Seasonal shift of post-fire sediment provenance from channels to hillslopes documented through high-resolution monitoring

Wildfire is a highly effective short-term disturbance agent in semi-arid mountainous settings, where runoff and associated sediment yields may increase by orders of magnitude during the first few post-fire storm events. This brief period represents a critical time for downstream stakeholders, when understanding the threat from sediment-laden flows to receiving water resources and infrastructure during intense rainfall is crucial. Despite its importance in hazard characterization, the relative importance of hillslope and channel processes in the production of sediment-laden flows and the evolution of these sediment pools prior to recovery is still poorly understood. In this study, we seek to understand how storm runoff magnitude and sequence affects sediment availability and transfer from hillslopes to channelized domains where material may be stored or exported downstream.

In order to achieve this, we monitored surface changes, precipitation, soil moisture, and runoff over the course of the first wet season following wildfire in a steep (>32 degrees) ~1 ha watershed burned by the 2018 Holy Fire in Southern California. Change detection was performed using terrestrial lidar scans to capture changes on the hillslopes and fringes of channelized features along with UAV-based structure-from-motion photogrammetry. The first effective events occurred four months following the burn and generated extensive erosion with sediment yields equivalent to ~300-400 tons/ha, with approximately half of this sediment derived from channelized erosion (>10cm depth). The next change detection sequences bracketed storms with more sustained rainfall intensities that yielded greater runoff at the hillslope plot scale with commensurate increases in shallower hillslope sediment yields relative to channels. These results indicate that amplified runoff from earlier, lower magnitude storms was sufficient to evacuate channels, while hillslopes persisted as a sediment source for sediment-laden flows later in the post-fire wet season. Predictive models of post-fire sediment export would be improved by more explicitly incorporating the role of storm sequences in controlling sediment availability in channels and on hillslopes.