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.

Benthic habitats on the continental shelf are strongly influenced by exposure to the effects of surface ocean waves, and tidal, wind and density driven ocean currents. These processes combine to induce a combined flow bed shear stress upon the seabed which can mobilise sediments or directly influence organisms disturbing the benthic environment. Output from a suite of numerical models predicting these oceanic processes have been utilised to compute the combined flow bed shear stresses over the entire Australian continental shelf for an 8-year period (March 1997- February 2005 inclusive). To quantify the relative influence of extreme or catastrophic combined flow bed shear stress events and more frequent events of smaller magnitude, three methods of classifying the oceanographic levels of exposure are presented: 1. A spectral regionalisation method, 2. A method based on the shape of the probability distribution function, and 3. A method which assesses the balance between the amount of work a stress does on the seabed, and the frequency with which it occurs. Significant relationships occur between the three regionalisation maps indicating seabed exposure to oceanographic processes and physical sediment properties (mean grain size and bulk carbonate content), and water depth, particularly when distinction is made between regions dominated by high-frequency (diurnal or semi-diurnal) events and low-frequency (synoptic or annual) events. It is concluded that both magnitude and frequency of combined-flow bed shear stresses must be considered when characterising the benthic environment. The regionalisation outputs of the Australian continental shelf presented in this study are expected to be of benefit to quantifying exposure of seabed habitats on the continental shelf to oceanographic processes in future habitat classification schemes for marine planning and policy procedures.

A nationally-consistent wave resource assessment is presented for Australian shelf (<300 m) waters. Wave energy and power were derived from significant wave height and period, and wave direction hindcast using the AusWAM model for the period 1 March 1997 to 29 February 2008 inclusive. The spatial distribution of wave energy and power is available on a 0.1° grid covering 110'156° longitude and 7'46° latitude. Total instantaneous wave energy on the entire Australian shelf is on average 3.47 PJ. Wave power is greatest on the 3,000 km-long southern Australian shelf (Tasmania/Victoria, southern Western Australia and South Australia), where it widely attains a time-average value of 25-35 kW m-1 (90th percentile of 60-78 kW m-1), delivering 800-1100 GJ m-1 of energy in an average year. New South Wales and southern Queensland shelves, with moderate levels of wave power (time-average: 10-20 kW m-1; 90th percentile: 20-30 kW m-1), are also potential sites for electricity generation due to them having a similar reliability in resource delivery to the southern margin. Time-average wave power for most of the northern Australian shelf is <10 kW m-1. Seasonal variations in wave power are consistent with regional weather patterns, which are characterised by winter SE trade winds/summer monsoon in the north and winter temperate storms/summer sea breezes in the south. The nationally-consistent wave resource assessment for Australian shelf waters can be used to inform policy development and site-selection decisions by industry.

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.

Basin evolution of the Vlaming Sub-basin and the deep-water Mentelle Basin, both located offshore on the southwest Australian continental margin, were investigated using 2D and 3D petroleum system modelling. Compositional kinetics, determined on the main source sequences, were used to predict timing of hydrocarbon generation and migration as well as GOR evolution and phase behaviour in our 2D and 3D basin models. The main phase of petroleum generation in the Vlaming Sub-basin occurred at 150 Ma and ceased during following inversion and erosion episodes. Only areas which observed later burial have generated additional hydrocarbons during the Tertiary and up to present day. The modelling results indicate the likely generation and trapping of light oils for the Jurassic intervals for a variety of structural traps. It is these areas which are of greatest interest from an exploration point of view. The 2D numerical simulations in the Mentelle Basin indicate the presence of active hydrocarbon generating kitchen areas. Burial histories and generalized petroleum evolutionary histories are investigated.

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.

Geoscience Australia provides spatial information of seabed environment to support Australian marine zone management. Central to this approach is the prediction of Australia's seabed biodiversity from spatially continuous data of seabed biophysical properties. Seabed hardness is an important environmental property for predicting marine biodiversity and is often inferred from multibeam backscatter data. Although seabed hardness can be measured based on video images, they are only available at a limited number of sampled locations. In this study, we attempt to predict the spatial distribution of seabed hardness using random forest based on video classification and available marine environmental properties. We illustrate the effects of cross-validation methods including a new cross-validation function on the selection of optimal predictive models. We also test the effects of various predictor sets on the predictive accuracy. This study provides an example for predicting the spatial distribution of environmental properties using random forest in R.

The subsurface of the Earth is a complex system, one that we are yet to fully understand and model. It is hence impossible to automate the process of mapping and modelling, and the input of user experience and knowledge ('prior knowledge') is required to produce meaningful and useful outputs. This form of solution does not lend itself to a simple programmatic approach. However, by taking advantage of advances in computer technology and the application of numerical methods for modeling complex environments, we can do much to improve upon past results. Introduction As Australia's national geoscience organisation, Geoscience Australia (GA) plays an important role in the creation and delivery of fundamental geoscientific information. Studies are carried out at a wide range of scales, from a continental perspective to highly detailed local site investigations. In most situations, direct geological observations are supplemented by the inferences that can be made from geophysical measurements. Observations of the Earth's gravity and magnetic fields contain signals from subsurface materials, and extensive holdings of these measurements are commonly used to help create 3D subsurface models. With sparse hard constraints and incomplete, insufficient, noisy observations, knowledge workers or experts continue to play an important role in providing implicit prior constraints on any system to model this volume. The interface to these people becomes an important part of any set of tools for performing geological modeling of gravity and magnetic data. Users constantly demand a better experience and better outcomes when modeling the subsurface. Some of their recurring requests are for: * A simpler, more intuitive user-experience * Higher resolution * Models with larger extents * Faster processing * Inclusion of a greater number of geological and rock property constraints * Estimates of the uncertainty in the outcomes * Improved 3D visualisation * Tracking of input provenance and subsequent processing that is carried out * Organised management of 3D models Integration of the elements is a key consideration when developing solutions, as users are loathe to adopt procedures that become more involved and more difficult to understand and to piece together. Today, developments to produce world-class solutions typically take place across multiple agencies, involving many people, and at locations spread around the globe. This in itself is a challenge! We have focused our efforts on the following: * The management and delivery of rock properties * Spherical and Cartesian coordinate gravity and magnetic modelling software * Use of High Performance Computing (HPC) facilities * Use of a virtual globe application for 3D visualisation

In 2008, the performance of 14 statistical and mathematical methods for spatial interpolation was compared using samples of seabed mud content across the Australian Exclusive Economic Zone (AEEZ), which indicated that machine learning methods are generally among the most accurate methods. In this study, we further test the performance of machine learning methods in combination with ordinary kriging (OK) and inverse distance squared (IDS). We aim to identify the most accurate methods for spatial interpolation of seabed mud content in three regions (i.e., N, NE and SW) in AEEZ using samples extracted from Geoscience Australia's Marine Samples Database (MARS). The performance of 18 methods (machine learning methods and their combinations with OK or IDS) is compared using a simulation experiment. The prediction accuracy changes with the methods, inclusion and exclusion of slope, search window size, model averaging and the study region. The combination of RF and OK (RFOK) and the combination of RF and IDS (RFIDS) are, on average, more accurate than the other methods based on the prediction accuracy and visual examination of prediction maps in all three regions when slope is included and when their searching widow size is 12 and 7, respectively. Averaging the predictions of these two most accurate methods could be an alternative for spatial interpolation. The methods identified in this study reduce the prediction error by up to 19% and their predictions depict the transitional zones between geomorphic features in comparison with the control. This study confirmed the effectiveness of combining machine learning methods with OK or IDS and produced an alternative source of methods for spatial interpolation. Procedures employed in this study for selecting the most accurate prediction methods provide guidance for future studies.

Geothermal energy has been harnessed in Australia for several decades for both direct use applications and power generation, but only at very small scale installations. Australia's geothermal resources are amagmatic and unconventional by the accepted definitions in other parts of the world centred on active volcanism or plate margin collision. Worldwide, there is a lack of experience in exploring for and developing unconventional resources, and few "deposit" or resource models to aide exploration. The conceptualisation of a range of geological environments amenable to geothermal resource development will underpin the large scale development of geothermal utilisation in Australia. This will include developing exploration models spanning the range of unconventional geothermal resources; from "EGS" or "Hot (Dry) Rock" where permeability stimulation is a pre-requisite, to "Hot Sedimentary Aquifer" where no permeability stimulation is required.