Small intermittently open lagoons are the most common type of estuary in NSW but the majority of them have no seagrass despite apparently suitable physical and water quality conditions. This paper describes the consequences of a catastrophic drought on trophic structure of primary producers in two intermittently open coastal lagoons that have negligible anthropogenic stress. It examines the roles of various factors influencing primary producers and hence ecosystems in intermittently open lagoons, in both short and long timeframes. Two estuarine lagoons with undisturbed catchments (Durras Lake and Nadgee Lake) experienced decadal scale shift in production dominance from benthic to pelagic and back to benthic following a catastrophic reduction in water levels during drought conditions. Water quality deteriorated significantly (high turbidity, ammonia and chlorophyll) after loss of macrophytes and then re-established to "normal" conditions within 2-3 years. Despite this, macrophytes did not re-appear for 6-7 years. The high proportion of intermittent estuaries with no seagrass was concluded to be a consequence of recruitment limitation, and estuarine morphology shapes the ecology of these lagoons, rather than an inherent inability for seagrass and macrophytes to live in the estuaries. Recruitment of macrophytes required coincident occurrence of multiple rare stochastic events, making chance an important aspect of re-establishment of biological assemblages in disturbance driven ecosystems. Conceptual models derived for northern hemisphere, riverine, surface-water fed estuaries tend to dominate thinking about estuaries in the literature, but are clearly inappropriate for typical southern-hemisphere low-flow and groundwater-dominated intermittent estuaries. Climate change is likely to increase the frequency of occurrence of the conditions that led to the loss of seagrass and macrophytes and therefore potentially reduce the number of intermittent estuaries in NSW with macrophytes. In the systems described, maintenance of regional seagrass populations is reliant on the stochastic survival of some populations to act as source populations. Climate change is predicted to make what are currently "extreme" events more common. This could, in turn, mean that the likelihood of the sort of loss described in our two cases studies could become more common, increasing the proportion of systems with no seagrass and progressively closing the window of opportunity for external recruitment. Potentially, this could lead to localised extinction of seagrass in intermittent estuaries.
