The Santa Maria dataset contains ten sources of spatially exhaustive covariates. These are two area-class soil maps, two geologic maps, two land use maps, two digital elevation models, and imagery of two orbital satellite sensors.
The first area-class soil map was published at a scale of 1: and depicts five mapping units \citep{AzolinEtAl1988} (Figure \ref{fig:soil-map-azolin}). The data used in its production came from existing technical reports and area-class soil maps prepared at smaller scales, aerial photographs (scale of 1:), topographic maps (scale of 1:), and sparse point soil data collected along the road network by convenience sampling.