In many areas of the world, vegetation dynamics in semi-arid floodplain environments have been seriously impacted by increased river regulation and groundwater use. With increases in regulation along many rivers in the Murray-Darling Basin, flood volume, seasonality and frequency have changed which has in turn affected the condition and distribution of vegetation. Floodplain vegetation can be degraded from both too much and too little water due to regulation. Over-regulation and increased use of groundwater in these landscapes can exacerbate the effects related to natural climate variability. Prolonged flooding of woody plants has been found to induce a number of physiological disturbances such as early stomatal closure and inhibition of photosynthesis. However, drought conditions can also result in leaf biomass reduction and sapwood area decline. Depending on the species, different inundation and drought tolerances are observed. Identification of groundwater-dependent terrestrial vegetation, and assessment of the relative importance of different water sources to vegetation dynamics, typically requires detailed ecophysiological studies over a number of seasons or years as shown in Chowilla, New South Wales [] and Swan Coastal Plain, Western Australia []. However, even when groundwater dependence can be quantified, results are often difficult to upscale beyond the plot scale. Quicker, more regional approaches to mapping groundwater-dependent vegetation have consequently evolved with technological advancements in remote sensing techniques. Such an approach was used in this study. LiDAR canopy digital elevation model (CDEM) and foliage projected cover (FPC) data were combined with Landsat imagery in order to characterise the spatial and temporal behaviour of woody vegetation in the Lower Darling Floodplain, New South Wales. The multi-temporal dynamics of the woody vegetation were then compared to the estimated availability of different water sources in order to better understand water requirements.

This report provides background information about the Ginninderra controlled release Experiment 2 including a description of the environmental and weather conditions during the experiment, the groundwater levels and a brief description of all the monitoring techniques that were trialled during the experiment. Release of CO2 began 26 October 2012 at 2:25 PM and stopped 21 December 2012 at 1:30 PM. The total CO2 release rate during Experiment 2 was 218 kg/d CO2. The aim of the second Ginninderra controlled release was to artificially simulate the leakage of CO2 along a line source, to represent leakage along a fault. Multiple methods and techniques were then trialled in order to assess their abilities to: - detect that a leak was present - pinpoint the location of the leak - identify the strength of the leak - monitor how the CO2 behaves in the sub-surface - assess the effects it may have on plant health Several monitoring and assessment techniques were trialled for their effectiveness to quantify and qualify the CO2 that was release. This experiment had a focus on plant health indicators to assess the aims listed above, in order to evaluate the effectiveness of monitoring plant health and the use of geophysical methods to identify that a CO2 leak may be present. The methods are described in this report and include: - soil gas - airborne hyperspectral surveys - plant health (PhenoMobile) - soil CO2 flux - electromagnetic (EM-31) - electromagnetic (EM-38) - ground penetrating radar (GPR) This report is a reference guide to describe the Ginninderra Experiment 2 details. Only methods are described in this report with the results of the study published in conference papers and future journal articles.

The dry-tropics of central Queensland has an annual bushfire threat season that generally extends from September to November. Fire weather hazard is quantified using either the Forest Fire Danger Index (FFDI) or the Grassland Fire Danger Index (GFDI) (Luke and McArthur, 1978). Weather observations (temperature, relative humidity and wind speed) are combined with an estimate of the fuel state to predict likely fire behaviour if an ignition eventuates. A high resolution numerical weather model (dynamic downscaling) was utilised to provide spatial texture over the Rockhampton region for a range of historical days where bushfire hazard (as measured at the Rockhampton Airport meteorological station) was known to be severe to extreme. From the temperature, relative humidity and wind speeds generated by the model, the maximum FFDI for each simulated day was calculated using a maximum drought factor. Each of these FFDI maps was then normalised to the value of the FFDI at the grid point corresponding to Rockhampton Airport (ensemble produced). The annual recurrance interval (ARI) of FFDI at Rockhampton Airport for the current climate was calculated from observations by fitting Generalised Extreme Value (GEV) distributions. For future climate, we considered three downscaled General Circulation Models (GCM's) forced by the A2 emission scenario for atmospheric greenhouse gas emissions. The spatial pattern of the 50 and 100 year ARI fire danger rating for the Rockhampton region (current and future climate) was determined. In general, a small spatial increase in the fire danger rating is reflected in the ensemble model average for the 2090 climate. This is reflected throughout the Rockhampton region in both magnitude and extent through 2050 to 2090. Cluster areas of higher (future climate) bushfire hazard were mapped for planning applications. Handbook MODSIM2013 Conference

Geoscience Australia and the CO2CRC operate a greenhouse gas controlled release facility at an experimental agricultural station maintained by CSIRO Plant Industry in Canberra, Australia. The facility is designed to simulate surface emissions of CO2 (and other greenhouse gases) from the soil into the atmosphere. Over 10 different near surface monitoring techniques were trialled at the Ginninderra controlled release site during 2012-2013. Different climatic conditions for the early 2012 release experiment (wet) and late 2013 release experiment (dry) resulted in markedly different sub-surface plume behaviour and surface expression of CO2. Gaseous CO2 was released 2 m below the ground surface from a slotted, 100 m long horizontal well at a rate of 144 kg/d for at least 8 weeks for both experiments. The most obvious difference between the two release experiments was that CO2 leakage expressed at different locations along the well for the two experiments. As also observed in other controlled release experiments internationally, the surface expression of CO2 during these experiments, as measured using a portable soil flux meter, was restricted to localised spots. For the 2012 (wet) release experiment, the leakage was limited to a small intense primary leak (approximately 12 m in diameter) and a neighbouring small secondary leak. In contrast, the leak from the 2013 (dry) release experiment was broader, spread over a longer length of the release well, and did not attain the very high flux intensities observed in the previous year. An array of 1 m deep soil gas wells provided insight into the migration pathways of CO2 in the sub-surface, showing a much broader dispersion of CO2 in the sub-surface compared to the surface CO2 expression. Krypton tracers confirmed that the spread of the introduced gases in the sub-surface was much greater than the surface expression, with different behaviour observed between the 2012 and 2013 experiments. The differences between the years are attributed to changes in groundwater levels, drier conditions, and a larger vadose zone during the 2013 experiment. Eddy covariance (EC) towers were deployed at the site for both experiments with the objective to detect and quantify CO2 emissions. CO2 leaks were detected above the background and the direction of the leak confirmed. However, analysis showed that current methods of EC are not appropriate for quantifying the CO2 leak, as much of the CO2 flux is lost through advection and diffusion below the measurement height. This is because the footprint of the leak is much smaller than the EC tower's footprint, resulting in a highly heterogeneous system that breaches EC's key assumptions. The results suggest that quantification using EC may not be possible for CO2 leaks with small footprints. An array of atmospheric CO2 sensors was also deployed at the site during the experiments. Application of atmospheric tomographic techniques using the point source sensors appears to be a more effective approach than EC for quantifying CO2 emissions. Broad scale leak detection technologies are necessary for surveying areas beyond high risk sites and is the subject of ongoing research at Ginninderra. Airborne hyperspectral and thermal scanning measurements were taken over CO2-impacted, mature wheat and field pea crops. The CO2 impact on plants was characterised through biochemical analysis and observed changes in plant morphology. High resolution ground-based hyperspectral and thermal measurements were taken over tillering barley and wheat, as well as field pea and canola seedlings. Dry conditions and crop stage strongly influenced the effectiveness of the remote sensing techniques for CO2 leak detection. A comparison between the high resolution ground-based and airborne hyperspectral measurements for detecting CO2 impacted plants will be presented as well as an overall assessment of the leak detection techniques. Submitted to the GHGT-12

We have developed a Building Fire Impact Model to evaluate the probability that a building located in a peri-urban region of a community is affected/destroyed by a forest fire. The methodology is based on a well-known mathematical technique called Event Tree (ET) modeling, which is a useful graphical way of representing the dependency of events. The tree nodes are the event itself, and the branches are formed with the probability of the event happening. If the event can be represented by a discrete random variable, the number of possible realisations of the event and their corresponding probability of occurring, conditional on the realisations of the previous event, is given by the branches. As the probability of each event is displayed conditional on the occurrence of events that precede it in the tree, the joint probability of the simultaneous occurrence of events that constitute a path is found by multiplication (Hasofer et al., 2007). BFIM contains a basic implementation of the main elements of bushfire characteristics, house vulnerability and human intervention. In the first pass of the BFIM model, the characteristics of the bushfire in the neighboring region to the house is considered as well as the characteristics of the house and the occupants of the house. In the second pass, the number of embers impacting on the house is adjusted for human intervention and wind damage. In the third pass, the model examines house by house conditions to determine what houses have been burnt and their impact on neighboring houses. To illustrate the model application, a community involved in the 2009 Victorian bushfires has been studied and the event post-disaster impact assessment is utilized to validate the model outcomes. MODSIM 2013 Conference

Here we report on the results of a study undertaken in the Flinders Commonwealth Marine Reserve (southeast Australia) designed to test the benefits of two approaches to characterising shelf habitats: (i) MBES mapping of a continuous (~30 km2) area selected on the basis of its potential to include a range of representative seabed habitats , versus; (ii) a novel approach that uses targeted mapping of a greater number of smaller, but spatially balanced, locations using a Generalized Random Tessellation Stratified sample design. We present the first quantitative estimates of habitat type on the shelf of the Flinders reserve, using both survey approaches, based on three MBES analysis techniques. We contrast the quality of information that both survey approaches offer in combination with the three MBES analysis methods. We then consider the implications for future inventory of benthic habitats in shelf environments in the context of monitoring extensive offshore marine reserves.

The Collaborative Australian Protected Areas Database (CAPAD) 2012 provides both spatial and text information about government, Indigenous and privately protected areas for continental and marine Australia. State and Territory conservation agencies supplied data, current to 31 December 2012, to Australian Government Department of the Environment.

A video for the launch of new Great Barrier Reef bathymetry data on 30 November 2017.

In this study, we aim to identify the most accurate methods for spatial prediction of seabed gravel content in the northwest Australian Exclusive Economic Zone. We experimentally examined: 1) whether input secondary variables affect the performance of RFOK and RFIDW, 2) whether the performances of RF, SIMs and their hybrid methods are data-specific, and 3) whether model averaging improves predictive accuracy of these methods in the study region. For RF and the hybrid methods, up to 21 variables were used as predictors. The predictive accuracy was assessed in terms of relative mean absolute error and relative root mean squared error based on the average of 100 iterations of 10-fold cross validation. In this study, the following important findings were achieved: - the predictive errors fluctuate with the input secondary variables; - the existence of correlated variables can alter the results of model selection, leading to different models; - the set of initial input variables affects the model selected; - the most accurate model can be missed out during the model selection; - RF, RFOK and RFIDW prove to be the most accurate methods in this study, with RFOK preferred; and these methods are not data-specific, but their models are, so best model needs to be identified; and - Model averaging is clearly data-specific. In conclusion, model selection is essential for RF and the hybrid methods. RF and the hybrid methods are not data-specific, but their models are. RFOK is the most accurate method. Model averaging is also data-specific. Hence best model needs to be identified for individual studies and application of model averaging should also be examined accordingly. RF and the hybrid methods have displayed substantial potentials for predicting environmental properties and are recommended for further test for spatial predictions in environmental sciences and other relevant disciplines in the future. This study provides suggestions and guidelines for improving the spatial predictions of biophysical variables in both marine and terrestrial environments.

A major concern for regulators and the public with geological storage of CO2 is the potential for the migration of CO2 via a leaky fault or well into potable groundwater supplies. Given sufficient CO2, an immediate effect on groundwater would be a decrease in pH which could lead to accelerated weathering, an increase in alkalinity, release of major and minor ions and heavy metals (particularly Pd, Ni and Cr) as well as CO2 mobilisation of trace organic contaminants. These scenarios potentially occur in a high CO2 leakage event, therefore detection of a small leak, although barely perceptible, could provide an important early warning for a subsequent and more substantial impact. Different approaches are required for the detection and quantification of these low level leaks and are the subject of this paper. A 3 year groundwater survey was recently completed in the Surat Basin, which provided comprehensive water and isotopic analysis of groundwaters together with their exsolved gases. The gases were analysed for composition, -13CCO2, -13CCH4 and -2HCH4. Methane is prevalent in the major Surat Basin aquifers (e.g. Mooga, Gubberamunda and Hutton sandstones) and is invariably associated with a bacterial (methanogenic) carbonate reduction source, evident from its isotopic signature ('13CCH4 ~ -70', '2HCH4 ~ -220'). In addition to methane and low levels of CO2, trace ethane is common. Two neighbouring wells, however, were quite different to the other 85 wells surveyed. Their exsolved gases contained comparatively high ethane, but also C1-C6 hydrocarbons in addition to methane. Methane isotope systematics were significantly different from other groundwater wells completed in the same formation. The -13C of the CO2 was similar to the surrounding groundwater wells, but the relative proportion of CO2 in the gas was significantly higher. Combined, these characteristics are consistent with hydrocarbon biodegradation. There was little difference in the groundwater chemistry for these wells compared to the regional baseline. The study provides a useful analogue study for detection, at various scales, of a leaky well associated with a geological storage site. Compositional and isotopic analysis of exsolved gases from groundwater samples could be used to demonstrate non-equilibrium conditions and intrusion of exogenic CO2. Abstract for the 2013 International Association of Hydrologist Congress, Perth