See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/242219231

3D Geologic Modelling for Groundwater Applications in the Condamine

Headwater Sub-catchments, Queensland

Article

CITATIONS READS

4 294

4 authors, including:

Pauline English J. Levett

Geoscience Australia
3 PUBLICATIONS   8 CITATIONS   
32 PUBLICATIONS   511 CITATIONS   
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Pauline English on 24 September 2014.

The user has requested enhancement of the downloaded file.

 G E O S C I E N C E A U S T R A L I A

3D Geologic Modelling for

Groundwater Applications in
the Condamine Headwater
Sub-catchments, Queensland
P. English, J. Levett, S. Lewis and J. Sandow.

Record
2008/03

S PAT I A L I N F O R M AT I O N F O R T H E N AT I O N
3D Geologic Modelling for Groundwater
Applications in the Condamine
Headwater Sub-catchments, Queensland

A Pilot Project for the Condamine Alliance

P. English1, J. Levett2, S. Lewis1 and J. Sandow1

GEOSCIENCE AUSTRALIA RECORD 2008/03

1. Geospatial & Earth Monitoring Division, Geoscience Australia GPO Box 378 Canberra ACT 2601
2. Mira Geoscience, 310 Victoria Ave, Suite 309, Westmount, H3Z 2M9, QC, Canada
Department of Resources, Energy and Tourism
Minister for Resources, Energy and Tourism: The Hon. Martin Ferguson, MP
Secretary: Dr Peter Boxall

Geoscience Australia
Chief Executive Officer: Dr Neil Williams

© Commonwealth of Australia, 2008

This work is copyright. Apart from any fair dealings for the purpose of study, research, criticism,
or review, as permitted under the Copyright Act 1968, no part may be reproduced by any process
without written permission. Copyright is the responsibility of the Chief Executive Officer,
Geoscience Australia. Requests and enquiries should be directed to the Chief Executive Officer,
Geoscience Australia, GPO Box 378 Canberra ACT 2601.

Geoscience Australia has tried to make the information in this product as accurate as possible.
However, it does not guarantee that the information is totally accurate or complete. Therefore, you
should not solely rely on this information when making a commercial decision.

ISSN 1448-2177
ISBN 978-1-921236-75-4 (web)
ISBN 978-1-921236-76-1 (hardcopy)

GeoCat # 65449

Bibliographic reference: English, P., Levett, J., Lewis, S., Sandow, J., 2008. 3D Geologic
Modelling for Groundwater Applications in the Condamine Headwater Sub-catchments,
Queensland. Geoscience Australia. Record 2008/03. 45pp.

ii
Contents

Executive Summary v
Introduction 1
Rationale 1
The project 1
Scope 1
Existing Information 3
The study area 3
Geology 5
Hydrogeology 8
Groundwater flow systems 8
Groundwater processes 8
3D Model Development 12
Preliminary stages 12
Overview of the 3D hydrogeological model 12
Hydrostratigraphic data analysis 14
3D modelling using Gocad 17
Importing data into Gocad 17
Surface modelling 20
3D Model Enhancement and Further Analysis 27
X3D representation of the Gocad model 27
The groundwater system of the Condamine headwater catchments 27
Upgrading the existing 3D model 29
Further 3D modelling in the Condamine catchment 31
Key Recommendations 34
Acknowledgments 34
References 35

Appendices
Appendix 1 – Lithological Borehole Codes 40
Appendix 2 – 3D Grid Construction 41
Appendix 3 – Gocad Object Definitions 43

iii
List of Figures
1. Location diagram and topography 3
2. SPOT5 image of study area 4
3. Eastern Darling Downs agricultural landscape – basalt terrain 6
4. Irrigation in Hodgson Creek sub-catchment 9
5. Conceptual model of hydrogeological processes 10
6. Distribution of bores used for 3D model-building 14
7. Bore distribution and drainage lines 15
8. Example of bore data in Gocad 15
9. Mapped contact between Quaternary alluvium and Main Range Volcanics 16
10. Oblique view of the DEM 18
11. Detail of contrasting DEM resolution across the catchment boundary area 19
12. Preliminary Walloon Coal Measure surface in Gocad 21
13. Top of Walloon – Base of Main Range Volcanics (MRV) surface 22
14. Side view of the main surfaces and boreholes in Gocad 23
15. Oblique view showing the DEM, base of the model and boreholes 23
16. Oblique view showing the three main surfaces in the model 24
17. Side view showing the relief and boreholes with aquifers highlighted 24
18. Close-up view showing Hodgson boreholes and variable aquifer distribution 25
19. Close-up view showing Kings Creek boreholes and variable aquifer distribution 26

List of Tables
1. Geology and landforms 7
2. Log for RN119852 deep borehole sunk in Toowoomba in 2007 8

iv
Executive Summary
Geoscience Australia has developed a computer-generated 3D model which conceptualises the key
hydrogeological components of the Hodgson and Kings Creek sub-catchments in south-eastern
Queensland. These are headwater tributaries of the Condamine River, and are situated in the
agriculturally important Darling Downs region south-west of Toowoomba. Groundwater is an
important resource for rural communities and landholders in the area, and is used extensively for
crop irrigation, stock and domestic purposes, and town water supplies. Our principal objective in
building the 3D model was to help promote greater understanding and appreciation of the region’s
groundwater resources among the numerous local people and communities that rely upon
groundwater, but who may not be familiar with the key concepts and principles which govern its
intrinsic behaviour. Additionally, the model can be applied by surface water and groundwater
scientists, water planners and natural resource managers to assist in the ongoing study, management
and administration of the region’s important aquifer systems.

Approximately 80 % of groundwater extracted from the Hodgson and Kings Creek sub-catchments
is derived from the Main Range Volcanics (MRV). The MRV comprises Tertiary basalts and minor
tuffs which are widespread across the Darling Downs, and have a nominal thickness of up to 200 m
in the study area (and likely up to ~300 m beneath the nearby Great Dividing Range). A significant
weathered basalt profile typifies the landscape surface, comprising fertile soils and lateritic material.
Local and intermediate groundwater flow systems occur in the fractured rock aquifers of the MRV.
Bore yields are highly variable due to the complex and irregular nature of high permeability
networks which control the movement of groundwater. Subhorizontal water-bearing zones are
predominantly hosted in weathered or vesicular basalt, commonly in the upper parts of individual
layers. Significant subvertical discontinuities also provide important fluid pathways within the
MRV; these include faults and fractures, columnar or radial joint networks, and the overlapping
edges of terminal palaeo-lava flows.

The Jurassic Walloon Coal Measures unconformably underlie the MRV and provide locally
important aquifers in the Hodgson and Kings Creek sub-catchments. Most water-bearing zones in
this system occur in coarse-grained sandstone and conglomerate beds, or in narrow coal seams
(mostly < 2 m thick). Regional groundwater flow systems occur in these Jurassic rocks, which have
only limited hydraulic connection with the overlying volcanic rocks. Minor volumes of groundwater
are also sourced from Cainozoic alluvium in both sub-catchments, particularly thicker such deposits
in the valley floors. The sporadic distribution and discontinuous nature of these unconsolidated
sediments however, tend to limit the usefulness and reliability of alluvial aquifers.

Development of the 3D computer-generated model relied heavily on water bore data (borehole logs)
obtained from water drilling operations conducted in the local area over the past 60 years. These
data were of variable quality and consistency and required significant analysis and modification to
enable modelling. Data from approximately 1100 water bores was selected to build the
hydrogeological model. Critical components of our 3D model include borehole collar locations and
total hole depths, the significant water-bearing intersections in each hole (colour-coded based on
hydrostratigraphic unit), and the boundaries of the main hydrostratigraphic units, such as the
unconformable contact between the Walloon Coal Measures and the MRV. The 3D model also
accurately depicts the surface topography across the sub-catchments (digital elevation model), and
the location of significant natural and cultural landmarks such as streams, towns and roads.
Additionally, a variety of 2D maps and images can be draped across the DEM; these include
regional surface geology maps, remotely sensed satellite images, and various available coarse-

v
resolution geophysical datasets that cover the area, at least partially, e.g., airborne magnetics and
electromagnetics survey data.

Further development of the 3D conceptual model described here is warranted to address specific
natural resource management issues. Future modelling could potentially take many forms, and
depends largely upon the needs and priorities of key stakeholder groups and end-users. We
recommend trialling the 3D modelling approach in other sub-catchments of the Condamine River,
especially those with markedly different geological structure and composition, i.e., not volcanic
rock-dominated. Eventually, the entire Condamine River basin could be modelled, although this
should proceed on a sub-catchment basis. Other suggestions for future work include addressing
specific hydrological problems focussed on smaller areas of the Hodgson and Kings Creek sub-
catchments, where more comprehensive and reliable borehole data exists, e.g., modelling temporal
changes in the standing water level across multiple monitoring bores. Additionally, the 3D model
could also be used to investigate hydrological processes associated with groundwater dependant
ecosystems, or to better understand the nature of surface water and groundwater interactions within
the sub-catchments or between these relatively well-studied areas and adjacent/subjacent aquifer
systems.

vi
Introduction
RATIONALE
The recent prolonged drought and associated water crisis in eastern Australia has highlighted the
need to effectively plan the allocation of limited water resources. This requires a strategic
understanding of the impacts of over-extraction of the connected groundwater–surface water
resource. In areas where groundwater is a significant component of the connected water system,
there is often extensive ‘point’ information about aquifers from driller’s descriptions of the materials
encountered while drilling groundwater boreholes. It is not straightforward, however, to integrate
this information across the landscape. Understanding the distribution of subsurface water-bearing
zones, and the likely movement of groundwater within aquifers, requires considerable interpretation
between water bores. This can be extremely challenging for specialists and non-specialists alike,
particularly in complex hydrogeological areas such as the upper Condamine catchment, south-east
Queensland.

A 3-dimensional geologic model (computer-generated) for groundwater applications has been

developed by Geoscience Australia as a pilot project for the Condamine Alliance, to provide a
conceptual framework for better understanding groundwater in the Hodgson and Kings Creek sub-
catchments. The pilot project aims to integrate existing datasets, such as bore logs and digital
elevation models, to construct a conceptual 3D geological model appropriate for land and water
applications, particularly for future hydrogeologic modelling. It is hoped that this conceptualisation
will advance understanding of the upper Condamine groundwater system and encourage better
communication between hydrogeologists, catchment managers and land users.

THE PROJECT

The pilot project comprised two stages for the first half of 2007:

STAGE 1: Data review and feasibility assessment; and

STAGE 2: Model design and construction.

These stages are documented in the present report. Potential further work, to build 3D models of
additional sub-catchments or a ‘whole-of-catchment’ groundwater model, is contingent upon the
outcomes of these initial phases and stakeholder requirements.

SCOPE

The project builds on the 3D visualisation capacity at Geoscience Australia which has augmented
various geologic, ore body definition and estuarine projects1, e.g., Ruperto and Nicoll, 2007. The
project also draws upon recent advances in the USA and Canada2 where 3D geologic mapping for
groundwater systems has been underway for approximately five years, particularly work coordinated
by the Illinois State Geological Survey, e.g., Thorleifson et al., 2005; Bajc and Newton, 2005.

The Condamine 3D Project is Geoscience Australia’s first 3D application to a hydrogeologic or

groundwater problem. The objective here is to provide a useful interactive tool to communicate

1
http://www.ga.gov.au/map/web3d/
2
http://www.isgs.uiuc.edu?3DWorkshop/nu_3d_welcome.html

1
complex geological–hydrogeological relationships to the Condamine catchment managers, other
scientists and the wider community. At the technical level, the objective is to be able to turn data
layers on and off and manipulate the model by rotating, zooming and panning to explore
hydrogeological relationships from any aspect. These functions will aid interpretation of the
groundwater distribution and enhance understanding of the hydrogeologic system and groundwater
processes. Given that the project is a new application of 3D geologic modelling, it has experimental
components for which no pre-existing templates are available. The study area in the upper
Condamine catchment was selected because it is data-rich and agriculturally important; thus, an
enhanced understanding of the groundwater system is highly desirable. Initial stages of the project
included a field reconnaissance of the Hodgson and Kings Creek study area with hydroscientists
from the Queensland Department of Natural Resources and Water (QDNRW). The project brings
together relatively new 3D geologic modelling techniques – developed largely for the mining
industry – with subsurface geologic and hydrogeologic datasets from a dynamic hydrologic system.

This report is a synthesis of two Geoscience Australia client reports written for the Condamine
Alliance during the pilot project:

REPORT 1: ‘3D Groundwater Conceptual Model Pilot Project: Feasibility Report for
the Condamine Alliance, Queensland’ (May 2007).
REPORT 2: ‘3D Groundwater Conceptual Model – Pilot Project for the Condamine
Alliance: 3D Modelling and Visualisation of the Hydrostratigraphic Systems in the
Hodgson and Kings Creek sub-catchments’ (August 2007).
Sections of Report 1 that outlined feasibility phases and ‘teething problems’ are excluded from the
present report. The project has relied heavily on data provided by the QDNRW, and on Earth
Decision Suite software (Gocad) and the expertise of Mira Geoscience3. This report summarises the
task of generating a 3D model using Gocad and outlines the multi-step process, providing:

1. An overview of the rationale and scope of the project;

2. A summary of the geology and hydrogeology of the study area;

3. Concepts and features of 3D geologic modelling;

4. Analysis of the hydrostratigraphic data which provided the foundation for building the
model;
5. An overview of Gocad software, routines and methods used to construct the 3D
components of the model, including a selection of random screen snapshots;
6. An overview of the groundwater system in the Condamine headwater sub-catchments;
7. Recommendations for further work which may enhance the 3D model or for modelling
wider reaches of the Condamine catchment; and

8. Appendices summarising bore data analysis, some of the technical procedures, and
contour plots of total aquifer thickness and yields.

3 Mira Geoscience: http://www.mirageoscience.com/products/gocad.php

2
Existing Information
THE STUDY AREA
The pilot study area covers Hodgson Creek and Kings Creek sub-catchments, in the eastern part of
the Condamine River basin, south-eastern Queensland (Figure 1). These adjacent sub-catchments are
located immediately south and west of Toowoomba and form part of the eastern Darling Downs, in
the northern headwaters of the Murray-Darling catchment. Topographic relief ranges from close to
700 m AHD on the Great Dividing Range near Toowoomba, to 400 m AHD at the confluence of
Hodgson Creek with the Condamine River. Hodgson Creek sub-catchment covers an area of 792
km2, and Kings Creek sub-catchment 796 km2.

Figure 1: A shaded relief contour map showing sub-catchment boundaries (black) of Hodgson
Creek and Kings Creek, the upper Condamine River drainage network and the many local
communities (north to top of page). Location in south-east Queensland is shown in the inset box.
Map dimensions are approximately 100 km east-west and 80 km north-south.

Umbiram and Emu Creeks are the main tributaries in the Hodgson Creek sub-catchment. A synopsis
of Hodgson Creek catchment, with respect to conceptualising groundwater–stream interaction is
provided in Reid et al. (2007). The landscape varies from rocky ridges and scattered volcanic cones
to undulating plains. The region has fertile soils and is a highly productive agricultural part of the
Murray-Darling Basin. The area was first settled by Europeans around 1840 when it was used
primarily for grazing. Floodplain and lowland regions have mostly been cleared of native vegetation
including riparian and floodplain woodlands; remnant vegetation cover is around 20% (Figure 2).
Mountain Coolabah and Ironbark trees are common on basalt hills, with grazing occurring under
some woodland (Rattray et al., 2002). Patches of Brigalow/Belah, Poplar Box, Ironbark, Bulloak and

3
Cypress Pine are associated with sandstone areas. Cleared land is used for dryland and irrigated
cropping, pastures, grazing and horticulture. Winter crops (wheat, barley, oats and chickpeas) and
summer crops (sorghum, sunflower and maize) are supported by the fertile soils and a favourable
climate. Cattle and sheep are commonly grazed.

Regional issues associated with water quality include: pesticides in cropping areas; high sediment
and nutrient loads in runoff from cropping, grazing and urban development; and accelerated bank
and gully erosion in some areas. Valley infilling occurred after agricultural development by
extensive sheet, rill and gully erosion. Groundwater is an important source of water for crop
irrigation, and for stock and domestic water supplies in the rich eastern Darling Downs agricultural
region. It is also a secondary source of water supply for Toowoomba (population 100 000, the
second largest city in the Murray-Darling catchment after Canberra, and the largest inland city in
Queensland) and surrounding communities, supplementing surface water stores, and is a strategic
resource for future expansion of the peri-urban area (Willey, 2003). Average annual rainfall varies
from over 900 mm in Toowoomba to 650 mm in the upper Condamine River valley, and is summer
dominant. A strong rain shadow exists on the Eastern Darling Downs, in the lee of the Great
Dividing Range. Irrigation water is largely sourced from basalt aquifers from a range of depths,
shallow to ~ 150 m. The extent of irrigated land is greater in Kings Creek sub-catchment (20%)
compared to Hodgson Creek sub-catchment (<10%). Urban expansion has placed pressure on both
premium agricultural lands and water resources.

Figure 2: SPOT5 image of the study area (2.5m resolution; 15 Sept. 2006), showing the Great
Dividing Range in the northeast, the Condamine River in the southwest and the largely
agricultural landscape. Area shown is approximately 80 km east-west and 60 km north-south
(Source: Geoscience Australia National Mapping Group).

4
GEOLOGY

The regional geology has been mapped and described by Wells and O’Brien (1994a and 1994b) and
is also summarised by Whitaker and Green (1978). Early work on the volcanic rocks was
documented by Stevens (1968), and more detailed geologic information and mapping in the study
area provided by Willey (1992). The regional geomorphology was interpreted by Ollier and
Harworth (1994). The area is dominated by two main rock units: a near-horizontal to gently dipping
Middle Jurassic sedimentary sequence of the Moreton-Clarence Basin, part of the Great Artesian
Basin, and the mid-Tertiary Main Range Volcanics (MRV), extending up to 50 km westward from
Toowoomba. These strata unconformably overlie a concealed basement of Palaeozoic rocks, the
New England Orogen.

Middle Jurassic sedimentary sequence

Jurassic strata crop out in the lower parts of the sub-catchments near the Condamine River; these are
dominantly Walloon Coal Measures (shale, siltstone, sandstone, coal, mudstone and limestone). The
Marburg Formation (sandstone, siltstone, shale, grits, conglomerate and oolitic ironstone)
conformably underlies the finer-grained Walloon sediments although, in the study area, it only sub-
crops in the lower part of Kings Creek sub-catchment4. The Marburg Formation conformably
overlies the medium- to coarse-grained, quartz-rich Triassic-Jurassic Helidon Sandstone in the
northern part of the study area. Recent water bore drilling in Toowoomba (RN119852) penetrated
coarse sandstone of the Helidon Sandstone at 562 – 620 m (pers. comm., A. McKay, DNRW, May
2007). Faulting of Jurassic strata is revealed in coal seams in the gently dipping Walloon Coal
Measures. In the southern part of the study area, the Marburg Formation unconformably overlies
basement rocks, namely the Carboniferous Texas Beds that outcrop on the south-western bank of the
Condamine River, immediately outside the study area. The Marburg Formation and underlying strata
and basement rocks are not included in the 3D model because they are either absent or poorly
described in the water bore logs used to build the model.

Mid-Tertiary Main Range Volcanics (MRV)

MRV basalts and derived soils make up over 80% of the study area. The MRV olivine basalts and
subordinate basaltic tuffs are well exposed along the Great Dividing Range. Numerous eruptive
centres are scattered across the rolling uplands (Figure 3); these comprise former lava pools that
once infilled tuffaceous crater cones; the latter have commonly been subjected to post-eruptive
erosion (Willey, 1995).

The MRV lie unconformably on Moreton-Clarence Basin sediments. The unconformity occurs at
around 440 – 460 m AHD, indicating a nominal basalt thickness ~ 200 m, with a maximum
thickness of approximately 280 m in the elevated uplands (Kalma, 1995). Basalts around
Toowoomba were investigated in detail by Stevens (1969). Numerous small overlapping basaltic
lava flows are represented. The prominent edges of the flows have been mapped by Willey (1992)
and at least 16 flows documented (Willey, 2003); individual flows extend up to 80 km2. Vertical
columnar joints, and vesicular, amygdaloidal or “bubbly” flow tops produced by successive
eruptions, significantly impact on hydrogeologic processes. Silcrete and ferricrete is locally
associated with MRV basalt, particularly around Toowoomba. Land surfaces and soils formed on
basalt flows before, during and after discrete extrusive events. Sloping MRV landscapes are mantled
with shallow soils and basaltic colluvium. The heterogeneity of more porous and permeable MRV
units impacts greatly on the local groundwater system (this is discussed further below).

4
Descriptions of sediments in logs for drill-hole in the study area do not readily distinguish between Walloon and Marburg
formations. Minor outcrops of the Marburg Formation mapped at the southern edge of Kings Creek sub-catchment are not
penetrated by reliable bore logs so this unit is not differentiated in the 3D model.

5
Figure 3: A prominent hill composed of Main Range Volcanics, representing a former eruptive
centre, and a QDNRW monitoring bore in the Hodgson Creek sub-catchment.

Cainozoic alluvium and soils

Valleys incised into both the MRV units and basal Mesozoic sediments are infilled with Cainozoic
colluvium and alluvium. Valleys are relatively steep-sided in up-catchment areas, grading to mature
broad alluvial valleys downstream near the Condamine River. Extensive sheet, rill and gully erosion
following agricultural development enhanced valley infilling in recent times. Colluvium derived
from the MRV typically blankets outcrops of the Walloon Coal Measures on gentle lower slopes.
Alluvium forms floodplains and river terraces in lower parts of the sub-catchments but functions as
only a thin or discontinuous aquifer within the study area. Condamine River alluvium and sediments
comprise a complex system of fluvial channel deposits and floodplain accumulations sourced from
basalts and Jurassic sediments as well as regional basement rocks exposed in the south (Willey,
1992); sand and gravel beneath clays in the Condamine River alluvium function as aquifers.

Two general soil types are represented: red soils of the Toowoomba Plateau and black-brown soils
of the eastern Darling Downs (Thompson and Beckman, 1959; Willey, 2003). Both of these soils
are highly fertile. Macnish et al. (1987) argued that lateritisation or bauxitisation of weathered MRV
basalt was less important in the development of the red ferrosoils than derivation from widespread
inter-basaltic tephras of similar composition. A strong correlation is observed between pre-exposure
weathering of the substrate and soil chemistry, i.e., soils with very "old" chemistry forming in the
contemporary environment, because the parent material was heavily weathered prior to being buried
by subsequent basalt flows, and then re-exposed. Accordingly, these soils forming in situ today
cannot be regarded as palaeosols; although true palaeosols occur around Toowoomba, they are not
common (pers. comm. A. Biggs, DNRW, May 2007).

Additional soil types, namely skeletal soils of the rocky escarpment and sandy-gravely soils
developed on alluvium, have relatively low fertility. The approximately 60 mm difference in rainfall
between the east and west of Toowoomba is reflected in associated differences in soil chemistry and

6
characteristics (pers. comm. A. Biggs, DNRW, May 2007). Despite contrasting pedogenesis, the
soils are typically deeply weathered and highly permeable throughout the region, a factor relevant to
hydrogeologic functioning.

Landforms
The landforms are closely related to the lithology and geologic history of the region. Many
geological units display characteristic landform features which influence hydrology and land use
(Table 1).

Table 1: Geology and landforms (after Biggs, 2000).

GEOLOGICAL UNIT DOMINANT LANDFORM
Quaternary alluvium Level to very gently inclined plains with levees, stream
channels, ox-bows, river terraces and prior streams,
commonly complex systems; associated Quaternary
sandy soils derived from erosion and re-distribution
can be extensive.
MRV basalts Complex landform patterns. In the east and north are
very steep mountains and low hills associated with the
Great Dividing Range. A large plateau is present on
the top of the range around Toowoomba. Undulating
rises and rolling low hills with some steep hills and
gently undulating plains. A dissected landscape has
resulted from stream incision associated with uplift
and eustatic change.
Walloon Coal Measures Gently undulating plains and rises to undulating to
(fine grained sediments) steep, low hills and rises; boundary between some
low slope areas with associated alluvia is often
indistinct.
Marburg Subgroup Undulating to rolling rises and plains with some steep
(coarse grained sediments) low hills, generally steeper than the terrain associated
with finer sediments.

7
Hydrogeology
GROUNDWATER FLOW SYSTEMS
The following groundwater flow systems have been described within the Hodgson and Kings Creek
sub-catchments (Biggs and Power, 2003):

• Cainozoic basaltic rocks: intermediate flow systems in basaltic rocks (MRV) on rises and
low hills;

• Cainozoic basaltic rocks: local flow systems in basaltic rocks (MRV) on plateaux;

• Cainozoic unconsolidated sediments intermediate flow systems in unconsolidated

sediments on alluvial plains; and

• Terrestrial and carbonaceous sedimentary rocks: regional flow systems in terrestrial

sedimentary rocks (Jurassic strata) on rises and low hills.

Additionally, regional flow systems associated with coarse-grained Marburg and Helidon strata of
the Moreton-Clarence Basin underlie the Condamine River alluvial plains and occur at substantial
depth beneath the Main Range Volcanics. For example, recent drilling in Toowoomba (RN119852)
encountered groundwater in the Helidon Sandstone of the Great Artesian Basin, at 562-658 m; the
SWL is around 250 m depth (Table 2).

Table 2: Log for RN119852 borehole in Toowoomba, drilled in 2007.

STRATA STRATA
TOP (M) BASE (M) STRATA DESCRIPTION UNIT NAME
0 2 top soil soil
2 48 basalt Main Range Volcanics
48 85 puggy clay and shale Walloon Coal Measures
85 113 sandy shale
113 151 grey shale
151 160 coarse sand and white clay
160 188 coal bands and grey shale
188 255 grey shale
255 562 sandstone and clay Marburg Formation
562 620 coarse sandstone Helidon Sandstone
620 658 clay bands, very fine sandstone

GROUNDWATER PROCESSES

Within the Cainozoic MRV systems, vesicular basalts and associated red weathering products/soils
are important horizontal water-bearing zones (van der Zee and Macnish, 1979). Irrigation water is
largely sourced from basalt aquifers, e.g., Figure 4. The presence of numerous small, overlapping
lava flows and columnar joints associated with former eruptive centres (e.g., Figure 3) creates
vertical and horizontal hydraulic discontinuities compared to the groundwater system in the
underlying Walloon Coal Measures (Kalma, 1995). The upper weathered basalt unit at the top of the
MRV is important for groundwater recharge during heavy precipitation events, with a delay period
of about a month between rainfall and recharge. This major recharge zone feeds creekbeds incised
into the permeable pallid base of the weathered unit, even during dry periods (Willey, 2003). The
dominant stream alluvium in upper valleys is cracking clay; baseflow here relates to springs

8
discharging at the edges of streambed aquitards. The main features of the hydrological system in the
Hodgson and Kings Creek sub-catchments are shown in Figure 5.

Groundwater recharge occurs following rainfall rates of 100–150 mm/month, with 3–6 days delay
between rainfall events and peak groundwater levels in the upper catchments (Willey, 2003), and lag
times of the order of 4–7 months for bores in basalt beneath deep clay alluvium (Silburn et al.,
2006). Unlike drier regions elsewhere in the Murray-Darling catchment, the unsaturated zone here
does not appear to have a large historic moisture deficit that needs to be filled for deep drainage to
become recharge. Thus, deep drainage recharges aquifers within several months and excess
groundwater is accordingly transferred down-gradient to the land surface or streams within a year
(Cresswell et al., 2006). Surface permeability has decreased across the Toowoomba Plateau through
urban expansion, for example sealed roads and rooftops now intercept rainfall and reduce recharge.

Figure 4: Sprinkler irrigation drawing on groundwater, Hodgson Creek sub-catchment.

Watertable depths in monitoring bores range from <2 to 50 m below ground (Natural Surface = NS).
In Hodgson Creek sub-catchment the watertable height decreases from 517 m to 426 m AHD over a
distance of ~25 km, a hydraulic gradient of 0.003. Groundwater salinity increases progressively
down-gradient, indicating continuity of the flow path (Cresswell et al., 2006).

Hodgson Creek is a ‘gaining stream’, receiving discharging groundwater as streamflow, although

dry periods may see a reversal, with streambeds locally recharging underlying strata. The presence
of baseflow is strongly influenced by the depth of stream incision. Baseflow no longer occurs in
several streams, for example, in Hodgson Creek downstream of Cambooya, below the area of
greatest groundwater use (Reid et al., 2007). Relatively stable groundwater levels occur at shallow
depths throughout most of the alluvium (< 5m). Depth of stream incision and groundwater use for
irrigation largely control the presence or absence of baseflow, e.g., groundwater is generally not far
beneath the creekbed in non-irrigation areas. The system is strongly predisposed to groundwater–
surface water interaction. The Kings Creek system shares similar characteristics (Huxley 1982).

9
Compared to the southern parts of the Murray-Darling Basin, salinisation is a minor problem.
Typically in the upper Condamine catchment, saline outbreaks occur where groundwater flow is
restricted at:

a. Impermeable interbeds in the MRV or where there are permeability contrasts between
different basalt layers;

b. The boundary between basalt and underlying Walloon Coal Measures in lower parts of sub-
catchments; and

c. Creeks incised into alluvial aquifers where the hydraulic conductivity of a valley is very
low, causing groundwater levels to rise (Kalma, 1995).

A notable saline outbreak occurs in the upper part of Umbiram Creek where a constriction in the
MRV inhibits groundwater flow.

Evapotranspiration

Rainfall recharge Bore abstraction Recharge infiltration through

eroded volcanic cone (rainfall)

Surface inflow
via creek

Watertable Perched watertable

Runoff
atop shale seam
Vesicular basalt flow top
(water-bearing) Spring

Regional groundwater inflow Regional groundwater

in Jurassic strata flow in Jurassic strata

Surface outflow to
Condamine River

Legend

Ground surface Main Range Volcanics (Tertiary) Walloon Coal Measures (Jurassic)

Irrigation area * Weathered basalt, laterite and soil * Fine-grained sandstone and shale

Natural spring * Basalt colluvium on slopes * Coal seam

Water flow lines * Fresh basalt and minor tuff Alluvium (Recent)

Figure 5: Conceptual model of groundwater systems in the Hodgson and Kings Creek district.
The local to intermediate groundwater flow systems in the Main Range Volcanics (MRV) supply
most of the agricultural and domestic water users in the region. Yields are highly variable in the
MRV due to the complex and irregular distribution and interaction of subvertical and
subhorizontal flow-paths, such as columnar jointing, vesicular zones and other structural
discontinuities. Groundwater interaction occurs readily across the boundary between the MRV
and alluvial sediments, but regional flow systems are dominant in the underlying Jurassic strata
(diagram modified after Willey, 2003).

Aspects of the hydrology, hydrochemistry and salinity of the area, focussing on Hodgson Creek, are
outlined by Dutta and Silburn (2005), McNeil and Silburn (2005), Silburn et al. (2005, 2006) and

10
Cresswell et al. (2006). Recharge in Hodgson Creek sub-catchment is estimated at around 14
mm/year; groundwater pumping rates (for crops, stock and domestic use) vary depending upon
locality, but are in the order of 7–8 mm/year, with the difference contributing to baseflow in streams
(Dutta and Silburn, 2005; Silburn et al., 2006). High recharge rates are likely to be promoted in the
vicinity of the large number of MRV eruptive centres where columnar jointing is common.

The following observations by Cresswell et al. (2006) are based on hydrochemical data for over 470
bores in Hodgson Creek sub-catchment. Groundwater salinity increases down-gradient, ranging from
300 to 1300 mg/l Total Dissolved Ions (TDI). Most groundwaters are of Mg-Na-HCO3-Cl type,
reflecting the influence of weathering of the basaltic host rocks. More Na-Cl-rich groundwaters in
the upper slopes may relate to evaporative concentration of infiltrated rainfall stored in the upper
weathered unit at the top of the catchments. Stable isotope data indicate that the aquifers have been
recharged dominantly from slightly evaporated rainwater (Cresswell et al., 2006). Radiocarbon (14C)
determinations indicate that most groundwaters are modern (<100 years). Very low 14C
concentrations (indicative age ~20 000 ka) from one deep bore may represent regional groundwaters
in Walloon Coal Measures beneath the MRV (Cresswell et al., 2006). Streamwater samples that are
dominated by groundwater input are more saline (EC ~1500-1800 µS/cm2) than those dominated by
overland flow. Average stream salinity is around EC 300 µS/cm; salt export from the near-surface
system is flushed from the sub-catchments during a few large annual flood events.

11
3D Model Development
PRELIMINARY STAGES
Initial phases of the project included field reconnaissance of the Hodgson – Kings Creek study area
and a workshop in Toowoomba with hydroscientists from QDNRW, Geoscience Australia, CSIRO
and Flinders University. Preliminary data review for the feasibility assessment involved the
following datasets, information sources and tasks:

• A 25 m resolution digital elevation model (DEM);

• Surface geology map;

• Drainage and infrastructure map;

• Landsat TM data;

• Excel spreadsheets of water bore logs: Registered Numbers, coordinates, elevations,

stratigraphic logs, bore construction details, groundwater monitoring data, etc;

• Assessment of surveyed bore collar elevation measurements with respect to the DEM;

• Creation of a 3D GIS environment; and

• Comprehensive literature review.

OVERVIEW OF THE 3D HYDROGEOLOGICAL MODEL

The model developed in Earth Decision Suite (Gocad) provides a functional 3D display of the
Hodgson and Kings Creek sub-catchments. Methods used for this project are only an example
of the tools available to 3D modellers, and it must be noted that model outputs are wholly
dependent on the nature and quality of input data. The on-screen display can be viewed from
any user-defined orientation, as the model can easily be moved, rotated and zoomed with simple
commands. The following summary outlines the main components of the model:
Topographic features
1) Digital Elevation Model (DEM) showing the topography of the Hodgson Creek and Kings
Creek sub-catchments. Created from a detailed 25 m resolution model and interlaced with
coarse-scale DEM (300 m resolution) to cover discontinuities and expand the model beyond
the catchment boundaries;

2) Surface hydrological features such as stream channels, water bodies and the location of
(sparse) natural springs;

3) The boundaries of the Hodgson and Kings Creek catchments; and

4) The main cultural features such as towns, roads and railways;

12
Water bores
1) Surface location of all water bores in the Hodgson and Kings Creek catchments selected
for use in this modelling exercise (core dataset); and,

2) Vertical down-hole representations of borehole depths, with marker zones down each
hole trace showing the upper and lower boundaries of each stratigraphic unit and individual
water-bearing zones.

Geological units
1) The unconformable geological boundary between the uppermost surface of the Walloon
Coal Measures and the lower Main Range Volcanics (MRV). This important geological
horizon is represented as a surface across the entire region, i.e., a former planation surface
on the Clarence-Moreton Basin, developed prior to extrusion of the basalts;

2) The interpreted base of the Walloon Coal Measures. This surface is a theoretical boundary
only as most water bores in the area do not penetrate the base of the Walloon Coal
Measures. The distribution of bores that penetrate to underlying strata is too sparse to
generate additional surfaces. Such bores are included wholly in the ‘Walloon Coal
Measures’ layer. This lowermost surface forms the base of the geological strata in the 3D
model5;

3) The layer of Quaternary sediment cover and soil which blankets most of the region is too
thin to be accurately represented as a continuous surface and has thus not been modelled;
and

4) Surface geological boundaries showing outcrop zones of the Quaternary sediments and
MRV across the catchments, as depicted on the 1:500,000 geological map (Wells and
O’Brien, 1994), Figure 8.

Geophysical datasets
1) Gravity data. A map of gravity data obtained from the on-line Geophysical Archive and
Data Delivery System (GADDS) coincident with the catchment boundaries;

2) Magnetics data. Airborne magnetics data acquired by the Queensland Geological Survey
and stored in GADDS, partially coincident with the catchment boundaries; and

3) Remote Sensing data. A SPOT5 satellite image of the Hodgson and Kings Creek region
showing land use distribution and many natural and cultural features of interest (Figure 2).

5
The lowermost surface provides a base for the 3D model; it does not represent the lower limit of potential aquifers in the region
nor a stratigraphic contact, but rather a surface that encompasses the bottom of bores used for model-building.

13
HYDROSTRATIGRAPHIC DATA ANALYSIS
Construction of the 3D model relied on water bore data for Hodgson and Kings Creek sub-
catchments provided to GA by the QDNRW. These data required substantial modification and
rationalisation before they could be used in the modelling software. This work involved scrutiny of
approximately 2200 borehole logs to select a core dataset (~1100 logs in total) which satisfied the
following criteria:

a. Adequate geological information for modelling;

b. Representation of the main water-bearing units (aquifers);

c. Broad geographic distribution of boreholes across each catchment; and

d. Availability of accurate coordinate and elevation data, e.g., GPS or survey

measurements.

The distribution of bores used for model generation is shown in Figures 6–8, and bore data is
summarised in Appendix 1.

Figure 6: Distribution of bores across the Hodgson and Kings Creek sub-catchments (blue
spheres) used for 3D model generation. Plan view shown on the DEM surface. North to top of
page, with sub-catchment boundaries shown as red lines. The combined sub-catchment area is
approximately 70 km east-west and 45 km north-south.

14
Figure 7: Bore distribution and drainage lines. Plan view on the DEM; bores shown as dark
blue squares, streams in blue, sub-catchment boundaries as thick red lines and roads as thin red
lines.

Figure 8: Detailed example of borehole distribution in the Gocad model (area about 50 km2),
Hodgson Creek sub-catchment (RN = registered number).

15
The original bore log dataset incorporated approximately 35 rock type codes; these were
subsequently rationalised to 15 main lithological types for modelling purposes (Appendix 1). Most
rock units also have water-bearing zones (designated ‘WB’ in the lithological code index). For the
Hodgson Creek catchment 323 bores were selected, whereas 761 bores were used for the Kings
Creek district. For the purposes of constructing the 3D model from these data, all rock codes were
further simplified into three main stratigraphic units:

• Quaternary Alluvium, e.g., gravel, sand and silt6;

• Tertiary Main Range Volcanics (MRV), dominated by basalt (weathered, massive or

vesicular) and volcanic ash;

• Jurassic Walloon Coal Measures, e.g., coal, sandstone and shale, representing the bulk of
Clarence-Moreton strata/aquifers within the study area.7

Figure 9: Mapped surface contact (‘surface curves’) between Quaternary alluvium and MRV
shown as green lines. Streams in blue and catchment boundaries in red on the DEM. The
geologic contact shown here is derived from the 1:500 000 geological map of the region (Wells
and O’Brien, 1994).

6
Quaternary alluvia functioning as aquifers proved to be limited in distribution and thickness, typically < 4 m.
7
Bore RN119852, sunk in Toowoomba in March 2007, to 658 m depth immediately outside the study area, is too deep to be
included in a model dominated by ~70 m deep bores (average depth). Surfaces for the base of the Marburg Formation – top of the
Helidon Sandstone were not generated in the model.

16
The relevant borehole data were compiled into Excel spreadsheets for use in the modelling software.
These modified datasets comprised:

a. Bore number;

b. Geographic coordinates (UTM easting and northing);

c. Elevation (Australian Height Datum in metres above sea-level);

d. Total borehole depth;

e. Downhole geological composition (based on rationalised lithological codes and main

stratigraphic groups); and

f. Aquifer details, e.g., downhole water-bearing intervals, standing water levels (SWL), water
yield and potability8.

3D MODELLING USING GOCAD

Gocad is an advanced 3D modelling and visualisation package used extensively in the mining and
petroleum industry and can be described as a 3D Geographic Information System (3D GIS); it is
used extensively by Geoscience Australia. The program is particularly well suited to constructing,
visualising and manipulating 3D models based on actual geological, geophysical and
hydrogeological datasets, e.g., borehole logs. The software also supports 3D vector and raster
overlays such as geophysical or remote sensing images, and a variety of spatial queries, e.g.,
volumes and distances. These properties make it particularly well suited to building and viewing the
conceptual 3D groundwater model developed for this project.

IMPORTING DATA INTO GOCAD

Borehole data
Data import was carried out using a combination of native Gocad import routines and proprietary
import filters created by Mira Geoscience. As Geoscience Australia has only one node-locked
licence of the Mira Mining Utilities application, most data were imported using basic Gocad filters.
However, this created some conflict errors in borehole orientations; for example, modelling is
typically undertaken with the depth-positive axis upwards, whereas the default option when
modelling in Gocad has the depth positive axis downwards (depth below the natural ground surface
increasing downwards). Use of the proprietary Mira import filters is therefore recommended, if
available, as they readily correct such conflict errors.

Importing the required borehole data was relatively straightforward as all holes were assumed
vertical. However, in the initial model-building stages there was uncertainty about which geological
contacts would be modelled, and how this would be achieved in Gocad, e.g., modelling stratigraphic
boundaries as either upper or lower surfaces. It was eventually decided that modelling the MRV-
Walloon unit contact as the top of the Walloon unit would be the most efficient way to achieve the
desired model using the structural modelling workflow (SMW) routine in Gocad. The uppermost

8
These datasets are not available for every borehole; unfortunately, they have not been systematically recorded in all drillers’ logs
in the QDNRW archive.

17
Quaternary alluvium layer, logged in most boreholes as a thin unconsolidated veneer, proved too
problematic to model accurately, especially in the undulating north-eastern part of the catchment.

Lithological marker horizons (rock unit zones) were initially defined in each borehole using the three
main stratigraphic units: Quaternary alluvium, MRV and Walloon Coal Measures. This stratigraphic
column was defined in Gocad and applied to all boreholes to standardise the modelling process.
Software functionality using the stratigraphic column approach also allows contact surfaces to
remove unrealistic geological horizons.

Correlating borehole collars to the topographic surface created by the DEM proved difficult and time
consuming. A wizard routine in Gocad was used to shift individual hole collars vertically (either up
or down as needed), to ensure they were properly located on the topographic surface. This is an
important step, especially where thin surface layers such as Quaternary alluvium unit are present.
Borehole collars placed incorrectly in vertical space may position contacts deeper than they actually
are, or even arbitrarily remove them completely if the collar is located above the DEM surface. For
each borehole, markers were assigned for the geological contacts and for each water-bearing unit.

Topographic data (Digital Elevation Model)

The proprietary Mira utilities routine allows the direct import of ESRI ASCII grid files to a 2D voxet
(modelling box). This object can easily be resampled at lower resolution, and modelled surfaces
generated directly. A 25 m ESRI ASCII grid in UTM format of the catchment area was imported
into Gocad and resampled to 50 m resolution. This was then used to create the topographic DEM
surface for the model (Figure 10).

Figure 10: Oblique view of part of the DEM showing the fine triangular mesh generated in
Gocad from 25 m resolution data; a x 10 vertical exaggeration is shown; sub-catchment
boundaries/watersheds shown in red, roads in grey; orange areas depict outcropping Walloon
Coal Measures.

Unfortunately, the available 25 m DEM coverage had previously been cropped along the north-
eastern edge of the sub-catchments, creating an uneven edge in the modelled topographic surface
particularly along the Great Dividing Range. To compensate for these irregularities, and also to

18
match the dimensional extents of the structural and stratigraphic grid models (created using a
straight-edged volume of interest, VOI), a relatively low resolution DEM (approximately 300 m-
scale) of the adjacent area was interlaced with the 50m DEM (Figure 11). Although the joining seam
between these two DEMs is not a perfect match, the surface properties of the final grid are
interpolated only within the catchment boundaries. Thus, the join does not directly affect the broad-
scale geologic/hydrogeologic model. To overcome this problem for future projects, high-resolution
DEM data should have geographic extents slightly greater than the area of interest, i.e., not cropped
precisely to the catchment boundaries.

Figure 11: Oblique view of part of the DEM showing the contrasting resolution of DEM data
that allowed extension of the model beyond the catchment boundaries, as described in text.

Remotely sensed images

Remotely sensed data such as satellite imagery can be imported into Gocad as world files using
standard routines in the Mira utilities application. This allows for correct geographic registration of
imagery. World file images can also be correctly registered using the native “resize voxet with
vectors” command in Gocad.

ESRI shapefiles
Using either the native Gocad shapefile importer or Mira utilities application, useful ESRI shapefile
data were imported and edited so that surface curves mapped the stratigraphic contacts as unit bases.

19
Geophysical data
Colour contoured gravity and airborne magnetics data partially coincident with the catchment area
were obtained from Geoscience Australia’s Geophysical Archive Data Delivery System (GADDS).
These datasets were imported into Gocad using Mira utilities import tools, to provide optional new
data layers to the model. The new data layers were added to the model from ERMapper (image
processing software package) grid files via conversion to Geosoft ASCII GXF file format. Spot
location gravity data were also downloaded from GADDS and imported into the model using Gocad
column file import utilities. Unfortunately, due to their relatively coarse resolution, the geophysical
datasets did not aid construction of the 3D model.

SURFACE MODELLING

Structural modelling using Gocad workflows

The structural modelling workflow (SMW) routine in Gocad is designed as a specific aid for
constructing 3D common-earth models. Workflows guide novice users step-by-step through the
modelling procedure, and also allow more advanced users the benefit of decision-driven modelling
while still retaining full control of the model-building environment. Workflow routines also fully
document each modelling step, permitting the user to add notes and snapshots whenever necessary.
The SMW reporting methodology was fully utilised during this project and can be referred to for this
stage of model construction (Appendix 2).

For boreholes which terminate in the MRV (thus lacking a marker for the top of the Walloon unit),
the main geological surface (uppermost Walloon contact) was modelled to pass smoothly below the
bottom of the borehole. In the absence of less arbitrary data or additional information about the
depth of the MRV for these boreholes, this represents the most appropriate modelling procedure.

Final surface modelling

Local editing of surfaces constructed using SMW’s is commonly required in most Gocad models.
This section discusses some important observations relating to the MRV-Walloon boundary surface,
and describes how major problems have been resolved.

To examine the surface constructed using the workflow process, the vertical thickness from the
DEM surface to the top of Walloon contact was computed using basic Gocad commands. The final
output was then colour contoured (Figure 12).

The uncoloured (white) parts of Figure 12 shown along the south-western margin and near the
eastern catchment boundary represent either:

• Upper Walloon borehole markers located at the landscape surface, indicating outcropping
Walloon Coal Measures and no alluvial cover (as in the south-western margin of the sub-
catchments);

• The upper Walloon boundary is located close enough to the surface (minimal cover
sequences) that local topographic variations smoothly fit the borehole data and project
above the topographic surface. This phenomenon causes the anomalous uncoloured region
in the topographically rugged area along the north-eastern boundary, and also the flat low-
lying area in the south west, or;

20
 • The edges of the modelled surface project above the DEM surface beyond the edges of the
model area, but at a relatively steep slope, such that the surface within the catchment area is
affected. This is best observed to the immediate north of Hodgson Creek catchment.

Figure 12: Contour plot of total vertical thickness (in metres) from the natural topographic
surface, as modelled by the DEM, to the top of the Walloon Coal Measures. Relevant borehole
markers are displayed as small red squares, and the boundaries of the Hodgson and Kings
Creek sub-catchments as red lines. Regions where the uppermost Walloon Coal Measures are
outcropping or virtually exposed/subcropping at the modelled topographic surface in the south-
western area are shown as white areas, explained in the text below.

Furthermore, although local windows through the Quaternary cover are likely, sharp changes in
contact elevation significantly affect the geometry and are considered unrealistic. The extensive
region to the north of the catchment which was originally modelled as projecting above the
topographic surface provides a good example. The local boreholes have scant Walloon Coal
Measures in proximity to the topographic surface, indicating that the modelled geometry was
unrealistic. On closer inspection of the primary data, an error was detected in the stratigraphic
column for a single borehole (basaltic rocks erroneously assigned as Walloon Coal Measures), thus
causing the modelling error. When this borehole error was corrected and the surface remodelled,
simply by running the SMW once more with amended input data, the new iteration provided a
significant improvement in the model. In addition, the same area was locally edited so that the
uppermost Walloon boundary was consistently below the topographic surface (Figure 13).

21
 Toowoomba
Great Dividing Range

Kings Creek catchment

Hodgson Creek catchment

Figure 13: Tilted isometric view looking towards the north-east (with 10-times vertical
exaggeration applied). This Gocad image shows the disposition of the DEM topographic
surface (green, vertically stretched) and the uppermost surface of the Walloon Coal Measures
(pink). Note the irregular-shaped exposures of the Walloon sedimentary unit exposed in the
lower reaches of the sub-catchments, and the relatively flat-lying Walloon strata continuing
eastward beneath thick MRV units of the Great Dividing Range. Stream channels shown as blue
lines and catchment boundaries as red lines.

The disposition of the uppermost surface of the Walloon Coal Measures generated from the bore
data is consistent with exposed Jurassic strata of the Clarence-Moreton Basin mapped by Wells and
O’Brien (1994) near the Condamine River in the south-west, and also north-eastward beyond the
present study area at the base of the escarpment (east of the Great Dividing Range).

To overcome most of the modelling problems outlined above, additional data from boreholes located
outside of the area of interest (beyond the catchment boundaries) enabled a more accurate and
realistic contact surface geometry.

The lowermost surface (base of model) was created by replicating the upper Walloon Coal Measures
surface and translating this vertically downwards to a depth just slightly below the deepest borehole
in the study area (Figures 14–16). Although this only represents a theoretical base, it allows the
model volume boundary to be created in the absence of actual data (recall that none of the boreholes
used to create the model could confidently be shown to have penetrated the base of the Walloon
Coal Measures)9. Relevant Gocad procedures are described further in Appendix 2. The
heterogeneous distribution of water-bearing zones is shown in Figures 17–19.

9
Refer to the previous Data Import section for discussion of the method used to edit the DEM and ensure the correct geographic
extents for the model.

22
Figure 14: Oblique north-north-east view showing the semi-transparent DEM surface (blue-
mauve), base of MRV-top of Walloon Coal Measures contact (pink), and the base of model
(orange); sub-catchment boundaries in red.

Figure 15: A random side view of the model showing the topographic relief as red outlines
which largely represents the top of the MRV. Boreholes are shown in dark blue, the MRV-
Walloon contact in yellow, and the base of the model in pink. The top of each borehole
corresponds with the landscape surface at each bore location, as conveyed in Figures 14 and 16
using a transparent DEM. The base of the model does not represent the Walloon-Marburg
contact but, rather, the bottom of bore holes in the study area.

23
Figure 16: Oblique west-north-west view across the two sub-catchments showing the DEM
surface (50% transparency), boreholes (dark blue), and base of the model (orange surface), the
latter with constructional triangles revealed.

Figure 17: A random side view of the model showing a vertically exaggerated semi-transparent
DEM, Hodgson boreholes (dark blue), Kings boreholes (red), MRV aquifers (purple), and
Walloon aquifers (green); dry bores are present in both sub-catchments.

24
Figure 18: Side view of part of the model for the Hodgson Creek sub-catchment looking beneath
the DEM (grey line-work showing triangular construction mesh); boreholes (dark blue) and
MRV aquifers (purple). This view reveals the extreme heterogeneity of water-bearing zones
within the MRV, within a relatively small area (several kilometres). Several of the bores shown
here penetrate numerous water-cuts of variable thickness, ranging from relatively narrow zones
less than 2 metres up to several tens of metres thick. There are also several boreholes which
have failed to intersect any water-bearing zones and are dry. This heterogeneity is caused by the
very irregular nature of basaltic discontinuities (both sub-horizontal and sub-vertical) within the
Main Range Volcanics which play an important role in controlling the nature and distribution of
water-bearing zones.

25
Figure 19: Side view of part of the model for the Kings Creek sub-catchment looking beneath the
DEM (grey line-work). Boreholes are shown in red, MRV aquifers are purple zones on the
drillhole stems, and Walloon aquifers are green zones. This view again reinforces the extremely
heterogeneous nature of the sub-surface distribution of water-bearing zones within the
stratigraphic units of the Condamine headwater sub-catchments.

26
3D Model Enhancement and Further Analysis
X3D REPRESENTATION OF THE GOCAD MODEL
The Gocad Model has been converted to an X3D (Extensible 3D) interface to provide the
Condamine Alliance and other stakeholders with web-based access, in the absence of the specialised
software used to generate the 3D model. The X3D interface – and its precursor, Virtual Reality
Modelling Language (VRML) – can display large volumes of complex 3D geological and
hydrogeological information (data volumes up to 22 MB) via the internet and World Wide Web.
X3D is still under development; it is an XML (eXtensible Markup Language) syntax for VRML,
with improvements, and is interoperable with other XML formats. X3D is an open and free ISO
standard for 3D graphics on the internet, and is supported by the Web3D Consortium. This enables
management and exchange of large amounts of data and a wide range of geometric features to
represent geoscientific features. X3D format enables substantial file compression, compared to
VRML, and allows easy interaction with other XML formats. The interfaces are user-friendly and
provide users with a high level of interaction with complex information (Beard, 2005; Beard et al.
2005). Geoscience Australia has published complex 3D geologic models on the web since 200110.
X3D will be one of the tools that Geoscience Australia will continue to use to deliver 3D content
over the internet in the coming years, in conjunction with other tools such as 3D PDF and Google
Earth applications.

THE GROUNDWATER SYSTEM OF THE CONDAMINE HEADWATER CATCHMENTS

For the purposes of building the 3D model, the respective geological units were regarded as
essentially flat-lying with sub-horizontal stratigraphic contacts, an inference which holds true for
most sedimentary strata of the Clarence-Moreton Basin (Wells and O’Brien, 1994). As previously
mentioned, this conceptualisation was borne out when constructing the MRV-Walloon contact from
the Hodgson and Kings Creek bore logs.

On a regional scale, the basaltic rocks of the MRV erupted across a relatively flat planation surface
over the Jurassic strata (Ollier and Haworth, 1994). However, the MRV were not formed as thick
and laterally continuous flood lava flows – as originally perceived – but rather as multiple eruptions
from discrete volcanic cones scattered across the region (Free, 1989), e.g., Figure 3. These form
distinctive topographic features across the present land surface, a reflection of differential erosion
between massive basalt (former lava pools in vents) and surrounding pyroclastic deposits and
volcanic breccias. Additional to the numerous eruptive centres and fissures, a chain of large shield
volcanoes trend north-west to south-east through the region, including the Bunya Mountain Volcano
north-west of Toowoomba and Tweed Volcano to the south-east in coastal NSW (Ollier and
Harworth, 1994). The study area encompasses adjacent parts of two such shield volcanoes; Hodgson
Creek is a radial drainage line of the erstwhile Toowoomba Volcano, and Kings Creek is a gutter
drainage-line formed at the junction of two coalescing volcanoes, the Toowoomba Volcano and the
Main Range Volcano to the south-east. These original large volcanoes are now severely eroded, with
their main craters further degraded by westward retreat of the Great Escarpment (Ollier and
Haworth, 1994). It is possible that significant volumes of volcanic material (both lava flows and
pyroclastic ejecta) were erupted concurrently across a significant area and period of time (Free,
1989).

10
Geoscience Australia Web-based 3D geologic models: http://www.ga.gov.au/map/web3d.

27
Around Toowoomba much of the basaltic rock is relatively compact and impervious; however, the
weathered zone and intense jointing – particularly vertical joints – are highly significant features for
groundwater processes (Armstrong, 1974). Fractured rocks of the MRV produce the highest
groundwater yields from individual water bores, for example: max. 35 L/sec, compared to 30 L/sec
from unconsolidated alluvial sediments, 12 L/sec from fractured Walloon Coal measures and 5 L/sec
from the Marburg Formation Sandstones. According to Carberry and Walker (1993) all aquifers in
the region have low average minimum yields of ≤1 L/sec. Notwithstanding, bore data analysed for
this project showed sparse scattered bores with high yields, to 25 L/sec. Bore RN44659 in Kings
Creek sub-catchment, sunk into fractured and vesicular basalt at 38 m depth, yielded 56 L/sec during
pumping tests conducted in 1974. Sparsely distributed high-yielding bores most likely correspond
with significant structural features such as intersecting discontinuities (faults or conjugate joints).
Such structures are likely to promote vertical connectivity between aquifers by providing secondary
conduits. In places however, especially where water-bearing zones appear to end abruptly, structures
and stratigraphic offsets likely impede groundwater flow and connectivity.

Basalts worldwide are renowned for their anisotropic and heterogeneous nature; a wide range of
hydraulic conductivities are recognised, generally ranging from 10-2 – 10-7 m/s (Freeze & Cherry,
1979). Locally, the presence of intrusive and extrusive volcanic features and significant lateral
discontinuities in the MRV are further compounded by widespread vertical and horizontal joints,
e.g., radial and, potentially, concentric fractures. Such discontinuities impact significantly on the
storage and transmission of groundwater in the local area (Free, 1989). Typically, the MRV consists
of an upper unconfined weathered and fractured zone, representing a semi-permanent aquifer, a mid-
level semi-confined aquifer, and a lower confined vesicular and fractured zone (Free, 1989;
QDNRW pers. comm., 2007). Varying degrees of vertical continuity exist between these respective
zones, as well as between individual aquifers in each zone (Free, 1989). Multiple buried weathering
surfaces, former ash deposits, and possible sedimentary interbeds between basalt flows exacerbate
complexities and substantially influence groundwater storage and movement.

The high degree of aquifer heterogeneity precludes generation of a realistic dynamic 3D model at
sub-catchment scale. A 3D model for a relatively small area, encompassing ~20 bores for example,
could be built using detailed logs and reliable lithological information to incorporate primary and
secondary properties and groundwater features. This type of high-resolution system could potentially
accommodate detailed monitoring of watertable levels aimed at generating a dynamic model to show
groundwater responses to variable climatic factors and groundwater use scenarios. However, this
degree of stratigraphic and hydrogeologic detail could not plausibly be extrapolated across the entire
catchment, nor would it be appropriate to inform decision-making about the distribution of water
resources over large spatial and temporal scales. Inconsistencies and discontinuities in or between
water-bearing units revealed in the present 3D model are representative of the varied and intrinsic
characteristics of the MRV, such as the aquifer heterogeneity illustrated in Figures 17-19.
Accordingly, the 3D model represents the sub-surface heterogeneity as faithfully as possible at the
sub-catchment scale but cannot represent fine-scale anisotropy or heterogeneity in the MRV or
associated groundwater complexity.

The 3D model reinforces the classic hydrogeologic interpretation for the Condamine headwater
region (e.g., Free, 1989) while portraying the actual magnitude of variability of the subsurface
geologic and hydrogeologic environment. If the 3D model is further developed in the future,
additional insight into the distribution of groundwater and hydrogeologic processes in the region is
likely to be revealed. While it is widely known that there is a substantial degree of connectivity
between MRV aquifers and streams in the headwater region, an outstanding issue relates to

28
improving knowledge of the magnitude of connectivity between MRV and underlying Walloon
sequence aquifers.

The Toowoomba Bypass tunnel currently being excavated through the Main Range to the north of
Toowoomba is providing new information about the properties of MRV basalt such as texture,
degree of weathering, porosity, permeability and hydraulic conductivity. This information, currently
being assessed and described by QDNRW, in conjunction with the analysis and 3D synthesis of bore
data undertaken for this project, will significantly enhance our understanding of the local
hydrogeology. Further collaboration between Geoscience Australia and QDNRW hydrogeologists is
expected to result in increasingly reliable depth to water table and potentiometric models and
groundwater yield data that accommodate seasonal fluctuations and decadal-scale trends; these are
important considerations for water resource management in the upper Condamine catchment.

UPGRADING THE EXISTING 3D MODEL

Geoscience Australia has developed a relatively simple but realistic and effective 3D computer-
generated model of the Hodgson and Kings Creek catchments. This model accurately shows the
topographic and hydrogeological architecture of the region, and provides a novel visualisation tool
which could prove useful for groundwater management. The current version of the model may be
further enhanced through integration of other datasets such as geophysical and remote-sensing
images, and could be further refined by undertaking more detailed work, such as:

Improving the contact geometry of the MRV – Walloon boundary

The current modelled surface between the MRV and the Walloon Coal Measures is the smoothest
possible 3D boundary given the available borehole data. However, by integrating down-hole and/or
regional geophysical datasets it may be possible to improve the accuracy of the surface geometry. In
particular, the following techniques could prove useful:

a. 3D inversion of gravity data – open-file gravity survey data across parts of the
Hodgson and Kings Creek catchments are freely available through GADDS. It may
be possible to re-process some datasets and apply current inversion methods to
potentially re-model the contact surface (and possibly other surfaces such as the
base of the Walloon Coal Measures).

b. Electro-magnetic (EM) data – At least one Airborne EM survey has previously

been conducted across parts of the study area (Kalma, 1995). It may be possible to
re-process AEM data using new analytical routines, and potentially improve the
quality of this data-set, especially as it relates to groundwater resources and depths
of various lithological contacts.

c. Seismic reflection profiles – regional seismic investigations have previously been

undertaken across the Clarence-Morton Basin. At least one survey line is likely to
have passed either directly across or proximal to the study area. Although the
seismic data would likely be ‘noisy’ at the relatively shallow subsurface depths
relevant to this project, it may help to elucidate deeper geological boundaries and
provide information about potential deep aquifers in the basin.

29
Geostatistical simulation of 3D aquifer geometries
The 3D geometry of the main water-bearing zones in the MRV and Walloon Coal Measures could be
modelled using geostatistical simulations of the rock units. This type of modelling is analogous to
facies modelling undertaken by the petroleum industry where the various rock types identified from
drillhole data (sandstone, shale etc.) are modelled to generate a 3D visualisation of petroleum-
bearing zones. A possible approach for this project would be to simulate four distinct
hydrogeological facies, these being: dry MRV zones, water-bearing MRV zones, dry Walloon zones,
and water-bearing Walloon zones. The overall objective would be to model possible connectivity
between isolated water-bearing zones and represent these as 3D surfaces. As the model currently
stands, the thickness of water-cuts are marked on respective drillhole paths – however, it would
require a detailed geostatistical analysis to actually model lateral (and vertical?) water-bearing
connectivity between drillholes. Ultimately, this approach would be difficult to implement due to
the extreme spatial heterogeneity of water-bearing zones in the MRV.

Modelling potential sub-vertical fluid pathways

As postulated in the section describing the groundwater system of the MRV, considerable sub-
vertical connectivity between water-bearing zones is inferred. This is especially likely where
discontinuities associated with successive volcanic events are common, e.g., columnar cooling
joints, eruptive cones and associated concentric fractures, variable weathering patterns. Present-day
remnants of volcanic centres and fissures act as preferential recharge zones due to the high degree of
sub-vertical connectivity. Although there are no large-scale faults mapped in the area, the Jurassic
sedimentary rocks underlying the MRV are likely to be cross-cut by localised faults and fracture
zones which also act as preferential fluid flow paths. Detailed analysis of relevant geophysical
datasets (as outlined previously) could help better delineate crustal discontinuities which are locally
important zones for surface water infiltration and groundwater flow.

Structural lineament analysis could also be applied where the model reveals abrupt disconnections or
irregularities between multiple water-bearing layers. High-resolution satellite (e.g. ASTER) and
RadarSat imagery, potentially coupled with appropriate airborne geophysical data, may reveal
distinct structural zones that act as conduits for (or barriers to) groundwater flow, e.g., columnar or
concentric jointing associated with eruptive centres.

Scope exists in the future to apply more complex remote sensing and GIS techniques to defining
groundwater recharge and discharge areas (and other processes) over the entire Condamine
catchment. For example, recent work focused on Tertiary basalt aquifers in the Glenelg–Hopkins
catchment of western Victoria (Tweed et al., 2007), along with other work conducted by the
Cooperative Research Centre for Landscape Evolution and Mineral Exploration (CRC LEME), may
be applicable to future hydrogeologic projects in the Condamine region.

Integrating other hydrological applications with the 3D model

We suggest exploring the usefulness of the 3D model as a platform for other hydrological analytical
methods. For example, it may be possible to integrate the existing conceptual model with numerical
groundwater simulations and thus improve the usefulness of both techniques (previously used
independently). Discussions with experienced numerical modellers would initially be required to
evaluate the viability of integrating numerical groundwater solutions with the 3D conceptual model.

Incorporating hydrochemical data into the model is another potential application for which few
precedents currently exist. An extensive hydrochemical dataset of the groundwater resources in the
Hodgson Creek catchment has been compiled by various research groups over the past few years (R.

30
Cresswell, pers. comm., 2007). These data could be readily integrated into the existing 3D model as
extra data layers. For example, water quality could be colour-coded to show spatial variation of
potable versus stock-quality groundwater within the catchment – this would be especially useful if
3D groundwater simulations were developed and integrated with the existing model.

FURTHER 3D MODELLING IN THE CONDAMINE CATCHMENT

This report captures various static screen snapshots of the Gocad model to convey recognition of the
various datasets and provide an indication of the model’s substance and potential (Figure 10–18).
Export of the Gocad model to the web-enabled X3D version maximises access and utility for
Condamine catchment managers and other stakeholders, while presentation of the Gocad version
using 3D visualisation facilities optimises two-way communication between modellers and
catchment users.

The present 3D model shows unique sub-surface perspectives of Hodgson and Kings Creek
catchments. The model also promotes understanding of the spatial distribution of groundwater-
bearing zones beneath any part of this important agricultural area, where the level of dependence on
groundwater resources is high. The model graphically reinforces the existing perception of highly
complex aquifer systems and heterogeneous groundwater distribution, gained from traditional
hydrogeological studies over recent decades. The model enables and enhances groundwater
conceptualisation in 3D space to promote communication among stakeholders. For example, the
model may aid informed site selection of future boreholes, and guide ongoing monitoring of aquifers
which are currently exploited, particularly those in areas of intense irrigation and peri-urban
development.

Based upon the results of this pilot study, and the lessons learnt during design, development and
implementation, further 3D modelling in the Condamine region is considered possible. However,
any future 3D groundwater modelling project would need to address questions such as:

Can the model as it stands be used to calculate groundwater volumes in

aquifers?
The 3D model accurately represents all water-bearing zones logged in over 1000 of the most reliable
and representative boreholes from Hodgson and Kings Creek sub-catchments. Enormous
heterogeneity of lateral aquifer distributions and vertical extents has subsequently been revealed.
Commonly there are several or many thin water-bearing zones (<2 m thick) proximal to significantly
thicker saturated zones of limited spatial extent. Additionally, completely dry boreholes also occur
close to bores that intersect multiple water-bearing units of variable thickness. These observations
were heralded by the early hydrogeologic work of Free and others, who recognised the typically
non-uniform lithological character of the Main Range Volcanics (MRV) and the magnitude of
conspicuous fracturing, faulting and structural disruption around volcanic eruptive centres. As the
random snapshots showing distribution of water-bearing zones over small parts of the study area
indicate (Figure 16–18), it is unrealistic to simply join up water-bearing units (sub-horizontally) to
form laterally extensive 'layer-cake' surfaces for estimating contained aquifer volumes. Thus, even
though Gocad is capable of such analytical simulations, this would be a completely contrived
exercise for the Hodgson–Kings 3D model. This volume estimation procedure would be perfectly
feasible in other groundwater systems characterised by relatively uniform and laterally extensive
aquifers; the reliability of estimated groundwater volumes would then be dependent upon the
homogeneity of the aquifer’s various hydraulic properties.

31
On the positive side, the spatial heterogeneity of the Hodgson–Kings aquifers is well conveyed in the
3D model; this is perhaps more important to catchment managers, users of groundwater resources,
and water planners than a contrived and non-representative numerical volume. Depiction of each
borehole, (for example, Figures 17–19) provides a reliable indication of the constituent water-cuts
and the potential groundwater storage capacity beneath the landscape, particularly if information
about aquifer hydraulic parameters and yields is also considered. Used in conjunction with existing
bore yield data and future pump tests, a general impression of underlying groundwater resources can
be obtained for small-scale areas. Although not precise in volumetric terms, this information is more
relevant than the likely misinformation conveyed if laterally continuous layer-cake aquifers with
homogeneous or predicable hydraulic parameters (such as storativity, transmissivity and yield) are
assumed in the sub-surface.

Can Standing Water Levels (water tables/groundwater levels) be conveyed, and

updated SWL scenarios represented in the model?
Standing water level (SWL) measurements for the suite of over 1000 bores are haphazard, spanning
some 50 years of observations and encompassing seasonal and decadal-scale fluctuations, and highly
variable extraction rates. Creating a single catchment-wide SWL surface based on the available data
for Hodgson and Kings Creeks would be unrealistic and non-representative. If the SWL of every
bore used in the model was measured at approximately the same time then a 'typical SWL surface'
could be generated. This could, for example, potentially show seasonal watertable fluctuations
associated with climate variations. However, such SWL surfaces would not be representative of the
sub-catchments for long periods of time. Similarly, scenarios of extraction and recharge, or
comparisons between dryland, irrigation and native vegetation areas, cannot be represented in the
model without appropriate time-series data.

At present, the 3D model is a conceptual model that represents the stratigraphic architecture and
distribution of water-bearing zones (albeit from historical records spanning some 50 years) as
accurately as possible in x-y-z space, i.e. the ‘whole-of-catchment architecture’. The groundwater
system essentially represents a 4th dimension of the model, dependent on the physical constraints of
the 3D geologic environment but independently responding to hydrologic and meteorologic fluxes
and land-use changes over time. The existing datasets currently preclude including more realistic or
dynamic representations of the watertable and groundwater processes. However, temporal variations
in groundwater levels could potentially be included in the 3D model if the necessary datasets were
available, although this would present new challenges in model design and development.

Is the process likely to work for the whole of the Condamine catchment?
The heterogeneity disclosed in the 3D modelling exercise for Hodgson and Kings Creek sub-
catchments indicates that, in areas of the upper Condamine dominated by the MRV, 3D modelling is
more suited to small- or medium-sized areas rather than broader regions. Developing the present 3D
model of the Hodgson and Kings Creek sub-catchments required considerable analysis and
rationalisation of lithological data from the original bore logs. For larger regions of the Condamine
catchment, or indeed for the entire Condamine Basin, the magnitude of analysing and consolidating
borehole descriptions would be extremely time-consuming and labour-intensive, as it would likely
involve many thousands of bores and numerous stratigraphic units.

Downstream in the Condamine catchment, away from the heterogeneous MRV basalts and the
structural complexity imposed by numerous volcanic eruptive centres and the uplifted Great
Dividing Range, 3D modelling is likely to present a more straightforward exercise and significantly
different outcomes. Seismic data may be more useful and readily available to help model the basin

32
architecture and major structural zones. Relatively thick strata of the Cainozoic Surat Basin and
Jurassic Clarence-Morton Basin also likely correspond with thicker and more laterally extensive
water-bearing units than occurs in the MRV. If the latter proved to be the case, then calculation of
aquifer volumes incorporating temporal variations is possible. Since the data density in other
Condamine sub-catchments is less than in the Hodgson and Kings sub-catchments interpolation
between data points would be greater. This would increase uncertainty involved in generating
catchment-wide geological boundaries and the thicknesses, continuity and lateral extents of the
contained aquifers.

Ideally, 3D modelling in the Condamine would be best to proceed at the sub-catchment scale. A
geologically different sub-catchment is most appropriate for the next phase of 3D model generation,
i.e., a sub-catchment located away from the inherent complexities of the MRV and the upland areas,
and for which there is an adequate distribution of representative and well-documented boreholes and
groundwater monitoring data. A 3D model of the whole Condamine catchment is technically
feasible. For example, broad-scale, data-rich regional 3D geologic models for groundwater
applications have been developed in recent years in North America, e.g., Thorleifson et al., 2005;
Bajc and Newton, 2005.11 Such models require large resources in terms of personnel, expertise and
time, especially as thousands of bore logs and supporting geologic/hydrogeologic data requires
significant analysis and interpretation to generate a useful conceptual model. Modelling a large area
invariably implicates increased generalisation and simplification – or ‘lumping of variability’ to
manageable levels – of stratigraphic units, geologic contacts, aquifer distribution and properties, and
groundwater data. Such generalisation involves concomitant lessening of reliability and applicability
for catchment or water resource management issues. The decision would always be whether to
represent and understand small sub-catchments reasonably well or a large catchment in a potentially
more schematic manner.

11
http://www.isgs.uiuc.edu?3DWorkshop/nu_3d_welcome.html

33
Key Recommendations
The 3D hydrogeological model of the Hodgson and Kings Creek sub-catchments developed for this
project has shown the feasibility of building such models, and their usefulness as tools for better
understanding and evaluating groundwater systems. However, the current version of the model is
clearly not definitive, and should only be considered as a stage in the ongoing development and
application of this technology. Further work is contingent on the requirements of the main
stakeholders, and should be undertaken in close consultation with the Condamine Alliance.

Building upon the comments provided in the previous section, our key recommendations for future
studies related to the Hodgson and Kings Creek 3D model are to:

1. Develop a similar 3D hydrogeological model in another Condamine River sub-catchment,

focussing on an area with markedly different hydrogeological characteristics to those of the
Hodgson and Kings Creek district, i.e., groundwater flow systems not hosted in volcanic
rock-dominated domains;

2. Integrate other sources of geological and geophysical data into the development of this (and
other) 3D models to improve their accuracy and scientific validity. These could include
geological data from mineral exploration or coal-seam methane exploration drillholes, or local
electromagnetics or radiometrics surveys;

3. Consult closely with other natural resource management sectors with a vested interest in the
Condamine region, in order to broaden the scope of the current 3D model and make it more
useful to a wider environmental audience. Specific work examples could include tackling
problems associated with groundwater dependant ecosystems such as wetlands or springs, or
integrating overlays of soils or landuse data with the digital elevation model; and

4. Further develop the model by focussing the model on a smaller scale area of either the
Hodgson or Kings Creek sub-catchment, where reliable monitoring bore data from multiple
sites is available. This would allow better resolution of hydrogeological properties, e.g.,
modelling individual water-bearing zones based on the lithological and hydraulic parameters
of the water-bearing zone. Modelling the standing water-level and investigating temporal
variations in the watertable level could also form part of a more detailed modelling
investigation.

Acknowledgments
This project has received substantial datasets and support from the Queensland Department of
Natural Resources and Water (QDNRW), Toowoomba. We particularly thank Mira Geoscience for
initial model building and technical assistance in developing the Gocad capacity at Geoscience
Australia. Ashwin Dyall conducted preliminary data analysis and GIS work for the project. Support
and assistance from Geoscience Australia colleagues: Malcolm Nicoll, Chris Fitzgerald, Jane Coram,
Fiona Watford, Terry Brennan and Tim Johns; is gratefully acknowledged.

34
References
Armstrong, D., 1974. Recharge and the Groundwater regime of the Toowoomba City Catchment.
Report to the Toowoomba City Council.

Bajc, A.F. and Newton, M.J., 2005. 3D Modeling of Quaternary Deposits in Waterloo region,
Ontario: A Case Study using Datamine Studio Software. Workshop Extended Abstracts:
Three-Dimensional Geologic Mapping for Groundwater Applications: October 15, 2005,
Annual Meeting, Geological Society of America, Salt Lake City, Utah.

Beard, D., 2005. Using VRML to Share Large Volumes of Complex 3D Geoscientific Information
via the Web. Web3D 2006 11th International Conference on 3D Web Technology, Columbia,
Maryland, 18-21 April 2006. ACM 163-167.

Beard, D.J., Hay, R.J., Nicoll, M.G. and Edge, D.O., 2005. 3D Web Mapping – 3D Geoscience
Information Online. In Proceedings of SSC 2005 Spatial Intelligence, Innovation and Praxis:
The national biennial Conference of the Spatial Sciences Institute, September 2005.
Melbourne: Spatial Sciences Institute.

Biggs, A.J.W., 2000. Geology and Landform. Chapter 3: Central Darling Downs Land Management
Manual: Understanding and Managing Land. CD produced by the Queensland Department
Natural Resources and Water.

Biggs, A.J.W. and Power, R.E., 2003. A review of salinity occurrences in the Queensland Murray-
Darling Basin, 2002. Department of Natural Resources and Mines Report:
http://www.nrw.qld.gov.au/salinity/pdf/qmdb_salinity_01.pdf

Biggs, A.J.W., Power, R.E., Silburn, D.M., Owens, J.S., Burton, D.W.G. and Hebbard, C.L., 2005.
Salinity audit – Border Rivers and Moonie Catchments, Queensland Murray-Darling Basin.
Department of Natural Resources and Mines, Queensland. Report QNRM05462.

Carberry, D. and Walker, B., 1993. Hodgson Creek Catchment: A Situation Statement. Report to the
Hodgson Creek Catchment Committee.

Cresswell, R.G., Silburn, D.M., Biggs, A.W, Dighton, J.C., deVoil, R.C., Rassam, D. and McNeil,
V.H., 2006. Hydrogeochemistry of Hodgson Creek Catchment, Queensland Murray-Darling
Basin. 10th Murray-Darling Basin Groundwater Workshop: ‘Does the science hold water?’
Canberra 18-20 September 2006.

Dutta, S and Silburn, D.M., 2005. Streamflow modelling, groundwater balance and recharge
estimates for Hodgson Creek, Condamine Catchment, Queensland. Queensland Department
of Natural Resources and Mines, Toowoomba/Brisbane. Report QNRM05514.

Free, D., 1989. Geology of the Main range Volcanics and Hydrogeological Implications:
Toowoomba District, South East Queensland. Water Resources Report.

Huxley, W.J., 1982. An Assessment of Groundwater Availability in the Tributary stream Areas:
Condamine River Valley Groundwater Investigation. Queensland Water Resources
Commission Record.

35
Kalma, S.J., 1995. An Evaluation of Airborne Geophysics for Salinity Assessment in Property
Management Planning: Pittsworth Airborne Geophysical Survey, Queensland Department of
Primary Industry.

Macnish, S.E., Koppi, A.J., Little, I.P. and Schafer, B.M., 1987. The Distribution, Nature and Origin
of Some red Sesquioxide Materials in Southeastern Queensland, Australia Geoderma, 41: 1-
27.

McNeil, V. and Silburn, D.M., 2005. Hodgson Creek salinity (EC), salt load and flow calculations.
Queensland Department of Natural Resources and Mines, Toowoomba/Brisbane Report
QNRM05515.

Murphy, G.A., 1980. Report on Groundwater Investigations: Darling Downs – Nobby Area.
Queensland Water Resources Commission.

Natural Resources and Water, 2000. Central Darling Downs Land Management Manual:
Understanding and Managing Land. CD produced by the Queensland Department NWR.

Ollier, C.D. and Haworth, R.J., 1994. Geomorphology of the Clarence-Moreton Basin. In: Wells,
A.T. and O’Brien, P.E. (Eds.). Geology and Petroleum Potential of the Clarence-Moreton
Basin, New South Wales and Queensland. AGSO Bulletin 241.

Rattray, D.J., Freebairn, D.M., Silburn, D.M., 2002. Water quality in grain farming catchments –
towards understanding the impact of management and scale.
http://www.riverfestival.com.au/2002/files/symposium/RattrayDJ.doc

Reid, M., Cheng, X., Banks, E., Jankowski, J., Jolly, I., Kumar, P., Lovell, D., Mitchell, M., Mudd,
G., Richardson, S., Silburn, M., and Werner, A., 2007. Catologue of Conceptual Models for
Groundwater-Stream Interaction: Case study 4: Hodgson Creek, Darling Downs Main Range
Volcanics (Basalts): Layered or complex system. eWater CRC Initial Project, May 2007: p.
80-88.

Ruperto, L. and Nicoll, M., 2007. 3D Modelling of the Callabonna Sub-Basin (Southern Curnamona
Region). CRC LEME Open File Report 184.

Silburn, D.M., Moss, J.B., deVoil, R.C. and Huggins, J., 2005. Assessment of the potential for
salinity and waterlogging in the Felton Valley Landcare area. QNRM05343. Department of
Natural Resources and Mines, Queensland.

Silburn, D.M., Owens, J.S., Cresswell, R.G. and McNeil, V., 2006. Hodgson Creek, Queensland
Murray-Darling Basin – salinity and recharge studies and 2CSalt modelling. 10th Murray-
Darling Basin Groundwater Workshop: ‘Does the science hold water?’ Canberra 18-20
September 2006.

Stevens, N.C., 1969. The Tertiary volcanic rocks of Toowoomba and Cooby Creek, South-east
Queensland. Proc. R. Soc. Qd: 80/7:85-96.

Thompson, C.H., and Beckmann, G.G., 1959. Soils and land use in the Toowoomba area, Darling
Downs, Queensland. CSIRO Australia: Soils and Land Use Series 28, 78 pp.

36
Thorleifson, L.H., Berg, J.A., Tipping, R.G., Malolepszy, Z., Harris, K.L., Lusardi, B.A.,
Setterholm, D.R. and Anderson, F.J., 2005. 3D Geological Modeling in support of Ground
Water Inventory in the Fargo-Moorhead region, Minnesota and North Dakota. Workshop
Extended Abstracts: Three-Dimensional Geologic Mapping for Groundwater Applications:
October 15, 2005, Annual Meeting, Geological Society of America, Salt Lake City, Utah.

Tweed, S.O., Leblanc, M., Webb, J.A., Lubczynski, M.W., 2007. Remote sensing and GIS for
mapping groundwater recharge and discharge areas in salinity prone catchments, southeastern
Australia. Hydrogeology Journal, 15: 75-96.

Van der Zee, J.J. and Macnish, S.E., 1979. The Geology and Geomorphology of an area of basaltic
soils near Pittsworth on the Darling Downs. Queensland Department of Primary Industry,
Division of Land Utilisation. Report 79/5.

Wells, A.T. and O’Brien, P.E., 1994a. Geology of the Clarence-Moreton Basin (1:500 000 scale).
Australian Geological Survey Organisation, Canberra.

Wells, A.T. and O’Brien, P.E. (Compilers and Editors), 1994b. Geology and Petroleum Potential of
the Clarence-Moreton Basin, New South Wales and Queensland. AGSO Bulletin 241.

Whitaker, W.G. and Green, P.M., 1978. Moreton Geology: 1:500 000 scale map. Geological Survey
of Queensland, Department of Mines.

Willey, E.C., 1992. Geology of the Hodgson Creek Catchment. Report to Hodgson Creek Catchment
Committee (Unpublished).

Willey, E.C., 2003. Urban geology of the Toowoomba conurbation, SE Queensland, Australia.
Quaternary International, 103: 57-74.

37
38
APPENDICES

Appendix 1: Lithological Borehole Codes

Appendix 2: 3D Grid Construction

Appendix 3: Gocad Object Definitions

39
Appendix 1 – Lithological Borehole Codes
Code Number of uses Type Stratigraphic unit
TSOIL 701 Topsoil Quaternary

WBSLT 728 Weathered Basalt MRV

WBSLTWB 200 Weathered Basalt - Water-bearing MRV

BSLT 2216 Basalt MRV

BSLTWB 127 Basalt - Water-bearing MRV

FBSLT 144 Fractured Basalt MRV

FBSLTWB 166 Fractured Basalt - Water-bearing MRV

HBSLT 534 Honeycomb Basalt MRV

HBSLTWB 707 Honeycomb Basalt - Water-bearing MRV

BSLTSOAP 40 Basalt with soapstone MRV

BSLTSOAPWB 4 Basalt with soapstone - water-bearing MRV

BSLTASH 39 Basaltic Ash MRV

BSLTASHWB 2 Basaltic Ash - Water-bearing MRV

CLAY 881 Clay ALL

CLAYWB 13 Clay - Water-bearing ALL

CLAYG 43 Clay - with gravel Quaternary

CLAYGWB 14 Clay with gravel - Water-bearing Quaternary

CLAYS 43 Clay with sand/silt Quaternary

COAL 157 Coal Walloon

COALWB 198 Coal - Water-bearing Walloon

GRAV 115 Gravel ALL

GRAVWB 71 Gravel - Water-bearing ALL

SAND 24 Poorly consolidated sand Quaternary

SANDWB 32 Sand - Water-bearing Quaternary

SDST 474 Sandstone Walloon, MRV

SDSTWB 90 Sandstone - Water-bearing Walloon, MRV

SHLE 257 Shale Walloon, MRV

SHLEWB 12 Shale - Water-bearing Walloon, MRV

40
Appendix 2 – 3D Grid Construction
Preparation for creating the 3D grid
The Sgrid object in Gocad (an application for creating stratigraphic grids) was successfully used to
generate the main stratigraphic units in the Condamine 3D model. The advantage of using this
method is that the cells of the SGrid can be deformed and eroded to represent the true geometry of
the stratigraphic units.

Surfaces used for constructing Gocad SGrids must conform to the following rules:

• They must have the same lateral extents and the same number of border segments (hence
the need to use the low resolution DEM data to extend the topographic surface beyond the
catchment boundary);

• They must be continuous across the entire model area, i.e., windows through a geological
unit must be locally represented by 0 m thickness; and

• Faults must be represented on the uppermost and lowermost surfaces from which the SGrid
is constructed (not an issue in this version of the model as no faults are modelled).

With regard for these rules, the topographic surface had to conform to the extents of the upper
Walloon surface as constructed in the structural modelling workflow (SMW). Thus, all ‘windows’ in
the top of Walloon surface had to coincide with the topographic surface. This was achieved by
mutually cutting the two surfaces with each other, and retaining the necessary parts of these surfaces
to create a new and continuous surface.

SGrid Construction
The SGrid construction workflow in Gocad has the same philosophy as the SMW; it is a decision
driven process and automatically generates a report as the workflow progresses. Unfortunately the
resolution of the 50 m DEM was too detailed for the SGrid routine to process, so the 100m DEM
was used instead.

SGrids are specifically designed to mimic stratigraphic units. The corners of each cell hold their
XYZ locations so that the cells can be deformed to more accurately represent the true sub-surface
geometry of the geological units. Cells at erosional contacts are able to be collapsed to 0 m
thickness if the erosion has completely removed that particular cell in the column.

Two approaches were attempted to determine the best way to create the SGrid model. The most
successful approach involved constructing individual SGrids for the MRV and Walloon Coal
Measures as:

1. The spatial extent of the 3D model when the Hodgson and Kings Creek catchments
are combined is too large to generate a functional SGrid model; and

2. The stratigraphic column for the model was too simple to define discrete layers
within the SGrid workflow.

To improve model functionality the cell layering in the MRV and Walloon grids varies slightly. In
the MRV grid (termed ‘strat_grid_4’) cells are 2 m thick and parallel to the modelled MRV-Walloon

41
contact. These are then eroded by the topographic surface (remembering that the thin upper
Quaternary layer could not be effectively modelled) to represent the possible basalt flow geometry as
closely as possible. Zero volume cells (i.e., topographically eroded cells) are registered as ‘invalid’,
so that in future property modeling only the valid cells will be used. In the Walloon grid (termed
‘strat_grid_5’), the cells are also 2 m thick, but distributed proportionally between the upper and
lower surfaces. Ideally, as the MRV-Walloon interface is an unconformity, the cells should also
parallel the bottom surface; however, as the base is a downward-translated copy of the upper
Walloon surface, the proportional distribution is considered the best representation.

Given the way in which the geometries of the cells are stored, the grid uses more computer memory
than the other type of grid object in Gocad, i.e., voxets. As a result, maintaining the detailed vertical
resolution (2 m) in the grids resulted in decreased lateral resolution. This required the lateral
resolution to be set at 400 m x 400 m. Although this is not ideal, it is well correlated with the
average distance between the main source of input data, i.e., the water bore data.

42
Appendix 3 – Gocad Object Definitions
The objects constructed in Gocad are organised into “Favourites” folders within the project.
Arranging objects like this allows them to be grouped according to the data types as opposed to
object types. This makes it easier for users to find particular objects when they may or may not
know the object type that support the data that they are most interested in. The following
descriptions explain the main data types used in the 3D model.

Borehole data
Drillholes_Hodgson – a drillhole group containing all holes in the Hodgson catchment area.
Contains markers for the tops and bottoms of all stratigraphic and aquifer units as well as continuous
zones for the stratigraphy and zones for the aquifers.

Drillholes_Hodgson_MRV – a curve object representing the drillhole zones for the MRV unit on the
Drillholes_Hodgson group. The same object exists for each of the zones on the drillholes.

Drillholes_Hodgson_paths – 1 single curve object what represents the drillholes traces. Included as
a lighter representation of the drillhole locations for visualization purposes.

The same nomenclature has been used for all the drillholes in the Kings Creek catchment.

Walloon_top_marker_points – points from top of Walloon markers within the drillholes.

Geology
3_Walloon_base_sgrid – the surface representing the base of the SGrid is below the deepest aquifer
from all drillholes.

3_Walloon_top_ts_edited_for_100m_DEM – the final surface from the Structural Modelling

Workflow (SMW), but with portions of the 100m DEM surface inserted where there is 0 m thickness
of MRV.

Strat_grid_4 – the stratigraphic grid created to model the MRV over the two catchment areas.
Vertical cell resolution 2m, lateral resolution 400 m x 400m. Cells parallel to
3_Walloon_top_ts_edited_for_100m_DEM. Some sections may show graphical artefacts where the
grid has 0 m thickness, i.e., MRV absent.

Strat_grid_5 – the stratigraphic grid created to model the Walloon over the two catchment areas.
Vertical cell resolution 2m, lateral resolution 400 x 400m. Cells proportional to
3_Walloon_top_ts_edited_for_100m_DEM and the surface 3_Walloon_base_sgrid.

Surface Data / Base_– the curve edited from a shape file representing the contact on the 1:500,000
mapsheet.

Surface Data / Base_Quaternary – the curve edited from a shape file representing the contact on the
1:500,000 mapsheet.

Walloon_top_marker_points – points from top of Walloon markers within the drillholes.

43
Geophysics
(Object names are derived from the data files supplied directly from GADDS)

Gravity / gravmapDec2005_cropped – GADDS gravity grid cropped over model area

Gravity / gravmapDec2005_regional – GADDS gravity grid.

Gravity / V_GRAV_JETSTREAM – GADDS gravity points data. Properties are Bouguer, FreeAir
and Gravity from the processing stream. Bouguer is set to be displayed by default.

Magnetics / GSQP1108tmig – GADDS magnetic grid

Magnetics / GSQP792tmig – GADDS magnetic grid

Remote sensing
Condamine_SPOT5_pseudonatural_resampled_small – Spot 5 satellite image

Topography
Cultural Data / Boundary – catchment boundaries

Cultural Data / dam_wiers_56 – points representing dam weirs within the model area imported
from ArcGIS. Same for Rail, Roads, Springs3D, Streams, and towns_56_objects.

DEM_100m_surface – Topography surface created from resampled 25m resolution DEM with
low resolution areas (~300 m) stitched onto the surface where data was absent. Properties
I0_4/J0_4/K0_4/tmp__I0_4/temp__J0_4 all created within the SGrid workflow. Thickness
properties are Depth_to_base_sgrid / Depth_to_Walloon_contact / Depth_Walloon_to_base.

Additional procedures
To import a drillhole path – the file “deviation_final.txt” is in the correct format – hole_id, depth,
azimuth, dip. Things to remember are that the final depth entry needs to be the length of the hole
and a vertical dip entry is 0 while a horizontal one is 90.

The command is found at File > Import Well Data > Path > from column-based file.

Import everything at 0,0,0. The holes will be correctly located in the next step.

File > Import Well Data > Locations from column-based file – the file “collar.txt” is in the right
format – hole_id, X, Y, Z (the Depth entry is superfluous). Note that KB (or Kelly Bushing) is the Z
value, use the option “move existing wells”.

Once this command has been executed, then the drillholes will be in the correct location.

The next steps are to add properties and geological information as needed.

Adding properties – The file “assay.txt” is a combination of Mira and Gocad format. Mira uses
from-to fields to plot the property value at the mid-point whereas the Gocad import filter plots the
property at the mid point (To_mid in this file). So the two filters plot the data at the same point, they
get to the same result in slightly different ways. The Gocad command is found at: File > Import

44
 Well Data > Logs > From column based file. Do not use the option to add all the properties, it is best
to manually add the properties in the next window of the filter.

Adding geology zones – The file “geology.txt” is the correct format for both the Mira and EDS
importers – Hole-id, from, to, lithology. The command to create well zones is found at File > Import
Well Data > Well Zone and Values

Adding geology markers – The file “geology.txt” is used for this command as well which is
found at File > Import Well data > Markers > Column based file, depending on whether you
want the marker at the top or the bottom of the interval, you select the From or To field as your
Measured Depth column. Remember, placing the marker at the bottom of the interval will place
a marker at the very end of the drill hole which almost certainly ends in the middle of the
lithological unit, these markers should not be used to constrain a surface.

45

View publication stats