The Northern Eurasian Riverbank Migration multi-tool dataset (NERM)

Abstract. Riverbank erosion monitoring and modeling have a long-standing tradition in Earth system science. The current body of research primarily relies on observations at the basin and site levels. We endeavored to compile a comprehensive dataset of riverbank migration observations using a variety of measurement techniques, with both field-based and remote sensing data. The dataset comprises information from 12 extensive river basins situated in Northern Eurasia, encompassing rivers that drain into the Baltic Sea, the Arctic and Pacific oceans, and the Caspian Sea, specifically the catchments of the Vistula, Volga, Ural, Ob, Nadym, Yenisey, Lena, Indigirka, Yana, Kolyma, Amur, and Kamchatka rivers. The rivers included in the dataset vary in terms of environmental conditions and have average discharges of between 0.3 and 19 700 m3 s−1. This study examined approximately 140 000 km of rivers in Northern Eurasia, covering small, medium, and large rivers, with data from up to 70 years of water classifications obtained from satellite images, including those from Landsat and Keyhole, across 626 772 river channel segments. The dataset collected average and maximum bank retreat rates (m yr−1), average areas of bank retreat (m2 yr−1), and gross bank-erosion sediment yield (t yr−1). It also recorded possible drivers, encompassing both hydrological and catchment factors like permafrost, natural land zones, and geology. Our study showed that river discharge and permafrost distribution are the primary indicators of riverbank erosion in Northern Eurasia. These data will enhance comprehension of bank erosion processes and their underlying factors, thereby facilitating the development of more accurate predictive models of river channels. The dataset is available open access via the Zenodo repository (https://doi.org/10.5281/zenodo.15965461) (Chalov et al., 2025a).