processing and inversion of walyarta (mandora marsh) aem data · 2.1 tempest aem system the tempest...

Processing and Inversion of Walyarta (Mandora Marsh) AEM Data: Enhanced TEMPEST data processing and inversion.

Camilla Soerensen, Tim Munday, Jasmine Rutherford, and Kevin Cahill

Technical Report EP 178999

10 November 2017

MINERAL RESOURCES

Citation:

Soerensen,C.C., Munday, T.J., Rutherford, J.L. *and Cahill, K. 2017. Processing and Inversion of Walyarta (Mandora Marsh) AEM Data: Enhanced TEMPEST data processing and inversion. CSIRO Technical Report EP 178999; CSIRO, Australia.

*J.L. Rutherford is an employee of the Department of Biodiversity, Conservation and Attractions.

Copyright

© Commonwealth Scientific and Industrial Research Organisation 2017. To the extent permitted by law, all rights are reserved and no part of this publication covered by copyright may be reproduced or copied in any form or by any means except with the written permission of CSIRO.

Important disclaimer

CSIRO advises that the information contained in this publication comprises general statements based on scientific research. The reader is advised and needs to be aware that such information may be incomplete or unable to be used in any specific situation. No reliance or actions must therefore be made on that information without seeking prior expert professional, scientific and technical advice. To the extent permitted by law, CSIRO (including its employees and consultants) excludes all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication (in part or in whole) and any information or material contained in it.

CSIRO is committed to providing web accessible content wherever possible. If you are having difficulties with accessing this document please contact [email protected].

Front Cover Image: Interval conductivity image from inverted TEMPEST data over Mandora Marsh

mailto:[email protected]

4 | Processing and Inversion of Walyarta (Mandora Marsh) AEM Data:

Foreword

This report provides a summary of the processing and inversion of TEMPEST Airborne electromagnetic (AEM) data covering Mandora Marsh in the Northern Part of WA. The work was undertaken to assist the West Australian Government Department of Biodiversity, Conservation and Attractions, in the development of its hydrogeological conceptual model for the area.

Processing and Inversion of Walyarta (Mandora Marsh) AEM Data: | 5

Contents

Foreword 4

Acknowledgments ........................................................................................................................... 8

1 Introduction ...................................................................................................................... 10

1.1 Objective: ............................................................................................................. 10

1.2 Workflow: ............................................................................................................ 11

1.3 Study area ............................................................................................................ 11

2 TEMPEST Data Processing and Inversion .......................................................................... 12

2.1 TEMPEST AEM system ......................................................................................... 12

2.2 Data processing ................................................................................................... 13

2.3 AEM data inversion ............................................................................................. 15

2.4 Inversion model ................................................................................................... 17

2.6 Depth of investigation ......................................................................................... 18

2.7 Results ................................................................................................................. 18

3 References ........................................................................................................................ 26


Figures Figure 1: The location of the Mandora Marsh study area in NW Western Australia. Inset photographs relate to site locations A, B, C and D indicated on the image. ................................ 10

Figure 2: TEMPEST AEM flight lines processed as part of this study. Flight lines are overlain on a DEM ............................................................................................................................................... 11

Figure 3: The standard TEMPEST AEM system in operation using the CASA 212 fixed wing aircraft. .......................................................................................................................................... 12

Figure 4. TEMPEST Z component amplitude response (top set of curves) and X component amplitude response data (bottom set of curves) for part of a flight line. The greyed out areas are where noise has been removed either through automatic filtering or through a manual process. ......................................................................................................................................... 15

Figure 5: Schematic representation of TEMPEST fixed-wing time domain EM data acquisition and interpretation. A) Data are acquired along parallel flight lines, with data recorded at fixed time intervals (soundings or fiducials) along each flight line; B) The EM receiver located behind and below the transmitter measures the secondary responses from the ground; C) The measured response is used to determine the conductivity-depth function by transformation or inversion. Conductivity-depth values can be calculated for each observation, taking account of the elevation of the system above the ground; D) The modelled conductivity structure for each sounding is then stitched together into conductivity-depth sections to provide a representation of the 2D variation of conductivity, sometimes referred to as a “para-section”. Conductivity depth models can be gridded to produce horizontal depth slices (or interval conductivity images) to show the spatial distribution of conductivity as it varies with depth. These maps can be shown as elevations (mAHD) or as depth intervals below the ground surface ground surface. .......................................................................................................................................... 16

Figure 6: Comparison between stitched conductivity sections for a smooth 30-layer (top panel) and a 7 layer blocky (lower panel) 1D inversion for the same line of TEMPEST data in the study area. .............................................................................................................................................. 17

Figure 7: A diagrammatic representation describing the principle of laterally constrained inversion. Individual conductivity models of the subsurface, derived from the inversion of individual soundings measured by the AEM system, are laterally correlated in the along-line direction to enhance the definition of laterally continuous conductive layers. .......................... 19

Figure 8: Stitched conductivity-depth section for a north-south line (line 4002804). The depth of Investigation (DOI) is defined by the black line below which the section has an opaque overlay. The conductivity model structure below this line should be interpreted with caution. As an example, the deep conductive unit between 500 and 5000m along the line could not be reliably interpreted as being stratigraphically related as it falls below the DOI. Similar features were excluded from a stratigraphic interpretation of the data. .................................................. 19

Figure 9: A comparison between smooth model 1D inversion results for Mandora Marsh. The top panel shows a stitched 1D conductivity-depth section derived from a VPem1D inversion (Fullagar and Pears, 2007; Fullagar et al., 2013). The lower panel shows the results from application of the AarhusInv inversion code ( Auken et al. 2015) for the same line, with a depth


of investigation also defined. The lower inversion results are fitted to the 1 sec SRTM DEM. The two sections are scaled over the same colour range. .................................................................. 20

Figure 10: A comparison between smooth model 1D inversion results for the same line over Mandora Marsh. The top panel shows a stitched 1D conductivity-depth section derived from a VPem1D inversion (Fullagar and Pears, 2007; Fullagar et al., 2013). The lower panel shows the inversion results from application of the AarhusInv inversion code ( Auken et al. 2015) for the same line, with a depth of investigation also defined. The lower inversion results are fitted to the 1 sec SRTM DEM. The two sections are scaled over the same colour range. ........................ 21

Figure 11: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from below the surface for 10m intervals, starting from surface to 40m. The pseudocoloured conductivity images are overlain on the SRTM DEM. .............................................................................................................................................. 22

Figure 12: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from below the surface, starting at 40m and extending to 80m. The pseudocoloured conductivity images are overlain on the SRTM DEM. .. 23

Figure 13: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from 80 – 120m below the surface. The pseudocoloured conductivity images are overlain on the SRTM DEM. ....................................... 24

Figure 14: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from 120 – 150m below the surface. The pseudocoloured conductivity images are overlain on the SRTM DEM. ....................................... 25

Tables Table 1. System characteristics of the Standard TEMPEST airborne electromagnetic system. ... 13

Table 2: TEMPEST window information for 25Hz base frequency ............................................... 14


Acknowledgments

This work was funded as a collaborative project between the Department of Biodiversity, Conservation and Attractions and the CSIRO, and we acknowledge the support of these institutions in helping conduct the work. BHP is also acknowledged for supporting the project.


1 Introduction

This report summarises work undertaken with a subset of the TEMPEST airborne electromagnetic (AEM) (see Section 2.1 for system description) survey data acquired over the West Canning Basin by the Department of Water (CGG 2015). Specifically it addresses the optimisation of a ground conductivity model derived from the AEM data over Walyarta (Mandora Marsh) mound springs, and its surrounding environment, extending to the coast (Figure 1). This study, commissioned by the Department of Biodiversity, Conservation and Attractions (DBCA), had as its focus the enhancement of the near surface conductivity structure in the vicinity of mound springs, and the deeper conductivity structure that provides information relating to the stratigraphy. The aim was to provide additional information to assist the Department in development of its hydrogeological conceptual model for the area. The study was not concerned with an interpretation of the derived conductivity structure as defined through the inversion of the AEM data.

Figure 1: The location of the Mandora Marsh study area in NW Western Australia. Inset photographs relate to site locations A, B, C and D indicated on the image.

1.1 Objective:

Using Raw TEMPEST data, undertake non-linear inversion of AEM data, trialling several optimisations to enhance lateral (sediment-related boundaries), and vertical/sub-vertical (possibly fault–related) changes in conductivity structure in the vicinity of Walyarta (Mandora Marsh) mound springs. The inversion of the TEMPEST AEM data was intended to improve the models of near surface conductivity, critical to the understanding of the upper part of the profile which supports the vegetation communities associated with the springs and other parts of the Marsh. It was also carried out to understand the deep conductivity


structure which was important for helping define the stratigraphy of the area. In that regard, a detailed understanding of the depth of investigation for the system was also important. The study did not involve the interpretation of the derived conductivity models, which was the remit of DBCA.

1.2 Workflow:

1. Process and invert raw TEMPEST data to take account of system geometry in defining conductivity structure;

2. Trial smooth, and blocky (few layer) inversions to define sediment package boundaries/discontinuities; 3. Test variations in lateral and vertical constraints to pick up vertical (fault-related) changes and to pick up

sub-horizontal layering; 4. Optimise inversions to maximise contrasts in vertical conductivity structure and to focus on variations in

the near surface; 5. Define the indicative depth of investigation for the TEMPEST system over the area studied; 6. Generate GIS compatible conductivity-depth sections and interval conductivities (elevation and depth

below ground surface) and deliver these to DBCA for their interpretation.

To develop optimal results for the subsequent interpretation of the AEM data, several approaches to their inversion (blocky and smooth), were undertaken. The study also modelled system geometry to ensure near surface structure was well defined.

1.3 Study area

This study concentrated on the processing and inversion of the flight lines indicated in Figure 1. These data were acquired as part of a larger survey undertaken in the region between December 2014 and 28th June 2015 (CGG 2015).

Figure 2: TEMPEST AEM flight lines processed as part of this study. Flight lines are overlain on a DEM


2 TEMPEST Data Processing and Inversion

2.1 TEMPEST AEM system

The TEMPEST system is a fixed-wing time domain airborne electromagnetic system developed in 1998 by CRC-AMET established by the Australian Government’s cooperative research centres (CRC) program (Lane and Pracilio 2000). The system has been widely used in Australia for both prospect-scale surveys (Beckitt 2003), catchment management (Lane et al. 2000) and groundwater mapping (Sattel and Kgotlhang 2004). It has also been deployed for surveys covering large regional areas, including, for example, the Paterson in Western Australia (WA), Pine Creek in the Northern Territory, the Frome Embayment in South Australia, and the Capricorn region of WA (Roach 2010, 2012; Craig 2011 and Ley-Cooper et al. 2017).

For this survey the TEMPEST AEM system (Lane et al. 2000) was mounted in a fixed-wing aircraft (a CASA 212) with the transmitter loop is draped around the wingtips, tail and nose of the aircraft, while the receiver coil, hosted in a bird, was towed approximately 120m behind and 40m below the aircraft (Figure 3). The transmitter height is typically a nominal 120m above the ground. The system measures both the inline (Z) and the vertical (X) components (Table 1). The deconvolved ground response is converted to a 100% duty cycle square-wave B-field response, to yield better resolution of near surface detail in the ground conductivity structure. The response is then binned into fifteen windows (Table 2).

More detailed system specifications can be found in Tables 1 and 2, and in the acquisition and logistics report provided by CGG (CGG 2015).

Figure 3: The standard TEMPEST AEM system in operation using the CASA 212 fixed wing aircraft.


Table 1. System characteristics of the Standard TEMPEST airborne electromagnetic system.

TEMPEST System STANDARD

Base frequency 25 Hz Transmitter area 244 m2 Transmitter turns 1 Waveform Square Duty cycle 50% Transmitter pulse width 10 ms Transmitter off-time 10 ms Transmitter turnoff 42µs Peak current 280 A Peak moment 68,320 Am2 Average moment 34,160 Am2 Sample rate 75 kHz on X and Z Sample interval 13 microseconds Samples per half-cycle 1,500 System bandwidth 25 Hz to 37.5 kHz Nominal Flying Height (subject to safety considerations)

120 m EM sensor Towed bird -3 component dB/dt coils Tx-Rx horizontal separation 117 m (nominal) Tx-Rx vertical separation 41.5 m (nominal) Stacked data output interval 200 ms (~12 m) Number of output windows 15 Window centre times 13 µs to 16.2 ms

2.2 Data processing

Although both X (inline) and Z (vertical) component data are recorded (CGG 2015), only the Z component data were used in the modelling of the TEMPEST data over Walyarta. The Aarhus Workbench (Auken et al. 2015) was used for their processing and subsequent inversion. For this study we differentiate between the processing of the AEM data and the procedures involved in their subsequent inversion. A subset of the full TEMPEST dataset flown in the region was imported into Aarhus Workbench and then processed prior to inversion.

The following workflow was employed for the data processing:

1. Import data to Workbench using contractor supplied line files and an associated file with system specifications including system waveform, number of channels/gates or windows, number of turns in the transmitter, filters etc.;

2. Import of SRTM 1 sec elevation data and application to the AEM data set 3. Manual editing of the automatically processed data to remove any cultural and late time noise which

the automatic filter settings did not account for; 4. Run preliminary lateral constrained inversion to assess if the manual processing was adequate, and to

help choose the optimal inversion model parameters


Table 2: TEMPEST window information for 25Hz base frequency

Window # Start sample

End sample

No of samples

start time (s) End time (s) centre time (s)

centre time (ms)

1 1 2 2 0.000007 0.000020 0.000013 0.013 2 3 4 2 0.000033 0.000047 0.000040 0.040

3 5 6 2 0.000060 0.000073 0.000067 0.067

4 7 10 4 0.000087 0.000127 0.000107 0.107

5 11 16 6 0.000140 0.000207 0.000173 0.173

6 17 26 10 0.000220 0.000340 0.000280 0.280

7 27 42 16 0.000353 0.000553 0.000453 0.453

8 43 66 24 0.000567 0.000873 0.000720 0.720

9 67 102 36 0.000887 0.001353 0.001120 1.120

10 103 158 56 0.001367 0.002100 0.001733 1.733

11 159 246 88 0.002113 0.003273 0.002693 2.693

12 247 384 138 0.003287 0.005113 0.004200 4.200

13 385 600 216 0.005127 0.007993 0.006560 6.560

14 601 930 330 0.008007 0.012393 0.010200 10.200

15 931 1500 570 0.012407 0.019993 0.016200 16.200

2.2.1 Topographic data

A digital elevation model for the area was required to relate the inverted conductivity-depth models to topography. For the areas of interest several options were available, specifically the one second Shuttle SRTM DEM for the area or the DTM derived from the TEMPEST survey itself. As related work employed the SRTM data, those data were employed.

2.2.2 Processing of amplitude response data

The processing of the amplitude data supplied by the contractor undergoes a two-step process. The first is the application of an automated process, and the second a more time consuming manual process

Automated data processing

The automatic data processing applied to the TEMPEST data set involved the application of an averaging trapez filter to aid removal of late time noise, and the choice of the sounding distance (distance between soundings) for the averaging of data. The width of the trapez filter was chosen so that it would average noise while still honouring the lateral structures seen in the amplitude response data. The data as supplied from the contractor has already been through some averaging and therefore the filters have been kept as narrow as possible while still allowing the late times to be averaged adequately to improve the signal to noise ratio, keeping in mind the need to retain as much information as possible.

Manual processing

During the manual processing step results from the application of automatic filters are assessed and noise still left in the data are then removed manually. This is done for both late time noise and noise caused by anthropogenic features such as powerlines or buildings. Figure 4, shows amplitude response data for part of a TEMPEST flight line. The greyed out areas are where data has been filtered (manually or automatically)


because it is interpreted as being noisy. These data have to be removed to ensure that when they are inverted, the modelling doesn’t attempt to fit a conductivity model to noise. If they are left in erroneous conductivity models can result.

Figure 4. TEMPEST Z component amplitude response (top set of curves) and X component amplitude response data (bottom set of curves) for part of a flight line. The greyed out areas are where noise has been removed either through automatic filtering or through a manual process.

2.3 AEM data inversion

Given that one of the primary objectives of the Walyarta study was to map and interpret spatial variations in conductivity it was necessary to convert measurements of the electromagnetic response to ground conductivity, and present the results in a form that allowed their ready analysis against existing or new ground data. This is achieved through a process known as inversion.

The inversion of AEM data, and their presentation as maps or sections detailing a conductivity distribution in the subsurface is now commonplace. The representation of essentially continuous and gradational conductivity distributions as discrete conductive “units” or bounding layers is an effective way of summarising information from large AEM surveys. It enables users to quickly and reliably visualise conductivity variations in the subsurface in three dimensions (Lane 2000, 2002, Lane and Pracilio, 2000).

To determine true conductivity variation with depth the data must be modelled. This entails taking data from each measurement point (also referred to as a sounding or fiducial) along a flight line and estimating the parameters of a layered-earth, conductivity-depth model, which would produce the observed response. Through the application of approximate transforms or layered inversions, conductivity values with depth can be calculated for each observation (sounding) made by the AEM system, and then stitched together into sections to provide a representation of the two dimensional variation of conductivity. This is sometimes referred to as a “para-section”. Further, the conductivity depth profiles can be combined into a 3D gridded volume from which arbitrary sections, horizontal depth slices (or interval conductivity images) and iso-surfaces can be derived. The schematic in Figure 5 summarises the process of acquiring AEM data, inverting the resulting data, and presenting the results as conductivity images.


2.3.1 The 1D assumption

Airborne electromagnetic (AEM) data acquired for exploration or environmental applications are commonly modelled using algorithms such as conductivity depth transforms (CDT’s) or Layered Earth Inversions (LEI’s) that assume a one dimensional (1D) Earth (Sattel 1998, 2005). Presently, the application and relevance of full 2 or 3D inversion of AEM data remains undetermined, although a recent studies (are indicating these approaches may have application in areas of complex geology (see, for example, Yang et al. 2014, Silic et al. 2015). It may be unnecessary to consider the application of these higher order inversion approaches, particularly for hydrogeological investigations in many Australian basins, where it is reasonable to assume that the subsurface can be represented as a series of horizontal layers. The 1D model assumption produces results that are only slightly distorted by 2D or 3D effects which may be induced by faults, fractures, or other geological phenomena (Auken et al., 2005; Newman et al., 1987; Sengpiel and Siemon, 2000). The geology of Walyarta area comprises a sub-horizontal, layered sedimentary sequence which on the local scale is laterally contiguous and extensive. However it is also known to be characterised by fault blocks that may juxtapose conductive sedimentary units against those being more resistive. In such situations the application of 1D inversions may crate artefacts at these boundaries. If finer scale information is required around the edges of fault blocks it may be appropriate to consider the application of higher order inversion codes. However, for the broad scale definition of the subsurface conductivity structure, the geology of the area suggests that it is a good environment in which to apply an inversion with a 1D assumption.

Figure 5: Schematic representation of TEMPEST fixed-wing time domain EM data acquisition and interpretation. A) Data are acquired along parallel flight lines, with data recorded at fixed time intervals (soundings or fiducials) along each flight line; B) The EM receiver located behind and below the transmitter measures the secondary responses from the ground; C) The measured response is used to determine the conductivity-depth function by transformation or inversion. Conductivity-depth values can be calculated for each observation, taking account of the elevation of the system above the ground; D) The modelled conductivity structure for each sounding is then stitched together into conductivity-depth sections to provide a representation of the 2D variation of conductivity, sometimes referred to as a “para-section”. Conductivity depth models can be gridded to produce horizontal depth slices (or interval conductivity images) to show the spatial distribution of conductivity as it varies with depth. These maps can be shown as elevations (mAHD) or as depth intervals below the ground surface ground surface.


2.4 Inversion model

The Mandora AEM survey was inverted using a smooth and blocky (few-) layer model.

The smooth model inversion typically consists of 15-30 layers with fixed thicknesses, often increasing with depth. The amount the conductivity of one layer can vary to the next is defined by a vertical constraint. The large number of layers and the gradual change in conductivity in this type of model makes the resulting conductivity models appear continuous. This in turn can make it difficult to pick layer boundaries as these may appear rather diffuse.

Blocky (few-) layer inversions typically comprise 2-7 layers with the layers varying thickness and conductivity. These models can appear abrupt and whilst layer boundaries may be easier to define, sometimes the inversion assigns layers to areas that can be difficult to interpret. A comparison of results from the different approaches is shown in Figure 6 for a line of TEMPEST data acquired in the study area.

Figure 6: Comparison between stitched conductivity sections for a smooth 30-layer (top panel) and a 7 layer blocky (lower panel) 1D inversion for the same line of TEMPEST data in the study area.

For the purposes of this study, two models were generated after assessing the value of different inversion strategies with DBCA. For the east-west lines (Figure 2), a 25 layer (smooth) model was used for the inversion. The first layer thickness was chosen to be 5m with logarithmically increasing thicknesses to a depth of 500m which is the depth of the last layer boundary. The starting model (of 40 Ωm) was a homogenous half space with an auto calculated conductivity, which is calculated as the mean of the apparent resistivity for each sounding. The regularisation constraints (smoothness constraints) were set to a vertical constraint of 3, a value which allows some vertical structure, without introducing artefacts caused by overfitting the data. The horizontal (lateral) constraint was set to 1.8 for all layer intervals.

For the shorter north-south survey lines over Walyarta (Mandora Marsh) (Figure 2) a 19 layer (smooth) model was used for the inversion, with the first layer thickness was chosen to be 5m, logarithmically increasing in thicknesses to a depth of 300m which is the depth of the last layer boundary. A similar starting model to that used with the east-west lines was employed.


2.5 Laterally constrained inversion (LCI)

The TEMPEST datasets were inverted using the laterally constrained inversion (LCI) methodology employing the AarhusInv 1D inversion code (Auken and Christiansen 2004; Auken et al. 2005). The spatial constraints, which are defined for adjacent soundings, allow prior information (e.g., the expected geological variability of the area) to migrate along the flight lines (Figure 7). The use of constraints along lines enhances the connection of layer parameters between adjacent soundings. In the context of the Mandora area this approach encourages the definition of laterally continuous conductive layers which is an aid to target definition and geological interpretation.

2.6 Depth of investigation

The presentation of conductivity models derived from AEM systems can be misleading if there is no attempt made to qualify the depth of investigation (DOI) of the measurement system. The depth of investigation is a complex quantity, being a function of the power, sensitivity and accuracy of the acquisition system, environmental noise levels (e.g. sferic and powerline sources), geologic complexity, the host conductivity and the target characteristics (e.g. a discrete object or an extensive layer, conductivity contrast to the surrounds) and the inversion procedure used (Lane 2000; Christiansen and Auken 2012). To ensure that the observed variations in measured conductivity reflect changing ground conditions, rather than inversion or model dependent changes arising from the inversion process, an estimate of the depth of investigation is calculated and presented on the conductivity-depth sections. This information assists the interpreter, helping to quickly evaluate the results and their validity. The DOI provides a depth to which the model is the most reliable, and model information below the DOI should be used with caution.

The DOI determination used here is based on the cumulative sensitivity of the actual model output from the inversion (it includes the full system response and geometry) and is described in Christiansen and Auken (2012). The data noise and the number of data points are integrated into the calculation, which is based on the final inversion model output, and a recalculated sensitivity (Jacobian) matrix. In general terms the more conductive the ground, the ability to resolve deeper variations in conductivity (or the depth to which the inverted model is reliable) decreases. In more resistive ground, the system is able to resolve those variations more reliably to greater depths.

2.7 Results

Results were presented as conductivity-depth intervals below the ground surface and as conductivity–depth sections, thereby providing a spatial picture of changes in ground conductivity across the survey area. Stitched 1D conductivity depth sections were generated for each line of TEMPEST data within the study area.

An example of a stitched section resulting from the inversion of one of the north-south lines is shown in Figure 8. A brief comparison between the initial 1D Inversion results for the TEMPEST data over Mandora Marsh (see Mira Geoscience 2015) and the optimised 1D inversion undertaken as part of this study was undertaken. A representative conductivity section for one north-south line over the Marsh is shown in Figures 9 and 10.


Figure 8: Stitched conductivity-depth section for a north-south line (line 4002804). The depth of Investigation (DOI) is defined by the black line below which the section has an opaque overlay. The conductivity model structure below this line should be interpreted with caution. As an example, the deep conductive unit between 500 and 5000m along the line could not be reliably interpreted as being stratigraphically related as it falls below the DOI. Similar features were excluded from a stratigraphic interpretation of the data.

Figure 7: A diagrammatic representation describing the principle of laterally constrained inversion. Individual conductivity models of the subsurface, derived from the inversion of individual soundings measured by the AEM system, are laterally correlated in the along-line direction to enhance the definition of laterally continuous conductive layers.


Although the overall modelled structure is comparable, small but significant differences are observed in the near surface with the AarhusInv results placing shallow conductive layers associated with the occurrence of salt lakes at the surface, as is expected in such environments.

Figure 9: A comparison between smooth model 1D inversion results for Mandora Marsh. The top panel shows a stitched 1D conductivity-depth section derived from a VPem1D inversion (Fullagar and Pears, 2007; Fullagar et al., 2013). The lower panel shows the results from application of the AarhusInv inversion code ( Auken et al. 2015) for the same line, with a depth of investigation also defined. The lower inversion results are fitted to the 1 sec SRTM DEM. The two sections are scaled over the same colour range.

VPem1D inversion section

AarhusInv 1D LCI inversion section

Depth of investigation


Figure 10: A comparison between smooth model 1D inversion results for the same line over Mandora Marsh. The top panel shows a stitched 1D conductivity-depth section derived from a VPem1D inversion (Fullagar and Pears, 2007; Fullagar et al., 2013). The lower panel shows the inversion results from application of the AarhusInv inversion code ( Auken et al. 2015) for the same line, with a depth of investigation also defined. The lower inversion results are fitted to the 1 sec SRTM DEM. The two sections are scaled over the same colour range.

Interval conductivity images for all layers calculated in the inversion were generated. Examples of these results are presented in Figures 11-14. Examples of the spatial conductivity structure of the sub-surface are shown for the very near surface and deeper.

VPem1D inversion section

AarhusInv 1D LCI inversion section


Figure 11: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from below the surface for 10m intervals, starting from surface to 40m. The pseudocoloured conductivity images are overlain on the SRTM DEM.


Figure 12: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from below the surface, starting at 40m and extending to 80m. The pseudocoloured conductivity images are overlain on the SRTM DEM.


Figure 13: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from 80 – 120m below the surface. The pseudocoloured conductivity images are overlain on the SRTM DEM.


Figure 14: Representative interval conductivity images from inverted TEMPEST AEM data over the Mandora Marsh study area. Depths are from 120 – 150m below the surface. The pseudocoloured conductivity images are overlain on the SRTM DEM.


3 References

Auken, E. and Christiansen, A. V., 2004, Layered and laterally constrained 2D inversion of resistivity data. Geophysics, 69, 752-761.

Auken, E., Christiansen, A. V., Jacobsen, Bo. H., Foged, N. and Sørensen, K.I., 2005, Piecewise 1D laterally constrained inversion of resistivity data. Geophysical Prospecting, 53, 497–506.

Auken, E., Christiansen, A. V., Kirkegaard, C., Fiandaca, G., Schamper, C., Behroozmand, A. A., Binley, A., Nielsen, E ., Effersø, F., Christensen, N. B., Sørensen, K., Foged, N. and Vignoli, G., 2015, An overview of a highly versatile forward and stable inverse algorithm for airborne, ground-based and borehole electromagnetic and electric data. Exploration Geophysics, 46, 223–235.

Beckitt, G., 2003, Exploration for unconformity uranium in Arnhem Land (NT). Exploration Geophysics, 34, 137–142.

CGG, 2015, Logistics and Processing Report, Department of Water TEMPEST Geophysical Survey, West Canning Basin Survey, Western Australia: Unpublished Report to Department of Water Western Australia, CGG Job # 2474.

Chen, J. and Raiche, A., 1998, Inverting AEM data using a damped eigen parameter method. Exploration Geophysics, 29, 128–132.

Christensen, N.B., 2002, A generic 1-D imaging method for transient electromagnetic data. Geophysics, 67, 438-447.

Christiansen, AV and Auken, E, 2012. A global measure for depth of investigation: Geophysics, 77, WB171–WB177.

Craig, M.A. ed., 2011, Geological and energy implications of the Pine Creek region airborne electromagnetic (AEM) survey, Northern Territory, Australia. Geoscience Australia Record 2011/18, 292.

Davis, A.C., Munday, T.J. and George, R.J., 2015, Gascoyne River airborne electromagnetic survey II: Inversion Report. CSIRO: Water for a Healthy Country Flagship Technical Report; CSIRO Technical Report EP-13-11-758. 158pp.

Farquharson, C.G. and Oldenburg, D.W., 1998, Non-linear inversion using general measures of data misfit and model structure. Geophysical Journal International, 134, 213-227.

Fitzpatrick, A., Munday, T.J., Cahill, K. and Stelfox, L., 2011, A regional, reconnaissance scale AEM survey to aid the definition of surface water - groundwater processes beneath a large unregulated river system: Fitzroy River, Western Australia. CSIRO Technical Report No: CESRE P2010/1235.

Fullagar, P. K. and Pears, G. A., 2007, Towards geologically realistic inversion: In Milkereit, B., (Ed.), Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration, 444-460.

Fullagar, P.K., Pears, G.A., and Reid, J.E., 2013, Hybrid 1D/3D geologically constrained inversion of airborne TEM data: Extended Abstract, Australian Society of Exploration Geophysicists 23rd Conference, Melbourne.

Halkjaer, M, Sorensen, KI, Christiansen, NB, and Auken, E, 2006. SkyTEM – Status and development. Australian Earth Sciences Convention 2006, Melbourne, Australia

Lane, R., and Pracilio, G. 2000. Visualisation of Sub-Surface Conductivity Derived From Airborne EM. SAGEEP 2000 Conference Proceedings.


Lane, R., Green, A., Golding, C., Owers, M., Pik, P., Plunkett, C., Sattel, D., Thorn, B., 2000, An example of 3D conductivity mapping using the TEMPEST airborne electromagnetic system: Exploration Geophysics, 31, 162-172.

Ley-Cooper, A.Y., Munday T.J. and Ibrahimi, T., 2017, Inversion of the Capricorn Orogeny regional Airborne Electromagnetic (AEM) survey, Western Australia. CSIRO Technical Report EP166290, pp. 37.

Mira Geoscience, 2015, Inversion and preliminary interpretation of airborne electromagnetic data from the West Canning Basin – Sandfire TEMPEST survey, Mira Geoscience Project #4399, report prepared for the Department of Water, Western Australia, 86pp.

Newman, G. A., Anderson, W. L., and Hohmann, G. W., 1987, Interpretation of transient electromagnetic soundings over three-dimensional structures for the central-loop configuration. Geophysical Journal of the Royal Astronomical Society, 89, 889-914.

Roach, I.C. ed., 2010, Geological and energy implications of the Paterson Province airborne electromagnetic (AEM) survey, Western Australia. Geoscience Australia Record 2010/12, 318.

Roach, I.C. ed., 2012. The Frome airborne electromagnetic survey, South Australia: implications for energy, minerals and regional geology. Geoscience Australia Record 2012/40 – DMITRE Report Book 2012/00003, 296 pp.

Sattel, D., 1998, Conductivity information in three dimensions. Exploration Geophysics, 29, 157-162.

Sattel, D., 2005, Inverting airborne electromagnetic (AEM) data with Zohdy’s method. Geophysics 70 (4), G77-G85.

Sattel, D. and Kgotlhang L., 2004, Groundwater exploration with AEM in the Boteti Area, Botswana. Exploration Geophysics, 35, 147-156.

Sengpiel, K. P., and Siemon, B., 2000, Advanced inversion methods for airborne electromagnetic exploration. Geophysics, 65, 1983–1992.

Silic, J., Paterson, R., FitzGerald. D. and Archer, T., 2015. Comparing 1D and 2.5D AEM inversions in 3D geological mapping using a new inversion solver. 14th International, Congress of the Brazilian Geophysical Society, Extended Abstracts.

Yang, D., Oldenburg, D. and Haber, E., 2014, 3-D inversion of airborne electromagnetic data parallelized and accelerated by local mesh and adaptive soundings, Geophys. J. Int., 196 (3), 1492-1507


CONTACT US t 1300 363 400 +61 3 9545 2176 e [email protected] w www.csiro.au

AT CSIRO, WE DO THE EXTRAORDINARY EVERY DAY We innovate for tomorrow and help improve today – for our customers, all Australians and the world. Our innovations contribute billions of dollars to the Australian economy every year. As the largest patent holder in the nation, our vast wealth of intellectual property has led to more than 150 spin-off companies. With more than 5,000 experts and a burning desire to get things done, we are Australia’s catalyst for innovation. CSIRO. WE IMAGINE. WE COLLABORATE. WE INNOVATE.

processing and inversion of walyarta (mandora marsh) aem data · 2.1 tempest aem system the tempest...

Documents