Abstract # : 1479734 Paper # : GP43B-1142 Session : GP43B Potential-field and EM methods for geologic problems of the mid and upper crust Developments for 3D gravity and magnetic modeling in spherical coordinates Richard Lane - Geoscience Australia - rjllane@gmail.com Qing Liang - China University of Geosciences (Wuhan) - qingliang.cug@gmail.com Chao Chen - China University of Geosciences (Wuhan) - chenchao@cug.edu.cn Yaoguo Li - Colorado School of Mines - ygli@mines.edu At Geoscience Australia (GA), Australia's Commonwealth Government geoscientific agency, we perform gravity and magnetic modeling at a range of scales, from broad regional crustal studies with thousands of kilometer lateral extent and tens of kilometer vertical extent, to detailed local studies with kilometer or less lateral extent and meters to hundreds of meters vertical extent. To achieve greater integration and coherence, and to better understand the geological significance of this work, we are investing in a number of development projects; * Spherical coordinate gravity and magnetic modeling, * Modeling using High Performance Computing facilities, * Utilizing rock property data as an input into the modeling and interpretation of gravity and magnetic data, * Better management of geoscience data and models, and * Visualization of spatial data in a Virtual Globe format. In collaboration with the Colorado School of Mines (CSM) and the China University of Geosciences (CUG), we are developing a capability to model gravity and magnetic data in a spherical coordinate framework. This will provide more accurate calculations and permit us to integrate the results into a single framework that more realistically reflects the shape of the Earth. Modeling gravity and magnetic data in a spherical coordinate framework is far more compute intensive than is the case when performing the corresponding calculations in a Cartesian (rectangular) coordinate framework. To reduce the time required to perform the calculations in a spherical coordinate framework, we will be deploying the modeling software on the National Computational Infrastructure (NCI) High Performance Computing (HPC) facility at the Australian National University (ANU). This will also streamline the management of these software relative to the other main option of establishing and maintaining HPC facilities in-house. We are a participant in the Deep Exploration Technologies Cooperative Research Centre (DET CRC). In combination with this involvement, we are expanding our support for systematic management of rock property data, and developing a better understanding of how these data can be used to provide constraints for the modeling work. We are also using the opportunities afforded through the DET CRC to make progress with documentation and standardization of data storage and transfer formats so that the tasks of management, discovery and delivery of this information to users are simplified and made more efficient. To provide the foundations of integration and analysis of information in a spatial context, we are utilizing and customizing 3D visualization software using a Virtual Globe application, NASA World Wind. This will permit us to view the full range of information types at global to local scales in a realistic coordinate framework. Together, these various development activities will play an important role in the on-going effort by Geoscience Australia to add value to the potential field, rock property, and geological information that we possess. We will then be better able to understand the geology of the Australian region and use this knowledge in a range of applications, including mineral and energy exploration, natural hazard mitigation, and groundwater management.

Geoscience Australia is supporting the exploration and development of offshore oil and gas resources and establishment of Australia's national representative system of marine protected areas through provision of spatial information about the physical and biological character of the seabed. Central to this approach is prediction of Australia's seabed biodiversity from spatially continuous data of physical seabed properties. However, information for these properties is usually collected at sparsely-distributed discrete locations, particularly in the deep ocean. Thus, methods for generating spatially continuous information from point samples become essential tools. Such methods are, however, often data- or even variable- specific and it is difficult to select an appropriate method for any given dataset. Improving the accuracy of these physical data for biodiversity prediction, by searching for the most robust spatial interpolation methods to predict physical seabed properties, is essential to better inform resource management practises. In this regard, we conducted a simulation experiment to compare the performance of statistical and mathematical methods for spatial interpolation using samples of seabed mud content across the Australian margin. Five factors that affect the accuracy of spatial interpolation were considered: 1) region; 2) statistical method; 3) sample density; 4) searching neighbourhood; and 5) sample stratification by geomorphic provinces. Bathymetry, distance-to-coast and slope were used as secondary variables. In this study, we only report the results of the comparison of 14 methods (37 sub-methods) using samples of seabed mud content with five levels of sample density across the southwest Australian margin. The results of the simulation experiment can be applied to spatial data modelling of various physical parameters in different disciplines and have application to a variety of resource management applications for Australia's marine region.

Geoscience Australia is supporting the exploration and development of offshore oil and gas resources and establishment of Australia's national representative system of marine protected areas through provision of spatial information about the physical and biological character of the seabed. Central to this approach is prediction of Australia's seabed biodiversity from spatially continuous data of physical seabed properties. However, information for these properties is usually collected at sparsely-distributed discrete locations, particularly in the deep ocean. Thus, methods for generating spatially continuous information from point samples become essential tools. Such methods are, however, often data- or even variable- specific and it is difficult to select an appropriate method for any given dataset. Improving the accuracy of these physical data for biodiversity prediction, by searching for the most robust spatial interpolation methods to predict physical seabed properties, is essential to better inform resource management practises. In this regard, we conducted a simulation experiment to compare the performance of statistical and mathematical methods for spatial interpolation using samples of seabed mud content across the Australian margin. Five factors that affect the accuracy of spatial interpolation were considered: 1) region; 2) statistical method; 3) sample density; 4) searching neighbourhood; and 5) sample stratification by geomorphic provinces. Bathymetry, distance-to-coast and slope were used as secondary variables. In this study, we only report the results of the comparison of 14 methods (37 sub-methods) using samples of seabed mud content with five levels of sample density across the southwest Australian margin. The results of the simulation experiment can be applied to spatial data modelling of various physical parameters in different disciplines and have application to a variety of resource management applications for Australia's marine region.

Integration of disparate sets of geophysical data, such as Bouguer gravity, magnetotelluric and seismic travel time data for a robust interpretation of architectural settings of the subsurface is carried out. At the outset, the layered 2D model space is appropriately gridded. A spline node layer boundary parameterisation with a sigmoid basis function is used to relate local 1D layered model parameterisation to the 2D model space. Joint 1D inversion of seismic travel time and magnetotelluric data is carried out at the spline nodes using empirical relationships between seismic velocity and resistivity. The two objective functions corresponding to each of the input data types are combined through a weighting factor, the appropriate value of which is determined using the L-curve technique. The particle swarm optimization scheme is used as a robust optimiser for the layer depths and property values. The inverted velocity model is transformed to a density model using a second empirical rock property relationship. A 2D inversion of Bouguer gravity data is then carried out producing adjustments to the depths for the layer boundaries. This completes the initialisation phase of the procedure. A second iterative phase during which only the depths to the layer boundaries are modified is carried out to build a coherent model which is consistent with all three kinds of data. This involves re-inverting jointly the seismic travel time and magnetotelluric data, where parameters corresponding to the rock properties are kept unaltered, but the depth to the layer interfaces are updated. The method is trialled with a synthetic situation and is implemented to interpret the architectural settings of newly discovered Millungera basin of North Queensland, Australia.

Despite growing concerns about potential enhancement of global warming and slope failure by methane produced by gas hydrate dissociation, much uncertainty surrounds estimates of gas hydrate reservoir sizes, as well as methane fluxes and oxidation rates at the sea floor. For cold seep sediments of the eastern Mediterranean Sea, depth-dependent methane concentrations and rates of anaerobic oxidation of methane (AOM) are constrained by modeling the measured pore-water sulfate profile. The calculated dissolved methane distribution and flux are sensitive to the advective flow velocity, which is estimated from the depth distributions of conservative pore-water constituents (Na, B). Near-complete anaerobic oxidation of the upward methane flux is supported by the depth distributions of indicative biomarkers, and the carbon isotopic compositions of organic matter and dissolved inorganic carbon. Pore-water and solid-phase data are consistent with a narrow depth interval of AOM, 14-18 cm below the sediment-water interface. Based on an isotopic mass balance, the biomass of the microbial population carrying out oxidation of methane coupled to sulfate reduction at the given methane flux represents about 20% of the total organic carbon, which is a significant pool of in situ formed organic matter. Model results indicate that the asymptotic methane concentration is reached a few meters below the sediment surface. The predicted asymptotic concentration is close to the in situ saturation value with respect to gas hydrate, suggesting that the rate of shallow gas hydrate formation is controlled by the ascending methane flux. The proposed model approach can be used to predict the formation of gas hydrate, and to quantify methane fluxes plus transformation rates in surface sediments where fluid advection is an important transport mechanism.

To perform a realistic 3D inversion of gravity data covering a significant proportion of the surface of the Earth, it is necessary to take into account the curvature of the Earth. We have developed an algorithm for inverting gravity data in spherical coordinates and have demonstrated this using data covering the continental mass of Australia and surrounding ocean areas. The density structures evident in the crust and uppermost mantle of the resultant 3D inversion model are in broad agreement with knowledge of the geological features for the region and with variations in the depth to the Moho that are present in the AusMoho model.

The Collaborative Research Centre for Greenhouse Gas Technologies (CO2CRC) Program 3.2 Risk Assessment is working toward a risk assessment procedure that integrates risk across the complete CCS system and can be used to meet the needs of a range of stakeholders. Any particular CCS project will hold the interest of multiple stakeholders who will have varied interests in the type of information and in the level of detail they require. It is unlikely that any single risk assessment tool will be able to provide the full range of outputs required to meet the needs of regulators, the general public and project managers; however, in many cases the data and structure behind the outputs will be the same. In using a suite of tools, a well designed procedure will optimize the interaction between the scientists, engineers and other experts contributing to the assessment and will allow for the required information to be presented in a manner appropriate for each stakeholder. Discussions of risk in CCS, even amongst the risk assessment community, often become confused because of the differing emphases on what the risks of interest are. A key question that must be addressed is: 'What questions is the risk analysis trying to answer?' Ultimately, this comes down to the stakeholders, whose interests can be broken into four target questions: - Which part of the capture-transport-storage CCS system? - Which timeline? (project planning, project lifespan, post closure, 1,000 years, etc) - Which risk aspect? (technical, regulatory, economic, public acceptance, or heath safety and environment) - Which risk metric? (Dollars, CO2 lost, dollars/tonne CO2 avoided, etc.) Once the responses to these questions are understood a procedure and suite of tools can be selected that adequately addresses the questions. The key components of the CO2CRC procedure we describe here are: etc

A key component of marine bioregional planning is to map the spatial patterns of marine biodiversity, often measured as species richness, total abundance or abundance/presence of key taxa. In this study, predictive modelling approaches were used to map soft bottom benthic biodiversity on the Carnarvon Shelf, Western Australia, using a range of physical surrogates. This surrogacy approach could also explicitly link physical environmental attributes to the marine biodiversity patterns. The statistical results show that between 20% and 37% of variances on the two biodiversity measures (Species Richness and Total Abundance) were explained by the Random Forest Decision Tree models. The best statistical validation performance was found at the Point Cloates area. This was followed by the Gnaraloo area, then by the Mandu Creek area. The models identified different individual physical surrogates for the three study areas and the two biodiversity measures. However, it was found that the infaunal biodiversity at the three study areas of the Carnarvon Shelf were driven by similar ecological process. Sediment properties were the most important physical surrogates for the infaunal biodiversity. Coarser and heterogeneous sediments favour higher infaunal species richness and total abundance. The prediction maps indicate the highest infaunal biodiversity at deeper water of the Point Cloates area. In contrast, the majority of the Mandu creek area has low infaunal biodiversity. This may be due to the much narrower shelf width (e.g., ~6 km) in this part of Carnarvon Shelf than the Point Cloates and Gnaraloo areas. The narrow shelf would limit the space for oceanographic processes to work on the sediment and develop heterogeneous sediment properties that support diverse and productive infaunal species.

Seismic reflection and gravity potential field data acquired during the Geoscience Australia SW Margins Survey 2008-09 was used to investigate the distribution of volcanic facies and large-scale structural architecture of the Mentelle Basin, located on the southwestern margin of Australia. Based on structural differences the basin is subdivided into the Western (WMB) and Eastern (EMB) Mentelle Basins. Isopach and seismic facies maps were used to identify the thickness and distribution of volcanic facies. These maps show that volcanism is generally confined to the Western Mentelle Basin, with two distinct areas of thick volcanic deposits occurring in the central and northern portion. Two and three dimensional gravity forward models were used to investigate the structural architecture of the Mentelle Basin. Two dimensional gravity modelling shows that the crust is extremely thin in the WMB (c.10km), associated with two mantle highs. The crust thickens from the EMB (>20km) towards mainland Australia. The two modelled mantle highs coincide with the two seismically defined areas of thick volcanic deposits. Analogue models indicate that rift related volcanism is generally confined to the locus of extension where crustal thinning and strain are greatest. Thus results of gravity modelling and seismic interpretation have been interpreted to indicate that Jurassic - Cretaceous extension was focussed in the WMB. This thinning of the crust and presence of mantle highs suggests that the WMB is a failed rift formed during the initial breakup between Australia and Greater India and abandoned at the onset of spreading in the Perth Basin.

A regional scale 3D geological map of the upper crustal sequence in the West Arnhem Land region, Northern Territory, was compiled from surface mapping, limited drilling information, and liberal amounts of geological inference. Modelling of the gravity and magnetic field response of this map was proposed as a means of evaluating the viability of this geological hypothesis. A relatively good supply of mass density and magnetic property data were available to constrain the transformation of 3D geological maps into property models in preparation for potential field modelling. The presence of numerous relatively thin magnetic horizons, dykes, and sills provided many challenges for producing geologically-realistic magnetic property models at a regional scale. Modelling of the gravity field at this scale was far more straight forward and successful. A stochastic procedure was used to derive a large number of geological maps by making small changes to the highly uncertain interpretive parts of the original 3D geological map. A subset of these derived geological maps had associated mass density models that could adequately reproduce the gravity field observations. The common characteristics of the geological models in this subset were isolated using statistical techniques and used to refine our portrayal of the regional scale 3D geological features.