Mapping debris-flow channels in the southern Central Andes using high-resolution topographic data

Resolving Earth’s surface at the meter scale is essential for an improved understanding of topographic signatures generated by debris-flow activity in high-relief mountainous terrains. Here, we explore the applicability and potential of digital elevation models (DEMs) derived from stereo-photogrammetry for debris-flow detection in the southern Central Andes of NW Argentina. Our analysis relies on a high-resolution (3 m) DEM created from SPOT-7 tri-stereo satellite data. We carefully validated DEM quality with ~5000 differential GPS points for an area of 245 km² in the Quebrada del Toro basin within the Eastern Cordillera. 

We build upon previous work that suggests that debris flows have a distinct signature in the drainage area and slope framework: debris-flow channels exhibit a nearly constant slope (no channel curvature), while channels dominated by fluvial transport processes show a negative power-law behavior in log-log space. Drainage-area approaches in geomorphic analysis are fast and efficient tools to distinguish signatures of debris-flow and fluvial transport processes, yet they might introduce an averaging bias because upstream areas are analyzed jointly. 

For a more precise localization and assessment of debris-flow activity, we evaluate topographic signatures of individual channels. We present a new approach that relies on connected components of similar slope that can be attributed to different transport regimes. Debris-flow activity reflects particularly steep segments of medium connected-component lengths in small drainage areas. The spatial occurrence and lengths of these segments are controlled by geologic and lithologic boundary conditions and we find that the highest debris-flow activity corresponds with steep slopes in areas documented Quaternary tectonic activity and the exposure of pervasively fractured bedrock. Comparing our results to topographic signatures of the corresponding catchments in log-log space, we show that individual channel approaches allow to better detect intra-catchment variability. These are imperative for understanding erosion and sediment-transport processes in the river channel. Since high-resolution data are needed to reliably resolve debris-flow channels, our meter-scale DEMs greatly improve the localization and prediction of debris-flow activity. Thus, for evaluations of recurring hazardous debris-flow activity in extensive, remote, and sparsely vegetated mountainous landscapes, stereo-photogrammetry presents a very suitable and cost-efficient alternative to airborne lidar data.