ASEG Extended Abstracts - 25th International Conference and Exhibition – Interpreting the Past, Discovering the Future, 2016

A significant scientific and engineering challenge for the energy resources industry is to monitor injected or produced fluid at depths of hundreds or thousands of metres, and over time-scales of hours to years. A new approach using surface magnetotelluric (MT) methods has been developed over the last five years to map deep-fluid pathways by virtue of their electrical resistivity changes, both spatially and temporally. This is a cheap technology as it uses natural electromagnetic source-fields and does not require drilling. However, is this method really effective for industry for economic reasons and for social and environmental compliance? In other words, is it a disruptive technology or a damp squib?

This paper reviews the physics of the approach, and demonstrates the feasibility of the MT method for monitoring unconventional energy resource development. A number of case studies will be shown, including shallow coal seam gas de-pressurisation, deep hydraulic stimulation of a shale gas reservoir, and enhanced geothermal system development.

Water bores in Australia are important capital assets used for local town water supply, agricultural stock supply and irrigation or simply as an indicator of the region’s aquifer health. Water bore construction methods used in previous decades has led to a pervasive problem of surface leakage and/or sub surface leakage and aquifer contamination particularly in relation to the Great Artesian Basin. Wireline logging methods are available to assess if a new water bore meets current design requirements to prevent leakage/contamination issues (hence certification) or to assess the current internal and “in situ” condition of old water bores. Knowledge of the current “in situ” condition of the water bore will be able to direct a rehabilitation workover programme. Methods range from a simple 3 arm caliper for internal casing inspection and full wave sonic logging is assess the presence/absence of cement in the annulus of the casing through to high resolution acoustic (multi-fingered caliper) and optical imaging of the internal casing. Methods are discussed and examples provided.

A significant scientific and engineering challenge for the energy resources industry is to monitor injected or produced fluid at depths of hundreds or thousands of metres, and over time-scales of hours to years. A new approach using surface magnetotelluric (MT) methods has been developed over the last five years to map deep-fluid pathways by virtue of their electrical resistivity changes, both spatially and temporally. This is a cheap technology as it uses natural electromagnetic source-fields and does not require drilling. However, is this method really effective for industry for economic reasons and for social and environmental compliance? In other words, is it a disruptive technology or a damp squib?

This paper reviews the physics of the approach, and demonstrates the feasibility of the MT method for monitoring unconventional energy resource development. A number of case studies will be shown, including shallow coal seam gas de-pressurisation, deep hydraulic stimulation of a shale gas reservoir, and enhanced geothermal system development.

We present a status report on the next generation data assimilation techniques for Resource Discovery using a new multidisciplinary fundamental science approach. We combine a recent multiphysics, multiscale geodynamic theory with laboratory and modern computational assisted petrophysics and material science concepts with the aim of linking it “on the fly” to geological and geophysical field data acquisition. This solid science base is designed to build the platform for enabling a data intensive paradigm for the resource industry. Such a physics-based big data interpretation opportunity has not yet been exploited in current geoscience applications. In other disciplines the approach is, however, fully realized. Owing to its major impact it has been hailed by the US National Research Council [ICME, 2008] as a transformational discipline for improved competitiveness and national security. The approach has been pioneered in polymer sciences, the automotive and aerospace engineering as well as in computational biomechanics, via a so-called “Integrated Computational Materials Engineering” (ICME) cyber-infrastructure. An ICME system unifies materials information into a multi-scale system that is linked by means of software integration tools to a designer knowledge base containing tools and models from different scales and different disciplines.

The identification and evaluation of tight gas sandstone reservoirs faces a great challenge due to the characteristics of ultra-low porosity, ultra-low permeability, complicated pore structure and strong heterogeneity. To improve tight gas sandstone reservoir evaluation, the pore structure should be first quantitatively evaluated. NMR logs are advantageous in formation pore structure evaluation, but may not always be practicable as they are only acquired in limited wells. This study is based on the analysis of morphological characteristics of experimental mercury injection capillary pressure (MICP) curves for 54 core samples from tight gas sandstone reservoirs of central Sichuan Basin, southwest China. A new method is proposed to construct pseudo capillary pressure curves from conventional porosity and permeability. The corresponding models of predicting pseudo Pc curves from conventional logs are established. The reliability of this technique is verified by comparing constructed capillary pressure curves with laboratory measured results. After this technique is extended to field application, consecutive pseudo capillary pressure curves are constructed, and the parameters associated with reservoir pore structure evaluation, such as pore throat radius distribution, the average pore throat radius, the threshold pressure, and so on, are estimated. Combining with these parameters, tight gas sandstone reservoirs pore structures are quantitatively evaluated, and effective formations can be easily identified. The proposed technique is more universally applicable in tight gas sandstone reservoir pore structure evaluation as it does not require the acquisition of field NMR logs.

Based on the simultaneously applied mercury injection capillary pressure (MICP) and nuclear magnetic resonance (NMR) laboratory experimental results for 20 core samples from tight gas sandstone reservoirs of the Sichuan basin, the relationships of the piecewise power function between nuclear magnetic resonance (NMR) transverse relaxation T2 time and capillary pressure (Pc) are established. A novel method, which is used to transform NMR reverse cumulative curve as pseudo capillary pressure (Pc) curve is proposed, and the corresponding model is established based on formation classification. By using this model, formation pseudo Pc curves can be consecutively synthesized. The pore throat radius distribution, and pore structure evaluation parameters, such as the average pore throat radius (Rm), the threshold pressure (Pd), the maximum pore throat radius (Rmax) and so on, can also be precisely extracted. After this method is extended to field applications, several tight gas sandstone reservoirs are processed, and the predicted results are compared with core derived results. Good consistency between evaluated results with core derived results illustrates the dependability of the proposed method. Comparing with the previous methods, this presented model is much more theoretical, and the applicability is much improved.

Three major developments provide improvements in the spatial resolution, long wavelength response and noise reduction in airborne gravity gradiometer data. The world’s first commercial strap down gravimeter, sGrav, has achieved the same levels of data quality at long wavelengths as stabilized platform systems. With a repeatability of 0.71 mGal RMS at 300 s filtering, it is ideally suited to augment the long wavelength airborne gravity gradiometer data. The sGrav data is incorporated into the airborne gravity gradiometer processing stream by a modified conforming technique that first converts the gravity data to gravity gradients and then merges the result with the gradients from the gravity gradiometer. This improved process has advantages for conforming gravity gradiometer data to any regional gravity data.

Processing gravity gradient data routinely involves low-pass filtering which limits the spatial resolution. A new processing method increases this resolution by splitting the acquisition noise from the geologic noise and then removing the acquisition noise. This method has reduced noise amplitude densities, in some cases by 50%.

Downhole NMR data was collected through a fibreglass cased section of water well NG3A that intersected the semi-consolidated sediments of the Yaragadee Formation which is one of the main aquifers supplying domestic water for the city of Perth. The main purpose of this work was to empirically derive formation specific NMR processing parameters to match the hydraulic property estimates to a direct core measurements, to calibrate the system to deliver a more detailed log of porosity and permeability in the hole as well as other holes logged in a similar geologic environment.

Equations for NMR permeability estimation generally include lithology specific calibration coefficients. We show that application of “generic” calibration coefficients derived for unconsolidated aquifer materials overestimate permeability in the Yaragadee Formation when compared to core permeability measurements. This result is expected given that the Yaragadee formation is not unconsolidated but is partially consolidated/cemented. More appropriate site specific coefficients were derived by scaling the calibration coefficients to produce NMR permeability estimates that match the measurements derived from the core samples. The site specific coefficients determined for the Yaragadee are consistent with those derived in previous studies for partially consolidated aquifers in the United States.

The NMR T2 distribution is very sensitive to pore size changes which often reflect subtle changes in the sedimentary geology of the formation. This can provide additional geological information which may not always be apparent in mud or core logging. Detailed knowledge of both the sedimentary geology as well as the hydraulic properties of the formation that can be obtained by NMR are likely to be useful in identifying the best place to place screens during the design of a production water well to generate the best possible yield.

The Leite target is located in Carajas Mineral Province and has a magnetic anomaly with 140 nT of amplitude, elongated in the northwest-southeast direction. Four exploratory drillholes were performed to test the magnetic anomaly. The test showed that the source of the anomaly is a narrow magnetite hydrothermal alteration zone bearing copper mineralization up to 2%. In addition, geologic and geochemical data, magnetic susceptibility (MS) measurements were collected to identify the lithotypes with ferromagnetic minerals. We use three different techniques to estimate the depth and geometry of the magnetic source: standard Euler deconvolution, total field magnetic anomaly modeling, and magnetic amplitude inversion. When visualized in 3D, the depth of solutions from Euler deconvolution crossed the real magnetic layer with less inclination. The modeling, using the solutions from Euler deconvolution, was performed, and the magnetic anomaly produced by the body modelled achieved a low misfit. The body used in the forward modeling is geometrically similar to the geologic magnetic layer. The magnetic amplitude inversion successfully recovered the MS distribution. Finally, we carried out a borehole magnetic survey in two drillholes to validate the obtained models and investigate the magnetic source. This survey confirmed that the models were intercepted and the magnetic anomaly was associated, a hydrothermal alteration zone, with magnetite intercepted by drillholes. In this study, we demonstrated that the use of those techniques was effective in Greenfield exploration programs

The growth, interaction and controls of gravity-driven normal faults is somewhat understudied. Using three-dimensional (3D) and two-dimensional (2D) seismic reflection data, located at the present-day shelf-edge break and into the deepwater province of the Otway Basin, southern Australia, we aim to temporally and spatially constrain the development of a normal fault system and determine the controls on growth. The Otway Basin is a Late Jurassic to Cenozoic age, rift-to-passive margin basin. The seismic reflection data images a gravity-driven fault array, consisting of ten fault segments, striking NW-SE (128-308), located within Upper Cretaceous clastic sedimentary rock. We analyse the growth of a gravity-driven hard-linked fault assemblage interacting with basement normal faults. Our analysis shows that the fault assemblage is linked to major basement faults and displays Turonian-Santonian nucleation, continued growth until the latest-Maastrichtian and a maximum throw of 1.74 km. High variability of throw along-strike and down-dip of the fault assemblage indicates growth via lateral and vertical segment linkage. We interpret that the spatial and temporal evolution of the fault assemblage is the result of rifting basement fault control during Upper Cretaceous resumed crustal extension in the Otway Basin. The control of the rifting basement faults on these gravity-driven normal faults has implications towards the growth and petroleum prospectivity of gravity-driven normal faults on passive margins such as the Niger Delta and Gulf of Mexico, but also towards gravity-driven normal faults developed in supra-salt sedimentary rock in rift basins, such as the North Sea and Suez Rift.

Modelling of magnetic rock properties from magnetic field observations has been an important practice in resource exploration for decades. However, the application of this practice has been limited by conventional thinking that assumes rock magnetization is dominated by induced magnetization such that magnetization direction is aligned with the geomagnetic field. Convention has also accepted that we are unable to model for magnetic remanence without a-priori knowledge of remanence direction and strength.

Recent practical successes in directly modelling magnetization vector direction and strength using Magnetization Vector Inversion (MVI) have challenged these conventions, and MVI modelling is proving useful in practical exploration scenarios. The addition of new information, namely the direction and amplitude of magnetization, demands new thinking and approaches to understanding what this information means, and how to use the modelled direction of magnetization in practical situations.

This paper presents a new statistical and quantitative approach to define and discriminate different magnetization domains within a full 3D MVI voxel model. Our studies show that modelled vector direction is meaningful even without prior knowledge of remanence (and other) magnetization characteristics. We also demonstrate that reasonable magnetization direction can be recovered from both weakly and strongly magnetized source rocks.

Tectonically deformed coal is a key factor affecting the phenomena of gas outbursts in coal mines. Seismic attributes associated with coal-bed reflections can represent this kind of coal, but the representation is indirect and ambiguous since both coal-bed thickness and lithology can affect the attributes. We propose a model-based function to calibrate seismic attributes from thin coal-bed reflections, and use the calibrated attributes to estimate the distribution of tectonically deformed coal. To eliminate the influence of coal-bed thickness on seismic attributes, we first build a synthetic model to simulate true geological condition of the coal bed. Then, we extract seismic attributes from the synthetic section and cross plotted the correlations between attributes and thickness. After the fitting of cross-plotted correlations with a Fourier series, we estimate the calibration functions of the attributes. Finally, true attributes of a mining zone are calibrated with those functions. The results presented here show much more observable correlation with the thickness of tectonically deformed coal than uncalibrated attributes.

Seismic facies mapping is an underutilized tool for getting extra geology in to your exploration, or development. It is a key component of palaeogeography mapping that help show different types and locations of seals and reservoirs.

Both manual and automated facies mapping, on 2D and 3D seismic, when calibrated to local wells, can help de-risk crucial components of a play.

In many areas today where deep water or new frontier exploration is undertaken, there is often little or no well control and seismic facies mapping, allied to analogues, may be your only tool for getting at the geology in your palaeogeography maps. This talk will apply geological models to, and discuss, the construction of manual seismic facies maps which are then used to guide automated facies map work.

Manual facies mapping is undertaken once sequence boundaries have been mapped and named. Each different facies identified is given the name of upper and lower boundaries and the facies name and then the horizon is picked so that the facies extent is shown on a base map. Once completed, polygons encompass this facies. The final map uses all polygons constructed, along with well logs for that sequence.

Automated facies maps start with RMS and maximum amplitude maps of the sequence, then possibly of proportional slices within the sequence and then we may construct waveform facies maps or spectral decomposition maps of the interval. All these maps are considered with the interval’s isochron map and at least gamma log windows at wells, with appropriate geology transcribed to finally construct the palaeogeography map of the interval.

Studies have shown that Grounded Electrical-Source Airborne Transient ElectroMagnetics (GREATEM) is a promising method for resistivity structures investigating in coastal areas, in addition to inaccessible areas such as volcanoes, mountains and deep forest cover. To expand the application of the GREATEM system, a three-dimensional (3-D) resistivity model that considers large lateral resistivity variations is required. In this paper, we present a frequency- domain 3-D electromagnetic (EM) inversion approach that can be applied to time domain data from GREATEM. In the frequency-domain approach, TEM data were Fourier-transformed using a smooth-spectrum inversion method, and the recovered frequency response was then inverted. To deal with a huge number of grids and a wide range of frequencies in airborne datasets, a method for approximating sensitivities is introduced for efficient 3-D inversion. Approximate sensitivities are derived by replacing adjoint secondary electric fields with those computed in the previous iteration. These sensitivities can reduce the computation time without significant loss of accuracy. Firstly, we verified both of our forwarding and inversion solutions. We then applied this approach to the GREATEM survey data from Kujukuri beach, central Japan. The inverted results of the field data are well fit with the previous study results at Kujukuri area, suggesting the applicability of this inversion approach for constructing 3D resistivity models from the GREATEM field survey data in the future.

Mistakes in processing gravity data lead to errors in the final product. This can mean that overlapping gravity surveys are often incompatible, and can lead to incorrect geological interpretations. In this paper I demonstrate the magnitude of the errors introduced at various stages of the gravity reduction process. I have focussed on errors relating to calibration factors, time zones and time changes, height, geodetic datums, gravity datums, and the equations involved therein. The errors range from below the level of detection, to many milligals.

The results highlight the need to not only be diligent and thorough in processing gravity data, but also how it is necessity to document the steps taken when processing data. Without properly documented gravity surveys they cannot be reprocessed should an error be identified.

In 1998 Denmark initiated a national groundwater mapping campaign in order to obtain knowledge of the aquifers with respect to their location, distribution, extension, interconnection and to acquire maps detailing groundwater vulnerability. The aim was to establish site-specific groundwater protection zones to prevent groundwater contamination from urban development and agricultural sources in agreement with The EU Water Framework Directive. The mapping campaign involved a dense data acquisition typically comprising boreholes, electromagnetic surveys - both airborne and land based and geoelectrical surveys. The data serve as basis for constructing 3D hydrogeological and groundwater models from which site-specific protection zones are establish. At present time dense geophysical mapping covers approximately 45 % of Denmark.

Based on a dense Airborne ElectroMagnetic (AEM) survey in combination with boreholes, three fault systems in the northern part of the island of Langeland, Denmark are mapped. Two of the fault systems were unknown prior to the mapping campaign. The two unknown fault systems are interpreted as a normal fault and graben structures, respectively. The presence of the hanging-wall block in the fault systems can be observed in the AEM data as a low resistivity layer that clearly distinguish from the underlying and surrounding high resistivity fresh water saturated limestone (footwall block) and the overlying glacial clay till. Soil descriptions from a borehole confirm that the low resistivity layer can be correlated to Palaeocene clay deposits. The fault systems were most likely initiated in the early Neogene during the Alpine orogeny. The fault systems are observed to alter the hydrology significantly and are therefore important to map.

Buried tunnel valleys are elongated depressions eroded into the substratum during the Pleistocene glaciations. Nine such valleys are mapped on- and offshore in a 300 km² area located at the Danish North Sea coast. The delineation of the buried valleys is based on an extensive data set consisting of on- and offshore 2D seismic data, TEM (Transient Electro-Magnetic) soundings, Schlumberger soundings, and boreholes. The valleys are observed as discrete incisions with three overall orientations: SSE - NNW, ESE - WNW, and SSW - NNE. They have depths between 75 and 185 m, widths up to 1.8 km, and lengths from 7 to 25 km. The infill comprises till, glaciofluvial sand, and glaciolacustrine clay and silt. Younger tunnel valleys are found to often re-use pre-established valleys generating cut-and-fill structures which are clearly revealed on the reflection seismic profiles. Cross-cutting relations, preferred orientations, and morphology support that three of the tunnel valleys cross the North Sea coastline. It is suggested that the nine valleys were formed during at least six events that occurred through one or more pre-Weichselian glaciations.

Magnetotelluric (MT) data were collected across the Habanero Enhanced Geothermal System project in the Cooper Basin, South Australia. A baseline regional MT survey consisting of two profiles was collected to delineate the pre-injection resistivity structure. Two dimensional inversions of the MT data reveal three main resistivity structures to a depth of 5 km. The low resistivity surface layer (about 1.5 km thick) is interpreted as poorly consolidated sediments of Lake Eyre and Eromanga Basins. Below the conductive layer, a zone with relatively high resistivity with thickness of 2 km can be correlated to consolidated Cooper Basin sediments. A high resistivity zone below depths of 3.5 km is interpreted as the hot intrusive granodiorite (granite) of the Big Lake Suite with low porosity and permeability. This deep structure is also related to the Habanero EGS reservoir.

The second MT survey was conducted during stimulation of Habanero-4 well by Geodynamics Ltd, where 36.5 ML of water with a resistivity of 13 Dm (at 25°C) was injected at a relatively continuous rate of between 27-53 L/s over 14 days at a depth of almost 4 km. Analysis of pre- and post-injection residual phase tensors for periods greater than 10 s indicate conductive fractures oriented in a N/NNE direction. Apparent resistivity maps also revealed that injected fluids possibly propagated towards N/NNE direction. This result is in agreement with the micro-seismic events with an area of 4 km observed during fluid injection, as well as orientation of pre-existing N-S striking sub-horizontal fractures susceptible to slip due to stimulation. The MT responses close to injection show on average 5% decrease in apparent resistivity at periods greater than 10 s. The main reasons for observing subtle changes in resistivity at Habanero EGS is the screening effect of the conductive thick sedimentary cover (about 3.6 km thick) and the presence of preexisting saline fluids with resistivity of 0.1 Dm (equivalent to salinity of 16.1 g/L at 240°C) in the natural fractures. Overall, the MT monitoring at Habanero EGS highlights the need for favourable geological settings to measure significant changes in resistivity in EGS reservoirs.

3-D resistivity monitoring surveys are used to detect temporal changes in the subsurface using the measurements repeated over the same site. The positions of the electrodes are measured at the start of the survey program and perhaps at occasional intervals. In areas with unstable ground, the positions of the electrodes can be displaced by ground movements. If this occurs at times when the positions of the electrodes are not measured, they have to be estimated from the resistivity data. The smoothness-constrained least-squares optimisation method can be modified to include the electrodes positions as additional unknown parameters. 3-D resistivity surveys present a special challenge due to the greater computational requirements for the forward modelling routine and the possible movements of the electrodes in three directions. To reduce the calculation time, a fast adjoint-equation method is used to calculate the Jacobian matrices required by the least-squares method. It is several orders of magnitude faster than the simpler perturbation method previously used for 2-D problems. In areas with large near-surface resistivity contrasts, the inversion routine sometimes cannot accurately distinguish between electrodes displacements and subsurface resistivity variations. To overcome this problem, the model for the initial time-lapse data set (with accurately known electrodes positions) is used as the starting model for the inversion of the later-time data set. This greatly improves the accuracy of the estimated electrode positions compared to the use of a homogeneous half-space starting model.