Exploring the Impacts of Experimental Floods on Riparian Vegetation in the Lower Spöl River (Swiss Alps): Insights from Remote Sensing- and Uncrewed Aerial Vehicle- (UAV-) Based Monitoring

This work aims to investigate the short- and longterm effects of experimental floods (i.e., high flows released from a dam for river restoration purposes) on riparian vegetation in a regulated alpine river. The study uses RGB orthoimages from UAV (2018-2023) and historical aerial imagery (2006-2022) to characterise riparian vegetation composition and mortality in the Spöl River (Swiss Alps). Riparian tree stands were digitised based on a field survey, where the species composition was recorded along with the proportion of dead or dying trees. The orthoimages were used to compute an index based on the red and blue bands (KANA index), and non-parametric zonal statistics were applied to each riparian tree stand. The performance of this index varied substantially across datasets: it is likely that image acquisition conditions and sensor characteristics exerted a strong influence on index values and comparability. However, in recent UAV-derived images, results showed that KANA could effectively discriminate between 1. vegetation types, with broadleaved stands exhibiting higher values than coniferous stands, and 2. levels of mortality, with decreasing values corresponding to increased proportions of dead trees in the stand. Historical aerial imagery proved unsuitable for applying KANA to characterise vegetation. It appears, however, that in suitable conditions, KANA can be a useful alternative to vegetation indices such as NDVI if near-infrared data are unavailable. Furthermore, RGBbased analyses allow to examine a considerably larger corpus of data than multispectral analyses. We therefore recommend further exploring possible applications and limitations of this index for cost-effective assessment of riparian vegetation composition and health conditions.