We re-mapped the soils of the Murray–Darling Basin (MDB) in 1995–1998 with a minimum of new fieldwork, making the most out of existing data. We collated existing digital soil maps and used inductive spatial modelling to predict soil types from those maps combined with environmental predictor variables. Lithology, Landsat Multi Spectral Scanner (Landsat MSS), the 9-s digital elevation model (DEM) of Australia and derived terrain attributes, all gridded to 250-m pixels, were the predictor variables. Because the basin-wide datasets were very large data mining software was used for modelling. Rule induction by data mining was also used to define the spatial domain of extrapolation for the extension of soil–landscape models from existing soil maps. Procedures to estimate the uncertainty associated with the predictions and quality of information for the new soil- landforms map of the MDB are described.


The pattern of soil in a landscape reflects the evolution of that landscape over long periods. It is the expression of the spatial integration of earth surface processes active in pedogenesis; this is especially true in Australia, a continent that has been relatively stable over the last 55 Ma at least (since it separated from Antarctica; Ollier, 2001). The interactions between lithologic parent material, topography, availability of water, biosphere, factors of soil formation after Jenny (1941), as reflected by surficial weathering and transport processes over time, determine the pattern of soil distribution(Paton et al., 1995).

Methods of soil survey are generally empirical and revolve around the mental development of correlation between soils and underlying rocks, landforms, vegetation and air-photo patterns. The conceptualization of soil–landscape models to predict the spatial distribution of soils has been termed the ‘soil–landscape paradigm’ (Hudson, 1992). Operationally, soil–landscape models for a region can be derived from a few, high intensity, detailed maps and extrapolated over the broader region (Thompson and Beck- man, 1959; Favrot, 1989). The mental model(s) derived from the small, high intensity, detailed studies are usually not stated (Hewitt, 1993). The operational usefulness of explicitly stated predictive models in soil survey is starting to be recognized (McKenzie, 1991; Lagacherie, 1992; McSweeney et al., 1994; Chen, 1997).