Mapping and classifying groundwater dependent ecosystems by incorporating remote sensing data and mapped vegetation communities within a hydro-geomorphological framework

Background/Question/Methods

An outcome of the 1994 Council of Australian Government’s Water reform framework, was that water allocation planning is required to protect ecosystems that have an important function or conservation value, of which Groundwater Dependant Ecosystems (GDEs) are included. While there is a world wide acceptance of the presence and importance of GDEs, within the Australian context the focus has largely been on iconic GDEs, such as the mound springs of the Great Artesian Basin and the Banksia woodlands of Western Australia. Such a limitedawareness of GDEs thus presents substantial challenges for GDE management policy development. While in some instances site specific studies detail the function of GDEs and their water requirements, such methods are not economical or practical for discerning GDEs at catchment and regional scales.

Therefore, to develop an understanding of Victorian GDEs in terms of their functionality and their relative groundwater dependency, it is first necessary to identify the location and extent of GDEs throughout the state of Victoria.
Major biophysical attributes that all GDE’s have in common were classified into GIS layers. They are;
1. Access to groundwater, either through root uptake or groundwater discharges into aquifers, lakes and rivers. This layer was produced by comparing groundwater, geological and stream gauge data with water balance modelling to create a regional groundwater interaction with the surface map.
2. Vegetation use of groundwater layer. Remote sensing data sets were interrogated to provide the commonly used normalised difference vegetation index (NDVI) which identifies the variation in photosynthetic activity, to highlight areas of constant activity.
3. Hydro-geomorphological setting. Current knowledge on the location of flora and fauna assemblages and remote sensing outputs were incorporated into the groundwater interactive map detailing the landscape setting for each potential GDE type.
These data layers were then incorporated into a Bayesian model. A series of rules were developed for the model, which rated the likelihood of a GDE existing in each landscape on a scale from high to low.

Results/Conclusions

The integration and analysis of remote sensing, hydrogeological datasets and ecological knowledge provided a powerful tool that enhances the understanding of each of these separate disciplines. This approach provides an efficient means of mapping vegetation communities reliant on groundwater at broader scales and provides a foundation on which to build detailed understanding of the management needs of at risk GDEs.