Modelling groundwater dependent ecosystems in the Willunga Basin, South Australia

The challenge of groundwater management is to establish sustainable extraction regimes that provide acceptable levels of protection of economic, social and environmental assets and values that depend on the resource. A key knowledge gap relates to how groundwater resource management affects the integrity and survival of groundwater dependent ecosystems (GDEs). This study involves the development of habitat suitability models that assess the suitability of the groundwater regime in the Willunga Basin for supporting GDEs. GDE species in the Basin are classified into five functional groups according to their water requirements and tolerances. Habitat suitability index curves were developed for the five groups based on species observations and descriptions reported in literature. The index curves are described by a set of constraints that quantify the minimum and maximum bounds of values, reflecting the uncertainty of the relationship between groundwater regime and species habitat suitability. Alternative mathematical translations of descriptions of the groundwater requirements of species were also tested. The models were applied to assess the habitat suitability of the five functional groups at 35 sites in the Willunga Basin in terms of the groundwater regime (results are summarised in the figure below). The models were able to identify sites that have unsuitable or poorly suited habitat for most of the functional groups with high confidence. However, there was low confidence in identifying sites with good habitat. The results of test runs of the alternative constraints suggested that the models were relatively robust. The sites were ranked by partial order, however if the plausible index bounds of sites overlapped, uncertainty remained about the ranking implied. Large bounds indicate gaps in knowledge that require further research. The model was applied to detect areas where GDE habitat may have been lost through time. Potential applications of the model include identifying sites for further monitoring or research. Future work should include the incorporation of more attributes of the groundwater regime, water quality and other environmental factors to describe habitat suitability, the use of ecological data for calibrating or validating the model, and the involvement of experts to set and review the model constraints. This modelling approach allows sites to be evaluated from an ecological point of view even with high uncertainty.