An extensive AEM survey recently commissioned by Geoscience Australia involved the use of two separate SkyTEM helicopter airborne electromagnetic (AEM) systems collecting data simultaneously. In order to ensure data consistency between the two systems, we follow the Danish example (conceived by the hydrogeophysics group from Aarhus University) of using a hover test site to calibrate the AEM data to a known reference. Since 2001, Denmark has employed a national test site for all electromagnetic (EM) instruments that are used there, including the SkyTEM system. The Lyngby test-site is recognised as a well-understood site with a well-described layered-earth structure of 5 layers. The accepted electrical structure model of the site acts as the reference model, and all instruments are brought to it in order to produce consistent results from all EM systems. Using a ground-based time-domain electromagnetic (TEM) system which has been calibrated at the Lyngby test site, we take EM measurements at a site selected here in Australia. With sufficient information of the instrument, we produce a layered-earth model that becomes the reference model for the two AEM systems used in the survey. We then bring the SkyTEM systems to the hover site and take soundings at multiple altitudes. From the hover test data and the ground based model, we calculate an optimal time shift and amplitude scale factor to ensure that both systems are able reproduce the accepted reference model. Conductivity sections produced with and without calibration factors show noticeably different profiles.

The Australian Solid Earth and Environment Grid (SEEGrid) is an eResearch infrastructure established to link diverse and distributed datasets in the geosciences, enable seamless interoperability between these, and undertake remote data processing. We present an integration between the GPlates plate-tectonic geographic information system and SEEGrid. Such a linkage is for the first time providing the necessary computational aids for abstracting an enormous level of complexity required for frontier solid-Earth research, in particular 4D metallogenesis. We present a continental reconstruction case study involving a proterozoic link between the greater Northern and Southern Australian cratons by combining evidence from several data sets. Faults are extracted from SEEGrid via Web Feature Services, and are used in conjunction with gravity anomaly data to test competing spatial alignment models of the reconstructed cratons. Additional information obtained from palaeomagnetic poles, granite geochemistry, geochronology, age-dated igneous provinces and other geophysics datasets can be used to further constrain the reconstruction. The metallogenic consequences of the best-fit reconstruction are profound, since they raises the possibility that the mineral systems hosting the giant Olympic dam, Broken Hill and Mt Isa could be linked in a particular geometry, resulting in a revised metallogenic map. The flexibility and extensibility of this spatio-temporal data analysis platform lends itself to a wide range of use-cases, including linking high-performance geodynamic modelling to kinematic reconstructions, creating the framework for future 3D and 4D metallogenic maps.

Since the 1989 Newcastle earthquake, the city of Newcastle, Australia, has become an extensive focus for earthquake hazard and risk assessment. The surficial geology varies between deeper alluvial deposits near the Hunter River, to shallower soils overlying weathered rock on the valley margins. Ambient vibration techniques, based on the dispersion property of surface waves in layered media, is one promising method for assessing the subsurface geophysical structure, in particular the shear-wave velocity (Vs). Using one such technique, the Spatial Auto-Correlation (SPAC) method, we characterise soil deposits at 23 sites in and around the city of Newcastle. Results show that values for soil overlying bedrock ranges from 200 m/s to 1000 m/s, with the higher velocity values observed in shallow soils which are relatively consolidated and far from river deposits. Bedrock depth varies from 6 to 56 m, but an accurate quantification is hampered by the low frequency picks (< 2 Hz) which are either unavailable or of dubious quality. Some shear-wave velocity profiles show two abrupt changes in Vs, the first ~ 4-15 m and the second ~19-56 m. Low Vs values are of particular interest as they may indicate areas of higher seismic hazard.

Government Geological Surveys study the Earth at the regional, province or national scale, and acquire vast volumes of technically complex data. These data must be high quality, fit for purpose, durable, and readily accessible and usable by industry. Increasingly, users require the geological information contained within the data as well as the data itself. High performance computers facilitate a step-change in advanced processing and modelling of large, complex data, and will help Government deliver more sophisticated products to industry and researchers. Data enhancement and manipulation are no-longer limited by the computational effort required, and there are no artificial limits to the size of the data or model, or the data resolution that can be processed. Geoscience Australia is collaborating with the National Computational Infrastructure facility (NCI) at the Australian National University to develop advanced methods for extracting the maximum geological information from large data volumes. The new methods include: Modelling of potential field data in spherical coordinates to create continental-scale reference models of density and magnetic susceptibility; Inversion of magnetotelluric tensor data to a full 3D mesh of resistivities, and; Monte Carlo inversion of AEM responses to assess the reliability and sensitivity of conductivity-depth images. These algorithms are being implemented in a new Virtual Geophysical Laboratory where government data and advanced processing methods are brought together in a single high performance computer environment.

Details initial algorithm development of a trans-dimensional Bayesian inversion methodology applied to remote sensing data - Shallow water bathymetry case study.

A detailed sequence stratigraphic study has been undertaken on the three wells in the Houtman Sub-basin: Gun Island 1, Houtman 1 and Charon 1. The study focussed on the Early-Late Jurassic Cattamarra Coal Measures, Cadda Formation and Yarragadee Formation succession. Wireline log character, cuttings, sidewall core and conventional core lithologies and palynological data were used to identify facies and paleoenvironments. Palynology for all wells has been reviewed, including new data collected by Geoscience Australia for Gun Island 1. Facies stacking patterns were used to define systems tracts and subsequently ten third-order depositional sequences. At the second-order (supersequence) level, the Cattamarra Coal Measures record a transgression culminating in maximum flooding in the Cadda Formation followed by highstand aggradation and regression in the Yarragadee Formation. The third-order sequences characterised in this study overprint this supersequence and control the local distribution of facies. The relative dominance of a facies may be either enhanced or diminished depending upon its position within the larger second-order supersequence. For example, a number of transgressive systems tracts within the dominantly non-marine Yarragadee Formation and Cattamarra Coal Measures record multiple, dinocyst-bearing, minor marine incursions into the Houtman Sub-basin. These marine incursions are not evident in the Yarragadee Formation in Charon 1, indicating a lack of accommodation space or proximal sediment input in the north during the mid-late Jurassic. The combined influence of these third-order and second-order sequences on facies distribution has significant implications for the distribution of potential reservoirs and seals in the Houtman Sub-basin and for regional palaeogeographic reconstructions of the Perth Basin.

We propose a surface cover change detection system based on the Australian Geoscience Data Cube (AGDC). The AGDC is a common analytical framework for large volumes of regularly gridded geoscientific data initially developed by Geoscience Australia (GA). AGDC effectively links geoscience data sets from various sources by spatial and temporal stamps associated with the data. Therefore, AGDC enables analysis of generations of consistent remote sensing time series data across Australia. The Australian Reflectance Grid 25m is one of the remote sensing data sets in the AGDC. The data is currently hosted at the high performance computational cloud at the National Computational Infrastructure. The proposed change detection system takes advantage of temporally rich data in the AGDC, applying time series analysis to identify changes in surface cover. The proposed system consists of various modules, which are independent of each other. The modules include: - a pixel quality mask and time series noise detection mask, which detects and filters out noise in data; - spectral classification modules based on random forests algorithm, which classifies pixels into specific objects using spectral information; - training modules which create classification modules using known surface cover data; - time series analysis modules, which models and transforms time series data into coefficients relevant to change detection targets; - temporal and spatial classification modules, which classify pixels into predefined land cover classes. A typical work flow for a change detection system includes sequential integration of the above mentioned modules. The system has been tested for applications in shallow water coastal zones and reforestation / deforestation detection, and displays a good potential for further development. This paper summarises development of the work flow and the initial results from example applications, such as reforestation / deforestation detection.

In the literature of remote sensing image analysis, an endmember is defined as a pixel containing only one land cover substance. However, with the varying resolutions of available sensors, in most cases a single pixel in a satellite image contains more than one type of land cover substance. One challenge is to decompose a pixel with mixed spectral readings into a set of endmembers, and estimate the corresponding abundance fractions. The linear spectral unmixing model assumes that spectral reading of a single pixel is a linear combination of spectral readings from a set of endmembers. Most linear spectral unmixing algorithms rely on spectral signatures from endmembers in pre-defined libraries obtained from previous on-ground studies. Therefore, the applications of these algorithms are restricted to images whose extent and acquisition time coincide with those of the endmember library. We propose a linear spectral unmixing algorithm which is able to identify a set of endmembers from the actual image of the studied area. Existing spectral libraries are used as training sets to infer a model which determines the class labels of the derived image based endmembers. The advantage of such approach is that it is capable of performing consistent spectral unmixing in areas with no established endmember libraries. Testing has been conducted on a Landsat7 ETM+ image subset of the Gwydir region acquired on Jun 22, 2008. Three types of land cover classes: bare soil, green vegetation and non-photosynthetic are specified for this test. A set consisting of 150 endmember samples and a number of ground abundance observations were obtained from a corresponding field trip. The study successfully identified an endmember set from the image for the specified land cover classes. For most test points, the spectral unmixing and estimation of the corresponding abundance are consistent with the ground validation data.

Mapping the variations between average air temperature and ground surface temperature

The National Dynamic Land Cover Dataset (DLCD) classifies Australian land cover into 34 categories, which conform to 2007 International Standards Organisation (ISO) Land Cover Standard (19144-2). The DLCD has been developed by Geoscience Australia and the Australian Bureau of Agricultural and Resource Economics and Sciences (ABARES), aiming to provide nationally consistent land cover information to federal and state governments and general public. This paper describes the modeling procedure to generate the DLCD, including machine learning methodologies and time series analysis techniques involved in the process.