How satellites can enrich river geomorphic characterization

Satellite images-derived data can change the way geomorphology describes patterns and processes of extremely dynamic systems such as rivers. This study focuses on how Sentinel-2 image data can be implemented in traditional classification, to obtain new hints on fluvial dynamics of Po River, in northern Italy. A traditional classification of the River has been performed, based on orthophotos with centimetric resolution, collected in 2020. The classification for Po River has been done following the guidelines of IDRAIM handbook, and calculating three fundamental indexes: anabranching, braiding and sinuosity, for the distinction of morphological characteristics. Secondly the normalized drainage area (W*) and slope has been calculated. The subdivision of Po River into fundamental units called reaches, and the identification of the most representative of them has been performed. For each reach identified, Sentinel-2 images, with a maximum resolution of 10 meters, has been downloaded for the period between January 2018 to December 2022. On those, a fuzzy classification has been performed, assigning to each pixel a percentage value of membership to one of the three semantic classes analyzed, which are: water, sediment and vegetation. The results of traditional classification have been interesting, showing distinctive features of each river morphology defined. W* has shown to be able to distinguish among different morphology even if it was not used in their definition, proving its value as river attribute and its functional links with other river morphological features. Transitional and anabranching patterns are easily distinguishable from straight, sinuous and meandering, due to their high sediment content (considering the same drained area). On the other hand sinuosity, braiding and anabranching indexes are able to distinguish among all the morphologies. From Sentinel-2 classification more specific trend emerged, that highlight a general decrease of water throughout the 5 years analyzed, and more specific seasonal trends, pretty evident for vegetation. Moreover has been possible to see the vegetation recover time after a flood event verified in December 2019. Riparian vegetation restore the habitat at the end of summer season, with a strong grew around August-September. For this classification the comparison between water and sediment is very representative of the morphology. Anabranching and transitional reaches are characterized by high amount of sediment and less water, that corresponds with a higher distribution of exposed sediment considered as a morphological unit. Therefore the process that leads to such a distribution is different from the one that creates a single channel, with few exposed sediment, characteristic of straight, meandering and sinuous patterns. A further consideration can be linked with the anthropization of the area, fostering riverbed mining in the past, and channel confinement nowadays, creates an almost sharp distinction between single-channel and multiple-channels morphologies. The results of the two classifications helped for a deeper understanding of the system. On one hand traditional classifications provide accurate information on the morphology, allowing to have centimetric resolutions. On the other hand satellite-image classification helps with the understanding of the processes that occur in a river system, thanks to their temporal resolution. Both classifications provide valuable information, that allow a deep understanding of the system, and may help with the creation of models that will try to predict future trends.