3.2.5 Information management and processing

A very large number of multi-dimensional and multi-scaled datasets are used in the impact and risk analysis for each BA, including model outputs, and ecological, economic and sociocultural data from a wide range of sources. To manage these datasets and produce meaningful results, a consistent spatial framework is needed that permits rapid spatial and temporal analyses of impacts without compromising the resolution of the results. The datasets for this BA are organised into an impact analysis database (Bioregional Assessment Programme, Dataset 2) to enable efficient management. The purpose of this database is to produce result datasets that integrate the available modelling and other evidence across the assessment extent. These datasets are required to support three types of BA analyses: analysis of hydrological changes, impact profiles for landscape classes, and impact profiles for assets. The results of these analyses are summarised in this product, with more detailed information available at www.bioregionalassessments.gov.au. The impact analysis database is also available at data.gov.au.

The datasets used in the impact and risk analysis database (Bioregional Assessment Programme, Dataset 3) include the assets, landscape classes, modelling results (groundwater, surface water and receptor impact modelling), coal resource development ‘footprints’ and other relevant geographic datasets, such as the boundaries of the subregion, assessment extent and zone of potential hydrological change. All data in the impact and risk analysis database (and the results derived from it) meet the requirements for transparency.

The impact and risk analysis requires the geoprocessing of complex queries on very large spatial datasets. To overcome the computational load associated with this task a relational, rather than geospatial, approach was utilised. All dataset geometries are split against a universal grid of assessment units that exhaustively cover the assessment extent (Dataset 2). An assessment unit is a geographic area represented by a square (1 km²) polygon with a unique identifier. Assessment units were used to partition asset and landscape class spatial data for impact analysis. The gridded data can be combined and recombined into any aggregation supported by the conceptual modelling, causal pathways and model data.

Normalising the database included calculating impact area, length and counts for individual features (e.g. stream reaches, individual assets, groups of assets or landscape classes) at the assessment unit level. Selecting the assessment units of interest and summing the pre-calculated values of area, length or count for the required dataset provides individual analysis result. This approach of front-loading the geospatial analysis through grid base attribution is fundamental to enabling the volume of calculations required to complete the assessment. The approach uses the source geometries in calculation and hence does not impact on the analysis calculations. In a few cases where source geometries created geospatial errors, resulted in the exclusion of these units. Removing invalid geometries did not, in any case, affect the analysis results more than a combined total area of one assessment unit (i.e. 1 km²) per analysis calculation.

The surface water modelling generates results at points that are extrapolated to links (see Section 3.3.4), so there is a need to map streams to assessment units. For assessment units with only a single stream reach, the assessment unit stores the information associated with this stream segment. However, where the assessment unit contains multiple stream reaches (e.g. at the confluence of two streams), it is necessary to prioritise which stream reach informs the value of the assessment unit for representing the surface water modelling results. The general rules for prioritising a stream reach take into account:

Some streams have insufficient hydrological information for assignment of hydrological variables. In this case, the BA identifies these streams and reports information based on area overlays only, that is, no receptor impact model exists in these areas.