AbstractThe goal of the project discussed in this paper is to produce small‐scale maps of dryland salinization potential for Victoria, Australia; maps which putatively identify areas where salinization effects might be expected to worsen or where effects have yet to become evident. This paper assumes: (a) that current stream salinity levels can be modelled empirically using a relationship reported by Greig and Devonshire (1981), and (b) that a subset of these levels are indicative at an early stage of themaximumlong‐term severity (meaning extensiveness) of proximate dryland salinization should water‐tables rise. This paper uses computer induction to relate modelled stream salinity to observed salinization and hence to dryland salinization pote