ASEG Extended Abstracts - ASEG2010 - 21st Geophysical Conference, 2010

Crosswell seismic is an emerging technology which provides highly detailed images of the subsurface at reservoir scale. The technology has the potential to delineate complex structure, help verify existing interpretation and also can be used to monitor changes in the reservoir caused by enhanced recovery strategies.

The technology employs tomographic surveying, whereby the transmitter and receiver are deployed in separate wells. Interwell structure and velocity can be obtained from reflection processing and direct wave processing respectively.

In this paper an example of crosswell seismic is discussed where the technology produced high-resolution images of the geology between wells at 8 to 10 times the resolution of surface 3-D seismic data and revealed numerous faulted compartments in the reservoir previously unseen in the surface seismic. These results have lead to a new understanding of the field and improved the well design process.

The fractured granite basement is the primary oil and gas reservoir in the Cuu Long Basin, Vietnam. Due to the complexity of this non-layered target, seismic data quality within the basement is very low. For these reasons, it is important to apply improved seismic data processing workflows to improve the fracture imaging quality.

Our studies in the fractured granite basement of the Cuu Long Basin show that, application of Radon and τ-p filters improves the signal-to-noise ratio by suppressing multiples, thereby revealing the top of the faulted basement. Using a multiarrival-solution Controlled Beam Migration further improves the signal-to-noise ratio, and helps image steeply dipping discrete fracture events. Applying geometric attributes such as apparent dip, energy gradient, and curvature further delineates these faults and fractures. Mapping fracture intensity and orientation assist delineating sweet spots and aid in planning horizontal wells.

For sustainable water management, understanding shallow aquifers and the unsaturated zone is critical. Therefore the spatial distribution of hydraulic properties is of great interest for development of accurate recharge distribution models. Logging of shallow boreholes and measurements made on soil samples provide an insight into hydraulic properties with depth. However, they do not provide an adequate image of the spatial variations of key parameters. This may lead to erroneous assumptions about overall distribution of soil properties. In this study, we illustrate how Ground Penetrating Radar (GPR) can be used to image layers that impede the recharge of shallow aquifers. We discuss some of the different attributes of these layers and show how these can be characterized by Ground Penetrating Radar over the Gnangara Mound.

Several hundred line kilometres of GPR have been acquired over the Gnangara Mound. Water retentive layers are easily identified and differentiated from the regional water table within the GPR sections. However, it is difficult to constrain the local 3D nature and the lateral extent of these layers from the very long sparse 2D GPR transects. Small pseudo 3D surveys at key locations have been completed. We demonstrate how these small pseudo 3D GPR surveys reveal the local consistency of water retentive layers and how the small high density surveys help understand the distribution of shallow hydraulic properties along the long transects.

Magnetotelluric data were acquired for Geoscience Australia by contract along the north-south 08GA-C1 Curnamona seismic traverse to the east of Lake Frome from November 2008 to January 2009 as part of the Australian Government’s Onshore Energy Security Program. 25 sites were spaced an average of 10 km apart, and five-component broadband data were recorded with a frequency bandwidth of 0.001 Hz to 250 Hz and dipole lengths of 100 m.

Apparent resistivity and phase plots are presented, along with dimensional analyses of the data based on rotational invariants, the representation of the data by the phase tensor, and Parkinson arrows. These analyses provide insight into the complexity of the Earth conductivity giving rise to the MT responses and are a useful precursor to modelling.

Early in the morning of the 29th of May 2006, hot mud started erupting from the ground in the densely populated Porong District of Sidoarjo, East Java. With initial flow rates of ~5000 cubic meters per day, the mud quickly inundated neighbouring villages. After almost four years, the ‘Lusi’ eruption has expelled over 73 million cubic meters of mud at an average rate of approximately 64000 cubic meters per day and at maximum rates of 170000m3/day. The mud flow has now covered over 700 hectares of land to depths of over 25 meters, engulfing a dozen villages and displacing approximately 40000 people. In addition to the inundated areas, other areas are also at risk from subsidence and distant eruptions of gas. However, efforts to stem the mud flow or monitor its evolution are hampered by an overall lack of knowledge and consensus on the subsurface anatomy of the Lusi mud volcanic system. In particular, the largest and most significant uncertainties are the source of the erupted water (shales versus deep carbonates), the fluid flow pathways (purely fractures versus mixed fracture and wellbore) and disputes over the subsurface geology (nature of deep carbonates, lithology of lithological unit between shales and carbonates). This study will present the first balanced overview of the anatomy of the Lusi mud volcanic system with particular emphasis on these critical uncertainties and their influence on the disaster.

The EMFlow program is an industry standard for the rapid transformation of AEM data to a conductivitydepth section. Written in the 1990’s, it provides a means to deconvolve data from any definable AEM system to the time-constant domain, and thereafter to predict conductivity-depth structures. The deconvolution to time-constant domain requires constraints for optimum performance: these constraints are user-controlled and can include causality, power-law or exponential extrapolation and geometrical amplitude constraints.

Geotem Deep and VTEM data collected over the Table Hill prospect in WA showed very small amplitude responses, which when processed with commercial EMFlow indicated an extensive conductor at depth. Initial drilling encountered manganese mineralisation consistent with this deep conductor. However, step-out drilling did not encounter similar mineralisation even though the holes were sited to intersect the imaged good conductor. The data were therefore reprocessed with an improved deconvolution method in an RMIT version of EMFLow. The specific improvement trialled was a novel matrix conditioning method that provided significantly more stable CDI results than simple matrix normalisation used in the commercial EMFLow program. The resulting CDI sections showed considerable detail within the deep conductive layer, where the most conductive region was consistent with the discovery hole, and less conductive regions were coincident with the lack of mineralisation in the step-out drilling. Improvements in data processing therefore were shown to provide very useful if not essential geological detail at depth.

In recent developments a full tensor magnetic gradient system has been deployed in South Africa. The instrument is made by IPHT (Institute of Photonic Technology, Jena, Germany) and flown by helicopter. In developing a ‘custom’ solution for taking the raw signal through to a final located geophysically sound data base, one of the biggest challenges has proven to be the compensation of aircraft / bird rotational movements, so that the reported magnetic curvature gradients in the world co-ordinate system are as free of these rotational errors as possible.

Issues that arise include drift correction of the Euler Angles from the INS (Inertial Navigation System), removal of flux jumps from the SQUID and recovering the three components of the ‘B’ field.

Compensation issues are limiting the achievable resolution of the field gradients.

A novel least-squares rotational adjustment of the tensor signal in a moving window is proposed as an extra compensation step to achieve higher spatial coherency of the curvature gradients.

The Australian geothermal industry is moving rapidly, and in that process requires a lot from geophysics to aid in characterising regional prospectivity for exploitable heat resources.

Various groups are using hybrid methods to estimate ‘Curie point’ temperatures at depth, or alternatively, the temperature at 5 kilometres below the surface. Deep drilling observations and airborne magnetic compilations are the key components, together with a basement geology interpretation. Several generations of this work are already published with more to come.

A method to test these maps and also help characterise uncertainty is proposed based upon a deep 3D continental model scale, extending to the lithosphere. Variable surface temperature and heat flow grids, based upon remote sensing are used, together with a simple lithosphere boundary condition. The heat diffusion is then employed to test the temperature-depth maps. Progress on applying this method to Australia is reported.

A deep seismic reflection profile and coincident refraction and magnetotelluric data were acquired across the northern Eyre Peninsula, Gawler Craton in 2008 to enhance the prospectivity of this region for uranium and geothermal energy, by establishing the crustal architecture and regional geodynamic framework. The seismic line crossed several tectonic domain boundaries, including the Kalinjala Mylonite Zone, which is a key structure in the southern and central Gawler Craton. In the vicinity of the seismic section, this mylonite zone dips moderately to the east and appears to be a major crustal-scale fault which cuts through to the Moho; it separates middle to lower crust of differing seismic reflectivity. The changes in seismic character correspond with the boundary between the Olympic Fe-oxide Cu-Au-U Province (host of Olympic Dam and Prominent Hill deposits) in the east, and the Central Gawler Au Province in the west, and also are suggested to limit the extent of the South Australian Heat Flow Anomaly.

During 1D inversion of TEM data it is desirable to constrain the inversion with all available information. Accordingly a program has been developed to perform geologically-constrained 1D TEM inversion. Three different types of inversion are permitted: geometry inversion, homogeneous property inversion, and heterogeneous unit inversion. During geometry inversion, the layer boundaries are allowed to move, except at fixed points, e.g. drilled contacts. During homogeneous property (“stratigraphic”) inversion, the conductivity of geological units is optimised, with each unit assumed uniform in conductivity across the entire survey area. During heterogeneous unit inversion, the conductivity is allowed to vary within geological units. These inversion options allow the interpreter flexibility to explore geological scenarios. The inversion program operates in a fully 3D geological context, but a 1D model is extracted at each location for adjustment. The method is illustrated via application to airborne TEM data recorded over shallow coastal waters. Inversion is constrained by sonar and marine seismic surveys, and by conductivity measurements of seawater and of sediment core. In this environment it is reasonable to assume that the sediment is fairly uniform in conductivity, and therefore an inversion sequence was designed to define the model with minimum intra-sediment conductivity variation. Seawater depth and conductivity were treated as known. After construction of a 3D starting model, inversion proceeded in three stages: first, the optimal uniform sediment conductivity was defined; then the geometry of the sediment-bedrock contact was adjusted; finally, internal conductivity variations were permitted in the sediment. At the end of this inversion sequence the surviving internal conductivity variations within the marine sediment are considered indicative of actual lateral variations in the sediment.

We present a revised set of Euler poles describing the relative motion between Australia and Antarctica from the onset of continental rifting at ~160-140 Ma to the reorganisation of the spreading system at ~50 Ma. Our revised reconstruction addresses two key issues that remain unresolved in current plate reconstructions.

Firstly, we present new estimates of the pre-rift plate boundary locations for the conjugate Australian-Antarctic margins. These reconstructions are truly palinspastic, incorporating estimates of crustal thickness along these margins, derived from gravity inversions. Integrating the crustal thickness along tectonic flowlines results in the pre-rift location of the continental plate boundary. This integration relies on defining the present-day extent of stretched continental crust within the margin, which is the subject of the companion paper, Williams et al. [2010]. Once restored, we are then able to use the pre-rift plate boundary positions to compute ‘full-fit’ poles of rotation for Australia relative to Antarctica. This approach allows us to model a deforming passive margin, with implications for understanding and modelling the formation of basins and deposition of sediments along passive margins.

Secondly, reconstructions for plate motions since ~83 Ma have been revised to obtain a better fit along the entire Australian-Antarctic conjugate margins, which extend from at least the Kerguelen Plateau and Broken Ridge in the west to Tasmania, Australia and Cape Adare, Antarctica in the east. Previously published reconstruction models for the period 83-50 Ma had resulted in a poor fit between the two plates at the extreme ends of the conjugate margins.

Geo-steering of the Longtom-3 and Longtom-4 development wells was based on an AvO inversion attribute that has, to date, proven reliable in highlighting thick, gas-filled sands. In-field and nearfield opportunities exist for adding upside to the Longtom field development, although attribute support has been hampered by coherent noise and resolution limitations. Reprocessing of the Northern Fields 3D within VIC/P54 was un dertaken with the specific objectives of reducing multiples and flattening gathers to improve upon and extend the existing AvO inversion dataset.

The reprocessing also provided an opportunity to trial the new Shallow Water Demultiple (SWD) technique in an ideal setting to gauge its ability relative to common demultiple methods. The application of SWD resulted in a stunning improvement to the stack with 2D SRME, tau-p deconvolution and high-resolution Radon demultiple delivering only modest further improvements. Three iterations of horizon-consistent velocity analysis ensured flat gathers, particularly near volcanic horizons whose strong interval velocity contrasts were not consistently honoured in the original processing. Anisotropic Kirchhoff prestack time migration was parameterised with a smoothly varying eta function and this resulted in some large fault plane movements, which could increase our reserves estimation for the Longtom field.

Revisiting the AvO inversion is expected to bring measurable improvements stemming from the use of multiple substacks as input, the extra well log data that is now available and by incorporating seismic interval velocities into the low frequency input model.

Te chnical Area: Seismic Data Processing - 3D technologies

PRESENTER PROFILE

Jarrod Dunne is a geophysicist at Nexus Energy in Melbourne. Prior jobs with Shell International and Woodside contributed to his special interest in seismic amplitude interpretation and its role in exploration and development. In 1996 he completed a Ph.D at Melbourne University focussing on seismic processing of deep seismic data from the Gippsland Basin. More recently his interests have broadened into petrophysics, rock physics, seismic interpretation and portfolio management. He is a member of the ASEG and SEG.

Email: [email protected]

Geothermal exploration plays an increasing role worldwide and in Australia in order to meet carbon emission targets and to provide green and renewable energy alternatives to coal. In Australia, EGS systems are widely used to extract geothermal energy from the subsurface. Initially, the stimulation process involves pumping large amount of fluids into the subsurface at depths of reasonable temperature to ensure a maximum energy output. The high-pressured fluids penetrate from the borehole into the surrounding sedimentary rocks and increase the permeability of the system until the stimulation is finished. Subsequently, a second borehole is put in place to extract the heated fluids.

Magnetotelluric (MT) measurements will be undertaken at the Paralana drill site in 2010, where the stimulation is conducted by Petratherm. The measurements take place in three stages in order to image the extent of the fluid body at depths of 3.5-4 km. To improve the outcome of the deployments, about 50 stations measure the MT response between periods of 100-0.01s prior to the stimulation to provide an initial model of the area. This is then used to compare and constrain the responses collected during and after the stimulation of the fluids into the reservoir. The aim of the survey is to delineate the extent of the water-filled reservoir.

During the stimulation, time-series of magnetic and electric field changes will also be recorded for comparison with events from the Micro Earthquake Array. These studies will look into the seismoelectric effect associated with the seismic events during the stimulation.

This paper describes the use of flexible electrode combinations for multi-borehole-surface resistivity surveys. Traditionally resistivity imaging adopts standard electrode arrays, such as Schlumberger, Wenner or dipole-dipole. These arrays were originally designed for use in surface surveys. For cross-hole or boreholesurface resistivity surveys, no analogous standard array design exists. Therefore we introduce a new method to complete a multi-borehole-surface resistivity survey, which has significant application in civil engineering and environmental exploration. For the method, electrodes may be positioned in both surface and boreholes positions with almost no limitation on the amount of electrodes able to be used in the boreholes. Real cases are presented showing that the multi-borehole-surface resistivity method is capable of solving some very difficult civil engineering exploration problems. This method is based on advanced resistivity inversion techniques, flexible electrode combinations, along with a fast multi-channel data acquisition system. It can easily be used for 3D resistivity multi-borehole-surface surveys.

The Ord Valley Airborne Electromagnetics (AEM) Interpretation Project was co-funded by the Australian Government and the Western Australian Government to provide information in relation to salinity and groundwater management in the Ord River Irrigation Area (ORIA). The project area covers the existing ORIA Stage 1, and the ORIA Stage 2 areas earmarked for irrigation extension. The project included the acquisition of 5,936 line km of AEM data acquired using the SKYTEM time domain system.

The SkyTEM AEM system successfully mapped key elements of the hydrogeological system over most of the project area. In general terms, the modelled conductivity structure defined from the SkyTEM smooth model Layered Constrained Inversion (LCI) matches that defined from available bore data exceptionally well, with an adjusted R² = 0.843 determined.

Overall, the AEM survey has provided enhanced spatial delineation of key elements of the hydrostratigraphy in 3D, including sand- and gravel-filled palaeochannels, and clay and silt distribution, as well as salt stores and groundwater quality. The study found significant areas of high salinity hazard in several of the Stage 2 areas earmarked for irrigation development, with salt stores and groundwater salinity often higher than in the Stage 1 areas.

This study has demonstrated the effective role that AEM methods can play as part of a ‘hydrogeological systems’ approach to the management of groundwater in existing and future irrigation developments in Northern Australia. The study has also demonstrated the potential for ‘calibrated’ AEM systems and Fast Approximate Inversion software to significantly shorten AEM project timelines.

Engineering and Community

• Geophysics role in increasing innovative engineering opportunities

• Better delineating groundwater resources

• Case histories in environmental geophysics

37C11LUI7

In 2007 the Australian Government funded an AEM survey (acquired with the RESOLVE frequency domain system) to provide information in relation to salinity management issues in the River Murray Corridor, including the Gunbower Island to Barr Creek sub-project area.

The study found that healthy vegetation along the Murray River is generally associated with river ‘flush zones’, where fresh groundwater is present to depths of up to 20m in zones up to 1.5km in width (eg in Gunbower Forest). Groundwater and salinity gradients suggest salt mobilisation from irrigated lands is adversely affecting vegetation health along the western boundary of the Forest. Overall, in the southern half of the project area the Murray River would appear to be a losing system, with a low short to medium term risk of off-site movement of salt into the River through the groundwater system.

In the northern half of the area there are discontinuous flush zones, the groundwater levels are closer to surface, and the river system may be a gaining system in places, with a moderately high salinity risk. Also, the Loddon River, Murrabit River and Barr Creek have no flush zones developed, and are potentially gaining systems at increased salinity risk.

Overall, the AEM and remote sensing data analysed in this study provide a new spatial context for assessing salinity hazard and risk, although there is a need for further ground validation and analysis including hydrodynamic modelling. The data might be used to guide plans for future salt interception in this area.

All geologists or geophysicists want to know the depth of investigation (DOI) for their final survey models. For diffusive methods, such as groundbased or airborne EM, there is no specific depth below which there is no information on the resistivity structure, but the question is to what depth the model is most reliable. We present a new robust concept for the calculation of DOI that is valid for any 1D EM geophysical model.

The method is based on the actual model output from the inversion and includes the full system response, contrary to assuming e.g. planar waves over a homogeneous halfspace. Equally important, the data noise and the number of data points is integrated in our calculation. Our methodology is based on a recalculated sensitivity (Jacobian) matrix of the final model and it can thus be used on any model type for which a sensitivity matrix can be calculated.

Contrary to other sensitivity matrix methods we define a global and absolute threshold value contrary to defining a relative, say 5%, sensitivity limit. The threshold limit will apply to all 1D inverted data and will thus produce comparable numbers of DOI.

It is widely accepted that, depending on acquisition technique, vertical seismic profiling (VSP) can provide us with extremely valuable information on seismic velocities (both P and S), attenuation and anisotropy of s eismic properties of the earth. Subsurface images, obtained with offset or 3D VSP could demonstrate superior resolution, signal-to-noise ratio and repeatability of surveys in comparison to standard surface seismic. These potential benefits are of great importance for both reservoir characterisation and time-lapse seismic monitoring. However in order to achieve them VSP surveys should be acquired with proper survey design and optimal source and receiver characteristics.

Modern 3C downhole VSP tools (such as VSI or similar array seismic equipment) are able to record seismic signal in wide frequency range (3-200 Hz) and provide high sensitivity and vector fidelity. Thus the main hardware factor affecting quality of VSP data is the seismic source. It can have different power, stability of the wavelet, frequency content and variable pattern of emitted wavefield including coherent noises.

In this paper we analyse land VSP data acquired with various seismic sources in order to determine signal quality and its repeatability. The data were acquired within the scientific program of the CO2CRC Otway Pilot Project in 2007-2009 and includes zero offset and offs et VSP in two neighbouring wells (CRC-1 and Naylor-1, distance between boreholes is ~300 m) and 3D VSP data in CRC-1. A wide range of seismic sources was used to acquire these vertical profiles; they include weight drop sources, MV vibroseis (6000 lbs), IVI Mini-Buggy vibroseis (16000lbs) and a limited amount of explosive shots.

We evaluate each source from a seismic data quality perspective (i.e. S/N ratio, image resolution, etc.) as well as acquisition performance and environmental impact.

Technical Area: Geophysics role in increasing innovative engineering opportunities

The CO2CRC Otway Project is Australia’s first demonstration of the deep geological storage of carbon dioxide. Its test site is located in Victoria, onshore. Within the bound of the first phase of the Otway project ~61000 tonnes of CO2/CH4 mixture were injected into depleted gas reservoir located at a depth of 2 km. Second phase of the project is dedicated to injection of small amount (up to 10 000 tonnes) of CO2-rich gas to saline aquifer at depth of ~1.3 km. Time-lapse seismic is a powerful tool for imaging of changes in the subsurface such as migration of CO2 within reservoir and assurance monitoring of possible leakage to other formations. However imaging of gas-into-gas injection (phase I) or injection of very small amounts (phase II) using land seismic are complicated problems. To meet these challenges a comprehensive seismic program is developed and being implemented. It includes surface and borehole time-lapse seismic surveys. First 3D survey was acquired in year 2000 to find gas fields in the area, two more surveys were shot in 2008 (pre-injection baseline) and 2009 (first monitor, ~31 000 tonnes of gas were injected to the date of survey), next survey is scheduled at January, 2010. A number or repeated 2D surveys were acquired over last several years to optimize 3D seismic acquisition technique and investigate repeatability of land seismic data.

In this presentation we discuss problems and preliminary results of time lapse seismic monitoring of CO2 sequestration in Otway.