Woody-perennial vegetation forms a critical component of floodplains in dryland regions, which are commonly managed with environmental flows to mitigate the impacts of flow regulation. Drone-based RGB and LiDAR imagery analysed with machine learning provides an accurate and repeatable approach to monitor and evaluate vegetation responses in environments with limited access such as floodplains and wetlands. This presentation will introduce two spatial frameworks: 1) integrating drone RGB imagery, Convolutional Neural Networks (CNNs) and GIS, and 2) drone RGB and LiDAR point clouds integrating the Visible Light Difference Vegetation Index (VDVI), machine learning (k-Nearest Neighbours) and ecological features (canopy height model, CHM) to classify and estimate floodplain vegetation and evaluate its response to wetting and drying – with a focus on the keystone floodplain shrub Duma florulenta (tangled lignum).
Drone RGB and LiDAR imagery, collected from 18 sites in 2023 and 2024 in the Mallee Region, North Western Victoria, Australia were used as the inputs for the frameworks. Four sites in the Macquarie Marshes, New South Wales, were also used to validate and demonstrate the method's generalization. Both methods had an overall accuracy of at least 90% and outperformed commonly used frameworks. The condition state of D. florulenta was assessed using high-quality (green and vigorous plant material) and low-quality (brown and dormant plant material) condition categories, recognising the life-history of the species, with percent cover shown to be related to recent flooding history. The outputs and data derived from the two methods and advantages of each will be explored in the presentation. These methods provide an effective, accurate and repeatable way to detect a response in woody-floodplain vegetation to flooding and drying and environmental flows and are now integrated into large, long-term monitoring programs in Australia. These approaches to monitor and evaluate the condition of floodplain vegetation can be transferred to other systems and species.