Exploring the 4D scales of eco-geomorphic interactions along a river corridor using repeat UAV Laser Scanning (UAV-LS), multispectral imagery, and a functional traits framework.

The importance of vegetation within the river corridor is well known and has been subject to a considerable body of research. The interactions between riparian vegetation and river morphology are typically complex, co-dependent, highly dynamic, and vary across both spatial and temporal scales. Vegetation diversity can typically be attributed to fluvial influences such as flood regimes and morphology, whilst simultaneously influencing the flow of water and sediment transport. However, adequately capturing the spatial and temporal complexity of vegetation characteristics has been a considerable challenge, and so a number of unresolved questions with regard to the spatial and temporal interactions of vegetation and river flow remain.

Within this research, we seek to establish the relationship between vegetation presence and geomorphic response over 2 years of data collection on a 1 km stretch of the River Teme in the United Kingdom. Functional vegetation traits of different plant forms relevant to hydrological research are extracted using UAV based laser scanning and multispectral imagery. These traits are then upscaled to reach scale functional group classifications, whereby they can be compared to geomorphic change occurring throughout the reach. Our framework moves beyond typical species level classification, as vegetation is instead grouped on the potential geomorphic impact that it may have due to their characteristics. Such methods are beginning to be established in fluvial research, but are often constrained by the need for extensive ground surveying or they focus on how traits vary in response to fluvial controls.

Our results show six distinct functional groups are obtained from a mix of laser scanning and imagery data, before being upscaled across the study area with a classification accuracy of 80% using random forest methods. Plant structure was subsequently used to assess spatially varying and seasonal changes in excess vegetative drag based on reference flow depths across the study site during a peak flow event. These variations could be used to assess the aggregated geomorphic response of the system based on flow conditions and vegetation type, and begin to unpick different feedbacks between them. Such methods could be used on similar river systems, to improve wider reach classifications using both airborne laser scanning and imagery, as well as in different geomorphic research where there is interactions between flows and vegetation.