Review and assessment of remote mapping methods of fluvial terraces: Whitewater River, Minnesota.

Landform mapping is a fundamental first step in understanding the spatial and topographic context of landforms within a landscape. For the majority of the history of geomorphic inquiry, these features were mapped primarily "on the ground", via surveying techniques and interpretation. Advancements in GPS/GNSS technology and surveying equipment have improved the efficiency of field mapping, but it is still a time-intensive process. Recent improvements in remote data accuracy and availability have significantly reduced the time necessary to map landforms. In addition, geospatial software and programming have produced several automated mapping techniques. Despite recent technological innovations, understanding uncertainty, efficiency, and accuracy in these methodological approaches is not well constrained.

Fluvial terraces are morphostratigraphic markers that aid in creating river paleo-longitudinal profiles and mark locations for geochronologic dating that are crucial to reconstructing landscape evolution. They are mapped by identifying relatively low-relief terrace treads that mark the former elevation and slope of paleo-floodplains. Here we review and conduct a preliminary assessment of remote methods (automated feature extraction and manual digitizing), focusing on the Whitewater River, Minnesota USA. The Whitewater River, a tributary to the upper Mississippi River (UMR), resides in an incised valley that contains terraces that formed following post-glacial incision of the UMR. We manually digitize these terraces using 1m DEM hillshade and slope maps, along with aerial imagery and soil maps. The terraces identified through remote manual digitizing are compared to terraces extracted using LSDTopoTools. We compare and contrast efficiency, accuracy, and uncertainties in these approaches. Preliminary results reveal that automated methods greatly reduce the time necessary to identify terraces. A manual approach is necessary to refine extracted terrace features based on the objectives of a particular study and parameters used to identify terraces in a particular landscape. We suggest a combination approach as an effective option to optimize efficiency and accuracy in mapping fluvial terraces.