An automated method to extract fluvial terraces from LIDAR based high resolution Digital Elevation Models: The Oiartzun valley, a case study in the Cantabrian Margin

Abstract

Combined climatic and tectonic conditions controlled the Quaternary incision of rivers, and the fluvial terraces are the most significant landforms to untangle their evolutionary history. Increasing availability of high-resolution topography data over large areas offers an opportunity to increment greatly the quantity and the quality of information feasible for the development of semi-automated techniques for identification and characterization of these features. In this study, we apply a methodology based on Demoulin et al. (2007) for the detection of fluvial terraces in Digital Elevation Models (DEM), with the novelty of using DEMs derived from high resolution LiDAR data that are processed using Geographical Information System (GIS) routines. The method is based on the recognition of slope variations and change of sign of profile curvature across the hillslopes. Considering scatter plots over valley segments of given length, every minimum of slope value as a function of either distance to or elevation above the base of the valley is considered a possible terrace remnant. The compilation of these slope minima retained within the successive segments covering the whole length of the valley will finally provide a two-dimensional plot of terrace remnants in respect to the present-day long profile of the river. We successfully tested this method in the Oiartzun valley in the eastern area of the Cantabrian margin (western Pyrenees), where terrace maps were previously developed via aerial photo-interpretation and field-mapping campaigns. Despite the highly erosive nature of the Cantabrian fluvial systems, the poor preservation of fluvial landforms and the abundant vegetation cover, our method obtained good results due to its objectivity, exhaustiveness, and rapidity, allowing fast and coherent analysis of fluvial valleys and their comparison.