basalt, as indicated by its different trace element geochemistry. A 40Ar/39Ar cooling age of about 3.47 Ga indicates that these rocks may be the same age as the Talga Talga Subgroup of the Warrawoona Group, to which the North Star Basalt belongs. Only a small fraction of the dykes that occur in the area, is genetically related to the extrusive pile; the majority has been emplaced later, probably during regional extension at ca 3.3 Ga. Granite intrusions at ca 3.3 Ga post-date emplacement of all of the dyke suites, and have destroyed the lower section of the greenstone sequence. There is no firm evidence for large displacements on any of the structures within the unit. Therefore the Talga Talga Anticline may still be a suitable type area for the North Star Basalt, but the presence of low angle unconformities should not be disregarded.
The Marble Bar Greenstone Belt is located in the Archaean Pilbara Craton in Northwestern Australia (Figure 4.1). The Pilbara Craton is an early to mid-Archaean terrane comprising granitoids and greenstone belts (e.g. Hickman, 1983), overlain by the late Archaean volcano-sedimentary sequence of the Hamersley Basin. The studied area is located about 25 km north of Marble Bar and 150 km to the southeast of Port Hedland (Figure 4.1). It is well accessible, and the outcrop is generally good. The basal Warrawoona Group forms an anticlinal structure in the Marble Bar Belt, west of the Mount Edgar Granitoid Complex. This structure is referred to as the Talga Talga Anticline (Figure 4.2) and hosts the type area for the Talga Talga Subgroup (Hickman, 1977; 1983). The local stratigraphy is summarized in Table 4.1. The basal formation of the Talga Talga Subgroup is the North Star Basalt which comprises a ca 1 km thick sequence of dominantly basalt.
The North Star Basalt is thought to be a potential ophiolitic sequence (see comment in De Wit (1998)). Zegers (1996) suggested it was part of the Coonterunah Group. If it does belong to either the Coonterunah or Warrawoona Groups, correlation with the work of Green et al. (2000) suggests that it cannot be oceanic. Green et al. (2000) indicated that the basalts of the Coonterunah and upper Warrawoona groups were deposited in a continental environment. This would indicate that the East Pilbara has been a stable continental block since at least the deposition of the ca 3515 Ma Coonterunah Group. The long term stability of the East Pilbara has implications for the evolution of the West and Central Pilbara and is important to the central theme of this thesis.
In this paper the results of a detailed lithological, structural and geochemical study of the North Star Basalt are presented. The study is based on mapping (Figure 4.3) and was undertaken to investigate the tectonic setting in which the extrusive pile was formed. The possibility of an oceanic origin for the North Star Basalt will be discussed after an investigation into the extent to which it has been structurally disturbed. In fact, Trendall (1995) suggested that the North Star Basalt presents an area that potentially contains some answers to important questions about the geological history of the Pilbara Craton.
4.2 Regional Geology
The Warrawoona Group in the East Pilbara is the most wide spread greenstone sequence in the East Pilbara Craton; it unconformably overlies the Coonterunah succession (see Table 4.1). U-Pb zircon ages indicate the Warrawoona Group was deposited between ca 3475 and 3435 Ma (e.g. Thorpe et al., 1992, Nelson 1996-2002). It comprises mainly volcanic assemblages that were possibly deposited in volcanic-arc and near-arc environments (Barley, 1997). In most of the East Pilbara the Gorge Creek Group unconformably overlies the Warrawoona Group. It consists mainly of clastic sediments, pillow basalts and minor komatiites.
The North Star Basalt is the basal formation of the Talga Talga Subgroup and also of the Warrawoona Group. The lower boundary of the North Star Basalt in the Marble Bar Belt (Figure 4.1) is the intrusive contact with the Mount Edgar Granitoid Complex. Its stratigraphic position and regional correlations have been the subject of debate, because its age is not well constrained. The mafic rocks of the Talga Talga Subgroup have a Sm-Nd model age of approximately 3560 ± 32 Ma (Hamilton et al., 1981) or 3712 ± 98 Ma (Gruau et al., 1987), however, it is not clear exactly which rocks suites have been included in the calculations for these model ages. In the most recently published stratigraphy of the east Pilbara Craton (Van Kranendonk et al., 2002) the North Star Basalt is still regarded as the oldest formation of the Talga Talga Subgroup (see Table 4.1).
Nowhere a conformable lower contact has been observed for the North Star Basalt (or the base of the Warrawoona or Coonterunah Group for that matter), so the nature of the basement onto which the sequence was extruded, remains enigmatic. However, Barley (1986) reported crustal contamination of the Warrawoona basalts, and Green et al. (2000) reported granitic xenoliths in basalts of the upper Warrawoona Group, suggesting those basalts were extruded onto a continental basement.
A pillow basalt near the top of the North Star Basalt was reported to contain zircons with ages of 3326 ± 10 Ma, and a syn-volcanic dolerite sill has been dated at 3238±6 Ma (McNaughton et al., 1993), but these zircons most likely have a hydrothermal origin. The stratigraphic locations of these published data are indicated in Figure 4.4. Nelson (2000) has dated a rock from the overlying McPhee Formation (Table 4.1). A rock described as a vitric tuff, but with a petrographic description of a cherty mylonitic schist, gave an age of 3477 ± 2 Ma. A felsic lens in the overlying Mount Ada Basalt (Table 4.1) was dated at 3469 ± 3 Ma (Nelson, 2000). In this paper the results of 40Ar/39Ar dating in the North Star Basalt will be presented.
Zegers et al. (1996) suggested that the Warrawoona Group was deposited during a phase of continental extension based on interpretation of syn-depositional extensional brittle faults in the Duffer Formation in the Coongan Greenstone Belt (Figure 4.1). They related these structures to ductile detachment shears observed at deeper levels in the stratigraphy and within batholiths, in an architecture similar to modern extensional core complexes. Although similar observations have been made in the North Pole Dome and Marble Bar Belt by Nijman et al. (1998), the role of extension has been questioned by White and Hickman (public discussion at 4IAS, Perth, 2001).
Van Haaften and White (1998) reported a major east-directed thrusting event during which they suggest Talga Talga Anticline was formed by thrust stacking, and therefore suggested that the area is not suitable as a type section for the Talga Talga subgroup. However, van Kranendonk et al (2001) commented that there is no evidence for large displacement on the structures found by Van Haaften and White (1998), and they suggest that the Talga Talga Anticline was formed by diapiric activity of the Mount Edgar Granitoid Complex (as did Hickman, 1984; Collins, 1989).
4.3.1 The extrusive stratigraphy
The Talga Talga area was mapped at a scale of 1:5000 with the aid of aerial photographs and Landsat imagery. A detailed map is shown in Figure 4.3 and the stratigraphic column is given in Figure 4.4. The basal unit of the North Star Basalt is the Lower Ultramafic Unit. The lower part of the unit consists of fine-grained massive mafic rocks, doleritic and gabbroic patches. Several lenses of up to 100 m long and 10 m thick occur on small hills: these lenses have a pyroxenite texture but their mineralogy consists mainly of actinolite, chlorite, epidote, serpentine and carbonates. The main lithologies of this unit are fine grained massive mafic rocks, composed of 30% chlorite, 20% magnesio-carbonates, 10% plagioclase and accessory titanite and opaques. The top of the unit is formed by a ca 30-50 m thick talc-carbonate-serpentine schist with a greenish-white appearance, which dips ca 20 degrees to the northwest, as shown in the map and cross section in Figure 4.3. Locally altered olivine spinifex textures up to 30 cm long are preserved (Figure 4.5), indicating that those rocks are not strongly deformed. These rocks are composed of chlorite, dolomite, magnesite, serpentine and opaque. On the basis of geochemistry (section 4.4) and texture, this unit is interpreted to be a komatiite.
The Main Basalt (Figure 4.4) forms the dominant lithology of the North Star Basalt. The sequence contains massive flows and pillows. Individual flows are up to several meters thick. The pillows and flows contain amygdales that consist of a quartz rim and chlorite center. The rocks contain mainly actinolite, albite, chlorite and minor quartz. The basaltic lavas have low vesicle contents, however, primary igneous textures are still preserved. In the pillows the vesicles form a radiating pattern, in flows they are locally elongated. The concentration of vesicles is greater in the top of flow and pillows, than in the bottom. This was used to establish that the Main Basalt has a consistent younging direction to the northwest. The unit dips 15-25 degrees to the northwest, as shown in the map and cross section in Figure 4.3.
The Thin Chert overlies the Main Basalt. It is a 5-10 m thick banded chert that dips 25-35 degrees to the northwest. This slightly steeper than the bedding below, but parallel to the bedding above. The lower contact is obscured by schistose and strongly weathered rocks. The chert consists of very finely laminates quartz and opaques. It could possibly be a silicified slate horizon. This chert is overlain by the Middle Ultramafic Unit. The lower part of this unit consists of massive ultramafic rocks. The upper part of this unit is a carbonate-serpentine schist.
The Middle Ultramafic Unit (Figure 4.4) is overlain by a Clastic Sedimentary Unit with variable thickness and composition. The Clastic Sediments display two distinct sub-units. A fine grained lower sub-unit consists of dark colored, very finely laminated slatey material. This sub-unit forms a marker bed throughout the area, and is about 10 m thick. The upper sub-unit consists of a very coarse matrix-supported breccia. Locally the breccia coarsens upward, but in most outcrops the size of the clasts varies randomly. The breccia contains a wide variety of clasts. Chert clasts are angular to subrounded, mostly elongated, and 1-30 cm in length. Basaltic clasts are typically rounded and 1 to 15 cm in size. The matrix of the breccia is silicified. A weak foliation sub-parallel to the bedding is defined by a preferred orientation of the clasts. Locally the upper sub-unit reaches a maximum thickness of 30 meters, while in other localities it is absent.
The top of the North Star Basalt is formed by the Upper Extrusive Unit (Figure 4.4), which contains alternating massive flows and vesicular pillow basalt. The rocks are locally altered to serpentine-carbonate schist, while in other places they are relatively unaltered basalt. Altered and unaltered packages alternate in this unit. A 20 m thick banded chert, the Big Chert, unit caps the formation. Like the Thin Chert halfway in the stratigraphy, the Big Chert overlies foliated and weathered rocks and dips at a slightly steeper angle than the bedding below. It forms the base of the overlying McPhee Formation (Figure 4.4).
Low in the stratigraphy, between the Lower Ultramafics and the Main Basalt, a gabbroic sill has been emplaced (Figure 4.4). This sill is internally differentiated: it has a melanogabbroic main body now consisting of mainly actinolite, plagioclase, epidote and relics of pyroxene, and a leucogabbroic top which is richer in plagioclase and contains magmatic hornblende which ahs been dated (section 4.6). Green Diorite Sills have intruded the contact of the Middle Ultramafic Unit and the Clastic Sedimentary Unit. The diorite locally crosscuts the clastic sediments, and diorite sills up to 20 m thick occur above the Clastic Sedimentary Unit (Figure 4.4). A felsic unit was emplaced at a small angle in the Main Basalt, and its system of feeder dykes have intruded the Lower Ultramafic Unit and the Gabbro (see Figure 4.3). The felsic unit is the youngest intrusion in this area. It is related to the ca 3300-3320 Ma granite suite of the Mount Edgar Granitoid Complex which forms the lower contact of the North Star Basalt (McNaughton et al., 1993).
The North Star Basalt is crosscut by several suites of mafic dykes. Some of the dykes form composite systems of mafic and ultramafic dykes, whereas others form separate sets of only dolerite dykes. All dykes are sub-vertical; their orientations are shown in Figure 4.6a and b. It is shown that the separate dolerite dykes form two sets, with a main trend of 340 and a minor trend at 320. The minor set are small and thin (up to 50 cm) and occur at a right angle to the basaltic bedding. They are interpreted to be the local feeders of the basalt, because some of the dykes have been observed to be connected to flows. The other set of dolerite dykes is younger. The composite dyke suite also forms separate sets. The trend of the main set is 300. In Figure 4.3 it can be observed that the composite dykes cut not only the volcanic stratigraphy but also the later diorite sills. None of the dyke suites cut the granite or the felsic sill, so they all pre-date ca 3320 Ma (McNaughton et al., 1993). Importantly the dykes in the North Star basalt form several distinct suites that formed at different times in the geological history.
Some of the alteration of the rocks may have occurred at an early stage in the history of the North Star Basalt. Yardley (1989) suggested that hydration and alteration are characteristic of basalts that have experienced sea floor alteration at low greenschist facies conditions. However, the greenschist facies metamorphism has been recorded throughout the area (Hickman, 1983). This thermal event may also be related to the regional emplacement of large granitic bodies. Zegers et al. (1996) suggested that this kind of metamorphism in greenstones could be the effect of the emplacement of mid-crustal rocks against the greenstones along a detachment fault in a setting similar to modern metamorphic core complexes. They associated the observed metamorphic event with a phase of continental extension at ca. 3.46 Ga.
The mineral assemblages of the metamorphic, formerly mafic and ultramafic rocks (epidote, chlorite, carbonate, serpentine) indicate greenschist grade metamorphism (Miyashiro, 1994). There is no variation in the metamorphic grade throughout the study area, except towards the contact with the granite. This indicates no significant gaps or duplications of the metamorphic field gradient and consequently no significant tectonic disturbance of the section post-dating the metamorphic event.
The contact of the North Star Basalt with the Mount Edgar Granitoid Complex (see map in Figure 4.3) is mostly exposed, however, the general dip of the contact could not be determined, due to the irregular intrusive nature of the contact. It is interpreted to be quite steep, based on the limited extent of the contact aureole. The structural effect of the intrusion is intense fracturing of the lower 10 meters of the mafics, allowing granitic melt and fluids to penetrate and silicify this bottom part of the section intensely. This effect drops off further away from the contact. The contact aureole shows a steep gradient from hornfelsing very close to the contact, to very limited overprinting about 500 meters away from the contact.
Contact metamorphism also occurred adjacent to the Green Diorite sills in the top of the stratigraphy. Thin section studies have shown that a greenschist metamorphic assemblage of albite, actinolite and epidote is overprinted by chlorite and calcite due to the heat and fluid flux associated with the intrusion. This contact metamorphic aureole reaches up to 30 meters into the host rock.
4.3.5 Lithology - discussion
A summary stratigraphic column is shown in Figure 4.4. The Lower Ultramafics were formed by extrusion in water, as indicated by the presence of pillows. Komatiites (now carbonate schists) were extruded, or emplaced as a shallow sill. This was followed by the deposition of the Main Basalt, accompanied by the formation of its own feeder dykes in the lower and central part of the exposed basalt sequence. During a period of quiescence, the Thin Chert was deposited on a low angle unconformity. This chert is now deformed and foliated and is interpreted to represent silicified and ferrugenized slate. The quiet period is followed by the extrusion of an alternation of ultramafic and mafic lavas forming the Upper Extrusives. The North Star Basalt is capped by a chert unit of the overlying McPhee Formation, following a second phase of quiescence and subsidence accompanied by syn-depositional faulting (at a high angle to the bedding) resulting in a low-angle unconformity at the top of the formation (see map in Figure 4.3.). This chert is now deformed and foliated and is interpreted to represent silicified and ferrugenized slate. The gabbro sill low in the stratigraphy was emplaced early in the history, the diorite sills and the felsics were emplaced later as indicated by crosscutting relationships with the dykes (see map in Figure 4.3.)
Three dyke suites were recognized. The first are a set of minor and thin dolerite dykes in the lower and central part of the stratigraphy, that appear to be associated with the Main Basalt as local feeders. This is confirmed by their geochemistry (section 4.4). The second are a suite of composite (mafic-ultramafic) dykes which have a northwest trend. They are up to 10 meters wide and can be traced through all structures in the North Star Basalt and into overlying formations, but not into the Mt Edgar Granitoid Complex. A clearly later and separate set of dolerite dykes has a northwest trend. They are up to 1 meter wide and have been observed to cut the older two sets and all older structures, but cannot be traced into the Mount Edgar Granitoid Complex. They are approximately orthogonal to the extensional direction in the upper plate at ca 3300 Ma, when the Mount Edgar Marginal Shear was under extension (Kloppenburg et al., 2001). That is, they are the result of a phase of continental extension rather than extension associated with a mid-ocean ridge. They can therefore not all be part of an ophiolite complex.
4.4.1 Analytical methods
Fifty-four representative samples were selected for geochemical analysis. All mafic lithologies in the North Star Basalt are represented and the least weathered samples were selected. The samples were located away from faults and fractures. Any secondary quartz and carbonate veins were removed before grinding. XRF analyses were performed on pressed powder pellets for the VK-samples samples, on fused glass discs for the JW-numbered samples. The XRF analyses were carried out at the Free University of Amsterdam, on a Philips PW1404/10 spectrometer. ICP-MS trace element analyses were performed at ACTLAB in Ontario, Canada.
Sample locations and petrographic descriptions and can be found in Appendix 4.A.1. All samples show evidence for low greenschist grade metamorphism, however, in most rocks the original igneous textures can still be observed. Rocks containing amphiboles, albite, epidote, chlorite and quartz are interpreted to be metamorphosed mafic rocks.
The major element data (Appendix 4.A.2) are given in weight %, recalculated to a total dry sum of 100% and with total Fe represented as Fe2O3. The uncorrected total sum for all major element analyses deviated no more than 2% from 100%. The trace element data can be found in Appendix 4.A.3. To check the reliability and reproducibility of the analyses, in-house and internationally certified standards (BAS1, BHVO-2) were measured as well as sample duplicates. Analytical errors are reported with the data.
4.4.3 Interpretation of geochemical data
Alteration and metamorphism will have affected the geochemistry of the analyzed rocks. Of the major elements only Al, Ti, Mn and P can be regarded as immobile. Therefore, classification of the rocks on the basis of major elements is problematic. The LILE elements (Rb, Ba, K, Th, U, Sr) are expected to be most affected by alteration. The other trace elements (HFSE and REE) are expected to have remained immobile (Rollinson, 1993). Certain trace element ratios can be used to check for the degree of alteration. Sr and Ba are less mobile than Rb, which is extremely mobile. Alteration commonly results in anomalously high Ba/Rb ratios and low Rb/Sr ratios. In general Ba/Rb ratios >15 and Rb/Sr ratios <0.03 are an indication of severe alteration (Hofmann and White, 1983). In Figure 4.7 these ratios are normalized and plotted; it can be observed that some of the samples show signs of loss of mobile elements. The LOI values also give an indication of the degree of alteration, as especially in ultramafic rocks the unstable mineral assemblages may have been altered to hydrous minerals. Early hydration under greenschist conditions appears to have been important. Metamorphic processes have changed the mineral assemblages but this does not necessarily imply that the bulk rock chemistry has changed. The recognition of primary igneous textures in the field and in thin section, and primary igneous trends in terms of Al2O3 and TiO2 versus MgO (Figure 4.8) is an indication of isochemical changes.
In order to interpret the geochemical data, the following approach was taken. The samples have been ordered into groups, based on the field lithology as shown in the stratigraphic column (Figure 4.4). On the basis of REE patterns some of these groups have been re-combined (Figure 4.10). Primitive mantle-normalized trace element patterns are shown (Figure 4.11), and MORB-normalized spidergrams of incompatible elements are used to determine the similarities and differences between the basaltic samples, and E- type and N-MORB rocks (Figure 4.12). Further more, selected trace elements are used to place further constraints on the tectonic regime and associated petrological processes (Figure 4.9, Figure 4.13).
The AFM major element diagram (Figure 4.9) is used to distinguish between tholeiitic and komatiitic rocks (Irvine and Baragar, 1971). Two representative major element variation diagrams, of Al2O3 and TiO2 versus MgO, are used to investigate fractionation trends in the basaltic series (Figure 4.8). The minor element oxides TiO2, MnO and P2O5 can also be used to subdivide basalts. These elements are regarded immobile in hydrothermal alteration and metamorphism up to lower amphibolite facies (Mullen, 1983). In addition, the trace elements La, Y and Nb can be used to discriminate between different types of basalt (Cabanis and Lecolle, 1989). Altered and metamorphic rocks may show some distortion relative to the La apex, because La is not fully immobile under greenschist metamorphic conditions.
It should be noted that the discrimination fields in the presented diagrams are based on modern tectonic regimes. Their use for Archaean studies may be limited, because the Archaean crust and mantle were probably different in their chemical and thermal properties. Firstly, the samples have experienced alteration and metamorphism, although it is shown that this is limited. Furthermore, tectonic and crustal processes may have been different in the Archaean. The geochemical signature of a rock records the physical-chemical condition of the source region, i.e. the composition of the source and the temperature, depth and degree of melting at which the melt was formed, and possibly contamination that occurred during magma migration. A similar geochemical signature does not necessarily imply a similar tectonic regime as is the modern case.
All basaltic samples plot in the tholeiitic field on the AFM diagram (Figure 4.9.a). Their spread relative to the A apex may be caused by mobility of the alkali components. In terms of their Mn-Ti-P variation, the basalts they plot in the arc tholeiitic field (Figure 4.9.b), whereas the Y-La-Nb composition points to the continental to back-arc field (Figure 4.9.c). The Y-Nb-Zr compositions plot in the MORB and intraplate-tholeiite fields (Figure 4.9.d). An evolution of minor olivine followed by mainly plagioclase fractionation can be recognized in the major element variation diagrams (Figure 4.8). The Ni content of these rocks is relatively high for basalts (70 ppm).
All basalts, irrespective of their stratigraphic position, are slightly LREE enriched and show a Eu and Sr anomaly confirming plagioclase fractionation (Figure 4.10). The primitive mantle normalized pattern (Figure 4.11) is relatively flat, but overall enriched. The N-MORB normalized trace element pattern shows a slope, however, the E-MORB normalized pattern is flat around a value of 1 (Figure 4.12). The messy appearance on the left side of these diagrams is caused by the mobility of these LILE elements, however, there may in fact be some real enrichment in fluid mobile elements. The trace element geochemistry of these basalts of the North Star Basalt shows resemblance to modern E-MORB, however, there are distinct negative Ta, Nb and Ti anomalies. This points to crustal contamination. This is also shown in Figure 4.12.a, where the basalts are plotted against a background of the values found by Green et al. (2000), who suggested continental crustal contamination for the upper Warrawoona Group. In Figure 4.12.b the data are plotted against continental flood basalts, showing a difference in enrichment.
On the basis of similar REE patterns (Figure 4.10), the pyroxenite, dolerite and ultramafics are grouped into one suite. The Lower Ultramafic Unit is peridotitic-komatiitic: it has a MgO contents between 15% and 35% and all samples plot in the komatiitic field on the AFM diagram (Figure 4.9.a). These extremely high MgO contents are characteristic of olivine and pyroxene cumulates. The series shows an evolution of olivine crystallization, followed by minor pyroxene and plagioclase at the end of the fractionation process (Figure 4.8). Small negative Sr and Eu anomalies confirm minor plagioclase fractionation. In the field pyroxene cumulate lenses have been observed (Figure 4.5.a). In Figure 4.9.b (Mn-Ti-P) the samples plot in the arc tholeiitic field, in Figure 4.9.c (Y-La-Nb) in the continental to back-arc field, and in Figure 4.9.d (Y-Nb-Zr) in the MORB and intraplate-tholeiite fields. Their scatter largely overlaps with that of the basalts, suggesting a common source.
The ultramafic suite does not have the negative Nb, La and Ti anomalies observed in the primitive mantle normalized spidergrams of the basalts (Figure 4.11). The Upper Extrusive unit contains an ultramafic series that has slightly lower MgO contents, but shows a similar evolution. The Upper Extrusive Ultramafics and the ultramafic dykes have slightly higher trace element concentrations than the lower units. All ultramafic rocks have extremely high Ni and Cr contents. The overall concentration of trace elements is low, indicating that the melts were extracted from their source by a high degree of partial melting (Figure 4.11).
The gabbroic suite can be divided in two groups based on the chondrite normalized REE patterns. One group has LREE enrichment, the other group has a flat to slightly LREE depleted pattern and is overall less enriched (Figure 4.10). The smaller patches of gabbro identified in the field, show more LREE enrichment than the larger gabbroic bodies. The LREE enriched group is interpreted to be genetically related to the main basaltic stratigraphy (sub-volcanic sills). The second group is interpreted to represent a separate suite of intrusions.
These separate gabbros show large variation in their major element compositions, but they all fall within the tholeiitic field (Figure 4.9.a). The other triplots again suggest a common source, but they do confirm that the separate gabbro is overall less enriched than the basalt. In Figure 4.8 a fractionation pattern of olivine and pyroxene followed by plagioclase fractionation can be recognized. The primitive mantle normalized trace element pattern for the separate suite of gabbro does not show the negative Nb and Ti anomalies recognized in the gabbro that is similar to the basalt (Figure 4.11).
The dioritic rock suite is represented by the diorite sills and their feeder dykes. They show a large amount of variation in their MgO values, but the trends are interpreted to be of magmatic origin. They plot as a range of tholeiites in Figure 4.9.a. The early magma evolution is dominated by olivine fractionation, later followed by plagioclase fractionation (Figure 4.8). This is confirmed by their slight negative Eu and Sr anomalies. The primitive mantle normalized trace element pattern shows even more strongly developed negative Nb and Ti anomalies than those recognized in the basalt (Figure 4.11).
The ultramafic parts of the composite dykes have been included in the ultramafic suite. In the mafic dykes, two groups can be recognized on the basis of the chondrite-normalized REE pattern (Figure 4.10). The first group is identical to the basaltic rocks in the area, and is interpreted to be genetically related. This is confirmed by the field relations (section 4.3.3). The second group is different in its flatter pattern and slight LREE depletion, and lack of negative Nb and Ti anomalies (Figure 4.11). It is similar to the Gabbro Sill and interpreted to be genetically related to that gabbro.
4.4.4 Discussion of geochemical data
The basalts and related dolerite feeder dykes and gabbro sills of the North Star Basalt are typical Archaean tholeiites. They are slightly LREE enriched with a weak negative Eu anomaly. The negative La, Nb and Ti anomalies reflect that the magmas have been enriched by a crustal component (Thompson et al., 1984). This is also illustrated with incompatible element ratios (Figure 4.13.b and c). This suggests that the magmas were either derived from a subduction-related enriched mantle source, or that the crustal component was added by assimilation during passage of the magma through continental crust. Our results are very similar to the data published by Green et al. (2000) for the Coonterunah and upper Warrawoona Groups in the Pilgangoora Belt, as illustrated in Figure 4.12.c. They showed by REE modeling that assimilation is a good explanation for the observed geochemistry, confirming what Barley (1986) has also suggested. In addition, they found granitic xenoliths. Therefore it may be concluded that the North Star Basalt was erupted onto a chemically evolved silicic basement. However, the Main Basalt is less enriched than modern continental flood basalts (Figure 4.12.d).
The Lower Ultramafics including the pyroxenite are distinct in their geochemistry, especially the very high Mg contents. Differentiation trend are characteristic of olivine fractionation. Low incompatible trace element concentrations and LREE depletion (Figure 4.10) indicate a large degree of partial melting of a source already depleted by removal of clinopyroxene. High Ni and Cr contents also indicate these rocks formed by a large degree of partial melting (Figure 4.13.a).
The Lower Ultramafics are interpreted to have a different magma source than the main basalts in the stratigraphy. The ultramafics were differentiated at greater depth as indicated by the absence Eu and Sr anomalies indicative of plagioclase fractionation. Their difference from the basalts is confirmed by incompatible trace element ratios (Figure 4.13). The Lower Ultramafics have no negative Nb and Ti anomalies, so they somehow did not assimilate much crustal material during magma migration. They can be classified as Barberton-type (low-Al) komatiites (Nesbitt and Sun, 1976; Arndt and Brooks, 1980). We suggest that the Lower Ultramafics are not related to the Main Basalt, and therefore they may not be part of the North Star Basalt sensu stricto. Unfortunately, Green et al. (2000) could not distinguish between the Coonterunah and Warrawoona Basalts on the basis of major or trace elements, and therefore we cannot determine whether the Lower Ultramafics belong to the Coonterunah or Warrawoona Group.
The gabbroic sills and later north-northwest trending dykes with flat REE patterns, and the dioritic intrusions, are interpreted to be genetically unrelated to the extrusive stratigraphy. They show much stronger characteristics of crustal contamination, which again can be explained either by a subduction-enriched source or crustal assimilation. In the incompatible element plots (Figure 4.13.c and d) they plot distinctly separate from each other, and from the basalt and the ultramafics. Therefore it is thought that they come from different sources.
4.5 Structural geology
On a regional scale, the lower Marble Bar Belt is folded in the Talga Talga Anticline, as can be seen on the Landsat image (Figure 4.2). Away from the core of the anticline, the limbs are increasingly destroyed by intrusions of the Mount Edgar Granitoid Complex (see map in Figure 4.3). Generally, the greenstones on the western side of the mount Edgar Granitoid Complex dip radially away from the granite. In the area studied the average dip of the bedding is between 20 and 35 degrees to the northwest. The dips steepen towards the northwest.
The North Star Basalt contains two bedding-parallel zones of foliated rock consisting of serpentine-carbonate schist. The foliation within the schist has a variable orientation. The average strike of the foliation is same as the strike of the bedding, but the foliation dips more steeply than the bedding, at 45-65°. From this it may be concluded that the schistose zones are in fact shear zones in which case there may have been west-block-up thrust movement along them (in the present orientation). However, the local preservation of primary igneous (spinifex) textures (Figure 4.5) indicates that the amount of strain in these zones may not be very large. The fact that the S-foliation within the high strain zones still occurs at an angle to the shear plane, is also an indication that the amount of strain is low.
Thin Chert and the Big Chert dip 10-20° steeper than the bedding of the underlying rocks. This disconformity may be caused by vertical movement and reorientation of the extrusive pile before deposition of the cherts. This vertical movement could either be related to magmatism, e.g. effects of cooling of the extrusive pile or the underlying magma chamber, or it is caused by regional scale deformation of the basement during or just after formation of the North Star Basalt. This can not be resolved at the scale of this study.
Just below the both chert units, the underlying basalts are foliated. The geometry of the foliation in the field indicates normal movement occurred along the cherts: the foliation curves from subhorizontal, i.e. with a normal sense, into the cherts. The relative timing of movement within the carbonate schists and the chert units can not be resolved, because the structures have no cross cutting relations.
The composite dyke systems are sheared. The foliation at the edges of the dykes indicates that they are sinistral strike slip shears: together they form an array with the vertical main shears striking ca 310 and vertical minor shears striking ca 340 (Figure 4.6). Kinks folds in the foliation at the sides of the dykes indicate dextral overprinting.
The Thin Chert contains meso-scale cylindrical, open folds with a wavelength of about 2m. Their fold axes plunge to the northwest. In two locations a different type of folding was found within the Thin Chert: these folds have horizontal fold axes parallel to the strike of the bedding. They are asymmetrical; when looking towards the southwest they are Z-shaped. This may indicate normal movement occurred along the chert, as was also concluded from the foliation described above.
In the Clastic Sedimentary Unit a variety of folds can be observed: in many locations the finely laminated sub-unit is folded at cm-, dm-, and m-scale: the folds are open to tight. The fold closures occur in all directions, some are dismembered, and locally the layers are weakly boudinaged. These folds are interpreted to be slump folds.
Minor folds in the ultramafic part of the composite dyke systems are closed, cylindrical and symmetric, they locally have a box shape. Their fold axes are sub-vertical. The asymmetry of these folds as well as the offset of lithological units show that these structures have a large dextral offset. This dextral phase overprints the sinistral movement, and is interpreted to have occurred at the same time as the development of the dextral faults.
4.5.4 Brittle faults
In the lower massive ultramafics and in the main basalt two sets of small vertical faults were observed. The faults are 1-20 cm wide. Some are filled with quartz, which very occasionally shows slickensides plunging 40° to the northwest. The largest faults of this type are traceable for no more than 100-200m. The orientation of faults is shown in Figure 4.6.c. Within these faults different elements of the brittle Riedel Array can be recognized; most faults strike due west and show a dextral sense of shear. The set striking northwest show a sinistral sense of shear. These sets form a conjugate system, and the lineations indicate the NW block was brought up relative to the southeast. The amount of displacement on these structures could not be determined as they occur only in homogeneous parts of the stratigraphy.
A number of subvertical northwest trending faults were observed in the Upper Extrusive Unit (see map in Figure 4.3). These structures have a dextral strike-slip component. Some of them do not displace the Big Chert and must therefore be synvolcanic, formed before formation of the Big Chert (McPhee Formation). A second set of large faults occurs in the top of the stratigraphy: they do displace the Big Chert by 1-20m. As they are located in small valleys the outcrop of the actual faults could not be observed. The absolute sense of shear could not be determined, but the relative displacement of the Big Chert in the present orientation indicates that there was a dextral strike slip component. Some of these faults coincide with composite dykes, but there are also a number of faults that are separate.
The margins of the mafic dykes show evidence for both the dextral as well as sinistral movement. This indicates that this set of mafic dykes as well as the ultramafic dykes both predate the initial sinistral movement on the shear zones. The late set of dolerite dykes is undeformed, indicating it postdates the last deformation. However, its emplacement is interpreted to be related to regional northeast-southwest extension (Kloppenburg et al., 2001), forcing them in this orientation and creating the stepped outcrop pattern.
4.5.5 Discussion of structures
The relative timing of structures is constrained by field relations. The first structures that developed were northwest trending vertical faults in the Upper Extrusives, that do not cut the Big Chert. They can either be interpreted as synvolcanic faults or as normal faults related to regional extension.
The relative timing of the next two events (east-west extension and east-west compression) has not been determined because there are no cross-cutting relations. An east-west extensional event led to deformation along the chert units. The foliation in the rocks adjacent to the cherts is interpreted to be associated with these structures and indicates normal movement. Small scale folds in the clastic sedimentary unit indicate normal movement occurred on planes approximately parallel to the bedding. It is interpreted that movement in the cherts and the clastic sediments occurred during the same event. This east-west extensional event is possibly related to an early phase of extension at about 3.46 Ga as proposed by Zegers et al. (1996), but the extent of this event has been questioned by White and Hickman. The increase of the dip of the stratigraphy across the chert units is interpreted to be a primary feature: the cherts are thought to be deposited on low angle unconformities. This is also shown on the map (Figure 4.3).
Two ductile layer-parallel shears were observed in the komatiitic units, now carbonate schists. The foliation is steeper than the bedding and is interpreted as an S-fabric, indicating reverse movement occurred on the shears. This is interpreted to be related to a regional east-west compressional event that occurred between 3460 and 3350 Ma (Van Haaften and White, 1998). The thrusts are bedding-parallel and have not been observed to cross-cut the stratigraphy. Because primary igneous features such a spinifex textures can locally be recognized in the schist, and the foliation occurs at an angle to the shear plane, the amount of displacement was probably minor.
After these events, shears were initiated in the composite dykes, first with a sinistral shear sense; then the principal stresses shifted resulting in shear reversal on the sheared composite dyke system. The last tectonic event resulted in the formation of a conjugate set of small brittle faults due to northwest-southeast directed compression.
4.6 40Ar/39Ar Geochronology
Sample descriptions are summarized in Table 4.2. The mineral separation and 40Ar/39Ar step heating experiments were carried out at the isotope laboratory of the Vrije Universiteit van Amsterdam. We followed the analytical methods developed by Wijbrans et al. (1995).
Age spectra of the samples are shown in Figure 4.14. A hornblende-pyroxene lens in the Lower Ultramafics has an 40Ar/39Ar cooling age of 3472 ± 21 Ma, which falls within the generally accepted age of the Warrawoona Group (Table 4.1). The Gabbro Sill low in the stratigraphy has cooling ages of 3290 ± 18 Ma and 3272 ± 21 Ma. This may record slow cooling of the ca 3300 Ma granitoids nearby (McNaughton et al., 1993; Nelson, 2000). Alternatively this may correspond to tectonothermal events recorded elsewhere in the Pilbara Craton, in the Sulphur Springs and Roebourne Groups (Van Kranendonk et al., 2002). Another sample from the Lower Ultramafics has a strongly disturbed age spectrum with a semiplateau at 3127 ± 77 Ma. This does not appear to be related to any known intrusions in the area, however, it is a widespread 40Ar/39Ar resetting age throughout the Pilbara Craton (see Figure 8.2).
4.7 Tectonic synthesis
The origin of Archaean granitoid-greenstone terrains has been discussed in the context of a wide variety of modern analogues such as continental rifts, continental margins, oceanic island-arcs, submarine plateaus, hot spot type oceanic settings and mid ocean ridges (Kröner and Layer, 1992; Condie, 1994; De Wit and Ashwal, 1997). One of the aims of this study was to investigate a possible ophiolitic origin of the North Star Basalt. Modern ophiolites are an association of allochthonous rocks that, while typically found on continents, are interpreted to represent oceanic crust and upper mantle. A complete ophiolite sequence as defined by the Geological Society of America is several kilometers thick. It is an allochthonous complex consisting of peridotite tectonite, cumulate gabbro, a mafic sheeted dyke complex and a mafic volcanic complex (Anonymous, 1972).
Processes of oceanic crust formation in Archaean times may have differed from those at modern oceanic spreading centers (De Wit et al., 1992; Vlaar et al., 1994) and consequently oceanic crust may have had a different composition and stratigraphy (Bickle et al., 1994). Considering that Archaean oceanic crust is thought to be substantially thicker than present-day oceanic crust (Bickle et al., 1994), the thickness of the North Star Basalt of ca 1 kilometer is not great enough to contain a complete ophiolite. The absence of a peridotite tectonite sequence may be explained by the fact that the lower contact of the North Star Basalt is the intrusive contact with the Mount Edgar Granitoid Complex. Geochemistry has shown that most of the gabbro in the area is genetically unrelated to the Main Basalt sequence, but are later mafic intrusions. Finally, even though in the North Star Basalt a large number of mafic and ultramafic dykes is present, this does not represent a sheeted dyke complex. Field relations and geochemistry have shown that very few of them are genetically related to the basaltic sequence and they only account for a small percentage of the total volume of rock. The dykes were formed as a result of distinct tectonic events: they are not the product of a single phase of extension at a mid ocean ridge. Even though the North Star Basalt does not resemble a modern ophiolite, the rocks did form in an under water environment as recorded by the presence of pillowed basalts.
The major and trace element geochemistry of the extrusive pile and related dykes and sills resembles that of EMORB. The higher degree of partial melting in the Archaean mantle, in combination with a higher fertility, may have resulted in a higher Fe and REE content than in modern MORB (Ohta et al., 1996). However, the LREE enrichment and negative Nb and Ti anomaly indicate some involvement of more evolved rocks. This suggests that the magmas were either derived from a subduction-related enriched mantle source, or that the crustal component was added by assimilation during passage of the magma through continental crust. Green et al. (2000) came to similar conclusions for the Coonterunah and Upper Warrawoona Groups in the Pilgangoora Belt. They showed by REE modeling that assimilation is a good explanation for the observed geochemistry, confirming what Barley (1986) has also suggested. Bickle et al. (1994) showed that demonstrably continental sequences contain successions resembling ocean floor, with mafic lavas and gabbroic, peridotitic and dunitic cumulates in intrusive bodies. It may be concluded that the North Star Basalt was erupted onto a chemically evolved silicic basement.
Trace element geochemistry suggests that the lower part of the stratigraphy (the Lower Ultramafics) may not be part of the North Star Basalt as presently defined. The stratigraphic position of the Lower Ultramafics below the Main Basalt of the North Star Basalt could lead to the suggestion that it may actually be part of the Coonterunah Group. However, Green et al. (2000) showed that the Upper Warrawoona and Coonterunah Groups are geochemically very similar. Therefore it is concluded that the Lower Ultramafics represent a distinct volcanic sequence. However, an 40Ar/39Ar cooling age of about 3470 Ma suggests the lower ultramafics are age equivalent of the lower Warrawoona Group.
Structures in the North Star Basalt record five tectonic events (see Table 4.3) which can be dated by previously published geochronology. The oldest structures are vertical faults at a high angle to the bedding. They offset the extrusive pile which has an angular upper contact with the overlying chert of the McPhee Formation. The relative timing of the next two events is unclear. The formation of strike parallel normal faults adjacent to or in the chert units. These are possibly associated with a phase of continental extension (Zegers et al., 1996) involving regional greenschist grade metamorphism. This was followed by east-west directed compression that resulted in the activation of bedding parallel thrusts in komatiitic units, and the formation of the Talga Talga antiform (Van Haaften and White, 1998). However, this study has shown that displacement on these structures (within the North Star Basalt) was probably minor. After the compression, a set of northwest-trending composite dykes was emplaced, cross cutting older structures.
The next deformational event, which post-dates the emplacement of the composite dykes, was directed west-southwest east-northeast and initiated sinistral strike slip shears in the composite dykes. These sheared dykes were reactivated as dextral shears. The last event has created a conjugate set of minor brittle shears due to NW-SE directed compression. Finally, a set of north-northwest trending dolerite dykes was intruded. This possibly occurred during southwest-northeast extension at a time when the Mount Edgar Marginal Shear was under extension (Kloppenburg et al., 2001). Major granitic intrusions at 3.3 Ga (McNaughton et al., 1993) post-date emplacement of all of the dyke suites, and have destroyed the lower section of the greenstone sequence.
4.8 Summary and Conclusions
The North Star Basalt comprises five different components based on lithological and geochemical characteristics. The lower part of the stratigraphy consists of LREE depleted Barberton-type komatiites and pyroxene cumulates. They are genetically not related to the overlying basalts, and it is not known whether they belong to the Coonterunah Group or the Warrawoona Group. The main basalt stratigraphy is LREE enriched with negative Nb and Ti anomalies. The basalts are interpreted to be contaminated tholeiites, enriched with continental crustal material during magma migration. This is similar to the findings of Green et al. (2000) for the Coonterunah and upper Warrawoona Group in the Pilgangoora Belt. Gabbro dykes and sills, and dioritic intrusions make up two minor components of the North Star Basalt, and they are interpreted to be genetically unrelated to the extrusive stratigraphy. Two chert units are interpreted to be silicified slatey sediments. Some displacement occurred along them.
Lithologically and geochemically the North Star Basalt does not resemble a modern ophiolite, however, the rocks were erupted in water as recorded by the pillowed basalts. The majority of the dykes in the area were emplaced during distinct regional events, few are genetically related to the basalt. The complex structural history of the area involved syn-depositional faulting, regional extensional en compressional events at ~3450 Ma and ~3300 Ma respectively. Therefore the Talga Talga Anticline may still be a suitable type area for the North Star Basalt, but the presence of low angle unconformities should not be disregarded. Early normal faults developed along these unconformities, however, the displacement on all structures in the area was probably minor.
The field research was funded by grant 1997/14 of the Dr. Schürmann Fund. Marije van Koolwijk participated in the mapping and carried out the geochemical analyses as part of her M.Sc. project in 1997. She is thanked for making her geochemical data available for this paper. I would like to thank Dr. Kim Hein for reviewing an early version of this manuscript.