Abstract Although the potential impacts of rising water tables and secondary salinization on agricultural land in southern Australia have been recognized for some time, it is only recently that the impacts on native vegetation have been considered. Despite the likely extent and severity of the problem, no comprehensive approach to assessing the impact of salinity upon native vegetation has been attempted to date. In the present paper, we discuss the causes and impacts of rising water tables and dryland salinity, assess the levels of risk in different ecosystem types and consider the possibilities for the maintenance of biodiversity and ecosystem function in vegetation at risk. We examine the salinity risk to woodland vegetation in the Western Australian (WA) wheatbelt, and consider both broad-scale context and finer-scale variation within individual patches of vegetation. From this information, we develop a set of conceptual models of the potential impacts of shallow saline water tables on ecosystem structure and processes in remnant vegetation in agricultural areas, particularly in the WA wheatbelt. First, we suggest that fine-scale variability in surface topography and soil characteristics may play an important role in limiting the impacts of rising saline water tables. The outcome will depend on the interaction of the heterogeneity of the impact, species distribution in relation to small-scale environmental heterogeneity and variation in species response to hydrological change. Second, we suggest that shallow saline water tables can be considered to cause an ‘edge effect’, which moves inwards from the edge of remnants of native vegetation. Finally, we consider how saline surface flows exacerbate the effects of shallow saline water tables and hasten vegetation decline in remnant areas. We put these models forward as hypotheses to be tested in different situations. We contrast the situation of secondary salinization in Australian vegetation with that of naturally saline systems in Australia and elsewhere, and suggest that these systems may provide important signposts toward developing management approaches for vegetation at risk. In conclusion, we consider the need to set priorities for the protection and restoration of natural vegetation at risk from altered hydrology, based on an assessment of relative threat and probability of persistence or recovery. We highlight the urgency for action that protects native vegetation from the increasing risks of rising water tables.