Key aspects of the design and of the Australian Hydrological Geospatial Fabric (the Geofabric) represent a departure from the methods traditionally used to build and manage framework data sets. The most important differences are the recognition that no single physical data model can support all users of the data and the reliance on formally specifying data products prior to their generation rather than post generation.

An essential component of the Geofabric will be a conceptual information model. This implementation independent model will map all features of the waterscape and the relationships between these features. Importantly, the model will be modular in nature such that governance of each component can be undertaken by appropriate bodies.

Another key component of the Geofabric will be the reliance on formal data product specifications to describe the data sets used to build the Geofabric as well as products derived from it. These specifications will replace the current ad-hoc methods used to write data products specifications.

They will be consistent and hence comparable ensuring the ability to identify commonality between data sets. Importantly they will form part of the contract between data supplier and user. This presentation describes further the concepts underpinning the design of the Geofabric and discusses why these concepts have been necessary to support the Bureau's requirements.