Every year floods cause millions of dollars damage to buildings and infrastructure, as well as to agricultural land and crops. They also disrupt business, and affect the safety and health of communities. The losses due to flooding vary widely from year to year and are dependent on a number of factors such as the severity of a flood and its location. Between 1967 and 2005 the average annual direct cost of floods in Australia has been estimated at AUD$377 million (BITRE 2008). This figure is likely to have risen following the widespread and devastating floods across eastern Australia that occurred over the summer of 2010-11.

Geoscience Australia, the Western Australian Department of Planning and the Western Australian Planning Commission have collaborated through this study to develop a regional-scale inundation model capable of simulating combined storm tide and riverine flood scenarios within current and future climate conditions (sea-level rise influences only). Modelling scenarios were applied to the Busselton region of Western Australia.

The Australian Flood Studies Database is available on line by Geoscience Australia. The database provides metadata on Australian flood studies and information on flood risk with a digital version where available. The purpose of the document is to guide new users in data entry and uploading of flood studies to a level acceptable for inclusion in the database.

Poster showing the 2010 Floods in Queensland fill Lake Eyre

The map shows the spatial distribution of short-duration rapid-onset floods and long-duration slow-rise floods. The Great Dividing Range in eastern Australia provides a natural separation of slower, wider rivers flowing west from faster, narrower coastal rivers flowing east.

The response to emergency situations such as floods and fires demand products in short time frames. If you use remote sensing then the response typically involves detailed examination of imagery in order to determine the spectral bands, ratios and associated thresholds that map the desired features such as flood or burn extent. The trial and error process associated with manual threshold selection is often time consuming and can result in significant errors due to confounding factors such as clouds and shadowed areas. By modelling features such as flood waters or fire scars as Gaussian distributions, allowing for fuzzy thresholds with neighbouring features, the required thresholds can be automatically derived from the imagery and emergency events can have extents determined much more rapidly. Automatic threshold selection minimises trial and error, thereby dramatically reducing processing turn-around time.

The Australian Flood Studies Database is available on line by Geoscience Australia via the Australian Flood Risk Information Portal. The database provides metadata on Australian flood studies and information on flood risk with a digital version where available. The purpose of the document is to guide new users in data entry and uploading of flood studies to a level acceptable for inclusion in the database.

The increasing availability of high-resolution digital elevation models (DEMs) is leading to improvements in flood analysis and predictions of surface-groundwater interaction in floodplain landscapes. To produce accurate predictions of flood inundation and calculations of flood volume, a 1m resolution LiDAR DEM was initially levelled to the Darling River floodplain by subtracting interpolated floodplain elevation trend surface from the DEM. This produces a de-trended flood-plain surface. Secondly, the levelled DEM surface was adjusted to the water-level reading at the Darling-River gauging station (Site 425012) at the time when the LiDAR was acquired. Flood extents were derived by elevation slicing of the adjusted levelled DEM up to any chosen river level. River-level readings from historical and current events utilised NSW Office of Water real-time river data. The flood-depth dataset is an inverted version of the flood-extent grid. Predicted flood depth and extent were classified by depth/elevation slice ranges of the adjusted de-trended DEM with 25 and 50 cm increments. In summary, the extent and depth of water inundation across the Darling floodplain have been predicted under different flooding scenarios, and validated using satellite data from historical (1990) and recent (2010/11) flood events. In all cases imagery and photo validation proved that predicted extents are accurate. The flood-risk predictions were then applied to a number of river-level scenarios. The flood risk predictions maps have been used as an input into developing recharge potential maps, and are being employed in flood-hazard assessments and infrastructure planning.

The Australian Flood Risk Information Portal (the portal) is an initiative of the Australian Government, established following the devastating floods across Eastern Australia in 2011. The portal is a key component of the National Flood Risk Information Project (NFRIP), and aims to provide a single point of access to Australian flood information. Currently much of Australia's existing flood information is dispersed across disparate sources, making it difficult to find and access. The portal will host data and tools that allow public discovery, visualisation and retrieval of flood studies, flood maps, satellite derived water observations and other related information, all from a single location. The portal will host standards and guidelines for use by jurisdictions and information custodians to encourage best practice in the development of new flood risk information. While the portal will initially host existing flood information, the architecture has been designed to allow the portal content to grow over time to meet the needs of users. The aim is for the portal to display data for a range of scenarios from small to extreme events, though this will be dependent on stakeholder contributions. Geoscience Australia's Australian Flood Studies Database is the portal's data store of flood study information. The database includes metadata created through a purpose-built data entry application, and over time, information harvested from state-operated catalogues. For each entry the portal provides a summary of the flood study, including information on how the study was done, what data was used, what flood maps were produced and for what scenarios, as well as details on the custodian and originating author. If the study included an assessment of damage, details such as estimates of annual average damage, or the number of properties affected during a flood of a particular likelihood will also be included. During the last phase of development downloadable flood study reports and their associated flood maps have been added to the portal where available. As the portal is populated it will increasingly host mapped flood data, or link to flood data and maps held in authoritative databases hosted by State and Territory bodies. Mapping data to be made accessible through the portal will include flood extents and to a lesser degree information on water depths. The portal will also include water observations obtained from Geoscience Australia's historic archive of Landsat imagery. This data will show whether a particular location was 'wet' at some point during the past 30 years. While this imagery does not necessarily represent the peak of a flood or show water depth, the data will support the validation and verification process of hydrologic and hydraulic flood modelling. This work will prove useful particularly in rural areas where there is little or no flood information. The portal also provides flood information custodians with the ability to either upload mapped data directly to the portal or to make this data accessible via web services. Data management tools and standards, developed through NFRIP, will enable data custodians to map their data to agreed standards for delivery through the portal. A portal framework and supporting principles has been developed to guide the maintenance and development of the portal.

In this study, various hydrochemical approaches were used to understand recharge processes in shallow (<120m) unconsolidated alluvial sediments in a 7,500 km2 area of the Darling River floodplain. Pore fluids were extracted from sediments from 60 sonic-cored bores, and together with surface and groundwater samples, provided a hydrochemical dataset with over 1600 samples and 25 analytes. Major ion chemistry highlights a mixing signature between river waters, the shallow unconfined aquifer and the underlying semi-confined Calivil Formation aquifer. These represent the fresh groundwater resources near the river and are Na-(Ca-Mg)-HCO3-Cl waters. Away from the influence of river leakage, the regional groundwater is more saline and sodic with an evolved Na-Cl-SO4 watertype. The mixing associated with river leakage is also supported by age dating. Stable isotope data show that recharge is episodic and linked to high-flow flood events rather than continuous river leakage, as demonstrated by hydrographic monitoring. The combination of surface water and groundwater sampling, the pore fluid analyses and fuzzy-k means (FCM) cluster analysis, provides a novel, relatively simple but powerful tool to assist with interpretation of groundwater processes. The FCM cluster analysis used analytes that were present in at least 60% of samples and resulted in samples being classified into eight classes (or hydrochemical facies). Pore fluids and groundwater with the greatest affinity to the surface water samples were easily identified. In this way, sites with significant active recharge, principally by river leakage, were mapped. Downhole plots of the pore fluid FCM classes provided additional insights into groundwater processes. Comparing the FCM classification of pore fluids within the target (semi)confined aquifer with those from the overlying clay aquitard and shallow aquifer allowed the assessment of vertical inter-aquifer leakage.