Record 2015/11: The Sandstone greenstone belt, northern ......regional geophysical datasets, the results of regional geophysical inversions, and out-of-section geometrical features

Geological Survey

of Western Australia

RECORD 2015/11

THE SANDSTONE GREENSTONE BELT, NORTHERN CENTRAL YILGARN CRATON: 3D MODELLING USING GEOLOGICAL AND GEOPHYSICAL CONSTRAINTS

by RE Murdie, K Gessner, and SF Chen

Government of Western AustraliaDepartment of Mines and Petroleum

Record 2015/11

THE SANDSTONE GREENSTONE BELT, NORTHERN CENTRAL YILGARN CRATON: 3D MODELLING USING GEOLOGICAL AND GEOPHYSICAL CONSTRAINTS

byRE Murdie, K Gessner, and SF Chen

Perth 2015

MINISTER FOR MINES AND PETROLEUM

Hon. Bill Marmion MLA

DIRECTOR GENERAL, DEPARTMENT OF MINES AND PETROLEUM

Richard Sellers

EXECUTIVE DIRECTOR, GEOLOGICAL SURVEY OF WESTERN AUSTRALIA

Rick Rogerson

REFERENCE

The recommended reference for this publication is:

Murdie, RE, Gessner, K and Chen, SF 2015, The Sandstone greenstone belt, northern central Yilgarn Craton: 3D modelling using

geological and geophysical constraints: Geological Survey of Western Australia, Record 2015/11, 33p.

National Library of Australia Card Number and ISBN 978-1-74168-638-8

Grid references in this publication refer to the Geocentric Datum of Australia 1994 (GDA94). Locations mentioned in the text are

referenced using Map Grid Australia (MGA) coordinates, Zone 50. All locations are quoted to at least the nearest 100 m.

Disclaimer

This product was produced using information from various sources. The Department of Mines and Petroleum (DMP) and the State

cannot guarantee the accuracy, currency or completeness of the information. DMP and the State accept no responsibility and disclaim

all liability for any loss, damage or costs incurred as a result of any use of or reliance whether wholly or in part upon the information

provided in this publication or incorporated into it by reference.

Published 2015 by Geological Survey of Western Australia

This Record is published in digital format (PDF) and is available online at <www.dmp.wa.gov.au/GSWApublications>.

Further details of geological products and maps produced by the Geological Survey of Western Australia

are available from:

Information Centre

Department of Mines and Petroleum

100 Plain Street

EAST PERTH WESTERN AUSTRALIA 6004

Telephone: +61 8 9222 3459 Facsimile: +61 8 9222 3444

www.dmp.wa.gov.au/GSWApublications

iii

Contents

Abstract ..................................................................................................................................................................1 Introduction ............................................................................................................................................................1

Location ..........................................................................................................................................................2Geological setting ...................................................................................................................................................2

Tectonic setting ...............................................................................................................................................2Stratigraphy .....................................................................................................................................................3Structural history .............................................................................................................................................4Economic geology ...........................................................................................................................................6

Geophysical data ....................................................................................................................................................6Gravity .............................................................................................................................................................6Total magnetic intensity ..................................................................................................................................6Seismic reflection ............................................................................................................................................6Depth to Moho ................................................................................................................................................7Physical properties ..........................................................................................................................................7Magnetotelluric surveys ................................................................................................................................10

Software ...............................................................................................................................................................10GM-SYS ........................................................................................................................................................10GeoModeller .................................................................................................................................................10GOCAD ........................................................................................................................................................10

Input to the model ................................................................................................................................................11Project coordinate system .............................................................................................................................11Topography ...................................................................................................................................................11Surface geology .............................................................................................................................................13Cross-sections ...............................................................................................................................................152D potential-field forward modelling............................................................................................................15

Modelling in 3D ...................................................................................................................................................15Model stratigraphy ........................................................................................................................................15Building the model ........................................................................................................................................19Forward modelling of the block model .........................................................................................................19Property optimization ....................................................................................................................................19

3D modelling and inversion in GOCAD ..............................................................................................................20Gravity modelling .........................................................................................................................................20Magnetic modelling ......................................................................................................................................22

Implications of the modelling results for the regional structure ..........................................................................23Comments on the process of model building .......................................................................................................28Acknowledgements ..............................................................................................................................................28

References ............................................................................................................................................................29

Appendix

Metadata Table 1: Model object list, notes and data references ..........................................................................31

Figures

1. Tectonic setting of the Sandstone greenstone belt within the Yilgarn Craton ..............................................3

2. Simplified map of the interpreted geology of the study area .......................................................................4

3. Stratigraphic column from geological maps, related to the stylized stratigraphic pile used for the

3D model ......................................................................................................................................................5

4. Total magnetic intensity (TMI) over the study area .....................................................................................7

5. Bouguer gravity anomaly over the study area .............................................................................................8

6. Interpreted sections of the 10GA-YU2 seismic traverse line ......................................................................9

7. Flow diagram summarizing the model building process ...........................................................................12

8. Topography of the model area from SRTM ..............................................................................................13

9. Map showing the areas of outcrop over the Sandstone greenstone belt ....................................................13

10. Locations in 3D space of cross-sections used in building the model ........................................................14

11a. Forward 2.5D gravity model of the 10GA-YU2 seismic line ....................................................................16

11b. Forward 2.5D gravity model of the user-defined north–south section at 726000E ...................................17

11c. Forward 2.5D gravity and magnetic models of the Atley cross-section .................................................18

12. Construction points from each forward-modelled section for four geological units located in

3D space .....................................................................................................................................................20

iv

13. Development of the model using the Bouguer anomaly as a constraint ....................................................21

14. Results of the basement inversion viewed from three different angles .....................................................22

15. Results of the heterogeneous inversion showing the variations in density in the granites ........................23

16. Magnetic anomaly maps of the model during the building process ..........................................................24

17. Results of heterogeneous magnetic susceptibility inversions ...................................................................25

18. Surfaces and volumes in the final 3D model ........................................................................................26, 27

Tables

1. Table of densities (gcm-3) used throughout the modelling process ...........................................................10

2. Table of magnetic susceptibilities (SI) used in the modelling process ......................................................11

3. Dimensions of the 3D model using the UTM Zone 50 coordinate system ...............................................13

GSWA Record 2015/11 Sandstone greenstone belt, northern central Yilgarn Craton: 3D modelling

1

The Sandstone greenstone belt, northern central

Yilgarn Craton: 3D modelling using geological and

geophysical constraints

by

RE Murdie, K Gessner, and SF Chen

AbstractThis report documents the procedures used in, and results from, generating a 3D geological model of the Sandstone

greenstone belt in the northern central Yilgarn Craton. The modelling procedures have been documented in detail as

similar procedures will be adopted by the Geological Survey of Western Australia for producing other regional 3D models,

although availability of data and geological context are likely to result in minor variations on these steps. The Sandstone

greenstone belt is located on the boundary between the Southern Cross and Murchison Domains of the Youanmi Terrane

of the Yilgarn Craton. Geological mapping of the area, where extensive sheetwash and other regolith cover obscure

large portions of the bedrock, was completed in 2005. Projection of structural data to depth led to a hypothesis that

the Sandstone greenstone belt is a refolded syncline. Model generation relied on the contrasting densities between the

surrounding granites and the greenstone belt, and the strong magnetic susceptibility of the banded iron-formation units

within the greenstone. These physical characteristics were exploited during inversion of high-resolution potential-field

data to define subsurface density and susceptibility distributions. The extents of these dense and magnetic bodies were

defined by appropriate property value distributions and compared to profile information from a seismic reflection survey.

The resulting 3D model constrains the thickness of the greenstone belt and geometry of its internal structure, which

confirms the refolded syncline hypothesis. Although the Sandstone greenstone belt is located in the hanging wall of the

tectonic domain-bounding Youanmi Shear Zone, our model does not provide evidence for distinct differences between

the Murchison Domain in the west, and the Southern Cross Domain to the east of the Youanmi Shear Zone.

KEYWORDS: banded iron-formation, cross-sections, geological maps, geophysical modelling, granite, greenstone,

seismic profile, three-dimensional models

IntroductionThis report details the Geological Survey of Western Australia (GSWA) in-house workflow for producing a three-dimensional (3D) geological model, using the Sandstone greenstone belt as an example. The Sandstone model is the first 3D model produced by GSWA in a project to generate 3D structural representations of the crust and upper lithosphere in Western Australia. Three-dimensional models of the subsurface geology provide a crucial contribution towards understanding complex geological phenomena and verification of geological concepts. This is achieved by placing independent geoscience datasets in a spatial context and extrapolating features where there are no input data. This approach is of particular significance to mineral exploration because mineral deposits are commonly located in structurally complex regions of the Earth’s crust, and the search for mineral deposits expands into areas of limited available data. Information obtained through surface geological mapping and structural measurements taken in the field can indicate gross geological structure which can be

inferred to continue to depth. Cross-sections showing diagrammatically the implied subsurface geology are routinely produced with each 1:100 000 map published by GSWA. While useful, cross-sections only give a limited 2D visualization of the subsurface and, in more complicated environments, many cross-sections are needed to build a comprehensive visualization in 3D space. Furthermore, 3D models also take into account regional geophysical datasets, the results of regional geophysical inversions, and out-of-section geometrical features to verify and test geological and structural frameworks between 2D sections in complex terrains. 3D models also force reconciliation of inconsistencies in 1D and 2D datasets that may go undetected without the data being projected into 3D space.

High-quality 2D datasets, such as gravity and magnetic field surveys, can be used to infer the depth and shape of subsurface bodies. Gravity and magnetic surveys are routinely acquired by the various Australian geological surveys and, together with commercial open-file surveys, are publicly available through the Geophysical Archive

Murdie et al.

2

Data Delivery System (GADDS) accessible through the Geoscience Australia website <www.geoscience.gov.au>. Forward modelling of density and magnetic susceptibility distributions against these datasets can be used to infer the extent of geological formations. However, there is always inverse problem ambiguity between physical properties and boundaries of the causative body. Three-dimensional inversions of such data are useful in further defining property distributions and sources at depth, but carefully researched and measured input parameters and constraints are required to produce meaningful results (Oldenburg and Pratt, 2007). The results of inversions also need careful examination before geological structure can be inferred; for example, by checking against the original hypothesis, geological viability (e.g. stratigraphic order is maintained, faults have the correct sense of motion, intrusive relationships and structural observations are maintained), and checking against other geophysical data such as seismic surveys.

Direct indications of depth to geological interfaces come from imaging techniques such as seismic reflection surveys or inversion of electromagnetic soundings. Unfortunately, these typically only provide information on profiles unless 3D data are available. Although now routinely used in petroleum exploration, 3D seismic surveys are rarely used in bedrock areas because the steep structures commonly associated with mineralized areas are technically challenging to process and interpret, and the costs are relatively high. However, 2D information from seismic and electromagnetic surveys is very useful as an input constraint to inversions of potential-field data and for setting a starting depth for a boundary in forward modelling.

Used together, these geoscience datasets can be combined in 3D space to test hypotheses and be used as a template for predictions.

The Sandstone area was chosen to build a 3D model for several reasons:

1) Detailed surface mapping was most recently completed by Chen (2003) and Chen and Painter (2005).

2) A conceptual geological model based on surface mapping (Chen et al., 2003) had not been fully tested by other methods.

3) Seismic line 10GA-YU2 (Zibra et al., 2014) provides a depth dimension not previously available.

4) High-resolution gravity and magnetic data are available over the area (GSWA, 2013, 2014a).

5) It has a relatively simple geometry to use as a testing ground for the development of a 3D product.

The Sandstone area is part of a wider study of the Yilgarn Craton, which is of importance to the economy of Western Australia because of its mineral potential, in particular for the greenstone areas that are known to host gold, nickel and base metal resources (Cooper et al., 2015). The objective of imaging the Yilgarn with 2D seismic and magnetotelluric surveys (Zibra et al., 2014) was to investigate the present architecture and, by inference, the

geodynamic setting during the time of crust formation and mineralization. Of particular interest are the structure of the middle and lower crust, thickness of the greenstone belts, and orientations and depth penetrations of major shear zones. These shear zones are mapped at the surface, but questions remain concerning their penetration through the crust to the mantle and their potential to act as fluid pathways for mineralization.

Location

The Sandstone greenstone belt is located in Western Australia between latitudes of 27°30’S and 28°20’S and longitudes 119°E and 119°30’E within the granite–greenstone terrain of the Yilgarn Craton (Fig. 1). For convenience, the model stopped at the edges of 1:100 000 maps, although it may, in hindsight, been more geologically logical to extend it slightly to the east to encompass the eastern extent of the greenstone belt. The topography is typically flat, varying from 400–580 m on the Australian Height Datum (AHD), and the greenstone area has only subdued strike ridges and subrounded hills. The granite areas are generally rockier with flat pavements and breakaways up to 15 m high (Chen, 2005). The area was most recently mapped at 1:100 000 scale, published as the Sandstone and Atley sheets (Chen, 2003; Chen and Painter, 2005). Digital data covering the map areas is provided in GSWA (2008).

Geological setting

Tectonic setting

The Sandstone greenstone belt lies in the central-northern part of the Southern Cross Domain (Fig. 1; Cassidy et al., 2006), which is a part of the Youanmi Terrane within the Yilgarn Craton (Martin et al., 2015). The Sandstone greenstone belt is bounded to the northwest by the Youanmi Shear Zone (Myers, 1993; Myers, 1997; Myers and Swager, 1997; Fig. 2). This major structure forms the boundary between the Murchison Domain and the Southern Cross Domain, both of the Youanmi Terrane. To the northeast it is bounded by the Edale Shear Zone. These faults dip towards each other and meet at the apex of the syncline, cradling the greenstone rocks. The southern margin is intruded by several large granitic plutons (Chen, 2005).

Three other mafic successions outcrop in the project area: the Gum Creek greenstone belt to the northeast, the Barrambie Igneous Complex and associated supracrustal rocks to the north along the Youanmi Shear Zone, and the Unaly Hill greenstone belt to the southwest along the Youanmi Shear Zone (Fig. 2). The Barrambie Igneous Complex has been identified as belonging to the Meeline Suite of mafic–ultramafic layered intrusions. However, because of the steeply dipping stratigraphy, poor outcrop and strong deformation within shear zones, its relationship to the greenstone sequences is obscured (Ivanic et al., 2010). The North Cook Well greenstone belt is apparent in the gravity signal east of the Edale Shear Zone but


3

Figure 1. Tectonic setting of the Sandstone greenstone belt in the red box within the Southern Cross Domain of

the Youanmi Terrane of the northern Yilgarn Craton. Inset shows location of Yilgarn Craton within Western

Australia

does not outcrop within the model area (Chen, 2005). All adjacent mafic belts have been included in the study, but without any detail applied.

Stratigraphy

The Sandstone syncline is a discrete greenstone belt and, although similar greenstone belts, for example Forrestania (Perring et al., 1996) and Marda–Diemals (Chen et al., 2003), are known from elsewhere in the Southern Cross Domain, it is difficult to correlate the regional stratigraphy. Differences in the geochemistry and stratigraphic position of komatiites within the Sandstone greenstone belt compared with other greenstone belts of the Youanmi Terrane, and poor geochronological control, make correlations between the belts difficult (Chen et al., 2006). The Sandstone greenstone belt appears to be unlike other greenstones in that the komatiites are in the upper part

of the stratigraphy, whereas elsewhere they typically are much lower in the stratigraphy. However, faulted contacts and the lack of outcrop make it difficult to assign their exact stratigraphic position. A comprehensive stratigraphic succession has been established for the northern part of the Murchison Domain to the west where there is better outcrop and significantly better geochronological control (Van Kranendonk et al., 2013).

The age of the Sandstone greenstone sequences is not well constrained, but comparison with the Murchison stratigraphy suggests that they were deposited after 3.0 Ga.

The greenstone stratigraphy (Fig. 3) comprises a lower mafic-dominated succession intercalated with clastic sedimentary rocks, which are locally intruded by gabbroic sills. Greenstones adjacent to the granite contact are strongly deformed, but this effect decreases away from the contact. Stratigraphically above the mafic-dominated succession there is a thin banded iron-formation (BIF) and

Murdie et al.

4

Figure 2. Simplified map of the of the interpreted geology of

the study area taken from the 1:100 000 Geological

Series maps ATLEY (Chen 2003) and SANDSTONE (Chen

and Painter 2005)

chert unit which is intercalated with clastic sedimentary rocks and gabbro sills (Chen, 2003, 2005; Chen and Painter, 2005).

The BIF and chert unit is overlain by a tremolite–chlorite(–talc) schist. Overlying this is a massive, weakly deformed tholeiitic basalt, followed by a major interval of BIF and chert with gabbroic sills. In the core of the syncline, mafic rocks with thin BIFs are poorly exposed.

In the northern area, a fine-grained, clastic sedimentary layer with intercalated BIF and chert overlies the lower mafic layer. At the southern end there are widespread ultramafic rocks (dominantly komatiite) and subordinate BIF with minor basaltic and sedimentary rocks. The southern ultramafic succession appears to be in a structurally higher position than the basal mafic-dominated succession but the relationship is unclear due to poor exposure. Most information for this part of the succession has come from first intercepts in exploration drillholes (Chen, 2005).

The majority of the study area comprises granitic rocks, dominantly monzogranites, which are strongly foliated with a steep to subvertical subplanar fabric within the major shear zones.

All the rocks have been metamorphosed. Within the greenstone belt, high-strain zones on the margins of the belt are characterized by mid-greenschist to lower amphibolite assemblages, whereas lower metamorphic grade of lower greenschist to prehnite–pumpellyite facies at lower strain are typical for the greenstones away from the sheared margins (Ahmat, 1986; Chen, 2005).

A full description of the geology can be found in Chen (2005).

Structural history

Three deformation events have been recognized in the central-northern part of the Southern Cross Domain (Chen and Wyche, 2001; Chen et al., 2001; Greenfield, 2001; Wyche et al., 2001; Riganti and Chen, 2002; Chen and Wyche, 2003; Chen et al., 2003; Riganti, 2003; Chen et al., 2004; Wyche et al., 2004). D1 north–south compression produced easterly trending thrusts and folds and an F1 easterly trending syncline. D2 east–west shortening produced the major northerly trending upright folds in the greenstones. In the Sandstone syncline, this produced the doubly-plunging F2 box-fold geometry (Chen, 2005) and the two macroscopic north-northeasterly trending anticlines near the southern margin. Granites were intruded during the D2 and D3 deformation events (Chen, 2005). The present overall geometry of the greenstones and intervening granites was generated during the progressive and inhomogeneous D3 east–west shortening, whereby the competent granitic blocks impacted the less competent greenstone belts. This resulted in isolated greenstone belts with arcuate shear zones between them (Chen and Wyche, 2001; Chen et al., 2003; Chen et al., 2004). In the Sandstone area this produced the dextral Youanmi Shear Zone and sinistral Edale Shear Zone.


5

Figure 3. Stratigraphic column from Chen (2005) and Chen and Painter (2005) in black and white on

the right showing how it relates to the stylized stratigraphic pile used for the 3D model

Murdie et al.

6

These are several-kilometre-wide high-strain zones exhibited by strongly foliated granitic rocks. Such major crustal structures are of prime importance in understanding Archean tectonics and controls on mineral deposits.

As a consequence of this deformation sequence, the Sandstone syncline has a triangular geometry that is about 40 km wide on its southern margin and 35 km north to south. The overall structure has been described as a refolded fold (Chen, 2005).

Several prominent mafic dykes have been mapped from aeromagnetic features striking east and east-northeast (Fig. 4), and these are thought to be post-D3 in age. Equivalent mafic rocks are seen as sills in the seismic profile (Ivanic et al., 2014a,b).

Economic geology

The Sandstone greenstone belt has produced over 25 t of gold at mines such as Adelaide (1903–21 producing 6361 kg), Bulchina (1999–2004 producing 7152 kg), Havilah (1905–29 producing 1053 kg), and Sandstone (1904–93 producing 6508 kg). However, although there are known resources in the area, there are currently no active mines in operation (GSWA, 2014b).

Nickel mineralization was discovered in 2010 by Western Areas Ltd and further drilling is planned (Western Areas Ltd, 2015).

Geophysical data

Gravity

The Bouguer gravity anomaly grid was cut from the WA State grid, which is gridded at 400 m cell size using all open-file gravity data within the State (GSWA, 2013). Gravity data points were extracted from the Australian National Gravity Database (ANGD; Wynn and Bacchin, 2009). Coverage in the Sandstone region is good with data points on an approximate 2.5 km grid over the whole area, and closer spacing of 400 m along the 10GA-YU2 seismic line. Another traverse with 1 km spacing crosses the southern margin of the greenstone (Fig. 5).

The Bouguer anomaly map (Fig. 5) shows the higher density of the greenstone belts surrounded by lower density granites. The Sandstone greenstones are well defined by a gravity high that closely matches the spatial extent of outcrop. Other greenstones are interpreted to have subsurface extents as indicated by the high Bouguer anomalies extending into areas otherwise mapped as granite at the surface.

Total magnetic intensity

Aeromagnetic total magnetic intensity (TMI) data were taken from the WA State merge of open-file and government surveys over the area, which was gridded at

80 m (GSWA, 2014a). Over the Sandstone greenstone belt, there have been company surveys of 100 m line spacing or better. Additionally, there are company and multiclient survey data of 200 m spacing and government surveys at 400 m spacing. Since the TMI response shows a lot of detail, a smoothing filter (9 x 9 symmetric least squares) was applied to the observed TMI for the purpose of 3D modelling. This removed the short wavelength anomalies which add unnecessary detail that cannot be adequately modelled.

The TMI map (Fig. 4) shows the strong positive magnetic signature caused by the BIF, which defines the shape of the greenstone and reveals some internal structure (cf. Fig. 2). It is suspected that there is a component of remanent magnetism seen in the TMI anomalies in both the Sandstone and Gum Creek greenstones; however, no measurements have been made to quantify this. Field specimens of the granites around the Agnew greenstones (Williams, 2009) were measured and showed a significant remanent magnetic component. By analogy, this may explain some high susceptibilities seen in the inverted models developed in this project.

The TMI map shows the foliations within the granites, especially within the shear zones, and the Proterozoic mafic dykes clearly cutting across the Archean stratigraphy.

Seismic reflection

A deep seismic reflection survey was conducted by Terrex Seismic in 2010 and preliminary results were published in Wyche et al. (2014). Of the three lines, only 10GA- YU2 crosses the study area, traversing the northern part of the Sandstone greenstone belt. The source was three IVI Hemi-60 vibrators shooting 2–3 sweeps at spacing of 80 m or 40 m over the greenstone areas. Receiver groups were every 40 m along a 12 km spread. Processed images were produced down to 20 s two-way travel time (TWTT; Costelloe and Jones, 2014).

Interpretation of the reflection profiles in the vicinity of the Sandstone area by Zibra et al. (2014; Fig. 6) was the most important factor in deducing the subsurface structure for the 3D model.

The seismic interpretation shows the shape of the greenstones and the later granites in profile (Fig. 6). It also shows the Youanmi Shear Zone truncating the Edale Shear Zone and continuing down to the lower crust. The upper crust is relatively nonreflective compared to the middle crust, which exhibits a strong seismic fabric. The unexposed middle crust layer was named by Korsch et al. (2014) the Yarraquin Seismic Province. Underlying this is a thin, reflective lower crust, below which a distinct change of character indicating the mantle lies at about 33 km depth (Korsch et al., 2014).

Strong reflectors that cut across the stratigraphy are thought to be caused by Proterozoic sills, several of which cross the Sandstone area (Ivanic et al., 2014b).


7

Figure 4. Total magnetic intensity (TMI) over the study area: a) TMI image; b) line drawing of geology (from Fig. 2) overlain

on TMI image

Depth to Moho

The depth to the Moho was taken from Korsch et al. (2014) along the seismic line, but across the rest of the region, it was taken from the Australian Seismic Reference Model (Kennett et al., 2011; Kennett and Salmon, 2012; Kennett, 2014). Along the entire seismic line the mantle boundary is sharp and flat, and this was assumed to be the case for the rest of the model space.

Physical properties

Starting densities and magnetic susceptibilities for the initial 2D forward modelling of the geological formations were taken from hand specimens examined by Williams

(2009) who measured dozens of samples from the Agnew–Wiluna greenstone belt located 120 km to the east.

Physical properties were also taken from gravity and magnetic models created along the seismic profile (Gessner et al., 2014), who had used average values from Rudnick and Fountain (1995).

Since each modelled unit is typically a collection of different rock types, and full understanding of population statistics was not well known, average densities for the assemblages were used in the initial models. During the geophysical inversions, the starting mean density distribution was revised to reflect a more realistic aggregate for the modelled unit. Tables 1 and 2 show the densities and susceptibilities used through the various stages of the modelling.

Murdie et al.

8

Figure 5. Bouguer gravity anomaly over the study area: a) high-resolution Bouguer anomaly with gravity measurement points,

including the location of the gravity profile of Shevchenko (2005); b) line drawing of geology (from Fig. 2) overlain

on Bouguer anomaly image


9

Fig

ure

6.

Inte

rpre

ted

secti

on

s o

f th

e 1

0G

A-Y

U2 s

eis

mic

tra

vers

e l

ine (

Ivan

ic e

t al., 2014)

cro

ssin

g t

he s

tud

y a

rea s

ho

win

g:

a)

mig

rate

d s

eis

mic

secti

on

; b

) in

terp

rete

d g

eo

log

ical

cro

ss-

secti

on

. Th

e p

art

used

in

th

e S

an

dsto

ne m

od

el is

ou

tlin

ed

in

th

e b

ox

Murdie et al.

10

Unit Williams (2009)Gessner et al.

(2014)Final forward 3D

modelBasement inversion (GOCAD)

Heterogeneous inversion (GOCAD)

Proterozoic dyke 2.8 2.76 2.76

Bald Rock Supersuite 2.53 ± 0.09 2.72 2.68 ± 0.020

D2 granite 2.53 ± 0.09 2.71 2.61 ± 0.077

Metasedimentary unit 2.79 ± 0.16 2.81 2.86

Seven units

modelled as a

homogeneous

unit of Sandstone

greenstones

2.86

Seven units

modelled as a

homogeneous

unit of Sandstone

greenstones

Ultramafic unit 2.87 ± 0.39 3.1 2.77

Mafic unit with BIF 2.79 ± 0.23 3.03 3.01

Banded iron-formation (BIF) 2.91

Massive mafic unit 2.79 ± 0.23 3.03 3.0

Tremolite schist unit 2.79 ± 0.16 2.9

Basal mafic unit 2.79 ± 0.23 3.03 2.96

Unaly Hill greenstone 2.92 2.86 2.86

Gum Creek greenstone 2.95 2.86 2.86

Barrambie Igneous Complex 2.97 2.86 2.86

North Cook Well greenstone 2.93 2.86 2.86

Foliated granite 2.53 ± 0.09 2.68 2.53 2.62 ± 0.079

Southern Cross granite 2.53 ± 0.09 2.73 2.7 2.53 2.57 ± 0.092

Murchison granite 2.53 ± 0.09 2.72 2.68 2.53 2.59 ± 0.086

Yarraquin Seismic Province 2.85 2.85

Lower crust 2.9 2.9

Mantle 3.3 3.3

Table 1. Table of densities, in gcm-3, used throughout the modelling process. Model units with a heterogeneous density

distribution are given as mean ± one standard deviation

Magnetotelluric surveys

A magnetotelluric (MT) survey was previously conducted along the Youanmi seismic line (Milligan et al., 2014). However, this information was not used as the resolution of the MT data is too coarse to significantly contribute to a 3D crustal model in this area.

Software

GM-SYS

GM-SYS was used in the modelling of gravity profiles. It is a specific application for potential-field forward modelling in 2.5D run from within the Geosoft Oasis Montaj software suite. It allows the extraction of potential-field data along a profile from gridded data. The user can then build a 2.5D model from polygons assigning appropriate physical properties to each polygon of the model. The model is adjusted either by changing the shape of the polygons or the physical property value so that the computed response matches the observed field more closely.

GeoModeller

The software used to initially build the 3D model was GeoModeller 2013, an implicit code, developed by Intrepid Geophysics, Melbourne and BRGM, France.

An implicit code uses a continuous mathematical representation to satisfy a set of x, y and z values across a volume. Hence, addition of more data refines the function that defines a surface.

GeoModeller has been designed for building complex, steady-state, 3D geological models. It constrains models with primary geological data such as contacts, orientation data (such as dip and strike), and drillhole intercepts described in 3D space, and interpolates between points using a ‘potential-field method’ (Lajaunie et al., 1997). Stratigraphic and cross-cutting relationships are honoured by use of logic matrices (Calcagno et al., 2008). Faults and other structural and stratigraphic relationships are also honoured. It allows forward 2D and 3D modelling and inverse modelling in 3D of the geology with direct reference to geophysical survey data including gravity, magnetic, seismic and thermal data. GeoModeller was used in the initial stages of model building as it allowed fast revision of the model when updating and modifying input.

GOCAD

The GOCAD v. 2009.4 software by Paradigm is an explicit subsurface modelling package with many advanced functionalities primarily targeted for the oil and gas industry. An explicit function only generates a surface where there is data available and models to each point. The result is that it does not extrapolate between or beyond the data. It enables wireframe and block modelling with


11

Table 2. Table of magnetic susceptibilities in SI used in the modelling process. Model units with a multi-modal distribution

are given as the mean values of the peak susceptibility and percentage of cells with this value peak. Model units with

heterogeneous distributions are given as a mean ± one standard deviation

Unit Williams (2009)Final forward 3D model

Homogeneous property optimization

Heterogeneous inversion (GeoModeller)

Heterogeneous inversion (GOCAD)

Proterozoic dyke 0.01 0.011 0/80

0.07/5

0.095/15

0.013

Bald Rock Supersuite 0.004 ± 0.003 0 0.003 0.0012/80

0.0045/20

0.003 ± 0.005

D2 granite 0.004 ± 0.003 1 0.007 0.0012/50

0.01/50

0.025 ± 0.011

Metasedimentary unit 0.001 ± 0.0006 0.0005 0.018 0.018 0.005

Ultramafic unit 0.024 ± 0.022 0.1 0.08 0.06/50

0.1/50

0.035

Mafic unit with BIF 0.0015 ± 0.0016 0.005 0.05 0.01 0.016 ± 0.021

Banded iron-formation (BIF) 0.3 0.18 0.135/90

0.3/10

0.111 ± 0.067

Massive mafic unit 0.0015 ± 0.0016 0.00001 0.025 0.005/80

0.085/20

0

Tremolite schist unit 0.002 0.022 0.001/85

0.02/15

0

Basal mafic unit 0.0015 ± 0.0016 0.00001 0.008 0.0001/80

0.016/20

0

Unaly Hill greenstone 0.005 0.01 0.001/70

0.02/30

0.01

Gum Creek greenstone 0.04 0.063 0.001/80

0.15/20

0.063 ± 0.152

Barrambie Igneous Complex 0.02 0.05 0.007/90

0.15/10

0.06

North Cook Well greenstone 0.04 0.039 0.007/40

0.09/60

0.04

Foliated granite 0.004 ± 0.003 0.00001 0.011 0.001/70

0.002/5

0.007/25

0.012 ± 0.032

Southern Cross granite 0.004 ± 0.003 0.001 0 0.016 ± 0.012

Murchison granite 0.004 ± 0.003 0.00005 0 0

associated physical property modelling. Here it was used with the Mira Geoscience Mining Suite add-in and the VPmg potential-field modelling and inversion software by Fullagar Geophysics.

Input to the modelThe workflow used to build the model is shown in Figure 7. The input data are described in detail below, followed by a description of each of the modelling steps.

Project coordinate system

The project area lies within Sandstone (Chen and Painter, 2005) and Atley (Chen, 2003) between longitudes 119°00’ and 119°30’E and latitudes 27°30 and 28°30’S.

All modelling packages used require the data to be in square grid coordinates so the working coordinate system was translated from geographical latitude and

longitude into metres in the Universal Transverse Mercator Projection using the Geocentric Datum of Australia 1994 (GDA94) system, Map Grid of Australia Zone 50.

The project area was 51 km east–west by 87 km north–south and 51 km deep. This allowed for 1 km elevation and 50 km depth, relative to AHD (Table 3).

Topography

Topographic data for the area were extracted from the Shuttle Radar Topography Mission (SRTM; Jarvis, 2008; Fig. 8). The original data were supplied with a resolution of 90 m grid cells, but for optimization purposes, GeoModeller extracted only every third point. This gave coverage of 214 cells in the east–west direction and 364 cells in the north–south direction, with a cell size of 240 by 240 m. Since the topography of the area is largely flat to undulating, this downsampling was reasonable and freed up computer resources from interpolations of the topography.

Murdie et al.

12

Figure 7. Flow diagram summarizing the model building process, and showing input data (rhomboids at top), process steps

(rectangles), and movement along the diagram governed by review steps (diamonds)


13

Surface geology

The geology was digitized into GeoModeller from Sandstone (Chen and Painter, 2005) and Atley (Chen, 2003) using the surface and solid geology maps (GSWA, 2008). Although good geological exposure is found locally in the west, south and east of the Sandstone greenstone belt, the cover overall is extensive and the centre of the area is covered by lateritic colluvium and sheetwash (Chen, 2003; Fig. 9). In the generation of the interpreted bedrock geology, Chen (2003) relied extensively on the potential-field data to interpolate between outcrop information.

Min (m) Max (m) Length (km)

Easting 696000 747000 51

Northing 6863000 6950000 87

Depth -50000 1000 51

NOTE: Datum for depth (z-axis of the model) is AHD

Table 3. Dimensions of the 3D model using the UTM Zone 50

coordinate system

Figure 8. Topography of the model area from SRTM

Figure 9. Map showing the areas of outcrop over the

Sandstone greenstone belt. In areas with no colour,

there is sufficient cover to obscure the bedrock.

Murdie et al.

14

Figure 10. Locations in 3D space of cross-sections from the maps of Chen (2003) and Chen and Painter (2005), and

the interpretation of the seismic line 10GA-YU2, used in the generation of the model. A selection of other

cross-sections, generated during model building are shown with north–south oriented sections in blue,

east–west oriented sections in pink, and diagonal sections in purple.

Due to the large model size and the resolution limit of the modelling package, very thin beds such as individual banded iron-formation units, mapped as lines on the 1:100 000-scale maps or discontinuous lenses within units less than 200 m thick, could not be imaged in a meaningful way. The divisions were based on a physical property contrast which would be useful in the geophysical modelling. Hence, the stratigraphic column was modified to reflect the major stratigraphic units (Fig. 3). The stratigraphic names used in this model are based loosely on the geological description as they include more than one geological unit.

The scale of the model meant it could not effectively utilize the detail seen in a borehole; therefore, no data from drillholes were used explicitly in the input data. However, the many shallow drillholes in the area informed

elements of Chen’s (2005) subcrop geology interpretation, such as the presence of ultramafic units in the centre of the syncline.

Average or representative structural surface measurements (dips) were selected from WAROX. The full resolution dataset did not provide a meaningful input to the model building, as these data represent smaller scale structures than can be resolved by the model. However, the full structural dataset led to understanding the overall architecture of the region and was incorporated into the model implicitly.

Other mafic belts within the area, such as the Unaly Hill greenstone belt, have been modelled in this study as homogeneous bodies with no stratigraphy applied as they are not present in the model with enough outcrop to warrant detailed internal modelling.


15

Cross-sections

Each of the 1:100 000 GSWA geological maps includes one or more representative cross-sections. Although these cross-sections are schematic, they were used as a starting point for generating the profiles of the 3D block (Fig. 10), including a profile across the main syncline on Atley and across the northern syncline on Sandstone.

The only depth-constrained information used in building the model was the interpreted section of the 10GA-YU2 seismic line (Ivanic et al., 2014a). The seismic line ran between the Atley and Sandstone cross-sections and was almost coincident with the latter at the eastern end (Fig. 2).

Other cross-sections were added as necessary to better constrain the model (Fig. 10).

2D potential-field forward

modelling

A valid geological model must honour all geological and geophysical observations. There is uncertainty in the interpretation of the seismic data and the inverse theory of potential-field modelling (whereby a smaller, denser or more magnetically susceptible body may have a similar potential field to a larger, less dense or less magnetically susceptible body) means that gravity and magnetic models are non-unique with respect to the geological entities that they may indicate. However, they provide some constraints that can be used as a starting point for depths and shapes of bodies. The more representative the starting model is of nature, the less we need to rely upon inherently ambiguous geophysical data for the solution.

Forward modelling of gravity data in 2D using GM-SYS was undertaken to provide depth constraints on the boundaries in key sections before a 3D model was constructed.

The first profile to be modelled was the seismic line, using the densities as per the gravity models of the whole Youanmi seismic line. The major features of the observed gravity profile were fitted by the profile of calculated gravity response of the model and the RMS difference between the two profiles was minimized (Fig. 11a). Using these densities and hand specimen data of Williams (2009), the other key sections were then modelled and geometries adjusted accordingly (Fig. 11b,c). As shown in the later section on 3D modelling, these 2D models were iteratively adjusted through inversions before the final model was achieved.

Magnetic forward modelling was also performed in GM-SYS, with input from initial estimates of the magnetic susceptibilities taken from Williams (2009). This was very crude modelling as there are large magnetic sources just offline and possible remanence effects, both of which would have an effect on the magnetic profile. The magnetic profile for the seismic and Atley sections (Fig. 11a,c) show the magnetic peaks produced by the BIF. Magnetic peaks in the granites were not

modelled. The granites show inhomogeneous magnetic characteristics with areas of strong susceptibility as seen in the TMI image. However, in the forward modelling, the polygons representing the granites were given a uniform susceptibility. Also, the internal structure of the granites was not a priority in the modelling.

After each profile was modelled to a minimum RMS misfit while retaining geological feasibility, each shape was input and assigned to the appropriate stratigraphic unit, and then added to the corresponding profile within the GeoModeller model.

Modelling in 3DGeoModeller, which was used for the initial stages of 3D model construction, requires certain information as determined by the implicit algorithm: specifically, a stratigraphy, and corresponding orientation and contact observations for each stratigraphic unit (Calcagno et al., 2008). Orientation and contact observations define the geometry of each unit, whereas the model stratigraphy defines where each unit is relative to each other. Model stratigraphy can be difficult to define in complex regions such as the Sandstone greenstone belt, and therefore requires a more detailed examination.

Model stratigraphy

The stratigraphic sequence, known in GeoModeller as the ‘stratigraphic pile’, determines the relative age of modelled lithological units. This ensures that the units always remain in order, are maintained ‘the right way up’, and the chosen stratigraphy is honoured. Units are grouped into series if they are conformable and share structural orientation data. A series can ‘erode’ or ‘onlap’ with the series below or above. In each series within a pile, the thickness and continuity of each bed is approximately maintained. This allows beds to be extrapolated into regions where there are no data. The Sandstone stratigraphic pile used in this model is shown in Figure 3. Within the greenstone package, the units have been represented by a basal mafic unit, a schist, a massive mafic unit, a BIF and an upper mafic unit with thin BIFs, and an upper ultramafic unit with overlying metasedimentary rocks.

Whereas GeoModeller assumes that the youngest rock lies on the top of the pile in this typical Archean setting, in fact the older greenstones overlie younger granites with intrusive and faulted contacts. Therefore, for model building purposes, a break from convention was used, in that the sheared, foliated granite and granite provinces (Murchison and Southern Cross ‘crustal’ units to the west and east of the Youanmi Shear Zone, respectively) were placed stratigraphically lower than the greenstones and the D2 granite unit (Fig. 3). The D2 granite unit is contained within F2 folds of the D2 event within the greenstone sequence, implying emplacement before the Youanmi and Edale Shear Zones were initiated during east–west shortening of the D3 event (Chen, 2005).

Murdie et al.

16

a)


17

Figure 11a. (opposite) Forward 2.5D gravity model of the

segment of the 10GA-YU2 seismic line crossing

the study area. Geometries are limited by the

interpretation of the seismic lines and initial

densities used the hand specimen measurements

from Williams (2009): i) Original interpretation

from seismic line 10GA-YU2; ii) lithology

section as defined in 2D by the gravity forward

model; iii) observed and calculated gravity

anomaly profile with error line; iv) profile of

density per lithology; v) observed and calculated

magnetic anomaly profile with error line;

vi) profile of mean magnetic susceptibility per

lithology.

Figure 11b. (above) Forward 2.5D gravity model of the

user-defined north–south section at 726000E.

Geometries are constrained only where this line

intersects the seismic line 10GA-YU2. Similar

densities were used in all modelled profiles:

i) Observed and calculated gravity anomaly

profile with error line; ii) lithology section

as defined by the gravity forward model;

iii) profile of density per lithology.

b)

Murdie et al.

18

Figure 11c. Forward 2.5D gravity and magnetic models of the ATLEY cross-section: i) Cross-section from

Chen (2003); ii) lithology section as defined by the gravity forward model; iii) observed

and calculated gravity anomaly profile with error line; iv) profile of density per lithology;

v) observed and calculated magnetic anomaly profile with error line; vi) profile of mean

magnetic susceptibility per lithology.

c)


19

The D2 granite is probably similar to the other Southern Cross granites, but has been assigned a different unit here for modelling purposes, as it sits on the other side of the Edale fault from the Southern Cross granite unit. GeoModeller can correctly displace faulted series that have an offset quantified by a horizon on either side of the fault. It was therefore relatively simple to include the smaller faults within the greenstone series. However, using a fault to completely remove an entire formation in part of the model is more difficult. In this case, the major faults, the Youanmi and Edale Shear Zones, although present in the model, do not displace any formation. Hence, it was necessary to place the greenstones stratigraphically higher than the granite components.

The Gum Creek, Unaly Hill and North Cook Well greenstone belts, and Barrambie Complex, were modelled as homogeneous bodies rather than with detailed stratigraphy. Only the Gum Creek greenstone belt has continuous mapped stratigraphy. The others are present as very small components within the major shear zones, with outcrops whose internal features were below the resolution of the model. As little information on their density was available, an average density for all greenstone units of the Sandstone greenstone belt was used. Subsequent optimizations and inversions were used to determine their extents.

Building the model

The geometries of the sections modelled in GM-SYS were digitized onto the appropriate sections within GeoModeller. Following the depth guides of these modelled sections, other sections were generated to help guide GeoModeller in the regions between the modelled sections (Fig. 10). The geological map was included as a horizontal section using the interpretation of the observed and inferred geology. Structural measurements, such as dip, strike and plunge, were also included (Fig. 12). GeoModeller’s implicit algorithm calculates the contact and its orientation together with the stratigraphic and fault relationship to generate a model.

Forward modelling of the block

model

GM-SYS only models standalone profiles. Profiles from the same observed potential-field measurements have no link to each other. This can lead to inconsistencies arising from the inverse ambiguity between profiles such densities, depths and thicknesses of bodies. Hence, forward gravity modelling of the 3D volume can pick out discrepancies between base levels in the different profiles.

To forward model a volume, the volume can be discretized into many rectangular prisms, known as voxels within

GeoModeller. Each voxel is assigned a density and susceptibility value, from which the potential-field response of the entire 3D model can be calculated. Once a 3D block is generated within GeoModeller, the application can calculate the potential-field response of the entire 3D model. This enables the modeller to rebuild the model to remove any regional differences between the individual profiles. Figure 13 shows the development of the model using the Bouguer anomaly as the main constraint.

The total gravity or TMI response across the whole (or part) of the model was calculated, and maps of the observed survey data and calculated response from the model were produced. The difference between the observed and calculated fields is the residual, and this is higher where larger misfits between model and data are observed. The model geometry or properties were manually adjusted, the model recalculated, and the residual displayed in order to find get a better fit to the observed data and a better model.

Property optimization

Once the modelled Bouguer anomaly closely resembled the observed Bouguer anomaly using the initial densities and adjusting the geometry, an adjustment for the bulk densities of the bodies was made. GeoModeller offers two levels of property optimization for geophysical properties: homogeneous and heterogeneous. The homogeneous method looks at the average bulk property across the whole formation and calculates a mean value with associated standard deviation. A heterogeneous optimization distributes higher and lower densities across voxels in the body within a certain density range to so that the final model has a density structure within each heterogeneous body. In each optimization, each cell property can vary according to a mean distribution defined by a mean and standard deviation. The optimization calculation uses a least mean-square method to invert the model to obtain the best average property per geological formation, and thus to achieve the optimal fit to the observed data.

The homogeneous property-optimized values suggested by the density optimization should be similar to, but slightly different from, the starting densities. If the starting densities are too far from the true value, the optimization will not be able to reach a sensible optimization point and this would be cause to revisit fundamental elements of the starting model. Also, since the units within GeoModeller may contain several rock types, it is unlikely that the units would model with the exact density of a hand specimen.

Once a reasonable geometry and density model was obtained, the model was transported into GOCAD for further refinements.

Murdie et al.

20

Figure 12. Construction points (constraints) from each forward-modelled section for four geological units

located in 3D space before generation of the model. Each point has an associated structural

orientation: an azimuth (pointed by the top of the cone) and a dip angle (indicated by the

orientation of the base of the cone).

3D modelling and inversion in

GOCADGOCAD builds models explicitly, which means each surface has to be built manually, and rebuilt each time the geometries are changed. Hence, for initial modelling when the ideas are still evolving significantly, it is easier to build the model in GeoModeller because recalculating the model is a trivial task in the implicit scheme. Once the basic framework, with geometries and densities that fit well with the observed geophysical data, had been obtained in GeoModeller, the VPmg inversion routines within GOCAD were employed to refine the model and generate the final model.

Gravity modelling

The GeoModeller model surfaces were exported and the model rebuilt in GOCAD. The density difference between the granites and the greenstones was used to determine the final shapes of the greenstone. A basement inversion routine within the VPmg suite of inversions was used. This aims to find the depth to the interface between a low-density cover and a high-density basement. However, in this case, it was used upside down as the denser material was above rather than below (Fig.14). This gave the outline of each of the mafic areas, which were retained during the magnetic modelling. A density difference of 0.33 gcm-3 was used between the granites and the greenstones. This provided the best fit for the shape of


21

Figure 13. Development of the model using the Bouguer anomaly as a constraint: a) the observed Bouguer anomaly;

b) calculated response of an initial model and residual below; c) calculated response of the final forward

model produced in GeoModeller and residual below; d) calculated response of the homogeneous

property optimization run in GeoModeller and residual below; e) calculated response of the GOCAD

basement inversion and residual below; f) final heterogeneous gravity inversion and residual below.

Note the histogram-equalized scale bar.

the greenstone belt when compared with the seismic line. Lower density differences resulted in excessive depth of the greenstones as a greater mass of greenstone is then required to achieve the observed gravity anomaly, and higher density difference values resulted in the greenstones being too shallow. Unexposed greenstones also showed in the inversion as depressions in the model surface, which was taken as their depth extent. However, they were modelled with a thin granite cover to account for their lack of exposure.

From this inversion, the gravity response of the model showed that the model adequately represented the bulk geometry of the greenstone belt. Hence, the external geometry of the Sandstone greenstone belt was not altered in subsequent modelling.

As the bounding geometry of the granites was defined by the geometry inversion, a further heterogeneous property inversion was carried out on the granites to pick out areas of higher and lower density (Fig. 15). Due to the nature of the inversion routine, this produced vertical prisms of regions with differing densities within the granites. One of these is an area of higher density within the Murchison granites, which could be correlated with a post-tectonic granite, the Bald Rock Supersuite, outlined in the seismic section.

The final modelled gravity and misfit of the final density-inverted model are given in Figure 13. The fit of the gravity data are very good as could be expected with a heterogeneous inversion. The biggest errors lie at the boundary of the greenstone. However, these are only a few mgals or less than 5% of the original dynamic range.

Murdie et al.

22

Figure 14. Results of the basement inversion showing the

depth of the Sandstone, Unaly Hill and Gum Creek

mafic areas from three different angles: a) view

towards northeast and down; b) view towards

northwest and down; c) view towards northwest

and from underneath. The Barrambie and North

Cook Well mafic areas also show as depressions

in the model. The depths were taken as indications

of the depth to the bottom of the mafic rocks.

A thin cover of granite was added on top of the

mafic body as there is no significant outcrop of

these rocks in the model area.

Magnetic modelling

As the gravity data were too coarse to image the formations inside the greenstone belt, these geometries were further refined using magnetic inversions as the magnetic data had much higher resolution.

The original pixel resolution of the TMI grid was 80 m, so the high magnetic response of the BIF was used to delineate the internal structure of the greenstones. A smoothing filter (9 x 9 symmetric least squares) was applied to the observed TMI for the purpose of 3D modelling (Fig. 16). This removed the short wavelength anomalies that add unnecessary detail and cannot be adequately modelled. It was not necessary to use a reduced-to-pole magnetic image as the VPmg code makes a correction for magnetic inclination and declination within the inversion.

Although the structural relationships between the different rock types in the greenstone are not known, it was assumed that:

1) the basic shape of the greenstone was a refolded syncline

2) the lower units were approximately parallel

3) the ul t ramafic uni t was only shal low and unconformable (Chen, 2005).

In summary, the surface mapping and the shape of the BIF were used to estimate the internal structure of the basic greenstone volume. Measurements of magnetic susceptibility from hand specimens (Williams, 2009) were used as initial data for the susceptibility modelling within GeoModeller. Forward models and some inversions were run within GeoModeller even though the final model was produced in GOCAD. GeoModeller allows for a bimodal or trimodal susceptibility distribution within a mode unit. Table 2 shows the range of values investigated for various stages on the modelling process.

Using the outline of the greenstones as obtained by the gravity inversions, a homogeneous susceptibility inversion was run to get a better approximation of the bulk formation susceptibility. Following that, a geometry inversion on the units inside the greenstone was run using these susceptibilities. As this made little difference to the final misfit, a heterogeneous property inversion was performed using the results from the homogeneous susceptibility inversion as a starting model. This clearly identified areas of high susceptibility. In some cases, this is readily explained by natural variation within the unit. In other cases, it was used as an indication that the geometry of the BIF could be better modelled. The model was adjusted where the susceptibilities indicated a likely presence of BIF and the inversion sequence run again. As the granites also showed variations within the TMI map, these were included in the heterogeneous property inversions. The only units that were left as homogeneous units were those which had consistently showed zero susceptibility in all other inversions. The final susceptibility distribution is shown in Figure 17.


23

Figure 15. Results of the heterogeneous inversion showing the variations in density

in the granites

Three faults, mapped by Chen (2003), were included within the Sandstone greenstone units in the model to achieve an approximation to the offsets seen in the magnetic response (Fig. 18a). Other faults mapped by Chen (2003) were omitted as their inclusion would add another level of complexity which was considered unnecessary for this model.

Two faults were added within the Southern Cross granite. These correspond to mapped faults which are part of the Edale Shear Zone and bound the extent of the foliated granite.

Implications of the modelling

results for the regional

structureThe Youanmi Shear Zone is considered to be a domain-bounding structure between the Southern Cross and the Murchison Domains, which share a relatively similar

geological history (Cassidy et al., 2006, and references therein). Our modelling confirms the overall similarity of these domains in the vicinity of the Sandstone greenstone belt. Although the Youanmi Shear Zone is a major structure with considerable strike length, there is no evidence for a significant variation of rock properties across it. However, the shear zone is evident in the fabric seen in the TMI response in the granites, although those fabrics are beyond the resolution of this model.

The density modelling implies the greenstone belt is slightly wider than the seismic section would indicate. This results in a slightly steeper dip for the Youanmi Shear Zone than modelled in the seismic interpretation (Zibra et al., 2014). However, the resolution of the interpretation of the seismic line would allow this interpretation.

Shevchenko (2005) used a value of 2.85 gcm-3 for greenstone density, against an unspecified granite density, to determine the depth of the greenstone along a single northwest–southeast traverse. An average depth of 4 km to the base of the greenstone was obtained. However, Shevchenko’s (2005) traverse, although following a line

Murdie et al.

24

Figure 16. Magnetic anomaly maps of: a) observed TMI; b) smoothed TMI;

c) calculated TMI of an initial model; d) calculated TMI the final forward

model; e) calculated response of the homogeneous property inversion

in GeoModeller; f) calculated response of the heterogeneous property

inversion in GeoModeller; g) calculated response of the heterogeneous

property inversion in GOCAD; h) residual TMI observed and calculated

response of the heterogeneous property inversion in GOCAD.

Note different scales for the observed and modelled TMI versus the residual

TMI on the histogram equalized scale bar.

of high-density gravity measurement points, missed the thickest part of the greenstone sequence, as it crossed the western limb (Fig. 5). The inversion carried out in this study used a slightly higher density of 2.86 gcm-3 (Gessner et al., 2014) for the mafic component, and a value of 2.53 gcm-3 for the granites. A maximum depth of greenstone of 6.5 km was obtained, mainly along the eastern limb of the belt (Fig. 18a,b). Along the western side, a maximum depth of 3.8 km was obtained, which is similar to the modelling of Shevchenko.

The regions assigned to the Murchison and Southern Cross Domains extend to considerable depth and make up most of the upper crust. These granitic domains were assigned a background density of 2.5 gcm-3 for the basement inversion. The heterogeneous inversion altered values to

within 0.05 gcm-3. These are within the ranges given by Emmerson (1990) and Schön (1998), but also as used by Gessner et al. (2014) for upper crustal densities within the Murchison region.

It is difficult to differentiate between the granites on either side of the Youanmi Shear Zone on density, but there appears to be an area of higher density granite on the Murchison side of the shear zone, which in the seismic section is termed post-tectonic granite and belongs geochemically to the Bald Rock Supersuite (Van Kranendonk, 2013). This was defined by a positive density anomaly of 0.06 gcm-3. However, there is negligible susceptibility difference between this rock unit and other granites of either domain.


25

Figure 17. Results of heterogeneous magnetic susceptibility inversions showing the

regions of high and low susceptibility.

The D2 granite unit has a higher density than the others. This may indicate either that parts of the higher density Sandstone greenstone belt are incorporated within the granite, or that this unit has a more mafic, for example granodioritic, component. High-density areas in the sheared granite around the Unaly Hill greenstone belt and Barrambie Igneous Complex may indicate that there are small parts of these intrusions entrained in the sheared granite, although these would be too small to be modelled. Shevchenko (2005) also modelled the Unaly Hill greenstone belt as ‘blind greenstone bodies’ to the west of the Sandstone greenstone belt.

Our model also confirms the overall form of the previously interpreted fold geometry of the Sandstone greenstone belt (Fig. 18a,b). This geometry is reproduced in the gravity and magnetic forward models (Fig. 11a–c). Initial forward modelling followed the outline of the greenstone as indicated by the seismic interpretation. However, later inversions showed that the greenstone had probably been eroded by the D2 granite, as inferred by Chen (2003) in the Atley cross-section.

The internal geometry of the greenstone belt was initially tested in forward magnetic models (Fig. 11a–c). These were later revised after consultation with the geologist who had mapped the area, and after applying magnetic

inversions. The TMI heterogeneous inversion (Fig. 17) shows a strong BIF component around the northern side of the syncline, dipping steeply to the south. Other high susceptibility areas on the southern side of the syncline appear to dip to the north. The presence of a semicontinuous BIF on the southern margin supports the interpretation that this is a synclinal structure. (This is best seen in the model by viewing the volumes of the units rather than the bounding surfaces). There are also well-developed BIF components, which lie in the centre of the syncline, with orientations at angles to the main syncline. These are probably part of the mafic unit with BIF (Fig. 3), or may be stringers of the upper ultramafic unit. The lower units (basal mafic unit, tremolite schist, and massive mafic unit) are not predicted by our model to appear in the south.

There are other areas of high susceptibility within the D2 granite unit. These may be detached portions of the Sandstone greenstones entrained deeper in the granite. Alternatively, Williams (2009) measured susceptibilities and remanence of the granites around the Agnew area and found that many samples had significant susceptibility and remanent magnetization. This may be another explanation for the high susceptibilities in this area. The Murchison and Southern Cross granite units also show heterogeneous susceptibilities as expected from the variations within the original TMI image.

Murdie et al.

26

a)

Figure 18. Surfaces and volumes in the final 3D model: a) surfaces in the model, including bottoms

of formations and faulted surfaces; b) regions filled with geological formations


27

b)

Figure 18. continued

Murdie et al.

28

The North Cook Well and Unaly greenstone belts, and the Barrambie Igneous Complex, are too small and dismembered for any internal structure to be delineated. However, the Gum Creek greenstones were delineated by a gravity contrast and the susceptible units were highlighted by the TMI inversion. Although the average susceptibility of the Gum Creek greenstones is low, there is a subvertical layer with a very high susceptibility.

An image of the surfaces in the model is given in Figure 18a and an image of the model regions in Figure 18b.

Comments on the process of

model buildingThe basic workflow for the modelling is given in Figure 7, and this is typical of the 3D modelling approach (Jachens, 2001; Rawling et al., 2011; Russell, 2011). However, the exact process for any model depends on:

• available data

• data quality

• data resolution

• geological setting

• the questions being asked.

Many models have ample data, such as from drillholes, to constrain them. However, the input data for the model in this study were restricted to surface mapping, high-resolution potential-field data (gravity and magnetic), and a seismic line.

A basement inversion is typically used in the case of a low-density cover or basin overlying a denser basement. The modelling of relatively dense Archean greenstones surrounded by lower density granites is a special situation from the point of inversion modelling. However, it is possible to use the same technique to model this arrangement. Potential-field interpretations are inherently non-unique and additional constraints are required to reduce the number of possible solutions to observed data. The seismic line provided a constraint along one profile. It was used to obtain an appropriate bulk density for the greenstone units during the basement inversion routine, so that the depth of the modelled greenstone matched that seen along the seismic line. However, there are also ambiguities in the interpretation of the seismic line and lack of detail at the scale of this model, especially within the greenstone belt. Other constraints came from the surface structural and lithological observations and from physical property measurements.

Although the internal structure of the greenstones was initially modelled in the 2D forward modelling of the gravity data, it was decided that the resolution of the original data (an average point spacing of 2.5 km), the

relatively small density differences of the greenstone units, and the abundance of thin and mixed units within the greenstones precluded the running of density of inversions across the greenstone for better internal structure. It was more appropriate to take advantage of the higher resolution TMI data (80 m grid cells from 200 m or better flight lines) and the strong magnetic response of the BIF, compared to the relatively weak response of the other mafic units, to resolve beds that are 500 m thick or less. It is possible that future drilling and more detailed mapping will provide a better understanding of relationships within the greenstones.

A large part of this study was devoted to preliminary modelling before running the final inversions and property optimizations. Some of this effort, such as modelling the internal structure of the greenstones with gravity and forward magnetic modelling, was superseded by later processes; this is seen in the difference between the initial 2D forward models (Fig. 11a–c) and the final product (Figs 16 and 18). Changes in the approach were made as the capabilities of each software package became apparent. However, each step explored different aspects of the model and helped to distil ideas into the final model, as well as improving familiarity with the functionality of each software package.

The choice of software used was our personal preference and many functions could have been undertaken in several of the packages. GM-SYS was preferred for the forward modelling as it has a dynamic update on the calculated potential-field response as properties or geometries are changed, making it quick and easy to produce 2D forward models. GeoModeller is ideal for building models where there are few data and it allows easy accommodation of new ideas as data are either removed or added, or additional data become available. Also, the forward potential-field calculations in 3D are useful for bringing the 2D profiles together. However, the potential-field inversion routines of the VPmg suite within GOCAD were preferred for ease of use over those offered by GeoModeller. These inversions allow development of the model away from profile areas. Also, GOCAD has the best options for presentation of the models including a free viewer for general viewing. Other commercial packages achieve very similar functions and all packages are continuously being upgraded. Personal preference may be the best guide in many cases.

All models incorporate the best available information, but users should be aware of the inherent errors and uncertainties, as we have outlined above, that went into building this model.

AcknowledgementsThe authors thank Mark Lindsay, Mark Duffett, Daniel Bombardieri and Stephen Wyche for their useful comments in reviewing this manuscript.


29

ReferencesAhmat, AL 1986, Metamorphic patterns in the greenstone belts of the

Southern Cross Province, Western Australia, in Professional papers

for 1984: Geological Survey of Western Australia, Report 19, p. 1–21.

Calcagno, P, Chilès, JP, Courrioux, G. and Guillen, A 2008, Geological

modelling from field data and geological knowledge: Part I.

Modelling method coupling 3D potential-field interpolation and

geological rules: Physics of the Earth and Planetary Interiors

v. 171, p. 147–157.

Cassidy, KF, Champion, DC, Krapež, B, Barley, ME, Brown, SJA,

Blewett, RS, Groenewald, PB and Tyler, IM 2006, A revised

geological framework for the Yilgarn Craton, Western Australia:

Geological Survey of Western Australia, Record 2006/8, 8p.

Chen, SF 2003, Atley, WA Sheet 2741: Geological Survey of Western

Australia, 1:100 000 Geological Series.

Chen, SF 2005, Geology of the Atley, Rays Rocks, and southern

Sandstone 1:100 000 sheets: Geological Survey of Western Australia,

1:100 000 Geological Series Explanatory Notes, 42p.

Chen, SF, Libby, JW, Greenfield, JE, Wyche, S and Riganti, A 2001,

Geometry and kinematics of large arcuate structures formed by

impingement of rigid granitoids into greenstone belts during

progressive shortening: Geology, v. 29, no. 3, p. 283–286.

Chen, SF, Libby, JW, Wyche, S and Riganti, A 2004, Kinematic nature

and origin of regional-scale ductile shear zones in the central

Yilgarn Craton, Western Australia: Tectonophysics, v. 394, no. 3–4,

p. 139–153.

Chen, SF, Morris, PA and Pirajno, F 2006, Komatiites in the Sandstone

greenstone belt, north-central Yilgarn Craton, in GSWA 2006

extended abstracts: promoting the prospectivity of Western Australia:

Geological Survey of Western Australia, Record 2006/3, p. 20–21.

Chen, SF and Painter, MGM 2005, Sandstone, WA Sheet 2742:

Geological Survey of Western Australia, 1:100 000 Geological Series.

Chen, SF, Riganti, A, Wyche, S, Greenfield, JE and Nelson, DR 2003,

Lithostratigraphy and tectonic evolution of contrasting greenstone

successions in the central Yilgarn Craton, Western Australia:

Precambrian Research, v. 127, p. 249–266.

Chen, SF and Wyche, S (compilers) 2001, Archaean granite–greenstones

of the central Yilgarn Craton, Western Australia — a field guide:

Geological Survey of Western Australia, Record 2001/14, 76p.

Chen, SF and Wyche, S 2003, Geology of the Bungalbin 1:100 000 sheet:

Geological Survey of Western Australia, 1:100 000 Geological Series

Explanatory Notes, 27p.

Cooper, RW, Abeysinghe, PB, Strong, CA and Irimies, F (compilers)

2015, Western Australia atlas of mineral deposits and petroleum

fields, 2015: Geological Survey of Western Australia, 50p.

Costelloe, RD and Jones, LEA 2014, Youanmi seismic survey 2010:

acquisition and processing, in Youanmi and southern Carnarvon

seismic and magnetotelluric (MT) workshop 2013 compiled by S Wyche, TJ Ivanic and I Zibra: Geological Survey of Western

Australia, Perth, Record 2013/6, p. 1–6.

Emmerson, DW 1990, Notes on mass properties of rock — density,

porosity, permeability: Exploration Geophysics, v. 21, p. 209–216.

Geological Survey of Western Australia 2008, Central Yilgarn, 2008

update: Western Australia Geological Survey, 1:100 000 Geological

Information Series.

Geological Survey of Western Australia 2013, Gravity anomaly grid

(400 m) of Western Australia (2013 – version 2), 11 November 2013

update: Geological Survey of Western Australia, digital data layer.

Geological Survey of Western Australia 2014a, Magnetic anomaly grid

(80 m) of Western Australia (2014 – version 1), 16 September 2014

update: Geological Survey of Western Australia, digital data layer.

Geological Survey of Western Australia 2014b, MINEDEX: mines and

minerals deposits database: Department of Mines and Petroleum,

Perth, Western Australia, viewed 16 October 2014, <www.dmp.

wa.gov.au/minedex>.

Geological Survey of Western Australia 2014c, Radiometric merged grid

(80 m) of Western Australia (2014): Geological Survey of Western

Australia, digital data layer.

Gessner, K, Jones, T, Goodwin, JA, Gallardo, LA, Milligan, PR, Brett, J

and Murdie, RE 2014, Interpretation of magnetic and gravity

data across the Southern Carnarvon Basin, and the Narryer and

Youanmi Terranes, in Youanmi and Southern Carnarvon seismic

and magnetotelluric (MT) workshop 2013 compiled by TJ Ivanic,

S Wyche and I Zibra: Geological Survey of Western Australia, Record

2013/6, p. 65–77.

Greenfield, JE 2001, Geology of the Lake Giles 1:100 000 sheet:



Ivanic, TJ, Korsch, RJ, Wyche, S, Jones, LEA, Zibra, I, Blewett, RS,

Jones, T, Milligan, P, Costelloe, RD, Van Kranendonk, MJ,

Doublier, MP, Hall, CE, Romano, SS, Pawley, MJ, Gessner, K,

Patison, N, Kennet, BLN and Chen, SF 2014a, Preliminary

interpretation of the 2010 Youanmi deep seismic reflection lines

and magnetotelluric data for the Windimurra Igneous Complex,

in Youanmi and southern Carnarvon seismic and magnetotelluric

(MT) workshop 2013 compiled by S Wyche, TJ Ivanic and I Zibra:

Geological Survey of Western Australia, Perth, Record 2013/6,

p. 97–111.

Ivanic, TJ, Wingate, MTD, Kirkland, CL, Van Kranendonk, MJ and

Wyche, S 2010, Age and significance of voluminous mafic–ultramafic

magmatic events in the Murchison Domain, Yilgarn Craton:

Australian Journal of Earth Sciences, v. 57, p. 597–614.

Ivanic, TJ, Wingate, MTD, Korsch, RJ, Blewett, RS, Jones, LEA, Wyche,

S, Zibra, I, Doublier, MP, Romano, SS, Pawley, MJ, Van Kranendonk,

MJ, Gessner, K, Hall, CE, Chen, SF, Patison, N and Costelloe, RD

2014b, Preliminary interpretation of the Youanmi deep seismic

reflection lines for Proterozoic intrusive rocks, in Youanmi and

Southern Carnarvon seismic and magnetotelluric (MT) workshop

2013 compiled by S Wyche, TJ Ivanic and I Zibra: Geological Survey

of Western Australia, Record 2013/6, p. 81–85.

Jachens RC, Wentworth CM, Gautier DL and Pack S 2001, 3D geologic

maps and visualization: A new approach to the geology of the Santa

Clara (Silicon) Valley, California. Digital Mapping Techniques '01

Workshop Proceedings: U.S. Geological Survey Open-File Report

01-223.

Jarvis, A, Reuter, HI, Nelson, A and Guevara, E 2008, Hole-filled SRTM

for the globe Version 4, available from the CGIAR-CSI SRTM 90m

Database: CGIAG Consortium for Spatial information, <http://srtm.

csi.cgiar.org>.

Kennett, BLN 2014, The nature of the lithosphere in the vicinity of the

Youanmi seismic lines, in Youanmi and Southern Carnarvon seismic

and magnetotelluric (MT) workshop 2013 compiled by S Wyche,

TJ Ivanic and I Zibra: Geological Survey of Western Australia,

Record 2013/6, p. 27–32.

Kennett, BLN and Salmon, M 2012, AuSREM: Australian seismological

reference model: Australian Journal of Earth Sciences, v. 59,

p. 1091–1103.

Kennett, BLN, Salmon, M, Saygin, E and AusMoho Working Group

2011, AusMoho: the variation of Moho depth in Australia:

Geophysical Journal International, v. 187, no. 2, p. 946–958,

doi:10.1111/j.1365-246X.2011.05194.x.

Murdie et al.

30

Korsch, RJ, Blewett, RS, Wyche, S, Zibra, I, Ivanic, TJ, Doublier, MJ,

Romano, SS, Pawley, MJ, Johnson, SP, Van Kranendonk, MJ,

Jones, LEA, Kositcin, N, Gessner, K, Hall, CE, Chen, SF, Patison, N,

Kennett, BLN, Jones, T, Goodwin, JA, Milligan, P and Costelloe, RD

2014, Geodynamic implications of the Youanmi and southern

Carnarvon deep seismic relection surveys: a ~1300 km traverse from

the Pinjarra Orogen to the eastern Yilgarn Craton, in Youanmi and

southern Carnarvon seismic and magnetotelluric (MT) workshop

2013 compiled by S Wyche, TJ Ivanic and I Zibra: Geological Survey

of Western Australia, Perth, Record 2013/6, p. 147–166.

Lajaunie, C, Courrioux, G and Manuel, L 1997, Foliation fields and 3D

cartography in geology: Principles of a method based on potential

interpolation: Mathematical Geology v. 29, p. 571–584.

Martin, DMcB, Hocking, RM, Riganti, A and Tyler, IM 2014, 1:500 000

State interpreted bedrock geology of Western Australia, 2014:

Geological Survey of Western Australia, <www.dmp.wa.gov.au/

geoview>.

Martin, DMcB, Hocking, RM, Riganti, A and Tyler, IM 2015, 1:500 000

tectonic units of Western Australia, 2015: Geological Survey of

Western Australia, <www.dmp.wa.gov.au/geoview>.

Milligan, PR, Duan, J, Fomin, T, Nakamura, A and Jones, T 2014, The

Youanmi magnetotelluric (MT) transects, in Youanmi and southern

Carnarvon seismic and magnetotelluric (MT) workshop 2013

compiled by S Wyche, TJ Ivanic and I Zibra: Geological Survey of

Western Australia Record 2013/6, p. 13–25.

Myers, JS 1993, Precambrian history of the Western Australian Craton

and adjacent orogens: Annual Review of Earth and Planetary

Sciences, v. 21, p. 453–485.

Myers, JS 1997, Preface: Archaean geology of the Eastern Goldfields

of Western Australia — regional overview: Precambrian Research,

v. 83, p. 1–10.

Myers, JS and Swager, C 1997, The Yilgarn Craton, in Greenstone belts

edited by M de Wit and LD Ashwal: Clarendon Press, Oxford, UK,

p. 640–656.

Oldenburg, D.W. and Pratt D.A. 2007, Geophysical inversion for

mineral exploration: a decade of progress in theory and practice, in

Proceedings of Exploration ’07 edited by B Milkereit: Fifth Decennial

International Conference on Mineral Exploration, Toronto, 2007,

p. 61–99.

Perring, CS, Barnes, SJ and Hill, RET 1996, Geochemistry of komatiites

from Forrestania, Southern Cross Province, Western Australia:

evidence for crustal contamination: Lithos, v. 37, p. 181–197.

Rawling TJ, Osborne CR, Mclean MA, Skladzien PB Cayley RA,

Williams B 2011, 3D Victoria Final Report, Geoscience Victoria

3D Victoria Report 14: Department of Primary Industries, Victoria.

Riganti, A 2003, Geology of the Everett Creek 1:100 000 sheet:



Riganti, A and Chen, SF 2002, Geology of the Jackson 1:100 000 sheet:



Rudnick, RL and Fountain, DM 1995, Nature and composition of the

continental crust: a lower crustal perspective: Reviews of Geophysics,

v. 33, p. 267–309.

Russell HAJ, Berg RC, and Thorleifson LH (compilers) 2011, Three-

dimensional geological mapping; Workshop extended abstracts,

Minneapolis, Minnesota – October 8, 2011: Geological Survey of

Canada, Open File 6998, p. 80 doi:10.4095/239609

Schön, JH 1998, Physical properties of rocks: fundamentals and

principles of petrophysics: Pergamon Press, Michigan, USA, 583p.

Shevchenko, S 2005 Geophysical modelling of the Sandstone greenstone

belt, in Geology of the Atley, Rays Rocks, and southern Sandstone

1:100 000 sheets compiled by SF Chen: Geological Survey of Western

Australia 1:100 000 Geological Series Explanatory Notes, p. 21–22.

Stewart, AJ, Williams, IR and Elias, M (compilers) 1983, Youanmi,

Western Australia Sheet SH 50-4: Geological Survey of Western

Australia, 1:250 000 Geological Series Explanatory Notes, 58p.

Van Kranendonk, MJ, Ivanic, TJ, Wingate, MTD, Kirkland, CL and

Wyche, S, 2013, Long-lived, autochthonous development of the

Archean Murchison Domain, and implications for Yilgarn Craton

tectonics: Precambrian Research, v. 229, p. 49–92.

Western Areas Ltd 2015, Sandstone: Western Areas Ltd, Perth, Western

Australia, viewed 1 April 2015, <http://www.westernareas.com.au/

exploration-a-growth/forrestania-regional/sandstone.html>.

Williams, NC 2009, Mass and magnetic properties for 3D geological and

geophysical modelling of the southern Agnew–Wiluna greenstone

belt and Leinster nickel deposits, Western Australia: Australian

Journal of Earth Sciences, v. 56, no. 8, p. 1111–1142.

Wyche, S, Ivanic, TJ and Zibra I 2014, Youanmi and southern Carnarvon

seismic and magnetotelluric (MT) workshop 2013: Geological Survey

of Western Australia, Perth, Record 2013/6, 180p. + 4 plates.

Wyche, S, Chen, SF, Greenfield, JE and Riganti, A 2001, Geology of

the Johnston Range 1:100 000 sheet: Geological Survey of Western

Australia, 1:100 000 Geological Series Explanatory Notes, 31p.

Wyche, S, Nelson, CS and Riganti, A 2004, 4350–3130 Ma detrital

zircons in the Southern Cross granite–greenstone terrane, Western

Australia: implications for the early evolution of the Yilgarn Craton:

Australian Journal of Earth Sciences, v. 51, p. 31–45.

Wynn, P and Bacchin, M (compilers) 2009, Index of gravity surveys (2nd

edition): Geoscience Australia, Record 2009/007, 1832p.

Zibra, I, Gessner, K, Pawley, MJ, Wyche, S, Chen, SF, Korsch, RJ,

Blewett, RS, Jones, T, Milligan, P, Jones, LEA, Doublier, MP,

Hall, CE, Romano, SS, Ivanic, TJ, Patison, N, Kennett, BLN and Van

Kranendonk, MJ 2014, Preliminary interpretation of deep seismic

line 10GA-YU2: Youanmi Terrane and western Kalgoorlie Terrane,

in Younami and Southern Carnarvon seismic and magnetotelluric

(MT) workshop 2013 compiled by S Wyche, TJ Ivanic and I Zibra:

Geological Survey of Western Australia, Record 2013/6, p. 87–96.

GS

WA

Record

2015/11 S

andstone g

reenstone belt, northern central Yilg

arn Craton: 3D

mod

elling

31

Model object list Notes (see accompanying Record for further information)Data reference (see accompanying Record for further information)

3D model

Structure surfaces

F1_Youanmi_Shear_Zone The Youanmi Shear Zone is the domain boundary between the Murchison and Southern Cross

Domains of the Yilgarn Craton. The surface represents the western margin of the several kilometre-

wide shear zone farthest from the greenstone. This also forms the base of the foliated granite

derived from the Murchison granite.

Wyche et al. (2014); magnetic anomaly grid (GSWA, 2014a)

F2_Edale_Shear_Zone This marks the eastern extent of the foliated granite, derived from Southern Cross granite, which

bounds the east side of the greenstone belt. F3 and F4 are also part of the Edale Shear Zone.

Wyche et al. (2014); magnetic anomaly grid (GSWA, 2014a)

F3_Granite_fault

F4_Granite_fault

Faults within the Southern Cross Domain which form the boundary of the foliated granites. SANDSTONE 1:100k map sheet (Chen and Painter, 2005)

F5_Greenstone_fault

F6_Greenstone_fault

F7_Greenstone_fault

Selected faults within the Sandstone greenstone belt. ATLEY 1:100k map sheet (Chen, 2003, 2005)

Lower bounding surfaces

S01_Proterozoic_dyke The envelope of the Proterozoic dykes or sills that cross the model. Magnetic anomaly grid (GSWA, 2014a); Ivanic et al. (2014b)

S02_Bald_Rock_Supersuite The envelope of a post-tectonic granite belonging to the Bald Rock Supersuite. Wyche et al. (2014)

S03_D2_granite The envelope of an undated granite which lies between the Youanmi and Edale Shear Zones. ATLEY 1:100k map sheet (Chen, 2003, 2005)

S04_Metasedimentary The base of the metasedimentary unit seen in the northern apex of the greenstone belt. SANDSTONE 1:100k map sheet (Chen and Painter, 2005)

S05_Ultramafic The base of the ultramafic unit which unconformably overlies the rest of the greenstone belt. ATLEY 1:100k map sheet (Chen, 2003, 2005)

S06_Mafic_with_BIF The internal boundary within the Sandstone greenstone belt between the stratigraphically higher

mafic unit with BIF and the BIF unit.

SANDSTONE 1:100k map sheet (Chen and Painter, 2005);

ATLEY 1:100k map sheet (Chen, 2003, 2005)

S07_Banded_iron_formation The internal boundary within the Sandstone greenstone belt between the stratigraphically higher

BIF and the lower massive mafic unit.



S08_Massive_mafic The internal boundary within the Sandstone greenstone belt between the stratigraphically higher

massive mafic unit and the lower tremolite schist.



S09_Tremolite_schist The internal boundary within the Sandstone greenstone belt between the stratigraphically higher

tremolite schist and the lower basal mafic unit.



S10_Base_of_greenstone The boundary between the different greenstone units and the surrounding foliated granite. SANDSTONE 1:100k map sheet (Chen and Painter, 2005);


S11_Unaly_Hill_greenstone The envelope of the Unaly Hill greenstone which lies to the southwest along the Youanmi

Shear Zone.


S12_Gum_Creek_greenstone The lower surface of the Gum Creek Greenstone. Individual units, although mapped on the

SANDSTONE 1:100k map sheet, are not differentiated here.

SANDSTONE 1:100k map sheet (Chen and Painter, 2005)

S13_Barrambie_Igneous_

Complex

The envelope of the Barrambie Igneous Complex which lies to the north along the Youanmi

Shear Zone.


Ap

pen

dix

Metadata Table 1: Model object list, notes and data references.

Table of information to accompany 3D model of the Sandstone greenstone belt and associated GSWA Record: Murdie, RE, Gessner, K, and Chen, SF 2015, The Sandstone greenstone

belt, northern central Yilgarn Craton: 3D modelling using geological and geophysical constraints: Geological Survey of Western Australia, Record 2015/11, 33p.

Murd

ie et al.

32


S14_North_Cook_Well_

greenstone

The envelope of the North Cook Well Greenstone which lies just east of the Edale Shear Zone, but

does not outcrop in the model.

Magnetic anomaly grid (GSWA, 2014a)

S17_Murchison_granite The base of the Murchison granite, equivalent to the base of the low reflectivity crust as seen in

Zibra et al. (2014) to the west of the Youanmi Shear Zone.

Wyche et al. (2014)

S18_Yarraquin_Seismic_Province The base of the Yarraquin Seismic Province, equivalent to the base of the high reflectivity crust as

interpreted from the seismic data in Zibra et al. (2014).

Wyche et al. (2014)

S19_Lower_crust The base of the lower crust, equivalent to the Moho. Wyche et al. (2014)

Volume surfaces (wrapped, hollow objects)

V01_Proterozoic_dyke The Proterozoic dykes or sills that cross the model. Magnetic anomaly grid (2014a); Ivanic et al. (2014b)

V02_Bald_Rock_Supersuite A post-tectonic granite belonging to the Bald Rock Supersuite. Wyche et al. (2014)

V03_D2_granite An undated granite which lies between the Youanmi and Edale Shear Zones. ATLEY 1:100k map sheet (Chen, 2003, 2005)

V04_Metasedimentary The metasedimentary unit seen in the northern apex of the greenstone belt. SANDSTONE 1:100k map sheet (Chen and Painter, 2005)

V05_Ultramafic The ultramafic unit which unconformably overlies the rest of the greenstone belt. ATLEY 1:100k map sheet (Chen, 2003, 2005)

V06_Mafic_with_BIF The mafic unit with BIF which is a basaltic unit with thin layers of BIF. SANDSTONE 1:100k map sheet (Chen and Painter, 2005);


V07_Banded_iron_formation The BIF unit consisting of a major unit of BIF with cherts, gabbro and basalt lenses. SANDSTONE 1:100k map sheet (Chen and Painter, 2005);


V08_Massive_mafic The massive mafic unit which is a tholeiitic basalt with gabbro and komatiitic basal lenses. SANDSTONE 1:100k map sheet (Chen and Painter, 2005);


V09_Tremolite_schist The tremolite schist unit which consists of a thin BIF intercalated with clastic sedimentary rocks

and intruded by gabbro sills sitting below a tremolite–chlorite(–talc) schist.



V10_Basal_mafic The basal mafic unit consisting of a basal foliated basalt and a basalt unit with local amphibolite

and gabbro lenses.



V11_Unaly_Hill_greenstone The Unaly Hill greenstone which lies to the southwest along the Youanmi Shear Zone. ATLEY 1:100k map sheet (Chen, 2003, 2005)

V12_Gum_Creek_greenstone The Gum Creek Greenstone. Individual units, although mapped on the SANDSTONE 1:100k map

sheet, are not differentiated here.


V13_Barrambie_Igneous_

Complex

The Barrambie Igneous Complex which lies to the north along the Youanmi Shear Zone. SANDSTONE 1:100k map sheet (Chen and Painter, 2005)

V14_North_Cook_Well_

greenstone

The North Cook Well Greenstone which lies just east of the Edale Shear Zone, but does not

outcrop in the model.

Magnetic anomaly grid (GSWA, 2014a)

V15_Foliated_granite Granites of both the Murchison and Southern Cross Domains which form the actual shear zones of

the Youanmi and Edale Shear Zones.



V16_Southern_Cross_granite The low reflectivity crust as seen in Zibra et al. (2014) to the east of the Youanmi Shear Zone. This

encompasses granites of the Southern Cross Domain.

Wyche et al. (2014)

V17_Murchison_granite The low reflectivity crust as seen in Zibra et al. (2014) to the west of the Youanmi Shear Zone. This

encompasses granites of the Murchison Domain.

Wyche et al. (2014)

V18_Yarraquin_Seismic_Province The high reflectivity crust as interpreted from the seismic data in Zibra et al. (2014). Wyche et al. (2014)

GS

WA

Record

2015/11 S

andstone g

reenstone belt, northern central Yilg

arn Craton: 3D

mod

elling

33


V19_Lower_crust The lower crust. Wyche et al. (2014)

V20_Mantle The mantle below the Moho. Wyche et al. (2014)

3D geology Units as per surface volumes.

Geological Maps

500k_bedrock_geology Based on 1:500 000 State interpreted bedrock geology, 2014, with detail from 1:100 000 mapping. Martin et al. (2014); Chen (2003); Chen and Painter (2005)

Gravity

Sandstone_grav_boug_grid Geosoft grid of Bouguer gravity anomaly data. Gravity compilation (GSWA, 2013)

Sandstone_grav_boug Bouguer gravity anomaly data utilized in inversion modelling in this study. Gravity compilation (GSWA, 2013)

VPmg_gravity_inversion_results A points dataset showing the observed, calculated and residual gravity data for the 3D model.

Magnetics

Sandstone_TMI_grid Geosoft grid of total magnetic intensity (TMI) data. Magnetic anomaly grid (GSWA, 2014a)

Sandstone_TMI_colour Total magnetic intensity (TMI) data image. Magnetic anomaly grid (GSWA, 2014a)

VPmg_magnetic_inversion_results A points dataset showing the observed, calculated and residual TMI data for the 3D model.

Radiometrics

Sandstone_KTU Radiometric KTU (potassium, thorium, uranium) data displayed in a ternary inverse CMYK image. Radiometric grid (GSWA, 2014c)

Relief

Sandstone_SRTM_elevation_grid Geosoft grid file of the topography. Jarvis et al. (2008)

Sandstone_SRTM_elevation Flat image of the topography. Jarvis et al. (2008)

Sandstone_SRTM_elevation_3D A surface of the topography.

Sections

Map sections

Atley_xsec-2741_ab

Sandstone_xsec-2742_ab

Sandstone_xsec-2742_cd

Cross-sections from published SANDSTONE and ATLEY 1:100 000 geological map sheets. GSWA mapping and WAROX data; Chen (2003); Chen and

Painter (2005)

Seismic reflection surveys

Sandstone_int_10GA-YU2 Interpreted seismic line image on a flat plane approximately through the centre of the trace of the

seismic line.

Wyche et al. (2014); Ivanic et al. (2014a)

Sandstone_int_10GA-YU2_3D Interpreted seismic line image correctly located along the common depth point line (only in

GOCAD model).

Wyche et al. (2014); Ivanic et al. (2014a)

10GA_YU2 Common depth point locations for the seismic reflection survey. Wyche et al. (2014)

THE SANDSTONE GREENSTONE BELT, NORTHERN CENTRAL YILGARN CRATON: 3D M

ODELLING USING GEOLOGICAL AND GEOPHYSICAL CONSTRAINTS

RECORD 2015/11M

urdie et al.

This Record is published in digital format (PDF) and is available as a free

download from the DMP website at

<www.dmp.wa.gov.au/GSWApublications>.

Information Centre

Department of Mines and Petroleum

100 Plain Street

EAST PERTH WESTERN AUSTRALIA 6004

Phone: +61 8 9222 3459 Fax: +61 8 9222 3444

www.dmp.wa.gov.au/GSWApublications

Further details of geological products produced by the

Geological Survey of Western Australia can be obtained by contacting:

Documents

Record 2015/11: The Sandstone greenstone belt, northern ......regional geophysical datasets, the results of regional geophysical inversions, and out-of-section geometrical features