The potential role of pedogenesis modelling in digital soil mapping

Mapping from field measured soil data generally relies upon empirical relationships (e.g., various multivariate regression models, geostatistics, Artificial Neural Networks (ANN)) and relationships with physical properties of the landform (e.g., slope length, gradient) and other soil-forming factors. Recent advances in the mathematical modelling of pedogenesis mean that it is now becoming feasible to use numerical and analytical models of soil process (weathering, erosion, etc.) to directly predict the grading of soils down through the soil profile and across the landscape. We are also able to simulate the history of pedogenesis and its interaction with climate so we are able to make an assessment of the equilibrium or non-equilibrium state of the soil relative to the environmental forcings on pedogenesis. By using the full grading distribution generated by the model we are able to directly use many pedotransfer functions to make high-resolution maps of soil functional properties and examine their spatial organisation. In this paper we will present an example of this approach using a mine spoil we have extensively studied. We have calibrated models for weathering and erosion of this spoil using laboratory and field studies. We will demonstrate the potential of the approach using these models.