Groundwater dependent ecosystem assessment using remote sensing. Exploring for the Future - Upper Darling River floodplain, New South Wales

This report presents the findings of a study conducted in the upper Darling River floodplain, aimed at improving optical and interferometric synthetic aperture radar (InSAR) remote sensing products for groundwater dependant vegetation (GDV) characterisation. The research was part of the Upper Darling Floodplain (UDF) groundwater study, funded by the Exploring for the Future program. This work tests the suitability of two novel remote sensing methods for characterising ecosystems with a range of likely groundwater dependence: combined wetness and greenness indices derived from Landsat products available through Geoscience Australia’s Digital Earth Australia platform, and an InSAR derived index of vegetation structure (known as SARGDE), which has been so far tested only in northern Australia. In addition, the relationship between the Normalised Difference Vegetation Index (NDVI), a remotely sensed proxy for vegetation condition, and water availability from surface water flows, rainfall, and groundwater was tested for sites with a range of low to high likely groundwater dependence.  The key findings of this work, and potential implications, are: • A multiple lines of evidence approach, drawing on persistence of wetness/greenness and vegetation structure, and correlation between inferred vegetation condition and groundwater levels, gives high confidence in the groundwater dependence of parts of the floodplain, particularly within the riparian zone. These indices require calibration with ground condition data to be applied in different regions, but a combined index could provide a high confidence measure of groundwater dependence. • Combined greenness and wetness, SARGDE, and the relationship between NDVI and groundwater levels all showed areas classified as ‘moderate’ likelihood of groundwater dependence having signatures comparable to areas classified as high likelihood. This could address a shortcoming of the groundwater dependence classification methodology, which, when groundwater level information is missing, classifies some vegetation types as moderate. • A combined index taking into account both greenness and wetness was able to better delineate vegetation types with a range of groundwater dependence previously not achievable using remote sensing products.  This work has provided improved methodology for applying remote sensing products to groundwater dependent vegetation characterisation in the study area. The methods are likely to be applicable to other regions with groundwater dependent vegetation. The results add to the evidence that it is necessary to better integrate surface and groundwater resources in water sharing plans at a basin scale. Further work is required to quantify the frequency and magnitude of flow events required to replenish alluvial groundwater sufficiently to maintain existing groundwater dependent ecosystems.