Controls on channel morphologic reach type in paraglacial regions—a study based on a Nordic channel classification in Sweden

River management policies commonly rely on geomorphic classifications to understand the form and function of rivers. The geomorphic reach type can be an important component of river ecological status reports and informs management decisions. However, geomorphic classification at the national level is difficult as field-based geomorphic assessment across large regions is costly and time consuming; thus, we need to find predictors based on remote data to classify geomorphic reach types on large spatial scales. A shortcoming in many geomorphic classification schemes is that they are designed for classification of alluvial rivers. Hence, these classifications may not accurately reflect rivers with a semi-alluvial or non-fluvial history, e.g. glacial legacy. Our objective was to determine which remotely-derived parameters predict channel type in paraglacial regions with past continental glaciation and abundant till deposits. We visually field-classified the channel type of 366 reaches (with drainage areas > 10 km²) in seven catchments across a latitudinal gradient of ca.1,500 km in Sweden. Prior to field work, reaches were stratified to obtain a representative sample of drainage areas, channel slope, valley confinement, and surficial geology type. Here we present results from the two northernmost catchments, the Torne and Vindel River catchments. Reaches were classified using the existing Swedish Agency for Marine and Water Management (SwAM) scheme in addition to a novel Nordic hydromorphological classification that takes into account semi-alluvial channels common in Fennoscandia. Using nationally available data — 50- and 2-m digital elevation (DEM) models and surficial geology and lithology maps — we characterized over 18 parameters for each reach, including DEM-based metrics of channel slope, surface roughness, and valley confinement, and surficial geology at various distances from the reach. We used a Random Forest algorithm to elucidate which parameters best predict reach channel type. Preliminary results show that surficial geology type is more important than most DEM-derived variables for several channel types; however, surficial geology maps are quite coarse and may not reflect the field conditions. Several DEM-derived variables were also important predictors of channel type, including channel slope and surface roughness. These results indicate that the glacial legacy, as reflected by the surficial geology, plays an important role in determining channel morphology and thus channel type. This work also supports the need for a classification system that considers the unique post-glacial legacy of Nordic rivers. Not only does this help our understanding of the physical and biological processes in these rivers, it is a tool for practitioners to understand and implement ecological measures such as restoration projects in dynamic systems.