Soil acidity is a natural process that can be exacerbated in farming systems. Current knowledge and data on the extent and severity of acidic soils in south-western Victoria is limited. This makes inferences on the impacts to production across the region difficult. Furthermore, improved mapping is required in order to define the opportunities to address soil acidity in southern Victoria and increase production potential. The availability of soil site data managed in the Victorian Soil Information System (VSIS) and spatially exhaustive ancillary datasets (i.e. environmental covariate map data such as elevation, rainfall and gamma radiometrics) support the application of predictive modelling techniques to produce soil pH maps at finer scales and qualities previously unattainable.
The digital soil maps of soil pH for the South West region of Victoria have been produced by modelling the spatial relationships between points (soil sites) of measured or estimated soil pH and their environment (defined by a comprehensive set of covariates). A 10-fold cross validation procedure was used to produce average predictions for the upper, lower and mean values. The mapping provides predictions of soil pH at 50 m pixel resolution for six set depths from the surface down to two metres. The six set depths have been chosen to align to the Global Soil Map specifications, www.globalsoilmap.net.
In total, data from 3,668 sites were identified for application in spatial models across south-western Victoria. This data has been sourced from land studies dating back to the 1950s and the 670 samples collected by this project are now accessible as part of this larger dataset. Spatial covariate datasets using in modelling includes climate (e.g. annual rainfall, evaporation, Prescott index), landscape (e.g. clay mineral maps), organisms (e.g. MODIS time series, LANDSAT scenes), relief (e.g. elevation, slope, topographic wetness index) and parent material (e.g. terrain weathering index). In total, 71 covariate raster datasets have been used in generating soil pH maps.
The maps are for soil pH measured in a 1:5 soil-to-water suspension (pHw) with possible addition of a salt solution (typically Calcium chloride, CaCl2). The raster datasets (maps) include a mean, lower and upper uncertainty prediction for each depth interval.
