Spatially continuous predictions of seabed hardness are important baseline environmental information for sustainable management of Australia's marine jurisdiction. Seabed hardness is often inferred from multibeam backscatter data with unknown accuracy, can be inferred based on underwater video footage at limited locations. It can also be predicted to two classes. In this study, we classified the seabed into four classes based on two new seabed hardness classification schemes (i.e. hard90 and hard70) for seabed video footage by. We developed optimal predictive models to predict the spatial distribution of seabed hardness using random forest (RF) based on point data of hardness classes and spatially continuous multibeam backscatter data. Five feature selection (FS) methods that are variable importance (VI), averaged variable importance (AVI), the combined, Boruta, and RRF were tested. Effects of highly correlated, important and unimportant predictors on the accuracy of RF predictive models were also examined. Finally, the most accurate models were used to predict the spatial distribution of the hardness classes and the predictions were visually examined and compared with the predictions based on two-class hardness classification. This study confirms that: 1) hard90 and hard70 are effective seabed hardness classification schemes; 2) seabed hardness can be predicted into a spatially continuous layer with a high degree of accuracy; 3) the typical approach used to pre-select predictors by excluding highly correlated predictors needs to be re-examined when using machine learning methods, at least, for RF, in the environmental sciences; 4) the identification of the important and unimportant predictors provides useful guidelines for further improving the predictive models; 5) FS is essential for identifying an optimal RF predictive model and the RF methods select the most accurate predictive model(s) instead of the most parsimonious ones, and AVI and Boruta are recommended for future studies; and 6) RF is an effective modelling method with high predictive accuracy for multi-level categorical data, can be applied to `small p and large n problems in environmental sciences, and is recommended for future studies. In addition, automated computational programs for AVI need be developed to improve its computational efficiency and caution should be taken when applying filter FS method in selecting predictive models in future studies.

Acoustic backscatter from the seafloor is a complex function of signal frequency, seabed roughness, grain size distribution, benthos, bioturbation, volume reverberation and other factors. Angular response is the variation in acoustic backscatter with incident angle and it is considered be an intrinsic property of the seabed. The objective of the study was to illustrate how the combination of a self-organising map (SOM) and hierarchical clustering can be used to develop an angular response facies map for Point Cloates, northwest Australia; demonstrate the cluster visualisation properties of the technique; and highlight how the technique can be used to investigate environmental variables that influence angular response.

This study aims to quantify sponge biodiversity of the Van Diemen Rise and eastern Joseph Bonaparte Gulf in northern Australia and to examine spatial and environmental patterns associated with differences in community structure of sponges.

Geoscience Australia carried out marine surveys in Jervis Bay (NSW) in 2007, 2008 and 2009 (GA303, GA305, GA309, GA312) to map seabed bathymetry and characterise benthic environments through colocated sampling of surface sediments (for textural and biogeochemical analysis) and infauna, observation of benthic habitats using underwater towed video and stills photography, and measurement of ocean tides and wavegenerated currents. Data and samples were acquired using the Defence Science and Technology Organisation (DSTO) Research Vessel Kimbla. Bathymetric mapping, sampling and tide/wave measurement were concentrated in a 3x5 km survey grid (named Darling Road Grid, DRG) within the southern part of the Jervis Bay, incorporating the bay entrance. Additional sampling and stills photography plus bathymetric mapping along transits was undertaken at representative habitat types outside the DRG. jb_s3 is an ArcINFO grid of southern part of Jervis Bay survey area (south3 is part of Darling RD grid) produced from the processed EM3002 bathymetry data using the CARIS HIPS and SIPS software

Discerning how marine ecosystems are linked through larval dispersal is essential for understanding demographic flow, investigating the development of population genetic structure, and for evaluating the potential responses of communities to climate change. This information is of critical importance when designing reserve networks, identifying key locations for restoration, controlling invasive species, and administering transboundary resources. As part of Geoscience Australia's commitment to the National Environmental Research Programme's Marine Biodiversity Hub, we have developed a fully four-dimensional (3D space x time) individual-based model that embeds artificially intelligent particles within real-world ocean flow fields, making it possible to examine expected dispersal patterns of marine larvae under a variety of conditions. The model fuses strategic biological behaviour with physical equations in a flexible manner through the use of object-oriented programming. We will discuss aspects of model development and testing, as well as practical issues relating to computing on the National Computing Infrastructure, and addressing large-scale data storage. We will also identify potential avenues for data analysis that can be used to inform environmental decision-making.

A defining characteristic of the seabed is the proportion that is hard, or immobile. For marine ecosystems, hard seabed provides the solid substrate needed to support sessile benthic communities, often forming 'hotspots' of biodiversity such as coral and sponge gardens. For the offshore resource and energy industry, knowledge of the distribution of hard versus soft seabed is important for planning infrastructure (pipelines, wells) and to managing risk posed by geo-hazards such as migrating sand waves or mass movements on steep banks. Maps that delineate areas of hard and soft seabed are therefore a key product to the informed management and use of Australia's vast marine estate. As part of the Australian Government's Offshore Energy Security Program (2007-2011) and continuing under the National CO2 Infrastructure Plan (2011-2015), Geoscience Australia has been developing integrated seabed mapping methods to better map and predict seabed hardness using acoustic data (multibeam sonar), integrated with information from biological and physical samples. The first method used was a two-stage, classification-based clustering method. This method uses acoustic backscatter angular response curves to derive a substrate type map. The angular response curve is the backscatter value as a function of the incidence angle, where this angle lies between the incident acoustic signal from the normal. The second method was a prediction-based classification, using a machine learning method called random forest. This method was based on bathymetry, backscatter data and their derivatives, as well as underwater video and sediment data. The techniques developed by Geoscience Australia offer a fast and inexpensive assessment of the seabed that can be used where intensive seabed sampling is not feasible. Moreover, these techniques can be applied to areas where only multibeam acoustic data are available. Importantly, the identification of seabed substrate types in spatially continuous maps provides valuable baseline information for effective marine conservation management and infrastructure development.

A biophysical dispersal model was used to simulate hydrodynamic connectivity among canyons located within Australia's South-west marine region. The results show that exchange among canyons in this area is greatly influenced by the Leeuwin current, transporting larvae in a unidirectional manner around Cape Leeuwin, and continuing eastwards along the Great Australian Bight. Larvae within canyons tend to remain within them, however if they are transported above the canyon walls, they then have the opportunity to be transported significant distances (thousands of kilometres). Analysis of the variability in connectivity patterns reveals concentrated flow near the shelf break, with increasing levels of variability leading offshore from the canyons. While the average potential flow distance and duration between canyons were approximately 550 kilometres and 33 days respectively, the average realized flow distance and duration were approximately 30 kilometres and 6 days respectively. This study provides the first consideration of connectivity among submarine canyons and will help improve management of these features by providing a better understanding of larval movement, transboundary exchange and the potential spread of invasive species.

As part of Geoscience Australia's commitment towards the National Environmental Programme's Marine Biodiversity Hub, we have developed a fully four-dimensional (3D x time) Lagrangian biophysical dispersal model to simulate the movement of marine larvae over large, topographically complex areas. The model operates by fusing the results of data-assimilative oceanographic models (e.g. BLUELink, HYCOM, ROMS) with individual-based particle behaviour. The model uses parallel processing on Australia's national supercomputer to handle large numbers of simulated larvae (on the order of several billion), and saves positional information as points within a relational database management system (RDBMS). The model was used to study Australia's northwest marine region, with specific attention given to connectivity patterns among Australia's north-western Commonwealth Marine Reserves and Key Ecological Features (KEFs). These KEFs include carbonate terraces, banks and reefs on the shelf that support diverse benthic assemblages of sponges and corals, and canyons that extend from the shelf edge to the continental slope and are potential biodiversity hotspots. We will show animations of larval movement near canyons within the Gascoyne CMR; larval dispersal probability clouds partitioned by depth and time; as well as matrices of connectivity values among features of interest. We demonstrate how the data can be used to identify connectivity corridors in marine environments, and how the matrices can be analysed to identify key connections within the network. Information from the model can be used to inform priorities for monitoring the performance of reserves through examining net contributions of different reserves (i.e. are they sources or sinks), and studying changes in connectivity structure through adding and removing reserve areas.

Geoscience Australia's GEOMACS model was utilised to produce hindcast hourly time series of continental shelf (~20 - 300 m depth) bed shear stress (unit of measure: Pascal, Pa) on a 0.1 degree grid covering the period March 1997 to February 2008 (inclusive). The hindcast data represents the combined contribution to the bed shear stress by waves, tides, wind and densitydriven circulation. Ecological data collected from Torres Strait suggests that bed shear stresses exceeding 0.4 Pa are important in determining the species present (Long, Bode, & Pitcher 1997). Although this data may not be representative of other regions or benthic communities, it has been utilised to calculate two parameters for determining the relationship between shear bed stress and the benthic community. One of the parameters, which is denoted by , and is calculated using; represents the stress in excess of 0.4 Pa integrated over time as a proportion of the total stress integrated over time, and is intended to represent the proportion of the total integrated stress that has some control on the benthic community (Hughes & Harris 2008).

Geoscience Australia carried out marine surveys in Jervis Bay (NSW) in 2007, 2008 and 2009 (GA303, GA305, GA309, GA312) to map seabed bathymetry and characterise benthic environments through colocated sampling of surface sediments (for textural and biogeochemical analysis) and infauna, observation of benthic habitats using underwater towed video and stills photography, and measurement of ocean tides and wavegenerated currents. Data and samples were acquired using the Defence Science and Technology Organisation (DSTO) Research Vessel Kimbla. Bathymetric mapping, sampling and tide/wave measurement were concentrated in a 3x5 km survey grid (named Darling Road Grid, DRG) within the southern part of the Jervis Bay, incorporating the bay entrance. Additional sampling and stills photography plus bathymetric mapping along transits was undertaken at representative habitat types outside the DRG. jb_s1 is an ArcINFO grid of southern part of Jervis Bay survey area (south1 is part of Darling RD grid) produced from the processed EM3002 bathymetry data using the CARIS HIPS and SIPS software