Partitioned by Process: Measuring Post-Fire Debris-Flow and Rill Erosion with Structure from Motion Photogrammetry

Document Type


Publication Date



After wildfire, hillslope and channel erosion produce large amounts of sediment and can contribute significantly to long-term erosion rates. However, pre-erosion high-resolution topographic data (e.g. lidar) is often not available and determining specific contributions from post-fire hillslope and channel erosion is challenging. The impact of post-fire erosion on landscape evolution is demonstrated with Structure from Motion (SfM) Multi-View Stereo (MVS) photogrammetry in a 1 km2 Idaho Batholith catchment burned in the 2016 Pioneer Fire. We use SfM-MVS to quantify post-fire erosion without detailed pre-erosion topography and hillslope transects to improve estimates of rill erosion at adequate spatial scales. Widespread rilling and channel erosion produced a runoff-generated debris-flow following modest precipitation in October 2016. We implemented unmanned aerial vehicle (UAV)-based SfM-MVS to derive a 5 cm resolution digital elevation model (DEM) of the channel scoured by debris-flow. In the absence of cm-resolution pre-erosion topography, a synthetic surface was defined by the debris-flow scour's geomorphic signature and we used a DEM of Difference (DoD) to map and quantify channel erosion. We found 3467 ± 422 m3 was eroded by debris-flow scour. Rill dimensions along hillslope transects and Monte Carlo simulation show rilling eroded ~1100 m3 of sediment and define a volume uncertainty of 29%. The total eroded volume (4600 ± 740 m3) we measured in our study catchment is partitioned into 75% channel erosion and 25% rill erosion, reinforcing the importance of catchment size on erosion process-dominance. The deposit volume from the 2016 event was 5700 ± 1140 m3, indicating ~60% contribution from post-fire channel erosion. Dating of charcoal fragments preserved in stratigraphy at the catchment outlet, and reconstructions of prior deposit volumes provide a record of Holocene fire-related debris-flows at this site; results suggest that episodic wildfire-driven erosion (~6 mm/year) dominate millennial-scale erosion (~5 mm/Ka) at this site.


Correction in: Earth Surface Processes and Landforms, 15 June 2020, 45(7), 1691-1691. In the original published version, Figure 4 was incorrect and appears to be a duplicate of Figure 2. The figure caption is correctly presented but the corresponding figure was wrongly inserted. See erratum publication for details at https://doi.org/10.1002/esp.4877