Many of Australia’s river systems support extensive floodplains and wetlands but are threatened by water resource development and water extraction. To return beneficial environmental flows to these systems a better understanding of the temporal and spatial dynamics of the flood pulse is required.Remote sensing provides a source of data to assess past flood events critical to understanding the system processes; however in expansive floodplain and wetland systems with high landscape complexity, rapid vegetation response, canopy cover and shallow water depths, conventional analysis of open water may not be possible. Using readily available time-series of Normalised Difference Vegetation Index (NDVI), a decision tree was used to identify and map the likely flood extent using the vegetation response in the Gwydir wetlands, Australia. This ‘effective inundation’ over a range of flood events can be defined by flood frequency and connectivity to provide a simple conceptual model of the floodplains and wetlands. The results show that high NDVI response occurs following flooding in the floodplain and wetlands and was significantly correlated to 40 day inflow volumes. For Landsat and MODIS derived vegetation indices, an NDVI response greater than 0.75 was significant, while NDVI greater than 0.52 was significant for AVHRR data. In the Gwydir wetlands floods of up to 50 GL have occurred in 60% of years since the completion of Copeton Dam (1977), with almost 3000 ha of effective inundation. In contrast, moderately large floods of 100–200 GL occur in 30–50% of years and result in effective inundation of over 95 000 ha. The Gwydir wetlands could be simplified to 17 discrete patches, with each patch representing a connected landscape at defined flood frequency. This conceptual model can be used to develop hydrological models over the system at scales relevant to environmental flow management.
POWELL, S., Jakeman, A., & Croke, B. (2014). Can NDVI response indicate the effective flood extent in macrophyte dominated floodplain wetlands? Ecological Indicators, 45, 486-493. https://doi.org/10.1016/j.ecolind.2014.05.009